1990 State/Federal natural resource damage assessment and restoration plan for the Exxon Valdez oil spill

[From the U.S. Government Printing Office, www.gpo.gov]

The 1990 State/Federal
Natural Resource Damage Assessment
and Restoration Plan
for the Exxon Valdez Oil Spill
Volume 1: Assessment and Restoration Plan
Appendices A, B, C

August 1990

Dear Reviewer:

This document describes the second year of studies being undertaken
to determine the injury to natural resources resulting from the
Exxon Valdez oil spill. These studies are being conducted by the
State of Alaska and the United States to assess related damages and
develop restoration plans.

The 1990 plan has benefitted greatly from the many thoughtful
public comments on the "State/Federal Natural Resources Damage
Assessment Plan for the Exxon Valdez Oil Spill, August 1989. 11 The
current plan was assembled through the-cooperative efforts of the
State of Alaska acting through the Department of Fish and Game and
the United States acting through the Federal Departments of
Agriculture and the Interior, the National Oceanic and Atmospheric
Administration, and the U.S. Environmental Protection Agency.

Public comment on this document will assist the Trustee Council in
developing future injury assessment and restoration efforts.

Questions concerning the plan and its distribution should be
directed to U.S. Department of Agriculture, Forest Service Public
Affairs Office (907) 586-8806.

Comments should be received by October 15, 1990, at the following
address:

Trustee Council
P. 0. Box 20792
Juneau, AK 99802

We appreciate your interest and look forward to your participation
in this important process.

Sincerely,

Walter Stieglitz
Commissioner Director
Alaska Department of Fish and Game Alaska Region
Fish and Wildlife Service
Department of the Interior

n P'
teve JSoyer
Regional Forester Director
Alaska Region Alaska Region
Forest Service National Marine Fisheries
Department of Agriculture Service
r
teve

THE 1990 STATE/FEDERAL NATURAL RESOURCE
DAMAGE ASSESSMENT AND RESTORATION PLAN
FOR THE EXXON VALDEZ OIL SPILL

TABLE-OF CONTENTS

INTRODUCTION .................................................. 1

PART I

Injury Determination/Quantification,

Coastal Habitat Injury Assessment ......................... 10
Air/Water Injury Assessment ............................... 21
Fish/Shellfish Injury Assessment .......................... 52
Marine Mammal Assessment .................................. 199
Terrestrial Mammal Injury Assessment ..................... 245
Bird Injury Assessment ................................... 272
Historic Properties and Archaeological Resources .......... 308
Technical services ....................................... 311

PART II

Economics ...................................................... 318

PART III

Restoration Planning ......................................... 333

PART IV

Budget ........................................................ 354

APPENDICES

A. Quality Assurance/Quality Control
B. Histopathology Procedures
C. Glossary of Terms and Acronyms
D. Response to Comments on 1989 Plan
(Bound Separately)

INTRODUCTION

The March 24, 1989, grounding of the tanker Exxon Valdez in
Alaska's Prince William Sound caused the largest oil spill in U.S.
history. Approximately 11 million gallons of North Slope crude oil
moved through the southwestern portion of the Sound and along the
coast of the western Gulf of Alaska (see map, Fig. 1). The spill
resulted in injury to fish, birds and mammals and a variety of
other forms of marine life and habitats.

This plan describes the second year of the process by which damages
will be assessed so that funds to restore impacted resources or the
services the resources provided, can be sought from those
responsible for the Exxon Valdez oil spill (EVOS). The State of
Alaska acting through the Alaska Department of Fish and Game
(ADF&G) and the United States acting through the federal
Departments of Agriculture (DOA), Commerce (DOC), through the
National oceanic and Atmospheric Administration (NOAA), and
Interior (DOI) are acting together as Natural Resource Trustees as
provided by the Comprehensive Environmental Response, Compensation,
and Liability Act (CERCLA), and the Clean Water Act (CWA), and
other state and federal authorities. The Environmental Protection
Agency (EPA) is assisting in damage assessment and is coordinating
the federal restoration efforts with those of the State of Alaska.

The 1990 damage assessment studies plan builds on the 1989 damage
assessment studies. These studies-are designed to determine the
nature and extent of the injuries, loss or destruction to resources
and will lead to a determination of damages. The assessment of
damages for injury to natural resources requires consideration of
(1) the nature of the resources at risk, (2) the nature of the oil
in the aquatic environment, (3) the exposure of the resources to
the oil, and (4) oil-related damages to important resources. The
data provides a base for developing a restoration plan.

The purpose of determining damages--the estimated monetary value of
the injured resources and the cost to restore those resources and
the services they provided--is to pursue a claim against parties
responsible for the spill. Funds received as the result of the
claim will be used to restore, replace or acquire the equivalent of
the injured natural resources and services and to reimburse
agencies for relevant costs incurred. The U.S. Department of
Justice and Alaska Department of Law represent the federal and
state governments, respectively, in pursuits of claims.

In 1989 the Trustees developed a damage assessment plan
incorporating 63 studies in ten categories. The Trustee Council
monitored the assessment process to ensure that study objectives
were met.

In order to identify studies that should be continued, terminated

or new studies that should be initiated, the Trustee Council
considered the extensive public comments on the initial plan, and
consulted damage assessment investigators, other agency scientific
staff, legal counsel, , and independent outside expert reviewers.
The studies were evaluated from five perspectives: (1) immediate
injury, (2) long-term alteration of populations, (3) sublethal
effects, (4) ecosystem-wide effects and (5) habitat degradation.
As a result of the review, 47 of the studies were continued, 26 of
the studies were discontinued or merged into other studies, and 4
new studies were initiated (Table 1). Many of the continuing
studies were modified.

The studies described in this plan fall into ten categories: (1)
Coastal Habitat, (2) Air/Water, (3) Fish/Shellfish, (4) Marine
Mammals, (5) Terrestrial Mammals, (6) Birds, (7) Technical Services
(including chemistry, histopathology, and an integrated geographic
information system, complete with mapping) to support the resource
studies, (8) Restoration, (9) Historic Properties and Archeological
Resources, and (10) Economic Studies. The cost f or the studies f or
the 1990 oil spill year (March 1, 1990 February 28, 1991) is
approximately $37 million.

The Coastal Habitat study measures spill-related changes in the
supratidal, intertidal, and. shallow subtidal zones. it is
designed to document injury to resources that rely on these
habitats, and to assess damages for the loss of services provided
by these habitats.

The Air/Water studies determine the distribution and composition of
petroleum hydrocarbons or their environmental conversion products
in water,, sediments, and living resources. Information gathered on
the distribution and nature of the hydrocarbons and their
conversion products provides a basis for documenting exposure and
for determining injury to resources. The combined results of the
Coastal Habitat and Air/Water studies also form a basis for
estimating rates of recovery of natural resources and the potential
for accelerating recovery.

The Fish/Shellf ish studies focus on identifying potential injury to
their various life stages in areas affected by the oil spill.
Species were selected for study based on their respective niche or
overall importance within the ecosystem, ability to be sampled, and
the existence of an historic data base.

Marine mammal studies include direct observations of injury (e.g.,
through carcass counts) as well as estimates of population effects
based on pathologic and toxicologic indicators (as is being
undertaken with otters and seals). In addition, the direct
observational data allows for inferences to be made about injuries
to populations.

Terrestrial mammals near the coast may have been exposed to

hydrocarbons by breathing fumes and eating oiled carcasses or
vegetation. The studies will determine the presence of hydrocarbons
in tissues of dead animals, and the ef f ects, if any, of oil
exposure on local populations of brown bears, Sitka black-tailed
deer and river otters. Studies of reproduction in laboratory mink
are also being conducted to serve as a model for assessing injury
to other potentially affected species.

The plan for determining injury to birds is organized into four
units: (1) surveys and censuses, (2) raptors, (3) sea birds, and
(4) waterfowl, shorebirds, and passerines. The information
obtained will contribute to an understanding of mortality,
population changes, and other factors essential for the damage
assessment process. Studies proposed for birds focus on improving
the accuracy of mortality estimates and collecting data on survival
and reproductive success in relation to exposure to hydrocarbons
and conversion products. These and other data will be gathered on
birds potentially most affected by the spill or best serving as
indicators for impacts on other important ecological components.

The technical services category includes activities which provide
process support or information services to all studies in the areas
of analytical chemistry, quality assurance/quality control for the
damage assessment process, histopathology, and an integrated
geographic information system, complete with mapping.

The restoration plan describes the strategy and scope of the
restoration process and feasibility studies planned for the second
oil spill year. Restoration measures will be implemented as soon
as it becomes ecologically feasible, appropriate methods are
identified, and funds are available.

Studies on historic properties and archaeological resources will
proceed in two steps: (1) inventory, description, and
classification; and (2) qualitative and quantitative descriptions
and measurements of changes detrimental to the archeological
resources related to the spill.

The value of lost or injured natural resources, and the goods and
services they provide humans, are based on results from economic
studies. In this regard, damages forming the basis of the
Trustees' claim against the potentially responsible parties are
calculated by considering (1) the reduction of these goods and
services, including intrinsic values, resulting from the spill, and
(2) the cost of restoring these goods and services to their
pre-spill level, replacing them or acquiring their equivalent.

The Trustees emphasized in a March 1990 letter to the Exxon
Corporation and Exxon Shipping, Inc. their desire to place all
state and federal injury assessment data in a public repository,
providing that the two corporations do likewise. In a letter dated
June 6, 1990 to the Trustees, Exxon proposed a more limited data

sharing arrangement. On July 19, 1990, the Trustees and Exxon
representatives discussed the creation of a public data repository.
The parties agreed to establish a technical committee to seek
agreement on an exchange of detailed study plans and to develop a
plan based on an initial set of damage assessment studies that
could be used as an example of how data would be placed in the
repository.

TABLE ONE: STUDIES AUTHORIZED IN 1989 AND 1990

Study
Category Number Title 1989 1990

Coastal CH1 comprehensive X X
Habitat Assessment

Air/Water AW1 Geographical Extent X
in Water

AW2 Injury to Subtidal X X
Sediments

AW3 Hydrocarbons in X X
Water

AW4 Injury to Deep X 1/
Water

AW5 Injury to Air X

AW6 Oil Toxicity X

Fish/Shellfish FS1 Salmon Spawning X X
Area Injury

FS2 Egg and Preemergent X X
Fry Sampling

FS3 Coded-Wire Tagging X X

FS4 Early Marine Salmon X X
Injury

FS5 Dolly Varden Injury X X

FS6 Sport Fishery X
Harvest & Effort

FS7 Salmon Spawning Area X X
Injury, outside PWS

FS8 Egg & Preemergent Fry X X
Sampling, Outside PWS

FS9 Early Marine Salmon X
Injury, Outside PWS

IMY

.... .......... ..........
ALASKA

ANCHORAC

.. ................. ..........

Ml="

..........

.. ........
IRRGIONAL

A4r4P4:P0.RP$,

f
Mg
P,
-0-
...................
.......... .. .
.................

..............-......... ....
. ..........
f0l- . ....... .............
K

. ..................
olu
VALDEZ OIL
EXXON

--MbLIFE
:REFUCE 0
KODIAK
M Cumulative extent of ADEC
aerial oil spill observations
WILDUM:
E] National Parks, Forests,
WOUF
and Wildlife Refuges
AEFUCE:

...............
Alaska State Parks, Game
Refuges, and Critical 'Habitats
Other uplands
@4
Scale 1 inch 54.5 miles

Study
Category Number Title 1989 1990

FS10 Dolly Varden & Sockeye x
Injury, Lower Cook Inlet

FS11 Herring Injury x x

FS12 Herring Injury, x
Outside PWS

FS13 Clam Injury x x

FS14 Crab Injury x

FS15 Spot Shrimp Injury x x

FS16 Injury to oysters x

FS17 Rockfish Injury x x

FS18 Trawl Assessment x x

FS19 Larvae Fish Injury x

FS20 Underwater Observations x

FS21 Clam Injury, Outside PWS x 2/

FS22 Crab Injury, outside PWS x x

FS23 Rockfish Injury, x 3/
Outside PWS

FS24 Trawl Assessment, x x
Outside PWS

FS25 Scallop Mariculture x
Injury

FS26 Sea Urchin Injury x

FS27 Sockeye Over-Escapement x

FS28 Run Reconstruction x

FS29 Life History Modeling 4/

FS30 Salmon Database Mgmt x

Study
Category Number Title 1989 1990

Marine MM1 Humpback Whale x x
Mammals
MM2 Killer Whale x x

MM3 Cetacean Necropsy x

MM4 Sea Lion x x

MM5 Harbor Seal x x

MM6 Sea Otter Impact x x

MM7 Sea Otter Rehabilitation x x

Terrestrial TM1 Injury to Sitka Black- x x
Mammals Tail Deer

TM2 Injury to Black Bear x x

TM3 Injury to River Otter x x
and Mink

TM4 Injury to Brown Bear x x

TM5 Injury to Small Mammals x

TM6 Reproduction of Mink x x

Birds B1 Beached Bird Survey x x

B2 Censuses & Seasonal x x
Distribution

B3 Seabird Colony Surveys x x

B4 Bald Eagles x x

B5 Peale's Peregrine Falcons X x

B6 Marbled Murrelets x

B7 Storm Petrels x

B8 Black-legged Kittiwakes x

Study
Category Number Title 1989 1990

B9 Pigeon Guillemots X

B10 Glaucous-winged Gulls X

B11 Sea Ducks X X

B12 Shorebirds X

B13 Passerines X X

B14 Exposure to North X
Slope Oil

Technical TS1 Hydrocarbon Analysis X X
services
TS2 Histopathology X X

TS3 Mapping X X

Archeology ARCH1 Archeological Resources 5/ X

Restoration RP1 Restoration Planning X X

Economics ECON1 commercial Fisheries X X
Losses

ECON2 Fishing Industry Costs X 6/

ECON3 Bioeconomic Models X 6/

ECON4 Public Land Effects X X

ECON5 Recreation Damages X X

ECON6 Subsistence Losses X X

ECON7 Intrinsic Values X X

ECON8 Research Program Effects X X

ECON9 Archeological Damage X X
Quantification

1/ AW4 Combined with AW2
2/ FS21 Combined with FS 13
3/ FS23 Combined with FS17
4/ FS29 Combined with FS28
5/ Part of Econ. 9
6/ Combined with Econ. I

PART I

INJURY DETERMINATION/QUANTIFICATION

COASTAL HABITAT STUDY NUMBER I

Study Title: Comprehensive Assessment of Injury to Coastal Habitats

Lead Agency: USFS

Cooperating Agencies: NOAA, DEC, NPS, FWS, ADF&G, DNR

INTRODUCTION

The purpose of the Coastal Habitat Injury Assessment is to document
and quantify injuries to biological resources found in the shallow
subtidal, intertidal, and supratidal zones throughout the shoreline
areas affected by EVOS.

Study sites are selected and ground-truthed during Phase I. Phase
II is an intensive evaluation of the study sites to determine the
extent of injury to natural resources. The objective of this study
is to estimate the effects of various degrees of oiling on the
quantity (abundance and biomass), quality (reproductive condition
and growth rate), and composition (diversity and proportion of
population) of key species in the critical trophic levels of
coastal communities. These data are expected to provide evidence
of injury to the overall health and productivity of these critical
coastal habitats, and provide information necessary to the more
species-specific studies on the effects of the oil spill on
affected mammals, birds and fish that use these habitats.

PHASE I

This study uses a stratified random sample design to select basic
experimental units called study sites. Oiled study sites were
selected from shorelines which were affected by the oil spill and
control study sites were selected from shorelines which were not
oiled. The shoreline was subdivided into ten strata; five habitats
multiplied by two oiling types. The study sites are grouped by
strata within three geographic regions: Prince William Sound (PWS) ,
Cook Inlet/Kenai Peninsula (CIK) , and Kodiak Archipelago/Alaska
Peninsula (KAP). Consolidating a wide range of habitat and oiling
characteristics into ten strata, coupled with the relatively coarse
resolution of the available habitat and oiling data base, resulted
in variances between ground-truthed classifications and mapped
classifications. These variances were addressed by confining the
additional, inductively selected sites in 1990 from the top 40
randomly ranked sites in PWS and the top 50 ranked sites in CIK and
KAP, respectively. These sites now represent a simple random
sample of the shoreline and will continue to be used to make
inductive inferences to the universe of all possible sites. There
are a total of 102 sites to be studied in 1990.

Sites in the very lightly and lightly oiled strata are not included

for additional site selection or comprehensive sampling in 1990.
This will enable the assessment,to focus on moderately and heavily
oiled sites and their respective controls.

Approximately 40 additional study sites will be deductively (non-
randomly) selected in 1990 to provide additional spatial and
habitat coverage in strata where a full array of inductive study
sites could not be obtained in 1989. Approximately 18 of the
additional sites will be new control sites selected to match the
physical and biological characteristics of existing inductively
selected oiled sites. The site selection process for control sites
in 1990 will be confined to those shorelines which match, as
closely as possible, the biological and physical attributes of
their respective oiled sites.

In the CIK and KAP regions, additional sites (approximately 8) will
be selected, where appropriate, from sites that were occupied by
the National Park Service in 1989. The remaining sites will be
selected using the original data base, as supplemented with
September oiling classifications, aerial photos, and additional
ground-truthing.

OBJECTIVES

1. To maintain a statistically valid study site selection
strategy and identify additional study sites using existing
map-based coastal habitat and oil impact classification
schemes.

.2. To ground-truth potential study sites to evaluate map-based
habitat and oil impact classifications.

3. To describe and mark approximately 45 study sites in addition
to the 57 sites that have been identified for comprehensive
sampling in 1990.

METHODS

Four sites, representative of habitat types appropriate to each
region in the original stratified random sample for the 1989 data
base and ranked moderately and heavily oiled, are selected for
study in 1990. Each site will be matched with a control site
deductively chosen to approximate closely the physical and
biological attributes of the oiled site. The selection of control
sites will begin with the existing pool of randomly selected non-
oiled sites that were identified and surveyed in 1989.
Photographs, maps, and geomorphology and biological descriptions of
the sites will be used as the bases for selection.

if a suitable control cannot be found from the existing pool of
1989 sites, candidate control sites will be identified by

consulting the original data base and oiling classifications and by
locating equivalent habitat types in closest proximity to the oiled
sites. Approximately 5 candidate control sites will be identified
for each oiled site. Sites will be visited in order of their
distance from the oiled site until a suitable match is found.
Sites will also be examined for any physical characteristics unique
to the oiled site that have not been accounted for in the five
habitat types. In PWS, control sites must be located on islands to
match oiled sites that are located on islands.

Supplemental oiled sites and matched controls will be sought and
added if there are fewer than three replicates of a particular
habitat type in the existing sample within each region (excluding
f ine-textured habitats in PWS) . Supplemental sites will be
deductively selected to include intertidal study sites occupied by
the NPS in 1989, "set-aside" (untreated) sites in PWS, and areas
which provide additional spatial distribution of a particular
stratum in a region. Data from supplemental sites will not be used
to expand inductively estimates of injuries to the universe of
potential sites. Rather, it will be used to estimate the effects
of oiling and treatment on a given habitat type.

The methods for locating and marking additional study sites will
follow the methods used in the 1989 reconnaissance survey.

. BIBLIOGRAPHY

Alaska Department of Environmental Conservation. 1989. Oil spill
impact maps. Unpublished preliminary data being developed
under Air/Water Study Number 1.

Cochran, William G. 1977. Sampling techniques. Third Edition.
John Wiley and Sons. New York. Chapter 11, pp. 292-324.

Environmental Systems Research Institute (ESRI). ARCINFO
geographic systems software. Version 5.0. Redland,
California.

Gundlach, E.R. and M.O. Hayes. 1989. Vulnerability of coastal
environments to oil pollution. 12 Marine Technology Society
Journal pp. 18-27.

Hayes, M.O. and C.H. Ruby. 1979. Oil spill vulnerability index
maps, Kodiak Archipelago. Unpublished maps. 47 leaves.

Hayes, M.O., E.R. Gundlach, and C.D. Getter. 1980. Sensitivity
ranking of energy port shorelines. Proceedings of a specialty
conference on ports. American Society of Civil Engineers.
New York. Pp. 697-709.

Research Planning Institute (RPI), Inc. 1983a. Sensitivity of
coastal environments and wildlife to spilled oil, Prince
William Sound, Alaska, and atlas of coastal resources.
Prepared f or National Oceanic and Atmospheric Administration.
Office of Oceanography and Marine Services, Seattle,
Washington. 48 leaves.

Research Planning Institute (RPI), Inc. 1983b. Sensitivity of
coastal environments and wildlife to spilled oil, Shelikof
Strait Region, Alaska, and atlas of coastal resources.
Prepared for National Oceanic and Atmospheric Administration,
Office of Oceanography and Marine Services, Seattle,
Washington. 431eaves.

Research Planning Institute (RPI), Inc. 1985. Sensitivity of
coastal environments and wildlife to spilled oil, Cook
Inlet/Kenai Peninsula, Alaska, and atlas of coastal resources.
Prepared for National Oceanic and Atmospheric Administration,
Office of Oceanography and Marine Assessment, Seattle,
Washington. 64 leaves.

Research Planning Institute (RPI), Inc. 1986. Sensitivity of
coastal environments and wildlife to spilled oil, Southern
Alaska Peninsula, an atlas of coastal resources. Prepared for
National Oceanic and Atmospheric Administration, National
Ocean Service, Alaska Office and U.S. Department of the
Interior, Minerals Management Service, Alaska OCS Region. 69
leaves.

BUDGET: ADF&G

salaries $ 14.7
Travel 3.0
Contracts 131.5
Equipment & supplies 7.5

Total $ 156.7

PHASE II

PART A: Injury Determination

Coastal habitats are unique areas of high productivity supporting
a diverse array of organisms, including many commercially and
ecologically important species. These habitats are particularly
vulnerable to oil spill impacts because of the grounding of oil in
the intertidal zone, the persistence of oil in intertidal and
subtidal sediments, and the effects of associated clean-up
activities.

Oil may affect coastal organisms directly by coating or ingestion,
with toxic effects leading to death or reproductive failure.
Indirectly, oiling may cause decreased productivity, accumulation
of toxic effects through the food chain, and loss of microhabitat
such as algae beds. Assessment of injuries to coastal habitat
resources and determination of rates of recovery require
consideration of the various coastal geomorphologic types, the
degree of oiling, the affected habitat, and their trophic
interactions. coastal habitats consist of three interactive zones
(supra-, inter-, and subtidal) . Animals may use multiple zones,
necessitating a coordinated study of the effects of oiling over the
entire habitat. The complexity of this system requires expertise
in many disciplines. Therefore, an interdisciplinary team with the
appropriate expertise, including plant and systems ecology, marine
biology, and statistical analysis, has been established.

Initial field studies were completed by November 1, 1989.
Processing of samples and data analysis is being conducted to
determine the variance and magnitude of changes between non-oiled
and moderately and heavily oiled sites.

OBJECTIVES

A. Estimate the quantity (abundance and dry weight biomass),
quality (reproductive condition and growth rate), and
composition (diversity and proportion of standing crop) of
critical trophic levels (and subsequent impact on trophic
interactions) in moderately and heavily oiled sites relative
to non-oiled sites.

B. Estimate hydrocarbon concentrations in sediments and soils.

C. Establish the response of these parameters to varying degrees
of oiling and subsequent clean-up procedures.

D. Extrapolate impact results to the entire spill-affected area.

E. Estimate the rate of recovery of the habitats studied and
their potential for restoration.

F. Provide linkages to other studies by demonstrating the
relationships between oil, trophic level impacts, and higher
organisms.

METHODS

Vertical transects will be established at each of the study sites
selected in Phase I. Work will be conducted along these transects
in the supratidal and intertidal zones. For this study, the
intertidal extends from the 11011 tide mark to Mean High High Water
(MHHW), and the supratidal is from MHHW or where terrestrial
vegetation begins (if below MHHW) to the highest extent of possible
oil occurrence. The intertidal transects will be extended into the
supratidal zone at locations in the KAP where coastal plant
communities occur. Primarily, this will be in fine-textured,
coarse-textured, and sheltered estuarine habitats. Beach sediment
texture will be determined as part of Phase I. Community
composition, cover, and standing crop by trophic level will be
estimated. Key species (dominant producers and food sources) will
be determined and studied according to the methods listed below, to
estimate the quantity, quality, and composition at each trophic
level, and to collect samples for determination of hydrocarbon
contamination. Sediment samples will be collected by DEC for
analyses of hydrocarbon composition and changes in concentration
over time. Using a geographic information approach, the impact (by
habitat type and degree of oiling) over the entire area affected by
the oil spill will be integrated and field-verified.

In 1990, sampling in the supratidal zone will occur in the KAP
where there is extensive vegetation and sufficient wave exposure to
move oil above the tide line.

Subtidal sites will be selected independent of the supra- and
intertidal sites. Subtidal sites will include three physio-
geographic types: bays, points and runs (straight lines). The
physio-geographic areas will be further divided into three habitat
types: Nereocystis beds, Zostera beds, and Laminaria beds. These
will be further divided into strata selected according to
ecological importance, potential impact, and extent of habitat
within the oiled region.

Specific methods for each component of the study were developed as
follows:

Coastal
1. Initial Site Survey
2. Locating Transects
3. Sample Identification and Chain of Custody

Supratidal
1. Quadrant Location
2. Determination of Plant Productivity

a. Analysis of Vegetation Nutrient
Content
b. Analysis of In Vitro Digestibility
3. Analysis of Soil/Sediment Microbial Activity
4. Sampling of Soils and Sediments for
Hydrocarbon Concentration

Intertidal

Invertebrates
1. Locating 1 Quadrants
2. Swath Surveys
3. Reproductive Condition
4. Growth and Survivorship
5. Hydrocarbon Sampling Procedures
6. Experimental Work
7. General Laboratory Sorting Procedures
8. Subsampling of Intertidal Samples
9. Processing of Histological Samples

Fish
1. Locating Transects
2. Locating Quadrants
3. Sampling Quadrants
4. Minnow Trap Sampling
5. Sample Storage and Identification
6. Fish for Hydrocarbon Analysis

Plants
1. Introduction
2. Study Plan
a. Stratified Sampling
b. Site Experiments at Selected
Habitats
C. Field Experiments

Subtidal
1. Sampling
2. Field Schedule
3. Laboratory Procedure for Benthic
Invertebrates
4. Hydrocarbon Sampling Procedures
5. Data Analysis

Analysis of samples obtained in 1989 is still underway and will
continue as additional samples are collected. Samples from 1990
will be processed as rapidly as possible after they are returned
from the field. The data from all of the component studies are
being entered into the INGRES database management system. This
system is widely used, and has good data security features. Use of
this data base system will therefore maximize both internal
integration and availability of the data to related damage

assessment projects.

BIBLIOGRAPHY

AOAC. 1980. official Methods of Analysis of the A.O.A.C., 13th ed.
Chipperfield, P.N.J. 1953. Observations on the breeding and
settlement of Mytilus edulis (L.) in British waters. J. Mar.
Biol. Ass. U.K. 32:449-476.

Johnson, R.D. and H.L. Bergman. 1984. Use of histopathology in
aquatic-toxicology: A Critique. Pp. 19-36. In Containment
Effects on Fisheries, V.W. Cairns, P.V. Hodson and J.0.
Nriagu, eds. John Wiley and Sons.

A.L. Page, R. H. Miller, and D.R. Keeney (eds.). 1982. Methods of
Soil Analysis, Part 2. Chemical and Microbiological
Properties. Second Edition. Am. Soc. Agron.

Ropes, J.W. 1968. Reproductive cycle of the surf clam, Spisula
solidissima, in offshore New Jersey. Biol. Bull. 135:349-365.

Seed, R. 1969. The ecology of Mytilus edulis L.
(Lame 11 ibranch iata) . I. Breeding and Settlement. Oecologia.
3:277-350.

Sheehan, D.C. and B.B. Hrapchak. 1980. Theory and Practice of
Histotechnology. 2nd Ed. C.V. Mosby Co.

Tietge, J.E., R.D. Johnson and H.L. Bergman. 1988. Morphometric
changes in gill secondary lamellae of brook trout (Salvelinus
fontainalis) after long-term exposure to acid and aluminum.
Can. J. Fish Aquat. Sci. 45: 1643-1648.

Tranter, D.J. 1958. Reproduction in Australian pearl oysters. II.
Pinctada albina (Lamarck): gametogenesis. Aust. J. Mar.
Freshwtr. Res. 9: 144-158.

United States Department of Agriculture Research Service. 1970.
Forage Fiber Analysis. Agric. Handbook No. 379.

Wilson, B.R. and E.P. Hodgkin. 1967. A comparative account of the
reproductive cycle of 5 species of marine mussels (Bivalvia:
Mytilidae) in the vicinity of Freemantle, W. Australia. Aust.
J. Mar. Freshwtr. Res. 18: 175-203.

PART B: PRE-SPILL AND POST-SPILL CONCENTRATIONS OF
HYDROCARBONS IN SEDIMENTS AND MUSSELS AT INTERTIDAL SITES WITHIN
PRINCE WILLIAM SOUND AND THE GULF OF ALASKA.

Damage assessment of the oil spill in PWS and the GOA requires
information on hydrocarbon contamination levels in water, sediment
and biota prior to the spill (baseline), and at various times after
the spill in order to determine the potential impact and duration
of impact. Hydrocarbon baseline information is available for
several sites in PWS prior to oil transport and for the first 4
years of oil shipment. The intertidal baseline for hydrocarbon
levels in mussels, sediment, water, and fish had been established
at 10 sites from 1977 to 1981. All sites are located on low
energy, low gradient beaches, often at the head of embayments, and
most sediments transects are associated with eel grass. All sites
have adjacent bands of mussels (Mytilus trossulus).

Because of the potential persistence of hydrocarbons in sediments
in temperate and subarctic intertidal and subtidal environments,
sampling may be continued to document depuration and recovery.
Concentrations of the full range of individual aliphatic and
aromatic hydrocarbons in sediments and mussels from intertidal
sites will be reported. Abundance of mussels and other epifauna
along sediment and mussel transects will be photographically
recorded during each sampling period. These data will provide a
basis for estimating temporal and spatial impact to other biota of
the nearshore environment.

OBJECTIVES

A. Sample and estimate hydrocarbon concentrations in mussels and
sediment from 20 sites within 10% of the actual concentration
95% of the time, when total aromatic concentrations are
greater than 200 ng/g dry wt.

B. Test the null hypothesis that hydrocarbon contamination of
sediments and mussels is the same for the pre-spill and post-
spill period.

C. Document changes in abundance and distribution of intertidal
epifauna and test the null hypothesis that no differences
occur at oiled and non-oiled sites.

METHODS

Ten intertidal sites in PWS and Port Valdez were sampled for
sediments, mussels, water, and fish annually from 1977 to 1981 to
establish a baseline against which future changes in hydrocarbon
concentrations can be compared. Sites were initially sampled in
spring, summer and fall to determine if short-term changes occurred
during the warm season. These sites were resampled in March of
1989, immediately before several of them were impacted by the EVOS.

Ten additional sites were established to cover areas in the
trajectory of the oil path. Four of these sites were on the Kenai
Peninsula (KP) and the remaining six were in PWS. Sediment and
mussel samples were taken. Photo documentation was initiated along
mussel and sediment transects at each site. These sites were re-
sampled several times during the summer of 1989 to document the
appearance of and changes in hydrocarbon contamination from the
EVOS.

Sediments: Transect lines thirty meters (m) in length are located
parallel to the water line at -0.75 m to +0.75 m (depending on
specific site). Sediment samples will be collected in triplicate
at each site by compositing 10 cores (dia 3.2 cm x depth 1.25 cm)
taken at random along a 30-meter transect for each sample.
Composite sediments will be placed in chemically clean 4-oz. jars,
*placed in an ice chest with artificial ice and transported. These
will be frozen within 2-3 hours of collection. one blank sample
will be taken at each site.

Mussels: These transects are located in mussel bands, parallel to
the water line, usually just above (+1 m tide level) the sediment
transects. Triplicate mussel samples will be collected by taking
approximately 30 2-5 cm. mussels (enough to produce >10 gms tissue)
at random along the 30-meter transect. Samples in 16 oz. jars will
be cooled, transported and frozen in the same manner as the
sediment samples.

Photo Documentation: Close-range views will be photographed of the
strata, macroflora and epifauna. Photos will be taken every 4 or
8 m along the sediment transect and every 2 or 4 m along the mussel
transect line beginning at 1 meter. Macrophyte cover as well as
epifaunal occurrence and density will be recorded from photographs
taken of 625 cm2 quadrants placed along the sediment and mussels
transect lines. A grid of 100 random dots projected on each slide
will be used to estimate the occurrence and percentage of surface
area covered by macrophytes and epifauna. Macrophytes and epifauna
will be identified to species where possible.

Data Analysis:

Random sample and subsample collection will ensure that
hydrocarbons present in the sample represent the average
concentration at each site. "Hot spots" of hydrocarbon
concentration over the 30 meter transects should be canceled out by
this procedure. Selected triplicate samples will be analyzed, the
mean concentrations and deviations from these means determined, and
appropriate statistical tests applied - either ANOVA or paired
comparisons (Tukey or Shef f e I tests). Digital tables of individual
hydrocarbons will be reported.

Macrophyte and epifauna occurrence and cover will be analyzed using
one-way ANOVA or paired comparisons (oiled versus non-oiled where

strata are similar). They will be tested at the .05 level of
significance.

BIBLIOGRAPHY

Connell, Joseph H. 1970. A predator-prey system in the marine
intertidal region. 1. Balanus qladula and several predatory
species of Thais. Ecol. Monog. 40:49-78.

Gundlach, Erich R., Paul D. Boehm I Michel Marchand, Ronald M.
Atlas, David M. Ward, and Douglas Wolfe. 1983. The fate of
Amoco Cadiz oil. Science 221:122-129.

Karinen, John F., L. Scott Ramos, Patty G. Prohaska, and William D.
MacLeod, Jr. In Preparation. Hydrocarbon distribution in the
marine environment of Port Valdez and Prince William Sound,
Alaska.

Warner, J.S. 1976. Determination of aliphatic and aromatic
hydrocarbons in marine organisms. Anal. Chem. 48:578-583.

BUDGET: USFS

PART A:

Salaries $ 56.0
Travel 14.0
Contracts 8,818.0
Equipment 65.0

Total $ 8,953.0

PART B:

Salaries $ 34.0
Travel 36.0
Contracts 82.0
Equipment & Supplies 8.0

Total $ 160.0

BUDGET SUMMARY

ADF&G $ 156.7
USFS 9,113.0

Total $ 9,269.7

AIR/WATER RESOURCES INJURY ASSESSMENT

The evaluation of injury to air, water, and sediment resources is
a critical component in assessing the overall damage to natural
resources caused by the EVOS. Three studies addressing impacts on
water column and bottom sediments will continue in the second year
of the damage assessment effort. Air resources were studied in the
first year immediately after the oil spill, but will not be
continued because of the short term impact to the air resource.

Water and Sediment Resources

Assessment of the concentrations,of petroleum hydrocarbons in the
water column of PWS and the Kenai Fjords region began almost
immediately after the EVOS. oil affected pelagic and nearshore
waters, benthic sediments, intertidal habitats and adjoining
habitats above high tide. Quantifying hydrocarbon levels in the
water column was most critical during the first few weeks following
the EVOS, when dissolution of soluble components was most rapid and
the likelihood of toxic exposure was highest. As wind and current
spread the oil and carried it farther away from the spill site,
concern shifted from immediate impacts in the water column to
longer-term effects from shoreline oiling on nearshore and subtidal
sediments, and to chronic low-level hydrocarbon contamination of
the water column.

Marine water quality is protected under state and federal water
quality standards which include classifications for such uses as
growth and propagation of fish and wildlife, aqua'culture, and human
uses such as recreation. Moreover, State of Alaska water quality
standards for petroleum hydrocarbons establish criteria for water
.habitats. I

The three water and sediment studies funded for this year are
designed to reveal the continuing extent of hydrocarbon
contamination remaining after the spill. Chronic, low-level
contamination of the water column is expected to continue through
the bleeding-off of oil from impacted shorelines. Nearshore and
offshore sediments continue to be exposed to further contamination
from the suspension and sinking of oily beach materials.
Documenting the extent of continuing contamination will assist in
demonstrating injury to the water resource along with chemical
exposure of marine mammals, birds, intertidal and shallow subtidal
communities, fisheries, and terrestrial mammals dependent on beach
habitats.

The Air/Water (A/W) studies are integrated with the Coastal Habitat
studies to provide data on injury to habitats for other studies
that address injury to biological resources. A/W studies will also
help establish the basis for restoration.

The three continuing water quality studies focus on:

1. Petroleum hydrocarbon-induced injury to subtidal marine
sediment resources and injury to benthic infauna.

2. Geographic and temporal distribution of dissolved and
particulate petroleum hydrocarbons in the water column.

3. The toxicity of weathered oil and the fate and effects of oil
transformation compounds within the marine environment.

A/W Study 1, which is not continued in 1990, documented the extent
of the surface oiling, and mapped the results for use by other
studies. Further oil fingerprinting will be conducted under
response activities, or by Technical Services Study Number 1. A
final report mapping the distribution of surface slicks in the
first year of the spill will be produced this year.

A/W Study 2 is now combined with elements of A/W Study 4 into one
integrated sediment contamination study. This study will continue
to document the presence, persistence, and chemical composition of
petroleum hydrocarbons in subtidal marine sediments. These data
will assist in quantifying injury to the sediment and will provide
the chemical linkage needed to assess biological injury. Shallow
subtidal oil concentrations will be compared with oil
concentrations in adjacent intertidal areas to better understand
the fate of oil. This study will continue to determine the degree
of injury to the benthic infaunal resource and the duration of any
documented injury. Additionally, microbial screening techniques of
subtidal sediments will be employed to determine the presence,
toxicity, and degradation rates of oil. Sediment sampling stations
extend outside PWS to include the Kenai Fjords, Katmai, Cook Inlet,
Kodiak, and the Aleutian chain.

A/W Study 3 will continue to document hydrocarbon concentrations in
the water column at a range of depths and locations. Trends in
ambient water quality will be determined using the blue mussel as
a biological indicator of low-level, chronic water quality
contamination to supplement chemical measurements. Sediment traps
will be deployed to measure sedimentation and associated
hydrocarbon inputs to subtidal sediments.

A/W Study 5 was completed.

A/W Study 6 is a new study designed to address the concerns f or
long-term contamination and toxicity of weathered oil and its
degradation products to selected test organisms, and to integrate
the results of several projects into a mass-balance budget for the
distribution, transport, transformation, and persistence of spilled
oil in Alaska coastal environments.

AIR/WATER STUDY NUMBER 2

Study Title: Petroleum Hydrocarbon-Induced Injury to Subtidal
Marine Sediment Resources

Lead Agency: NOAA, State of Alaska

INTRODUCTION

A proportion of -the oil that entered the water (either the original
crude oil derived from the spill, oil leaching from contaminated
shorelines, and/or oil dispersed into receiving waters via
shoreline remediation procedures) probably has reached, or will
reach, the bottom as a result of physical (Boehm et al. 1987) and
biological processes.

Benthic data collected in polluted waters elsewhere suggest that
changes in number and diversity of species, as well as abundance
and biomass of species, can be expected if sizable amounts of oil
settle to the bottom. These changes can have serious trophic
implications since many subtidal benthic invertebrates are
important food resources for bottom-feeding species such as
pandalid shrimps, crabs, bottomfishes and sea otters. Further, the
larvae of most benthic organisms in PWS move into the water column
(in March through June) and are utilized as food by large
zooplankters and larval and juvenile stages of pelagic f ishes,
small salmon f ry, and herring. Thus, damage to the benthic system
by hydrocarbon contamination can af f ect f eeding interactions of
important species both on the bottom and in the water column.

Continuation of this study will evaluate the extent of subtidal
hydrocarbon contamination in PWS, along the LKP, and near Kodiak
Island. The purpose of this study will be to determine to what
depth petroleum hydrocarbons have been transported over the winter
months of 1989/90, to continue the time-course of data acquisition
necessary to answer the question of persistence of petroleum
hydrocarbons in subtidal sediments, and to determine the impact of
oiling upon subtidal resources. Fewer sites will be studied this
year. However, intensity of sampling will be increased.

Three projects formerly funded under A/W Study 4 will be included
in this study. The first enumerates hydrocarbon oxidizing bacteria
and assesses the maximum potential for 'in situ' biodegradation of
selected hydrocarbons at various sites within and outside of PWS.
Coupled with data on ambient hydrocarbon concentrations, the
microbial data will allow a gross estimate of the maximum possible
rate of bacterial hydrocarbon oxidation 'in situ' to be made. The
second project will screen sediments for petroleum hydrocarbons
using ultra-violet fluorescence spectrophotometry and will assess
the toxicity of marine sediments using the luminescent marine
bacterium Photobacterium phosphoreum to test for aqueous toxicants

(Schiewe et al. 1985). The third project titled "Injury to Deep
Benthos" will examine the injury, if any, to infaunal communities
below a depth of 20m in bays adjacent to eel grass beds. The
sampling for all projects included in A/W Study #2 will be
conducted from the same vessel and time (June, July). The present
study will also coordinate closely with the subtidal project of the
Coastal Habitat Study. Sediment and microbiological samples will
be collected at the identical eelgrass sites where the Coastal
Habitat study will sample shallow subtidal benthos.

OBJECTIVES

A. Determine occurrence, persistence, and chemical composition of
petroleum hydrocarbons in subtidal marine sediments.

B. Provide marine sediment data to assist agencies in mass
balance calculations on the fate of oil in the marine
environment.

C. Relate subtidal oil concentrations to adjacent intertidal
concentrations.

D. Screen sediments for oil contamination and estimate the toxic
ef f ects of petroleum hydrocarbons using bacterial bioassays of
sediment samples collected from oiled and nonoiled habitats.

E. Enumerate hydrocarbon oxidizing bacteria and assess the
maximum potential for 'in situ' biooxidation of selected
hydrocarbon substrates in subtidal marine sediments at oiled
and nonoiled sites within and outside of PWS.

F. Determine if changes occurred in the macro-benthos by
comparing species richness, species diversity, general
abundance and biomass, and trophic composition of the benthic
biota living on similar substrata at approximately 40, 100,
and >100m below sea grass beds between oiled and unoiled bays.

G. Determine if temporal changes will occur in the macro-benthos
between oiled and unoiled bays by comparing species richness,
species diversity, general abundance and biomass, and trophic
composition of the benthic biota at specific stations.

H. If changes are detected in the infauna, examine the
relationship between the accumulation and retention of
hydrocarbons in sediments and the effect on the benthic biota.

METHODS

The methods employed by the three agencies cooperating in this

study are described separately below.

National Marine Fisheries Service:

Auke Bay Laboratory

Sediments will be sampled at 16 sites in PWS (four reference sites
and 12 contaminated sites). Sampling will be conducted during
three periods (May, June/July and September) . Six sites will be the
same as those to be sampled by the subtidal project of Coastal
Habitat Study. Outside PWS eight sites will be sampled. Six sites
will be on the Kenai Peninsula and two sites will be near Kodiak
Island. These sites will be sampled in July.

Three samples, each a composite of eight subsa*mples collected
randomly along a 30 m transect laid parallel to the shoreline will
be taken at each intertidal site. These samples will be collected
at low tide or by divers. Intertidal collections will be made at
a single tidal height in the range of +1 to -1 m relative to mean
lower low water (MLLW) depending on the distribution of fine
sediments.

Subtidal sediment collections will be made at depths of 3, 6 and 20
m below MLLW in May and September and at 3, 6, 20, 40 and 100 m in
June/July. Collections at 3, 6 and 20 m will be made by divers on
transects laid along the appropriate isobath and sampled in the
same way as described above for the intertidal transects. The
subtidal project of Coastal Habitat Study Number I will sample
sediments, infauna and epifauna in the same depth range at six of
the PWS sites. Samples taken at depths below 20 m will be
collected with a Haps corer. A Smith-McIntyre grab will be used to
sample those sediments which cannot be effectively sampled with the
Haps corer. Three cores will be taken at each depth. Four
subsamples will be removed at randomly selected points within each
core. The subsamples will be combined to form one sample per core.
The samples will be taken at the same sites as the benthos (see
deep benthos sampling methods below) , however sediments will not be
taken from the same core/grab as the benthos samples because the
volume removed for sediment hydrocarbon analysis will jeopardize
the quality of the benthos samples.

Northwest Fisheries Center

Surface sediment samples for establishing levels of petroleum
hydrocarbon residues and sediment-associated toxicity will be
collected June through July, 1990 (Table 1) . Sites will be located
in potentially oil-impacted areas and also in unimpacted areas in
PWS and LCI.

Selected sediment samples will be analyzed for petroleum
hydrocarbons and other organic contaminants. After rapid
extraction of sediments, relative aromatic hydrocarbon levels in

sediment extracts will be measured using liquid chromatography
coupled to a fluorescence detector. Sediment toxicity will be
estimated using the Microtox bioassay that utilizes the luminescent
marine bacterium Photobacterium Rhosphoreum (Schiewe et al. 1985).
The test involves exposing suspensions of the bacterium to saline
solutions of organic solvent extracts of sediment samples and
measuring the effect of exposure on the amount of light emitted
f rom the bacteria. The results of the test can be used to rank the
relative toxicity of sediment samples. The relationship between
Microtox results and contaminant levels in sediments will be used
to provide support f or the rankings of sediment toxicity by the
Microtox bioassay.

Sampling activities will be conducted at 27 sites in PWS and LCI,
including oiled and nonoiled sites (Table 1) . Samples will be
collected at water depths of 0 (intertidal), 3, 6, 20, 40, and 100
meters. At each site, sediment samples will be collected with a
box corer, Van Veen or Smith-McIntyre grab. Each of three
replicate sediment subsamples for each depth will be placed in two
20 ml scintillation vials and stored at - 200 C. The coordinates
and depths of each station will be recorded.

Three sediment replicates are composited, the excess water is
decanted and the sediment is stirred to homogenize and placed into
a tared 100-ml centrifuge tube. Sodium sulfate, methylene
chloride, and activated copper are added. Clumps are broken up
with a tef lon stirring rod, if necessary. Each tube is capped
tight enough to prevent leakage. Each mixture is placed in a
sonic bath or sonicated with a sonic probe. The sonicated samples
are centrifuged for 5 minutes at 1,500 rpms. Each extract is then
decanted into 50-ml labeled concentrator tubes. To the sediment
remaining in the centrifuge tubes, another 10 ml of methylene
chloride is added. The mixture is stirred with a teflon rod,
capped, and sonicated in bath or with probe and centrifuged for 5
minutes. The solution is decanted into the original concentrator
tube, another 10 ml methylene chloride is added, stirred, and
sonicated and centrifuged again as described previously. The third
extract is added to the first two, a boiling chip is added and the
solution concentrated to exactly 10 ml. The sample is divided into
two 5.0-ml portions, one for HPLC screening and one for Microtox
analysis.

To the sediment portion for high pressure liquid chromatography
(HPLC), polystyrene internal standard is added and about I ml of
the mixture is transferred to a vial for the autosampler (the
remainder is stored in another vial in the freezer). The
analytical procedure for the detection of aromatic hydrocarbons
(AH) in the sediment is similar to that of Krahn et al. (1988a,b),
but analytical columns, are used instead of preparatory columns.
The sediment extract (150 ul) is injected onto the HPLC columns and
isocratically eluted with methylene chloride. The internal
standard is detected with a UV detector and the aromatic compounds

with a fluorescence detector (phenanthrene wavelengths--260/380 nm
and an additional wavelength pair to be selected). To quantitate
the total Ahs, the total f luores6ence area is integrated during the
time when the fraction would be collected for the prep cleanup
(Krahn et al. 1988b) and converted to phenanthrene equivalents (the
concentration of phenanthrene that would result in an equivalent
integrated area) or to other AH equivalents, respectively (as
determined from above).

Initially, dose-response studies using oil and o i 1 -contaminated
sediments will be evaluated by the Microtox bioassay in conjunction
with the ultraviolet fluorescence (UVF) screening methods.
Subsequently, five ml of each sediment extract in methylene
chloride will be obtained as described. Samples (volumes will be
determined in the dose-response studies) will be exchanged into 1
ml of ethanol by solvent evaporation under constant heat. Microtox
assays of the organic extracts will then be conducted as described
in the Puget Sound Estuary Program protocols for sediment bioassays
(EPA 1988).

Alaska Department of Environmental Conservation:

Microbiology sediment samples from the intertidal and shallow
subtidal areas will be obtained from the shore parties and divers
collecting samples for hydrocarbon chemistry analysis.
Microbiology samples from deeper subtidal areas will be obtained
f rom the core or grab sampler at the same stations and times as
those collected for chemistry analysis.

Sediment samples for microbiological analysis will be collected in
sterile Whirlpak bags as composites in triplicate along the same
horizontal transects from which the chemistry samples are obtained.
Care will be taken to avoid contamination of samples by the
sampling personnel and cross-contamination between different
sediment samples. Sampling apparatus should be thoroughly rinsed
with water between samples and, where possible, disinfected with
alcohol or other disinfectant. Samples obtained from the deeper
water grabs will be collected from the center of the core to avoid
surface contamination incidental to sample handling.

Hydrocarbon biodegradation potential associated with sediment
microbes will be assayed by adding radiolabelled aliphatic and
aromatic substrates to sediment samples. (14C) -hexadecane, (14C) _
phenanthrene and (14C) -benzo(a]pyrene will be the three hydrocarbons
substrates used. Each substrate will be monitored for
biodegradation by the evolution of radio-C02 from the samples after
two incubation periods. The incubation periods will be chosen
appropriately to show biodegradation activity for the given
substrate (e.g., the benzopyrene incubations will be longer than
for hexadecane).

A total of 20 grams of sediment from each sample will be needed for

this assay. Each sediment sample assayed for hydrocarbon
degradation will first be mixed 1:10 with sterile seawater
augmented with mineral nutrients. Ten ml aliquots of the resulting
slurry will then be placed in sterile 40-ml incubation vials fitted
with silicone septa. For each substrate, on selected sediment
samples, two concentrations will be used to investigate the effect
of hydrocarbon concentration on biooxidation rates. The substrate
of interest will be added at either 1 or 10 ppm (ug/ml slurry)
concentrations by injection via syringe through the septa. The
substrates will be added in an acetone carrier (Bauer and Capone,
1988). Two replicate vials for each substrate/sediment
sample/incubation time combination will be prepared with a "time
zero" killed control also prepared for each substrate and
triplicate set. All vials will be placed on a rotatory shaker for
24 hours and then incubated at ambient temperatures for the
duration of the incubation period.

Following incubation of the sample for the appropriate period (or
initially in the case of the controls) , substrate biodegradation in
the sample vials will be halted by the addition of 1 ml 1ON NaOH
through the septum. This will result in a Ph greater than 13,
killing the culture of degraders and sequestering any evolved C02
in the form of carbonates in solution. The extent of hydrocarbon
degradation will be monitored by measuring the radio-CO evolved
from each vial (Foght et al., 1989). After transport to the
analytical facility at the University of Alaska, the sample vial
contents will be acidified by addition of concentrated HC1 via
syringe through the septum. The headspace will be purged of radio-
C02 and the effluent gas will be passed first through an organic
vapor trap and then through phenethylamine scintillation cocktail
to trap the evolved C02 (Fedorak et al., 1982). The mean of each
set of biodegradation samples for each substrate, concentration and
incubation period will be compared with the "time zero" controls to
assess for losses due to volatilization in transit or any possible
abiotic C02 evaluation. The extent of biodegradation will be
expressed as a percentage of the total radiocarbon added to the
sample after correction for abiotic losses and ambient hydrocarbon
concentrations.

In addition to the biooxidation potential assay, populations of
hydrocarbon oxidizing bacteria will be enumerated using a "dilution
to extinction" technique. The Most Probable Number (MPN)
statistical enumeration technique, as modified for oil degrading
bacteria and shipboard space constraints (e.g., using sterile, 24-
well tissue culture plates), will be used.

Aliquots of slurry taken from the dilution bottles generated for
the biooxidation potential study will be serially diluted ten-fold
several times, giving a range of dilutions from 1:10 to 1:109.
Five replicate 100-ul aliquots of these dilutions will be placed
into the microliter plates' wells filled with sterile, carbon-free
marine broth, producing five identical inoculations of each

dilution. Then a "drop" of sterile Prudhoe Bay crude oil will be
added to each inoculation well. The crude oil serves as the sole
carbon source for any bacteria in the inoculum, selecting only
those able to grow on crude oil. A positive indication of growth
will be emulsification of the slick formed on addition of the oil
to the wells. The most probable number of hydrocarbon degraders
will then be calculated using a standard MPN table.

University of Alaska:

Five replicate samples will be taken at each of three stations
within six bays identified as oil-exposed sites and at three
stations within six bays determined to have been uncontaminated
(control) sites. All stations sampled will be at approximate
depths of 40, 100, and >100m on a transect extending below seagrass
(Zostera) beds within each of the identified bays. The intertidal
and shallow subtidal stations on the transect will be sampled for
biota. A total of 36 deep stations x 5 replicates will be
collected on a single cruise in early July in conjunction with
microbiological and hydrocarbon sampling projects that will be
underway from the same ship platform at the same or at
approximately the same time. Benthic samples at oil-exposed and
unexposed sites will be collected on bottoms that are as physically
similar as possible, based on chart and fathometer data and
preliminary grab samples to be made before actual sampling occurs.
Considerable amounts of ship time might occasionally be required
at some sites to ensure that similar bottom types are compared
between oiled and control sites.

The six oil-exposed sites that will be sampled for deep benthos are
Northwest Bay, Disk Island, Herring Bay, Bay of Isles, Snug Harbor,
and Sleepy Bay. The six unexposed (control) sites to be sampled
are West Bay, Rocky Bay, Zaikof Bay, MacLeod Harbor, and two sites
to be selected prior to the July cruise.

The benthic biological samples from approximately 40, 100, and
>100m will be collected with a O.lm2 van Veen grab weighted with
31.7 kg of lead to facilitate penetration. Five replicate samples
will be taken at all stations. Material from each grab will be
washed on nested 1.0 mm steel screens and preserved in 10%
formalin-seawater solution buffered with hexamine.

DATA ANALYSIS

The null hypotheses to be tested will depend on which of the
objectives listed above is under consideration. In general, for
sediment analyses the null hypothesis will state that the
concentration of petroleum hydrocarbons at particular depths or the
distribution of petroleum hydrocarbons with depth at oiled sites
does not differ from that at reference sites. All data will be
tested for heteroscedasticity with Bartlett's test or equivalent.

Data will be reported as means and 95% confidence intervals
calculated according to a standard formula (Sokal and Rohlf 1981).
Parametric statistics (Model I analysis.of variance with site and
depth as fixed factors and Scheffe's a posteriori test) will be
used to test for differences in hydrocarbon concentrations between
sites and depths if underlying assumptions of the parametric
procedures are met (with data transformation if required),
otherwise nonparametric tests (eg. the Kruskal-Wallis test) will be
employed. Key petroleum weathering and source ratios will be
calculated (Boehm et al. 1987).

The relationship of sediment toxicity to luminescent bacteria at
the study sites and hydrocarbon concentrations determined in
sediment will be compared statistically using appropriate tests
(Sokal and Rohlf 1981). Where significant differences are found,
the a value will be understood to be < 0.05.

Analysis of the data on the deep benthos will be completed using
previously written programs at the University of Alaska for
comparison of species (taxa), rank abundance and rank biomass of
species (taxon) . A diversity program will also be used to examine
differences and similarities between stations. Station groups and
species (taxon) assemblages for each year and for the combined data
collected on cruises in future years will be identified using the
technique of hierarchical cluster analysis. Principal coordinate
analysis will be used as an aid in interpretating of the cluster
analysis of the data and in identifying misclassifications of
stations by cluster analysis. Use of'both of these multivariate
techniques makes it possible to examine similarities (or
dissimilarities) between groups of stations, and will be useful
when comparing oiled vs unoiled bays.

A Kruskal-Wallis and a multiple comparison test for significance
will be used to test for differences in the total abundance and
biomass between the stations sampled in each year and in the multi-
year data sets. These same tests will be made on the abundance and
biomass of selected, dominant taxa at stations between years. The
taxa will be chosen from the rank abundance and biomass printouts
for each station, and taxa selected will generally be those
commonly present within bays being compared. However, taxa that are
common at stations within unoiled bays, but rare or missing at
stations within oiled bays, will also be tested. other statistical
tests, such as the two-tailed Wilcoxon signed ranks test for
pairwise observations, will be used to test differences between
stations at similar depths and bottom type within unoiled and oiled
bays.

Various measures of diversity will be calculated, and compared
qualitatively between stations at similar depths within unoiled and
oiled bays. The indices to be calculated and presented are:
Shannon Diversity (measures total diversity), Simpson Dominance
(useful f or identifying dominance by one or a f ew taxa at a

station), Evenness, and Species Richness.

The calculation of K-dominance curves f or the abundance and biomass
data will be used in an attempt to assess the effect of
hydrocarbons on benthic organisms in oiled bays. This is a
technique designed to detect pollution-induced disturbance on
marine benthic communities. Distributions of geometric classes of
abundance of species will also be calculated. Assessment of the
distribution of taxa in these abundance classes is often useful to
identify indicator species within a disturbed area.

The goal of the data analysis in this study is to determine the
effects of short and long-term accumulations of petroleum
hydrocarbons on benthic species composition, species diversity,
abundance, biomass, and trophic composition. The critical aspect
of the study is whether concentrations of petroleum contaminants
from the EVOS are present at concentrations which cause deleterious
effects on benthic organisms. Because the "deleterious effects"
criteria are complex and often require subjective interpretation,
a detailed comparison is ultimately required of the hydrocarbon
concentrations at which various biochemical, behavioral,
physiological, organismal, population, and ecological effects
occur. This only addresses certain aspects of the organismal,
population, and ecological effects on the benthic infauna.

BIBLIOGRAPHY

Abelson, P. H. 1989. Oil spills. Science 244:629.

Bauer, J.E. and D.G. Capone. 1988. Effects of co-occurring
aromatic hydrocarbons on degradation of polycyclic aromatic
hydrocarbons in marine sediment slurries. Appl. Env.
Micrbiol. 54:1644-1655.

Boehm, P. D., M. S. Steinhauer, D. R. Green, B. Fowler, B.
Humphrey, D. L. Fiest, and W. J. Cretney. 1987. Comparative
fate of chemically dispersed and beached crude oil in subtidal
sediments of the arctic nearshore. Arctic 40, supp. 1: 133-
148.

Environmental Protection Agency. 1987. Recommended protocols for
sampling and analyzing subtidal benthic macroinvertebrate
assemblages in Puget Sound. Environmental Protection Agency,
Final Report TC-3991-04.

Fedorak, P.M., J.M. Foght, and D.W.S. Westlake. 1982. A method for
monitoring mineralization of 14C-labeled compounds in aqueous
samples. Water Res. 16:1285-1290.

Foght, J.M., D.L. Gutnick, and D.W.S. Westlake. 1989. Effect of
Emulsan on biodegradation of crude oil by pure and mixed

bacterial cultures. Appl. Env. Microbiol. 55:36-42.

Hoberg, M. K. 1986. A numerical analysis of the benthic infauna
of three bays in Prince William sound, Alaska. M.A. Thesis,
Humboldt State University, Arcata, CA 153 pp.

Krahn, M.M., L.K. Moore, R.G. Bogar, C.A. Wigren, S-L, Chan, and
D.W. Brown. 1988a. A rapid high-pressure liquid chromato-
graphic method f or isolating contaminants from tissue and
sediment extracts. J. Chromatography. 437:161-175.

Krahn, M.M., C.A. Wigren, R.W. Pearce, L.K. Moore, R.G. Bogar,
W.D. MacLeod, Jr., S-L, Chan, and D.W. Brown. 1988b.
Standard analytical procedures of the NOAA National Analytical
Facility. New HPLC cleanup and revised extraction procedures
for organic contaminants. U.S. Dep. Commer., NOAA Tech. Memo
NMFS F/NWC-153, 52 p.

Schiewe, M.H., E.G. Hawke, D.J. Actor, and M.M. Krahn. 1985. Use
of a bacterial bioluminescence assay to assess toxicity of
contaminated marine sediments. Can. J. Fish. Aquat. Sci.
42:1244-1248.

Science Applications International Corporation. 1989. Screening
analysis f or petroleum hydrocarbons in sediments and sediment
pore waters by use of ultra-violet fluorescence
spectrophotometry for Exxon Valdez damage assessment. Draft
Final Report No. SAIC-89/7570&230* to the National Oceanic and
Atmospheric Administration. 36p.

Sokal, R. R. and F. J. Rohlf 1981. Biometry. W. H. Freeman and
company, San Francisco. 859pp.

Table 1. Location of sites in PWS and the GOA where intertidal
and subtidal sediment and biological samples will be collected in
1990. Samples of hydrocarbon-degrading bacteria will be collected
at all sites in each sampling period. Samples for microtox
bioassay will be collected at all sites in June/July only. Deep
benthos (D) will be collected at selected sites in June/July.
Depths to be sampled are: A intertidal (1) 3, 6 and 20 m; C I,
3, 6, 20, 40 and 100 m.

Location May June/July Sept

Prince William Sound'
Bay of Isles A C,D A
Block Island A C A
Chenega Island A C A
Disk Island A C,D A
Fox Farm A C A
Green Island A C A
Herring Bay A C,D A
Macleod Harbor A C,D A
NE Knight Island A C A
NE Port Fidalgo A C A
Northwest Bay A C,D A
Olsen Bay A C A
Rocky Bay A C,D A
Sleepy Bay A C,D A
Smith Island A C A
Snug Harbor A C,D A
West Bay A C,D A
Zaikof Bay C,D A

Gulf of Alaska
Agnes Cove C
Black Bay C
Chugach Bay C
Hallo Bay C
Katmai Bay C
Sunny Cove C
Tonsina Bay C
Windy Bay C

1. Two additional sites will be selected before June 1990 to
provide additional control sites for the deep benthos project.

BUDGET

NOAA

Salaries $159.8
Travel 20.9
Contracts 53.0
Supplies 50.5
Equipment 32.6
Vessel 150.0

Total $466.8

ADF&G

Salaries $174.7
Travel 5.7
Contracts 124.9
Supplies 23.1
Equipment 5.1

Total $333.5

AIR/WATER STUDY NUMBER 3

Study Title: Geographic and Temporal Distribution of Dissolved
and Particulate Petroleum Hydrocarbons in the
Water Column

Lead Agencies: NOAA, DEC

INTRODUCTION

This study will continue to assess the geographic and temporal
distribution of dissolved and particulate hydrocarbons in the water
column and deposited in sediments resulting from the EVOS.
Knowledge of these concentrations will determine whether violations
of State of Alaska Water Quality Criteria have occurred, and will
allow estimation of the exposure risk of subsurface marine biota to
petroleum hydrocarbons. This study extends work begun within one
week of the grounding of the Exxon Valdez and continued to date.

A. DEC

During the autumn of 1989 (November/December), DEC collected
interstitial water samples at target sampling sites in PWS and
deployed sediment traps at selected sites. It was determined that
no further interstitial water sampling would be conducted in 1990.
Studies related to hydrocarbonoclastic bacteria will be continued
by AW 2. An increased number and distribution of sediment traps is
planned.

B. NOAA

Trends in hydrocarbon concentration in the water column will be
studied by analyzing the hydrocarbon body burden of transplanted
bay mussels Mytilus trossulus. The use of a bioaccumulator
provides a time integrated indication of hydrocarbons available in
the water column. No further direct sampling of the nearshore
water column will be done because hydrocarbon concentrations in the
water column will likely be below detection levels in field samples
that are practical to analyze.

The products of this study will consist of estimates of aliphatic
and aromatic hydrocarbons in the matrices examined. These data will
be used to determine biological resource exposure to petroleum
hydrocarbons.

This study is coordinated with the other A/W studies and with the
CH 1 to provide information on petroleum hydrocarbon distribution
and movement in the nearshore water column to researchers assessing
biological and economic damage. In the 1990 f ield season this

study will share research platforms with AW 2 and FS 24. Several
sites of this study will coincide with sites from these two studies
and with at least one CH 1 subtidal control site. Data gathered at
these joint sites will provide a comprehensive picture of damage
and will be especially valuable to studies assessing biotic and
economic damage. Selection of NOAA AW 3 study sites was aided by
information produced by AW 1 and AW 4 (now part of AW 2).
Information on beach cleanup at study sites will be obtained from
the DEC Spill Response Office and NOAA HAZMAT.

OBJECTIVES

A. To determine if sediments settling out of the water column in
nearshore subtidal environments contain absorbed hydrocarbons
(DEC).

B. Determine hydrocarbon inputs in nearshore environments and
evaluate trends in ambient water quality using mussels
(Mytilus trossulus) as bioaccumulators (NOAA/NMFS).

METHODS

A. DEC

Subtidal particulate samples will be collected with sediment traps
for hydrocarbon analysis. sampling arrays containing three
sediment traps each will be placed in the subtidal zone adjacent to
target shorelines at no more than 20 meters below mean lower low
water. Results will generate information on sedimentation and
associated hydrocarbon inputs to subtidal sediments.

Currently, five sediment trap arrays are in place in four locations
in PWS: Sleepy Bay, Snug Harbor, Northwest Bay (2), and Northeast
Port Fidalgo. Each platform consists of three removable long-term
sediment traps. These traps will be picked up, and the number and
distribution of the traps in PWS will be increased. Sediment traps
will be placed in as close proximity as possible to the mussel
cages being deployed in the NMFS segment of this study.
Approximately fifteen additional emplacements (3 traps per site)
are proposed. Sediment trap arrays will be deployed in relation
to shoreline habitat types, according to the Environmental
Sensitivity Index (Gundlack and Hays 1982), in conjunction with
bioaccumulation where possible, and in relation to shoreline
treatment methods as deemed feasible.

Particulate samples from sediment traps will be screened for
hydrocarbon content by ultraviolet fluorescence spectrophotometry
after methylene chloride extraction of the samples in the field.
UVF is a semiquantitative method of analysis for hydrocarbons
(ASTM, 1982). Samples showing significant quantities of petroleum

hydrocarbons will be further analyzed for polynuclear aromatic
hydrocarbons (PAH) and total petroleum hydrocarbons (PHC) according
to procedures established by TS 1.

B. NOAA/NMFS

Mussels will be placed at all 1989 sites within PWS (Figure 1)
except Squire Island and The Needle. Except for Olsen Bay
(control), all sites were in the spill trajectory and subject to
varying degrees of oiling. Redeployment this year will indicate
changes in water column hydrocarbon concentrations at these sites
since deployed mussels were last collected in September 1989.
Seven additional sites are proposed: four at sites of maximum
original oiling as indicated by preliminary analysis of water
column samples (AW 3) and sediment pore water samples (AW 4), and
at Disk Island and Black Island where extensive cleanup activity is
anticipated. Mussels will also be deployed at a second control
site, McCleod Bay.

Outside PWS, redeployment is proposed at Sunny Cove (Resurrection
Bay), Black Bay, Tonsina Bay, Blue Fox Bay (Af ognak Island), Hallo
Bay and Kukak Bay (control) (Figure 2). The two new sites, Agnes
Cove and Windy Bay, are coincident with AW 2 sites.

Local siting will ensure a site depth of 34 m (to accommodate the
deepest mussel cage) and the best available protection from
prevailing weather and currents. Site substrates will be cobbled
to finer sediments to ensure that mooring anchors are set securely.
If treatment activity occurs, siting may be adjusted or additional
sites may be added so that beaches adjacent to sites represent both
treated and untreated beaches.

Physical data on location (geographic coordinates) , site depth,
sampling time, tidal stage, and temperature and salinity at
deployment depths will be recorded at each site.

Bay mussels will be collected from a hydrocarbon free site,
Admiralty Island in southeast Alaska, a few days before each new
deployment cruise. Mussels will be held in living stream tanks,
that have been rinsed with dichloromethane and flushed with ambient
unfiltered seawater at the rate of 2 liters/minute at least
overnight. Since mussel size influences hydrocarbon uptake (Bayne
et al., 1981), only mussels with shell length of 45-50 mm. will be
selected for deployment. A sample from each collection of at least
30 individuals will be measured for shell length, width, and height
and whole wet versus dry weight. Another 40-50 animals will be
taken immediately prior to shipment of mussels to a deployment
vessel as a reference sample of the population's base hydrocarbon
level and condition.

Mussels will be shipped to the field in layers of healthy Fucus sp.
seaweed in insulated coolers whose lids have been drilled with air

holes. Mussels will be kept aboard the deployment/ collection
vessel in coolers and the blue ice changed daily for up to 6 days.
on longer cruises, mussels will be resupplied by air or possibly
irrigated with the seawater. Samples of irrigation water will be
taken daily, extracted with dichloromethane,and frozen. A mussel
baseline sample will be taken before deployment of irrigated
mussels at each new site.

A deployment "cage" is a nylon mesh diver collecting bag held open
by a perforated polypropylene sheet that has been rinsed with
dichloromethane and fitted into the bag bottom. Twenty mussels
will be placed in each bag separately (i.e. byssal connections to
other mussels will be separated so that byssal development observed
when mussels are collected will have occurred during field
exposure.) Assuming some mortality during exposure, this number
was chosen to provide at least triplicate samples of 3 � .5 g of
issue for hydrocarbon analysis (Krahn et al 1988). At 4 sites in
PWS an extra cage containing 40 mussels will be deployed at 1 m to
be exposed for 8 to 10 weeks so that hydrocarbon uptake over the
longer period may be compared with uptake over the 2 shorter
periods at the same site. Filled bags will be closed and attached
to the mooring line with a halibut snap. At each site, bags will
be attached at the I m, 5 m, and 25 m depths. The 2 shallower cage
depths were chosen to correspond to water column depths sampled by
this study in the f irst 6 weeks after the spill; mussels at the
third depth will be exposed to the water column about 10 m above
the bottom at low tide. Reference samples of mussels will be
taken just after the final deployment on a cruise to determine any
hydrocarbon uptake or deterioration of general condition during
holding of mussels on the vessel. Exposure times will be 4 to 5
weeks, and 8 to 10 weeks at the four selected sites. There will be
three periods of exposure at PWS sites and one period at Kenai,
Afognak and Alaska Peninsula sites.

After exposed mussels are retrieved, the number of clumps of
mussels, the number of individuals per clump, comments regarding
the strength and elasticity of byssal threads, and the number of
alive, dead, or gaping animals will be recorded. Dead or gaping
animals will discarded. At least 1 hydrocarbon free 16 oz jar with
a Teflon lid will be filled with live animals from each bag, kept
in a cooler, and frozen at -18 OC as soon as possible. A field air
blank will be taken at the site and on the vessel, if sample jars
are filled aboard the vessel. Mussel cages will then be refilled
and redeployed.

Naturally occurring adult mussels will be collected in intertidal
areas adjacent to some deployment sites. These will be packed in
clean 16 oz jars and handled similarly to caged mussel samples.

Sample estimates are: 236 caged exposed mussel samples, 59 air
blanks, 20 native mussel samples, and 15 reference samples.

DATA ANALYSIS

A. DEC

Hydrocarbon concentration data will be tested for
heteroscedasticity (Bartlett's test) and reported as means and 95%
confidence intervals calculated according to a standard formula
(Sokal and Rohlf, 1981). Parametric statistics will be used to
test for differences between hydrocarbon concentrations between
sites, if the assumptions of 'parametric procedures are met.
Otherwise, nonparametric tests (e.g., the Kruskal-Wallis test) will
be employed.

B. NOAA/NMFS

ANOVA will be used to determine the significance of differences of
any hydrocarbons found in the collected samples.

Products of this study will consist of tables containing lists of
hydrocarbons found in the samples collected.

BIBLIOGRAPHY

ASTM D-3650-78. Standard Test Method for Comparison of Waterborne
Petroleum Oils by Fluoresence Analysis.

Bayne, B.L., K.R. Clarke and M.N. Moore. 1981. Some practical
considerations in the measurement of pollution effects on
bivalve molluscs and some possible ecological consequences.
Aquatic Toxicology 1:159-174.

Gundlach, E.R. and M.O. Hayes. 1982. The oil spill environmental
sensitivity index applied to the Alaskan coast: in 1�82
Arctic Marine Oil Pollution (AMOP) Seminar. EcTi-onton,
Alberta, Canada. pp.311-323

Krahn, M.M., C.A. Wigren, R.W. Pearce, L.K. Moore, R.G. Bogar, W.D.
MacLeod,Jr., S. Chan, and D.W. Brown. 1988. Standard
analytical procedures of the NOAA National Analytical
Facility, 1988, New HPLC cleanup and revised extraction
procedures for organic contaminants. NOAA Technical
Memorandum NMFS F/NWC-153. 52 pp.

Sokal, R. R. and R. J. Rohlf. 1981. Biometry. Freeman, San
Francisco.

.39

BUDGET: DEC

Salaries $ 19.6
Travel 1.4
Contracts 17.5
Supplies & Equipment 9.0

TOTAL $ 47.5

BUDGET: NOAA

Salaries $161.5
Travel 15.3
Contracts 12.3
Supplies 40.2
Equipment 43.2
Ship Costs: 200.0

TOTAL @-472.5

TOTAL both Projects $520.0

AIR/WATER STUDY NUMBER 6

Study Title: Fate and Toxicity of Spilled Oil from the EVOS

Lead Agency: NOAA

INTRODUCTION

Overview and Relation to other Studies

This study is designed to: a) assess the toxicity of weathered
EXXON VALDEZ oil and its degradation products to selected test
organisms; and b) integrate the results from selected other
projects into an overall budget for the distribution, transport,
transformation, and persistence of spilled oil in Alaskan coastal
environments. The study is very closely coordinated with A/W
Study 2 for its field work and toxicity studies, and will require
close interaction with all of the present and past A/W studies, the
Coastal Habitat study, and with related spill response studies for
completion of the spilled oil budget.

Toxicity of Crude Oil in Relation to the Weathering Process

Currently, limited information is available on the significance of
either the polar constituents of crude oil or the intermediate
oxidation products of petroleum hydrocarbons (whether from
photooxidation or biodegradation) in terms of their potential for
bioaccumulation and toxicity to resource organisms in the marine
environment. Since these compounds have undergone preliminary
oxidation and (sometimes) conjugation, they are more polar than
their parent hydrocarbons, and will as a result generally be more
subject to excretion or depuration, less subject to
bioaccumulation, more susceptible to further oxidation (or
biodegradation if accumulated), and more susceptible to dilution
and dispersion in the water column. Studies proposed here are
designed to help determine whether such polar constituents pose a
significant risk of toxicity or mutagenicity to Alaskan marine
organisms as a result of the EVOS.

Acute and Sublethal Toxicity of oil to Marine Organisms

A very considerable body of literature exists on the toxicity of
Alaskan crude oil to Arctic and subarctic marine organisms. The
data base is probably adequate for assessing the relative
sensitivities of different marine species to exposure and for
estimating the range of potential responses (at the organismic
level) that may result from a particular level of exposure in the
environment. However, very little of this prior toxicity research
has been directed specifically at the contribution of either
hydrocarbon metabolites or other oxidation products of oil that may

be produced by the processes.of biological or chemical weathering
in the environment.

Sources of Toxicity in Crude Oils

By the mid-1970's, it had been concluded that much of the acute
toxicity of oil was accountable directly to the content of soluble
aromatic compounds (Moore and Dwyer 1975; Neff et al. 1976), and
attention was being directed towards determining which fractions of
petroleum were most responsible for the toxicity observed in
laboratory and f ield exposures to oil. Based on the relative
concentrations of the low-molecular weight constituents in crude
oill, it has become generally accepted that most of the acute
toxicity effected by oil in the environment is derived from the
mono- and di-nuclear aromatics. When a water soluble fraction
(WSF) was simulated by mixing the 10 predominant aromatic
hydrocarbons at the same concentrations and proportions found a
true WSF of crude oil, however, the toxicity of the resulting
mixture was only 20-30% of the true WSF, suggesting that either
minor aromatic constituents, or components other than aromatic
hydrocarbons, also contribute significantly to the observed
toxicity (Rice et al 1984).

Polar Constituents and Oxidation Products of Oil

Petroleum in the marine environment is decomposed primarily through
the processes of microbial biodegradation and photooxidation or
autooxidation These processes are effective for oil in surface
slicks, in the water.column, in sediments, and in the atmosphere
(photooxidation of evaporated compounds) . In addition, parent
petroleum compounds are bioaccumulated, and metabolized by
macro-organisms. While the. eventual major products of these
oxidative reactions are carbon dioxide and water, some of the
oxygenated intermediates produced along the way may be more toxic
than their precursors.

oxidation products of photooxidation include fatty acids, alkylated
naphthols, and substituted 1- to 3-ring aromatic and heteroaromatic
acids, as well as alkylated benzothiophene sulfoxides (Overton et
al. 19790, 1980). Microbial biodegradation of alkanes,
cycloalkanes, and monoaromatics leads to the production of
alcohols, aldehydes, and carboxylic acids that are generally of
little concern from a toxicity standpoint. Condensed polyaromatic
hydrocarbons, however, may be transformed by microbial metabolism
to potential carcinogens or mutagens. Materials such as
benzo (a) pyrene and benzo (a) -anthracene, f or example, are oxidized
by eucaryotic organisms (macroorganisms, yeasts and molds) to
trans-dihydrodiols which are subsequently activated into oxides
that bind to DNA and are powerful mutagens.

Some metabolic products of high-molecular weight aromatics are
demonstrated mutagens or carcinogens and have been shown to bind to
DNA (Ahokas et al. 1979; Lech And Bend 1980; Varanasi et al. 1981).
These same materials are also associated with the prevalence of
liver lesions, including neoplasms (Varanasi and Stein 1990).

Asphaltenes and resins are two heterogeneous and poorly
characterized assortments of (non-hydrocarbon) compounds that
comprise only about 2% and 6%, respectively, of Prudhoe Bay crude
oil (Clark and Brown 1977). Asphaltenes are constituents of tar
that are highly resistent to biodegradation, and are not generally
considered to pose a risk of toxicity to marine organisms. Resins
include the polar and heterocyclic nitrogen sulfur oxygen (NSO)
compounds, such as phenols, cresols, thiophenes, dibenzothiophenes,
pyridines, and pyrroles. Some of them are likely to undergo
biodegradation, and very broad suites of NSO compounds have been
Adentified in contaminated marine environments (Krone et al. 1986,
Wolfe et al. 1981). Like hydrocarbon metabolites, many of these
compounds are moderately water-soluble and therefore subject to
dispersion in the water column. While some of these polar
compounds could be toxic at high concentrations, no studies have
been made of the levels of toxicity exerted by these materials
under oil spill conditions in the marine environment.

Fate of Spilled Oil: Budgets and "Mass Balance"

An accurate and complete mass balance has yet to be prepared for
any major oil spill in a marine environment. The quality of
estimates of the quantities and locations of oil affected by
different processes of transport or transformation have varied from
spill to spill, depending on the local circumstances of the spill
and the effort devoted to any particular process. Selected
observations at past spills have been summarized by Mackay (1981),
Gundlach et al. (1983), Jordan and Payne (1980), National Academy
of Sciences (1985), and Wolfe (1985, 1987). Information especially
pertinent for summarizing the fate of oil from the EXXON VALDEZ
spill has been and is being gathered by the Interagency Response
Team and the ADEC, and by certain projects under the NRDA.

OBJECTIVES

A. Document the toxicity of contaminated sediments and related.
environmental samples to selected marine biota

B. At selected sites, document and quantify the occurrence of
oxidized derivatives of EXXON VALDEZ oil

C. Determine the extent to which the observed toxicity of
oil-contaminated environmental samples may be attributable
to oxidation products of petroleum.

D. Construct a summary budget or "mass balance" summarizing the
fate of the spilled oil.

METHODS

A. Toxicity of oil-Contaminated Sediments and other
.Environmental Samples

Toxicity tests will be performed on sediment samples taken at
selected sites sampled by A/W Study 2. Two specific tests, both
following well-established protocols, are proposed: a sediment
elutriate test using larval mussels Mytilus edulis, and a whole
sediment test using Ampelisca abdita (or other suitable Alaskan
species of Ampelisca). Mytilus is an intertidal species whose
larval recruitment is vulnerable to interruption by toxic oil
residues remaining in intertidal sediments. Ampelisca inhabits
soft nearshore sediments that are possible sinks for petroleum.
Subtidal ampeliscid amphipods exhibited considerable sensitivity to
oil in the aftermath of the AMOCO CADIZ spill (Cabioch et al.
1982). Use of these two species should provide a direct measure of
the toxicity of the residual oil to actual marine species.

Sampling sites have been selected to represent the more heavily
oiled areas. At each of 20 of the sites, eight one-liter samples
of surficial sediments will be collected (2 each at the intertidal,
6-meter, 20-meter, and 100-meter depths) for toxicity testing with
Mytilus and Ampelisca. These samples will be stored at 0-4 degree
celsius and offloaded from the vessel at regular intervals for
shipment to the testing laboratory. Bioassays will be initiated
within 10 days of the collection of the samples.

B. Oxidation Products of Petroleum

The fractionation and toxicity testing of polar constituents from
weathered petroleum will be pursued in a tiered, stepwise fashion.
A limited pursuit of chemical fractionation and characterization
will be undertaken in association with the toxicity testing to be
performed under A/W2. This preliminary study will include two
major objectives, and will employ the following approaches:

1. Determine whether the toxicity (if any) of organic
extracts from Exxon Valdez o i 1 -contaminated sediments is
partitioned between polar and non-polar constituents:

Approach:

perform microtox on unfractionated extracts
perform one-step separation on PAC column to
fractionate polar from non-polar constituents
repeat -microtox on the two separated fractions
perform simple carcinogenicity/mutagenicity

bioassay (eg Ames test, S.O.S. Chromotest) on
selected fractions

2. Determine what types of polar constituents are present in
these extracts:

Approach:

perform HPLC/UV on polar (and non-polar) fractions
to identify presence of PNA derivatives
perform detailed GC/MS on selected polar fractions
to identify and quantify (major)constituents

The work should focus (a) on determining whether a
significant fraction of the observed toxicity or
mutagenicity can be ascribed to polar derivatives in a few
of the most biologically active samples, and if so, (b) on
a preliminary characterization of the polar constituents
that may be involved. The selection of samples for this
chemical fractionation and characterization will be guided
by the magnitudes of the Microtox and UVF signals.

If this preliminary work suggests that polar constituents
could account for significant toxicity in the marine
environment, more intensive testing will be performed. At
two heavily oiled sites, one untreated and one that has
undergone bioremediation (both yet to be selected), and one
lightly or unoiled site in PWS, special samples will be
taken to assess the concentrations and compositions of
petroleum oxidation products in intertidal sediments. These
samples will also be taken in conjunction with.A/W Studies
2 and 3, probably from the NOAA vessel COBB during the late
summer or fall of 1990. Large quantities of sediments
and/or interstitial water will be required to support the
necessary development of suitable techniques for bulk
fractionation of samples to be tested for toxicity (in C.
below) and for chemical characterization and quantification
of the polar metabolites.

Replicate (3) sediment samples on the order of 10kg (wet
weight) each will be collected at each site for exhaustive
extraction with methylene chloride. The chemical
fractionation procedure will include a succession of solvent
partitioning, absorption chromatography, HPLC using gel
exclusion, and GC-MS. Initial phases of fractionation will
follow closely the procedures outlined by MacLeod et al.
(1985) for separation and analysis of petroleum compounds,
but additional steps will be required for separation and
characterization of the more polar fractions.

Interstitial water samples should also be examined. Because
of uncertainties about the flux of interstitial water in

oiled beaches and the resultant levels of polar metabolites,
however, it is very difficult to estimate the volume of
sample that may be required for characterization and
quantification of polar metabolites. For initial trials, it
is suggested that "interstitial water" be pumped or siphoned
from a shallow "well" (i.e., a glass tube inserted to a
depth of about 30 cm in heavily oiled beach sediments) into
precleaned glass carboys (18-20 liters) containing acid and
methylene chloride to carry out the initial extraction and
to ensure preservation of the samples. An alternative
collection technique would be to allow "interstitial water"
merely to seep into an excavation on the beach and then to
dip the water into the sample carboys. At each of the sites
where sediments are collected for analysis, 100 liters of
"interstitial water" should be collected. Exhaustive
extraction with methylene chloride would be followed by
analytical steps similar to those used for sediments.

Target compounds of the fractionation and analytical scheme
will include phenolic, carbonyl, quinone, and carboxylic
derivatives of polynuclear aromatic hydrocarbons, for
example: 9-fluorenone, 9-fluorenone carboxylic acids,
phenanthraquinone as potential derivatives of phenanthrene.
Analogs related to naphthalene, anthracene, chrysenes,
benzanthracene, pyrene, and benzpyrene will also be sought
specifically, as Will oxidation products of
dibenzothiophenes. GC-MS data will be analyzed also for
other major constituents in associated polar fractions to
provide a general characterization of the polar compounds
found in these samples.

C. Toxicity of Oxidized Petroleum Fractions

Following the initial characterization and quantification of polar
constituents in oiled sediments and interstitial water, the
fractionation process will be scaled up to provide quantities of
material suitable for toxicity testing. Based on the results of
initial fractionation and chemical characterization, selected
polar fractions will be assayed for toxicity using the standard
Microtox bioassay with organic extracts (Schiewe et al. 1985).
Toxicity of polar fractions will be compared with the better known
toxicities of aromatic fractions and reference compounds. The
composition of all assayed fractions will be checked by GC-MS for
consistency with previous fractionations of earlier samples.

D. Mass Balance Budget for Fate of Spilled Oil (Budget)

This task is primarily a synthesis function. Information on the
distributicn and fates of EXXON VALDEZ oil needs to be assembled
from a number of sources and interpreted in light of existing
information and models.

The following compartments and processes are proposed for initial
analysis and inclusion in the Budget. Potential sources of data,
historical information, and modeling expertise are also noted:

1. Floating oil (Distribution in Time & Space)
2. Evaporation
3. Photooxidation in the atmosphere
4. Mousse formation
5. Beaching of oil & mousse (T&S)
6. Water column accommodation (T&S)
7. Photooxidation in water column, in slicks and on beaches
8. Biodegradation in water column
9. Transport to subtidal sediments
10. Biodegradation in sediments

Representatives of the above noted activities, along with other
recognized experts on oil weathering and fates, will be
consulted for recommendations on appropriate approaches to
synthesis, and for their judgments on the suitability and
adequacy of existing information for development of the Budget.
Apart from the information on polar constituents described
above, no original data is proposed for collection under this
project. Timely progress on the Budget will depend on the
availability of suitable information from other sources and
projects; chemical data, i.e., from T/S 1, will be of utmost
importance to the completion of this project. Where existing
information is found to be deficient, means will be explored
for gathering of improved information. The reliability of all
estimates will be assessed and qualified in the final analysis.

E. Ouality Assurance and Control

All samples will be taken with careful adherence to QA/QC Plan
for NRDA.

BIBLIOGRAPHY

Ahokas, J.T., H. Saarni, D.W. Nebert, and 0. Pelkonen. 1979.
The in vitro metabolism and covalent binding of
benzo(a)pyrene to DNA catalyzed by trout liver microsomes.
Chem. Biol. Interact. 25:103-111.

Brown, D.A., R.W. Gossett, and S.R. McHugh. 1987. Oxygenated
metabolites of DDT and PCBs in marine sediments and
organisms. PP. 61-69. In: J.M. Capuzzo and D.R. Kester
(eds) . Oceanic Processes in Marine Pollution. Vol. 1.
Biological Processes and Wastes in the Ocean. Krieger
Publishing Co., Malabar, Florida.

Cabioch, L., J.-C. Dauvin, C. Retiere, V. Rivain, and D.
Archambault. 1982, Les effets des hydrocarbures de 11AMOCO

CADIZ sur les peuplements benthiques des Baies de Morlaix
et de Lannion d'Avril 1978 a Mars 1981. PP. 205-228. In:
Ecological Study of the Amoco Cadiz Oil Spill. Rpt of the
NOAA-CNEXO Joint Scientific Commission. U.S. Dept.
Commerce, NOAA. Washington, D.C.

Clark, R.C., and D.W. Brown. 1977. Petroleum: Properties and
analyses in biotic and abiotic systems. PP. 1-89. In:
D.C. Malins (ed.) Effects of Petroleum on Arctic and
Subarctic Marine Environments and organisms. Vol. I,
Biological Effects, Academic Press, New York.

Dunn, B.P. 1980. Polycyclic aromatic hydrocarbons in marine
sediments, bivalves, and seaweeds: Analysis by
high-pressure liquid chromatography. PP. 367-377. In:
A. Bjorseth and A.J. Dennis (eds.) Proc. 4th Internatl.
Sympos. on Polynuclear Aromatic Hydrocarbons. Battelle
Press, Columbus, Ohio.

Gundlach, E.R., P.D. Boehm, M. Marchand, R.M. Atlas, D.M. Ward,
and D.A. Wolfe. 1983. The fate of AMOCO CADIZ oil.
Science 221: 122-129.

Jordan, R.R., and J.R. Payne. 1980. Fate and Weathering of
Petroleum Spills in the Marine Environment: A Literature
Review and Synopsis. Ann Arbor Science Publishers. Ann
Arbor, Michigan. 174 pp.

Karickhoff, S.W. 1981. Semi-empirical estimation.of sorption
of hydrophobic pollutants on natural sediments and soils.
Chemosphere 10:833-846.

Krone, C.A., D.G. Burrows, D.W. Brown, P.A. Robisch, A.J.
Friedman, and D.C. Malins. 1986. Nitrogen-containing
aromatic compounds in sediments from a polluted harbor in
Puget Sound. Environ. Sci. Technol. 20:1144-1150.

Lech, J.J. and J.R. Bend. 1980. The relationship between
biotransformation and the toxicity and fate of xenobiotic
chemicals in fish. Environ.. Health Perspectives 35:115.

Mackay, D. 1981. Fate and behaviour of oil spills. PP. 7-27.
In: J.B.Sprague, J.H. Vandermeulen, and P.G. Wells (eds.)
Oil Dispersants in Canadian Seas-Research Appraisal and
Recommendations. Environment Canada, Toronto.

MacLeod, W.D., Jr., D.W. Brown, A.S. Friedman, D.G. Burrows, 0.
Maynes, R. Pearce, C.A. Wigren, and R.G. Bogar. 1985.
Standard analytical procedures of the NOAA National
Analytical Facility, 1982-1986: Extractable toxic organic
compounds. 2nd Edition. NOAA Technical Memorandum NMFS
F/NWC-92. 121 pp.

Means, J.C., J.J. Hassett, S.G. Wood, and W.L. Banwart. 1979.
Sorption properties of energy-related pollutants and
sediments. PP. 327- In: P.W. Jones and P. Leber
(eds.) Polynuclear Aromatic Hydrocarbons. Ann'Arbor
Science Publishers, Ann Arbor, Michigan.

Means, J.C., S.G. Wood, J.J. Hassett, and W.L. Banwart. 1980.
Sorption of polynuclear aromatic hydrocarbons by
sediments and soils. Environ. Sci. Technol. 14:
1524-1528.

Melancon, M.J.,Jr., and J.J. Lech. 1979. Uptake,
biotransformation, disposition, and elimination of
2-methylnaphthalene and naphthalene in several fish
species. PP. 5-22. In: L.L. Marking and R.A. Kimerle
(eds.) Aquatic Toxicology. ASTM STP667. Am. Soc.
Testing and Materials. Philadelphia, Pennsylvania.

Moore, S.F. and R.L. Dwyer. 1975. Effects of oil on marine
organisms: A critical assessment of published data. Water
Research 8: 819-

National Academy of Sciences. 1985. Oil in the Sea. Inputs,
Fates, and Effects. National Academy Press, Washington,
D.C.

Neff, J.M., J.W. A nderson, B.A. Cox, R.B. Laughlin, Jr., S.S.
Rossi, and H.E. Tatum. 1976. Effects of petroleum on
survival, respiration and growth of marine animals. PP
516-539. In: Sources, Effects & Sinks of Hydrocarbons in
the Aquatic Environment. The American Institute of
Biological Sciences, Washington, D.C.

Overton, E.B., J.R. Patel, and J.L. Laseter. 1979. Chemical
characterization of mousse and selected environmental
samples from the AMOCO CADIZ oil spill. PP. 169-174. In:
Proceedings, 1979 Oil Spill Conference (Prevention,
Behavior, Control, Cleanup), American Petroleum
Institute Publication No. 4308, Washington, D.C.

Overton,. E.B., J.L. Laseter, W. Mascarella, C. Rashke, I.
Noiry, and J.W. Farrington. 1980. Photochemical
oxidation of IXTOC-I oil. PP. 341-383. In: Proc.
Sympos. Preliminary Results from the September 1979
RESEARCHER/PIERCE IXTOX-I Cruise. NOAA Office of Marine
Pollution Assessment, Boulder, Colorado.

Rice, S.D., A. Moles, J.F. Karinen, S. Korn, M.G. Karls, C.C.
Brodersen, J.A. Gharrett, and M.M. Babcock. 1984.
Effects of petroleum hydrocarbons on Alaskan aquatic
organisms: A comprehensive review of all oil-effects
research on Alaskan fish and invertebrates conducted by

the Auke Bay Laboratory, 1970-1981. U.S. Dept. Commerce,
NOAA Tech. Memo. NMFS/NWC-67. 128 pp.

Roubal, W.T., T.K. Collier, and D.C. Malins. 1977.
Accumulation and metabolism of carbon-14 labeled benzene,
naphthalene, and anthracene by young coho salmon
(Oncorhynchus kisutch). Arch. Environ. Contam. Toxicol.
5:513-529.

Schiewe, M.H., E.G. Hawk, D.I. Actor, and M.M. Krahn 1985.
Use of a bacterial bioluminescence assay to assess
toxicity of contaminated marine sediments. Canadian
Journal of Fisheries and Aguatic Sciences 42: 1244-1248.

Varanasi, U., and J.E. Stein. 1990. Disposition of xenobiotic
chemicals and metabolites in marine organisms.
Environmental Health Perspectives (in press).

Varanasil U., D.J. Gmur, and W.L. Reichert. 1981. Effect of
environmental temperature on naphthalene metabolism by
juvenile starry flounder (Platichthys stellatus) and rock
sole (Lepidopsetta bilineata). Arch. Environ.
Contam.Toxicol. 10:203-214.

Veith, D.G., D.L. Defoe, and B.V. Bergstedt. 1979. Measuring
and estimating the bioconcentration factor of chemicals in
fish. J. Fish. Res. Board Canada 36:1040-

Wolfe, D.A. 1985. Fossil Fuels: Transportation and marine
pollution. Chapter 2. pp. 45-93. In: Iver W.Duedall,
Dana R. Kester, P. Kilho Park and Bostwick H. Ketchum
(eds.), Wastes in the Ocean, Volume 4. Energy Wastes in
the Ocean. John Wiley & Sons, New York.

Wolfe, Douglas A. 1987. Interactions of spilled oil with
suspended materials and sediments in aquatic systems. pp.
299-316. In: K.L. Dickson, A.W. Maki, and W.A. Brungs
(eds.), Fate and Effects of Sediment-bound Chemicals in
Aguatic Systems. Proceedings of the 6th Pellston
Workshop, August 12-17, 1984. Florissant, Colorado.
Pergamon Press, Oxford, England.

Wolfe, D.A., R.C. Clark, Jr., C.A. Foster, J.W. Hawkes, and
W.D. MacLeod, Jr. 1981. Hydrocarbon accumulation and
histopathology in bivalve molluscs transplanted to the
Baie de Morlaix and the Rade de Brest. PP. 599-616. In:
AMOCO CADIZ. Internatl. Symposium on the Amoco Cadiz:
Fate and Effects of the Oil Spill. Brest, France. 19-22
Nov 1979. CNEXO, Paris, France.

BUDGET: NOAA

salaries $60.0
Travel/Shipping 67.0
Contracts 580* 0
Supplies 13.0
Equipment 0.0
Ship Costs 150.0 1

Total $870.0

FISH/SHELLFISH INJURY ASSESSMENT

The grounding of the tanker Exxon Valdez discharged crude oil into
one of the richest marine ii7sheries communities of the United
States. Although oil contamination was most severe within PWS, the
oil spread into large portions of the Gulf of Alaska (GOA), Lower
Cook Inlet (LCI) , Shelikof Strait, and other North Pacific ocean
waters off the coasts of Kodiak and the Alaska Peninsula. The fish
and shellfish populations inhabiting these marine and estuarine
waters form integral parts of a vast and complex ecosystem which
also includes various other invertebrate species, birds, and
mammals (including humans).

For example, the various life history stages of Pacific herring are
important forage species for various piscivorous fishes (e.g.
Pacific salmon, halibut, etc.), birds (gulls, cormorants, eagles,
loonst etc.), mammals (sea lions, seals, whalest etc.),
invertebrates (crabs), and are used for subsistence and commercial
purposes. . Regarding Pacific salmon, outmigrating smolts are
important seasonal prey items for a variety of predatory fish and
marine birds. Maturing salmon in the high seas and adult salmon
returning to inland waters are the major portion of the diet of
marine mammals such as sea lions,'seals, and killer whales. Salmon
are also the summer mainstay for eagles and many species of gulls.
Spawning adults in the streams constitute almost 100% of the summer
diet for bear and some land otter and are a very important link
between the marine and terrestrial ecosystems. Salmon carcasses in
streams, estuaries, and lakes are a crucial source of nutrients for
planktonic communities and benthic organisms which represent the
bottom rungs of the food chain for a wide variety of animals.

Various fish and shellfish species are also important components of
human subsistence, commercial and sport fishery harvests.
Communities such as Tatitlek, Chenega Bay, and English Bay depend
upon subsistence fisheries in PWS and LCI for the very existence of
their residents. The ex-vessel value of commercial fish and
shellfish catches within PWS and other affected areas was estimated
to be $1.3 billion in 1988. The largest recreational fisheries in
Alaska for salmon, halibut, and rockfish center in Homer and
Seward; a total of 300,000 angler days was recorded from these
areas in 1987. Finally, many non-consumptive users of fish and
wildlife also utilize the waters affected by the oil spill. Injury
to fish and shellfish populations and resulting alterations to
ecological communities would certainly diminish the value of the
area to this group of people.

Bioassays prior to EVOS using crude oil from Prudhoe Bay and other
areas have shown that exposure to concentrations as low as a few
parts per billion in seawater will cause loss of limbs in Tanner
crab, immediate death of eggs and larvae of herring, and death of
Dungeness crab and various shrimp species. To assess the type and

extent of injury done to marine fish and shellfish communities by
the EVOS, a series of Fish/Shellfish (FIS) studies was developed by
investigators from various State and Federal agencies. Species
were selected for study based on their value as indicators of
damage, their role as key species within the ecosystem, or their
direct importance to man as components of subsistence, commercial
or sport harvests.

Comparisons of the abundance of larvae, juveniles, or adults
between oiled and non-oiled waters were chosen as the basic
experimental units. In some studies, oiled and non-oiled waters
pertain to different geographic areas; in other studies these terms
relate to the same area or populations before and after the oil
spill; in the remaining studies these terms refer to different
areas and populations before and after the spill. Contamination of
individual fish and shellfish will be documented by analysis of
tissue samples, bile samples, and/or testing for induction of
specific enzymes associated with hydrocarbon exposure. Damages to
fish and shellfish populations resulting from the oil spill may be
expressed as lethal (e.g., mortality to specific life history
stages) or sublethal (e.g., decreased growth, reproduction
potential, etc.) injuries. Such injuries to populations could
cause losses in harvests and use of these species by man, and
result in undesirable alterations of natural communities which
might be difficult to restore.

Project proposals were reviewed and modified through input provided
by State and Federal agency staff members, State and Federal
attorneys, various experts retained by the State and Federal
governments, and many corporate and private individuals. Based on
these inputs and results from first year studies, a number of
changes were made for the 1990 fisheries program. salmon studies
FIS 1, 2, 3, 4, 7, and 8 were continued, as modified, another year
while F/S 9 (Early Marine Studies outside PWS) which could not be
initiated in 1989 was not approved for 1990. Dolly Varden and
cutthroat trout study FIS 5 which was conducted in 1989 was
approved for continuation, as modified, in 1990. The sport fish
harvest and effort study (F/S 6),conducted in 1989 was not approved
for continuation in 1990. A study on Dolly Varden and sockeye
salmon in LCI (F/S 10) which was approved for 1989 but could not be
implemented was not approved for 1990. The herring study (F/S 11)
was expanded and modified considerably from that of 1989 and
approved for continuation in 1990. Herring studies outside PWS
(F/S 12) were completed in 1989 and not proposed for continuation
in 1990. Clam study F/S 13 was combined with FIS 21 and approved
for continuation, as modified, in 1990. The crab study within PWS
(FIS 14) conducted in 1989 was not approved for continuation in
1990. The crab study outside PWS (F/S 22) was conducted in 1989
and was approved for continuation in 1990. The spot shrimp study
(F/S 15) was conducted in 1989 and was approved for continuation in
1990 with few modifications. The rockfish studies FIS 17 and 23
were combined and modified for continuation in 1990. Multi-species

trawl surveys, FIS 18 and 24 were conducted in 1989 and modified
considerably for continuation in 1990. The oyster study (FIS 16),
larval fish study (F/S 19), underwater observations (FIS 20),
scallop mariculture study (FIS 25), and the sea urchin study (F/S
26) were not approved for continuation in 1990. Three new studies
not conducted in 1989 were approved for implementation in 1990,
these being sockeye salmon over-escapement (FIS 27), salmon run
reconstruction (FIS 28), and salmon data base management (FIS 30).

FISH/SHELLFISH STUDY NUMBER 1

Study Title: Injury to Salmon Spawning Areas in PWS

Lead Agency: ADF&G

INTRODUCTION

Wild stock production of pink salmon in PWS has ranged from 10 to
15 million fish in recent years. Chum salmon returns have ranged
from 800,000 to 1,500,000. Much of the spawning for pink and chum
salmon (up to 75% in some years) occurs in intertidal areas.
Intertidal spawning areas are susceptible to marine contaminants
and the March 24, 1989, EVOS may adversely affect spawner
distribution and success in Prince William Sound. To detect injury
to pink and chum salmon stocks, intertidal contamination will be
documented and correlated with trends in adult returns. Return
estimates are based on accurate appraisals of catch and
escapements. This project is designed to document oil
contamination of intertidal spawning habitat; provide accurate
.estimates of escapements of wild stocks; and provide estimates of
the intertidal and upstream area available for spawning. F/S Study
3 provides estimates of the wild stock component of the commercial
catch. Results from F/S Study 3 and this study will be combined to
estimate total return of wild stocks. F/S Study 2 estimates eggs
and fry per square meter and egg to fry survival by tide zone in a
subset of the streams in this study. Egg and f ry density and
survival data from F/S Study 2 will be combined with stream bed
area estimates by tide zone from this study to estimate total egg
deposition and egg to fry survival by tide zone in 138 streams.

The ADF&G has performed spawning ground surveys of the major salmon
spawning streams in PWS since the late 1950's. An aerial survey
program provides weekly estimates of numbers of fish in 218
spawning streams and a ground survey program on a subset of
approximately 116 has provided corresponding estimates of numbers
during the peak of spawning. During 1987 and 1988, funding for the
ground survey program was severely curtailed and only 58 streams
were walked. This study includes a thorough and extensive ground
escapement survey program on salmon spawning streams for which
there are past ground survey data and includes additional oiled and
unoiled streams in western PWS. The study also includes ground
surveys of salmon streams to document the presence of oil in
intertidal spawning habitat.

In 1989 a total of 411 streams were surveyed for the presence of
oil in intertidal spawning areas and 138 streams were included in
the ground census of pink and chum salmon escapements. In 1990 the
oil survey will be limited to the 138 streams in the escapement
censuring portion of the project. The total area of intertidal and

upstream spawning habitat will be estimated for each stream.
Estimates of stream residence time (stream life) will be made for
pink and chum salmon in 11 of the 138 streams.

The results of the study will provide accurate estimates of the
pink and chum salmon escapement to each stream surveyed; will be
correlated with escapement estimates based on aerial counts to
estimate past and current year escapements for 218 streams included
in the ADF&G aerial survey program; will provide estimates of post
oil spill distribution of spawning within stream zones and among
streams; will estimate total available intertidal and upstream
spawning habitat for each stream; will estimate the average stream
life for pink and chum salmon in PWS; and will provide an atlas of
aerial photographs and detailed maps for important spawning sites.

OBJECTIVES

A. Determine the presence or absence of oil on intertidal habitat
used by spawning salmon through visual observation, aerial
photography, and hydrocarbon analysis of tissue samples from
intertidal mussels at stream mouth.

B. Document the physical extent of oil distribution on intertidal
spawning areas.

C. Estimate the number of spawning salmon, by species, within
standardized intertidal and upstream zones for 138 streams in
PWS.

D. Enumerate the total intertidal and upstream escapement of pink
and chum salmon through weirs installed on one or more
moderately large streams which are representative of streams
in the aerial and ground escapement survey programs.

E. Estimate the accuracy of aerial counts for the 218 aerial
index streams by comparison of paired ground and aerial counts
from 138 of the streams on the same or adjacent survey dates
and by comparison of aerial, ground, and weir counts on one
stream.

F. Estimate the average stream life of pink and chum salmon in at
least 11 streams in PWS using a variety of techniques.

G. Estimate 1961 through 1988 pink and chum salmon escapements to
the 218 aerial index streams using the average observed error
in the aerial survey method and on stream life data from 1989
and 1990.

H. Estimate the stream area available for spawning within
standardized intertidal and upstream zones for the 138 streams
surveyed.

I. Produce a catalog of aerial photographs and detailed maps of
spawner distribution f or the more important pink and chum
salmon streams of Prince William Sound for use in designing
sampling transects in the egg deposition and pre-emergent fry
studies.

J. Identify streams appropriate for enumerating and CWT pink
salmon fry.

METHODS

This project is an integral part of the study of impacts of the
EVOS on Pacific salmon populations in PWS. Streams examined by
this project are a subset of the anadromous salmon streams
monitored by the ongoing ADF&G aerial survey program. Two
additional F/S studies in PWS, pink and chum salmon egg deposition
and pre-emergent fry studies Study 2 and salmon coded-wire tagging
studies Study 3, will rely on information about salmon spawning and
distribution obtained from this project.

Three crews of two people each will perform foot surveys of
intertidal and upstream portions of 138 major pink and chum salmon
spawning streams. Each stream will be visited once prior to the
salmon returns to measure and mark tide levels and survey
intertidal areas in and adjacent to the stream for presence of oil.
Live and dead pink and chum salmon will be enumerated by ground
survey crews in standardized intertidal and upstream zones in each
stream. Streams will be enumerated three times at approximately
two week intervals'during the spawning season.

Streams to be surveyed will be selected according to the following
criteria:

1. Stream is included in the ADF&G aerial survey program.
2. Stream is included in the pink and chum salmon egg deposition
and pre-emergent fry project (F/S Study 2).
3. Stream was enumerated in prior spawning ground foot survey
programs.
4. Stream is representative of the early, middle, and late run
pink and chum salmon stocks in PWS.
5. Stream is representative of the spatial distribution of pink
and chum salmon stocks in PWS and include streams from oiled
and unoiled areas.

Maps of all streams in the program prepared from aerial photographs
prior to the 1989 field season and modified and corrected during
the three survey circuits in 1989 will be used and updated during
the 1990 field season.

The pre-season survey to mark tide zones and document the presence
of oil in the intertidal area at the stream mouth will be conducted
in June, prior to the return of the pink and chum salmon. The
location of tide levels 1.8f 2.4, 3.0, and 3.7 m above mean low
water will be measured from sea level using a surveyors's level and
stadia rod. Sea level at each site will be referenced to mean low
water with site specific, computer generated tide tables which
predict tides at five minute intervals. Tide zone boundaries will
be delineated with color coded steel stakes. The linear length of
the stream within each intertidal zone will be measured with a
surveyors chain or range finder. The linear length of the stream
in the upstream zone will be measured similarly on short streams
and estimated from accurately scaled aerial photos on long streams.
The average stream width will be determined from systematic width
measurements taken in each zone. The number of intervals in each
zone will depend on the length of the zone. Each measurement will
be recorded at the appropriate location on the stream maps prepared
in 1989.

Crews marking, measuring, and mapping tide zones will also conduct
foot surveys of the intertidal stream bed and adjacent beaches to
document, map, and classify any oil present.

During the escapement enumeration portion of the project, streams
will be surveyed visually from the ground in a systematic order.
During each stream survey the following data will be recorded:

- anadromous stream number and name (if available);
- latitude and longitude of the stream mouth;
- date and time (24 hour military time);
- tide stage;
- observer names;
- counts of live and dead salmon by species and tide zone (0.0-1.8
m, 1.8-2.4 m, 2.4-3.0 m, and 3.0-3.7 m above mean low water and
upstream); and
- weather and comments on visibility, lighting, and other survey
conditions.

All data will be recorded on pre-printed mylar data sheets which
will overlay a map of the stream. Maps will be improved and
modified during the survey to show spawner distribution within each
zone and the upstream limit of spawning. Particular attention will
be given to spawner density and distribution observations for the
48 streams to be sampled during FIS Study 2.

During the first survey circuit, a composite sample of mussels will
be collected at the mouth of each stream for hydrocarbon analyses.
Results of the analyses will be used to document the level of oil
impact that the stream sustained. Each sample will consist of
enough mussels to provide 10 grams of tissue (approximately 30
mussels) for analysis. The mussels will be collected in the zone
from 0-2 m above mean low water in the immediate vicinity of each

stream mouth and will be collected above water to avoid
contamination by hydrocarbons on the water surface. The samples
from- each stream will be stored in separate, properly cleaned,
glass jars with teflon lined lids. Appropriate chain of custody
forms will accompany each sample.

During all three circuits counts of live and dead salmon will be
made for the five tide zones (the intertidal zones < 1.8 m, 1.8-2.4
m, 2.4-3.0 m, 3.0-3.7 m above mean low water and the upstream zone)
from the 1.8 m tide level to the limit of upstream spawning on all
138 streams. Tide stage will be monitored continuously and survey
times and direction will be adjusted accordingly. If the tide
stage at the time of the walk is at or below the 1.8 m level the
stream walk will begin at the mouth of the stream and progress
upstream. The mouth or downstream limit of the stream will be
defined as the point where a clearly recognizable stream channel
disappears or is submerged by salt water. Fish seen below the
downstream limit will be included in an estimate of fish off the
stream mouth and noted as a comment on the data f orm. If the
intertidal portions of the stream above the 1.8 m level are
submerged at the time the walk begins, the crew will proceed to the
upstream limit of the walk, walk downstream, and coincide the end
of the walk with the time predicted for the tide to be at or below
the 1.8 m level. The upstream limit of a walk will be determined
by the presence of natural barriers to fish passage (i.e.
waterfalls), by the end of the stream, or by the upstream limit of
spawning. The upstream limit of spawning will be marked on U.S.
Geological Survey color aerial photos of each stream following each
survey.

For counts of live and dead fish on moderate size streams with a
single channel, crew members will walk together but independently
count live fish in each intertidal zone. Crew members will
individually enter their count on mechanical hand tallies. A
maximum of three replicate counts may be made in each zone at the
request of either observer. If the two counts differ by more than
10%, the zone will be recounted until counts differ by 10% or less.
Upstream counts in a single channel will be similarly conducted at
convenient stopping points (i.e., log jams or other clear counting
delineators) . To avoid confusion with counts of live f ish, counts
of dead fish will be recorded on the return leg of the stream walk.
For large braided or branched streams, each crew member will count
separate channels or upstream forks. To avoid perpetuating
counting biases within a counting crew, personnel will be rotated
between crews daily. When possible, crew members will not be
assigned to the same streams on succeeding survey circuits.

Tests for variability among observers and among counting crews
(observer pairs) will be conducted on 10 streams during each of
the three enumeration survey circuits. At test streams, all
observers will estimate numbers of live and dead pink salmon by
zone and will record their counts independently. Counts will be

compared after all test streams have been surveyed. Three crews of
randomly paired observers will also replicate counts on 10 stream
and results among observed pairs will be compared.

At 11 of the 138 streams in the ground survey program, fish will be
include in a stream lif e study. Stream life studies will be
modeled in part after previous studies in PWS ( .Helle et al. 1964;
McCurdy 1984). For each stream will be captured at the stream
mouths with beach seines and tagged with individually numbered
Peterson disk tags color coded for day of capture. Tagging will be
conducted at weekly intervals at each stream and during each
tagging episode, 80 fish will be tagged. If fewer than 80 fish
are available, all fish captured will be tagged. Daily counts of
live and dead pink and chum salmon will be made by tide zone in
each of the 11 streams. Live and dead fish bearing tags will be
enumerated separately by color code and tag n umber. Only fish that
have died since the previous count will be tallied in the daily
surveys. To prevent duplicate counts between surveys, tails and
tags of all dead pink and chum salmon observed will be removed.

At least one moderately large stream from among the 138 surveyed
will be weired. The weir will be installed at the six foot tide
level or the lower level of intertidal spawning. Weir crews will
be record daily passage through the weir and will tag a portion of
each day's escapement. Tags will be numbered sequentially and
unique color codes will be used for each weekly interval. The weir
crew will survey the intertidal and upstream portions of the stream
daily for stream life data. Counts of live and dead pink and chum
salmon will be made by tide zone. Fish bearing tags will be
enumerated by color code and the sequential number on tags from
dead fish will be recorded. only f ish that have died since the
previous count will be tallied in the daily surveys. To prevent
duplicate counts between surveys, tails and tags of all dead pink
and chum salmon observed will be removed

Steams will be divided into categories based on levels of
hydrocarbon contamination (as determined from visual observations
species, stream zone, and stream for each will be assigned to one
of the categories. Categorical data analysis techniques such as
log linear models using chi-square statistics will be used to
compare differences in spawning among streams and tide zones, and
related these disruptions to the level of hydrocarbon
contaminations. Count and spawner distribution data will also be
compared with historical stream survey data and related to the
level of hydrocarbon impact.

Steam life will be estimated using three methods. The f irst
estimate is the mean difference between date of tag recovery from
dead f ish and the tagging date and the estimate will be made of the
color coded tag lot to examine changes in stream lif e through
times. The second estimate will use weir data and will be based on
similar data for numeric tag codes from individual fish. The third

method will be based on the difference in the dates between peak
life count and the peak dead count.
I

BIBLIOGRAPHY

Helle, J.H., R.S. Williamson, and J.E. Baily. 1964. Intertidal
ecology and lif e history of pink salmon at Olsen Creek, Prince
William Sound, Alaska. U.S. Fish and Wildlife Service,
Fisheries No. 483. Washington D.C.

McCurdy, M.L. 1984. Eshamy District pink salmon streamlife study,
1984. Alaska Department of Fish and Game, Division of
Commercial Fisheries. Prince William Sound Data Report No.
94-18. Cordova

BUDGET: ADF&G

salaries $ 179.2
Travel 2.0
Contractual 145.3
Commodities 25.0
Equipment 40.0

Total $ 391.5

FISH/SHELLFISH STUDY NUMBER 2

Study Title: Injury to Salmon Eggs and Pre-Emergent Fry in PWS

Lead Agency: ADF&G

INTRODUCTION

Much of the spawning for pink and chum salmon (up to 75% in some
years) occurs in intertidal areas. Moles, Babcock, and Rice
(1987) have shown the adverse effects of oil on pink salmon
alevins, particularly in saltwater. The EVOS in PWS occurred
immediately prior to emergence of pink and chum salmon from
stream and intertidal spawning areas. Obviously, these areas
have the potential to be severely impacted by the oil spill.

This study along with FIS Studies I and 3 support a comprehensive
and integrated determination of injury to PWS salmon stocks.
Results will include documentation of oil in intertidal salmon
spawning habitat, pre-spill and post-spill estimates of total
adult returns of wild and hatchery stocks, wild stock spawning
success, wild stock egg to fry survival, and early marine
survival of wild and hatchery stocks. Information on the extent
and persistence of oil in the intertidal zone will be
supplemented by Coastal Habitat Study 1. The results of FIS
Studies 1 through 3 will be used by Economic Uses Study 3 to
determine the extent of damage to the,Prince William Sound salmon
resource.

The ADF&G has sampled pink and chum salmon pre-emergent fry since
the 1960's in order to predict the magnitude of future salmon
returns. The fry dig program has operated at a reduced level
since 1985. The oil spill has the potential to cause mortality
to the critical egg and fry life stages and thus an increased and
more comprehensive fry dig program is necessary. This project is
designed to meet this need by assessing the effect of the oil
spill on egg and fry of wild stock pink and chum salmon.

OBJECTIVES

1. Estimate the density of pink and chum salmon eggs (31
streams) and pre-emergent fry (48 streams) by tide zone in
study streams.

2. Estimate over-winter mortality of pink and chum salmon eggs
in oiled and control streams based on sampling of 31 natural
streams.

3. Assess reductions in adult returns (if any) associated with
increased egg to fry over-winter mortality in oiled streams.

4. Document hydrocarbon contamination using tissue
concentrations of hydrocarbons in alevins and mussels, -and
mixed function oxidase (MFO) levels in alevins and eggs from
study streams.

METHODS

There are approximately 900 anadromous fish streams in PWS. Pre-
emergent fry sampling from some of these streams has historically
provided an abundance index for pink salmon which is used to
forecast future pink salmon returns. In recent years, 25 index
systems considered representative of pink and chum salmon
producing streams in PWS have been sampled during the fry dig
program. Prior to 1985, sampling had been performed on as many
as 45 streams. This study is designed to compare rates of
mortality and abundance between areas with various levels of oil
impacts and with data from sampling prior to the oil spill.

Sampling will consist of egg-digs performed in late September and
early October, and pre-emergent fry digs conducted in mid-March
to mid-April. Preliminary sampling was performed on two
occasions during the spring of 1989 in an effort to assess fry
abundance prior to and immediately after oil impact. on the
first occasion the 25 streams in the ongoing ADF&G pre-emergent
index program were sampled along with 14 additional streams.
During the second event (approximately two weeks after the oil
spill), 14 of the streams were resampled (representing both oiled
and non-oiled areas) and an additional 16 streams were surveyed
to assess their potential as egg and pre-emergent study streams.
During September and October of 1989 egg digs were conducted on
31 of these streams.

spring fry digs in 1990 will be conducted on 48 streams. These
will include the 25 streams in the ongoing ADF&G pre-emergent
index program plus 23 additional streams. The additional streams
are located in Central to Southwest PWS where the majority of the
oiling occurred. New study streams were selected using the
following criteria:

sufficiently large adult salmon returns to indicate a
high probability of"success in egg/fry digging;
past history of egg/fry digging; and
streams which had low to no oil impact in the immediate
vicinity of high oil impact streams. This will help
account for possible variability due to differing
climatic/stream conditions.

The 48 streams span a range of oil impact and include streams in
the historic sampling program. Most of the streams with
suspected or obvious oil impact are new additions. The 30
streams in low impact areas include 27 with a history of

sampling, six streams suspected of having received some impact
including four with a history of sampling, and 12 streams with
oil visibly present in the intertidal zone, including five with a
history of sampling.

As in 1989, egg digs will be conducted in the fall on a subset of
the 48 streams sampled for pre-emergent fry. Streams included
in the fry sampling program, but not the egg sampling, are
traditional fry sampling streams located on the eastern and
northern shore of PWS. These streams are spatially quite
distinct from the streams studied for oil impact effects. The 13
streams in low impact areas which were left in the egg dig
program include four with a history of sampling. The streams
suspected of having experienced some impact and the streams where
impact is visibly obvious are the same as in the fry sampling
program.

Sampling methods are identical for the pre-emergent fry and egg
digs and are modeled in part after procedures described by Pirtle
and McCurdy (1977). On each sample stream, four zones, three
intertidal and one above tidal influence, will be identified and
marked by crews conducting stream surveys under F/S Number 1
(PWS). The zones are 1.8-2.4 m, 2.4-3.0 m, 3.0-3.7 m above mean
low water, and upstream of tidal influence. Separate linear
transects 30.5 m in length will be established for the egg and
pre-emergent fry digs in each zone (one transect for each type
dig in each zone). The transects will run diagonally across the
river with the downstream end located against one bank and the
upstream end against the opposite bank. Overlapping of transects
will be kept to a minimum to control the influence of fall egg
digs on perceived abundance of fry during spring 2 sampling.
Fourteen circular digs (56 per stream), each 0.3 m. will be
systematically dug along each transect using a high pressure hose
to flush eggs and fry from the gravel. Eggs and f ry will be
caught in a specially designed net.

Numbers of live and dead fry by species as well as numbers of
live and dead eggs by species will be collected from each 0.3 m2
dig. Additional information such as date, time, zone, and a
subjective estimate of overall percent absorption of the fry egg
sacs in the sample will also be collected.

Tissue samples from pre-emergent pink salmon fry will be
collected from the intertidal channels of streams. Tissue samples
will be analyzed for the presence of hydrocarbons characteristic
of those found in oil from the Exxon Valdez.

Fry sampled for hydrocarbon analysis will be from the intertidal
stream bed at a level approximately 2.5 m above mean low water.
Samples will be collected when the tide stage is below that level
to avoid contamination from any surface film of oil on salt
water. A shovel or clam rake will be used to dislodge the fry

f rom the gravel and a stainless steel strainer which has been
pre-rinsed in dimethylchloride and dried, will be used to catch
the fry as they are swept downstream. Captured fry will be placed
in jars with teflon lined lids and frozen.

Fry from each tide zone will be collected for MFO analysis and
these samples will be selected systematically from the digs in
each transect. Sampled fry will be preserved in buffered
formalin solution in glass jars.

Pre-emergent fry/egg data will be summarized by date, stream,
level of hydrocarbon impact, stream zone, and number of alive and
dead, fry and eggs. A mixed effects analysis of covariance will
be used to test for differences in egg to fry mortality due to
oiling using the 31 streams sampled for both eggs and pre-
emergent fry. Degree of oiling and height in the tidal zone will
be treated as fixed effects. Height in the tide zone is nested
within stream, a random effect. Possible covariates will be
provided by hydrocarbon analysis of mussel populations in close
proximity to each stream.

If no suitable hydrocarbon data are available, analysis of
variance will be used. Degree of oiling as visually assessed by
the mapping portion of the assessment of intertidal spawning
areas will be used to post-stratify streams. Degree of oiling
and height in the tidal zone will again be treated as f ixed
effects. Height in the tidal zone is nested within streams, a
random effect.

Power of the test was estimated for the analysis of variance
using data from the 1975 and 1976 egg and pre-emergent fry digs
in PWS. This study indicated the ability to detect an increase
of 15% (e.g. 10% mortality to 25% mortality) in egg to fry
mortality at = 0.05, 95% of the time.

These studies will be used to test for 1) differences in egg to
fry mortality between streams which were oiled and those that
were not, and 2) increases in fry mortality in 1989 immediately
after oiling.

Specific statistics to be estimated are:

number of dead and viable eggs per square meter by
salmon species, stream, and stream zone;
number of dead and live fry per square meter by salmon
species, stream, and stream zone; and
egg to fry survival by salmon species, stream, and
stream zone.

BIBLIOGRAPHY

Moles, A., M.M. Babcock, and S.D. Rice. 1987. Effects of oil
exposure on pink salmon, 0. gorbuscha, alevins in a
simulated intertidal environment. Marine Environment
Research, 21:49-58.

Pirtle, R.B. and M.L. McCurdy. 1977. Prince William Sound
general districts 1976 pink and chum salmon aerial and
ground escapement surveys and consequent brood year egg
deposition and pre-emergent fry index programs. Alaska
Department of Fish and Game, Division of commercial
Fisheries, Technical Data Report 9, Juneau.

BUDGET: ADF&G

Salaries $ 120.0
Travel 4.0
Contractual 150.0
Commodities 10.0
Equipment 18.8

Total $ 302.8

FISH/SHELLFISH STUDY NUMBER 3

Study Title: Salmon Coded-Wire Tag Studies In PWS

Lead Agency: ADF&G

INTRODUCTION

Two questions must be answered to measure a loss in salmon
production due to EVOS: 1) which stocks were exposed to
contaminated waters and 2) to what extent did exposure reduce
production (catch plus escapement)? This study will contribute to
estimates of production and survival for hatchery and wild stocks
in oiled and unoiled areas by quantifying fry outmigration, the
adult component of the catch, and the escapement to hatcheries.

Wild stock returns of pink salmon in PWS have ranged from 10 to 15
million fish in recent years. Chum salmon returns have ranged from
800 thousand to 1,500,000. Additionally, returns of pink salmon to
four PWS hatcheries now average more than 20 million fish and
hatchery chum salmon returns in excess of 1.4 million fish are
expected.

Catch and escapement data for wild pink salmon in PWS have been
collected since 1961. In 19851 hatchery production became a
significant part of the total salmon return. Consequently, pink
salmon fry tagging was initiated at three area hatcheries in 1986
to estimate the survival of those stocks and their contribution to
the 1987 catch. Similar estimates were made for a fourth facility
based on tagging in 1987 and recoveries in 1988. FIS Study 3
estimated catch and survival rates of pink salmon released from
these four PWS hatcheries based on tags applied in 1988 and
recoveries of tags in the commercial, cost recovery and hatchery
brood stocks in 1989. Tags were also applied to pink, chum,
sockeye, coho, and chinook salmon releases from PWS area hatcheries
and to smolts from two wild stocks of sockeye salmon. Tagging in
1990 will include all the same stocks plus one more wild stock of
sockeye salmon and six pink salmon wild stocks. Tag recoveries are
expected for releases at all four pink salmon hatcheries in 1989,
releases of chum salmon from Main Bay Hatchery in 1986 and from
Main Bay and Solomon Gulch Hatcheries in 1987, releases of sockeye
salmon from the Main Bay facility in 1988, and releases of coho
salmon from Wallace H. Noeremberg (WHN) and Solomon Gulch
Hatcheries in 1989.

OBJECTIVES

1. Estimate catch, escapement, and survival rates of pink, chum,
sockeye, coho, and chinook salmon released from five
hatcheries in PWS. Outmigrating smolt and returning adults
from these facilities are exposed to oil in varying degrees.

2. Estimate catch of the combined wild stocks of pink salmon in
PWS and using escapement data from FIS Study 1, estimate
differences in relative survival rates between pre- and post-
spill brood years.

3. Estimate survival rates of wild pink salmon from three streams
with contaminated estuaries and three with uncontaminated
estuaries.

4. Provide marked salmon of known origin and oil exposure history
for recovery by researchers studying early marine migration,
growth, and survival (FIS Study 4).

5. Estimate survival rates of wild stocks of sockeye salmon, two
from oiled areas, one from an unoiled area.

METHODS

A subsample of fry or smolt from all hatcheries releasing salmon
into PWS will be tagged with a coded wire tag (Appendix A). Wild
stock pink fry and sockeye smolt from both oiled and non-oiled
areas of the Sound will also be tagged (Appendix B). Tags will be
applied at rates which will insure that, given a realistic recovery
effortl sufficient numbers can be recovered in the commercial
fishery, hatchery cost recovery harvests, and hatchery brood stock
collections (Appendixes) to allow researchers to estimate the
contribution of each tag release group by district, week, and
processor stratum. Release groups represent differences in release
timing or treatment (i.e. fed vs. unfed fry)

Tag application will be similar among all hatcheries and among all
wild stock systems. Fry or smolt will be randomly selected as they
emerge from incubators or outmigrate from streams, anesthetized in
a 1 ppm solution of MS-222, adipose f in clipped, and tagged. A
random sample of 100 fish will be graded for fin clip quality each
day. The proportion of bad clips in the sample will be used to
discount the daily release of tagged fish. Clipped fish will be
tagged and passed through a quality control device to test for tag
retention. Fish repeatedly rejected will be killed to minimize the
number of untagged but clipped fish in the release. Fish that
retain tags will be held for 24 hours to determine short term
mortality. A sample of tagged fish from each tagger will be taken
each day and graded for tag placement according to criteria

developed by Peltz and Miller (1988) Prior to release, a 200 fish
sample will be randomly sampled to estimate overnight tag
retention. The proportion of lost tags in the sample will be used
to estimate tag retention in the daily release.

At the three Prince William Sound Aquaculture Corporation (PWSAC)
hatcheries, tagged fish will be released directly into large
saltwater rearing pens with untagged f ish of the same release
group. At the Valdez Fisheries Development Association (VFDA)
Solomon Gulch Hatchery tagged fry will be Placed in small
enclosures within larger saltwater rearing pens for at least three
days to allow them to recover from tagging before being mixed with
unmarked fry from the same release group. At PWSAC hatcheries,
unmarked f ry entering large pens were counted with Northwest Marine
Technology counters. At VFDA, unmarked f ry in each pen will be
estimated from counts of eggs in incubators minus egg mortalities.
At all facilities, mortalities in the large pens will be estimated
visually prior to release. Mortality rates based on visual
estimates will be applied equally to tagged and untagged fish. The
total number of f ish in group t with valid tags at the time of
release will be estimated as

Tt (Tt - mt) - (Tt - Mt) Lt,
where Tt total number of fish tagged from group t,
Mt overnight mortality among fish tagged from
treatment group t,
Lt overnight tag loss among fish tagged in
treatment group t.

The VFDA estimate includes a term f or short term mortality of
tagged fish from treatment group t during saltwater rearing (st).
The number of tagged fish released becomes

Tt (Tt - Mt - St) - (Tt - Mt - St) Lt.

Hatcheries will release fry when plankton monitoring indices
indicate peak zooplankton abundance.

Four hatcheries released 13 groups of pink salmon in 1989. Only one
of these groups was not tagged. Each of the hatchery pink salmon
tag groups contained tagged fish at the rate of approximately one
tag per 570 fish released. The tag rate was held constant across
release groups to prevent confusion of differential tag mortality
with variation in survival between release groups (Peltz and
Geiger, 1988; Geiger and Sharr, 1989). In 1989, chum salmon were
tagged at the rate of approximately one tag per 60 f ish released at
the Solomon Gulch Hatchery near Valdez.

In 1990, hatchery pink and chum salmon tagging will continue at the
same level of ef fort with the addition of chum salmon at the Esther
Island Hatchery; approximately 250,000 of these chum salmon will be

tagged in one release group.

Wild pink salmon will be tagged from six stocks examined in FIS
Study 2. Fry will be captured as they emerge using various means.
The fry will be anesthetized with MS-222 and tagged with Northwest
Marine Technology tagging equipment and tags. The anesthesia and
associated trauma will require that the tagged fish be held
separate from their untagged cohorts, until they appear to have
fully recovered from the effects of tagging. The extent to which
the survival and behavior of the tagged fish can be extrapolated to
other groups of salmon will be assessed at the time of recovery.

Prior to tagging, hatchery chinook and coho salmon smolt in
hatcheries will be crowded using seines. A sample of smolt will be
drawn from each rearing appliance in approximate proportion to the
number of fish in that appliance. They will be anesthetized with
MS-222, their adipose fin excised, and a tag applied using
Northwest Marine Technology equipment and tags. A sample of fish
from each day's tag production will be retained to estimate short-
term tag loss and tag induced mortality. Following tagging, the
tagged fish will be returned to mix with untagged cohorts. All
mortalities during the first week after tagging will be examined
and the tag status noted. At the end of a week, the fish will again
be crowded, and a sample of approximately 2,000-4,000 fish from
each rearing appliance will be drawn. These fish will be
anesthetized, and run through a tag detector. Peterson abundance
estimates for all rearing appliances will be performed and any
major discrepancies from hatchery inventory records noted. Finally,
a written description of the tagging will be developed. This will
include a detailed description of each tag lot, the number of fish
tagged, the total number of fish in the release lot, the average
size of the fish at release, a profile of the exposure history of
the release lot to the oil spill, and all information required by
the ADF&G Coded-Wire Tag Laboratory which coordinated tagging in
Alaska.

In 1989 wild sockeye salmon were tagged at Eshamy and Coghill
Lakes. Smolt were captured in traps as they migrated to saltwater.
The smolts were anesthetized with MS-222 and tagged with Northwest
Marine Technology tagging equipment and tags. The anesthesia and
associated trauma required that the tagged fish be held separate
from their untagged cohorts until they appeared to have fully
recovered from the effects of tagging. As in the wild pink salmon
tagging, the extent to which the survival and behavior of the
tagged fish can be extrapolated to other groups of salmon will be
assessed at the time of recovery. The rate of tag occurrence in the
stock will be determined from counts at an adult salmon weir in
each of the systems. All fish passing through the weirs will be
enumerated and heads from fish with adipose fin marks will be taken
at the weir for tag removal and decoding. Hatchery produced sockeye
salmon smolts will be tagged using the methods described for
chinook salmon above.

The recovery samples are from a stratified sample (Cochran 1977),
by district and discrete time segments. The recovery will be
further stratified by processor as described in Peltz and Geiger
(1988). For each time and area specific stratum, 15% of the pink
salmon catch and a minimum of 20% of other salmon species catches
will be scanned for fish with a missing adipose fin. Catch sampling
will be done in four fish processing facilities in Cordova, one
facility in Seward, and three facilities in Valdez. When feasible,
sampling will occur at facilities in Kodiak, Kenai, Anchorage, and
Whittier and on large floating processors. All deliveries by fish
tenders to these facilities will be monitored by radio and by daily
contact with processing plant dispatchers to ensure that the catch
deliveries being sampled are district specific.

In addition to catch sampling at the processing facilities,
approximately 15% of the fish in the hatchery terminal harvest
areas will be scanned for fish missing adipose fins. There will be
a brood stock tag recovery effort at each of the three hatchery
facilities where tags were initially applied. A minimum of 50% of
the daily brood stock requirements of each facility will be scanned
for fish with missing adipose fins. Finally, there will be an
intensive survey of adult pink salmon returning to natural systems
where tagging was conducted, and a weir will be operated for
sampling adult sockeye salmon on those systems where sockeye salmon
were tagged.

In the catch, terminal harvest, brood stock, and natural system
surveys, the total number of fish scanned and the total number of
fish with missing adipose fin will be recorded. The heads will be
removed from fish with missing adipose fins. Each head will' be
tagged with uniquely numbered strap tags. Recovered heads will be
assembled and pre-processed in the Cordova area office. Heads will
then be sent to the FRED Division Coded-Wire Tag Laboratory in
Juneau for decoding and data posting.

A statewide coded-wire tag lab is located in Juneau and operated by
FRED Division of ADF&G. Coded-wire tag sampling forms will be
checked for accuracy and completeness. Sampling and biological data
will first be entered onto the laboratory's data base. Next, the
heads will be processed. This involves removing and decoding the
tags, and entering the tag code and the code assigned in the
recovery survey into the database. Samples will be processed within
five working days of receipt.

The first step in the coded-wire tag analysis will be to estimate
the harvest of salmon from each tag lot, in units of adult salmon.
Adult salmon from these tagged lots will be recovered in the common
property fishery, the hatchery cost recovery fishery, and the adult
brood stock. For the hatchery stock, a modification of the methods
described in an ADF&G technical report by Clark and Bernard (1987)
will be used. The specific methods, estimators, and confidence
interval estimators are described in ADF&G technical reports on two

previous studies of pink salmon in PWS: Peltz and Geiger (1988),
and Geiger and Sharr (1989). Additional references on methods of
tagging pink salmon in PWS can be found in Peltz and Miller (1988).
In the case of the wild stocks, the methods and estimators and
necessary assumptions are described by Geiger (1988).

The contribution of a particular tag lot, to a particular fishery
stratum, is estimated multiplying by the number of tags recovered
in the structured recovery survey, by both the inverse of the
proportion of the catch sampled (the inverse sampling rate) , and by
the inverse of the proportion of the tag lot that was actually
tagged (the inverse tag rate) . The escapement (brood stock) of each
tag lot is estimated using methods unique to the particular
situation. After the contribution to each fishery is estimated for
the tag lot, the survival is estimated by summing the estimated
harvest of the tag lot in each fishery, and the estimated
escapement (brood stock), and dividing by the estimated number of
fish represented by the tag code.

Total catches stratified by week, district, and processor were
obtained f rom summaries of f ish sales receipts (f ish tickets)
issued to each fisherman. The total hatchery contribution to the
commercial and hatchery cost recovery harvest is the sum of the
estimates of contributions in all week, district, and processor
strata:

Ct zi Xti ( Ni / Si ) pt-
where: Ct = catch of group t fish,
Xti = number of group t tags recovered in ith strata,
Ni = number of fish caught in ith strata,
Si = number of fish sampled in ith strata,
pt = proportion of group :t tagged.

For sampled strata, we used a variance approximation which ignores
covariance between release groups (Geiger 1988):
V (Cd ZiXtj(Nj/Sipt)2[j - (Ni/Sipt)-'].

The average tag recovery rate for all processors in a week and
district will be used to estimate hatchery contribution in catches
delivered to processors not sampled for that district and week.
Variances associated with unsampled strata will not be calculated.

BIBLIOGRAPHY

Clark, J.E. and D.R. Bernard. 1987. A compound multivariate
binomial hypergeometric distribution describing coded
microwire tag recovery from commercial salmon catches in
southeastern Alaska. Alaska Department of Fish and Game,
Division of Commercial Fisheries, Informational Leaflet 261.

Cochran, W. G. 1977. Sampling Techniques, 3rd ed. John Wiley and
Sons, New York, New York.

Geiger, H.J. 1988. Parametric bootstrap confidence intervals for
estimates of fisheries contribution in salmon marking
studies. Proceedings of the international symposium and
educational workshop on fish-marking techniques. University
of Washington Press, Seattle. In press.

Geiger, H.J. and S. Sharr. 1989. A tag study of pink salmon from
the Solomon Gulch Hatchery in the Prince William Sound
fishery, 1988. Alaska Department of Fish and Game, Division
of Commercial Fisheries. In press.

Peltz, L. and H.J. Geiger. 1988. A study of the effect of
hatcheries on the 1987 pink salmon fishery in Prince William
Sound, Alaska. Alaska Department of Fish and Game, Division
of Commercial Fisheries. In press.

Peltz, L. and J. Miller. 1988. Performance of half-length coded-
wire tags in a pink salmon hatchery marking program.
Proceedings of the international symposium and educational
workshop on fish-marking techniques. University of
Washington Press, Seattle. In press.

BUDGET: ADF&G

Salaries $ 902.0
Travel 21.0
Contracts 667.0
supplies 100.0
Equipment 300.0

Total $1,990.0

Appendix A. Coded-wire tagging goals for hatchery
releases of salmon in PWS, 1990.

Total
Release
Valid Number /Marked Number
Projected Tag Tags to Ratio Number of Tags Tag
Hatchery Species Release Goal order Goal Tag Codes \Code Length

Armin F. Koernig Pink 120,000,000 200,000 255,000 600 8 30,000 Half
1 15,000 Half
Cannery Creek Pink 150,000,000 250,000 277,000 600 7 37,000 Half
1 18,000 Half
Solomon Gulch Pink 125,000,000 208,333 225,000 600 5 45,000 Half

Wally Norenburg Pink 250,000,000 416,667 460,000 600 10 46,000 Half

GRAND TOTAL Pink 645,000,000 1,075,000 1,217,000 600 32 Half

Solomon Gulch Chum 6,000,000 40,000 40,000 150 1 30,000 Half
1 10,000 Half
Wally Norenburg Chum 50,000,000 100,000 100,000 500 4 25,000 Half

GRAND TOTAL Chum 56,000,000 140,000 140,000 400 6 Half

Ft. Richardson
Whittier Coho 100,000 20,000 20,000 5 1 20,000 Full
Cordova Coho 60,000 10,000 10,000 6 1 10,000 Full

Solomon Gulch Coho 1,000,000 30,000 30,000 33 1 30,000 Full

Wally Norenburg Coho 2,000,000 70,000 70,000 29 1 50,000 Full

GRAND TOTAL Coho 3,160,000 130,000 130,000 24 5 Full

Main Say Sockeye 2,500,000 100,000 100,000 25 8 12,500 Order Filled

GRAND TOTAL Sockeye 2,500,000 100,000 100,000 25 8 12,500 Order Filled

Wally Norenburg King 150,000 30,000 30,000 5 1 30,000 Full

GRAND TOTAL King 150,000 30,000 30,000 5 1 30,000 Full

GRAND TOTAL ALL 706,810,000 1,475,000 1,617,000 479

Appendix B. Coded-wire tagging goals f or wild stock of salmon
in PWS, 1990.

Total
Valid Release Number
Projected, Tag /Marked Number of of Tags Tag
System Treatment Species Release Goat Ratio Tag Codes \Code Length

Upper Herring B. Oiled Pink 1,000,000 40,000 25 2 25,000 Half

Hayden Ck. Oiled Pink 1,000,000 40,000 25 2 25,000 Half

Loomis Ck. Oiled Pink 1,000,000 40,000 25 2 25,000 Half

McClure Ck. Clean Pink 1,000,000 40,000 25 2 25,000 Half

O'Brien Ck. Clean Pink 1,000,000 40,000 25 2 25,000 Half

Totemoff Ck. Clean Pink 1,000,000 40,000 25 2 25,000 Half

GRAND TOTAL ALL Pink 6,000,000 240,000 25 12 300,000 Half

Coghill Clean Sockeye 1,000,000 20,000 50 1 20,000 Half

Eshamy Oiled Sockeye 1,000,000 20,000 50 1 20,000 FuLt

Jackpot Oiled Sockeye 200,000 20,000 10 1 20,000 Half

GRAND TOTAL ALL Sockeye 2,200,000 60,000 37 3 60,000 Both

GRAND TOTAL ALL ALL 8,200,000 300,000 27 15 360,000 Both

FISH/SHELLFISH STUDY NUMBER 4

Study Title: Early Marine Salmon Injury Assessment In PWS

Lead Agencies: ADF&G, NMFS

INTRODUCTION

The early marine period is a critical one for salmon because it is
at this time that the greatest mortality is sustained (Parker 1968;
Bax 1983, Hartt 1980; Foerster 1968; Ricker 1976; Nichelson 1986).
Mortality is considered to be inversely proportional to the rate of
growth, since a prolonged juvenile period will result in a pro-
longed vulnerability to predators (Parker 1971; Healey 1982; Taylor
1977; Walters et al. 1978). For a possible exception to this, see
Helle (1980). Therefore, factors that lower normal growth rates
during the early marine period, such as toxic effects of exposure
to hydrocarbons, reduction in prey populations, or increased energy
expenditures associated with the disruption of normal migratory
patterns, could have a strong influence on survival.

Juvenile salmon are especially susceptible to oil toxicity when
first in seawater (Rice et al. 1975; Rice et al. 1984). Sublethal
levels of hydrocarbons can affect metabolism and reduce growth of
juvenile salmon (Rice et al. 1975). Sublethal levels of water-
soluble hydrocarbons can also damage olfactory lamellar surfaces,
conceivably impacting migratory behavior and feeding patterns
(Babcock 1985). Oil can also be toxic to littoral and pelagic
macroinvertebrates (Caldwell et al. 1977; Gundlach et al. 1983).
Thus, mortality, reduction of reproductive potential, or growth
inhibition of prey populations could reduce growth rates of
juvenile salmon, and thus increase their exposure to predation.

During the past decade, five world-class hatcheries have been
established within PWS. These facilities, operated by the PWSAC
and the State of Alaska, produced approximately 535 million
juvenile salmon in 1989. The hatchery contribution represents
roughly half of the total number of juvenile salmon produced in PWS
this year. CWT program marked roughly 1.3 million juvenile salmon
this year. Approximately one in every 1,000 juvenile salmon in PWS
this year was expected to have a CWT. Recoveries of these marked
fish in PWS will play a major role in our assessment of the impact
of the oil spill.

In 1990, the impact assessment will be conducted by the ADF&G and
the National Marine Fisheries Service (NMFS) . Studies conducted by
ADF&G will focus on the impact of the oil on growth and migratory
behavior, and studies conducted by NMFS will focus on pairwise
comparisons of salmonid growth and behavior in oiled and unoiled
nearshore rearing habitats. Sampling will be coordinated to

produce a single cohesive data base of 1) coded-wire tag recoveries
and 2) zooplankton and epibenthos collections with associated
temperature data.

This study emphasizes a coordinated approach to attaining the
objectives. The studies are mainly complementary. A strong effort
is required because of 1) the high ecological and economic value of
the resource and 2) the wide range of habitats utilized by salmon
during the early marine phase.

OBJECTIVES

A. Estimate the effects of oil contamination on abundance,
growth, feeding habits, and behavior of juvenile salmon during
their early marine residence.

B. Describe migration patterns of juvenile salmon relative to
oiled and unoiled areas of western PWS.

C. Estimate hydrocarbon levels in tissues of j uvenile salmon
collected in oiled and unoiled areas in 1989.

D. Determine distribution, abundance, habitat utilization, size
and growth, and feeding habits of juvenile pink and chum
salmon, in order to compare these parameters with 1989
results.

E. Determine if sediment contamination has reduced the abundance
of primary prey species of harpacticoid copepods.

F. Determine if pollution of azoic sediments with hydrocarbons
will inf luence meiof auna colonization, especially harpacticoid
copepods, in terms of species distribution and abundance.

PART I: Impacts of Oil Spill on Migratory Behavior and Growth

The present study is designed to distinguish between the effect of
oil and other factors on growth and migration by resampling fry in
a few areas examined in 1989. It is expected that the major
difference between these areas in 1990 compared to 1989 will be a
lower level of oil contamination. In 1990, the study will focus on
tag lots that will have been released in a period of a week, or
less.

Portions of the 1989 sampling program will be discontinued in 1990
in order to f ocus attention on growth and migrations. Discontinued
studies will include tow net sampling because of low yields, fry
stomach analyses because growth can be determined by less expensive
means, and epibenthic sampling.

OBJECTIVES

A. Compare the growth of CWT salmon captured in oiled and un-
oiled areas in 1989 with fry captured in the same areas in
1990. Determine at the alpha=.05 level whether size and
condition factors are different in CWT fry collected in oiled
and unoiled years.

B. Document the impact of oil on the migratory path and speed of
migration of CWT salmon releases in PWS. Determine at the
alpha=.05 level whether migration speeds and patterns are
different in oiled and unoiled areas and in oiled and unoiled
years.

C. Document the hydrocarbon content of CWT fry collected in 1989.
Determine at the alpha=.05 level whether hydrocarbon content
differs in CWT fry collected in oiled and unoiled areas in
1989.

METHODS

Fry collections will be targeted on tag lots that are released
during a time of 1 week or less. Recovery of these salmon at later
times and in different places will allow relatively accurate
measurements of growth, and reasonable estimates of migration paths
and migration speeds. Approximately 1.3 million tagged fish will
be released. The goal will be to recover approximately 30 tagged
fish in each of several tag lots each sampling time period. Based
on 1989 recovery data, it is expected that during the proposed 6-
week field season, sufficient fry should be collected to evaluate
approximately six tag lots. The recovery effort will be targeted
on tag lots released from Esther and Armin F. Koernig (AFK)
hatcheries, because ('I) a good data set is presently available on
releases from these hatcheries in 1989, (2) the field crew knows
where fry released from these hatcheries can be collected, and (3)
the hatcheries are located in an unoiled area (Esther) and an oiled
area (AFK).

Most collections, especially early in the season, will be made with
a beach seine (modified Auke Bay design) and 1811 diameter dip nets
on two or three beaches in each sampling area. A 120 ft. purse
seine (modified Auke Bay design) will be used in nearshore areas
where the beach seine can not be used. Water temperatures and
salinity will be measured with a salinometer at 0 m and 5 m. Tide
levels and directions will be recorded for each sampling site.

To avoid excessive mortality when large numbers of fish are caught,
fish will be placed in a holding tank until processing is
completed. Lots of approximately 300 ml of fish (measured by
displacement in a I-liter beaker) will be put through a 2-inch
tunnel tag detector (Northwest Marine Technologies) with a small

stream of salt water. When a tag is f ound in the lot, the lot will
be continuously divided until the tagged fish is found. One 300-ml
sample of fry will be sorted immediately to determine species
composition and released. Another sample of approximately 80 fry
will be preserved in buf f ered f ormalin f or later size measurements.
This number should be sufficient to identify different size groups
if they should occur. Remaining fry will be-released.

All tagged fish will be blot-dried, and measured (snout to fork).
Tags on these fish will be read later by the FRED Tag Lab. After
reading the heads will be preserved in 70% ethanol and archived in
case otolith analyses are desired at a later date. Chain of
Custody procedures will be used throughout transfer and storage of
these samples. Untagged fish will be left in buffered 10% formalin
for at least 30 days to standardize shrinkage. These fry will be
rinsed in buffered sea water, blot-dried, weighed and measured.

Analyses will test the null hypothesis of no difference between CWT
pink salmon fry collected in oiled and unoiled areas at the alpha
= 0.05 level. A Chi-square analysis will test the hypothesis of no
difference in presence/ absence of oil among the different sampling
areas.

Analysis-of-variance of CWT fry will help separate the effects of
oil from other variables contributing to variation in growth rates
(change in body weight per unit time) and condition factor. other
variables include year, hatchery of origin, tag code lot, sampling
time, and sampling area. Those variables contributing
significantly to differences in growth and condition will be
further analyzed to assess the relationship. Where applicable
analyses will use tag code lot as a blocking variable to evaluate
effects of oil. In addition, apparent growth rate curves will be
analyzed using untagged fry caught at the same sample areas during
the different sample times.

Migration rates will be calculated using the minimum distance
between release and recovery sites and the average release date for
a given tag lot. Differences in migration rate, distance and
pattern will be analyzed with ANOVA as described in the above
section.

PART II. Impact of Oil Spill on Juvenile Pink and Chum Salmon and
Their Habitat

INTRODUCTION

In 1989, the NMFS component of the early marine salmon studies
focused on pairwise comparisons between oiled and non-oiled study
sites in PWS. The objectives were to determine if oil had affected
distribution, abundance, size and nominal growth ratesi feeding
habits, and prey abundance.

Epibenthic harpacticoid copepods produced in the intertidal and
upper subtidal reaches are an important food resource for juvenile
pink and chum salmon (e.g., Kaczynski et al. 1973; Healey 1980;
Godin 1981; Cooney et al. 1981; Landingham 1982; Cordell 1986;
Taylor et al. 1987; Landingham and Mothershead 1988). The trophic
link between the benthos and the salmon is the most likely route
for an impact on juvenile salmon in 1990. The contamination of the
littoral zone could reduce prey densities by direct toxicity to
harpacticoid copepods (Bonsdorff 1981; Bodin 1988), or by changing
harpacticoid species assemblages from those dominated by epibenthic
species to those dominated by inbenthic species which are not as
available to visual feeders (Stacey and Marcotte 1987). Uptake of
hydrocarbons by harpacticoids living in and- on the contaminated
sediments could also reduce growth. Contamination from prey
decreases growth and causes changes in'feeding behavior of juvenile
pink salmon (Schwartz 1985). Slower-growing juveniles are more
susceptible to size selective predation (Parker 1971; Hargreaves
and LaBrasseur 1985) and thus suffer higher mortality (Healey 1982;
Taylor et al. 1987; Taylor 1988).

Proposed research for continuation in 1990 will collect data on
distribution, abundance, habitat utilization, size and growth, and
feeding habits of juvenile pink and chum salmon, in order to
compare these parameters with the 1989 results. Resolution of
growth comparisons between oiled and non-oiled locations will be
increased by using otolith increment analysis (Volk et al. 1984).
Research will determine if sediment contamination has reduced the
abundance of primary prey species of harpacticoid copepods and will
determine if pollution of azoic sediments with hydrocarbons will
inf luence meiof auna colonization, especially harpacticoid copepods,
in terms of species distribution and abundance.

OBJECTIVES

(Letters refer to general objectives described above, as well as
three components listed above.)

D-1. Test, at alpha = 0.05, if the abundance of juvenile pink
and chum salmon does not differ between oiled and
non-oiled areas.

D-2. Compare distribution and habitat utilization by juvenile
salmon between 1989 and 1990.

D-3. Test, at alpha 0. 05, if the size and growth r *ates of
juvenile salmon do not differ between oiled and non-oiled
areas; to compare growth rates between 1989 and 1990.

D-4. Quantify the feeding habits of juvenile pink and chum salmon
in terms of fullness, frequency of occurrence, biomass, and
index of Relative Importance, and compare oiled and non-oiled
areas in 1990 and between 1989 and 1990.

D-5. Determine migratory behavior of juvenile salmon based on
coded-wire tag recoveries.

E-1. Examine if sediment contamination has reduced the abundance of
primary prey species (harpacticoid copepods).

E-2. Test, at alpha = 0.05, if the abundance of epibenthic prey
species for juvenile salmon does not differ between heavily
contaminated and lightly contaminated beaches.

F-1. Test, at alpha = 0.05, if the colonization of sediments by
harpacticoid copepods and other meiofauna is not affected by
the presence of oil in the sediments.

METHODS

Sampling Design

Component I (Objective D)

In order to make direct comparisons between years, the same four
oiled and four non-oiled locations sampled in 1989 in Western PWS
will be sampled again in 1990. The locations are categorized as
bays or migration corridors. At each location, three habitat types
will be sampled. These habitat types are grossly characterized by
grade and substrate: low gradient beach (<10% grade, granule-pebble
substrate) ; medium gradient beach (12-25% grade, pebble-cobble
substrate) ; and steep gradient beach (>50% grade, bedrock or large
boulder substrate). Particular beaches will be selected for

similarity between oiled and non-oiled areas in such
characteristics as wave exposure, macrophyte coverage, and
substrate. Two beaches of each habitat type will be sampled within
each location in 1990, for a total of 48 sites. To minimize
variability due to tide heights, sampling at the sites will be
restricted to the -1 to +3 tide range. The sites will be sampled
on each of four sampling cruises from mid-April to early June.

Component 2 (Objective E)

Abundance of important harpacticoid prey species will be compared
between "lightly oiled" and "heavily oiled" beaches within each of
three oiled embayments. Comparisons between oil contamination
levels within an embayment was chosen to minimize the effects of
geographic variability. Three 40 m transects at the 0 tide level
will be established for each of the two contamination categories
within each embayment. Random number series will be used to select
25 points along each transect for sampling with an epibenthic pump.
Transects for prey abundance will be sampled during the low-tide
series encompassed by sampling cruises 1,3, and 4; a different
embayment will be sampled on each of these cruises.

Component 3. (Objective F)

Colonization by meiofauna of azoic sediments will be compared
between oiled and control sediments. Standard dish pans with holes
to allow water drainage will be filled with control, low oil (0. 5%)
and high oil (2.0%) azoic sediiuents. The pans will be aged in
running freshwater for 1-2 weeks prior to use. Three pans for each
level (control, two treatments) will be buried in the lower
intertidal in two locations (a lightly oiled and a heavily oiled
location). The pans will be placed parallel to the water line,
approximately 5 pan widths apart. The sediments used in this
experiment will be collected in Auke Bay, and made azoic by
freezing. Approximately one*third will be used for the controls.
The remaining sediments will be divided in half, and mixed with
Prudhoe Bay crude oil to 0.5% and 2% concentrations. The pans will
be placed in PWS in late April, and be sampled for meiofauna after
I day and 4, 6, and 12 weeks.

Sample Collection

1. Fish Sampling

Fish sampling at study sites will be restricted to the -1 to +3
tide levels to minimize tidal effects between sites. Fish will be
captured using 37 m beach seines. Catches will be sorted by
species and enumerated; all salmon will be checked for the presence
of CW tags using an OMNI coded-wire tag detector. Each CW tagged
salmon will be measured, weighed and frozen. on each sampling
trip, up to 60 each juvenile pink and 60 juvenile chum salmon from
each sample site will be preserved in formalin for later length and

weight measurements; 10 of each species of these f ish will be
randomly selected during processing for diet analysis. In
addition, 50 juvenile pink salmon from each embayment site will be
retained for otolith analysis, as per standard operating
procedures. All other fish will be released.

As time permits, the shoreline within the general vicinity of the
habitat sites will be surveyed, and additional seine sets made.
Juvenile salmon collected in these sets will be enumerated, checked
for coded-wire tags, and used to supplement collections for otolith
and stomach analyses when insufficient numbers are collected at the
regularly sampled beaches. All other fishes caught in such sets
will also be identified and enumerated.

2. Zooplankton, Epibenthic Harpacticoids, and Meiofauna Sampling

In the offshore water adjacent to the habitats sampled, triplicate
samples of pelagic zooplankton will be taken with a 20-m vertical
haul of a 0.5 m diameter 243 micron net. Epibenthic harpacticoid
copepods will be sampled using a pump sampler (Cordell and
Simenstad 1989).

Meiofauna in the pan experiment (Component 3) will be sampled by
taking five core samples from each pan, using 50-ml syringes.
Samples will be f ixed in buf f ered f ormalin in 120-ml glass jars and
labeled and sealed in the same manner as the other prey samples.

3. Hydrocarbon Samples

Mussel samples and sediment samples will be taken for hydrocarbon
analysis at each location. Mussels will be sampled at or near one
of the habitats within each site, and frozen. Three replicate
samples will be collected in 120 ml glass jars from the top 2 cm of
sediments from 6-8 spots along the water line adjacent to the beach
seine site. For the epibenthic transects, the replicate samples
will be aggregated from within the six random quadrants selected
for photographing. Two core samples will be taken for hydrocarbon
analysis from the experimental sediment pans. A blank sample will
be provided from each site. All sediment samples will be frozen.

4. Environmental Data

Water temperature and salinity at 0.5 m depth, wave height, and
current measurements will be taken at each nearshore site regularly
sampled, and at each prey transect. Water temperatures at 1 m and
4 m will be taken in association with each set of zooplankton tows.
Temperature and salinity will be measured using a Beckman probe
conductivity-temperature meter. Current will be measured with a
Marsh-McBirney induction current meter. Wave height will be
measured with a meter stick. Extent of oil deposition and of
visible oil in the water will also be noted for each habitat. A
recording temperature/ salinity device will also be deployed at the

two sites where pans of experimentally oiled and control sediments
are placed, for hourly temperature records. The boundary of the
oxic-anoxic layer within the sediment pans will.be measured at the
end of the experiment.

Sample Processing

1. Fish Samples

Coded-wire tagged fish will be stored frozen until processing for
tags. The fish will be transported from field collection to the
Auke Bay Laboratory, then to the ADFG Tag Processing Laboratory in
Juneau. The tag lab personnel will decode the tags, and transmit
the information to NMFS and to the ADFG investigator coordinating
Early Marine Salmon Studies.

After being weighed and measured, each fish retained for stomach
analysis will be put into a labeled 20-ml vial filled with 50%
isopropyl alcohol or 70% ethanol. Subsequent analysis will involve
excising and weighing the foregut, removing the contents and
estimating stomach fullness, and reweighing the empty foregut to
get a measure of total content wet weight. The prey items will be
identified to a minimum of order level and counted.

2. Otolith samples

The sagittal otoliths will be removed from frozen samples of at
least 50 juvenile pink salmon from each of the four embayment
locations, for both 1989 and 1990 samples. Each otolith sample
will be assigned a sample number corresponding to the original
sample and the fork length of the individual fish. The otoliths
will be sent to a qualified contractor, who will process the
otoliths and determine: number of increments subsequent to the
hatching and saltwater entry check; width of these increments along
a standard axis in the posterodorsal quadrat of the otolith; mean
increment width and associated error term for each 50 fish group.

3. Zooplankton, Epibenthic Pump, and Meiofauna Samples

Upon transport to the Auke Bay Laboratory, the samples will be
logged in by sample number. A total of 96 zooplankton samples, 450
epibenthic pump samples, and 360 meiofauna core samples will
require processing.

4. Hydrocarbon samples

A total of 32 mussel samples and 472 sediment samples (counting
triplicates) will be collected. All hydrocarbon samples collected
in the course of this study will be prioritized by the Hydrocarbon
Analysis project as to if and when the samples will be processed.
Procedure for analysis of these samples is detailed in the
Hydrocarbon Analysis study plan. (Technical Services Study No. 1)

5. Sediment analysis

A total of 54"sediment samples each from component 2 and 54 from
component 3 will be collected for organic carbon and nitrogen
analysis and to quantify the sediment composition. Processing of
these samples will also be let to a qualified contractor.

DATA ANALYSIS

Component 1 (Oblective D)

Two approaches will be used to compare abundance and size of
juvenile salmon, and stomach fullness and relative biomass of
stomach contents: non-parametric comparisons of paired oiled and
non-oiled locations, and analysis of variance (ANOVA). The
Wilcoxin matched-pairs signed-rank test (Daniel 1978) contrasts the
differences between the a priori pairs of oiled and non-oiled
locations f or cells that match in terms of time and habitat. In the
full analysis of variance model, five factors will be considered:
oil/no-oil (fixed), time (fixed), bay/corridor (fixed), location
(specific sampling location nested within bay/corridor), and
habitat type (fixed). To confirm probability levels for the main
factor of interest (oil) , a randomization procedure will be used to
generate distribution-free significance levels.

Nominal growth rates between oiled and non-oiled areas will be
compared using a exponential growth model, and comparing the
regression slopes of Ln weight over time with analysis of
covariance (Zar 1974). ANOVA will be used to compare mean otolith
increment widths using a partially hierarchal design (Winer 1971)
involving three factors: year, oil/no-oil, and bays nested within
oil/no-oil. Condition of juvenile salmon will be compared between
oiled and non-oiled areas using least squares regression of the
natural logarithms of weight and length (Cone 1989). Percent
similarity indexes (Whittaker 1975) will be calculated for feeding
habits between oiled and non-oiled areas, between bays and corridor
locations, and between 1989 and 1990.

Component 2 (Oblective E)

Abundance, percent gravid females, and percent total harpacticoids
for primary prey species of juvenile salmon will be compared
between heavily oiled and lightly oiled beaches using ANOVA.
Because the three embayments will be sampled at different times,
differences between bays are not of interest as they could be an
artifact of sampling time. Thus each embayment will be considered
a separate experiment using a nested ANOVA to compare lightly and
heavily oiled transects. The transects will be nested within oil
contamination levels. An alternate analytical approach will be to
use regression to examine the relationship between abundance of the
to amount of oil in the sediment, as well as the substrate
composition, macrophyte coverage, and carbon and nitrogen levels in

the sediments.

Component 3 (Objec ive F)

Abundance of total meiof auna and harpacticoid copepods in the
experimental sediments will be compared using a three f actor,
fully-crossed ANOVA (Winer 1971). The factors are location of the
sediments, level of oil contamination in the sediments, and time.

Data and Sample Archival

All f ield and laboratory data forms generated through the course of
this study will be placed in notebooks numbered according to the
Auke Bay Laboratory Oil Spill Notebook Tracking System (NTS). All
f ield notes will be similarly cataloged. Trip reports for each
sampling cruise and study plans will also be archived within the
NTS. Copies of computer data f iles will be maintained on two
microcomputer hard drives, as well as rotating f loppy disk back-up
kept in a locked cabinet.

Table 1. Location of sample sites in Prince
William Sound listed as a priori pairs.

Location Type oil

Herring Bay Bay Yes

McClure Bay Bay No

Snug Harbor Bay Yes

Long Bay Bay No

Prince of Wales Corridor Yes

Passage

Culross Passage Corridor No

Knight Island Corridor Yes
Passage

Wells Passage Corridor No

BIBLIOGRAPHY

Bailey, J.E., B.L. Wing, and C.R. Mattson. 1975. Zooplankton
abundance and feeding habits of fry of pink salmon and chum
salmon in Traitor's Cove, Alaska,with speculations on the
carrying capacity of the area. Fish. Bull. 73:946-961.

Babcock, M. M. 1985. Morphology of olfactory epithelium of pink
salmon@ Oncorhynchus gorbuscha, and changes following exposure
to benzene: a scanning electron microscope study,p. 259-267,
In J. S. Gray and M. E. Christiansen (eds), Marine biology of
polar regions and stress on marine organisms. John Wiley &
Sons.

Barnard, D. R. 1981. Prey relationships between juvenile pink
(Oncorhynchus gorbuscha) and chum (Q. Keta) salmon in Prince
William Sound, Alaska. M. S. Thesis, University of Alaska,
Fairbanks.

Bax, N.J. 1983. Early marine mortality of marked juvenile chum
salmon released into Hood Canal, Puget Sound, Washington, in
1980. Can. J. Fish. Aquat. Sci. 40:426-435.

Bodin, P. 1988. Results of ecological monitoring of three beaches
polluted by the 'Amoco Cadiz' oil spill: development of
meiofauna from 1978 to 1984. Mar. Eco. Prog. Ser. 42: 105-
123.

Bonsdorff, E. 1981. The Antonio Gramsci oil spill impact on
the littoral and benthic ecosystems. Mar. Pollut. Bull.
12(9):301-305.

Caldwell, R. S., E. M. Caldarone, and M. H. Mallon. 1977. Effects
of a seawater-soluble fraction of Cook Inlet crude oil and its
major aromatic components on larval stages of the Dungeness
crab, Cancer magister Dana. p. 210-220 In D. A. Wolfe (ed),
Fate and effects of petroleum hydrocarbons in marine
ecosystems and organisms. Pergamon Press, Oxford.

Cooney, R. T., D. Urquhart, R. Neve, J. Hilsinger, R. Clasby, and
D. Barnard. 1978. Some aspects of the carrying capacity of
Prince William Sound, Alaska, for hatchery released pink and
chum salmon fry. Univ. Alaska, Sea Grant Rep. 78-4. 102 p.

Cooney, R.T., D. Urquhart, and D. Barnard. 1981. The behavior,
feeding biology and growth of hatchery-re leased pink and chum
salmon fry in Prince William Sound, Alaska. Alaska Sea Grant
Report 81-5. 114 pp.

Cordell, J.R. 1986. Structure and dynamics of an epibenthic

harpacticoid assemblage and the role of predation by juvenile
salmon. M.S. Thesis, Univ. Wash., Seattle. 134 pp.

Cordell, J.R. and C.A. Simenstad. 1989. Epibenthic harpacticoid
copepods as indicators of wetland fitness, p. 422-431, In
Proceedings First Annual Puget Sound Research Symposium.

Coyle, K. 0. and A. J. Paul. 1988. Zooplankton population and
biomass in Auke Bay during the spring diatom bloom, 1985-
1987. 1987 APPRISE report; University of Alaska SFOS APP87-
100.

Coyle, K. 0. and A. J. Paul. 1989. Zooplankton population and
biomass. 1988 APPRISE report; University of Alaska SFOS
APP88-200.

Daniel, W. W. 1978. Applied nonparametric statistics. Houghton
Mifflin Co., Boston. 510 pp.

Foerster, R.E. 1968. The sockeye salmon. Fish. Res. Bd. Can.
Bull. 162, 422 p.

Fulton, J. D. 1983. Seasonal and annual variations in net
zooplankton at Ocean Station P, 1956-1980. Can. Fish. Aqua.
Sci. Data Rept. No. 374. 65 pp.

Godin, J.-G.J. 1981. Daily patterns of feeding behavior, daily
rations, and diets of juvenile -pink salmon (Oncorhynchus
ctorbuscha) in two marine bays of British Columbia. Can. J.
Fish. Aquat. Sci. 38:10-15.

Gundlach, E. R., P. D. Boehm, M. Marchand, R. M. Atlas, D. M. Ward,
and D. A. Wolfe. 1983. The fate of Amoco Cadiz oil. Scierre
221: 122-129.

Hargreaves, N. B. and R. J. LeBrasseur. 1985. Species selective
predation on juvenile pink (Oncorhynchus gorbuscha) and chum
(g. keta) by coho salmon (2. kisutch). Can. J. Fish. Aquat.
Sci. 42:659-668.

Hartt, A.C. 1980. Juvenile salmonids in the oceanic ecosystem-
the critical first summer. In Salmonid ecosystems of the
North Pacific, W.J. McNeil and D.C. Himsworth, eds., p. 25-57.
Oreg. State Univ. Press.

Healey, M.C. 1978. The distribution, abundance and feeding habits
of juvenile Pacific salmon in Georgia Strait, British
Columbia. Fish. Mar. Serv. Tech. Rep. 788, 49 p. Dep. Fish.
Environ. Pac. Biol. Stn. Nanaimo, B.C.

Healey, M. C. 1980. The ecology of juvenile salmon in Georgia
Strait, British Columbia. In W. J. McNeil and D. C. Himsworth

(eds.) , Salmonid ecosystems of the North Pacific. Oregon State
Univ. Press, Corvallis. 331 pp.

Healey, M. C. 1982. Timing and relative intensity of size-
selective mortality of juvenile chum salmon during early sea
life. Can. J. Fish. Aquat. Sci. 39:952-957.

Helle, J.H. 1980. Influence of marine environment in age and size
at maturity, growth, and abundance of chum salmon from Olsen
Creek, Prince William Sound, Alaska. Ph.D. Thesis, Oreg.
State Univ.

Kaczynski, V. W., R. J. Feller, and C. Clayton. 1973. Trophic
analysis of juvenile pink and chum salmon in Puget Sound. J.
Fish. Res. Board Can. 30:1003-1008.

Landingham, J. H., and P. D. Mothershead. 1988. Feeding habits of
juvenile pink salmon in nearshore and offshore areas of Auke
Bay. p. 450-469 In APPRISE 1988 Annual Report, Vol. 1. School
of Fish. and ocean science, University of Alaska, Fairbanks.

LeBrasseur, R. J. and R. R. Parker. 1964. Growth rate of central
British Columbia pink salmon (Oncorhynchus gorbuscha) . J.
Fish. Res. Board Can. 21:1101-1128.

Milliken, G. A. and D. E. Johnson. 1984. Analysis of messy data,
volume I: designed experiments. Van Nostrand Reinhold, New
York.

Mortensen, D. G. and A. C. Wertheimer. 1988. Residency and growth
of juvenile pink salmon (onncorhynchus gorbusha) in Auke Bay,
Alaska. 1987 APPRISE report; University of Alaska SFOS
AAP87-100.

Murphy, M.L., J.F. Thedinga, and K V. Koski. 1988. Size and diet
of juvenile Pacific salmon during seaward migration through a
small estuary in southeastern Alaska. Fishery Bulletin,
86:213-222.

Nichelson, T. E. 1986. Influences of upwelling, ocean
temperature, and smolt abundance on marine survival of coho
salmon (Oncorhynchus kisutch) in the Oregon production area.
Can. J. Fish. Aquat. Sci. 43:527-535.

Parker, R.R. 1968. Marine mortality schedules of pink salmon of
the Bella Coola River, central British Columbia. J. Fish.
Res. Bd. Can. 25:757-794.

Parker, R. R. 1971. Size selective predation among juvenile
salmonid fishes in a British Columbia inlet. J. Fish. Res.
Bd. Can. 28:1503-1510.

Pielou, E.C. 1975. Ecological diversity. John Wiley and Sons.
New York. 165 pp.

Pinkas, L., M. S. Oliphant, and I.. L. K. Iverson. 1971. Food
habits of albacore, bluefin tuna and bonito in California
waters. Calif. Dep. of Fish and Game Bull. 152:1-64.

Rice, S. D., D. A. Moles, J. F. Karinen, S. Korn, M. G. Carls, C.
C. Brodersen, J. A. Gharrett, and M. M. Babcock. 1984.
Effects of petroleum hydrocarbons on Alaskan aquatic
organisms. NOAA Tech. Mem. NMFS F/NWC-67. 128 p.

Rice, S. D., D. A. Moles, and J. W. Short. 1975. The effect of
Prudhoe Bay crude oil on survival and growth of eggs, alevins,
and fry of pink salmon, oncorhynchus gorbuscha.p. 503-507,In
1975 Conference on prevention and control of oil pollution.
American Petroleum Institute, Washington, D. C.

Ricker, W.E. 1976. Review of the growth rate of and mortality of
Pacific salmon in salt water, and non-catch mortality caused
by fishing. J. Fish. Res. Bd. Can. 33:1483-1524.

Schwartz, J. P. 1985. Effects of oil-contaminated prey on the
feeding and growth rate of pink salmon fry Oncorhynchus
gorbuscha. Pp. 459-476 in, Vernberg, F. John, Frederick
Thurberg, Anthony Calabrese, and Winona Vernberg (eds.),
Pollution and Physiology of Marine Organisms. U. South
Carolina Press. Columbia, S.C. 545 pp.

Stacey, B. M. and B. M. Marcotte 1987. Chronic effect of No. 2
fuel oil on population dynamics of Harpacticoid copepods in
experimental marine mesocosms. Mar. Ecol. Prog. Ser.
40:61-68.

Taylor, C. C. 1953. Nature of variability in trawl catches. Bull.
U. S. Fish. Wildl. Serv. 83:144-166.

Taylor, S.G. 1977. The effect of timing of downstream migration on
marine survival of pink salmon. M.S. Thesis. Univ. Alaska,
Southeastern Senior College, Juneau, 40 p.

Taylor, S.G. 1988. Inter- and intra-annual survival of pink salmon
(Oncorhynchus gorbuscha) returning to Auke Creek, Alaska in
1986 and 1987. p. 545-572, In APPRISE Annual Report, Vol. 1.
School of Fish. and Ocean Science, University of Alaska,
Fairbanks.

Taylor, S. G., J. H. Landingham, D. G. Mortensen, and A. C.
Wertheimer. 1987. Pink salmon early life history in Auke
Bay: Residence, growth, diet and survival. p. 273-318, In
APPRISE Annual Report-1986. Vol. I: Technical Report.

School of Fisheries, University of Alaska, Juneau.

Urquhart, D.L. 1979. The feeding, movement, and growth of pink
salmon fry released from a hatchery in Prince William Sound,
Alaska. M.S. Thesis, University of Alaska, Fairbanks. 111 p.

Vogel, A. H. and G. McMurray. 1986. Seasonal population density
distribution of copepods, euphausiids, amphipods and other
holoplankton on the Kodiak shelf. Outer Continental shelf
Environmental Assessment Program. Final reports of Principal
Investigators 46:423-659.

Volk, E.C., R.C. Wissmar, C.A. Simenstad, and D.M. Eggers. 1984.
Relationship between otolith microstructure and the growth of
juvenile chum salmon (Oncorhynchus keta) under different prey
rations. Can. J. Fish. Aquat. Sci. 41:126-133.

Walters, C.J., R. Hilborn, R.M. Peterson, and M.M. Staley. 1978.
Model for examining early ocean limitation of Pacific salmon
production. J. Fish. Res. Bd. Can. 35:1303-1315.

Whittaker, R. H. 1952. A study of summer foliage insect
communities in the Great Smoky Mountains, Ecological
Monographs 2(l):1-44.

Whittaker, R. H. 1975. Communities and ecosystems. MacMillan
Publishing Co., N.Y. 385 pp.

Winer, B. J. 1971. Statistical principles in experimental design.
McGraw-Hill, New York. 907 pp.

Zar, J. H. 1974. Biostatistical analysis. Prentice- Hall, Inc.,
Englewood Cliffs, NJ.

BUDGET

BUDGET: ADF&G

Salaries $82.0
Travel 3.0
Contracts 48.0
supplies 5.0
Equipment 12.0

Total $150.0

BUDGET: NOAA Total

Labor $ 100.0
Travel 15.0
Contractual Services 100.0
Supplies and Materials 25.0
Equipment 10.0
Vessel Support: 150.0

Total $ 40d.0

TOTAL BUDGET

Category Budget

Personnel Services $182.0
Travel 18.5
Contractual 140.0
Supplies 30.5
Equipment 21.0
Vessel Support 150.0

Total $550.0

FISH/SHELLFISH STUDY NUMBER S

Study Title: Injury to Dolly Varden Char and Cutthroat Trout
In PWS

Lead Agency: ADF&G

INTRODUCTION

The goal of this study is to compare the survival and growth of
populations of Dolly Varden Char (char) and cutthroat trout (trout)
differentially affected by the oil spill in PWS. This project is
currently in the second year of a three-year study design. Trout
and char are estuarine anadromous species that inhabit PWS (Morrow
1980). Unlike anadromous Pacific salmon, trout and char utilize
nearshore and estuarine areas for feeding. Their marine
migrations are not as extensive as those of Pacific salmon (Morrow
1980). Some of the most important stocks of these species inhabit
areas that have been severely impacted by direct contact with oil
including Green and Montague Islands and Eshamy Bay (Mills 1988).
since these species commonly live to age 8 (Morrow 1980), the
potential exists for both short-term and long-term effects from
exposure to oil. Study of these species is crucial in that they
represent finfish species that inhabit the most oil-affected areas
throughout most of their lives.

The experimental design for this program is based upon the model
developed by Armstrong (1970, 1974, 1984) and Armstrong and Morrow
(1980) to explain the migratory behavior of anadromous char. This
model identifies two patterns of life history, fish spawned in lake
systems and fish spawned in non-lake systems. For both groups,
juvenile char remain in freshwater residence in their natal stream
f or up to f our years. During their last spring of freshwater
residence, they smolt to sea. During late summer or early fall,
f ish that were spawned in lake systems return to their natal stream
to overwinter in the f reshwater lake. During the spring, they
again emigrate into marine waters and annually return to their
natal lake system during late summer or early fall to spawn and
overwinter. Fish that were spawned in non-lake systems exhibit a
more complex migration. Upon smolting, juvenile-char search for a
lake system to overwinter. These fish then behave in the same
manner as do tish that originate in a lake system except that they
return to their natal stream to spawn and then return to their
selected lake system to overwinter.

The migratory habits of anadromous cutthroat trout are less well
understood than those of anadromous char in Alaska although it
appears that they exhibit similar migratory habits to char (Jones
1982). Trout, however, spawn in the spring as opposed to fall for
char.

It is hypothesized that two detrimental impacts on these species
could result from the presence of large amounts of crude oil in
marine waters: (1) reduced survival; and (2) reduced growth. To
test whether there was a measurable impact, three stocks of trout
and char that over winter in watersheds that issue into a marine
environment which has been directly exposed to oil (treatment
group) and two stocks of trout and char that over winter in
watersheds that issue in unoiled areas (control group) were
selected for study.

significant changes in stock abundance, composition, or dynamics
from the initial emigration of stocks within the treatment group as
compared to stocks from the control group is assumed to be due to
contact with the oiled marine waters. Evidence from the literature
indicates that marine migrations can range up to 116 kilometers for
char (Armstrong 1974) and 80 kilometers for trout (Jones 1982).
Armstrong's model of migratory behavior provides the basic
framework for this study. First, each of the study streams
represents a stock of fish that annually homes to that specific
over wintering stream. second, since over winter residency occurs
entirely in freshwater, fish sampled during the 1989 spring
emigration had not yet encountered oiled waters. Given this, the
first sample from each stream (the emigration during 1989) provides
the baseline data for stocks in control and treatment.

OBJECTIVES

A. Test if there is no difference in annual survival rates of
char and cutthroat trout between treatment and control groups
during 1989-90 (the test will be done given a level of
significance of alpha = 0.05.);

B. Test if there is no difference in annual growth rates of char
and cutthroat trout between treatment and control groups
during 1989-90 (the test will be done given a level of
significance of alpha = 0.05.);

METHODS

Trout and char were still in freshwater residence at the time of
the spill, and the opportunity existed to sample these fish during
their 1989 emigration prior to any potential exposure to an oiled
marine environment. Data collected during 1989 became the baseline
for each system. Therefore, in addition to comparisons between
treatment and control, comparisons are also possible for each
stream within treatment and control between subsequent years' data
and the 1989 baseline.

Each study stream consists of a freshwater lake-river system that:
(1) is a tributary to marine waters that were either impacted by
large quantities of oil (treatment) or received virtually no oil
(control) ; and (2) contains stocks of anadromous trout and char.

A weir will be installed and completely block each study stream
prior to the initiation ofthe 1990 spring emigration. A smolt
weir for sockeye salmon will operate at the outlet of Eshamy Lake
as part of FIS Study #3. Sampling for char and trout will be
conducted in conjunction with this project.

Char greater than 250 mm. in length will be initially clas sified as
mature (Blackett 1968). At the conclusion of this year's sampling,
length frequency data will be analyzed to identify more precise
classifications for immature and mature fish.

All emigrating trout and char at each weir site will be counted and
measured from tip-of-snout to fork-of-tail to the nearest
millimeter. Trout and char greater than 199 mm. will be tagged
with numbered Floy FD-68 anchor tags and fish under 199 mm and
greater than 149 mm, will be tagged with numbered Floy Fabric anchor
tags. All tagged fish will have their adipose fin removed to
estimate tag loss.

If all fish can be censused and examined for tags in all years,
survival will be known for each system. Annual survival will be
estimated for immature and mature trout and char. The mortality
rate of spawning char is known to be high, particularly for males
(Armstrong 1974) and it is probable that the same is true for
trout. Therefore, the rate of survival estimated for immature char
and trout will be used to test the hypothesis of equal survival
between treatment and control groups.

The hypothesis of equal survival will be tested using a chi-square
statistic. However, if unknown numbers of fish can be expected to
be lost past the weir (due to such events as weir washout) , it will
not be possible to directly estimate survival from the numbers
released and returned. Instead survival will be estimated using
mark-recapture methods. Estimates of survival (Seber 1982) from a
mark-recapture experiment with their 95% confidence intervals at
three levels of abundance were examined to estimate the sample
goals required to detect significant differences in survival.

The hypothesis of equal growth will be tested by analysis of
individual growth rate. Incremental growth for individuals will be
computed from recaptured fish. An Analysis of Variance will be
conducted with stocks of char or trout serving as replicates within
the treatment group. Years, and possibly initial length, will
serve as f actors in the design. Differences in average growth
rates between control and treatment groups will be attributed to
some external disturbance so long as initial length of f ish is used
as a covariate.

During the spring sampling, weirs will be used to count and sample
the emigration of trout and char from study streams. Weirs will be
installed approximately 0.5 km upstream from the saltwater terminus
of the streams. The weirs will be operated by a two-person crew

f rom mid- April to early July. Downstream live traps will be
installed.

All fish captured in the trap will be examined for presence or
absence of tags, tag scars, and adipose fins. Each fish containing
a tag from 1989, a tag scar, or missing its adipose fin will be
considered one recapture event. Recaptured fish with missing tags
will be retagged. Fish with no visible tag scar and containing
their adipose fin (not tagged in 1989) will also be tagged. Each
fish captured will be identified, counted, and measured (tip-of-
snout to fork-of-tail to the nearest mm). Scale smears will be
collected from the preferred area from all cutthroat trout and
placed individually on acetate slides in coin envelopes. Date,
species, sex (if identifiable from external maturation
characteristics), and length will be recorded for each fish.
Recapture events will be recorded separately for fish containing
tags and fish with missing tags. Tag numbers will be recorded for
each recapture and each fish tagged.

All fish found dead impinged on the weir or in the live box will be
examined for presence of tags and adipose fins, identified, and
measured as outlined above. Sex and maturity will be determined by
internal examination, and sagittal otoliths will be collected.
Datef species, sex, length, maturity, and tag number will be
recorded. Fish containing tags, tag scars or missing adipose fins
will be recorded as recaptures.

Estimates of annual survival will be computed for each study site
through analysis of tag returns. If all emigrating fish can be
examined for marks, the estimates of annual survival (S) can be
simply computed as:

S = M2/Rl

where:

m2 = number of fish recovered in year y+l
R1 = number of fish tagged in year y.

The Jolly-Seber three-sample method (Seber 1982) found in the
Appendix will be used in the event that each emigrating fish cannot
be examined at the weirs. Buckland's program RECAP (1980) will be
used to generate the estimates and variances.

The sampling event for the purposes of the mark-recapture
experiment is the emigration of trout and char past the weirs. All
emigrating fish must cross the weir and therefore are assumed to be
equally vulnerable to being sampled. The assumption of equal
survival of tagged fish will be tested for the different tag groups
and for the different length classes using chi-square statistics.
Tag loss will be estimated for fish tagged in 1989, as all tagged
fish will also have their adipose fin-clipped.

If all weirs hold, the hypothesis of equal survival will be tested
using a chi-square test for independence. Given a survival rate
from immature to mature fish of greater than 15% the test will be
able to detect differences of 7% or more (alpha = .05). If the
weirs do not hold and all char are not sampled at the weir, during
emigration during the second and third year of the experiment; the
95% confidence intervals of the survival estimated from the multi-
year mark-recapture experiments will be compared to test for
significant differences. In order to examine the effect of initial
length on subsequent survival, the tests and estimates will be
stratified by tagging length, and if possible a logistic regression
will be used to estimate this effect.

Annual individual growth will be calculated from the tag data as
the difference between length at time of release and length at time
of recovery. An Analysis of Variance will be used to test for
significant differences in growth between fish from control and
treatment groups. Variation due to differences in years and
initial length can be controlled for through the use of a block and
covariate in the linear model if necessary. The power to detect a
.5% difference in the growth rate of fish from treatment and control
areas is estimated to be 90%.

The assumption of normality will be tested using Kolomogorov's D
statistic. In all likelihood the data will not be normally
distributed and a logarithmic or a rank transformation will be
necessary.

The homogeneity of variance assumption will be tested with a
Bartlett's test. Again, if the assumption is not valid a
transformation will be used.

BIBLIOGRAPHY

Armstrong, R.H. 1970. Age, food, and migration of Dolly Varden
smolts in Southeastern Alaska. J. Fish. Res. Bd. Canada
27:991-1004.

. 1974. Migration of anadromous Dolly Varden (Salvelinus
malma) in southeastern Alaska. J. Fish Res. Board Can.
31:435-444.

. 1984. Migration of anadromous Dolly Varden char in
southeastern Alaska - a manager's nightmare. p. 559-570. In L.
Johnson and B.L. Burns [eds.) Biology of the Arctic char,
Proceedings of the International Symposium on Arctic Char,
Winnipeg, Manitoba, May, 1981. Univ. Manitoba Press,
Winnipeg.

Armstrong, R.H. and J.E. Morrow. 1980. The Dolly Varden char. p.
99-104. In Balon, E.K. [ed.] Chars: salmonid fishes of the
genus Salvelinus. Dr. W. Junk b.v., Publisher. The Hague,

Netherlands.

Blackett' R. F. 1968. Spawning behavior, fecundity and early life
history of anadromous Dolly Varden Salvelinus malma (Walbaum)
in southeastern Alaska. ADF&G Research Report. 6:85 p.

Buckland, S.T. 1980. A modified analysis of the Jolly-Seber
capture-recapture model. Biometrics 36: 419-435.

Clutter, R. and L. Whitesel. 1956. Collection and interpretation
of sockeye salmon scales. International Pacific Salmon
Fisheries Commission, Bulletin 9. 159 pp.

Hepler, K., A. Hoffmann, and P. Hansen. 1989. Injury to Dolly
Varden char and cutthroat trout in Prince William Sound.
State/Federal Natural Resource Damage Assessment Preliminary
Status Report Draft, January 1990. Fish/Shellfish Study
Number 5. Alaska Department of Fish and Game, Sport Fish
Division, Anchorage.

Jones, D.E. 1982. Development of techniques for enhancement and
management of cutthroat trout in southeast Alaska. Alaska
Department of Fish and Game. Annual Report of Progress,
Project AFS-42, 23(AFS-42-10-B): np.

Mills, M.J. 1988. Alaska statewide sport fisheries harvest
report. Alaska Department of Fish and Game, Fishery Data
Series No. 2. 142 pp.

Morrow, J. E. 1980. The freshwater fishes of Alaska. Alaska
Northwest Publishing Company, Anchorage, Alaska. 248 pp.

Roth, K.J., C. Whitmore, and P. Hansen. 1990. Prince William
Sound and Gulf of Alaska sport fishery harvest and effort,
1989. State/Federal Natural Resource Damage Assessment
Preliminary Status Report Draft, January 1990. Fish/Shellfish
Study Number 6. Alaska Department of Fish and Game, Sport
Fish Division, Anchorage.

Seber, G. A. F. 1982. Estimation of animal abundance and related
parameters. 2nd edition, Griffin & Company, London. 655 pp.

BUDGET: ADF&G

Salaries $ 228.0
Travel 6.0
Contracts 37.3
Supplies 18.7
Equipment 0.0

Total $ 290.0

FISH/SHELLFISH STUDY NUMBER 7a

Study Title: Injury to Pink/Chum Salmon Spawning Within
Lower Cook Inlet and Kenai Fjords

Lead Agency: ADF&G

INTRODUCTION

Wild stocks of pink and chum salmon are a major ecosystem component
in the outer Kenai Peninsula and Lower Cook'Inlet area, immediately
"down current" from PWS. Salmon represent a very important food
source for marine mammals (sea lion, seals), terrestrial mammals
(bear) , and a wide variety of bird species (eagles, etc.) . In
addition, these wild stocks of pink and chum salmon are harvested
commercially. In 1988, the year before the oil spill, the ex-
vessel value of the commercial catch of wild and hatchery stocks of
salmon from the lower CIK area was more than $8.2 million. Salmon
are also very important to the sport, subsistence, and personal use
fisheries. The future abundance of wild stocks of pink and chum
salmon in the lower CIK areas may be adversely impacted as their
intertidal spawning areas were affected by the oil spill.* This
project was designed to evaluate the distribution of pink and chum
salmon spawning in intertidal and upstream areas as a result of oil
contamination from the Exxon Valdez oil spill. This project also
provides spawner distribution data for F/S Study No. 8a).

OBJECTIVES

A. Count the numbers of spawning salmon by' species and by
intertidal and upstream areas for nine streams in the Lower
Cook Inlet/Kenai Fjords area.

B. Produce maps of spawner distribution for each stream sampled.

METHODS

This project is designed to evaluate changes in numbers and
.distribution of spawning salmon relative to oil contamination from
the Exxon Valdez spill of March 1989. Three two-person crews will
perform foot surveys of intertidal and upstream portions of nine
major pink and chum salmon spawning streams. Port Dick Creek and
Island Creek will be surveyed every other day by the first crew,
Humpy Creek will be surveyed daily by a second crew, and the other
six streams will be surveyed at least once a week on a rotating
basis by the third crew. The third crew will be flown in and out
by aircraft. All surveys will be conducted during low tide between
July 7 and September 7.

Streams to be surveyed will be selected using the following
criteria:

I The stream must be included in the existing aerial survey
program.
2. The stream was examined in past spawning ground survey
programs.
3. A significant fraction of spawning occurs in the
intertidal area.

The nine streams studied in the Lower Cook Inlet area during 1989
were Windy Creek Left, Port Dick Creek, Windy Creek Right, and
Island Creek in the Kenai Fjords area and Humpy Creek, China Poot
Creek, Seldovia River, Tutka Lagoon Creek, and Port Graham Creek.
All but one historical alevin density index stream, Rocky River,
was examined. Rocky River was not included in the 1989 study
because the effects of logging on that river would have complicated
analysis.

Of the streams studied in 1989, Windy Creek Left and Port Dick
Creek have had oil deposited near the stream mouths, Windy Creek
Right and Island Creek had oil floating offshore, and the remainder
had no visible impact.

Three of the non-oiled streams in Kachemak Bay, China Poot Creek,
Tutka Lagoon Creek, and Seldovia River, will not be studied during
1990 because of logistical problems that inhibit sampling efforts
at these sites.

Three new non-oiled streams in the Kenai Fjords area, Tonsina Creek
(in Resurrection Bay) , South Nuka River and James Lagoon Creek,
will be added to facilitate comparisons with oiled streams. All
three of these streams have an intermittent history of fry digs.

During each stream survey the following data will be recorded:

1. Stream name;
2. Date and time;
3. Counts of live and dead salmon by observer, species and
location in the stream C(1) 0.0-0.6 m, (2) 0.6-1.2 m, and
(3) 1.2-1.8 m below mean high water, (4) the upstream
(above tidal inundation) egg-fry dig area, and (5) the
upstream area above the egg-fry dig];
4. subjective assessment of count quality: (1) the counts
are reasonably accurate, (2) the counts are not accurate
and a recount may provide a better estimate (e.g. lots of
fish in a deep pool, a glare problem, etc.) , (3) the
counts are not accurate but a recount will not change the
results (e.g. bad viewing conditions due to siltation,
wind, rain, etc.).
5. Location of tagged fish, tag type, number, and color
(Port Dick Creek and Humpy Creek only).
6. Observers name(s).

During foot surveys the numbers of spawning salmon will be

100

estimated by stream zone f or each study stream. Stream zones
represent levels of tidal influence; (1) 0.0-0.6 m below mean high
tide, (2) 0.6-1.2 m below mean'high tide, (3) 1.2-1.8 m below mean
high water, (4) the upstream (above mean high tide) egg-fry dig
area, and (5) the remaining area above the upstream egg-fry dig
area.

Stream zones were marked at each stream during the 1989 season.
Intertidal zones in the Lower Cook Inlet/Kenai Fiords area were
measured from the mean high tide level due to large differences in
mean tidal height between the gulf of Alaska (4 m) and Cook Inlet
(6 m) sides of the Kenai Peninsula. The stream bed location of the
tide levels 0.0, 0.6, 1.2, and 1.8 m below mean high water were
marked with a 0. 3 m 2 fluorescent orange plywood rectangle. The
markers were numbered consecutively 1 through 4 with number 1
furthest downstream at the 1.8 m below mean high water level. A
fifth marker will be added during the 1990 field season to identify
the upstream end of the egg-fry dig area (NRDA PIS study 8a). A
commercial hand held tide computer (Conex Electro-Systems model
TF0290W TideFinder) with time and location corrections will be used
to determine tide heights.

Maps for each stream will be rechecked for boundaries between
stream zones, distances across the streams within zones, and
distances between zone boundaries. Areas of spawning concentration
and preference by species within each stream zone will be recorded
on the revised maps. These maps will be used later by the Pink and
Chum Salmon Egg and Pre-emergent Fry Sampling project (NRDA PIS
8a).

Surveys will be at low tide and progress upstream from the 1.8 m
below mean high tide marker (marker number 1) . The upstream limit
of a survey will be determined by the presence of natural barriers
to fish passage (e.g. waterfall, log jam, etci), the end of the
stream, or the absence of spawning salmon.

Counts from each crew member will be recorded as an independent
observation. Crew members may either walk together or on opposite
banks of stream channels depending on terrain and viewing
conditions. Both crew members will walk up stream forks together
and not split up. Also, the crew will not divide tasks (e.g. one
member counts only live salmon while the other counts only
carcasses, etc.). survey partners will be rotated on a weekly
basis to prevent counting bias from being perpetuated. Crews will
be assigned to a different stream on each succeeding circuit.

Crews will begin each survey with a "practice count" for a short
distance. If their counts differ by more than 10%, they will
retrace their steps and search for the cause of the difference
(fish in a deep pool not clearly visible to both crew members, sun
glare, deep shadow, overhanging vegetation, etc.) and recount as
many times as necessary until they are satisfied that they can

101

compensate for visibility problems peculiar to their vantage
points. Likewise, crew members may warn each other and are
encouraged to discuss counting conditions in anticipation of
problems before they begin counting (e.g. discuss depth and breadth
of school before counting a deep pool of fish, warn each other of
difficult viewing conditions, etc.). Thereafter, each member will
count and record their data independently.

Both live and carcass counts will be made while walking upstream.
Carcasses will be marked at Humpy Creek and Port Dick Creek (e.g.
tail removed) to prevent double counting. Hand tally counters
(with 4 banks) will be used when counting. Recounts or stops to
record counts can be requested at any time and anywhere by either
crew member. At a convenient stopping point (e.g. a log jam,
either end of a deep pool, the start or end of silty water, etc.),
each crew member will record their counts and rate their counts as
follows:

1) the counts are reasonably accurate and a recount would
give very similar results,
2) the counts are not accurate and a recount may provide a
better estimate (e.g. lots of fish in a deep pool, a
glare, wind, or rain problem, etc.),
3) the counts are not accurate but a recount will most
likely not change the results (e.g. bad viewing
conditions due to siltation, wind, rain, etc.).

If both crew members rate their counts -1 or 3, then proceed to the
next stream section and count to the next stopping point. if
either or both crew members rate their counts 2, then both crew
members recount that section and record the results of the second
count. If necessary, both crew members could recount and record
the results for a third time. After the third trial, proceed and
count to the next stopping point.

Pink salmon will be tagged with individually numbered tags to
determine stream life and movement for Port Dick Creek and Humpy
creek. These streams will be surveyed on an every other day basis
with stream location, tag type, color, and number information
recorded for every tagged salmon observed. Stream life is expected
to vary over time and between sexes. Fifth (50) fish of each sex
will be tagged every other day, over an 11 day period. A beach
seine will be used to collect fish at the mouth of the streams for
tagging. The color-number combinations for Port Dick Creek will be
as follows:

Day Male Female Tag

1 orange 1-50) yellow (#51-100) Peterson disk
3 orange 1-50) yellow (#51-100) adhesive tape
5 orange 1-50) yellow (#51-100) rubber band
7 red (#101-150) green (#151-200) Peterson disk

102

9 red (#101-150) green (#151-200) adhesive tape
11 red (#101-150) green (1151-200) rubber band

The color-number combinations for Humpy Creek will be as follows:

Day Male Female Tag

1 orange (#201-250) yellow (#251-300) Peterson disk
3 orange (1201-250) yellow (#251-300) adhesive tape
5 orange (#201-250) yellow (#251-300) rubber band
7 red (#301-350) green (#351-400) Peterson disk
9 red (#301-350) green (1351-400) adhesive tape
11 red (#301-350) green (#351-400) rubber band

The numbered Peterson disk will be obtained from commercial
sources. The second tag type will be self adhesive tape wrapped
around the caudal peduncle with the two free ends attached together
and protruding up in to the air with a number written on the tape
while the third tag type will be a rubber band around the caudal
peduncle with 15 cm of numbered survey's tape attached.

A weir will be installed at Humpy Creek to provide known numbers of
fish in the stream. Foot survey counts will be made on an every
other day basis while aerial survey counts will be made in
conjunction with the Commercial Fisheries Division aerial survey
program. There will be no attempt to adjust foot or aerial survey
counts to match weir counts as the two are expected to differ.
Humpy Creek will be mapped and marked for shallow areas with a
clear view of the sky, deep pools with a clear view of the sky, and
areas with an overhead canopy. Counts in these areas will be
recorded separately as aerial and foot survey counts in these areas
are also expected to differ.

Data Analysis

Total number of salmon by species present at the time of the foot
survey will be estimated using a simple stratified sampling scheme.
Each stopping point will be considered the end of a sampling
strata. Thus, a stream zone could encompass many sampling strata.
The average and variance of the independent counts will be used to
estimate the number of fish present within the strata and the
variance about the estimate. Averages will be summed across stream
zones to provide estimates of numbers of fish within the zone and
zones summed to estimate numbers of salmon present during the
survey. Variances will be weighted by the number of fish estimated
in the strata and summed across stream zones and stream.

Stream life and its variance will be the average and corresponding
variance for the number of days a particular tag lot survives in
the stream. Changes through time and by sex will be examined.

Total escapement to each stream will be estimated using the area

103

under the curve method (similar to that described by Johnson and
Barrett 1988). The point estimates and variances from the f oot
surveys along with stream life estimates and variances will be used
to estimate total escapement and the corresponding variance.

Statistics that will be estimated include:

1. Number of spawning and dead salmon by species, stream
zone, stream, and date and the corresponding variances.
2. Stream life and variance :
3. Total escapement and variance.
4. Adjustment factors to relate aerial and foot surveys to
weir counts.

A composite sample of mussels (Mytilus sp.) will be collected at
the mouth of each stream for hydrocarbon analysis. A field blank
(sample container opened at the collection site, closed and stored
as if it contained a sample) and two sample replicates will also be
collected. Results of the analysis will be used to document the
level of oil impact sustained by the stream. Each sample will
consist of enough mussels to provide 10 grams of tissue for
analysis. The mussels will be collected from the immediate
vicinity of all streams. Collectors will use wooden tongue
depressors when possible. All mussels will be above water when
collected to prevent contamination by surface hydrocarbons. The
sample containers will be pre-rinsed (with dicloromethane) glass
jars with teflon lined lids as supplied by I-Chem. The samples
will be stored in padlocked containers And kept in a freezer in the
Homer ADF&G office. Appropriate chain of custody forms will
accompany each sample.

Streams will be divided into 2-3 categories based on levels of
hydrocarbon contamination (as determined from 1989 visual
observations and hydrocarbon level in mussel tissues from the 1989
and 1990 samples) . Counts of salmon by species and stream zone for
each stream will be assigned to one of the hydrocarbon categories.
Counts and spawner distribution will be compared with historical
stream survey data and related to the level of hydrocarbon impact.

BIBLIOGRAPHY

Johnson, B.A. and B. Barrett. 1988. Estimation of salmon
escapement based on stream survey data. Alaska Department of
Fish and Game. Division of Commercial Fisheries. Regional
Information Report No. 4K88-35. Kodiak.

104

BUDGET: ADF&G

Salaries $ 96.9
Travel 7.5
Contract 9.6
Supplies 3.3
Equipment 0.3

Total $ 117.6

105

FISH/SHELLFISH STUDY NUMBER 7b

Study Title: Injury to Pink Salmon Spawning Areas Within the
Kodiak and Chignik Areas

Lead Agency: ADF&G
INTRODUCTION

Large escapements into Kodiak and Chignik streams in 1989 occurred
as a result of severely limited commercial fishing opportunities
caused by the EVOS. Total pink salmon escapements in 1989 were
20.0 (Kodiak) and 1.4 (Chignik) million fish, whereas target
escapement goals were 4.0 and 0.7 million pink salmon in the Kodiak
and Chignik areas, respectively. The magnitude of escapement
experienced during 1989 is unprecedented and has the potential for
adversely impacting future returns of pink salmon through density
related factors such as fungus and disease outbreaks at the egg
and/or fry stage.

Pink salmon are a major component of the Kodiak and Chignik area
ecosystem, providing an important food source for both marine
mammals, terrestrial mammals, birds, and other fish and shellfish.
Additionally they annually re-charge fresh water and near shore
marine environments with nutrients as carcasses decompose after
spawning. This species is used for subsistence, sport and
commercial purposes. Annual ex-vessel value (1978-1988) of the
pink salmon harvest is 14.2 and 1.45 million dollars for the Kodiak
and Chignik areas, respectively (Malloy 1989; Thompson and Fox
1989).

Pink salmon commercial fisheries are managed, in part, by
controlling escapement which is evaluated by aerial and weir
counting methods. Aerial surveys for pink salmon escapement indices
using fixed wing aircraft and trained observers have been conducted
annually for over 30 years in the Kodiak and Chignik management
areas. Total pink salmon escapement and spawner density estimates
by index stream, geographical region and management area will
provide the basis for quantifying the effects of large escapements
realized during 1989, on forthcoming brood year returns. Estimates
of total available pink salmon spawning habitat will permit
assessment of production from empirical escapement densities, and
allow for determination of optimum spawning density by geographical
area which will be useful for restoration efforts,.if needed.

OBJECTIVES

A. Estimate total pink salmon escapements for streams where
historic pre-emergent sac fry density data exist. This
includes 44 Kodiak and 18 Chignik streams.

106

B. Def ine the distribution of spawning pink salmon for index
streams within the Kodiak and Chignik management areas. This
entails mapping and photographing spawner distribution.

C. Estimate total available spawning habitat for index streams
within the Kodiak and Chignik management areas.

METHODS

There are two integral components of this investigation: 1) weir
counts and repeated aerial and foot surveys for escapement and
stream life calculations, and 2) stream surveys for collection of
spawning habitat data necessary for calculating total available
spawning habitat.

Trained observers will conduct aerial surveys using fixed wing
aircraft on the 44 Kodiak and 18 Chignik pre-emergent index
streams. Surveys will be conducted weekly on each index stream
with the program continuing until at least seven surveys over the
spawning period have been completed or when spawner counts have
decreased to less than 10 percent of the observed peak count.
Additional non-index streams (342 Kodiak and 72 Chignik) will be
surveyed as time and aircraft availability permit. The following
information for each survey will be collected: 1) stream name and
statistical number; 2) date, weather conditions, fish visibility
rating (poor, fair or good) and time; 3) observer, aircraft type
and pilot; 4) number of live and dead fish of each species (in bay,
mouth and stream); and 5) general survey comments. Data will be
recorded on standard forms suitable for data entry into the
regional survey database. The observer, upon completion of surveys
for a particular index stream, will map spawner distribution and
designate the percent of spawning area used. Aerial photographs
will be taken of spawner distribution for all index streams
surveyed.

counting weirs will be located at Akalura, Litnik, Saltery,
Paramanof, Barling, and Uganik (Pillar Creek will be monitored by
foot survey). Pink salmon weir counts will be made daily and
recorded on standardized forms. once every three days throughout
the spawning period, foot surveys will be conducted by weir crews
and additional personnel to enumerate live and dead fish by
species. Data collected during stream surveys will be recorded on
standardized forms. Daily weir count and foot survey data will not
be reported until after the project has been completed. Aerial
surveys will be completed for all weired index streams with a
minimum frequency of one survey per week. Data collected during
these surveys will be kept separate from the routine aerial survey
data. Information collected for this study component will allow
for calculation of stream life by system, in-stream population
estimates, and aerial survey calibration.

107

At each weir where stream lif e is being estimated, adult pink
salmon will be tagged with 0. 3m, color coded Floy tags with a
marking rate of 200 per week. Tagging will commence approximately
July 21st and continue for four weeks. Tags will be affixed on a
single day each week by capturing f ish in a trap located on the
upstream side of each weir. Each week will have a specific tag
color for identification of that tagging lot. Recovery effort will
consist of enumeration of specific color coded fish during foot
surveys and tags will be recovered and counted from mortalities.

During the 1989 field season, total available spawning habitat was
estimated for 31 Kodiak and 14 Chignik preemergent index streams.
The remaining 17 index streams will be surveyed in 1990 to
determine total available spawning habitat.

Relying upon previously constructed maps of Kodiak and Chignik
index streams with defined limits of historic spawner distribution,
total and available stream length will be calculated. The
available stream length component will be divided into 300 meter
sections, which will be randomly selected for surveying. Within
each section, 12 transects spaced 25 meters apart will be run
perpendicular to the stream bank.

Pink salmon spawning habitat can be envisioned as a continuum of
water velocity, depth, substrate size and embededness. According
to Raleigh and Nelson (1985), substrate size and water velocity
have the greatest influence on spawning success of pink salmon,
however, substrate embededness may also have an impact (Platts et.
al. 1983). At each transect surveyed, stream width, water
velocity, depth and substrate embededness will be assessed. Values
used for these criteria are founded upon averages derived from the
literature (Andrew and Geen 1960; Chambers 1956; Divinin 1952;
Krueger 1981; Neave 1966; Wilson et. al. 1981). Available spawning
habitat will be considered as an area along a transect where depth
is a minimum of 15.2 cm, substrate size is within 0.6 to 13.8 cm,
water velocity is 0.3 to 0.9 m/sec and substrate embededness is
such that gravel is displaced without excessive foot pressure.
Spawning habitat will be recorded as a percent of the total area
encompassed by a one meter band along the transect line. Data will
be entered onto standardized forms and later entered into the
regional database. All field personnel responsible for data
collection will have prior experience with assessing habitat
variables.

An additional task of the spawning habitat inventory program will
be to calculate total spawning area for all of the spawning riffles
sampled in the pre-emergent fry dig study (F/S Study 8b).
Estimates will be derived for 44 Kodiak and 18 Chignik Area
streams. The exact location and approximate dimensions for each
riffle will be obtained from detailed maps of individual index
streams and from consultation with pre-emergent fry dig personnel.
Length of the riffle will be measured on a straight line transect

108

measure, while width measures will be made one meter apart on both
sides of the line for the entire length of the riffle. All
measurements will be in meters and recorded on standardized forms.
only substrate size, velocity and substrate embededness variables
from the above methods will be used to delineate spawning habitat
in this context.

Stream life of pink salmon is the length of time an adult is alive
in the freshwater environment. In 1989, stream life estimates were
successfully derived from weir and foot survey counts incorporated
into the Johnson and Barrett (1988) model for two Afognak Island
systems. Weir and foot survey counts of both live and dead fish
collected during 1990 will provide for a maximum of 30 data points
per system to be available for stream life analysis. Two
analytical approaches will be used to calculate stream life in
1990. The first will be identical to that used in 1989, while the
second will use cumulative and periodic dead fish counts. The
first approach relies upon the cumulative weir count and periodic
foot survey counts being entered into the model and the stream life
value being iterated until the model output converges upon the
cumulative weir count. The second approach will involve the
cumulative dead fish count, periodic foot survey dead counts, and
the Johnson and Barrett (1988) model, following the steps outlined
above. This analysis, in addition to providing a second stream
life estimate, will also allow for calculating a combined washout
and predation rate (difference between cumulative dead and
cumulative weir counts) which will allow for calibrating counts
derived from stream systems surveyed without weirs. Computed
stream life values will be statistically tested and if significant
differences exist, system specific stream life values will be
evaluated based upon the variables stream order, orientation,
geomorphology and stream length.

Stream life estimated from 'tagging data will consist of determining
the point at which f if ty percent of the tagged f ish f or a
particular color code have been recovered. This approach will be
carried out f or each of the f our weeks f or which tags have been
affixed. An overall stream 1 -".fe estimate will consist of averaging
the estimates f or each of -he four weeks and also comparisons
between systems and weeks following the previously mentioned
analytical approach for between system tests.

Instantaneous live pink salmon population size will be estimated
from cumulative weir counts and periodic (every three days) foot
survey counts of dead pink salmon, Relying upon either a linear or
exponential model with in-stream population estimates and aerial
survey counts as parameters, precision of aerial survey counts by
geographical location and escapement magnitude can be quantified.
From this analysis, calibration factors can be determined for use
in total escapement estimation procedures.

Temporal aerial survey escapement counts of pink salmon in spawning

109

streams are, depending upon the frequency and timing of the
surveys, related to total or cumulative escapement. Defining and
quantifying variables or relationships that allow transformation of
aerial survey counts into reliable estimates of total escapement is
the major task. The Johnson and Barrett (1988) geometric model is
one such approach to estimating total escapement. Two data
components are required for the algorithm, escapement counts over
time and stream life. The unit of measurement is area of the
spawner abundance curve derived from a series of survey counts.
Two segments comprise the analytical phase of the model, the first
is calculating number of fish present between survey counts and the
second is deriving total escapement.

Total escapement accuracy, according to Johnson and Barrett (1988),
is related to precision of escapement counts and stream life
estimates. The influence that both escapement counts and stream
life estimates have on the total escapement estimate will be
evaluated. Total escapement estimates will be calculated for all
index and non-index streams for both the Kodiak and Chignik
management areas where aerial and or foot surveys have been
completed. Total escapement will also be estimated for index
streams based upon the historic aerial survey data base (1968-1988)
so that all data which will be utilized in future analyses will
have been derived in a similar fashion.

Available spawning habitat will be determined using equations
identified in Cochran (1977) and Wolter (1984), (personal
communication, Alan Johnson Regional Biometrician, ADF&G, Kodiak).
These equations will allow for calculation of the total available
spawning habitat and variances associated with the estimates. The
estimates can then be used for density estimates for spawning pink
salmon for the 1989 brood year in addition to assessing the
relationship between density and subsequent returns from previous
brood year spawning events.

To derive estimates of potential egg deposition to pre-emergent fry
survival, an estimate of total area available for pre-emergent fry
sampling is necessary. The area sampled for an individual fry.dig
when coupled with the total area estimate will allow for expansion
of the fry dig data to an estimate of survival. Total area of
sampling riffles will be estimated by using the habitat measures
for each riffle and summing over all riffles which are sampled for
pre-emergent fry. An estimate for each index stream sampled within
the Kodiak and Chignik management areas will be derived. These
estimates form the foundation of an analysis component conducted as
part of FIS Study 8b.

[email protected] distribution maps will be prepared, in part, by observers
conducting aerial survey escapement counts as was done for the 1989
return. In addition, aerial photographs will also be obtained and
cataloged for reference.

110

BIBLIOGRAPHY

Andrew, F.J. and G.H. Geen 1960. Sockeye and Pink salmon
production in relation to proposed dams in the Fraser River
system. International Pacific Fisheries commission Bull.
No.11.

Chambers, J.S. 1956. Research relating to study of spawning
grounds in natural areas. U.S. Army Corps. Eng. No. Pac. Div.
Fish Eng. Res. Prog. 6pp.

Cochran, W.G. 1977. Sampling Techniques. John Wiley, New York.

Divinin, P.A. 1952. The Salmon of South Sakhalin. Investia-Tinro.
37:69-108.

Frisell, C.A. and W.J. Liss 1986. Classification of stream habitat
and watershed systems in south central Oregon. Unpub. Progress
report. Oak Creek Lab. Corvalis Or.

Hankin, D.G. and G.H. Reeves 1988. Estimating total fish abundance
and total habitat area in small streams based on v i s u a 1
estimation methods. Canadian Journal of Fisheries and Aquatic
Sciences. 45:1413-1424.

Johnson, B.A. and B.M. Barrett 1988. Estimation of salmon
escapement based on stream survey data: A geometric approach
Alaska Department of Fish and Game, Division of Commercial
Fisheries, Kodiak. Regional Information Report. No. 4K88-35.
8pp-

Krueger, S.W. 1981. Freshwater habitat relationships, pink salmon
(Oncorhynchus gorbuscha). Alaska Department of Fish and Game,
Anchorage. 41pp.

Malloy, L.M. 1989. 1988 Kodiak area salmon management report to
the Alaska board of fisheries. Alaska Department of Fish and
Game, Division of Commercial Fisheries, Kodiak. Regional
Information Report. No. 4K89-5. 171pp.

Murphy, M.L., J.M. Lorenz, J. Heifetz, J.F. Thedinga, K.V. Koski,
and S.W. Johnson 1987. The relationship between stream
classification, fish, and habitat in Southeast Alaska. Tech.
Bull. 12., Tongass National Forest. R10-MB-10.

Neave, F. 1966. Salmon of the North Pacific Ocean- Part III. A
review of the life history of North Pacific pink salmon in
British Columbia. International North Pacific salmon
fisheries commission Bull. No. 18:71-78.

Platts, W.S., W.F. Megahan and G.W. Minshall 1983. Methods for
evaluating stream, riparian and biotic conditions. U.S. Dept.

III

of Agriculture, General Tech. Rept. Int-138.

Raleigh, R.F. and P.C. Nelson 1985. Habitat suitability index
models and instream flow suitability curves: Pink salmon.
U.S. Fish and Wildlife Service. Biol. Rept. 82(10.109). 36pp.

Thompson, F.M. and J.R. Fox 1989. Chignik management area annual
f inf ish management report, 1988. Alaska Department of Fish and
Game, Division of commercial Fisheries, Kodiak. Regional
Information Report NO. 4K89-5. 171pp.

Wilson, W.J., E.W. Trihey, J.E. Baldridge, C.D. Evans, J.G. Thiele,
and D.E. Trudgen 1981. An assessment of environmental effects
of construction and operation of the proposed Terror Lake
hydroelectric facility, Kodiak, Alaska. Arctic Environmental
Information and Data center, Univ. of Alaska. Anchorage.
419pp.

Wolter, K. M. 1894. An investigation of some estimators of variance
for systematic sampling. Journal of the American Statistical
Association. 79:781-790.

BUDGET: ADF&G

Salaries $251.7
Travel 5.0
Contracts 132.2
Supplies 49.3
Equipment 22.1

Total $460.3

112

FISH/SHELLPISH STUDY NUMBER Sa

Study Title: Injury to Pink and Chum Salmon Eggs and Pre-Emergent
Fry Within Lower Cook Inlet and Kenai Fjord

Lead Agency: ADF&G

INTRODUCTION

Wild stocks of pink and chum salmon are a major ecosystem component
in the outer'Kenai Peninsula and Lower Cook Inlet area, immediately
"down current" from Prince William Sound. Salmon represent a very
important food source for marine mammals (sea lion, seals),
terrestrial mammals (bear) , and a wide variety of bird species
(eagles, etc.) . In addition salmon are harvested commercially. In
1988, the year before the oil spill, the ex-vessel value of the
commercial catch of wild and hatchery stocks of salmon from the
lower Cook Inlet/Kenai Peninsula (CIK) area was more than $8.2
million. Salmon are also very important to the sport, subsistence,
and personal use fisheries. The future abundance of wild stocks of
pink and chum salmon in the lower CIK areas may be adversely
impacted as their intertidal spawning areas were affected by the
oil spill. This project continues the evaluation of pink and chum
salmon egg to fry survival in the intertidal spawning areas
affected by the EVOS.

OBJECTIVES

A. Estimate abundance of pink and chum salmon eggs and pre-
emergent fry by intertidal and upstream areas for nine streams
in the lower CIK. Six of the streams were studied in 1989.
Three unoiled streams in Kachemak Bay were dropped while three
Gulf of Alaska streams were added to provide a better
comparison of oiled and unoiled streams in the Gulf of Alaska
area.

B. Estimate overwinter mortality (egg to pre-emergent fry) of
pink and chum salmon eggs.

C. Estimate reductions, if any, in pink and chum salmon pre-
emergent fry abundance due to oiling.

METHODS

Sampling will be conducted in two phases: egg-digs performed in
October and pre-emergent fry digs conducted in March. The number
of streams to be studied is limited by the number of days in
October and November with low tides (maximum of +4.0 feet) during
daylight hours.

Streams were selected using the following criteria:

113

1 Sufficiently large adult salmon returns to indicate a
high probability of success in egg/fry digging.
2. Past history of egg/fry digging.
3. Streams covered by FIS Study 7a and aerial escapement
survey project.
4. Streams can be saf ely studied during the winter and early
spring months.

The nine streams studied during 1989 were Windy Creek Left, Port
Dick Creek, Windy Creek Right, and Island Creek in the Kenai Fjords
area and Humpy Creek, China Poot Creek, Seldovia River, Tutka
Lagoon Creek, and Port Graham Creek in the Cook Inlet area. All
but one historical alevin density index stream, Rocky River, was on
that list. Rocky was not included in the 1989 study because the
effects of logging would have confused the results.

of the streams studied in 1989, the first two have had oil
deposited near the stream mouths, the next two have had oil
floating offshore, and the remainder had no visible impact.

Three new non-oiled streams in the Kenai Fjords area, Tonsina Creek
(in Resurrection Bay), South Nuka River and James Lagoon Creek,
will be added to facilitate comparisons with oiled streams on the
Gulf of Alaska. All three were once considered non-index alevin
density streams with an intermittent history of fry digs.

Sampling methods are identical for the pre-emergent fry and egg
digs. On each sample stream, four zones, 3 intertidal and one
above tidal inundation, will be identified and marked by crews
conducting stream surveys during F/S 'Study 7a. The zones are 0.0-
0.6 m, 0.6-1.2 m, and 1.2-1.8 m below mean high water, and upstream
of tidal inundation.

Separate linear transects will be established in each zone (one
transect for each type dig). The ttansects will run the entire
length of the zone. Overlapping of transects will be kept to a
minimum to control the influence of fall egg digs on abundance of
fry during spring 2 sampling. Fourteen circular digs (56 per
stream) , each 0. 3 m in size, will be systematically dug along each
transect using a high pressure hose to flush eggs and fry from the
gravel. Eggs and fry will be caught in a specially designed net.
Areas where salmon were not observed spawning during the spawning
ground surveys (FIS Study 7a) will be avoided. Numbers of live and
dead fry by species as well as numbers of live and dead eggs by
species will be collected from each 0.3 m2 dig. Additional
information such as date, time, and zone will also be collected.

114

Eggs and fry will be collected for MFO analysis. A sample of 40
fish will be preserved in a buffered formalin solution.

A composite fry sample will be collected from the intertidal area
for hydrocarbon analysis. A field blank (sample container opened
at the collection site, closed and stored as if it contained a
sample) will also be collected. Each sample will consist of enough
fry to provide 10 grams of tissue (about 110 fry) for analysis.
The sample containers will be pre-rinsed glass jars with teflon
lined lids. The samples will be kept frozen until shipment for
processing in Auke Bay. Appropriate chain of custody forms will
accompany each sample.

A mixed effects analysis of covariance will be used to test for
differences in egg to fry survival due to oiling. The level of
hydrocarbon impact will be determined from hydrocarbon analysis of
mussels collected in 1989 and 1990 by F/S Study 7a.

Analysis of variance will be used if no suitable hydrocarbon data
are available. Degree of oiling as visually assessed by F/S Study
7a will be used to post-stratify streams. Degree of oiling and
height in the tidal zone will be treated as fixed effects. Height
in the tidal zone is nested within stream, a random effect.

The number of streams sampled is limited by the window of time
available for sampling. Power was estimated for the ANOVA using
data from the 1975 and 1976 egg and pre-emergent fry digs in PWS.
This analysis indicated the ANOVA could detect an increase of 20%
(e.g. 10% mortality to 30% mortality) in egg to fry mortality at a
= 0.05, 90% of the time.

An assessment of lost fry production will be made if differences in
egg to fry survival due to oiling are detected. Average survival
from unoiled areas will be used to estimate potential fry density
in oiled areas. observed and potential fry densities will then be
expanded to estimate total observed and potential fry. The
difference between the two estimates will be considered lost fry
production.

Specific statistics to be estimated are:

1. Number of dead and viable eggs per square meter by salmon
species, stream, and stream zone.
2. Number of dead and live fry per square meter by salmon
species, stream, and stream zone.
3. Egg to fry survival by salmon species, stream, and stream
zone.
4. Lost production by salmon species, stream, and stream
zone.

115

BUDGET: ADF&G

salaries $46.9
Travel 1.7
Contracts 19.7
Supplies 1.3
Equipment 1.4

Total $71.0

116

FISH/SHELLFISH STUDY NUMBER 8b

Study Title: Injury to Pink Salmon Egg and Pre-Emergent Fry In the
Kodiak And Chignik Management Areas

Lead Agency: ADF&G

INTRODUCTION

Large escapements of pink salmon into Kodiak and Chignik streams in
1989 occurred as a result of severely limited commercial fishing
opportunities caused by the EVOS. Total pink salmon escapements in
1989 were 20.0 (Kodiak) and 1.4 (Chignik) million fish, whereas
target escapement goals are 4.0 and 0.7 million pink salmon in the
Kodiak and Chignik areas, respectively. The escapement magnitude
experienced during 1989 is unprecedented and has the potential for
adversely impacting future returns of pink salmon through density
related factors such as fungus and disease outbreaks at the egg
and/or fry stage.

Pink salmon are a major component of the Kodiak and Chignik area
ecosystems, providing an important food source for both marine
mammals, terrestrial mammals, birds, and other fish and shellfish.
Additionally they annually re-charge fresh water and near shore
marine environments with nutrients as carcasses decompose after
spawning. This species is used for subsistence, sport and
commercial purposes. Annually, pink salmon comprise 78% and 31%
(1978-1988) of the Kodiak and Chignik salmon harvest, respectively.
Ex-vessel value (1978-1988) of the pink salmon harvest is 14.2 and
1.5 million dollars for the Kodiak and Chignik areas (Malloy 1989;
Thompson and Fox 1989).

A total of 386 Kodiak and 90 Chignik streams support populations of
pink salmon. Pre-emergent sac fry sampling has been conducted in
44 Kodiak and 18 Chignik streams periodically over the last 20
years. These streams, referred to as index streams, provide data
which are utilized for projections of returns and potential
harvest.

Potential damage caused by the 1989 brood year escapements upon
future brood year returns can be quantified by: 1) examination of
observed versus expected numbers of live fry/dig produced from
potential egg deposition; 2) comparison of 1989 potential egg
deposition to pre-emergent fry survival for the odd years 1969 to
present; 3) evaluation of numbers of live fry/dig for streams with
optimum spawning density versus streams with spawner densities
above optimum.

OBJECTIVES

A. Estimate potential egg deposition for all Kodiak and

117

Chignik pre-emergent index streams.

B. Estimate pink salmon fry density for Kodiak and Chignik index
streams.

C. Estimate pink salmon survival from potential egg
deposition to pre-emergent fry.

D. Assess changes, if any, of pink salmon pre-emergent fry
abundance in 1991 due to the oil spill.

E. Estimate the 1991 adult pink salmon return by using the 1990
fry index data.

METHODS

Potential egg deposition (PED) for each of the 62 Kodiak and
Chignik Management Area index streams will be determined using
index stream total escapements and fecundity data collected during
the 1989 field season. The PED estimates will be based upon
average fecundity derived from the relationship of fish length to
number of eggs carried and the total escapement estimates derived
using the Johnson and Barrett (1988) model.

Pre-emergent sac fry sampling will be conducted on 44 Kodiak and 18
Chignik index streams. A majority of these streams. have been
frequently and consistently sampled each year. Sampling station
(spawning riffles) selection is founded upon pink salmon spawner
distribution and specific habitat utilization as recorded from
aerial surveys. The number of sampling sites per stream is based
upon escapement magnitude, stream size and observed productivity of
individual index streams. Generally, smaller streams where
escapements average less than 20,000 will have 4-6 sampling
stations, 6-9 stations for the larger intermediate sized streams
and the most productive streams, will have 10-15 stations.
Historically, 10 digs have been completed for each station using a
pump and associated equipment which hydraulically remove pink
salmon eggs and fry (both live and dead) from the stream bed. A
collection frame is used to capture eggs and fry as they are
displaced from the gravel. Depth of stream bed sampling is 15. 0 to
46.0 cm with a duration of 1-3 minutes depending on substrate.
After eggs and fry (both live and dead) are enumerated the
collection frame is moved to the next dig location and the steps
repeated. Sampling is done in an X configuration with equal
numbers of digs done above and below the center of the X. Digs
which are at the extremes of the configuration are those which are
closest to the stream banks. Ancillary information recorded along
with egg and fry counts are stream temperature, predator presence,
stage of fry development, quantities of egg fragments and evidence
of stream bed scouring or shifts (Brennan 1990). only minor
modifications have occasionally beset the above sampling program

118

and were associated with water conditions, ice coverage or flood
events which had altered the stream channel. Presently the only
modification which will be imposed on the sampling program for 1990
will be that a minimum of 30 digs with at least one live fry be
obtained for each stream sampled, regardless of the historic number
of digs done for that system (Johnson 1990). Alternative stations
for additional digs to meet this constraint will be from
established sampling sites. Pre-emergent sac fry sampling will be
conducted in a time frame which will minimize the chances of fry
emigration prior to sampling.

Determination of egg to pre-emergent fry survival (1989-1990) will
be founded upon PED, live fry/dig, and habitat data collected from
F/S Study 7b. An estimate of spawning density for all spawning
riffles sampled for pre-emergent fry will be obtained from the
detailed spawner distribution maps.

Utilizing PED and number of live fry/dig data spanning the odd
years 1969 to 1989, analyses will consist of fitting and assessing
an empirical relationship. If required for quantifying possible
outlying data points, climatological variables (precipitation and
mean monthly temperature) will be assessed as possible causative
factors. Damage from this analytical standpoint would be live
fry/dig values which fall below (descending limb of curve) the
range of expected PED. Hypothesis testing using non-parametric
tests will be used to assess whether observed differences in number
of live fry/dig are statistically significant. A signif icance
level of 0.1 will be used for all statistical analyses.

The observed difference between potential eggs deposited and
resultant pre-emergent fry will be designated as fry survival for
a given year. The equation which will provide the estimate will be
from Snedecor and Cochran (1967, p.520) with N defined as spawning
area.

PED in this framework will be determined from the proportion of the
total escapement which during odd years utilize this fraction of
the overall spawning habitat. Again, data for the odd years 1969-
1989 will be used and comparisons of eggs and fry densities for
contrasting levels of escapements and years will be made. if
needed, climatological conditions will be assessed in relation to
calculated survival values. Damage, if any, due to escapement
levels experienced in 1989 will be quantified from this method.

This analysis component will take into account available spawning
habitat, total estimated escapement and pre-emergent fry dig data.
Control streams will be designated based upon a cumulative ranking
of escapement and total available habitat in which the overall
density of all streams will be 1.3 fish per M2 or less. All index
streams that fall outside of this classification will be designated
as treatment streams (those with spawner densities above the
calculated optimum of 1.3 fish/M2). There are 17 control and 14

119

treatment streams within the Kodiak area and 11 control and 3
treatment streams in the Chignik area. Analysis will consist of
comparing the live fry/dig data in composite for the control versus
treatment streams. An alternative method will use an independent,
mutually exclusive classification method for each index stream.
The above (control versus treatment) method incorporates streams
with density values that are above the defined optimum for the
control group. Index streams (13 Kodiak and 4 Chignik) without
available spawning habitat estimates will be included in this
analysis when estimates become available.

BIBLIOGRAPHY

Brennan, K.R. 1990. History of the pink salmon Pre-emergent fry
sampling program in the Kodiak and Chignik management areas.
Alaska Department of Fish and Game. Unpublished manuscript.
12pp.

Johnson, B.A. 1990. Detecting changes in pink salmon
(Onchorhynchus gorbuscha) fry density for Humpy Creek, Alaska,
and determination of sample size requirements. Alaska
Department of Fish and Game, Kodiak. Regional Information
Report No. 4k9O-3. 6pp.

Johnson, B.A. and B.M. Barrett 1988. Estimation of salmon
escapement based on stream survey data: A geometric approach.
Alaska Department of Fish and Game, Division of commercial
Fisheries, Kodiak Regional Information Report No. 4K88-35.
8pp.

Malloy, L.M. 1989. 1988 Kodiak area salmon management report to
the Alaska board of fisheries. Alaska Department of Fish and
Game Division of Commercial Fisheries, Kodiak. Regional
Info@mation Report No. 4k89-6. 72pp.

Snedecor, G.W., and W.G. Cochran 1967. Statistical Methods. Iowa
State University Press, Ames, Iowa.

Thompson, F.M. and J.R. Fox 1989. Chignik Management Area Annual
Finf ish Management Report, 1988. Alaska Department of Fish and
Game, Division of Commercial Fisheries, Kodiak. Regional
Information Report No. 4k89-5. 171pp.

BUDGET: ADF&G

salaries $ 86.4
Travel 1.1
Contracts 53.0
Supplies 8.8
Equipment 0.0

Total $ 149.3

120

FISH/SHELLFISH STUDY NUMBER 11

Study Title: Injury to PWS Herring

Lead Agency: ADF&G

INTRODUCTION

The oil spill in PWS coincided with the annual migration of Pacific
herring Clupea harengus to near-shore spawning areas. In 1989, a
significant portion of the spawning area in PWS was located within
areas contaminated by oil. Additionally, adult spawning herring and
newly hatched juveniles traversed areas impacted by oil and beach
cleaning activities.

It was hypothesized that the oil spill would adversely impact adult
fish through direct mortality, food shortages, slowed growth, and
a possible reduction in fecundity. In addition, herring eggs have
been shown to be particularly susceptible to hydrocarbon
contamination due to the affinity of hydrocarbon compounds for yolk
sac material. Although no significant acute mortality was observed
for adult fish in 1989, significant impacts were measured on egg
mortality, egg hatching success, and percent viable hatch. Any of
these adverse effects have the capacity to reduce the abundance and
availability of herring. Adult and juvenile herring,-as well as
herring eggs, often form an important item in the diet of marine
fishes (e.g. salmon and halibut), mammals (e.g. sea lions, seals,
and whales), and birds (e.g. cormorants, ducks, puffins, gulls).
Herring also support an important commercial fishery within PWS,
worth over 12 million dollars in 1988.

The goal of this project is to determine whether the EVOS will have
a measurable impact on populations of Pacific herring in PWS.
Accurate and precise estimates of population abundance, age
structure, weight, and length composition data are needed to
accomplish this goal. In addition, the direct effects of oil
contamination on spawning success and egg survival will be
determined.

OBJECTIVES

A. Expand the normal sampling of herring populations in PWS to
increase the precision of herring abundance, age composition,
weight, sex ratio, and fecundity estimates. Specifically we
intend to:

Continue to estimate the biomass of the spawning stock of
herring in PWS such that the estimate is within � 25% of the
true value 95% of the time;

121

Estimate the age, weight, length (AWL) , and sex composition of
herring in PWS during 1989 such that age composition estimates
are within � 10% of their true values 95% of the time;

B. Continue to document the occurrence of herring spawn in oiled
and unoiled areas, validating the sites with quantified oil
level information obtained from shoreline survey maps and
hydrocarbon analysis of 1989 and 1990 herring eggs and mussel
tissue.

C. Continue to estimate hydrocarbon contamination of, and
physiological impacts on, adult herring by analyzing tissue
samples:

Test the hypothesis that the level of hydrocarbons in herring
tissues is not related to the level of oil contamination of
the area from which the herring were sampled. The experiment
is designed to detect a difference of 1.6 standard deviations
in hydrocarbon content with the probability of making a type
I and type II error of 0.05 and 0.1, respectively.

Estimate the presence and type of damage to tissues and vital
organs of herring sampled from oil-impacted and un-impacted
areas.

Test the hypothesis that the level of hydrocarbons in herring
eggs is not related to the level of oil contamination of the
area from which the herring were sampled. The experiment is
designed to detect a difference of 1.6 standard deviations in
hydrocarbon content with the probability of making a type I
and type II error of 0.05 and 0.1, respectively.

D. Continue to estimate the proportion of dead herring eggs from
a subsample of study sites in oiled and un-oiled areas that
were utilized in the 1989 egg mortality study, expanding the
data base and providing sample sites for sample collection of
live and preserved eggs. In addition, add an egg loss study
at the egg mortality sites to increase the accuracy of the
spawn deposition biomass estimates.

E. Continue to estimate the hatching success, viable hatch,
occurrence of abnormal larvae, and collect embryonic and
larval tissue for sublethal testing including MFO,
cytogenetics, RNA/DNA ration analysis, and others by
collecting herring eggs from egg mortality sites and control
sites in Southeast Alaska (Sitka Sound) and rearing them under
laboratory observation.

METHODS

This project will be conducted in three parts: (1) herring spawn
deposition estimation; (2) herring age, weight, length, growth, and

122

fecundity estimation; and (3) herring egg survival and egg loss
estimation.

Herring Spawn Deposition Estimation

The management of the PWS herring stock is based on a harvest
policy established by the Alaska Board of Fisheries which specifies
a maximum 20% exploitation rate for the combined harvest of all
herring f isheries. The allowable harvest is based on biomass
estimates established the previous year modified by the expected
growth and survival over the year. While aerial surveys were used
to estimate biomass from 1973-87, spawn deposition surveys were
performed in 1983 (Jackson and Randall 1983) and 1984 (Jackson and
Randall 1984), and began to be used as the primary biomass estimate
in 1988 (Biggs and Funk 1988). Aerial surveys are easier to
perform than spawn deposition surveys, but aerial survey biomass
estimates are not as reliable because of the varying visibility of
herring schools from the air and because the residence time of
herring schools on the spawning grounds is unknown. Estimates of
precision are not available for aerial survey biomass estimates.
The ADF&G continues to conduct an annual aerial survey of spawning
biomass to provide in season indicators of run timing and location
and to collect information on the timing and distribution of
spawning activity that is used for planning the spawn deposition
survey.

This project represents an augmented program to assess the PWS
herring stock's response to the EVOS. The original goal of the
1989 herring spawn deposition survey was to estimate the spawning
biomass with a precision such that the biomass estimate would be
within � 25% of the true biomass estimate 95% of the time under
optimal survey conditions. Fishery managers determined that this
level of precision was acceptable for estimating exploitation rates
and forecasting future abundance. If weather or other logistic
problems hampered the spawn deposition survey sampling effort,
fishery managers were willing to tolerate reduced precision. The
EVOS introduced a potentially new and unknown level of mortality on
herring stocks. The accuracy and precision of estimates of stock
abundance need to be assured from both oiled and unoiled areas (as
reflected in objectives 1 and 2). The opportunity to estimate
herring biomass with spawn deposition surveys is only available
during a relatively narrow two week window. After the oil spill,
the number of divers involved in the survey was increased to assure
that even if weather problems restricted the available sampling
time, sufficient numbers of transects could still be performed.
The number of transects was also increased to provide a level of
precision such that the biomass estimate would be within � 25% of
the true biomass 95% of the time. The amount of time devoted to
skiff surveys of spawning areas was also increased. Skiff survey
delineation of spawning area boundaries should help to increase the
level of precision of spawn deposition surveys and provides
important documentation of the occurrence of herring spawn in oiled

123

and unoiled areas.

The aerial survey project will provide a map indicating the general
location of herring spawning areas. A skif f survey will then
delineate the boundaries of each spawning area in more detail.
Transects will be placed perpendicular to the shoreline at
locations selected randomly from the shoreline maps of spawning
areas. Divers will swim along the transects and systematically
place 0.1 m2 quadrants at 5 m intervals. Divers will estimate the
total number of eggs in each quadrant. All egg-containing
vegetation will be removed from a subset of the quadrants for later
enumeration of the number of eggs in a laboratory procedure. These
enumerated egg counts will be used to correct bias in diver-
estimated egg counts and estimate the precision of the diver
estimates. The survey design is described in detail by Biggs and
Funk (1988), and follows closely the two-stage sampling design of
similar surveys in British Columbia (Schwiegert et al. 1985), and
in Southeast Alaska (Blankenbeckler and Larson 1982, 1987). The
surveys use random sampling at the first stage (transects) , and
systematic sampling at the second stage (quadrants within
transects). Random sampling in the second stage is not feasible
because of underwater logistical constraints (Schwiegert et al.
1985). In addition to the two-stage design, the survey is
stratified by five areas within PWS (Southeast, Northeast, North
Shore, Naked Island and Montague), because of the geographic
separation of these areas and the potential for herring in these
areas to be discrete stocks.

mean egg densities along each transect will be combined to estimate
an overall average egg density. The observed widths of the
spawning bed along each of the transects will be used to estimate
the average spawning bed width. The average width, average
density, and total spawning bed shoreline length (verified from the
skiff survey) will be used to estimate the total number of eggs
deposited in each of five area strata established within PWS.
Using the average fecundity and sex ratio derived from the AWL
sampling portion of this project, the total number of eggs
deposited will be converted into population numbers and biomass.
Based on the variances obtained during the 1989 survey, 160
transects would be needed to insure that the estimated biomass
would have a 95% chance of being within 25% of the true biomass
(161 transects were conducted in 1989 with a 95% chance of being
within 19% of the true biomass).

Sampling Procedure:

The general locations of spawning activity will be derived from
visible milt observed in the water column during scheduled aerial
surveys. This information will be compiled and summarized on maps
showing spawning locations and the number of days on which milt was
observed.

124

Using this information, skiff surveys will be conducted in season,
by members of the spawn deposition dive team, to verify the
accuracy of spawning area maps derived from aerial survey data.
Diving where herring have spawned is not recommended for at least
5 days after spawning activity has ceased because of water
visibility problems caused by milt and because large numbers of sea
lions are usually present.

The shoreline area containing herring spawn on the map verified by
skiff survey will be divided into the smallest segments resolvable
on the scale of the map (0. 1 mile or less). A total of 160 of the
shoreline segments will be selected at random from all of the
spawn-containing shoreline segments. Each transect will be
assigned a number and its location drawn on waterproof field maps
that can be taken out in the dive skif f . The dive team leader will
determine the exact transect location within the randomly selected
shoreline segment by identifying a shoreline feature (tree, rock,
cliff, etc.) located above the high tide line as the dive skiff
approaches the shore, but before bottom profiles, bottom
vegetation, or herring spawn are visible from the skiff.
A 0.1 m 2 quadrant constructed of PVC pipe will be used for the
sampling frame. A depth gauge and compass will be fastened to the
quadrant. Data will be recorded on pre-printed single matte mylar
forms attached to PVC clipboards, using a large weighted
carpenter's pencil attached to the clipboard. Normally the dive
team leader will make egg density estimates and record data while
the assistant diver sets and follows the compass course, measures
distancest and carries and places the quadrant.

Sampling along the transects will occur in the following manner:

1. A compass course perpendicular to the shoreline at the
transect location will be set on the compass attached to
the sampling quadrant.

2. The first quadrant will be placed within the first 5
meters of spawn by tossing the quadrant.

3. The lead diver will estimate and record the number of
eggs in the quadrant. The number of eggs is normally
recorded in units of thousands. The vegetation type,
percent cover, substrate, and depth are also recorded.

4. The assistant diver will measure four complete 1 m hand-
spans offshore, along the compass course. Halfway
through the fifth hand-span, the assistant diver will
gently toss the quadrant ahead approximately one-half
meter and allow it to come to rest. The lead diver then
makes another estimate at the new quadrant location.

5. This process continues every 5 meters until the apparent

125

end of the spawn is found. Divers will verify the end of
the spawn by swimming at least an additional 20 m past
the end of the spawn, unless a steep drop-off is
encountered.

Data codes have been developed for the vegetation types and species
that are encountered in PWS. If more than one is present in the
quadrant sampled, the three most common are recorded on the data
forms. Percent cover is a simple estimate of the percentage of
plant cover that exists within the quadrant sampled (e.g., if half
the area is covered, the cover is 50%).

Approximately every fifth quadrant will be used as a special diver
calibration sample. Both divers will estimate the number of eggs
in the quadrant in a manner such that neither can see the other's
estimate. Divers will attempt to remove all egg-containing
vegetation and scrape eggs off rock substrate, placing the material
in numbered mesh bags. A sample size goal of 80 calibration
samples per diver was established, including 20 in each of four
vegetation categories (eelgrass, fucus, large brown kelp, hair
kelp), based on 1988 and 1989 survey results. Calibration samples
should also be spread over a wide range of egg densities. The
spawn deposition project leader will track the number of samples
collected by each diver by vegetation group and density to ensure
that sufficient calibration samples are taken in each category.
Upon completing a dive shift, calibration sample material will be
removed from the numbered mesh bags and placed in nalgene ziploc
bags. Gilson's solution will be poured over the sample so that all
material is completely immersed. A label will be made for each
sample (preferably in pencil on mylar) containing the transect
number, both diver's estimates, date, and vegetation type. Five or
6 calibration sample bags can be stored in a 5 gallon plastic
bucket. Samples should not be stacked over one another to prevent
spilling and mixing. Procedures for the enumeration of the number
of eggs in each calibration sample are described, including the
formulas used to prepare Gilson's solution and the other chemicals
used for sample processing.

Data Analysis:

Biomass Estimation

The 1990 spawn deposition survey was patterned after the 1988 and
1989 spawn deposition survey in Prince William Sound (Biggs and Funk
1988, Biggs In Press). The overall biomass estimator is:

(T BI)
B (1)
R)

where:

126

B = estimated spawning biomass in tonnes,
T = estimated total number of eggs (billions) deposited in an
area,
BI = estimated tonnes of spawning biomass required to produce one
billion eggs, and
R estimated proportion of eggs disappearing from the study
area from the time of spawning to the time of the survey
(egg loss).

The estimates for T and BI are derived from separate sampling
programs and are thus independent. Ignoring the unknown
variability in R, the estimated variance for the product of
the independent random variables T and B1, conditioned on R
is:

[T'Var(BI) + B"Var(T) - Var(T)-Var(B')]
Var(BIR) = (1-R)2 ; where (2)

Var(BI) = an unbiased estimate of the variance of BI; and
Var(T) = an unbiased estimate of the variance of T
(Goodman 1960).

Total Number of Eggs (T)

The total number of eggs deposited in an area is estimated
from a two-stage sampling program with random sampling at the
primary stage, followed by systematic sampling at the
secondary stage, using a sampling design similar to that
described by Schwiegert et al. (1985). In computing variances
based on the systematic second stage samples it is assumed
that eggs are randomly distributed in spawning beds with
respect to the 0.1 m 2 sampling unit. While this assumption
was not examined, in practice the variance component
contributed by the second sampling stage was much smaller than
that contributed by the f irst stage, so that violations of
this assumption would have little effect on the overall
variance. The total number of eggs (T), in billions, in an
area is estimated as:

T=N- 91- 10-6, where (3)

N = LIVO.1 = the total number of possible transects;
L = the shoreline length of the spawn-containing
stratum in meters;
VO.1 0.3162 m = width of transect strip;
average estimated total number of eggs (thousands)
-6 per transect; and
10 conversion from thousands to billions of eggs.

The average total number of eggs per transect strip (in
thousands) is estimated as the mean of the total eggs (in

127

thousands) for each transect strip using:

....ere (4)
n

~q9~1i = M~qi ~- ~2qY~qi; and
y~q,= average quadrant egg count in transect i (in thousands
of eggs);

i transect number;

~4qM~qi ~4qw~qi~ql~4qVO.1 = number of possible quadrants in transect i;
~0qw~qi transect length in meters; and

~'h number of transects actually sampled.

The average quadrant egg count within a transect, yi~,~-is
computed as:

~M~_
.=I y~qij F where (5)
~0qm~qi

quadrant number within transect i,
~0qm~i number of quadrants actually sampled in transect i, and
y~qi~qj adjusted diver-estimated egg count (in thousands of eggs)
from the diver calibration model for quadrant j in
transect i.

The variance of T is similar to that given by Cochran
(1963) for three stage sampling with primary units of
equal size, although in this case the expression is
modified because the primary units (transects) do not
contain equal numbers of secondary units (quadrants), and
the variance term for the third stage comes from the
general linear model used in the diver calibration
samples:
2~q(~ql~qo-6~q)2~q[ (1~-f ~q1) f 1 (1~-f 2) f~q1f2
Var(T) = N _ - S~q12 + ~_~_~qF~q_ ~- S2 2 + ~q_~6qF~_~q_ ~* ~S3~q1~q], (6)
n ~4qZ m~qi ~0qm~qi
~qi~q=~q1

n ~4q9~q,~0q)2
~24qZ (~8q9~q, i ~8q-
~qs~0qi2 variance among transects,
n-1

128

2 n 2 (Yij Y-,)2
s 71M variance among quadrants,
2 j=1 i j=1

2 n
s z Var(y,,) sum of the variances of the
3 individual predicted quadrant egg counts
from the diver calibration model,

n
f, - = proportion of possible transects sampled, and
N

m
f2 - = proportion of quadrants sampled within transects
Mi (same for all transects).

Diver Calibration

Diver observations of vegetation species will be
aggregated into four vegetation categories based on
structural and phylogenetic similarities of plants in the
quadrant: eelgrass, fucus, hair kelp, and large brown
kelp. Diver estimates of egg numbers are approximately
proportional to laboratory-enumerated counts! but
systematic biases in the diver estimates can be accounted
for by vegetation type and density (Biggs and Funk 1988,
Biggs In Press). Individual diver effects were not
significant in the 1988 and 1989 survey, but potential
differences among individual divers will be examined. The
basic form of models used to account for biases in diver
observations is:

a Dj Vk Bik
Yijk e * e e Xijk -e where (7)

a a constant;
Dj parameters representing the effect of j1h diver;
V k parameters representing the effect of the k"
vegetation type;
8 jk parameters controlling the functional form of the
relationship between the diver estimate and laboratory-
enumerated egg count for diver j in vegetation type k;
Yijk = the ith laboratory egg count in the vegetation-
diver stratum jk;
X ijk =the ith diver estimate in vegetation-diver stratum jk;
and
e a normally distributed random variable with mean 0 and
variance a2.

129

A multiplicative-effect model is chosen because relative
estimation errors are expected to change with egg density.
The distribution of laboratory-enumerated egg counts for
a given diver estimate was positively skewed in the 1988
and 1989 surveys (Biggs and Funk 1988, Biggs In Press), so
that the logarithmic transformation used to estimate the
parameters of the multiplicative-effect model also
stabilized the variance and corrected the skewness of the
egg density estimates. After a logarithmic transformation
model 7 becomes:

loge (Yijk) = a + Di + Vk + Bik. loge (Xijk) + 6 (8)
Bik = the slope of the relationship between the logarithm
of the diver estimate and the logarithm of the
laboratory-enumerated egg count.

In logarithmic form, the model comprises a linear analysis
of covariance problem with two factor effects (vegetation
and diver) and 1 covariate (diver-estimated egg number) .
The SAS Institute Inc. (1987) procedure for general linear
models will be used to obtain least squares estimates of
parameters and evaluate variance components. In addition
to the two factor effects and one covariate, terms for
diver-vegetation group interactions, density-vegetation
group interactions and density-diver interactions will be
considered in the analysis of covariance. Three-way and
higher level interaction effects will not be considered
because the objective is to derive a simple model with a
relatively small number of parameters. Backward stepwise
procedures will be used to determine subsets of the six
effects that explain the maximum amount of variability in
the data with the smallest number of parameters. During
the backward stepwise procedures, effects will be included
or eliminated from the model based on the probability
level of F ratios for partial sums of squares.

Translation of the predicted values from the logarithmic
model, equation (8), back to the original scale, equation
(7), requires a correction for bias. The bias in the
expected value of Yijk is eXp(12a2) when the true variance
of Y k a2, is known. Laurent (1963) gives an exact
expression for the bias correction that incorporates
additional terms when a2 is estimated from a sample. For
the diver calibration data, the biases in estimating a2
from a sample were less than 0.05% (Biggs and Funk 1988),
so expected values for Yijk are estimated from:

a Di Vk Bik 12S2
E (Yijk) e * e e Xijk - e where (9)

130

s the mean squared error from the general linear model.
The variance of individual predicted Y is estimated
from:

Var(Y ) = [e (2Y + a2) [e - 1 (10)

Although the above expression is appropriate when a is
known (Laurent 1963), s will assumed to be an unbiased
estimate of a for the 1990 study since only a small bias
was introduced into estimates of the mean when s was used
to estimate a in past years (Biggs and Funk 1988).

Spawning Biomass per Billion Eggs (B')

Catch sampling programs will be used to estimate the
relationship between spawning biomass and egg deposition.
The tonnes of spawning biomass required to produce one
billion eggs (B') will be estimated as:

W S
B' 10 where (11)
F(Wf)
W = estimated average weight in grams of all herring (male
and female) in the spawning population in an area;
S = estimated ratio of total spawning biomass (male and
female) to female spawning biomass;
F(W) = estimated fecundity at the average weight of females in
the spawning population in an area, in numbers of eggs;
and
10-6 conversion from grams to tonnes
103 units conversion factor = - =
10-9 conversion from eggs to billions

Estimates of average weight, sex ratios, and fecundies are not
independent. The variance of B' is approximately:

Var(B') = (10 ){ [S/F(W)] Var(W)
+ [W/F(W)] Var(S)
+ [WS/F(W) ] Var(F(W ))
+ 2Cov(W,S) [S/F(W )] [W/F(W )]

131

2COV[@R,F(@q-f)] [SIF(Rf)] [@@SIF(Rf)']

2Cov[S,FFWf)]-[W/FFWf)] -CWS/F(Wf)'] (12)

The covariance terms containing S, Cov(W,S) and Cov[S,F(Wf)lf
will not be included in the estimate for 1990. These terms
were not included in the estimate of Var(B') in 1988 and 1989
because S was estimated from either the same pooled AWL
samples or from a single AWL sample. However, Cov(9,S) and
Cov[S,F(wf)] probably contribute a small amount to Var(B')
since the term involving Cov['W F('Wf)) was very small in 1988
and 1989.

Correction for Egg Loss

The only component needed for the biomass estimate that has
not been estimated within the present study is egg loss (the
proportion of eggs disappearing from spawning areas between
the time of spawning and the time of surveys). Before the
extensive use of SCUBA diving to survey herring egg
deposition, estimates of egg loss were relatively high.
Montgomery (1958) estimated that egg loss was 25 to 40% for
Southeast Alaska, and Blankenbeckler and Larson (1987) used
similar estimates in their early egg deposition surveys in
Southeast Alaska. However, Haegele et al. (1981), counducting
diving surveys in British Columbia, argued that egg loss was
only about 10%. They based this assumption on the fact that
most spawn was deposited in the subtidal zone where egg loss,
primarily due to predation and wave loss, was probably less
than had been observed in the intertidal zone. Presently, egg
loss is assumed to be 10% in British Columbia, Southeast
Alaska and PWS since the timing of diving surveys in relation
to spawning has been standardized among these areas (W.
Blankbeckler, ADF&G, Ketchican, personal communication; Biggs
and Funk 1988). To test this assumption, an initial study of
actual egg loss within PWS will be conducted in conjunction
with the egg survival study during 1990.

Herring Age, Weight, Length, Growth and Fecundity Estimation

Mean Weight and Sex Ratio

Mean weight and sex ratio will be estimated from AWL samples
collected from the commercial catch and ADF&G test fishing
conducted before or after commercial openings. AWL samples
will be collected from the spawning population in each of the
spawn deposition summary areas (Southeast, Valdez Arm, North
Shore, Naked Island, and Montague Island). The approximate
timing of peak herring spawning in each summary area will be

132

determined from aerial survey sightings of milt and herring
schools. All herring AWL samples taken during the time of
peak spawning in each area will be pooled to obtain estimates
of mean weight and sex ratio for each summary area. Average
weights and sex ratios for all of PWS will be estimated as the
average of the estimates from each of the areas weighing by
the spawn deposition biomass estimate in each area.

The estimated sex ratio, S, is expressed as the ratio of the
number of herring of both sexes in the AWL samples to the
number of females. The binomial distribution will be used to
estimate the proportion of females, p, in samples, where S
1/p. The variance of S.is then given by:
S'(S-1)
Var(S) = 1 (13)
n

where n is the number of herring in the AWL sample.

EGG LOSS

Commercial and test fishing catches will be sampled for AWL,
fecundity, and roe maturity information. These data are used
to estimate spawning biomass and spawn deposition, forecast
herring returns, and evaluate effects of the oil spill on
survival. Information on fecundity, average weight of females,
and sex ratio are also important components of the spawn
deposition biomass estimator. AWL sampling will be intensified
in 1990 to increase the precision of biomass estimates and,
therefore, enhance the possibility of detecting oil spill
impacts upon herring stocks.

Sampling will begin as soon as concentrations of herring
appear in near shore areas that can be sampled with purse
seine gear. Efforts will be made to sample major
concentrations of herring throughout PWS at periodic intervals
throughout the spawning period. The major objective of this
portion of the study will be to determine the age, sex, and
size composition of all major herring concentrations in the
general areas of Valdez Arm and the Eastern District, the
North Shore, Naked Island, and Montague Island. Results of
the aerial survey program will be used to direct test f ishing
efforts within each area.

Each week during the sampling period, early April through
early May, six to eight samples of herring will be collected
through test fishing or from the commercial catch. A sample
of 403 herring is needed to simultaneously estimate the
proportion of at age of a multinomial population such that 95%
of the time the estimated proportions will be within �10% of
the true proportions (Thompson 1987). Therefore, efforts will

133

be made to obtain samples consisting of approximately 450
herring to allow for the occurrence of unreadable scales
(usually less than 5% of the sample). Herring samples will be
flown from the fishing grounds each day to Cordova for
processing. Augmentation of the standard AWL sampling program
will be needed to collect sufficient samples for hydrocarbon
analyses, fecundity estimates, and oocyte loss measurements.
All AWL data will be collected using personnel and funding
from the standard (i.e. non-oil spill related) AWL sampling
program conducted by ADF&G within PWS.

The following data will be collected for each herring sampled:

1. sex (determined by examination of gonads);
2. standard length (in mm);
3. weight (in grams);
4. age (determined by examination of scales);
5. capture information (date of capture, fishing district,
subdistrict,local name for the location, fishing vessel
name, gear type);
6. herring number on data form; and
7. data form number.

Fecundity

Additionally, a subsample of herring will be collected to
estimate f ecundity. The average fecundity at the average
female weight (F(Wf)) from expression (11) is a component of
the spawn deposition survey biomass estimator. The spawn
deposition survey attempts to estimate spawning biomass so
that the 95% confidence interval is within � 25% of the actual
biomass estimate. If fecundity sampling is to contribute no
more than 1% to the confidence interval width, a sample of 85
females of exactly the average weight of females in the
spawning population is needed. Since average female weight is
unknown at the time of sampling, more herring must be sampled
over a range of sizes. Based on the precision of 1989
fecundity sampling, a sample size of 130 herring would be
needed to provide the desired level of precision. An
additional 100 samples clustered around the average size of
females in 1989 will be taken to compare with the past year's
data. The average weight of a female in the fecundity sample
in 1989 was 119 grams. The predicted average weight for the
population in 1990 is 142 grams that translates to an average
predicted length of 215 to 225 mm. Therefore, sampling should
be clustered about the 210 mm to 230 mm length classes is
desirable.

Effects of the oil spill on fecundity will also be examined by
testing for differences in fecundity among five areas: (1)
Southeast Shore including Simpson and Sheep Bays, Port
Gravina, and Port Fidalgo; (2) Northeast Shore including

134

Valdez Arm and Tatitlek Narrows; (3) North Shore; (4) Naked
Island; and (5) Montague Island. While extensive mortality of
adult herring from the oil spill has not been documented, it
is possible that sublethal stresses could result in reduced
fecundity.

Herring fecundity samples will be collected concurrently with
AWL samples. To accomplish this, at least five individual
test purse samples will be subsampled. Females within these
purse seine samples will be randomly selected within 10 mm
length classes until stratum goals are reached. The roe sacs
from each selected females herring will be removed and placed
in a ziploc bag labeled with the AWL number corresponding to
that female. Each individually packaged roe sample will then
be placed in a larger plastic bag labeled with the sample date
and location. Standard laboratory procedures have been
developed to process fecundity samples.

Samples for hydrocarbon analyses will also be obtained from
herring collected at each of the four locations (Naked Island,
Galena Bay, Cedar Bay, and Stockdale Harbor):

1. three gut samples for hydrocarbons;
2. three viscera samples for hydrocarbons;
3. three muscle samples for hydrocarbons; and
4. three gonad samples for hydrocarbons.

General observations on the prevalence of nematodes, liver and
gall bladder condition, and fullness of gut will also be made
for each herring collected for hydrocarbon analyses. Standard
protocol, including sample sizes and collection strata, for
collecting herring eggs for hydrocarbon analyses will be
followed.

In addition to the 500 ovaries collected for fecundity
analysis, 50 ovaries will be collected and preserved in a
buffered formalin solution for oocyte loss measurements. An
additional 25 preserved ovaries will be obtained from Sitka
Sound, Southeastern Alaska, for use as a control. Atretic
eggs and histopathological damage in the sac roe of the adult
herring will be recorded during oocyte loss observations.

A linear relationship was found between fecundity and weight
for herring samples collected in 1988 and 1989 (Biggs and Funk
1988). In 1990, the fecundity-weight relationship will again
be examined using data pooled across all areas. Average
fecundity for each area will be estimated from the fecundity-
weight relationship using the average female weight from each
area. The average fecundity for each area will then be
applied to the spawn deposition biomass estimator (F(Wf) in
expression (11). The variance of estimated average
fecundities will be approximated using the variance of

135

predicted means from the fecundity-weight linear regression
(Draper and Smith 1981):
I (Wf WF7 I
Var[F(i@f) S2 1-n + q + z (Wi - WF) where (14)
2
s = residual mean square from the fecundity-weight linear
regression;
Wf = average weight of female fish in the spawning
population;
average weight of females in the fecundity sample;
Wi weight of individual females in the fecundity sample;
n total number of females in the fecundity sample; and
q total number of females in the AWL sample.

General Linear Model (GLM) extensions of linear ANOVA
techniques will be used to test for year and area effects in
growth and fecundity.

Herring Egg Survival and Egg Loss Estimation

oil contamination of herring spawning sites and exposure of
spawning herring to oil may cause mortality to herring eggs,
decrease hatching success, reduce larval viability, and impair
larval growth. The major objective of this portion of the
study will be to measure immediate, easily observable
mortality of herring eggs in a subsample of the sites used in
1989. In 1990, nine sites will be used to conduct the egg
loss study, collect hydrocarbon samples, collect live eggs for
the laboratory portion of the study, and to gather samples for
sublethal impact testing.

Three study transects will be re-established in each of three
areas used during 1989 (assuming those areas receive spawn in
1990): Naked Island, Fairmont Bay, and Rocky Bay on North
Montague Island. The ratio of live to dead eggs will be
determined along each transect from subsamples of 100 eggs.
Dead eggs turn an opaque white color and are easily identified
with low power magnification under a binocular microscope.
Mussel tissue samples will also be collected for hydrocarbon
analysis.

A 1990 laboratory egg incubation experiment, similar to the
one conducted in 1989, will be carried out by a private
consultant contracted by ADF&G. This experiment will
determine the survival of herring eggs and larvae collected
from the nine study sites in PWS and three control sites in
Sitka Sound, Southeast Alaska.

Divers will establish the location of mean low water (MLLW) at
the start of each dive. Each dive team will attempt to sample

136

three transects each day. Each transect will be sampled every
two days until most herring eggs have hatched (about 20 May) .
A total of twelve to sixteen dives will be made along each
transect over the course of egg development.

The location of each transect will be marked. Divers will
work along transects by following a compass course set
perpendicular to shore. During the first dive, five sample
stations at the +1, 0, -5, -15, and -30 foot depths will be
marked underwater with weighted floats anchored by a spike.
Station depths, corrected for tide stage, will be determined
using diver's depth gauges. Three samples of vegetation
containing at least 100 eggs will be collected at each depth
along the transect whenever possible.

The following data will be recorded the first time each
transect is sampled:

1. transect number;
2. site description (location, exposure, plant community);
3. number of depth strata from which herring eggs were
obtained; and,
4. original treatment category (high, medium, low, or no
oil-impact).

The following data will be recorded every time each transect
is sampled:

1. transect number and location
2. date;
3. dive time;
4. treatment level;
5. air and water temperature;
6. maximum depth; and,
7. number of live, dead, and other eggs per sample.

Herring eggs and mussels will be collected at each site for
hydrocarbon analysis on the first day. Three samples each of
eggs and mussels (six per transect) will be collected from
each sampling location, including the three control sites in
Sitka Sound, at the lowest tide stage at which mussels occur
(usually about 5 ft below MLLW) . Collection methods will
follow established protocol, including chain of custody forms.

During one of the sampling trips to each transect, herring
eggs and associated vegetation will be collected for the
laboratory incubation project. Herring eggs will be collected
at nine sites within Prince William Sound and three sites
within Sitka Sound. At each site, three samples of vegetation
containing at least 300 eggs will be collected at three depths
(MLLW, -5 ft, and -15).

137

Herring eggs will also be collected and preserved in a
phosphate buffered formalin solution, using sea water, for
biochemical analysis. Results of these analyses may help
determine the extent of oil exposure from determination of
sublethal effects.

Finally, herring egg samples will be collected from each of
the 12 study sites for cytogenetic analysis. Ten egg patches
consisting of approximately 1000 eggs each (5 ml) will be
preserved in a buffered formalin solution from each study site
(i.e. a total of 120 samples). A subsample of eggs will be
taken from each sample jar and analyzed for mitotic
aberrations in the embryonic and yolk cells. Detailed
methodology will be provided by the lab contracted to perform
the service.

Egg survival data will be summarized by level of hydrocarbon
impact, transect, depth, date of sample collection, and
proportion of live eggs. Several different analyses will be
conducted to test for differences in egg survival due to the
level or amount of oil. The first analysis will be a nested
mixed factor ANOVA incorporating all possible factors and
interaction effects like:

Yijkl = u + Ai + Bj (Ai) + Ck + Dt + ACtk + ADil + CDk1 + ACDikt +6ijkL' (15)

where,
Yijkt = the arc sin transformed proportion of live eggs;
U = grand mean;
Ai = oil impact level (treatment; fixed effect);
Bi = transect (random effect; nested within treatment);
Ck = depth (fixed effect);
Dt = time interval (days) between spawning and sample
collection (random effect);
AC1k + ADR + CD kt +ACD ikt = interaction terms; and
fijkt = error terms, which, after arsine transformation are
assumed to be normally distributed with mean 0 and
variance o.2

The second analysis will be an analysis of covariance (ANCOVA)
where both treatment (Ai) and time (D,) will be treated as
covariates. Treatment and depth will be treated as f ixed
effects, while transect (nested within treatments) and time
will be treated as random effects. This model will describe
the decrease in the proportion of live eggs over time, using
time as a covariate, and will reduce the number of parameters
that must estimated for the model.

Egg loss is the only component of the spawn deposition biomass
estimator that has not been measured. In the past, a 10% egg
loss factor was applied to all transect data to adjust the

138

total spawned biomass estimate. In 1990 a preliminary egg
loss study will be conducted in conjunction with the egg
survival study to determine whether the 10% egg loss factor is
appropriate for use at PWS study locations.

The same three transects used in each of three areas for the
egg survival study will be used in the egg loss study: Naked
Island, Fairmont Bay, and Rocky Bay on North Montague Island.
Egg loss will be estimated by observing changes in egg density
over time at these locations.

To avoid sampler bias in selecting samples, as was done for
the egg survival study, a marked leadline, 20 m or less in
length, will be used to select samples. The leadline will be
placed parallel to shore and to the left of each transect
station. Egg density estimates will be taken within 0.1 M2
sample quadrants using the same procedures described for spawn
deposition diver transects. For each transect, five egg
density estimates will be made at each of five depths (+l, 0,
-5,-15,-30) ft depths). Divers making egg density estimates
for the egg loss study will be calibrated in a similar manner
used for divers assisting in spawn deposition surveys. one
egg count calibration sample will be collected at each
transect and at each depth level. For the calibration sample,
all herring eggs and vegetation will be removed from a 0.1 M2
sample quadrant. Counts of eggs within the calibration sample
will be made in the laboratory at a later time. Egg density
estimates and egg counts will be conducted every other day
f rom the time of spawning in each area until the time of
hatching (a period of approximately 20-25 days). It should be
possible to obtain egg density estimates and egg counts for
about eight days during the study. This would result in a
total of approximately 1,800 egg density estimates (three
areas; 3 transects per area; five depths per transect; five
egg density estimates per depth; eight days) and 540 egg
counts (three areas; three transects per area; f ive depths per
transect; one egg count per depth; eight days) f or the season.

Egg loss data will be summarized by area, transect, depth,
date of sample collection, and estimated egg density. Egg
density estimates will be adjusted for observer (diver)
biases, following procedures set forth for diver calibration
in the spawn deposition survey, prior to analyses. The change
in egg density over time for each transect and depth will be
examined.

BIBLIOGRAPHY

Biggs, E.D., and F. Funk. 1988. Pacific herring spawning ground
surveys for Prince William Soundl 1988, with historic overview.
Regional Information Report 2C88-07, Alaska Department of Fish and
Game, Anchorage, 73 p.

139

Blankenbeckler, W.D. and R. Larson. 1982. Pacific herring (Clupea
harengus pallasi) spawning ground research in Southeastern Alaska,
1978, 1979, and 1980. Alaska Department of Fish and Game Technical
Report No. 69. 51 P.

Blankenbeckler, W.D. and R. Larson. 1987. Pacific herring (Clupea
harengus vallasi) harvest statistics, hydroacoustical surveys, aget
weight, and length analysis, and spawning ground surveys for
Southeastern Alaska, 1980-1983. Alaska Department of Fish and Game
Technical Data Report No. 202. 121 p.

Cochran, W.G. 1963. Sampling techniques. John Wiley and sons,
New York.

Draper, N.R. and H. Smith. 1981. Applied regression analysis.
John Wiley and Sons, New York.

Goodman, L.A. 1960. On the exact variance of products. Journal
of the American Statistical Association 55:708-713.

Haegele, C.W., R.D. Humphreys, and A.S. Hourston. 1981.
Distribution of eggs by depth and vegetation type in Pacific
herring (Clupea harengus pallasi) spawnings in Southern British
Columbia. Canadian Journal of Fisheries Aquatic Sciences 38:381-
386.

Hourston, A.S., H. Rosenthal, and H. von Westernhagen. 1984. Viable
hatch from eggs of Pacific herring (Clupea harengus pallasi)
deposited at different intensities on a variety of substrates.
Canadian Technical Report of Fisheries and Aquatic Sciences 1274.

Jackson, M. and R.C. Randall. 1983. Herring spawn deposition
surveys in Prince William Sound, 1983. Alaska Department of Fish
and Game, Prince William Sound Data Report No. 83-6. 15 p.

Jackson, M. and R.C. Randall. 1984. Herring spawn deposition surveys,
Prince William Sound, 1984. Alaska Department of Fish and Game,
Prince William Sound Data Report 84-16. 15 P.

Laurent, A.G. 1963. Lognormal distribution and the translation
method: description and estimation problems. Journal of the
American Statistical Association 58:231-235.

Montgomery, D.T. 1958. Herring spawning surveys if Southeastern
Alaska. United States Fish and Wildlife Service, Bureau of
Commercial Fisheries, Marine Fisheries Investigations Field Opera-
tions Report. 22 p.

Palsson, W.A. 1984. Egg mortality upon natural and artificial
substrata within Washington State spawning grounds of Pacific
herring (Clupea harencfus nallasi) , M.S. Thesis, University of

140

Washington, Seattle.

SAS Institute Inc. 1987. SAS/STAT Guide for personal computers,
version 6 edition. SAS Institute, Cary, North Carolina.

Schweigert, J.F., C.W. Haegele, and M. Stocker. 1985. Optimizing
sampling design for herring spawn surveys on the Strait of Georgia,
B.C. Canadian Journal of Fisheries and Aquatic Sciences 42:1806-
1814.

Thompson, S.K. 1987. Sample size for estimating multinomial
proportions. The American Statistician 41:42-46.

BUDGET: ADF&G

salaries $ 121.7
Travel 6.9
Contracts 398.0
supplies 16.8
Equipment 15.0

Total $ 558.4

141

FISH/SHELLFISH STUDY NUMBER 13

Study Title: Effects of Hydrocarbons on Bivalves

Lead Agency: ADF&G

INTRODUCTION

Bivalve mollusks are an important component of the f ood chain,
existing as prey f or bear and sea otters, and bivalves support
subsistence and sport fisheries in PWS. Because they are relatively
sedentary and occupy nearshore areas, bivalves may be particularly
susceptible to contamination by oil. In contrast to finfish species
which metabolize hydrocarbons at a much higher rate, bivalves
metabolize hydrocarbons at a reduced rate and are therefore much more
likely to bioaccumulate hydrocarbons. It is hypothesized that
increased hydrocarbons in nearshore sediments could affect bivalves
for a long period of time by increasing mortality, decreasing growth,
or causing sublethal injuries. The effects of oil on the growth and
survival of littleneck clam (Protothaca staminea) in particular and
other bivalves in general have been well documented (Anderson et al.
1982, Anderson et al. 1983, Augenfeld et al. 1980, Dow 1975, Dow
1978, Keck et al. 1978).

This study seeks to continue to evaluate the potential effects caused
by the oil spill by comparing data obtained from several beaches
representing different levels of oil contamination. The effects of
the mechanical cleaning of beaches following the spill will also be
evaluated. Documenting effects on littleneck clams, butter clams
(Saxidomus giganteus), and razor clams (Siliqua patula) is required
to determine the scope of impact by the oil spill on these species,
associated ecosystem elements, and current and future employment,
recreation, and lifestyles of coastal communities on PWS.

OBJECTIVES

A. Test if the level of hydrocarbons in bivalves and in sediments
is not related to the level of oil contamination of a beach.

B. Document the presence and type of damage to tissues and vital
organs of bivalves sampled from beaches such that differences of
�5% can be determined between impact levels 95% of the time.

C. Test if the growth rate of littleneck, butter and razor clams is
the same at beaches of no oil impact, intermediate or high
levels of oil impact and intermediate or high levels of oil
impact in areas which had been treated.

D. Test if the proportion of dead clams is not related to the level

142

of oil contamination or treatment at a beach.

E. Document numbers of young-of -the-year clams and test if the
proportion of young-of -the-year clams is not related to the
level of oil contamination or treatment at a beach.

METHODS

This study will be conducted by the ADF&G and represents a
consolidation of the former FIS Studies 13 and 21. During April
through June, 1990, beaches will be selected which, if possible,
coincide with those sampled during 1989. An emphasis will be placed
on beaches that are known to be important habitat for bear or sea
otters. It is possible that the baseline beaches sampled in 1989
will not be resampled because they do not constitute bear and/or
otter habitat. Sites will be chosen inside PWS which have not been
impacted by oil, which have received moderate to heavy oiling, and
which have been cleaned of oil by mechanical means. Sites outside of
PWS will be chosen which have not been impacted by oil and which have
received heavy to moderate oiling. Because razor clams are not found
in the same habitat as butter or littleneck clams, razor clam beaches
along the Kenai and Alaska peninsulas will be located which were
affected and unaffected by oil.

Beaches of known bear and/or sea otter habitat and known to contain
clams will be classified by oil contamination levels. Nine study
sites for littleneck or butter clams in PWS representing three levels
of oil contamination (subjectively rated as no contamination,
intermediate or high contamination and intermediate or high
contamination which had been treated by mechanical means) will be
sampled. Beaches with no oil contamination are Hell's Hole, Double
Bay, and Simpson Bay. Beaches with moderate or heavy oil
contamination which have been treated or untreated include Gibbon
Anchorage, Snug Harbor, Wilson Bay, North Chenega Island, Horseshoe
Bay and Green Island. Sites not sampled in 1989 are selected
contingent upon habitat suitability.

Since cleaning efforts were not as intense outside of PWS, and since
a smaller number of sites have been chosen theref the additional
sample level based on mechanical treatment will not be investigated
outside of PWS. Eight study sites for littleneck or butter clams
representing two levels of oil contamination (subjectively rated as
no contamination or moderate to high contamination) were chosen to be
sampled in the Lower Cook Inlet and Kodiak areas. Beaches with no
oil contamination are Jakalof Bay, Kachemak Bay and Seldovia Bay in
lower Cook Inlet and Port Bailey on Kodiak Island. Beaches with
moderate to heavy oil contamination are Windy Bay, Tonsina Bay and
Port Dick in Lower Cook Inlet and Kupreanof Strait on Kodiak Island.

Six sites have been chosen for sampling razor clam habitat
representing two levels of oil contamination (subjectively rated as
no or high contamination) . Beaches with no oil contamination are

143

Halibut Bay, Polly Creek and Augustine. Beaches with moderate or
heavy oil contamination are Swishak, Alinchak and South Nuka Island.

For each sample site, the following site description information will
be recorded: site orientation (N-NW etc.), latitude, longitude,
beach slope, low tide height, percent dominant substrate composition,
temperature and salinity of the water, weather and wave action.
Temperature and salinity of the water will be measured at a distance
of approximately 5 meters offshore from the sampled beach at the
daily low slack tide.

Beaches will be sampled for littleneck and butter clams at maximum
low tides for a monthly tidal cycle. For beaches which had been
sampled in 1989, 1990 tidal heights and time of year will be matched
with the 1989 values as closely as possible. At each beach, three
sampling transects will be run to insure complete coverage of the
beaches as distribution of oil on the beaches is unknown. Transects
will be perpendicular to the water's edge and parallel to each other
with a total distance between each transect of 15 meters. Transects
are perpendicular to the water to insure complete sampling of clam
habitat. The top of each transect is placed at the +1.6 meter tide
level and the bottom of the transect at the lowest tide level.

Prior to sampling, the upper distribution of clams will be determined
by removing sediment to a depth of 30 cm (12 in) along a trench
adjacent to the proposed transect. The trench is dug starting from
the top of the transect and continuing until clams are encountered.

A total of eight quadrants will be sampled from each transect to
obtain2hydrocarbon and necropsy specimens. Sample quadrants are each
0.25 m (0.5 m by 0.5 m). Additional sampling or complete sampling
of each transect (all possible sampling quadrants) may be necessary
if insufficient numbers of clams are recovered within the eight
sampling quadrants to meet project objectives. Quadrants will be
sampled from the top to the bottom of each transect as the tide
recedes. The distribution of clams will extend below the low tide
levels encountered during each sampling event. However, the bottom
of each transect and the bottom sampling quadrant will occur at the
daily low tide level. The upper layer of sediment will be removed
and washed through a 1 mm mesh screen to retain small young-of-the-
year clams. The remainder of the sediment is washed through a larger
3 mm mesh screen.

Razor clam habitat tends to be comprised of long and broad sandy
beaches. Because of the size of the area inhabited by razor clams,
and due to the time and manpower required for a full scale study of
this species, there will be no attempt made to estimate the abundance
of razor clams on the beaches being surveyed. The primary objectives
of this portion of the study are to obtain hydrocarbon and necropsy
samples, and to collect a sufficient number of razor clams for age
and growth determination.

144

Razor clams inhabitat the 0.91 m to -1.22 m tidal range (Quinn and
Jones ' 1989). In order to minimize sampling effort, a tidal height
known to contain a large number of clams will be established at
approximately 0 m to -0.33 m on each beach. A transect along this
tidal height will be dug with a high pressure pump (pre-emergent fry
pump) until the desired sample size has been collected.

A total of nine sediment samples will be collected from each beach
site (triplicates from each transect).. All sediment samples will be
collected before bivalve sampling is performed. The triplicate
hydrocarbon samples from each transect will be composite sediment
samples which will be collected by scooping one tablespoon (15 cc) of
sediment to a depth of 2 to 3 cm from each of the eight sample
quadrants on a transect.

All samples from each transect will be placed in 8 oz glass jars
rinsed with methylene chloride. Each-jar will be labelled with the
site name, latitude, longitude, date, "SEDIMENT", transect number,
sample number, names of the sampling team members, "BIVALVE", and
11ADF&G11. Data will be recorded on the appropriate form.

Triplicate composite sediment samples will be taken from the razor
clam beach transect. This will provide 3 samples per beach and 9
samples per treatment level.

The small sub-samples of sediment taken from each sampling quadrant
will provide a representative mixture of sediment composition and
contamination throughout the transect. Three composite sediment
samples for each transect at each site provides 27 composite samples
for each impact level (no, intermediate or high, and intermediate or
high with treatment). The industry standard is 8 samples for each
treatment level. A sample size of nine composite samples is
considered an adequate number of samples to detect a difference in
sediment contamination between impact levels at the desired a and 8
levels. This coverage level is being tripled.

Three common species of clam, littleneck clam Protothaca staminea
butter clam Saxidomus giganteus., and razor clam Siligua patula wili
be sampled for hydrocarbon analysis, necropsy, and age and growth
statistics.

Specimens for hydrocarbon analysis will be taken from each sampling
quadrant before any other specimen sampling is conducted. Bivalves
of each species will be randomly selected for hydrocarbon analysis
from sampling quadrants at each site.

0ne hydrocarbon sample for each species will be obtained from each
transect. For littleneck clams and butter clams, each hydrocarbon
sample will be composed of 14 specimens. The 14 specimens from each
transect (1 hydrocarbon sample) will be selected by randomly picking
two clams with a shell length of 2-5 cm from each of the eight
sampling quadrants and discarding two clams selected at random.

145

Each hydrocarbon sample for razor clams will optimally be composed of
six to eight individuals. Six to eight razor clams with a shell
length of 2-5 cm will be randomly collected at the beginning, middle
and end of the collection transect, for a total of three samples per
site.

Bivalve samples are being limited to a particular size range because
rates of uptake, metabolism, and depuration by clams probably change
with size. If specimens of the desired size are not found in each of
the sampling quadrants, then the desired number of additional
specimens will be collected from the other sample quadrants.

Combined tissue samples from each sampling quadrant will provide a
representative mixture of bivalve tissue composition and
contamination throughout the transect. The desired size of each
composite tissue sample is 15 qm. The number of bivalves to provide
this sample from each transect was estimated based on the average
size of individuals of each species. An estimate of three
hydrocarbon samples from each site is needed for detecting
contamination between impact levels. A sample size of nine composite
samples per three impact levels within the sound will allow the
detection of differences in hydrocarbon content of 1.9 standard
deviations with a and B levels of 0.05 and 0.1, respectively. A
sample size of 12 composite samples per two impact levels outside of
the sound will allow the detection of differences in hydrocarbon
content of 1.4 standard deviations with a and B levels of 0.05 and
0.1, respectively.

Collection of specimens for necropsy will begin only after all
hydrocarbon samples have been taken. Total sample size is 20 live or
moribund specimens of each species taken at random from each beach
site. Noticeable numbers of moribund animals will be documented and
sampled separately. With 20 bivalves sampled from each beach, the
total sample for each treatment (no, intermediate or high oil
contamination, and intermediate or high oil contamination which has
been treated) will be 60 within the sound and 80 for each of two
treatment levels (oiled and not oiled) outside the sound. This
sample size will allow detection of differences in presence of tissue
damage of �5% with 95% confidence between samples obtained from
beaches with different levels of oil impact. This sample size will
allow detection of gross differences between beaches with no, medium
or high oil impact and medium or high oil impact which have been
treated by mechanical means.

one specimen of each species will be randomly selected from each
sampling quadrant. This will yield a total of 24 specimens. Four
specimens from the 24 collected will be randomly selected and
discarded from the sample to achieve a sample size of 20 specimens.
Twenty razor clams will be collected at random along the beach
sampling transect.

For littleneck and butter clams, a total of 100 specimens will be

146

collected from each transect at each site. From.each transect six
sampling quadrants will be selected at random. From each of these,
12 specimens will be randomly sampled from the quadrant containers.
Fourteen specimens will be randomly sampled from the remaining two
quadrant containers.

The sample of 100 specimens per species from each transect will
provide 300 samples from each beach or 900 and 1,200 clams per
species for each level of beach impact inside and outside of PWS,
respectively. Sample size for growth is based on the difference
between mean shell height (width) for age i and age i+l clams,
variance in shell height for age i+l clams, probability of making a
type I error equal to .01 and probability of making a type II error
equal to .05 (Netter and Wasserman 1985). Data for mean shell height
and variance in shell height was taken from Paul and Feder (1973) for
littleneck clams and Nickerson (1977) for butter clams. Sample size
for detecting difference in growth at age of clams between impact
levels was estimated at 261-275 littleneck clams for each impact
level. This sample size was rounded up to 300 clams. The purpose of
3 sites for each impact level is to provide replicates at each impact
level. The sample size required for detecting difference in growth
at age was somewhat smaller for butter clams, however because not all
size ranges were represented in the available data, the larger sample
size of 300 clams was recommended for this species as well.

All shells will be collected from each quadrant and the number of
live clams, the number of dead hinged shells, and the number of half
shells will be recorded. One hundred hinged shells from dead clams
taken in the sampling quadrant located at the median tidal height
(quadrant 4) will be retained for age analysis. If possible, some
will have the microstructure analyzed to determine the year of death.
.If less than 100 hinged shells are found in the three mid-tidal
height sampling quadrants, additional shells will be collected at
random at this approximate tidal height until the sampling objective
of 100 dead shells is obtained.

The sample size for determining the age composition of razor clams is
based on data taken from Clam Gulch (Quinn and Jones, 1989). Quinn
and Jones recommend a sample size for determining age composition of
between 300 and 400 clams per stratum of interest. A minimum of 300
razor clams per beach will. be collected for size and age
determination.

A total of 600 clams will be submitted for microstructure analysis.
A random sample of 200 clams collected from each of three
representative beaches (no contamination, intermediate or high oil
contamination, and intermediate or high oil contamination which has
been treated) located within PWS will be analyzed. In particular,
this analysis will look for the presence of a "check" in the shell
material which has been laid down by clams as a possible response to
the oil spill. Growth which has occurred since the "check" will be
examined. Growth rates will be reported as well as estimated ages.

147

To further quantify oil impacts on clam growth and to discount site
effects, littleneck clams will be transplanted from oiled to non-
oiled areas and from non-oiled to oiled areas. Three oiled beaches
and three non-oiled beaches will be chosen for this purpose.
Criteria for selecting paired oiled/non-oiled beaches, to the extent
possible, will include similarity in profile, drainage and length-
frequency distribution of bivalves.

Two tidal heights will be selected, each of which has an adequate
number of specimens at paired beaches. Clams will be transplanted to
the same tidal height from which they originated. At each tidal
height, three locations will be established creating triplicate
sampling stations at each height. 2 Each location will consist of
three adjacent clearly marked 0.25 m plots. one plot will be marked,
but will not be disturbed until clams are sampled for growth.
Another plot will be dug to a depth of 0.3 m and all of the removed
clams and sediment will be replaced in the plot. Clams from this
plot will have a small notch filed into the ventral edge of the
valves to mark the time of disturbance. All clams will be removed
from the third plot which will be dug to a depth of 0.3 m and the
transplanted clams will be placed in this plot along with the
original sediment. The clams which have been removed will be
collected for comparison with the clams in the undisturbed plots at
the end of the experiment.

Clams to be transplanted will be obtained by digging a trench along
the prescribed tidal height of the donor beach until 150 clams
between 15 mm and 35 mm. in length have been collected. Fifteen
millimeters is considered to be the smallest size which can
effectively be tagged. Clams less than 35 mm are selected to narrow
the range of ages for which differences in growth are being
determined and because the maximum growth rate appears to occur
within this size range. A sample of 50 specimens from each of three
plots will provide 150 samples from each tidal height at each beach
and 450 clams for each tidal height and level of beach impact.
Sample size for growth is based on the difference between mean shell
height for age i and age i+1 clams, variance in shell height for age
i+1 clams, probability of making a type I error equal to . 01 and
probability of making a type II error equal to .05 (Netter and
Wasserman 1985). The sample size was determined after comparing data
for mean shell height and variance in shell height taken from Paul
and Feder (1973) and Nickerson (1977). The sample size for detecting
between impact level differences in growth at age of clams in the
size range of 15 mm to 35 mm was estimated at 133 clams from the Paul
and Feder data and at 85 clams from the Nickerson data for each
impact level. The higher estimate was rounded up to 150 clams by
including the next smaller size group (age 5-6). The purpose of 3
sites for each impact level is to provide replicates at each impact
level.

Transplanted clams will be identified by marking each clam with a
numbered floy tag secured with a quick-drying adhesive. All marked

148

clams will have a small notch f iled into the ventral edge of the
valves to mark the time of transplantation. Individual clams will be
measured at the beginning and end of the experiment. At the end of
the growing season (October 1990), clams will be removed from each of
the plots described above and analyzed f or growth. Wet and dry
weights of clams will also be recorded so that clam condition can be
compared in terms of a weight to height ratio. Hydrocarbon samples
will be taken during the experiment.

To address objective A (hydrocarbons in sediments and bivalve
tissues), an ANOVA will be used to test for differences in
hydrocarbon content in sediment between sites. Differences in
sediment hydrocarbon content will verify that control sites (areas of
no oil impact) are in fact "controls". These differences will also
permit post-stratification of sample sites according to level of
impact. An analysis of variance will be performed on the hydrocarbon
content of clam samples among sites. The results of this test will
be related to the level of sediment impact.

Objective B will be met through ANOVA contingent upon the processing
of necropsy samples. These samples will be processed if hydrocarbon
analysis is positive.

To provide baseline (pre-impact) information on variance in growth at
age among sites, an analysis of variance on growth parameters from
clams taken during 1989 between areas will be conducted. Growth
parameters will be determined for various growth curves, such as
Gompertz, von Bertalanffy, or polynomial equations. Growth
parameters will be presented for the most appropriate growth models
only. A similar ANOVA will be conducted on growth parameters from
clams taken during 1990 between areas. Those beach sites which are
resampled in 1990 will be subjected to an analysis of variance on
growth parameters obtained from fitting algorithms for clam growth
after impact (1990 and beyond) and will be compared to growth
parameters for clam growth prior to impact (approximately 1979-1989)
to resolve impact of oil contamination on growth (Objective C).
Graphics will be used to display differences in growth among areas
over time, including growth curves (size at age) and growth increment
at age by year for each beach.

To meet objectives D and E, a chi-square or an appropriate
nonparametric test will be used to test for significant differences
in proportions of dead clams (objective D) and young-of-the-year
clams (objective E) between treatment levels. Appropriate tests
involving relative abundance measures may also be used to meet
Objective E.

BIBLIOGRAPHY

Anderson, J.W., J.R. Vanderhorst, S.L. Kiesser, M.L. Fleishmann, and
G.W. Fellingham. 1982. Recommended Methods for Testing the Fate
and Effects of Dispersed Oil in Marine Sediments. In Oil Spill

149

Chemical Dispersants: Research, Experience, and
Recommendations. ASTM Special Technical Publication
840. Tom E. Allen Ed. Philadelphia, PA 19103. pp. 224-238.

Anderson, J.W., R.G. Riley, S.L. Kiesser, B.L. Thomas, and G.W.
Fellingham. 1983. Natural Weathering of Oil in Marine
Sediments: Tissue Contamination and Growth of the Littleneck
Clam, Prototheca staminea. Canadian Journal of Fisheries and
Aquatic Sciences. 40(Suppl. 2):70-77.

Augenfeld, J.M., J.W. Anderson, D.L. Woodruff, and J.L. Webster.
1980. Effects of Prudhoe Bay Crude 0 i 1 -Contaminated Sediments on
Protothaca staminea (Mollusca: Pelecypoda) : Hydrocarbon Content,
Condition Indexf Free Amino Acid Level. Marine Environmental
Research. 4(1980-81):135-143.

Dow, R.L. 1975. Reduced Growth and Survival of Clams Transplanted to
and Oil Spill Site. Marine Pollution Bulletin. 6(8):124-125.

Dow, R.L. 1978. Size-Selective Mortalities of Clams in an oil Spill
Site. Marine Pollution Bulletin. 9(2):45-48.

Keck, R.T., R.C. Heess, J. Wehmiller, and D. Maurer. 1978. Sublethal
Effects of the Water-soluble Fraction of Nigerian Crude Oil on
the Juvenile Hard Clams, Mercenaria (Linne). Environmental
Pollution. 15:109-119.

Neter, J., W. Wasserman, and M. Kutner. 1985. Applied Linear
Statistical Models. Richard D. Irwin, Homewood Illinois.

Nickerson, R.B. 1977. A Study of the Littleneck Clam (Prototheca
staminea Conrad) and the butter clam (Saxidomus giganteus
Deshayes) in a habitat permitting coexistence, Prince William
Sound, Alaska. Proceedings of the National Shellfisheries
Association. 67:85-102.

Paul, A.J. and H.M. Feder. 1973. Growth, recruitment, and
distribution of the littleneck clam, Protothaca staminea in
Galena Bay, Prince William Sound, Alaska. Fishery Bulletin
71 (3) :665-677.

Quinn, II, T.J., and N.F. Jones. 1989. Razor Clam (Siliqua patula,
Dixon) Investigations on the Eastside Cook Inlet Beaches. Juneau
Center for Fisheries and Ocean Sciences, University of Alaska
Fairbanks, Project Report UAF-JCFOS-8902. 165 p.

150

BUDGET: ADF&G

Salaries $ 121.4
Travel 5.0
Contracts 100.8
Supplies 0.0
Equipment 2.0

Total $ 229.2

151

FISH/SHELLFISH STUDY NUMBER 15

Study Title: Injury to PWS Spot Shrimp

Lead Agency: ADF&G

INTRODUCTION

This project will continue to determine possible damage to spot
shrimp, Pandalus platyceros, due to the EVOS. Spot shrimp are a
representative species of the deep water near shore benthic
ecosystem, serving as a food source for a variety of fish. They are
a commercially important species and also support subsistence and
personal use fisheries in PWS. This project is a continuation of FIS
Study 15 which was conducted during 1989-90.

Spot shrimp are known to be sensitive to oil contamination in both
the larval and adult phase, and the effects of oil on spot shrimp in
particular and shrimp in general are well documented (Anderson et al
1981, Brodersen et al 1977, Brodersen 1987, Mecklenburg, Rice and
Karinen 1977, Sanborn and Malins 1980, Stickle et al 1987,
Vanderhorst 1976). To determine the impacts that hydrocarbons from
the spill may have had on spot shrimp, samples will again be
collected from the three oiled and three non-oiled sites in western
PWS which had been surveyed in 1989. The data collected from the
samples will be analyzed to determine tissue hydrocarbon levels and
tissue damage. The collected data will also be tested to confirm or
reject the hypothesis that there is no significant difference in
hydrocarbon levels between the oiled and non-oiled areas. Relative
abundance, in terms of catch per unit effort, at each study site and
changes in relative abundance over time will be tested to determine
possible relationships with the level of oiling. A comparison with
historical records will also be made. The size composition of the
stock at each site will be estimated and, dependent upon recruitment
to the fishing gear, analyzed to determine whether the 1989 year
class suffered a high mortality rate in areas of high oil impact
relative to other year classes in non-oiled areas. Spot shrimp
fecundity will also be determined and tested for significant annual
and interannual differences between oiled and non-oiled sites.

OBJECTIVES

A. Estimate the relative abundance by weight and sex of spot
shrimp and the relative abundance by weight of incidentally
caught pink and coonstripe shrimp in oiled and unoiled areas and
compare these values to those obtained during the first
assessment survey in 1989.

B. Compare size and age frequencies (by sex and depth stratum)
between sites using mixture model analysis.

152

C. Estimate fecundity, egg mortality, and other sublethal effects
between oiled and non-oiled areas over time, and determine
whether those effects result in adverse changes in reproductive
viability.

D. Analyze tissue and egg samples for presence of hydrocarbons and
compare differences between oiled and non-oiled sites. Test the
hypothesis that the level of hydrocarbons is not related to the
level of oil contamination present at a site.

E. Document injury to tissues and compare differences between
oiled and non-oiled sites if warranted by results from tissue
hydrocarbon analysis.

METHODS

This project uses commercial spot shrimp pots of a standardized size
to catch spot shrimp in oiled and unoiled areas. Shrimp specimens
will be analyzed for Prudhoe Bay crude oil levels and necropsied to
determine if damage has occurred to tissues as a result of oil
contamination. As in the 1989 study plan, oiled and unoiled areas
will be sampled in two phases which correspond with two stages of egg
development. The first phase will occur in early November (1990)
following the fall molt and egg extrusion. The second phase will
occur in early March (1991) just prior to egg hatching. The sampling
strategy will be identical during both phases. Relative abundance
estimates of spot shrimp will be made using a stratified pot
deployment based on depth and location. Size distribution, species
composition, and reproductive data will also be collected. Previous
spot shrimp research in PWS is documented by Kimker and Donaldson
(1987), Donaldson (1989), Donaldson and Trowbridge (1989), and Kruse
and Murphy (1989).

This project will be carried out in two general areas. One will be
an area of little apparent impact, the northwestern portion of PWS.
This area includes Unakwik Inlet, the site of previous ADF&G research
on abundance and growth of spot shrimp. The second area will be
central and southwestern PWS, an area of generally high oil impact.
This area includes Green Island where ADF&G test fishing occurred in
1981. Within each of these two areas, fishing will take place at
three sites. In the northwestern sound, test fishing will occur in
Unakwik Inlet, Port Wells, and Culross Passage. In the central and
southwestern sound, test fishing will take place near Herring Bay,
Chenega Island, and Green Island. Shrimp distribution in these areas
has been established by surveying the commercial fleet.

Fishing will take place at six sites - three in oiled areas and three
in non-oiled areas. Each site will be stratified by depth. Stratum
1 will be shallow waters - 20 to 70 fathoms. Stratum 2 will be deep
waters - 70 to 120 fathoms. Based on past research, spot shrimp are
not abundant below those depth ranges. Because of the difficulty of
placing the gear at precise depths, it is impractical to divide the

153

depth into more than two strata. Strata span 50 fathoms in depth or
approximately 65 to 85 fathoms in width along the bottom at slopes of
75 to 100 percent. Fishing a 100 fathom string will span the width
of each strata and allow for a complete placement of gear over the
strata.

Eleven pots spaced 10 fathoms apart will be fished on a long line so
that each string of pots is 100 fathoms long. One 100 fathom string
of gear constitutes a sampling station. Two stations will be fished
in each stratum at each site for a total of 22 pots per stratum per
site, or 44 pots per site. Forty-four pots is the most that can be
fished in a day while collecting all of the various samples and data.
If necessary, pots will be redeployed an additional day at each site
and at each depth until a minimum of 500 shrimp are captured per
depth stratum. A total of 264 pots will be fished during each time
period.

Water temperature, salinity, and dissolved oxygen concentration by
depth will be recorded using a CTD, transferred from the CTD to a
micro-computer and stored on diskette. CTD casts will be at one
station in the deep stratum every day. The CTD will be lowered at a
rate of 60 meters per minute. Because of the configuration of the
CTD, only readings from the downcast will be used.

Total weight of catch, sub-sample weight, and the weight of each
species in a sub-sample will be recorded for each pot on a paper form
at the time the pot is retrieved. The total weight of shrimp per pot
will be determined by weighing the contents of each pot on an
electronic scale. The average number of shrimp per kilogram will be
determined. If less than 500 shrimp are estimated to be contained in
all of the pots, all of the shrimp will be sampled. If the pots are
estimated to contain more than 500 shrimp, a constant proportion by
weight of each pot will be sampled for a total sample of 500 shrimp.

Each sub-sample will be sorted by species. Weight and number of
animals will be recorded for each species. Only spot shrimp will be
retained for further data collection. All spot shrimp in the sub-
sample will be measured for carapace length to the nearest 0.1
millimeters using a digital caliper and sex will be determined as
male, transitional, or female. For female spot shrimp, egg color and
stage of development (eyed or uneyed) ; relative clutch size; presence
of breeding dress and egg parasites or parasitic externa will be
noted. Each female retained for fecundity analysis will be
identified with a code number to allow cross reference of fecundity
and other data.

Specimens for necropsy analysis will be taken after the catch is
weighed and processed. Twenty shrimp from a single station in each
stratum will be selected randomly to make up a necropsy sample.
Necropsy samples will be labeled with the date, station number,
latitude and longitude, sample number, project leader's name,
species, and agency.

154

To prevent contamination, specimens for hydrocarbon testing will be
taken from the pot immediately after removal from water and before
contents are weighed. Three spot shrimp will form one composite
sample. Each composite will be taken from a different pot. Two
replicates of the composite will be taken randomly from one station
in the stratum and the third replicate will come from the other
station. Three samples per site per depth stratum result in nine
samples per depth stratum (three sites X three samples) per impact
level and 18 samples per oil impact level (nine samples X two depth
strata). This will allow hypothesis testing to detect differences in
hydrocarbon levels of 1.2 standard deviations with the probability of
a type I or type II.error being 0.05 and 0.10, respectively.

The number of specimens for one hydrocarbon analysis is dependent on
the size of the specimens collected. Tissue volume based on the
average size of the species was estimated and the number of specimens
needed to provide 15 gm. of tissue was calculated to be three spot
shrimp. An estimate of three hydrocarbon samples from each treatment
level is needed for detecting contamination between levels.

Twenty five egg bearing females will be taken at random from each
station to estimate fecundity and egg mortality. A total of 24
stations will yield a total sample size of 600 females. Specimens
from each station will be individually labeled. Each sample bag will
be labeled with project leader's name, species name, "eggs", date,
station, and agency name.

Fecundity will be determined by removing the eggs from the pleopods,
drying each egg mass to a constant weight, weighing a sub-sample of
a known number of eggs, and expanding the sub-sample weight to the
weight of the entire clutch. Carapace length will be taken for each
specimen at the time the eggs are removed and recorded on the
fecundity form.

A minimum number of five shrimp from each station will be sampled for
fecundity which will allow an adequate sample (30 per depth strata
per oil impact level) to test for differences in fecundity between
depth strata and oil impact level.

Objective A will be addressed by estimating the average catch per pot
by weight, sex, and species. ANOVA will be used to test for
significant differences in each of these categories between strata
(depth) , sites, and oiled versus non-oiled areas. To def ine the
relationship between hydrocarbon levels and changes in relative
abundance, statistics for analysis of covariance or an appropriate
multivariate technique will be calculated to contrast differences in
hydrocarbon content and relative abundance in- oiled and non-oiled
areas. Changes in average catch per pot over time will also be
analyzed between different depth strata, sites, and oiled and non-
oiled areas.

A size frequency distribution will be made by species and sex to

155

address objective B. The hypothesis that there is no significant
difference between strata, and oil impact levels for size frequency
distribution will by tested using quantile-quantile plots, Chi-square
tests or other appropriate methods. A t test or a similar non-
parametric test will be used to test for similarity in means.
changes in size frequency distribution over time will be examined by
comparing data collected during, phase one and phase, two. A. t test
for means and an,appropriate method for comparing distributions will
be used to look for significant differences between time.periods as
well.

To meet objective C, the relationship between size and fecundity will
be examined. The percentage of spot shrimp females bearing eggs; the
stage of spot shrimp egg development (color and,presence or absence
of eyes); the percentage of spot shrimp egg fouling and egg
mortality; the fecundity by size; and the relative clutch size will
be determined for each station and each phase. Chi-square tests @will
be used to test for differences in stratal sites.and levels in data
which involve percentages and proportions. Differences between
strata, sites, and impact,levels:for fecundity and relative size of
clutch will be tested for using analysis of variance. ANOVA will
also be used to test for a significant difference in the above
measures between phase one and* phase two which may provide an
estimate of the number of eggs -dying over the course of the brood
period or estimates of differences in egg viability.

To address objectives D and E, the average levels of oil present in
spot shrimp tissue by strata and site will be estimated. Significant
differences in hydrocarbon concentrations between oiled and unoiled
sites will be tested by analysis of variance. To further define the
impact of hydrocarbon -levels on the stock, the percentage of animals
with abnormal tissues in oiled and unoiled areas*will be determined.
A chi-square test will be utilized to test for significant
differences in percentage of animals with abnormal tissues between
strata, sites, and impact levels.

BIBLIOGRAPHY

Anderson, J.W., S.L. Kiesser, R.M. Bean, R.G. Riley and B.L. Thomas.
1981. Toxicity of 'chemically dispersed oil to shrimp exposed to
constant and decreasing concentrations in a flowing system. In:
.1981 Oil Spill Conference (Prevention, . Behavior, Control,
Cleanup), Proceedings. Washington D.C. American Petroleum
Institute. Pp. 69-75.

Brodersen, C.C., S.D. Ricef, J.W. Short, T.A. Mecklenburg, and J.F.
Karinen. 1977. Sensitivity of larval and adult Alaskan shrimp
and crabs to acute exposures of the water-soluble fraction of
Cook Inlet crude oil. In: 1977 Oil Spill Conference (Prevention,
Behavior,, Control, Cleanup) , Proceedings. Washington, D.C.
American Petroleum Institute. pp. 575-578.

156

Brodersen, C.C. 1987. Rapid narcosis and delayed mortality in larvae
of king crabs and kelp shrimp exposed to the water-soluble
fraction of crude oil. Marine Environmental Research.
22(1987):233-239.

Donaldson, W. 1989. Synopsis of the Montague Strait experimental
harvest area 1985 - 1988. Alaska Department of Fish and Gamef
Division of Commercial Fisheries, Regional Information Report
No. 2C89-04. 21 pp.

Donaldson, W. and C. Trowbridge. 1989. Effects of rigid mesh panels
on escapement of spot shrimp (Pandalus I?latyceros) from pot
gear. Alaska Department of Fish and Game, Division of
Commercial Fisheries, Regional Information Report No. 2C89-05.
22 pp.

Kimker, A. and W. Donaldson. 1987. Summary of 1986 streamer tag
application and overview of the tagging project for spot shrimp
in Prince William Sound. Alaska Department of Fish and Game,
Division of Commercial Fisheries, Prince William Sound
Management Area Data Report 1987-07.

Kruse, G. and P. Murphy. 1989. Summary of statewide shrimp workshop
held in Anchorage during October 24-26, 1988. Alaska
Department of Fish and Game, Division of Commercial Fisheries,
Regional Information Report No. 5J89-##.

Mecklenburg, T.A., S.D. Rice, and J.F. Karinen. 1977. Molting and
survival of king crab (Paralithodes camtschatica) and coonstripe
shrimp (Pandalus hypsinotus) larvae exposed to Cook Inlet crude
oil water-soluble fraction. In: D.A. Wolfe (ed.). Fate and
Effects of Petroleum Hydrocarbons in Marine Ecosystems and
Organisms. Pergamon Press, New York, NY. pp. 221-228.

Sanborn, H.R. and D.C. Malins. 1980. The disposition of aromatic
hydrocarbons in adult spot shrimp (Pandalus platyceros) and the
formation of metabolites of naphthalene in adult and larval spot
shrimp. Xenobiotica. 10(3):193-200.

Stickle, W.B., M.A. Kapper, T.C. Shirley, M.G. Carls, and S.D. Rice.
1987. Bioenergetics and tolerance of the pink shrimp (Pandalus
borealis) during long-term exposure to the water-soluble
fraction and oiled sediment from Cook Inlet crude oil. In: W.B.
Vernberg, A. Calabrese, F.P. Thurberg, and F.J. Vernberg
(eds.). Pollution Physiology of Estuarine Organisms. Belle W.
Baruch Libr. Mar. Sci. 17, Univ. S. C. Press, Columbia. pp. 87-
106.

Vanderhorst, J.R., C. I. Gibson, and L J. Moore. 1976. Toxicity of No.
2 fuel oil to coonstripe shrimp. Marine Pollution Bulletin.
7(6):106-108.

157

BUDGET: ADF&G

Personal Services $ 44.0
Travel 1.4
Contractual 15.0
Supplies 4.6
Equipment 0.0

Total $ 65.0

158

FISH/SHELLFISH STUDY NUMBER 17

Study Title: Injury to Demersal Rockfish and Shallow
Reef Habitats in PWS and Along the Lower KP

Lead Agency: ADF&G

INTRODUCTION

In light of the findings of the first year of study of potential
impacts on rockfish populations conducted in 1989, continued study
of demersal rockf ish populations and shallow reef habitats is
warranted for 1990. Unlike many species of marine fish, demersal
rockfish complexes are relatively sedentary, residing near rocky
reefs and boulder fields. The potential impact of the oil spill on
various nearshore assemblages is dependent upon location of various
rockpiles. The potential uptake of various contaminants will be
related to the level of oil contamination and food web
characteristics of these reefs. of primary importance are
questions of transport of oil to subsurface habitats and the
RPtential f or residual persistence of this contamination. Khan
(1987) reports that crude oil can contaminate sediments and persist
for long periods of time in the environment.

Under these conditions, the petroleum hydrocarbons can exert a
broad range of effects on animals, from impaired feeding, growth,
reproduction, and changes in behavior; to tissue and organ damage,
damage to blood cells, changes in enzyme activity and changes in
parasite densities (Khan 1986; Khan 1987; Kiceniuk and Khan 1986;
Rice 1985; Wennekens et al. 1975; Malins et al. 1977; Rice et al.
1977; Gundlach et al. 1983; Hose et al. 1987; Spies et al. 1982).
These possible affects are especially critical to demersal rockfish
since they are long-lived, recruitment is low, and the potential
for long-term stock decline due to chronic exposure to crude oil is
high. Continuation of this study will help determine long term
histopathological effects on the fish and will quantify the extent
to which hydrocarbons persist in the environment.

Only limited baseline data are available for rockfish populations
in PWS and along the lower Kenai Peninsula (LKP). Rockfish were
studied as part of a study of nearshore f ish assemblages during the
years 1977-1979 in PWS (Rosenthal, 1980) and Morrison studied
select reefs along the LKP during 1980 through 1984. These
investigations provided descriptions of selected rockfish
populations including estimates of species and prey composition,
density, length and age composition.

OBJECTIVES

A. Determine the presence or absence of hydrocarbons in demersal

159

rockf ish, benthic prey species, benthic suspension feeders,
and sediments from two control and two treatment sites in PWS
and two control and two treatment sites along the LKP.

B. Determine the physiological effects resulting from oil
contamination through histopathological examination of five
organs, enzyme activity, examination of red blood cells for
circulating micronuclei; and the examination of developing
embryos.

C. Determine the feasibility of using toxicological analysis of
gonads and pituitary glands to ascertain effects of oil
contamination on growth and reproduction.

D. Determine the feasibility of using otolith microstructure to
evaluate depressed growth as a result of oil contamination.

METHODS

Eight sites (four treatment and four control) in PWS and along the
LKP will be sampled in 1990. Demersal species of rockfish, benthic
and epibenthic invertebrates, and finfish prey species,
unconsolidated benthic sediments and sessile suspension feeders
will be collected at each sample location for analysis of
hydrocarbons. From the results of these analyses the mechanism of
hydrocarbon uptake in demersal rockfish and the extent to which
hydrocarbons persist in reef ecosystems may be determined. The
effects of sublethal hydrocarbon contamination in demersal rockfish
may be determined through histopathological examination of five
organs; evaluation of enzyme activity; examination of red blood
cells for circulating micronuclei; and, the examination of
developing embryos. The feasibility of evaluating affects on
growth and fecundity through toxicological and biochemical analysis
of gonad and pituitary tissues, as well as determining depressed
growth through examinations of otolith microstructures, will be
explored. Results will be compared between control and treatment
sites. A pilot sampling trip will be made to determine what
species are present and to evaluate sampling techniques and site
selection.

Criteria for choosing sample reefs were based on: (1)
accessibility to boat and diving operations (ocean floor
surrounding the reefs were approximately 20 fathoms or less in
depth) ; (2) exposure of surface waters to oil; (3) location of
reported kills and/or sublethal contamination of demersal rockfish;
(4) occurrence of sampling by other oil spill assessment studies
relative to this study and (5) previous study sites of Rosenthal or
Morrison.

A systematic sampling design will be used to identify sampling
sites within each reef. Transects will be established at discrete

160

depths by deploying an anchor line along specific contours of the
reef and each end will be marked by anchored flag pole assemblies.
Coordinates, length, depth, and orientation of the transect will be
recorded. The actual number of sample sites will be depend on the
length of the transect and the orientation of the reef in the ocean
currents. During the pilot sampling trip prey and benthic species
that are common to all reefs and that are not transient will be
listed as target species. These species will then be collected at
each reef during the sampling trip. Sampling will be conducted
during late July and early August, the time frame that Rosenthal
(1980) identified as near the peak abundance of rockfish in
nearshore areas. Collection methods for finfish, prey species,
sediment, and sessile invertebrates are outlined below.

Twenty adult demersal rockf ish (target primarily yelloweye rockf ish
Sebastes ruberrimus) will be collected at each sample site using
hook and line jigging techniques. The sample size is based on the
number required for histological evaluation as determined by the
Histopathology Technical Group (Meyers, 1989). Baited lures will
be lowered to the substrate and raised enough to allow for adequate
jigging action. When a fish is on the line it will be retrieved
slowly in order to allow the air bladder to equilibrate and prevent
extrusion of the stomach and regurgitation of its contents. Where
excessive depths make this impractical, divers will enclose the
fish in a dive net to retain the stomach contents upon
regurgitation. Where hook and line techniques do not yield results
divers will verify the presence or absence of demersal rockfish
assemblages and if, present, collect them using spear guns.
Stomach contents will be collected to determine composition of the
prey species. Species identification of adult rockfish will be
accomplished using the methods of Kramer and O'Connell (1988) and
Hart (1973).

Fifty juvenile demersal rockfish will be collected using variable
mesh, monofilament gillnets set in the shallow areas of the reef
and in intertidal zones adjacent to the reefs. Given estimated
proportions of 0.6 and 0.2 respectively, sample size was determined
(Zar 1984) to be 50, where a =.05. Species identification of
juvenile rockf ish species will be accomplished using the methods of
Matarese et al. (1989).

Ten samples of prey species (Rice, 1990) at each reef will be
collected for hydrocarbon analysis. The sample size is based on
the number required for hydrocarbon analysis as established by the
Analytical Chemistry Group (Manen, 1989). The species to be
collected will be determined during the pilot sampling trip.
Additional information used to select prey species will be based on
the analysis of stomach content samples and previous food ecology
studies (Rosenthal et al., 1988; Rosenthal, 1980). Divers
outfitted with SCUBA gear will use an air-lift sampler to collect
benthic prey species at each site (Chess, 1978). The air-lift
sampler uses suction to collect all organisms from a square meter

161

area and deposits them into a sampling container. Additional food
organisms may be collected using a variety of other techniques
depending on the target species. crab pots and shrimp pots will be
used to collect crustacean species. Trammel nets, plankton tows,
and diver controlled nets will be used to capture appropriate
target species.

.Nine sediment samples (Rice, 1990) will be collected at each sample
site by divers outfitted with SCUBA equipment prior to the
collection of air-lift samples outlined above. Each sample will
consist of ten 2 cc scoops taken from the top 2 cm of the substrate
along a 10 m long transect. Excess water will be poured of f at the
surface and the sample will be frozen. Three sediment samples will
be collected at each reef.

Three samples of sessile filter feeders (Rice, 1990) will be
collected from each reef by divers outfitted with SCUBA equipment.
Each sample will consist of pieces of two or three sessile filter
feeders. Enough samples will be collected to at least half fill a
4 oz. hydrocarbon sampling jar.

Samples collected will be handled differently depending upon the
data required and type of analysis being conducted. The following
sections explain each type of preparation that will be used. Most
samples collected will be used for only one type of analysis,
however, each rockfish captured will be used or prepared for a
variety of purposes. Rockfish will be processed in the following
specific order: 1) immediately after capture blood samples will be
drawn and slides prepared; 2) rockfish will be measured to the
nearest millimeter (fork length) and weighed to the nearest gram
for calculation of condition factor; 3) tissue will be sampled for
hydrocarbon analysis and histopathological evaluation according
procedures outlined in proceeding sections; and, 4) otoliths will
be removed for later age determination.

Length (fork length) , to nearest millimeter, and weight, to the
nearest gram, will be used to calculate a relative condition
factor. Condition factors will be calculated for all rockfish
captured.

Ten of the 20 rockfish (Rice, 1990) collected at each reef will be
prepared for hydrocarbon analysis. All samples will be collected
from live fish. Bile samples will be collected first by removing
the whole gall bladder and emptying the bile into 0.5 oz. amber
sampling jars. Ten grams each of stomach, pyloric caeca, liver,
and muscle tissue will be collected from each rockfish. Each
tissue type will be stored in separate 4 oz. sampling jars.

Ten samples of prey species (Rice, 1990) will be collected at each
reef. Different preparation methods will be conducted depending
upon the prey species being collected. Larger fish will be handled
in the same manner as the rockfish. Smaller fish, and other small

162

organisms where dissection is not practical, will be collected
whole in sufficient numbers to fill a 4 oz. sampling jar half to
three-quarters full.

Twenty live demersal rockfish, including the ten sampled for
hydrocarbons, will be collected at each reef for histopathological
analysis and processed under the guidelines outlined by the
Histopathology Technical Group (Meyers, 1989). Blood samples will
be collected using a heparinized syringe inserted between the
vertebrae in the caudal peduncle and smears made and fixed for
later staining with May-Grunwald-Giesma stain. The liver will be
visually examined for discoloration, blotchiness, and firmness and
its condition recorded. One centimeter sections of tissue will be
removed from the following organs: liver-pancreas, kidney, gills,
gonads, and eyes. All developing embryos will be collected and
preserved in a neutral formalin solution.

Sagittal otolith pairs will be collected from fifty juvenile
yelloweye rockfish (measuring less than 200 mm) from each reef.
Age validation studies involving daily growth increments, such as
Boehlert and Yoklavich (1987), typically utilize otoliths from
juveniles because growth is deposited more rapidly, and
physiological checks and daily growth increments are more visible.
Upon collection, otoliths will be rinsed and stored dry in pairs in
coin envelopes.

Juvenile otoliths will be prepared for examination following
methods outlined by Boehlert and Yoklavich (1987). Otoliths will
be viewed under transmitted light with a compound microscope at
40OX magnification. Presence and location of hyaline zones
comprising annuli, daily growth increments, and checks resulting
from physiological factors including a reduction in growth will be
examined. The feasibility of distinguishing differences in the
type of zones will be explored by measuring the width of growth
zones deposited over consecutive periods of time (days and years).
Where physiological checks are clearly discernible from annuli, the
presence of checks will be determined with respect to annuli.
Checks deposited within the growth zone of the previous year will
be noted. The proportion of otoliths containing checks within this
growth zone will be compared between control and treatment groups.

DATA ANALYSIS

Data analysis will consist primarily of the comparison of results
between control and treatment groups for each of the following:

LeCren's relative condition factor (Kn) (Anderson and Gutreuter,
1983) will be calculated for each adult and juvenile rockfish.
The mean condition factor for adult and juvenile rockfish for each
reef will be calculated and differences between control and
treatment groups will be tested using ANOVA.

163

Rockfish tissues, prey species and sessile filter feeders will be
analyzed f or presence of hydrocarbons. Proportions of contaminated
samples in each category will be compared between control and
treatment groups.

For each species the proportion of treatment sites containing
contaminated samples will be compared to the proportion of control
sites with contaminated samples using a two-sampled z-test from Zar
(1984).

Tissues will be examined for histopathological abnormalities and
enzyme activity, and blood will be examined for circulating
micronuclei by the Histopathological Technical Group. The
proportion of samples showing evidence of histopathological
abnormalities will be compared between control and treatment groups
for each tissue type using the z-test from Zar (1984).

Otoliths from juvenile demersal rockfish will be examined as
described in the methods section. Proportion of otoliths
containing checks between the last two annuli will be compared
between control and treatment groups using the z-test from Zar
(1984). Age composition and mean length-at-age will be calculated
for each species of rockfish.

BIBLIOGRAPHY

Anderson, R.O., and S.J. Gutreuter. 1983. Length,
weight, and associated structural indices. Chapter 15 IN:
Fisheries Techniques, L.A. Neilson and D.L. Johnson eds.,
American Fisheries Society, Bethesda, Maryland.

Boehlert, G.W. and M.M. Yoklavich. 1987. Daily growth increments
in otoliths of juvenile black rockfish, Sebastes melanops: An
evaluation of autoradiography as a new method of validation.
Fishery Bulletin: Vol. 85, No. 4. pp. 826-832.

Chess, J.R. 1978, An airlift sampling device for in situ
collecting of biota from rocky substrate. Marine Technology
Society Journal, 12:20-23.

Gundlach, E.R., P.D. Boehm, M. Marchand, R.M. Atlas, D.M. Ward, and
Douglas A. Wolfe. 1983. The fate Amoco Cadiz oil. Science
221:122-129.

Hart, J.L. 1973. Pacific Fishes of Canada. Bulletin 180,
Fisheries Research Board of Canada. Ottawa, Ontario, Canada.
pp. 388-453.

Hepler, K., A. Hoffmann, and T. Brookover. 1989. Injury to
rockfish in Prince William Sound. State/Federal resource
damage assessment data summary report. Fish/Shellfish Study

164

Number 17. ADF&G Sport Fish. Anchorage, Alaska.

Hose J.E., J.N Cross, S.G. Smith and D. Deihl. 1987. Elevated
circulating erythrocyte micronuclei in fishes from
contaminated sites in California. Marine Environmental
Research, 22:167-176.

Khan R.A. 1986. Effects of chronic exposure to petroleum
hydrocarbons on two species of marine fish infected with
hemoprotozoan, Trypanosoma muranensis. Can. J. Zool. 65:2703-
2709.

Khan R.A. 1987. Crude oil and parasites in fish. Parasitology
Today, 3:99-102.

Kiceniuk J.W. and R.A. Khan. 1986. Effects of petroleum
hydrocarbons on Atlantic cod, Gadus Morhua, following chronic
exposure. Can. J. Zool. 65:490-494.

Kramer, D.E. and V.M. O'Connell. 1988. Guide to Northeast Pacific
Rockfishes Genera Sebastes and Sebastolobus. University of
Alaska Marine Advisory Bulletin No. 25.

Malins, D.C., E.H. Gruger, Jr., H.O. Hodgins, N.L. Karrick, and
D.D. Weber. 1977. Sublethal effects of petroleum
hydrocarbons and trace metals, including biotransformations,
as reflected by morphological, chemical, physiological,
pathological, and behavioral indices. OCS Energy Assessment
Program. Seattle, Washington.

Manen, C. A., Chairperson. 1989. State/federal damage assessment
plan, Analytical Chemistry Group, National Marine Fisheries
Service, Auke Bay, Alaska.

Matarese A.C., A.W. Kendall Jr., D.M. Blood, and B.V. Vinter.
1989. Laboratory guide to early life history stages of
northeast Pacific fishes. NOAA Tech. Rep. NMFS 80. National
Oceanic and Atmospheric Adm., National Marine Fisheries
Service. Seattle, Washington 98115. 625 pp.

Mey ers, T. R., Chairperson. 1989. State/federal damage assessment
plan, Histopathology Technical Group, Alaska Department of
Fish and Gamel Fisheries Rehabilitation, Enhancement, and
Development Division, Juneau, Alaska.

Ricef S.D., J.W. Short, and J.P. Karinen. 1977. Comparative oil
toxicity and comparative animal sensitivity. In: "Fate and
effects of petroleum hydrocarbons in marine organisms and,
ecosystems, Proceedings", Wolfe, Douglas A., ed., Pergamon
Press, New York. pp. 78-94.

Rice, S.D. 1985. Effects of oil on fish. Chapter 5 IN:

165

Petroleum effect in the Arctic environment. F.R. Engelhardt
ed. pages 157-182. Elsevier Applied Science Publishers,
London.

Rice, S.D. 1990. Personal communication. Analytical Chemistry
Group, National Marine Fisheries Service, Auke Bay, Alaska.

Rosenthal, R.J., Victoria Moran-O'Connell and Margaret C. Murphy.
1988. Feeding ecology of ten species of
rockf ishes (Scorpaenidae) f rom the Gulf of Alaska. Calif . Fish
and Game 74:16-37.

Rosenthal, R.J. 1980. Shallow water fish assemblages in
northeastern gulf of Alaska: habitat evaluation, species
composition,abundance, spatial distribution and trophic
interaction. Prepared for NOAA, OCSEAP Program. 84 pp.

Rubin, J. 1988. A review of petroleum toxicity and fate in the
marine environment, with implications for the development of
a penalty table for spilled oil. Institute for Marine
Studies, University of Washington. Seattle, Washington.

Spies, R.B., J.S. Felton, and L. Dillard. 1982. Hepatic mixed-
function oxidases in California flatfishes are increased in
contaminated environments by oil and PCB ingestion. Marine
Biology 70:117-127.

Wennekens, M. P., L. B. Flagg, L. Trasky, D. C. Burbank,
R. Rosenthal, and F. F. Wright. 1975. Anatomy and potential
costs of an oil spill upon Kachemak Bay. Alaska Department of
Fish and Game, Habitat Protection Section Anchorage, Alaska.

Zar, J.H. 1984. Biostatistical Analysis. Prentice Hall, Inc.,
Edgewood Cliffs, New Jersey.

BUDGET: ADF&G

Personnel $ 40.9
Travel 2.7
services 63.6
supplies 1.2
Equipment 1.0

Total $ 109.4

166

FISH/SHELLFISH STUDY NUMBER 18

Study Title: Prince William Sound Trawl Assessment

Lead Agency: NOAA

INTRODUCTION

This project is a continuation of a multispecies trawl survey to
collect samples from bottomfish for 'hydrocarbon analyses. Its
purpose is to determine if bottomfish in PWS are still exposed to
oil or oil components from the EVOS and, if so, the geographical
extent of the exposure. The study will be conducted for 12 days
during June 1990.

OBJECTIVES

A. Collect bile and tissue samples, and stomach contents from
bottomfish, especially the utilized species.

B. Use CTD instrument to profile water characteristics throughout
the sampling area.

C. Preserve any fish observed with abnormalities of any type for
subsequent analysis.

METHODS

Sample collection will be done from the RV John N. Cobb using 400-
mesh Eastern otter trawls at known trawl locations in the Sound.
CTD profiles will be taken at each trawl station to provide
information on the structure of water masses.

Because the samples must be obtained from live fish, the hauls will
be of short duration (probably 10 minutes or less) to avoid death
from capture. The fish will be placed in live tanks and samples
taken immediately after capture. Procedures for taking samples
will be the same as in the 1989 survey, with the exceptions that
the samples will be obtained only from live fish and that samples
will be frozen immediately after being taken.

Six tows per day are planned for each of the 10 days. The 60 tows
will be distributed among the areas and depth strata used for the
1989 summer survey. The 1989 summer survey sampled six areas but
the 1990 survey will sample only five of the six areas; one area,
Port Wells, had negligible catches in 1989 and this area will not
be sampled in 1990. The same depth strata used f or the 1989
sampling will be used for the 1990 sampling:

167

Stratum Depth (fm)

1 10-50

2 51-100

3 101-200

4 201-400

Eleven depth strata occur in the five areas:

Area Depth strata to be sampled

Hinchenbrook 2,3

Orca/Fidalgo 1,2

Central Basin 3,4

Knight Island 2,3,4

outside 2,3

The tows (60) will duplicate the 1989 stations to maximize the
number of successful tows and provide coverage throughout the
Sound. The stations to be sampled in 1990 include both oiled and
unoiled 1989 areas and extend from the central basin south to
Montague Strait. Species that will be sampled will include
halibut, walleye pollock, flathead sole, and Pacific cod. The
maximum number of live individuals that can be handled will be
processed prior to retrieving the next tow. Halibut and pacific
cod are not anticipated to be as common in the hauls as walleye
pollock and flathead sole. If time permits at the end of the
cruise, stations where halibut and walleye pollock were few in
number will be resampled to increase the number of hydrocarbon
samples for these species.

Exposure of fish to oil will be determined by measuring
concentrations of metabolites of aromatic petroleum compounds in
bile. Analytical procedures used for the bile metabolite assays
will likely include excitations/emission measurements at
wavelengths where naphthalene and phenanthrene fluoresce. if
exposure is documented through bile analysis, analysis of tissue
and stomach content samples will occur. Estimated exposures to
petroleum hydrocarbons will be available to other investigations
(particularly Study Number 24) to assess environmental damage using

168

statistical and simulation models. These other studies will meld
bile and tissue chemistry to establish relationships between
biological damage and estimated exposures to hydrocarbons

BUDGET: NOAA

salaries $ 65.4
Travel 10.7
Ship Coast 100.0
Supplies 10.0
Equipment 0.0

Total $ 186.1

169

FISH/SHELLFISH STUDY NUMBER 22

Study Title: Injury to Crabs outside PWS

Lead Agency: NOAA

INTRODUCTION

Dungeness crabs in Alaska occupy nearshore areas in protected bays
and estuaries. These habitats are usually characterized by fine
benthic sediments and minimal wave action. If oil becomes
incorporated in shallow subtidal sediments it persists and can
affect crab populations for several years after an oil spill (Krebs
and Burns 1977, Boehm et al. 1987). Dungeness crab may be
especially susceptible to contamination by petroleum hydrocarbons
because they often burrow into benthic sediments; ovigerous female
Dungeness crab, in particular, are known to spend long periods (up
to 10 months) burrowed into sediments while brooding their eggs.

Several studies have documented deleterious effects on crabs
exposed to petroleum hydrocarbons. Sublethal concentrations can
result in early post-molt autotomy of limbs, behavioral disorders
and reduced reproductive capacity (Karinen and Rice 1974, Krebs and
Burns 1977, Karinen et al. 1985 and Malan 1988). Sex and
reproductive state may determine responses of crabs to oil
pollution. Krebs and Burns (1977) noted a greatly reduced
proportion of females in populations of the fiddler crab, Uca
pugnax, at oil contaminated sites in Buzzards Bay, Massachusetts.
Reproductively active ghost crabs, Ocypode quadrata, are more
sensitive to the water soluble fraction of crude oil than are crabs
not in reproductive condition (Jackson et al. 1981).

This project is a continuation of work begun in 1989 and will
provide quantitative data regarding adverse impacts on populations
of Dungeness crab outside PWS as a result of the EVOS. The project
will provide information on hydrocarbon levels in benthic sediments
occupied by crabs as well as hydrocarbon levels in the tissues of
crabs in contaminated and uncontaminated areas near Kodiak Island
and the eastern Alaska Peninsula. It will also provide biological
data on fecundity, reproductive capacity, and distribution and
relative abundance of the crabs. These data will permit the
assessment of short-term losses caused by contamination of
harvestable crabs and long-term impacts owing to adverse effects on
crab reproduction. The data will also contribute to the long-term
data base for management of fisheries and assessment of future oil
spills.

Products will include estimates of the amount of petroleum
hydrocarbons taken up by the tissues of crabs inhabiting areas with
contaminated sediments, estimates of the impact of hydrocarbons
taken up by crab reproductive tissues on crab fecundity and

170

reproductive capacity and identification of possible contamination
pathways from sediments to crab reproductive tissues and developing
eggs.

OBJECTIVES

A. Determine the levels of hydrocarbons, if present, in Dungeness
crabs in oiled and unoiled sites along the eastern Alaska
Peninsula and/or near Kodiak Island.

B. Assess reproductive condition of crabs in oiled and unoiled
areas by measuring such variables as percentage of ovigerous
crabs, fecundity and egg loss, condition and development.

C. Determine the incidence of limb loss and of abnormalities in
newly formed crab exoskeletons in oiled and unoiled areas.

D. Compare the strength of larval settlement in oiled and unoiled
areas using artificial substrates.

E. Identify potential methods and strategies for restoration of
lost use, populations, or habitat where injury is identified.

METHODS

The study will be conducted at eight sites with populations of
Dungeness crab near Kodiak Island and the eastern Alaska Peninsula.
Five sites will be in oiled areas and three will be reference sites
in unoiled areas. Final site selection will, of necessity, be made
in the field and will depend on the presence of Dungeness crab
populations as determined by diver observations. A list of
candidate sites will be compiled prior to departure. Dungeness
crab will be sampled by diving. sex, carapace width, presence or
absence of an egg clutch and external physical condition will be
recorded for each crab.

A total of 30 live female crab' will be sampled from each site
during each sampling period. Divers will swim three transects to
collect female crabs. Three randomly-selected subsamples of ten
female crab each will be taken from the divers' total catch. The
specimen number, carapace width, fresh weight, clutch size and a

1 Determination of sample sizes for all variables covered
by this project depends on estimates of sample size for
comparable variables by Margaret C. Murphy in the Project
operational Plan on "The effects of hydrocarbons on reproduction
inlDungeness crab."

171

description of the physical condition of each female crab in the
samples will be recorded. The left fifth pleopod will be removed
from each crab and fixed in 5% formalin for subsequent processing
to estimate egg development, egg mortality and egg fouling. Three
crab will be randomly selected from the ten crab in each sample,
measured and sacrificed for ovaries and hepatopancreas. Three
ovaries will equal one composite hydrocarbon sample. Three
hepatopancreas will equal one composite hydrocarbon sample. One
composite hydrocarbon sample of eggs will be taken by clipping a
small portion of the egg clutch (4 g = 1/3 of pleopod) from the
right fifth pleopod of each of the three crab. The remaining seven
crab from the subsample of ten will be returned to the sea.

Artificial substrates will be used to assess the intensity of the
settlement of larval Dungeness crab at oiled and unoiled sites.
Ten artificial substrates will be installed at about 0.5 m above
the bottom at each site. At those sites with eelgrass beds the
collectors will be placed about 3 m apart along an isobath just
below the lower limit of the eelgrass; at those sites lacking
eelgrass beds the substrates will be placed 3 m apart along the 6
m isobath. The substrates will be put in place in mid-May and will
be sampled for the megalopae of Dungeness crab in July and August.
The substrates will be retrieved in August at the time of the
second sampling.

Sediment samples will be collected at all sites. Divers will
collect three composite sediment samples along a 30 m transect laid
parallel to shore in the area where divers collected the crabs.
Each composite sample will consist of eight subsamples collected
randomly along the 30 m transect. All sediment samples collected
by divers will be taken from the top 2 cm of the sediment column.

Physical oceanographic data will be collected at each site during
each sampling period using an instrument that measures CTD. CTD
will be recorded every 2 seconds as it is lowered to the bottom and
raised to the surface. The CTD measuring instrument will be
deployed once at each site during each sampling period.

Definitive analysis of the chemical composition of petroleum
hydrocarbons in the sediments, tissues and eggs will be
accomplished in the laboratory with gas chromatography/mass
spectrometry as directed by the Analytical Chemistry Control Group.
The types of analyses to be performed on the samples will be
determined by the Analytical Chemistry Group and will include 1)
characterization of oil in marine sediments and cra *b tissues, 2)
total organic carbon on selected samples, and 3) size fraction
analysis on representative sediment samples. Prescreening analyses
of collected samples will occur prior to full gas
chromatography/mass spectrometry analysis in areas of low
likelihood of oiling. Details of the methods used in the chemical
analyses are recorded under the Quality Assurance Program.

172

The number of specimens required for one hydrocarbon analysis
depends on the amount of tissue available in a crab and the need
for a composite sample. Three Dungeness crab are enough to provide
15 g of ovarian tissue. One pleopod from an average clutch would
provide 15 g of crab eggs, but a sample representative of more than
one crab is desirable. Therefore egg clips from the clutches of
three crab will be combined to form a composite sample for
hydrocarbon analysis. Three hydrocarbon samples from each site are
the minimum needed to detect contamination between oiled and non-
oiled sites. A sample size of 30 crab was estimated to be an
adequate number to determine differences in reproductive output
between impact levels.

All data will be tested for heteroscedasticity with Bartlett's test
,or equivalent. Data will be reported as means and 95% confidence
intervals calculated according to a standard formula (Sokal and
Rohlf 1981). Parametric statistics (analysis of variance and
Scheffe's a posteriori test) will be used to test for differences
in means between oiled and non-oiled sites if underlying
assumptions of the parametric procedures are met, otherwise
nonparametric tests (eg. the Kruskal-Wallis test) will be employed.
Variables to be tested will include hydrocarbon concentrations in
Dungeness crab tissues, the reproductive parameters of Dungeness
crabs, crab larval production and viability and hydrocarbon content
of sediments in crab habitat.

Further multivariate statistics (eg. analysis of covariance, rank
correlation coefficients, discriminate analysis) will be computed
if the above summary statistics indicate relationships may exist
between Dungeness crab hydrocarbon content, reproductive capacity,
sediment hydrocarbon content, and physical oceanographic factors.

BIBLIOGRAPHY

Boehm, P. D., M. S. Steinhauer, D. R. Green, B. Fowler, B.
Humphrey, D. L. Fiest, and W. J. Cretney. 1987. Comparative
fate of chemically dispersed and beached crude oil in in
subtidal sediments of the arctic nearshore. Arctic 40, supp.
1: 133-148.

Jackson, L., T. Bidleman" and W. Vernberg. 1981. Influence of
reproductive activity on toxicity of petroleum hydrocarbons to
ghost crabs. Mar. Pollut. Bull. 12: 63-65.

Karinen, J. F. and S. D. Rice. 1974. Effects of Prudhoe Bay crude
oil on molting tanner crabs, Chionoecetes bairdi. Mar. Fish.
Rev. 36: 31-37.

173

Karinen, J. F., S. D. Rice, and M. M. Babcock. 1985. Reproductive
success in Dungeness (Cancer magister) during long-term
exposures to o i 1 -contaminated sediments. Final Report-OCSEAP
P-Unit 3008, Anchorage, Alaska, 28 pp.

Krebs, C. T. and K. A. Burns. 1977. Long-term effects of an oil
spill on populations of the salt-marsh crab Uca pugnax.
Science 197: 484-487.

Malan, D. E. 1988. The effects of Qatar light crude oil on the
saltmarsh crab Sesarma catenata and its implications in the
field: toxicity to adults and larvae. S. Afr. J. mar. Sci. 7:
37-44.

Neter, J., W. Wasserman, and M. Kutner. 1985. Applied Linear
Statistical Models. Richard D. Irwin, Homewood, Illinois.

O'Clair, C. E. and J. L. Freese. 1988. Reproductive condition of
Dungeness crabs, Cancer magister, at or near log transfer
facilities in southeastern Alaska. Mar. Env. Res. 26: 57-81.

Sokal, R. R. and F. J. Rohlf 1981. Biometry. W. H. Freeman and
Company, San Francisco. 859pp.

BUDGET: NOAA

salaries $ 28.0
Travel 8.0
Ship Costs 50.0
Contracts 12.0
supplies 6.0
Equipment 6.0

Total 110.0

174

FISH/SHELLFISH STUDY NUMBER 24

Study Title: Assessment of Oil Spill Impacts on Fishery
Resources: Measurement of hydrocarbons and their
metabolites, and their effects, in important species.

Lead Agency: NOAA

INTRODUCTION

Preliminary analyses of data collected in 1989 indicates that
several nearshore fish species were exposed to petroleum or
petroleum derivatives in and around PWS, subsequent to the EVOS
(Varanasi et al., 1990). Because petroleum and its components can
cause severe damage to fishery resources, this study provides for
the continued monitoring of the nearshore fisheries resources of
PWS and adjacent areas. Such monitoring will include measurement
of petroleum exposure and short-term effects, as was done in the
summer and fall of 1989, but will also encompass an assessment of
long-term biological effects, including measurements of
reproductive dysfunction and histopathological lesions of liver,
gill, kidney, and gonad.

Certain petroleum components C e.g. aromatic hydrocarbons (Ahs) ) can
cause reproductive toxicity and teratogenicity in rodents (Shum et
al., 1979; Gulyas and Mattison 1979, Mattison and Nightingale,
1980). Similarly, reproductive impairment has been noted in
benthic fish residing in contaminated areas of San Francisco Bay
(Spies and Rice, 1988) and southern *California (Cross and Hose,
1988). Moreover, English sole from areas of Puget Sound having
high sediment concentrations of Ahs showed inhibited ovarian
maturation (Johnson et al., 1988), and fish from these areas that
did mature often failed to spawn after hormonal treatment to induce
spawning (Casillas et al., 1990).

In general, reproductive impairment (including reduced plasma
levels of the sex steroid hormone, estradiol) was found in English
sole which showed evidence of exposure to aromatic compounds.
Moreover, our laboratory studies have shown that plasma levels of
estradiol are reduced in gravid female English sole exposed to
chemical contaminants extracted from urban sediments (Stein et al.,
1990), and, more importantly, our preliminary studies have also
shown that exposure to Prudhoe Bay crude oil reduced plasma levels
of estradiol in gravid female rock sole.

In view of our findings last year that several nearshore fish
species in and around PWS have been exposed to crude oil
components, including Ahs, the assessment of possible reproductive
dysfunction in animals from impacted areas will be very important
in determining biological damage to living marine resources as a
result of the oil spill. ovaries of selected species will be

175

histologically examined to determine if ovarian maturation is being
affected in animals from oil-impacted areas, and to determine
fecundity and levels of plasma estradiol in these same animals.
Combined with measurements of metabolites in bile and enzyme
activities in liver, these studies will enable us to estimate the
degree of reproductive dysfunction which may be occurring in oil-
exposed fish.

Exposure of animals to crude oil can also result in changes at the
tissue and cellular levels (National Academy of Sciences, 1985).
Examples of such changes after exposure of fish to o i 1 -contaminated
sediments include liver hypertrophy and fatty liver in winter
flounder (Payne et al., 1988) and the occurrence of hepatocellular
lipid vacuolization in English sole (McCain et al., 1978) . Certain
Ahs (e.g., benzo(a]pyrene) are known carcinogens in rodents (Lutz,
1979), and studies with several bottomfish species show that, of
the xenobiotic chemicals in sediments, Ahs are most strongly
associated with high prevalences of liver lesions, including
neoplasms (Malins et al., 1984; Myers et al., 1987; Black et al.,
1983; Varanasi et al., 1987) . Generally, histopathological lesions
of the types noted above do not become manifest until at least
several months after exposure. However, by the summer of 1990,
fish in and around oil impacted sites will have potentially been
exposed to petroleum components for more than one year, and
juveniles of some species of salmon will have potentially been
exposed during most of their developmental period. Accordingly,
assessment of histopathological effects in selected species is
strongly warranted, and is included in this proposal.

This study will continue to measure exposure to oil and oil
components in the biota of PWS and other areas affected by the oil
spill, by determining levels of hydrocarbon metabolites in bile and
by measuring hepatic enzyme activities. A range of biological
effects, especially indicators of reproductive dysfunction and
histopathological effects will be measured. By employing such a
broad spectrum of state-of-the art chemical, biochemical and
biological methods, analytical data will be obtained to document
the degree of exposure and resultant biological effects of
petroleum hydrocarbons on economically and ecologically important
fish species. This information for important Alaskan fish
species, will be incorporated into models for use in estimating oil
spill impacts on fishery resources.

OBJECTIVES

A. Sample selected nearshore fish species from 14 sites inside
and outside PWS, with emphasis on sites outside PWS. Site
selection is primarily based on data from last year's sampling and
analyses. Representative sediment samples will also be taken from
each sampling site for subsequent chemical analysis.

B. Estimate the exposure to petroleum hydrocarbons by measuring

176

levels of hydrocarbon metabolites in bile of the above species from
oiled and nonoiled habitats to detect significant differences in
bile concentrations with a = 0. 05 and b = 0. 10. Additionally,
stomach contents of fish showing high levels of hydrocarbon
metabolites in bile will be analyzed for hydrocarbons, such to
detect signif icant dif f erences in concentrations with a = 0. 05 and
b = 0.10

C. Estimate the induction of hepatic aryl hydrocarbon hydroxylase
(AHs) activity or increased levels of cytochrome P-45OIAl in the
above species from oiled and nonoiled habitats such to detect
statistical differences in levels of effects with a = 0.05 and b
0.10.

D. Estimate the prevalence of pathological conditions in the above
species from oiled and nonoiled habitats such to detect statistical
differences in levels of effects with a = 0.05 and b = 0.10.

E. Estimate the levels of plasma estradiol, the degree of ovarian
maturation, and fecundity in adult females of two of the above
species (Dolly Varden char and yellowfin sole) from oiled and
nonoiled habitats such to detect statistically significant
differences with a = 0.05 and b = 0.10.

F. Estimate temporal changes in the parameters described in
Objectives B&C, by comparing data obtained in 1990 to data obtained
in 1989. In order to assess either recovery or increased damage of
habitats from the oil spill, trends in these parameters must be
statistically significant at a = 0.05 and b = 0.10.

G. Construct simulation models similar to those of Schaaf et. al.
(1987) for important Alaskan fish species for use in estimating oil
spill impacts on fishery resources. These models will incorporate
pre-spill information from the fisheries literature on mortality
and fecundity together with information on reproductive impairment,
pathological conditions, and biochemical effects in fish exposed to
petroleum hydrocarbons as a result of the spill.

METHODS

General Strategy and Approach

Samples of biota will be collected from approximately 14 sites
during 1990, from mid-May to mid-June. Sites will generally be the
same sites occupied last year, and will be located in potentially
oil-impacted and unimpacted areas in PWS, in CI, and in embayments
on the KP, Alaska Peninsula, and Kodiak Island. As feasible, the
sample locations will be coordinated with A/W Study 2. Dolly
Varden char and juvenile salmon will be sampled in intertidal
areas, whereas f latf ish (e. g. Pacif ic halibut, yellowf in sole, rock
sole and flathead sole) will be sampled in subtidal areas. Salmon
and halibut were selected primarily because of their economic

177

importance, and f lathead sole, yellowf in sole, rock sole, and Dolly
Varden were selected because of their wide geographical
distribution and year-round residency in the sampling areas.
surf icial sediment samples for establishing levels of petroleum
hydrocarbon residues will be collected at all sites, with analyses
projected to be done under A/W Study 2.

Petroleum exposure of fish will primarily be assessed by measuring
(a) concentrations of metabolites of aromatic petroleum compounds
in bile, and (b) AHs activities in liver. These types of
measurements are necessary because petroleum hydrocarbons in fish
are rapidly metabolized to compounds that are not detectable by
routine chemical analyses. AHs activity in fish is due primarily
to a single cytochrome P-450, apparently cytochrome P-4501Al
(Varanasi et al., 1986, Buhler and Williams 1989). Measurement of
hepatic AHs activity will provide a very sensitive indicator of
contaminant exposure of sampled animals (Collier and Varanasi,
1987). Moreover, the induction of AHs activity indicates not only
that contaminant exposure has occurred, but also that biological
changes have occurred as a result of the exposure. In addition to
measuring AHs activity, cytochrome P-4501Al will be directly
quantitated in selected liver or tissue samples by an
immunochemical method recently developed at the University of
Bergen (Collier et al., 1989; Goksoyr, 1990).

Other biological effects in fish will be estimated by examining
selected species for pathological conditions and by assessing
reproductive impairment in suitably mature female fish.
Pathological conditions will include grossly visible abnormalities
(e.g., fin erosion) and other lesions diagnosed by histological
procedures (e.g-, gill necrosis, liver cell necrosis).
Reproductive capacity will be estimated by examining the
developmental stages of ovaries and by measuring plasma levels of
certain reproductive hormones (Johnson et al., 1988), in addition
to measuring fecundity (Cross and Hose, 1988). The two primary
species for assessing reproductive impairment are Dolly Varden char
and yellowfin sole. It is anticipated that, during the proposed
sampling period (May/June), these two species will be at an
appropriate stage in their reproductive cycle for such assessments
to be done. Concurrent laboratory studies will be conducted to
determine the effects of known doses of oil and oil components on
reproductive processes in these or related species.

Samples of sediment, and selected stomach contents of fish (from
fish whose bile had evidence of oil exposure) will be analyzed
(sediment under A/W Study 2) for hydrocarbons by recently
developed, scientifically sound and cost-effective analytical
procedures involving high-performance liquid chromatography, gas
chromatography and mass spectroscopy (Krahn et al., 1988).

Environmental injury will be determined using statistical and
simulation models, which will be developed as part of these

178

proposed studies, as well as from other investigations with related
fish species. The bile and tissue chemistry data will be used to
establish relationships between biological damage and estimated
exposures to petroleum hydrocarbons.

Sampling Methods

sampling activities will be conducted at several sites in PWS and
the GOA. Sample collection will be performed from a NOAA vessel
(and its launches) at water depths of approximately 0 to 320
meters. At each site, sediment samples will be collected with a
box corer, VanVeen or Smith-McIntyre grab. Sediments will be
stored at - 200 C. The coordinates and depths of each station will
be recorded.

Fish will be collected with a bottom trawl, long-line gear, gill
nets! or beach seines. Bottom trawls will be performed with an
otter trawl. Tows will be of 5 to 15 minutes duration. In order
to reduce contamination of the catch by free oil, trawling will
avoid areas of surface films or slicks. Individuals of selected
target fish species will be sorted and examined for externally
visible lesions; up to 30 fish of selected species will be
measured, weighed, and necropsied; and tissue samples will be
excised and preserved in fixative for histopathological examination
or frozen for chemical analyses.

Laboratory Analyses

1. Bile Metabolite Assay (analyses done under Technical Services-
il

Samples of bile will be injected directly into a liquid
chromatograph and a gradient elution conducted (Krahn et al., 1984,
1986a, b, c) . Two fluorescence detectors are used in series. The
excitation/ emission wavelengths of one detector are set to 290/335
nm, where metabolites of naphthalene (NPH) fluoresce.
Excitation/emission wavelengths of the other detector are set to
260/380 nm, where metabolites of phenanthrene (PHN) fluoresce. The
total integrated area for each detector is then converted
(normalized) to units of either NPH or PHN that would be necessary
to give that into-grated area.

2. Liver Aryl Hydrocarbon Hydroxylase (AHs) Activity and Cytochrome
P-4501AI Analysis

Hepatic microsomes will be prepared essentially as described by
Collier et al. (1986) and microsomal protein will be measured by
the method of Lowry et al. (1951), using bovine serum albumin as
the standard. AHs activity will be assayed by a modification of
the method of Van Cantfort et al. (1977) as described by Collier et
al. (1986), using 14C-labeled benzo(a]pyrene as the primary
substrate. All enzyme assays will be run under conditions in which

179

the reaction rates are in the linear range for both time and
protein. Cytochrome P-4501Al will be measured by an ELISA
utilizing rabbit antibodies to cytochrome P-450c isolated from
Atlantic cod (Goksoyr,1990).

3. Histopathology

Histopathological procedures to be followed are described in the
report from the Histopathology Technical Group for Oil Spill
Assessment Studies in PWS, Alaska. Briefly, the procedures will
involve the following: (a) tissues preserved in the field will be
routinely embedded in paraffin and sectioned at five microns
(Preece, 1972) ; and (b) paraffin sections will be routinely stained
with Mayer's hematoxylin and eosin, and for further
characterization of specific lesions, additional sections will be
stained using standard special staining methods (Thompson, 1966;
Preece, 1972; and Armed forces Institute of Pathology, 1968). All
slides will be examined microscopically without knowledge of where
the fish were captured. Hepatic lesions will be classified
according to the previously described diagnostic criteria of Myers
et al. (1987). Ovarian lesions will be classified as described in
Johnson et al. (1988).

4. Reproductive Indicators

Reproductive activity will be assessed by examining the ovaries of
the sampled fish histologically to determine their developmental
stage, and for the presence of ovarian lesions that would be
indicative of oocyte resorption (Johnson et al., 1988). other
parameters associated with reproductive activity will also be
measured, including fecundity (Bagenal and Braum, 1971), plasma
vitellogenin (Gamst and Try, 1980; DeVlaming et al., 1984) and
estradiol (Sower and Schreck, 1982) levels, and gonadosomatic index
(ovary wt/gutted body wt x 100) . Relationships between ovarian
maturation, fecundity, plasma estradiol, plasma vitellogenin, and
petroleum hydrocarbon exposure will then be evaluated.

DATA ANALYSIS

Where possible, non-parametric statistical tests will be employed
to avoid assumptions that the data are normally distributed. The
principal non-parametric tests that will be used are Spearman rank
correlation, which has about 0.91% of the power of product-moment
correlation when the parametric assumptions are met (Zar, 1984),
and the heterogeneity-G statistic. Spearman rank correlation will
be used for estimating uptake and metabolism of petroleum
hydrocarbons from oiled and non-oiled habitats when an independent
measure of contamination (e.g. , levels of Ahs in sediment) is
available.

The heterogeneity-G statistic (Sokal and Rohlf, 1981) will be used
to study prevalence of pathological conditions at oiled and non-

180

oiled habitats. In addition, logistic regression (appropriate
where the outcome variable is binomial) will be used to model the
prevalences of pathological conditions in relation to
contamination.

The Kruskal-Wallis test (a non-parametric form of ANOVA) will be
used for supporting statistical analyses of variation in sediment
and fish hydrocarbon levels at sites sampled. If the null
hypothesis of no differences among sites is rejected at a = 0.05,
a non-parametric multiple comparison test (Dunn, 1964; Hollander
and Wolfe, 1973; Zar, 1984) will be used to determine differences
between sites at a = 0. 05. Principal components analysis and
LOWESS (Chambers et al., 1983) will also be employed for this
purpose; both are methods of exploratory data analysis rather than
inferential statistical methods. Cohen (1977) will be used for
computations of statistical power.

Products will include information on the distribution and
concentrations of petroleum hydrocarbons and their metabolites in
fish tissues and in sediments obtained from sites in Alaska; the
hepatic activities of AHs and levels of cytochrome P-45OIAl in fish
from sites in Alaska; and the distribution and prevalence of
histopathological disorders and reproductive impairment among
selected species from those sites. Chemistry data will be
submitted in the f orm of data tables and distribution maps, and all
data will be stored in computerized data management programs. Fish
pathology data will be reported in the form of distribution maps,
tables describing disease frequencies of each species examined,
photographs of gross and microscopic properties of abnormalities,
figures representing various types of biological data (e.g.,
length-weight, age-weight) and discussions of the relative
importance of the types of abnormalities found. comparisons of the
characteristics of these abnormalities will be made with similar
conditions previously reported in other marine areas of the world.
The data management formats were designed in cooperation with the
National Oceanographic Data Center.

BIBLIOGRAPHY

Armed Forces Institute of Pathology. 1968. Manual of Histologic
Staining Methods. Third Edition (L.G. Luna, ed.) McGraw-
Hill, New York, 258 p.

Bagenal, T.B. and E. Braum. 1971. Eggs and early life history.
In: Methods for the assessment of fish production in fresh
waters. International Biology Programme Handbook 3. (W.E.
Ricker, Ed.) Blackwell Sci. Pub. Oxford and Edinborough,
England, p. 165-198.

Black, J.J. 1983. Field and laboratory studies of environmental
carcinogenesis in Niagara River Fish. J. Great Lakes Res.
9:326-334.

181

Buhler, D.R. and D.E. Williams. 1989. Enzymes involved in
metabolism of PAH by f ishes and other aquatic animals:
oxidative enzymes (or Phase I enzymes). In: Metabolism of
polycyclic aromatic hydrocarbons in the aquatic
environment. (U. Varanasi, Ed.) CRC Press, Inc., Boca
Raton, FL, p. 151-184.

Casillas, E., D. Misitano, L.J. Johnson, L.D. Rhodes, T.K. Collier,
J.E. Stein, B.B. McCain, and U. Varanasi. 1990.
Inducibility of spawning and reproductive success of female
English sole (Parophrys vetulus) from urban and nonurban
areas. Submitted to Mar. Environ. Res.

Chambers, J. M., W. S. Cleveland, B. Kleiner, and P. A. Tukey.
1983. Graphical methods for data analysis. Belmont, CA:
Wadsworth International Group. 395 p.

Cohen, Jacob. 1977. Statistical power analysis for the
behaviorial sciences. New York: Academic Press. 474 pp.

Collier, T.K., J.E. Stein, R.J. Wallace, and U. Varanasi. 1986.
Xenobiotic metabolizing enzymes in spawning English sole
(ParoRhrys vetulus) exposed to organic-solvent extracts of
sediments from contaminated and reference areas. Comp.
Biochem. and Physiol. 84C:291-298.

Collier, T.K. and U. Varanasi. 1987. Biochemical indicators of
contaminant exposure in flatfish from Puget Sound, Wa. pp
1544-1549. In: Proceedings Oceans 187 IEEE, Washington,
D.C.

Collier, T.K., B.-T. L. Eberhart, and A. Goksoyr. 1989.
Immunochemical quantitation of cytochrome P450 IA1 in
benthic fish from coastal U.S. waters. Proc. Pac. NW Assoc.
Toxicol. 6:9. (Abstract).

Cross, J.N. and J. Hose. 1988. Evidence for impaired reproduction
in white croaker (Genyonemus lineatus) from contaminated
areas off Southern California. Mar. Environ. Res.
24:185-188.

DeVlaming, V., R. Fitzgerald, G. Delahunty, J. J. Cech, Jr., K.
Selman, and M. Barkley. 1984. Dynamics of oocyte
development and related changes in serum estradiol 17-8,
yolk precursor, and lipid levels in the teleostean fish,
Leptocottus armatus. Comp. Biochem. Physiol. 77A:599-
610.

Dunn, 0. J. 1964. Multiple contrasts using rank sums.
Technometrics 6: 241-252.

Gamst, 0. and K. Try. 1980. Determination of serum-phosphate

182

without deproteinization by ultraviolet spectrophotometry
of the phosphomolybdic acid complex. Scand. J. Clin. Lab
Invest. 40:483-486.

Goksoyr, A. 1990. An ELISA for monitoring induction of cytochrome
P-4501Al in fish liver samples. manuscript in
preparation.

Gulyas, B.J. and D.R. Mattison. 1979. Degeneration of mouse
oocytes in response to polycyclic aromatic hydrocarbons.
Anat. Rec. 193:863-869.

Hollander, M. and D. A. Wolfe. 1973. Nonparametric statistical
methods. New York: John Wiley. 503 p.

Johnson, L.J., E. Casillas, T.K. Collier, B.B. McCain, and U.
Varanasi. 1988. Contaminant effects on ovarian
development in English sole (Parophrys vetulus) from
Puget Sound, Washington. Can. J. Fish. Aquat Sci.
45:2133-2146.

Krahn, M.M., M.S. Myers, D.G. Burrows, and D.C. Malins. 1984.
Determination of metabolites of xenobiotics in bile of
fish from polluted waterways. Xenobiotica. 14:633-646.

Krahn, M.M., L.J. Kittle, Jr., and W.D. MacLeod, Jr. 1986a.
Evidence for oil spilled into the Columbia River. Mar.
Environ. Res. 20:291-298.

Krahn, M.M., L.D. Rhodes, M.S. Myers, L.K. Moore, W.D. MacLeod,
Jr., and D.C. Malins. 1986b. Associations between
metabolites of aromatic compounds in bile and occurrence
of hepatic lesions in English sole (Parophrys vetulus)
from Puget Sound, Washington. Arch. Environ. Contam.
Toxicol. 15:61-67.

Krahn, M.M., L.K. Moore, and W.D. MacLeod, Jr. 1986c. Standard
Analytical Procedures of the NOAA National Analytical
Facility, 1986: Metabolites of Aromatic Compounds in Fish
Bile. Technical Memorandum NMFS/F/NWC-102, 25 pp.
(Available from the National Technical Information
Service of the U.S. Department of Commerce, 5285 Port
Royal Road, Springfield, VA 22161).

Krahn, M.M., C.A. Wigren, R.W. Pierce, L.K. Moore, R.G. Bogar,
W.D. MacLeod, Jr., S.-L. Chan, and D.W. Brown. 1988.
Standard Analytical Procedures of the NOAA National
Analytical Facility, 1988: New HPLC Cleanup and Revised
Extraction Procedures for Organic Contaminants.
Technical Memorandum NMFS/F/NWC-153, 52 pp. (Available
from the National Technical Information Service of the
U.S. Department of Commerce, 5285 Port Royal Road,

183

Springfield, VA 22161).

Lowry, O.H., N.J. Rosebrough, A.L. Farr, and R.J. Randall. 1951.
Protein measurement with the Folin phenol reagent, J.
Biol. Chem. 193:265-275.

Lutz, W.K. 1979. In vivo covalent binding of organic chemicals
to DNA as a quantitative indicator in the process of
chemical carcinogenesis. Mutat. Res. 65:289-356.

MacLeod, W.D., Jr., D.W. Brown, A.J. Friedman, D.G. Burrows, 0.
Maynes, R.W. Pearce, C.A. Wigren, and R.G. Bogar. 1985.
Standard Analytical Procedures of the NOAA National
Analytical Facility, 1985-1986: Extractable Toxic
Organic Compounds, 2nd Ed. NOAA Technical Memorandum
NMFS F/NWC-92, 121 pp. (Available from the National
Technical Information Service of the U.S. Department of
Commerce, 5285 Port Royal Rd., Springfield, VA 22161;
PB86-147873).

Malins, D.C., B.B. McCain, D.W. Brown, S-L. Chan, M.S. Myers,
J.T. Landahl, P.G. Prohaska, A.J. Friedman, L.D. Rhodes,
D.G. Burrows, W.D. Gronlund, and H.O. Hodgins. 1984.
Chemical pollutants in sediments and diseases in bottom-
dwelling fish in Puget Sound, Washington. Environ. Sci.
Technol. 18:705-713.

Mattison, D.R. and M.S. Nightingale. 1980. The biochemical and
genetic characteristics of murine ovarian aryl
hydrocarbon (benzo(a]pyrene) hydroxylase activity and its
relationship to primordal oocyte destruction by
polycyclic aromatic hydrocarbons..Toxicol. appl.
Pharmacol. 56:399-408.

McCain, B.B., H.O. Hodgins, W.D. Gronlund, J.W. Hawkes, D.W.
Brown, M.S. Myers, and J.H. Vandermeulen. 1978.
Bioavailability of crude oil from experimentally oiled
sediments to English sole (Parophrys vetulus) and
pathological consequences. J. Fish. Res. Board Can.
35:657-664.

Myers, M.S., L.D. Rhodes, and B.B. McCain. 1987. Pathologic
anatomy and patterns of occurrence of hepatic neoplasms,
putative preneoplastic lesions and other idiopathic
hepatic conditions in English sole (Parophrys vetulus)
from Puget Sound, Washington, U.S.A. J. Natl. Cancer
Inst. 78:333-363.

National Academy of Sciences. 1985. Oil in the Sea; Inputs,
fates and effects. National Academic Press, Washington,
D. C. 601pp.

184

Preece, A. 1972. A Manual for Histologic Technicians. 3rd
edition. Little, Brown and Co., Boston, 428 pp.

Schaaf, W.E., D.S. Peters, D.S. Vaughan, L. Coston-Clements, and
C.W. Krouse. 1987. Fish population responses to chronic
and acute pollution: the influence of life history
strategies. Estuaries 10: 267-275.

Shum, S., N.M. Jensen, and D.W. Nebert. 1979. The murine Ah locus:
in utero toxicity and teratogenesis associated with genetic
differences in benzo[a]pyrene metabolism. Teratology
20:365-376.

Sokal, R. and F. Rohlf. 1981. Biometry. (Second Ed.) W.H.
Freeman and Co.: San Francisco, CA, 859 pp.

Sower, S. A. and C. B. Schreck. 1982. Steroid and thyroid
hormones during sexual maturation of coho salmon
(Oncorhynchus kisutch) in seawater or freshwater. Gen.
Comp Endocrine. 47:42-53.

Spies, R.B. and D.W. Rice, Jr. 1988. Effects of organic
contaminants on reproduction of the starry flounder
Platichthys stellatus in San Francisco Bay. II.
Reproductive success of fish captured in San Francisco
Bay and spawned in the laboratory. Mar. Biol. 98:191-200.

Stein, J.E., T. Hom, H.R. Sanborn, and U. Varanasi. 1990. Effects
of exposure to a contaminated-sediment extract on the
metabolism and disposition of 173-estradiol in English
sole (Parophrys vetulus). Manuscript in preparation.

Van Cantfort, J., J De Graeve, and J.E. Gielen. 1977. Radioactive
assay for aryl hydrocarbon hydroxylase. Improved method
and biological importance. Biochem. Biophys. Res. Commun.
79:505-511.

Varanasi, T.K. Collier, D.E. Williams, and D.R. Buhler. 1986.
Hepatic cytochrome P-450 isozymes and aryl hydrocarbon
hydroxylase in English sole (Parophrys vetulus).
Biochem. Pharmacol. 35:2967-2971.

Varanasi, U., D.W. Brown, S-L. Chan, J.T. Landahl, B.B. McCain,
M.S. Myers, M.H. Schiewe, J.E. Stein, and D.D. Weber.
1987. Etiology of tumors in bottom-dwelling marine fish.
Final Report to the National Cancer Institute under
Interagency Agreement Y01 CP 40507.

Varanasi, U., S-L. Chan, R.C. Clark, Jr., T.K. Collier, W.D.
Gronlund, M.M. Krahn, J.T. Landahl, and J.E. Stein. 1990.
Oil Spill Progress Report. Shellfish and Groundfish Trawl
Assessment Outside Prince William Sound. 30 p.

185

Zar, J.H. 1984. Biostatistical Analysis. Prentice-Hall:
Eaglewood Cliffs, NJ, 620 pp.

BUDGET: NOAA

Salaries $230.0
Supplies 35.4
Travel 19.7
Equipment (disposable) 14.9
Vessel support 150.0

$450.0

186

FISH/SHELLFISH STUDY NUMBER 27

Study Title: Sockeye Salmon Overescapement

Lead Agency: ADF&G

INTRODUCTION

commercial fishing for sockeye salmon in 1989, was curtailed in
upper CI, the outer Chignik districts, and the Kodiak areas due to
presence of oil in the fishing areas from the EVOS. As a result,
the number of sockeye salmon entering four important sockeye
producing systems (Kenai/Skilak, Chignik/Black, Red, and Frazer
Lakes) and two less important lake systems (Akalura and Af ognak or
Litnik lakes) greatly exceeded levels that are thought to be most
productive. Sockeye salmon spawn in lake associated river systems.
Adult salmon serve an extremely important role in the ecosystem,
providing food for marine mammals, terrestrial mammals, and birds.
Additionally, carcass decomposition serves to charge f resh water
lake systems with important nutrients. Juvenile salmon which rear
in lakes for one or two years serve as a food source for a variety
of fish and mammals. Sockeye salmon are also an important
subsistence, sport, and commercial species. The ex-vessel value of
the commercial catch of sockeye from these lake systems has
averaged about $42 million per year since 1979, with the 1988 catch
worth $115 million. Sockeye salmon returns to the Kenai River
system support some of the largest recreational fisheries in the
State.

Overly large spawning escapements may result in poor returns by
producing more rearing juvenile sockeye than can be supported by
the nursery lake's productivity (Kyle et al. 1988). In general,
when rearing f ish abundance greatly exceeds the lake I s carrying
capacity, prey resources are altered by changes in species and size
composition (Mills and Schiavone 1982, Koenings and Burkett 1987,
Kyle et al. 1988) with concomitant effects on all trophic levels
(Carpenter et al. 1985). Because of such changes, juvenile sockeye
growth is reduced, mortality increases, larger percentages holdover
for another year of rearing; and the poor quality of smolts
increases marine mortality. Where escapements are two to three
times normal levels, the resulting high juvenile densities crop the
prey resources to the extent that more than one year is required to
return to normal productivity. Rearing juveniles from subsequent
brood-years suffer from both the poor quality of forage and from
the increased competition for food by holdover juveniles (Townsend
1989). This is the brood-year interaction underlying cyclic
variation in the year class strength of anadromous fish.

This project will examine the effects of large 1989 spawning
escapements on the resulting progeny for a select subset of the

187

above mentioned sockeye nursery lakes. Three impacted lake systems
where the 1989 escapements were more than twice the desired levels
(Kenai/Skilak in upper CI; Red and Akalura lakes on Kodiak Island)
were selected. Upper Station Lake which is near the two impacted
lakes on Kodiak did not receive a large escapement and will be
examined as a control.

This study is necessary to obtain a more timely assessment of
impact as adult sockeye, produced from the 1989 escapement, will
not return until the 1994/1995 season. Further, total return data
are not available for individual Kodiak sockeye systems due to the
complex mixed-stock nature of the commercial fisheries and the
inability to estimate stock-specific catches.

OBJECTIVES

A. Estimate the number, age, and size of sockeye
salmon juveniles rearing in selected freshwater
systems.

B. Estimate the number, age, and size of sockeye
salmon smolts migrating from selected freshwater
systems.

C. Determine ef f ects of large escapements resulting
from fishery closures caused by the EVOS on the
rearing capacity of selected nursery lakes
through:

a. analysis of age and growth of juveniles and smolts
b. examination of nursery area nutrient budgets and
plankton populations.

METHODS

Numbers of adult sockeye salmon that entered selected spawning
systems outside PWS prior to and during 1989 have been estimated at
weir stations or by sonar. This information was collected during
projects routinely conducted by the ADF&G as part of their resource
management program. Optimal escapement levels, which on the
average should produce maximum sustained yield, have been based on
either past relationships between spawners and returning progeny or
the extent of available spawning and rearing habitat. The baseline
program will continue at each site including but not limited to
estimates of adult sockeye escapement and collection of scales for
age analysis.

For each of the 4 lake systems identified, the response (abundance,
growth, and freshwater age) of rearing juveniles from the 1989
escapement will be studied through its likely period of freshwater

188

residence, early summer 1990 to spring 1992.

The total number of juvenile sockeye in each lake will be estimated
through hydroacoustic surveys conducted during the summer (late
June) and fall (September-October) of 1990, 1991, and 1992. Age
and size information as well as diet items will be obtained from
samples of juvenile sockeye collected from concurrent mid-water
trawl netting surveys. Survey transect designs for hydroacoustic
sampling and tow-netting have been established for Kenai and Skilak
lakes (Tarbox and King 1989), and will be developed for each
additional lake in the study. The basic survey design will be a
stratified random sample where each lake is subdivided into areas
and survey transects randomly selected in each area. Such
programs, funded through other studies, are already in place for
Tustumena and Afognak lakes. Depending on densities of rearing
juvenile sockeye, estimates of fish densities will be made for each
transect either by echo integration or by echo counting. Total
f ish population estimates will be computed, by summing transect
populations, along with 95% confidence intervals (Kyle 1989).

Freshwater growth and age of sockeyp salmon rearing juveniles from
all study systems will be determined from scale and otolith
measurements made either by direct visual analysis of scales or on
an Optical Pattern Recognition system. In cases where data are
available (e.g., Kenai and Skilak Lakes) , growth of progeny from
the 1989 spawning escapements will be compared with growth (size)
of progeny produced from spawnings within these systems during
prior years.

The total number of smolt migrating from each system will be
estimated with a mark-recapture study during 1990, 1991 and 1992
using inclined plane traps after Kyle (1983), and Tarbox and King
(1989). Smolt will be captured in traps, sampled for age and size
information, marked with Bismark Brown Y (a biological dye), and
transported upstream of -the traps and released for subsequent
recapture (Rawson 1984). Periodic retesting will determine the
capture efficiency of the traps under changing river conditions
during the spring. Total population estimates (with 95% confidence
intervals) will be made using catch efficiencies, and weekly number
weighted smolt size and age information will be calculated using a
computer spreadsheet developed by Rawson (personnel communication,
1985). Size and ages of sockeye smolts from the 1989 spawning
escapements will be compared with smolt information from spawnings
within these systems during prior years. Finally, smolt programs
consistent to those for the study lakes are planned, under separate
funding, for Tustumena and Afognak Lakes.

Limnological studies will monitor the response of the lakes to the
high juvenile rearing densities and to estimate the carrying
capacity parameters of euphotic volume, nutrient budgets (carcass
enrichment) , and zooplankton biomass, body-sizes, and population
shifts. Approximately six limnology surveys will be conducted at

189

two stations, during 1990, 1991, and 1992, to determine zooplankton
species abundance and body-sizes, nutrient chemistry, and
phytoplankton abundance for Kenai/Skilak, Red, Akalura, and Upper
Station lakes. Carrying capacity parameters exist for Afognak and
Tustumena lakes based on ongoing studies by FRED and Commercial
Fish Divisions.

In cases where seasonal data are available (e.g., Akalura, Kenai,
and Skilak lakes), limnological parameters taken during residence
of the juveniles from the 1989 spawning escapements will be
compared to parameters within these systems during prior years. In
addition, randomized intervention analysis (RIA) will be used to
detect changes in the systems with large escapements relative to a
control or reference system (Carpenter et al. 1989).

In addition to RIA, the holistic approach proposed here involves
several evaluation procedures to assess the effects of sockeye
salmon overescapement.

First, fresh-water production from the 1989 escapements will be
assessed in Kenai/Skilak, Red, Akalura, and Upper Station lakes.
This will be accomplished through analysis of growth, freshwater
survival (in particular over winter survival), and freshwater age
of sockeye smolt populations. Also planktonic food sources will be
assessed through estimation of abundance of zooplankton prey
biomass and numbers of species. Any anomalies will be determined
by analysis of freshwater growth recorded on archived scales,
historical freshwater age composition, and modelled freshwater
survivals; and from results of previous studies as well as the 1990
smolt characteristics from each of the study systems.

Second, future sockeye salmon production from the 1989 parent year
and subsequent parent years will be estimated based on
spawner/recruit relationships incorporating a brood-year
interaction term. Losses of adult sockeye production from
subsequent parent years may result from negative effects of progeny
of the 1989 escapement on the lake's carrying capacity and/or from
continued high escapements due to the inability to harvest the runs
because of oil in the fishing area. The spawner/recruit
relationships will be estimated from historical stock specific
return data (where available) , and generalized spawner/recruit data
scaled to the carrying capacity parameters (i.e., euphotic volume
and zooplankton biomass) of the nursery lakes where stock specific
return data are not available (Geiger and Koenings 1990).

Third, experimental and empirical sockeye life history/production
models (Koenings and Burkett 1987, Koenings et al 1989) will be
used to compare salmon production by life-stage at escapement
levels consistent with management goals to the 1989 escapements.

190

BIBLIOGRAPHY

Carpenter, S. R., J. F. Kitchell, and J. R. Hodgson. 1985.
Cascading trophic interactions and lake productivity.
BioScience 35:634-639.

Carpenter, S. R., T M. Frost, D. Heisey, and T. K. Kratz. 1989.
Randomized intervention analysis and the interpretation of
whole-ecosystem experiments. Ecology 70:1142-1152.

Geiger, H. J., and J. P. Koenings. 1990. Escapement goals for
sockeye salmon with informative prior probabilities based on
habitat considerations. Journal of Fish Biology (in review) .

Koenings, J. P., and R. D. Burkett. 1987. Population
characteristics of sockeye salmon (Oncorhynchus nerka) smolts
relative to temperature regimes, euphotic volume, fry
density, and forage base within Alaskan Lakes. p. 216-234. In
H. D. Smith, L. Margolis, and C. C. Wood [ed. ] Sockeye salmon
(Oncorhynchus nerka) population biology and future
management. Can. Spec. Publ. Fish. Aquat. Sci. 96.

Koenings, J. P., J. E. Edmundson, G. B. Kyle, and J. M. Edmundson.
1987. Limnology field and laboratory manual: methods for
assessing aquatic production. Alaska Department of Fish and
Game, FRED Division Report Series No. 71:212 p.

Koenings, J. P., R. D. Burkett, M. Haddix, G. B. Kyle, and D. L.
Barto. 1989. Experimental manipulation of lakes for sockeye
salmon (Oncorhynchus ngrka) rehabilitation and enhancement.
Alaska Department of Fish and Game, FRED Division Report
Series No. 96: 18p.

Kyle, G. B. 1983. Cresent Lake sockeye salmon smolt enumeration
and sampling, '1982. Alaska Department of Fish and Game, FRED
Division Report Series No. 17:24 p.

Kyle, G. B. 1989. Summary of acoustically-derived population
estimates and distributions of juvenile sockeye salmon
(Oncorhynchus nerka) in 17 nursery lakes of Southcentral
Alaska. Alaska Department of Fish and Game, FRED Division
Report Series No. (In review).

Kyle, G. B., J. P. Koenings, and B. M. Barrett. 1988. Density
-dependent, trophic level responses to an introduced run of
sockeye salmon (Oncorhynchus nerka) at Frazer Lake, Kodiak
Island, Alaska. Canadian Journal of Fisheries and Aquatic
Sciences 45:856-867.

191

Mills, E. L., and A. Schiavone, Jr. 1982. Evaluation of fish
communities through trophic assessment of zooplankton
populations and measures of lake productivity. North
American Journal of Fisheries Management 2:14-27.

Rawson, Kit. 1984. An estimate of the size of a migrating
population of juvenile salmon using an index of trap
efficiency obtained by dye marking. Alaska Department of
Fish and Game, FRED Division Report Series No. 28:23 p.

Tarbox, K.E., and B.E. King. 1989. An estimate of
juvenile fish densities in Skilak and Kenai Lakes, Alaska
through the use of dual beam hydroacoustic techniques in
1989. Alaska Department of Fish and Game, Commercial Fish
Division Regional Information Report No. 2S90-1.

Townsend, C.R. 1989. Population cycles in freshwater fish.
Journal of Fish Biology 35(Supplement A):125-131.

BUDGET: ADF&G

Personnel Services $168.2
Travel 4.9
Contractual 88.7
Supplies 52.7
Equipment 77.5

Total $392.0

192

FISH/SHELLFISH STUDY NUMBER 28

Study Title: Salmon Oil Spill Injury Model and Run
Reconstruction

Lead Agency: ADF&G

INTRODUCTION

There are at least two approaches to the determination of damage
to PWS fishery resources. The first approach is the "bottom up"
view. Here one begins with escapements in the oiled and unoiled
areas and projects return based on the various life history
parameters observed for the two adult return areas. The damages
are the dif f erence in adult production between the oiled and
unoiled areas. The second approach is the "top down" view. Here
one begins with the returns (district catches plus escapements)
and estimates the stock specific return and return per specimen
for oiled and unoiled areas based on a reconstruction of the run.
Damages are defined as lost production of adults and are
estimated from the differences in return per spawner applied to
the parent escapements.

This study will estimate damages to PWS fishery resources based
on both the "bottom up" (i.e. life history modeling) and "top
down" (i.e. run reconstruction) views.

In the life history modeling approach, it is necessary to add
together the factors at various life history stages over several
individual river systems and salmon species. Examples of
potential factors include: reduced growth of fry, increased
mortality of eggs and fry, loss of spawning habitat, increased
early marine mortality, and overescapement. The magnitude of the
overall loss in productivity for an individual salmon stock can
best be understood by considering the survival at each life
history stage (egg, fry, smolt, subadult, adult) over the life
span of all fish of the same age class, and over all age classes
present in the population. Note that survival of future age
classes must be considered if the detrimental effects of oiling
are persistent. A bookkeeping program is necessary to take
advantage of the data already being collected, and to integrate
existing historical data into documentation of the actual and
potential damages due to oiling.

In the "top down" view, it is necessary to estimate the stock
specific return (i.e. catch plus escapement) so that return per
spawner in oiled and unoiled areas can be estimated. Because the
fisheries in PWS harvest mixed stocks of fish, it will be
necessary to reconstruct the stock specific abundance over time

193

in each of the fishing districts to estimate stock specific
catches.

OBJECTIVES

Life History Modeling.

A. Develop a computational framework to account for specific
effects of oiling on species, stock, and life history stages
of salmon populations in PWS, Cook Inlet, Kodiak, and the
Chignik areas.

B. Estimate the "status quo or no-oiling" values f or all
parameters implicit in the computational framework that are
most consistent with the scientific literature and give the
best description of the aggregate of stocks' historical
population dynamics.

C. Estimate the "oiling" values for all parameters implicit in
the computational framework that are most consistent with
the synthesis of the individual stocks responses as
identified in the NRDA studies and/or responses deduced from
the available scientific literature.

D. Develop estimates of salmon injury by comparison of future
simulations of salmon production under the "oiling" and "no-
oiling" model parameter values.

Run Reconstruction.

A. Develop a computational framework for estimating stock
specific abundance over time in the 8 fishing districts in
Prince William Sound. The approach will be a two
dimensional (multi-stock and multi-district) generalization
of the comprehensive timing model of Schnute and Sibert
(1983).

B. Analyze the historical timing data and tagging data
necessary to develop simplifying assumptions to derive
estimable parameters. Test the run reconstruction approach
by reconstructing the 1988 pink salmon run, develop
estimates of hatchery contribution, and compare those to the
hatchery contributions observed in the coded-wire tag (CWT)
study.

C. Reconstruct the 1990 pink salmon run to Prince William Sound
and develop estimates of return per spawner for oiled and
unoiled areas.

194

METHODS

The life history model and run reconstruction model will be
developed by a select group of experts under contract to ADF&G.

The life history model will have the following properties:

1. The salmon stocks and areas included in this computer
based mathematical model are those included in the
portions of F/S studies 1-10 as approved by the
trustees as well as any stocks that were observed to
have suffered overescapement in 1989 as a result of the
presence of oil.

2. The model will have sufficient temporal or life history
structure to account for all of the potential oil
related injuries that might affect the future
production of salmon.

3. The model will have stochastic elements to account for
natural variation.

4. The model will project future abundance of salmon by
individual stocks and will enable the comparison of
future scenarios of salmon abundance with and without
oiling.

With regard to the run reconstruction model, the following
relates the multi-stock and multi-district generalization of
Schnute and Sibert (1990) for PWS Pink Salmon. There are eight
fishing districts in Prince William Sound and eight stocks
consisting of the aggregate of spawning streams within the
respective fishing district.

Define the following:

ii(t) = cumulative entry to District j
Ej(t) = cumulative escapement in District j
Xij(t) cumulative entry to District j from District i
Pj(t) total number of fish in District j
Cj(t) cumulative catch in District j
Cij(t) cumulative catch of stock i in District j

8
Cj(t) E Ci. (1)
J=i

195

P j (t) I j (t) + E Xkj M - Cj(t) - E Xjt(t) - Ej(t) (2)
k 1

The timing functions I., Xi1j, and P. will be estimated by
reconstructing back from the cumulative catches (C,(t)) and
cumulative escapements (E,(t)). Several assumptions must be made
in order to solve the above generalization of Schnute and
Sibert's model. To do so requires extensive analysis of
historical timing and tagging information. Examples of
assumptions are: 1. Entry of pink salmon into Prince William
Sound occurs in Districts 6, 7, 8 (i.e. Southwestern, Montague,
and Southeastern, respectively). 2. Rate of exploitation is the
same for all stocks within a given District.

DISCUSSION

Note that the life history and run reconstruction models will
accommodate harvest in existing mixed stocks fisheries and will
enable the comparison of alternative commercial fisheries harvest
policies. This will facilitate the evaluation of fisheries
restoration strategies that attempt to rebuild damaged stocks by
reducing catch in fisheries that exploit stocks damaged and
stocks not damaged by the oil spill.

Activities during the first year include: analyses of historical
data, developing efficient software for computer simulation, and
synthesis of model parameters for the no-oiling scenario.
Activities during the second year include analysis of NRDA
results, including run reconstruction, to developing the best
estimate of the various salmon stocks response to oiling;
synthesis of these results into altered parameters in the model;
develop the best scientific. estimates of future salmon stocks
production under the no-oiling and oiling scenarios.

BIBLIOGRAPHY

Schnute, J and J. Sibert. 1983. The salmon terminal fishery: a
practical, comprehensive timing model. Can. J. Fish. Aquat.
Sci. 40:835-853.

BUDGET: ADF&G

Personnel $58.9
Travel 5.2
Contractual 100.0
Supplies 1.0
Equipment 10.0

Total $175.1

196

FISH/SHELLFISH STUDY NUMBER 30

Study Title: Data Base Management

Lead Agency: ADF&G

INTRODUCTION

Large quantities of data are being analyzed in order to demonstrate
the fact and extent of injury to natural resources due to oiling.
The purpose of this study is to make original data readily
available to agency and non-agency personnel so that data analysis
can be conducted, and so that all analyses can be accomplished in
an efficient and cost effective manner. The data to be placed
under the database management system (DBMS) will be drawn from two
categories; 1) historical data necessary to the interpretation and
implementation of the results of NRDA studies, and 2) data
resulting from NRDA studies.

OBJECTIVES

A. To construct a cost effective DBMS to readily retrieve and
order data from original data in electronic form according to
user specified criteria of time, space, and selection of
variables. The DBMS should be constructed to meet the
following criteria, in order of priority: 1) completeness of
contentst 2) speed of retrieval, and 3) ease of use in
assembling primary data into data sets for further analysis by
other software. Furthermore, the DBMS will take advantage of
existing DBMS applications currently available in the ADF&G.

B. To develop the structural facilities for individuals to access
data that is physically located at different sites. To
accomplish this, a LAN (local area network) facility must be
developed in the Cordova and Anchorage ADF&G offices, as well
as to develop a system for linking these with existing LAN's
in Juneau and Kodiak. Note that objective B, although a
necessity for. this project, will be met by a concurrent and
separately funded project "statewide data base system"
currently being implemented by ADF&G.

METHODS

A distributional data base management system, using SQL software,
will be developed. The system will be flexible to accommodate the
data physically located in Kodiak, Anchorage, Cordova, and Juneau.

The DBMS system will be accessed through a linked system of LAN's,
(Juneau, Anchorage, Kodiak, and Cordova) . The DBMS can be accessed
by any user with an IBM compatible PC that has access to the
Anchorage LAN. Interface software using "WINDOWS" will be

197

developed and made available to individuals to facilitate non-
programmer access to the DBMS systems.

The following data, for all species and from Prince William Sound,
Cook Inlet, Kodiak, and Chignik areas, will be incorporated into
the DBMS:

1. All NRDA project data.
2. Salmon escapement data, including weir counts, stream
counts, aerial survey counts, and sonar counts.
3. Biological data including age composition, size, sex,
growth, and stock composition.
4. Pre-emergent and egg density.
5. Groundfish and shellfish survey data.

In addition, the DBMS will have access to statewide fish ticket
system data which includes commercial fisheries catch and effort
data by area, species, and gear type.

This project will be developed concurrently with the development of
the ADF&G statewide data base system which is being funded with
State of Alaska general funds. It is the intent to develop LANS in
the Anchorage and Cordova ADF&G offices. These new LAN's will be
linked with existing Kodiak and Juneau LAN's to facilitate
statewide access to the above DBMS as well as to accommodate the
need to access data, currently in electronic form, located in
Kodiak and Juneau. For example the catch data cited above is
currently in the statewide f ish ticket system data base which
resides in Juneau. The network will accommodate all Commercial
Fisheries Division personnel and have the potential capacity to be
expanded to all departmental personnel.

BUDGET: ADF&G

Personnel Services $ 80.0
Travel 5.0
Contractual 0.0
Supplies 1.0
Equipment 34.0

Total $120.0

198

MARINE MAMMAL ASSESSMENT

Although the most visible impact of the EVOS on marine mammals was
the large number of dead sea otters, other marine mammal species
were potentially injured by the spill, including Steller sea lions,
harbor seals, killer whales, and endangered humpback whales.

In 1989, seven studies were assembled and implemented to gather
information on injury to marine mammals. Aerial surveys for
stranded cetaceans were also conducted. Additional data on
injuries to sea otters were gathered at the sea otter
rehabilitation centers.

All of these studies, except one, will be continued in 1990.
Marine Mammals Study Number 3, Cetacean Necropsies to Determine
Injury from the EVOS, is discontinued. No oil spill related
cetacean strandings are expected in the second year. Cancellation
of this study does not exclude the possibility of collecting
samples opportunistically, should fresh carcasses be encountered.

In many cases, the 1990 studies have been expanded and modified in
response to knowledge gained during the f irst year and comments
from reviewers and the public. The sea otter study is far more
extensive than last year and will look at possible physiological
and toxicological impacts that could result in long-term, sublethal
injuries. The assessment of population effects is also greatly
expanded and will look closely at pre and post spill populations,
population dynamics, and reproductive biology. Data from studies
on Steller sea lions and harbor seals will provide information on
toxicological effects of the EVOS. The ongoing cetacean studies
are intended to provide information on changes in cetacean use of
the spill zone, to assess impacts that may not become apparent
until the second year, and to corroborate information on injury to
killer whales gathered during the 1989 studies.

199

MARINE MAMMAL STUDY NUMBER 1

Study Title: Effects of the EVOS on the Distribution and
Abundance of Humpback Whales in PWS, Southeast
Alaska, and the Kodiak Archipelago

Lead Agency: NOAA

INTRODUCTION

During the f irst year of the humpback whale damage assessment,
photographs of individual humpback whales occurring in PWS and
Southeast Alaska were collected from May to September 1989 to
assist in determining the impact of the EVOS on humpback whale lif e
history and ecology. In PWS, f our dedicated research vessels
traversed 9,623 nautical miles to search for and photograph whales,
reflecting 260 days of field research. In Southeast Alaska,
researchers working from five different vessels spent 1,011 hours
searching for whales for a total of 230 days of field research. An
additional 155 hours were spent off Kodiak conducting marine mammal
sighting surveys.

Concerns about the North Pacific humpback whale stock encountering
or being exposed to oil is well founded based on evidence in the
literature. The humpback whale is currently listed as an
endangered species. Changes in abundance of humpback whales (after
being exposed for one season to the EVOS) would more likely occur
in the second year.

This study will obtain photographs of individual humpback whales
occurring in PWS from early June to late September 1990. Calves of
the year will be documented. Photographs collected will be
compared to the Alaskan photographic database for the years 1977 to
1989 to determine if changes have occurred in whale abundance,
seasonal distribution, continuity of habitat usage, and mortality
and natality rates. Results of this research will allow
determination of the extent of injury (displacement) or loss
(reduction in numbers) to humpback whale populations as a result of
the EVOS.

OBJECTIVES

A. Count and individually identify humpback whales entering PWS.

B. Test the hypothesis that humpback whale distribution and
abundance within PWS is similar to that reported for 1989 and
previous years.

C. Test the hypothesis that humpback whale natality has not
changed since the EVOS.

D. Test the hypothesis that humpback whale mortality rates have

200

not changed since the EVOS.

METHODS

Shore-based camps (shared with personnel from the killer whale
project) will be established in PWS to conduct photo-identif ication
studies on humpback whales from small boats (June through
September 1990). Camp locations will be sinilar to those set up in
1989. Early in the season, camps will be located in the
northwestern area of PWS (Naked Island); the southwestern region
at Squire Island (off the southwest side of Knight Island); and
either on Hinchinbrook Island or off the northern side of Montague
Island. Camps may be moved during the field season based on whale
distributional data collected during the study. Each humpback
whale camp will be staf fed by at least two biologists equipped with
one small boat. For consistency in data collection, key personnel
will remain in the field throughout the 4-month period.

Weather permitting, field personnel will spend an average of 8 to
10 hours per day conducting boat surveys searching for whales.
Effort will be comparable to the 1989 season. Specific areas,
known for whale concentrations, will be investigated first.
However, if reports of whales are received from other sources (e.g,
sighting network described below) these areas are examined. if
whales are not located in "known" areas and opportunistic sighting
reports are not available; a general search pattern will be
developed and implemented. Travel routes taken by whales will be
surveyed. When whales are sighted, researchers end their general
search effort and approach the whales to collect photo-
identification information. A humpback whale survey form is
completed for each encounter.

To obtain a high-quality photograph, an approach within 30-60
meters is required. Photographs are taken of the ventral surface
of the fluke and left side of the dorsal fin.

Daily effort logs are maintained each day which will permit 1)
quantification of the amount of time searching for whales versus
photographing whales, 2) quantification of search effort under
different weather conditions; 3) daily vessel trackline, and 4) an
estimation of the number of vessels/ aircraft encountered in the
study area.

To increase the sighting effort within PWS to ensure that all
whales are being seen and photographed, a marine mammal sighting
network will be organized throughout the PWS area. This network
will record all sightings of whales collected opportunistically
from Alaskan State Ferries and private aircraft and boaters. Whale
sightings are reported directly to the whale research vessels.
Field teams respond by searching out the area where whales were
reported to collect photographic data.

201

All photographs of humpback whales will be analyzed for individual
identification. An individual whale's ventral aspect of the fluke
is recorded (notes and sketches) . Photographs are then grouped by
individual. Each individual whale identified is then visually
compared to the historical photographic database. A second,
independent matching analysis will be performed to ensure accuracy.
Considerable expertise exists in recognizing individual whales
through computer matching of color patterns. Once all photographs
are properly cataloged and evaluated, it is then possible to
determine 1) the identification of individual whales and 2) if the
individual whales have altered their distributional patterns.

To avoid biases in data interpretation, it is important that the
amount of effort in searching for and photographing whales in 1990
is at least equal to (but not less than) that completed in previous
years. When comparing differences in sightings per unit effort,
either the Kolmogorov-Smirnov or Mann-Whitney test will be used.

Calves of the year will be noted and their mothers identified.
Natality (number of calves per adult female) will be calculated for
each area. Comparisons of natality among years will be made using
either Chi-square tests or Z tests for comparing differences
between two proportions (selection of test based on sample size).
Stranded animals found during the 1990 season will be reported.
Distributional comparisons will be made on a qualitative basis.

BIBLIOGRAPHY

The following humpback whale articles are pertinent to the studies
being conducted in Alaska.

Baker, C. S. 1985. The Population Structure and Social
organization of Humpback Whales (Megaptera novaeangliae) in
the Central and Eastern-North Pacific. Ph.D. Dissertation.
University of Hawaii. 306 pp.

Baker, C. S. and L. Herman. 1987. Alternative population
estimates of Humpback Whales (Megaptera novaeangliae) in
Hawaiian Waters. Canadian Journal of Zoology, 65: 2818-2821.

Hall, J. S. 1979. A Survey of Cetaceans of Prince William Sound
and Adjacent Waters--their Numbers and Seasonal Movements.
In: Environmental Assessment of the Alaskan Continental
Shelf. NOAA OCSEAP Contract No. 01-6-022-15670. 72 pp.

Hall, J. D. 1981. Aspects of the Natural History of Cetaceans of
Prince William Sound, Alaska. Ph.D. Dissertation. University
of California Santa Cruz. 148 pp.

Hall, J. D. i982. Prince William Sound - Humpback Whale
Population and Vessel Traffic Study. Final Contract Report
No. 81-ABG-00265 to National Marine Fisheries Service, 20 pp.

202

Johnson, J. J. and A. A. Wolman. 1984. The Humpback Whale
(Megaptera novaeangliae). Marine Fisheries Review 46(4):30-
.37.

Jurasz, C. M. and V. P. Jurasz. 1979. Feeding Modes of the
Humpback Whale (Megaptera novaeanaliae) in Southeast Alaska.
Sci. Rep. Whales Res. Inst. No. 31: 69-83.

Katona, S., B. Baxter, 0. Brazier, S. Kraus, J. Perkins, and H.
Whitehead. 1979. Identification of Humpback Whales by Fluke
Photographs. In: H. E. Winn and B. L. Olla (eds). Behavior
of Marine Animals - Current Perspectives in Research, Vol. 3:
Cetaceans: pp. 33-44. Plenum Press, New York.

Rice, D. W. 1978. The Humpback Whale in the North Pacific:
Distribution, Exploitation, and Numbers. In: Report on a
Workshop on Problems Related to Humpback Whales (Megaptera
novaeangliae) in Hawaii. NTIS Report PB-280 794. pp. 29-44.

Watkins, W. A., K. E. Moore, D. Wartzok, and J. H. Johnson.
1981. Radio Tracking of Finback (Balaenoptera physalus) and
Humpback (Megaptera novaeangliae) Whales in Prince William
Sound, Alaska. Deep-Sea Research 78: 577-588.

Wing, B. L. and K. Krieger. 1983. Humpback Whale Prey Studies
in Southeastern Alaska, Summer 1982. Report by Northwest and
Alaska Fisheries Center Auke Bay Laboratory, 60 pp. National
Marine Fisheries Service, NOAA, P. 0. Box 155, Auke Bay,
Alaska 99821.

von Ziegesar, 0. 1984. A Survey of the Humpback Whales in
Southwestern Prince William Sound, Alaska 1980, 1981, and
1983. A Report to the State of Alaska, Alaska Council on
Science and Technology, 68 pp.

von Ziegesar, 0. and C. 0. Matkin. 1989. A Catalogue of Prince
William Sound Humpback Whales Identified by Fluke Photographs
Between the Years 1977 and 1988. 2 8 Pages. North Gulf
Oceanic Society, P. 0. Box 15244, Homer, Alaska

BUDGET: NOAA

Salaries $ 0.0
Travel 6.0
Contracts 80.0
Supplies 2.0
Equipment 4.0

Total $ 92.0

203

MARINE MAMMAL STUDY NUMBER 2

Study Title: Assessment of Injuries to Killer Whales in PWS,
Kodiak Archipelago, and Southeast Alaska

Lead Agency: NOAA

INTRODUCTION

During the first year photographs of individual killer whales
occurring in PWS, Southeast Alaska, and the Kodiak Archipelago were
collected from May to September 1989 to assess the impact of the
EVOS on killer whale life history and ecology. In PWS, f our
dedicated research vessels traversed 9,623 nautical miles searching
and photographing whales, reflecting 260 days of field research.

This year's study will obtain photographs of individual killer
whales occurring in PWS and adjacent waters from early June to late
September 1990. Calves of the year will be documented.
Photographs collected will be compared to the Alaskan photographic
database for the years 1977 to 1989 to determine if changes have
occurred in whale abundance, seasonal distribution, continuity of
habitat usage, pod integrity, and mortality or natality rates.
Results of this research will allow determination of the extent of
injury (displacement) or loss (reduction in numbers) to killer
whale populations as a result of the EVOS.

OBJECTIVES

A. Count the number and individually identify killer whales
within PWS and adjacent waters.

B. Test the hypothesis that killer whale distribution within
PWS and adjacent waters is similar to that reported for
previous years (1984-1989).

C. Test the hypothesis that pre- and post-spill killer whale
pod structure and integrity have remained constant.

D. Test the hypothesis that killer whale natality rates
within PWS have not changed since the EVOS.

E. Test the hypothesis that killer whale mortality rates
within PWS have not changed since the EVOS.

METHODS

Shore-based camps will be established in PWS to conduct photo-
identification studies on killer whales from small boats (May
through September 1990). Camp locations will be similar to those
set up in 1989. Early in the season camps will be located in the

204

northwestern area of PWS (Naked Island) , the southwestern region at
Squire Island (off the southwest side of Knight Island); and either
on Hinchinbrook Island or off the northern side of Montague Island.
Camps may be moved during the field season based on whale
distributional data collected during the study. Each camp is
staf f ed by at least two biologists and one small boat. For
consistency in data collection, key personnel remain in the field
throughout the 4-month period.

Weather permitting, field personnel will spend an average of 8 to
10 hours per day conducting boat surveys searching f or whales.
Effort will be comparable to the 1989 season. Specific areas,
known for whale concentrations, are investigated first. However,
if reports of whales are received from other sources (e.g, sighting
network described below) , those areas are examined. If whales are
not located in known areas and opportunistic sighting reports are
not available, a general search pattern will be developed and
implemented. Travel routes typically taken by whales are surveyed.
When whales are sighted, researchers stop further search efforts
and approach the whales to collect photo-identification
information. A killer whale survey form is completed for each
encounter. When whales are encountered, researchers select a
vessel course and speed to approximate the animals' course and
speed to facilitate optimal photographic positioning.

To obtain a high-quality photograph, an approach within 30-60
meters is required. Photographs are taken of the left side of the
whale's dorsal fin and saddle patch. Any high-performance camera
system can be used to collect the data.

Daily effort logs are maintained each day which will permit 1)
quantification of the amount of time searching for whales vs
photographing whales, 2) quantification of search effort under
different weather conditions; 3) daily vessel trackline, and 4) an
estimation of number of vessels/aircraft encountered in the study
area.

To account for the possible displacement of killer whales to areas
outside PWS and to confirm that the missing whales are not
elsewhere (e.g., particularly the absence of the 22 individuals of
AN pod), photographic studies will be conducted off Kodiak Island.

205

A marine mammal sighting network will be organized throughout
Alaska which includes sightings collected opportunistically from
Alaskan State Ferries and private aircraft and boaters. To provide
extended coverage throughout the GOA, marine mammal sighting
information collected by NOAA ships and other research vessels that
have been working areas of interest will be examined. All killer
whale data will be extracted and summarized. If photographs were
collected; an attempt will be made to obtain them.

All photographs of killer whales will be analyzed for individual
identification. Each negative (or prints as needed) is placed
under a dissection microscope for identification purposes and notes
and sketches made. Photographs are then grouped by individuals.
Each identified whale is then visually compared to the historical
photographic database available at the Pacific Biological Station,
Nanaimo, British Columbia, Canada. Once all photographs are
properly entered and evaluated, it is then possible to determine 1)
if all members of the pod were present, and 2) if pod
structure/ integrity is similar to previous years. Missing animals
are noted. The stability of resident pods over time is such that
if an individual is listed as missing for at least one year, that
missing whale is considered dead.

To avoid biases in data interpretation, effort in searching for and
photographing whales in 1990 will at least be equal to (but not be
less than) that completed in previous years. For a large pod ( >12
animals), the liklihood of obtaining photographs of all individuals
are increased as the number of encounters are increased. Some
individuals, and certain pods, are more likely to approach vessels
making photographic documentation easier; while others keep a
considerable distance away making for more difficult conditions.
Whale behavior also plays a role when attempting to obtain
photographs of individual whales. If the pod is resting (typically
grouped together) , it is easier to obtain photographs of all whales
than when the pod is travelling (spread out through an area) .
Researchers with prior killer whale experience in a particular area
who are capable of recognizing individuals, will also enhance the
likelihood of accounting for all whales within a pod.

Calves of the year will be noted and their mothers identified.
Natality (number of calves per adult female) will be calculated for
each pod for each year and comparisons made between resident and
transient groups using descriptive statistics. Mortality rates
through 1989 will also be calculated for resident groups.
Mortality for transient pods will be calculated when necessary data
are available.

General location of whales will be recorded each time photographs
are taken, allowing comparisons of pod distributions among years.
Changes in normal distribution patterns will be reported.

206

BIBLIOGRAPHY

The following killer whale articles are pertinent to the studies
being conducted in Alaska.

Anon. 1982. Report on the workshop on identity, structure, and
vital rates of killer whale populations. Rept. Int. Whal.
Commn, 32: 617-631..

Balcomb, K. C. 1978. Orca Survey 1977. Final Report of a Field
Photographic Study Conducted by the Moclips Cetological
Society in Collaboration with the U. S. National Marine
Fisheries Service on Killer Whales (Orcinus orca) in Puget
Sound. Unpub. Report to the Marine Mammal Division, National
marine Fisheries Service, Seattle, Washington, 10 pages.

Bigg, M. A. 1982. An Assessment of Killer Whale (Orcinus orca)
Stocks off Vancouver Island, British Columbia. Rept. Int.
Whal. Commn., 32: 655-666.

Braham, H. W. and M. E. Dahlheim. 1982. Killer Whales in Alaska
Documented in the Platforms of Opportunity Program. Rept.
Int. Whal. Commn. 32: 643-646.

Calambokidis, J., J. Peard, G. H. Steiger, J. C. Cubbage, and R.
L. DeLong. 1984. Chemical contaminants in marine mammals
from Washington State. Natl. Oceanic Atmospheric Admin.,
Tech. Memo, NOS OMS, 6: 1-167.

Ellis, G. 1987. Killer Whales of Prince William Sound and
Southeast Alaska. A Catalogue of Individuals Photoidentif ied,
1976-1986. Sea World Research Institute/Hubbs, Marine Research
Center, Technical Report No. 87-200. April 1987. ,

Fowler, C. W. 1984. Density Dependence in Cetacean Populations.
In "Reproduction in Whales, Dolphins, and Porpoises". Eds. W.
F. Perrin, R. L. Brownell, and D. P. DeMaster. Rept. Int.
Whal. Commn., Spec. Issue 6: 373-380.

Hall, J. D. 1981. Aspects of the Natural History of Cetaceans
of Prince William Sound. Ph.D. Dissertation. University of
California - Santa Cruz. 148 pp.

Heyning, J. E. and M. E. Dahlheim. 1988. Orcinus orca.
Mammalian Species Account, No. 304, pp. 1-9, 4 figs.

Leatherwood, S., K. C. Balcomb, C. 0. Matkin, and G. Ellis.
1984. Killer Whales (Orcinus orca) of Southern Alaska -
Results of Field Research 1984 Preliminary Report. Hubbs Sea
World Research Institute Tech. Report No. 84-175, 59 pp.

Leatherwood, S., A. Bowles, E. Krygier, J. D. Hall, and S.

207

Ignell. 1985. Killer Whales (Orcinus orca) in Southeast
Alaska, Prince William sound, and Shelikof Strait; A Review of
Available Information. Rept. Int. Whal. Commn., SC/35/SM 7.,
10 pp.

Perrin, W. F. and *S. B. Reilly. 1984. Reproductive Parameters
of Dolphins and Small Whales of the family Delphinidae. In
"Reproduction in Whales, Dolphins, and Porpoises". Eds. W. F.
Perrin, R. L. Brownell, and D. P. DeMaster. Rept. Int. Whal.
Commn., Spec. Issue 6: 97-134.

von Ziegesar, 0., G. Ellis, C. Matkin, and B. Goodwin. 1986.
Repeated Sightings of Identifiable Killer Whales (Orcinus
orca) in Prince William Sound, Alaska 1977-1983. Cetus, Vol.
6, No. 2, 5 pp.

BUDGET: NOAA

Salaries $ 45.0
Travel 10.0
Contracts 180.0
Supplies 10.0
Equipment 10.8

Total $ 255.8

208

MARINE MAM14AL STUDY NUMBER 4

Study Title: Assessment of Injury to Steller Sea Lions in PWS
and the GOA

Lead Agency: NOAA

Cooperating Agency: ADF&G

INTRODUCTION

Steller sea lions (Eumetopias Jubatus) are the largest and one of
the most conspicuous pinnipeds inhabiting the GOA. The north GOA
contains a major portion of the worldwide habitat of this
species. Regularly used haulouts are located throughout PWS and
along the Gulf coast. Major breeding rookeries occur at the
entrance to PWS, along the eastern Kenai Coast, in the Barren
Islands, in the northern Kodiak area, at Chirikof Island south of
Kodiak, and in the Semidi Islands, south of Shelikof Strait
(Calkins and Pitcher 1982; Loughlin et al. 1984; Merrick et al.
1987).

Steller sea lions were present in large numbers in PWS during the
oil spill, were exposed to oil immediately after the spill, and
may continue to be exposed for several more years. Initial
observations indicated that sea lions did not attempt to avoid
the oil. Oiled sea lions were reported at haulouts by several
observers.

Steller sea lion populations have declined substantially in much
of their range since at least 1970 (Braham et al. 1980; Calkins
1985; and Loughlin et al. 1984). This decline appears to be
accelerating in the northern Gulf of Alaska (Loughlin et al.
1990; Calkins and Goodwin 1988; Merrick et al. 1987 Loughlin et
al. 1984). The NMFS has listed this population as threatened
under terms of the Endangered Species Act. Further reductions of
this species in this area could result in adverse ecological
impacts on the marine ecosystem.

Various studies and observations suggest that several thousand
sea lions move across the northern GOA in the spring; probably
most return to the large rookeries along the Kenai coast and
northern Kodiak to pup and breed (Calkins and Pitcher 1982).
Many of these animals use PWS during the period of March through
May (Calkins and Pitcher 1982).

This study addresses the impacts of the EVOS on the Steller sea
lion population in PWS and the GOA. Sea lion pups will be
counted at rookeries from Chowiet Island to Seal Rocks. These
counts will be compared between years for 1989 and 1990 as well
as compared to historical data of a similar nature. The counts
will be used to monitor relatively large changes which may occur

209

in the population. Premature pupping will be investigated by
comparing premature pupping rates at an area close to the oil
spill (Cape St. Elias) to an area a substantial distance from the
spill (Chirikof Island). It is assumed that, because of
proximity, the sea lions at Cape St. Elias had a higher exposure
rate to the oil than those at Chirikof Island. Premature pupping
has been suggested as a possible toxicological consequence of
exposure to hydrocarbons during gestation. This would not have
been a likely effect in 1989 because the spill occurred late in
the gestation period. Toxicological and histological examination
of tissues from sea lions will provide information on absorption
and possible damage caused by hydrocarbons including the pups
born prematurely. Tissues have been taken from both animals
collected specifically for that purpose and from animals found
dead in oiled areas. All animals collected for tissue analysis
were examined by a certified veterinary pathologist. Tissue
analysis may show if injury occurred to sea lions.

OBJECTIVES

A. Test the hypothesis that premature pupping occurs at a
higher rate at a hauling area nearer the oil spill.

B. Test the hypothesis that pup production is lower in the
vicinity of the oil spill.

C. Estimate hydrocarbon levels in sea lion tissues to within
10% of the actual value 95% of the time.

D. Test the hypothesis that tissue damage has occurred.

METHODS

Premature pupping has occurred historically in the GOA sea lion
population (Calkins and Goodwin 1988) and may be accelerated by
toxic effects of oil. observations and searches for aborted
fetuses will be made at all hauling areas and rookeries visited
after March 1989. Premature pupping will be measured at Cape St.
Elias and at Chirikof Island by stationing observers at these
locations for a 4 week period during April and May. Each
premature birth will be recorded and each fetus will be examined.
Tissues will be preserved from each animal examined for
hydrocarbon and histological analysis. Adults will be counted
daily at each location and a rate of premature births to adults
present will be determined. Daily observations will be conducted
using spotting scopes and binoculars.

Pup production will be measured by counting pups directly at the
six rookeries within the oil spill area from Chowiet Island to
Seal Rocks (Calkins and Pitcher 1982). Sea lions from all of
these rookeries could be assumed to be impacted. This count has
been conducted in June/July 1989 and will be conducted again in
1990.

210

In order to insure accurate hydrocarbon analysis of tissues, it
is important to preserve tissues within six hours after death.
Accordingly, sea lions will be collected under terms of a permit
issued by the NMFS, upon consultation with NMFS, if available
tissue analyses indicate further collections are warranted.

In accordance with established criteria of the histopathology
technical group, a Board certified veterinary pathologist will
perform histopathological analysis of all sea lion tissues and a
second board certified pathologist will perform an independent,
blind reading of a subsample of histology slides. Reference
histology slides will be retained and archived toxicological
samples will be frozen and stored.

Data analysis for comparing premature pupping between two areas
to determine if the proportion of premature pups born to adults
in an area close to the oil spill is higher than an area further
away from the oil spill will be tested with a two sample t-
statistic on rate (Snedcor and Cochran 1980) at alpha=0.05 in the
lower tail. The normality assumption will be examined with Q-Q
plots (Hoaglin, Mosteller and Tukey 1985) and if necessary, the
data will either be transformed to meet this assumption or a
Mann-Whitney, nonparametric statistic will be used (Conover
1980).

Analysis of pup counts will utilize a regression model to predict
expected numbers of sea lion pups in the absence of the EVOS.
Because sea lion pup numbers have been declining since 1979, the
1990 pup count will be compared to the 1989 count and historical
data to determine if it is lower than what the regression model
suggests-. A Hotelling's T2 statistic will be used to test if the
observed 1990 count is significantly lower than the predicted
value from the regression equation modeling of the pre-1989 sea
lion decline (Neter and Wasserman 1974). The validity of the
model will be tested using data from counts from unoiled areas.

It is assumed that the distribution of pup counts is normal. The
regression model would accurately predict pup numbers in the Gulf
of Alaska in the absence of the oil spill. The regression model
is correctly specified, and has constant variance. The
proportion of sea lions exhibiting hydrocarbon uptake will be
estimated and an exact 95% confidence interval determined using
the binomial distribution (Ostle and Mensing 1982).

BIBLIOGRAPHY

Braham, H. W., R. D. Everitt, and D. H. Rugh. 1980. Northern sea
lion population decline in the. eastern Aleutian Islands.
Fish. Bull. 44:25-33.

Calkins, D. G. 1985. Sea lion pup counts in and adjacent to
Shelikof Strait. Final report submitted to North Pacific
Fisheries Management Council, contract 84-1. Alaska

211

Department of Fish and Game, Anchorage Alaska. 13pp.

Calkins D. G. and E. Goodwin. 1988. Investigation of the
declining sea lion population of the Gulf of Alaska.
National Marine Mammal Laboratory contract NA-ABH-00029.
Alaska Department of Fish and Game, Anchorage, Alaska. 76pp.

Calkins, D. G. and K. W. Pitcher. 1982. Population
assessment,ecology, and trophic relationships of Steller sea
lions in the Gulf of Alaska. In: Environ. Assess. of the
Alaskan Cont. shelf. Final Reports. 19:445-546.

Conover, W. J. 1980. Practical nonparametric statistics
2nd ed. John Wiley and Sons, New York. 493 pp.

Hoaglin, D. C., F. Moesteller, and J.W. Tukey. 1985. Exploring
data tables, trends, and shapes. John Wiley and Sons, New
York. 527 pp.

Johnson, R. A. and D. W. Wichern. 1988. Applied multivariate
analysis. 2nd ed. Prentice Hall, Englewood Cliffs, New
Jersey. 607 pp.

Loughlin, T. R., D. J. Rugh, and C. H. Fiscus. 1984.
Northern sea lion distribution and abundance: 1956-80. J.
Wildl. Manage. 48:729-740.

Loughlin, T. R., A. S. Perlov, and V. A. Vladimirov. 1990.
Survey of northern sea lions (Eumetopias jubatus) in
the Gulf of Alaska and Aleutian Islands. NOAA
technical memorandum NMFS F/NWFC- 176. 26 pp.

Merrick, R. L., T. R. Loughlin, and D. G. Calkins. 1987.
Decline in abundance of the northern sea lion Eumetopias
jubatus in Alaska, 1956-86. Fish. Bull. 85:351-365.

Mendenhall, W., R. L. Scheaffer, and D. D. Wackerly. 1981.
Mathematical statistics with applications. Second
edition. Duxbury Press, Boston, Mass. 686pp.

Neter, J. and W. Wasserman. 1974. Applied linear statistical
models: regression, analysis of variance, and experimental
designs. Richard D Irwin, Inc. Homewood, Illinois. 842pp.

Ostle, B. and R. W. Mensing. 1982. Statistics in research,
3rd ed. The Iowa State Univ. Press. Ames, Iowa. 595 pp.

Snedecor, G. W. and W. G. Cochran. 1980. Statistical
methods. Seventh edition. The Iowa State Univ. Press, Ames
Iowa. 507pp.

212

BUDGET: NOAA

Personnel $ 107.4
Travel and per them 6.0
Services 45.5
Commodities 6.8 1
Equipment 5.5

TOTAL $171.2

213

MARINE MAMMAL STUDY NUMBER 5

StudyTitle: Assessment of Injury to Harbor Seals in PWS and
Adjacent Areas

Lead Agency: NOAA

Cooperating Agency: ADF&G

INTRODUCTION

Harbor seals (Phoca vitulina richardsi) are one of the most
abundant species of marine mammals in PWS and adjacent areas.
They are resident throughout the year, occurring primarily in the
coastal zone where they feed and haul out to rest, bear and care
for their young, and molt (Hoover 1988). They are used for
subsistence purposes by Native residents in the area. Unlike fur
seals (Callorhinus ursinus) and sea lions (Eumetopias Jubatus),
harbor seals do not form distinct rookeries during the pupping
and breeding season. Pups are born at the same locations as
those used as haulouts at other times of year. Some of the
largest haulout sites in PWS, and adjacent waters to those
haulouts, were directly impacted by substantial amounts of oil
during the EVOS. oil that moved into the GOA impacted harbor
seal habitat at least as far to the southwest as Tugidak Island.
Harbor seals swam through oil and breathed at the air/water
interface. On haulouts they crawled through and rested on oiled
rocks and algae. Pups were born on the haulouts in May and June
while some of the sites still had oil on them, resulting in pups
becoming oiled. The same locations were also used during the
molt in August and September.

Trend count surveys indicate that the number of harbor seals in
PWS declined by 40% from 1984 to 1988, and similar declines have
been noted in other parts of the northern GOA (Pitcher 1989).
Additional impacts on harbor seal populations are therefore of
particular concern.

Three ringed seals (Phoca hispida) exposed in the laboratory to
fresh Norman Wells crude oil all died within 71 minutes; six
seals exposed for 24 hours at a field site showed minor damage to
the eyes, kidneys, and liver (Geraci and Smith 1976).
Hydrocarbons were rapidly absorbed into the body fluids and
tissues when ringed seals were exposed to oil by either immersion
or ingestion (Engelhardt et al. 1977). In 1974, oil from an
unknown source came ashore at a grey seal (Halichoerus gryRus)
pupping beach in Wales. Two pups drowned when they became so
encased in oil that they were unable to swim, and oiled pups
reached a lower peak weight at weaning than did unoiled pups
(Davis and Anderson 1976).

214

Following the spill, f ield observations were made of seals in
oiled and unoiled areas of PWS. Carcasses of 39 seals were
necropsied and sampled; 19 that were f ound dead or died in
captivity, and 20 that were collected specifically for sampling.
Histopathological and toxicological analyses are in progress.

Last year aerial surveys were conducted during June to count the
number of harbor seal pups and non-pups on 25 oiled and uno 'iled
haulouts in PWS. Aerial surveys were also conducted at the same
25 haulouts'during the fall molt. Results of the fall surveys
have bee 'n compared to results,of surveys flown in 1984 and 1988
to determine whether trends in numbers are similar in oiled and
unoiled areas.

This project proposes to complete histopathological and
toxicological analyses of harbor seal tissues and to provide
.counts of harbor seals on haulouts in oiled and unoiled parts of
PWS and during pupping and molting in two additional years (1990
and 1991). . Data from aerial surveys will be used to evaluate
whether changes occurred in the 'distribution and abundance of
harbor seals following the, EVOS, and whether such changes
coincided with the presence or absence of oil in the Area or on
the haulouts. Toxicological analyses of tissues from oiled and
unoiled seals will allow an assessment of whether hydrocarbons
were assimilated by the seals and how contaminant levels changed
over time. Histopathological examinations will determine the
types and degrees of toxic damage. to tissues. Survey and
laboratory data, in combination with historical data for PWS,
will be used to evaluate whether the EVOS caused a reduction in
pup productivity at oiled sites in 1989 and 1990, and whether
changes in abundance during the 1989 fall molt were due to the
EVOS. This information can then be used to make recommendations
regarding restoration of lost use, populations, or habitat where
injury is identified.

OBJECTIVES

A. Test the hypothesis that harbor seals found dead in the
area affected by the EVOS died due to oil toxicity.

B. Test the hypothesis that seals exposed to oil from the
EVOS assimilated hydrocarbons to the extent that harmful
pathological conditions resulted.

C. Test the hypothesis that the abundance of harbor seals on
the trend count route during pupping and molting decreased
in oiled areas of PWS as compared to unoiled areas.

D. Test the hypothesis that pup production was lower in oiled
than in unoiled areas, or than in years not affected by
the EVOS.

215

METHODS

For one week during pupping in June 1990, small boats will be
used to observe seals and seal haulouts in oiled areas. Haulout
sites will be inspected for the presence of oil or dead animals.
Seals will be examined using 7 to 10-power binoculars and a 25-
power spotting scope to determine whether any have oiled pelage.
Seals observed will be classified as to the degree of pelage
oiling (heavy, moderate, light, or none). If any carcasses are
found that are in suitable condition, they will be necropsied by
trained biologists, veterinarians, or pathologists, and samples
will be obtained and preserved for toxicological and
histopathological examination.

A maximum of 12 additional harbor seals will be collected, under
a permit f rom NMFS. Most will be collected at or adjacent to
sites impacted by the EVOS. One or more seals will be collected
from an area not impacted by the spill, such as southeast Alaska.
Each animal will be necropsied as soon as possible after death by
qualified personnel.

Collected animals will be measured, weighed, and photographed;
time, date, location, and circumstances of collection will be
noted; and any gross abnormalities will be recorded. Blood
samples for serum, plasma, and whole blood analyses will be
taken. Samples will be taken for histopathology and toxicology.
Chain of custody will be maintained for all samples. Samples for
histopathology will be stored in formalin until they are
analyzed. Reference histology slides will be retained and
archived. Toxicology samples will be frozen and stored until
they are sent to an approved laboratory for analysis.

Aerial surveys will be conducted during pupping in June and
molting in September along a previously established trend count
route (Calkins and Pitcher 1984; Pitcher 1986, 1989) that covers
25 haulout sites and includes 6 sites impacted by the EVOS
(Agnes, Little Smith, Big Smith, Seal and Green islands, and
Applegate Rocks), 16 unoiled sites, and 3 intermediate sites that
were not physically oiled but were adjacent to oiled areas.
Visual counts will be made of seals at each site and photographs
taken of large groups for later verification.

During June, separate counts will be made of pups and non-pups.
Pupping surveys are needed in 1990 and should be done in 1991
since there are no historical data available from PWS during the
pupping season with which to compare the 1989 results. Breeding
and embryo implantation for 1990 pups occurred while seals were
still exposed to oil on haulouts and while hydrocarbon levels in
tissues may have been abnormally high.

Surveys during the molt in 1990 and possibly 1991 are necessary

216

to determine whether observed changes in the number of seals on
oiled sites between 1988 and 1989 persist.

All statistical tests for significance will use alpha = 0.05.
Statistical testing is not appropriate for all objectives. The
assessment of cause of death of animals found in areas impacted
by the EVOS (objective A) will require expert evaluation of
limited and varying toxicology and histopathology data sets.

Toxicological results for each seal collected will be entered
into a data base along with information on date and location of
collection; presence of oil in the area; degree of external
oiling of the seal; age, sex, size, and reproductive condition.
Hydrocarbon levels in the tissues will be tabulated by individual
and by groups based on age, sex, collection location, and degree
of oiling. Differences between groups will be tested where
possible using ANOVA (Neter and Wasserman 1974).

Types of pathology detected will be listed for each specimen and
will be grouped into tables by sex, age, collection location, and
degree of oiling. Incidence of pathology will be expressed as
the percentage of the total number of animals in the group that
exhibited a particular type of anomaly. Incidence of pathology
will be evaluated in light of toxicological results for each
specimen.

Harbor seal surveys must be conducted within biological time
windows imposed by the pupping and molting periods. While
results of previous harbor seal trend counts have indicated that
it is desirable to obtain 7-10 counts during a survey period
(Pitcher 1986, 1989), the actual number of counts is frequently
limited by the number of days suitable for flying. During
pupping, the survey window cannot be extended to accommodate
sample size needs since, as pups grow and are weaned, they become
increasingly difficult to differentiate from adults when observed
f rom the air. Similarly, during the molt it is necessary to
confine surveys to the period when maximum numbers are thought to
haul out.

Aerial surveys of harbor seals do not estimate the total number
of seals present since they do not account for seals that are in
the water or seals hauled out at locations not on the trend count
route. Surveys provide indices of abundance based on the number
of hauled out seals counted on the trend count route.
Interpretation of trend count surveys relies on the assumption
that counts of harbor seals on select haulout sites are valid
linear indices of local abundance. We assume that within a given
biological window, such as the pupping or molting period, haul
out behavior remains the same f rom one year to the next, and
counts can thus be compared. Standardization of procedures
minimizes the affects of variables such as tide and weather that

217

could influence the number of seals hauled out on a given day.

The trend count route includes haulouts impacted by the EVOS, as
well as haulouts that are north, east, and south of the primary
area impacted by oil. There is an adequate sample of both oiled
and unoiled areas.

There are no historical data on the distribution of harbor seals
in PWS during the pupping period. The first surveys during
pupping were conducted in June 1989 after the EVOS. In order to
gather these data it will be necessary to conduct surveys in at
least 1990 and 1991. These data will be used in a retrospective
analysis comparing counts of seals in oiled and unoiled sites
between years and using the same statistical techniques employed
for fall molting surveys (Frost 1990).

Fall molting surveys of the trend count route were conducted in
1983, 1984, 1988, and 1989. The 1984, 1988, and 1989 counts are
considered reliable and will be used for comparisons with data
collected in 1990 and possibly 1991. Analysis of count data and
comparisons to other years will be conducted following
statistical methodology used for 1989 molting surveys
(Frost 1990).

A repeated measures ANOVA (Winer 1971) will be conducted on the
trimean (Hoaglin et al. 1985) of the site count data in order to
examine trends in abundance at oiled versus unoiled sites. The
trimean statistic will be used as the measure of central tendency
because sets of counts at a single location sometimes show
bimodal distributions or extreme variations. This analysis
assumes random samples, constant variance, and normality of the
differences. If necessary, transformations (Snedecor and Cochran
1980) will be used to ensure constant variance and normality.
The test assumes that the mean proportion of the population
hauled out on the trend count route is constant over years.
Hypotheses addressing Objective C will be tested using orthogonal
contrasts derived from the ANOVA.

In order to compare pup production at oiled and unoiled sites, a
one-way analysis of co-variance (Neter and Wassermann 1974) will
be performed on the square roots of the trimeans (Hoaglin et al.
1985) of pup counts, using the square roots of non-pup counts as
a covariate. The square root transformation will be used to
correct for non-constant variation of the count data (Snedecor
and Cochran 1980). Linear contrasts (Neter and Wasserman 1974),
where the average number of pups is adjusted to a common number
of adults, will be used to test working hypotheses.

218

BIBLIOGRAPHY

Calkins, D. and K. Pitcher. 1984. Pinniped investigations in
southern Alaska:1983:84. Unpubl. Rep. ADF&G, Anchorage,
AK 16pp.

Davis, J. E. and S. S. Anderson. 1976. Effects of oil pollution
on breeding gray seals. Mar. Poll. Bull. 7:115-118.

Engelhardt, F. R., J. R. Geraci, and T. G. Smith. 1977. Uptake
and clearance of petroleum hydrocarbons in the ringed
seal, Phoca hispida. J. Fish Res. Board Canada 34:1143-
1147

Frost, K. J. 1990. Marine Mammals Study Number 5: Assessment
of injury to harbor seals in Prince William Sound,
Alaska, and adjacent areas. State-Federal Natural
Resource Damage Assessment for April-December 1989.
Unpubl. Prelim. Status Rep. ADF&G, Fairbanks, AK. 27pp.

Geraci, J. R., and T. G. Smith. 1976. Direct and indirect
effects of oil on ringed seals (Phoca hispida) of the
Beaufort Sea. J. Fish. Res. Board Canada 33:1976-1984.

Hoaglin, D. C., F. Mosteller, and J. W. Tukey. 1985. Exploring
data tables, trends, and shapes. John Wiley & Sons. New
York. 527pp.

Hoover, A. A. 1988. Pacific harbor seal. Pages 125-157 in J.
W. Lentfer (ed). Selected Marine Mammals of Alaska:
Species Accounts with Research and Management
Recommendations. U.S. Marine Mammal Commission,
Washington, D.C.

Neter, J., and W. Wasserman 1974. Applied linear statistical
models. Richard D. Irwin, Inc., Homewood, Illinois.
842pp.

Pitcher, K. W. 1986. Harbor seal trend count surveys in
southern Alaska, 1984. Unpubl. Rep. ADF&G, Anchorage,
AK. 10pp.

Pitcher, K. W. 1989. Harbor seal trend count surveys in southern
Alaska, 1988. Final Rep. Contract MM4465852-1 to U.S.
Marine Mammal Commission, Washington, D.C. 15pp.

Snedecor, G. W. and W. G. Cochran. 1980. Statistical Methods.
Iowa State University Press, Ames, Iowa. 507pp.

219

Winer, B. J. 1971. Statistical principles in experimental
design. 2nd Ed. Mcgraw-Hill, New York, New York.
907pp.

BUDGET: NOAA

salaries $ 82.7
Travel 15.1
Contracts 42.1
Supplies 5.4
Equipment 14.0

Total $ 159.3

220

MARINE MAMMAL STUDY NUMBER 6A

Study Title: Assessment of the Magnitude, Extent, and Duration
of Oil Spill Impacts on Sea Otter Populations in
Alaska.

Lead Agency: FWS

INTRODUCTIOX

In the first year following the EVOS, several hundred sea otters
are known to have died as a result of contamination by oil. Death
occurred from hypothermia and from severe liver, kidney, and lung
damage as a result of ingestion of oil and inhalation of toxic
aromatic compounds present during the early period of the spill.
Long-term or chronic effects of oil on sea otters are not known,
but initial results from the first year's studies indicate sea
otter populations have been detrimentally affected. Potential
effects may occur as the result of debilitating or sublethal
injury, accumulation of toxins, and loss or contamination of the
food supply. The capacity of the population to recover to pre-
spill levels is not known. This study will assess the impacts of
the oil spill on Alaska sea otter populations through (1) surveys
of wild populations living in oiled and unoiled areas, (2) genetic,
hematological, histopathological and toxicological analysis of
tissues collected from live and dead sea otters, (3) analysis of
survival, reproduction and movements of adult females and pups
living in oiled and non-oiled areas and (4) analysis of population
dynamics of dead and living sea otters recovered from or living in
oiled and non-oiled areas.

OBJECTIVES

A. BOAT SURVEYS

1. Test that differences in sea otter densities are not
significantly different between oiled and unoiled areas.

2. Test for differences in sea otter densities between pre-
and post-event surveys in oiled and unoiled areas.

3. Estimate the magnitude of any change between pre- and
post-event sea otter population estimates in PWS.

4. Estimate post-event sea otter population size and monitor
population trends of sea otters in PWS.

5. Estimate winter 1990 offshore densities of sea otters in
oiled and unoiled areas.

221

B. HISTOPATHOLOGY AND TOXICOLOGY

1. Test the hypothesis that sea otters residing in regions
that were not affected by the oil spill have lower levels
of hydrocarbons in their visceral f at and whole blood
than sea otters residing in areas that were affected by
oil.

2. Test the hypothesis that sea otter carcasses f ound in
oiled portions of the Alaska coastline subsequent to the
oil spill contain levels of hydrocarbon contamination
similar to those in sea otters killed immediately as a
result of the spill.

3. Test the hypothesis that sea otter carcasses f ound in
oiled areas subsequent to the spill contain higher
burdens of hydrocarbon contaminants than sea otter
carcasses found in non-oiled areas or those analyzed
before the spill.

4. Evaluate the nature and cause of death of sea otters that
died subsequent to the oil spill by performing complete
gross and histopathological examinations of carcasses
recovered after September 1, 1989.

C. CAPTURE OF ADULT FEMALE AND JUVENILE SEA OTTERS

1. Test the hypothesis that pup -survival pre-weaning is not
different between oiled and non-oiled areas.

2. Test the hypothesis that weanling survival at various age
intervals is not different between oiled and non-oiled
areas.

3. Test the hypothesis that survival of adult female sea
otters is not different in oiled and non-oiled areas.

4. Test the hypothesis that pupping rates of adult female
sea otters are not different between oiled and non-oiled
areas.

5. Evaluate the movements of weanling and adult female sea
otters with respect to areas in PWS that have been
affected by the oil spill.

6. Test the hypothesis that blood values (obtained from
complete blood counts and blood panel) do not differ
between samples collected from otters from oiled and non-
oiled areas.

222

D. CAPTURE OF ADULT MALE SEA OTTERS

1. Test the hypothesis that blood values (hematogram and
chemistry) do not differ among male sea otters living in
the oil spill zone and males living in non-oiled areas.

2. Test the hypothesis that variation of DNA content in
lymphocytes does not differ among male sea otters living
in the oil spill zone and males living in non-oiled
areas.

3. Test the hypothesis that DNA structure in sperm cells
(measured by the stability of nuclear chromatin) does not
dif f er among male sea otters living in the oil spill zone
and males living in non-oiled areas.

4. Test the hypothesis that spermatogenic function, measured
by a DNA profile of the testicular cells, does not differ
among male sea otters living in the oil spill zone and
males living in non-oiled areas.

5. Test the hypothesis that proportion of morphologically
normal sperm cells, measured by light microscopy, does
not differ among male sea otters living, in the oil spill
zone and males living in non-oiled areas.

6. Test the hypothesis that levels of hemoglobin adducts,
measured by isoelectric focusing and capillary
electrophoresis of hemoglobin, do not differ among male
sea otters living in the oil spill zone and males living
An non-oiled areas.

7. Test the hypothesis that levels of plasma proteins,
including haptoglobin, quantified by gel electrophoresis,
do not differ among male sea otters living in the oil
spill zone and males living in non-oiled areas.

E. ANALYSIS OF POPULATION DYNAMICS BASED ON CARCASSES IN MORGUE

1. Test the hypothesis that the sex and' age structure of
dead otters recovered during the 5-month period after the
oil spill did not dif f er among various geographic regions
and, hence, can be pooled for demographic analysis.

2. Test the hypothesis that age structure of dead otters
collected after the spill does not differ from the age
structure of otters which died of natural causes before
the spill.

3. Assess potential biases in the sample of dead otters
collected caused by differential mortality and/or
differential probability of carcass recovery.

223

4. Develop sex and age specific survival schedules that
ref lect natural survival in the populations prior to the
spill.

5. Develop age specific fecundity schedules that reflect
natural reproductive rates in the populations prior to
the spill.

6. Construct a population model for assessing population
recovery in areas affected by the oil spill.

METHODS

BOAT SURVEYS

Surveys will be conducted from small boats manned by an operator
and two observers.

A stratified random sampling design, including shoreline,
coastal/pelagic and pelagic strata, will be used to meet Objectives
A1-5. Approximately 29% of the shoreline and 25% of
coastal/pelagic and pelagic transects will be surveyed once in
March 1990 and three times (one survey each in June, July and
August) during summer 1990 jointly with Bird Study Number 2. All
sea otters within transect boundaries will be recorded. The
shoreline stratum includes all water within 200 m of any shoreline,
and will be surveyed by traveling 100 m offshore, parallel to the
coast, at 5-10 knots. The shoreline stratum is divided into
transects consistent with those used during 1984-1985 surveys.
(Irons et. al., 1988). Sampling strip width and protocols are
similar for pelagic surveys.

Strip transect sampling is conducted under the assumption that all
sea otters located within the transect are sighted. If this
assumption is not met, then population estimates are biased low.
If sufficient time and resources are available, an attempt will be
made to verify the boat-based observations with concurrent land-
based observations during the summer 1990 field season. The
sightability assumption is not critical to this study however,
since pre-spill observations were not corrected for this factor.
Results produced by this study should be considered "estimates of
surface abundance" or "population indices" rather than simply
"population estimates".

Abundance estimates will be calculated independently for shoreline,
coastal and pelagic environments using ratio estimation techniques
(Cochran, 1977). Estimates calculated from second-year surveys
will be compared to earlier estimates for the determination of
injury to the sea otter population within PWS. Differences in
otter densities will be tested using two sample t-tests and/or
ANOVA, dependent upon post-stratification of oil condition.

224

HISTOPATHOLOGY AND TOXICOLOGY

Tissue samples for histology will be collected from dead sea otters
recovered in or adjacent to habitats affected by oil. Histology
samples will be sent to the Armed Forces* Institute of Pathology for
processing and analysis. For toxicology, duplicate samples of
liver, kidney, skeletal muscle, bile and fat will be collected from
each sea otter carcass that is recovered from the oil spill zone or
areas outside of the oil spill zone that could serve as controls.
Collection procedures will follow strict guidelines outlined by the
Analytical Chemistry Working Group.

Tissues from carcasses will be graded for degree of necrosis. Mean
values for the degree of necrosis will be computed for tissues of
sea otters of various age, sex, and location parameters provided
sufficient sample sizes exist. Contaminant data from tissues will
be stratified by degree of necrosis, and the effect of necrosis
will be tested. A comparison of contaminants will also be made
between sea otter carcasses found in the oil spill zone and 4
control animals examined in 1986. Data will be checked for
normality; if needed, appropriate transformations will be made.
Blood values and contaminant values will be compared between
treatment and control groups using t-tests or ANOVAs, at a = 0.05.

CAPTURE OF ADULT FEMALE AND JUVENILE SEA OTTERS

In addition to sampling dead sea otters, fat and blood will be
sampled from free-ranging sea otters residing in areas affected by
the oil spill as well as from otters living in non-oiled control
areas. The sampling design calls for sampling of blood and fat
from a total of 100 reproductively mature females and 100 pups and
a total of 100 reproductively mature males. Up to 36cc of blood
will be collected from captured animals. At least 4cc of whole
blood will be frozen for toxicology, and the remaining blood will
be processed as needed for additional assays. From sea otters
which are implanted with transmitters, a small (1/2 inch diameter)
piece of visceral fat will be removed prior to closing the incision
and frozen for toxicology.

The experimental design for the capture and telemetry study takes
advantage of the fact that the oil from the EVOS directly covered
less than one-half of PWS. That situation has been used to develop
a treatment/ control study where the portion of PWS that was covered
by the spill is the treatment area, and the unaffected portion of
eastern PWS, specifically Port Gravina, Port Fidalgo, and Sheep
Bay, is the control area.

Intensive studies of sea otters using radio telemetry will
concentrate on reproductively mature females and large pups in each
area. The pup portion of the study will be initiated in late
summer, 1990. The female study was initiated in October 1989 but
only 23 females in non-oiled habitat and 9 females in oiled habitat

225

were instrumented. Additional radios will be put on females in
spring 1990. Up to 50 reproductively mature females and pups will
be instrumented in both the treatment and control areas.

Sea otters will be caught primarily in unweighted tangle nets or
dip nets. Tangle nets will be set in areas used by sea otters and
anchored at one end. The nets will be monitored closely to prevent
captured sea otters from fighting. Captured animals will be
removed from the nets and placed in holding cages and transported
to a temporary holding cage. Captive sea otters will be fed ad
libitum a combination of fresh frozen dungeness crabs and razor
clams.

The transmitters will be implanted into the body cavity by a
qualified veterinarian. Surgical procedures will follow Williams
and Siniff (1983) and Garshelis and Siniff (1983). Transmitters
measure 311 x 211 x 111, weigh 120 g, and are coated with an inert
material suitable for implantation in sea otters. The transmitters
contain a coiled antenna and are powered by batteries that provide
an operating life of up to 1,000 days. Following immobilization
with a combination of fentanyl citrate and azaperone (Kreeger et
al. 1989), abdominal surgery will be performed. During surgeryl
the animal's status will be monitored by observation of capillary
perfusion, color of mucous membranes, respiration rate and depth,
and heart rate. While the sea otters are still anesthetized they
will be marked with one Temple Tag in each of the flippers and
implanted with a passive glass transponder chip (10 mm x 2 mm)
injected under the skin in the gluteal area (Thomas et al. 1987).
A 30cc blood sample will be taken from each sea otter for blood
hemotograms and chemical analyses. A premolar will be removed from
each adult for age determination.

After release, attempts will be made to relocate each animal at
least bi-weekly from either a boat or airplane. Attribute data
for each relocation, including group size, whether or not the focal
animals have pups, behavior of focal animal, sea condition, and
presence or absence of tags, will be collected. During the pupping
season and shortly following that period, efforts will be made to
locate reproductively mature females at least weekly.

Following instrumentation, efforts will also be made to locate pups
at least weekly. Previous studies have shown relatively high
mortality at weaning in normal populations (Monnett, unpublished
data). Therefore, frequent relocations of weanlings will increase
the chances of recovery of carcasses as soon after death as
possible.

All fresh-dead sea otters found in either the treatment or control
areas will be sent immediately for necropsy. Samples of tissues
for contaminant and histology analysis will be collected.

Alternatives to the implanting of transmitters were considered,

226

including 1) radio tracking devices attached to the outside of the
animal, 2) dyes, 3) visual tags attached to flippers, and 4) no
marking. External telemetry devices have been tried in the past
but are easily damaged by the animal and the environment and have
only an average of about 60 days operational time compared to up to
3 years for internal implants. Dyes are not feasible in the marine
environment and would adversely affect the animals fur. Temple
Tags,and a glass transponder chip will be used in conjunction with
each implant but by themselves would not allow for the tracking of
the animals.

Using standard sample size calculations for testing the difference
between two proportions (Snedecor and Cochran, 1967; p. 221), it
was determined that a sample size of 50 gives a 79% chance of
finding a significant difference at - = .2, given that the
population proportion changes by .2 from an initial value of .5.
A sample of 50 represents a minimum number at which significant
differences between groups might be detected, and the maximum
number that can be realistically instrumented and radio tracked.

Reproductive rates are estimated by counting the number of females
observed with pups divided by the total number of females.
Estimates of survival and reproduction can be calculated over
various time intervals.

Reproductive data will be compared between treatment and control
areas using contingency tables analysis. Two-way contingency
tables will be used except when interactions among age, sex, or
location are of interest. In that case three-way or multi-way
contingency tables based on log-linear models will be used (Sokal
and Rohlf, 1981). Survival estimates will be obtained by the
product limit method and differences in survival patterns will be
tested with log-rank tests (Pollock et al., 1989).

Data on movements and dispersal will be compared between treatment
and control areas. Distance between successive locations, distance
between extreme locations and the minimum convex polygon will be
calculated for each radio-marked sea otter and stratified by sex,
age and reproductive status (Garshelis and Garshelis, 1984; Ralls
et al. 1988). Dispersal distance, here defined as the shortest
distance between the site of weaning (or the location of the last
sighting of females with their pups) and the midpoint of their
first established activity center will be compared for sea otter
pups.

CAPTURE OF ADULT MALE SEA OTTERS

Proposed approaches to damage evaluation in male sea otters include
analysis of blood panels, blood proteins, blood toxin levels, DNA
content and structure (in blood lymphocytes, sperm and testis
cells), and sperm morphology.

227

Blood panels (hematograms and chemistry) are a standard diagnostic
procedure that will be used. Information on chemical damage will
be obtained by examination of blood proteins. Specifically,
increased levels of hemoglobin adducts are indicative of chemical
exposure (Sabbioni and Neumann, 1990; Tornqvist et al., 1988) and
analysis of plasma proteins, with specific examination of
haptoglobin binding, can also be of diagnostic value in assessing
the health of an individual (Van Pilsum et al., 1986).

Nuclear DNA content of blood lymphocytes is a sensitive indicator
of damage to developing cells from clastogenic contaminants in the
environment. Cells can be measured by flow cytometry and, for
normal samples, the resulting frequency histogram of DNA content
should have a very low coefficient of variation. Deviations from
the normal DNA content are detected as an increase in the
coefficient of variation, reflecting damage to the chromosomes.

Spermatozoa are another cell type in which damage to DNA is readily
assayed by flow cytometry. (Evenson, 1986). The structural
stability of sperm nuclear DNA decreases after exposure to toxic
compounds (Evenson, et al., 1985, 1989). The stability of the DNA
is inversely related to male fertility (Ballachey et al., 1987;
Evenson et al., 1980). Morphology is an alternate indicator of
genotoxic damage to sperm cells (Wyrobeck et al., 1983), and thus
proportion of normal sperm in samples from otters living in oiled
versus non-oiled regions will be compared. Testicular cells will
also be collected by fine needle aspiration and examined by flow
cytometry to determine proportions of germ cells.

Males will initially be caught in two areas: 1) Western PWS; and
2) Eastern PWS. The first area was directly affected by spilled
oil and thus is the treatment area. The latter area will serve as
a control. Blood samples, testicular fine needle aspirations
(Hendriks et al., 1969; Thorud et@al., 1978; Nseys et al., 1984;
Sandqvist et al. , 1986; B. Purscell, pers. comm.) and electro-
ejaculated sperm cells (Salisbury et al., 1978; Howard et al.,
1986; Wildt et al., 1989) will be collected from each otter.
Following analysis of these samples a decision will be made whether
males should be caught in the following three areas: 1) The KP
(treatment) ; 2) Kodiak Island (treatment) and 3) Sitka, in
southeast Alaska (control).

It is estimated that a minimum sample size of 18 otters for each
control area and 18 otters for each treatment area will be required
to give an 80% chance of detecting a significant difference of .10
in the proportion of damaged sperm cells between the groups at a =
.05. Twenty animals from each treatment and control area will be
sampled.

Blood will be obtained by jugular venipuncture and'samples will be
handled according to established protocols for the given tests.
Complete blood counts and veterinary panels will be done on the

228

blood samples. A subsample of the blood will be allocated f or
measurement of DNA content of lymphocytes.

An additional subsample of blood will undergo assays f or hemoglobin
adducts. Plasma protein levels will be quantified. DNA in testis
and sperm cells will also be measured. For flow cytometry of sperm
cells, samples will be prepared as described for the SCSA by
Ballachey et al. (1988). For flow cytometry of the testicular
cells, the samples will be prepared as described by Thorud et al.
(1980). , A premolar will be taken from each otter for age
determination. Otters will be tagged and implanted with a
transponder chip prior to release.

A one-way MANOVA will be used to test for differences among the
geographic groups, using a significance level of a = .05. Linear
contrasts will be used to make specific comparisons of the groups.
Analyses on various subsets of variables will be handled separately
(i.e., 1) blood panels, 2) blood DNA/ lymphocytes, 3) blood proteins
and 4) sperm and testis cells). Toxicology data, when available,
will be analyzed in a similar manner. Prior to analysis, the
variables to be tested will be examined and transformed as
necessary to see that they meet the assumptions of the MANOVA.

ANALYSIS OF POPULATION DYNAMICS BASED ON CARCASSES IN MORGUE

All sea otter carcasses found in the spill zone have been kept in
frozen storage. All carcasses not yet examined will be removed
from the freezer and thawed. Degree of decomposition and amount of
oil on the carcass will be used to subjectively place each animal
into one of three categories, killed during the spill due to
exposure to oil, died before the spill, and died during the spill
but unrelated to oil exposure. Standard body measurements (total
length, weight, bacula length) will be recorded for complete
carcasses and sex will be determined based on external genitalia or
tooth measurements if the carcass is not intact. Sections from a
premolar and canine tooth extracted from the skulls of each carcass
will be stained and mounted on slides with the age of each dead sea
otter estimated to the nearest year by counting the number of
cementum, lines present (Schneider 1973, Garshelis 1984).
Reproductive tracts of all adult females which are not badly
decomposed will be examined for implanted fetuses, placental scars,
and corpora albicans. Where possible the approximate age and sex
of fetuses will also be determined. All data collected on each
carcass as well as information available on recovery date and
location and comments will be incorporated into a database which
will be used for analysis.

The sex and age data will be summarized using a 2 x 3 x 4
contingency table, representing 3 geographic areas (PWS, KP, AP)
and 4 age classes (pup, immature, mature, and old) . Log-linear
analysis will be used to test for differences between area, sex,
and age.

229

The age structure of dead otters collected after the spill in the
various geographic areas will be compared to the age structure of
otters collected on beaches in PWS prior to the spill (Johnson
1987). Significant post spill increases in prime age animals will
be indicative of a major mortality event unrelated to normal
mortality processes.

Results from the analysis of age structure will determine if
segments of the age structure data should be eliminated from
survival rate estimation because of possible sampling biases. From
these results the age structures will be constructed for survival
estimation using techniques described by Chapman and Robson (1960)
and Robson and Chapman (1961). An initial analysis will be
performed using all age classes and the model for constant
survival. Chi-square tests will be used to test if the model of
constant survival adequately fits the data. If constant survival
does not appear appropriate, the contribution of each age class to
the chi-square statistic will be examined to determine which age
classes the assumption of constant survival appears appropriate.
The "segment" method (Chapman and Robson 1960) will then be used to
estimate annual survival for these age classes. If data from other
sources indicates that the assumption of a "stationary" population
is not met the survival estimates will be adjusted using estimates
of rate of change in the population (Eberhardt 1988). Estimates of
the impact of senescence on the survival rates of the oldest age
classes will be obtained by using minimum chi-square or nonlinear
least square techniques to fit age structure data to the 3-
component survivorship model developed by Siler (1979) and modified
by Eberhardt (1985). Estimates of the parameters in the
survivorship model will then be used to construct an age-specific
survival schedule for incorporation into a population model.

Results of the examination of female reproductive tracts and the
variability in sizes of fetuses will be used to construct a
fecundity schedule. These data will then be fit to Eberhart's
(1985) fecundity model using minimum chi-square or nonlinear least
square techniques. Estimates of the parameters in the fecundity
model will then be used to construct an age-specific fecundity
schedule for incorporation into a population model.

A Leslie matrix (Leslie 1945, 1948) type population model will be
constructed using the survivorship and fecundity schedules
developed from the analysis described in objectives 4 and 5. The
stable age distribution will be calculated using Lotka's (1907)
equation as modified by Cole (1954) for populations where births
are concentrated in a short time interval each year (Eberhardt and
Siniff 1988). This stable age distribution will be used to
construct an initial population. Population projections will be
simulated using a commercial spreadsheet and the fecundity and
survival schedules developed from the carcass data. The
performance of the simulated population will be compared to data on
the general demographic characteristics of the PWS population

230

available from past and current telemetry studies. These
comparisons will suggest adjustments to the fecundity and survival
schedules and possibly incorporation of density dependent
mechanisms into the model. A series of simulation experiments will
then be conducted to explore possible recovery patterns of the PWS
population following the mortality event caused by the oil spill.

BIBLIOGRAPHY

Ballachey, B.E., D.P. Evenson, and R.G. Saacke. 1988. The Sperm
Chromatin Structure Assay: Relationship with alternate tests
of semen quality and heterospermic performance of bulls. Jnl.
Androl. 9(2):109-115.

Ballachey, B.E., W.D. Hohenboken, and D.P. Evenson. 1987.
Heterogeneity of sperm nuclear chromatin structure and its
relationship to bull fertility. Biol. Reprod. 36:915-925.

Ballachey, B.E., H.L. Miller, L.K. Jost, and D.P. Evenson. 1986.
Flow cytometry evaluation of testicular and sperm cells
obtained from bulls implanted with zeranol. Jnl. Anim. Sci.
63:995-1004.

Bickham, J.W., B.G. Hanks, M.J. Smolen, T. Lamb, and J.W. Gibbons.
1988. Flow cytometric analysis of the effects of low-level
radiation exposure on natural populations of slider turtles
(Pseudemys scrinta). Arch. Environ. Contam. Toxicol. 17:837-
841.

Chapman, D. G. and D. S. Robson. 1960. The analysis of a catch
curve. Biometrics 16:354-368.

Cochran, W.G. 1977. Sampling Techniques. John Wiley and Sons, Inc.
New York, New York. 428pp.

Cole, L.C. 1954. The population consequences of life history
phenomena. Quarterly Review of Biology. 29:103-137.

Eberhardt, L.L. 1985. Assessing the dynamics of wild populations.
J. Wildl. Manage. 49:997-1012.

Eberhardt, L.L. 1988. Using age structure data from changing
populations. J. Applied Ecology 25:373-378.

Eberhardt, L.L. and D.B. Siniff. 1988. Population model for Alaska
Peninsula sea otters. Minerals Management service, 88-0091.
94pp.

Evensonf D.P. 1986. Flow cytometry of acridine orange stained
sperm is a rapid and practical method for monitoring
occupational exposure to genotoxicants. Pp. 121-132 in:
Monitoring of Occupational Genotoxicants, M. Sorsa and H.

231

Norppa, eds. New York, Alan R. Liss Inc.

Evenson, D.P., R.K. Baer, and L.K. Jost. 1989. Long-term effects
of tr i ethyl eneme lamine exposure on mouse testis cells and
sperm chromatin structure assayed by flow cytometry. Envir.
Molec. Mutagen. 14:79-89.

Evenson, D.P., Z. Darzynkiewicz, and M.R. Melamed. 1980. Relation
of mammalian sperm chromatin heterogeneity to fertility.
Science 240:1131-1133.

Evenson, D.P., P.H. Higgins, D. Grueneberg, and B.E. Ballachey.
1985. Flow cytometric analysis of mouse spermatogenic
function following exposure to ethylnitrosourea. Cytometry
6:238-253.

Evenson, D.P. and M.R. Melamed. 1983. Rapid analysis of normal
and abnormal cell types in human semen and testis biopsies by
flow cytometry. J. Histochem. Cytochem. 31:248-253.

Fienberg, S.E. 1980. The analysis of cross-classified categorical
data. 2nd edition, MIT Press, Cambridge, MA. 198pp.

Garshelis, D.L. 1984. Age estimation of living sea otters. J.
Wildl. Manage. 48:456-463.

Garshelis, D.L. and D.B. Siniff. 1983. Evaluation of
radiotransmitter attachments for sea otters. Wildl. Soc.
Bull. 11:378-383.

Garshelis, D.L. and J.A. Garshelis. 1984. Movements and
management of sea otters in Alaska. J. Wildl. Manage.
48:665-678.

Garshelis, D.L., J.A. Garshelis, and A.T. Kimker. 1986. Sea otter
time budgets and prey relationships in Alaska. J. Wildl.
Manage. 50(4):637-647.

Hendricks, F.B., P.A. Lambird, and G.P. Murphy. 1969.
Percutaneous needle biopsy of the testis. Fertil. Steril.
20:478-481.

Howard, J.G., M. Bush, and D.E. Wildt. 1986. Semen collection,
analysis and cryopreservation in non-domestic mammals. In:
Current Therapy in Theriogenology, pp. 1047-1053. Ed. D.
Morrow. W.B. Saunders Co., Philadelphia.

Irons, D.B., D.R. Nysewander, and J.L. Trapp. 1988. Prince William
Sound sea otter distribution in relation to population growth
and habitat type. U.S. Fish and Wildlife Service. Unpubl.
report. 31pp.

232

Jameson, R.J. and A.M. Johnson. 1987., Reproductive characteristics
of female sea otters. Unpublished manuscript.

Johnson, A.M. 1987. Sea Otters of Prince William Sound, Alaska.
Unpubl. Report, U.S. Fish and Wildlife Service, Alaska Fish
and Wildlife Research Center, Anchorage, AK.

Kenyon, K.W. 1969. The sea otter in the eastern Pacific Ocean.
North Am. Fauna No. 68. 352pp.

Kreeger, T.J., C; Monnett, L.M. Rotterman, and A.R. DeGange. 1989.
Chemical immobilization of Sea otters (Enhydra lutris).
Unpubl. manuscript.

Lensink, C.J. 1962. The history and status of sea otters in
Alaska. Ph.D. Thesis, Purdue Univ., Lafayette, IN. 186pp.

Leslie, P.H. 1945. On the use of matrices in certain population
mathematics. Biometrika 33:183-212.

Leslie, P.H. 1948. Some further notes on the use of matrices in
population mathematics. Biometrika 35:213-245.

Lotka, A.J. 1907. Relation between birth and death rates. Science,
N.S. 26:21-22.

McBee, K. and J.W. Bickham. 1988. Petro-chemical related DNA
damage in wild rodents detected by flow cytometry. Bull.
Environ. Contam. Toxicol. 40:343-349.

Nseyo, U.0., L.S. Englander, R.P. Huben, and J.E. Pontes. 1984.
Aspiration biopsy of testis: another method for histologic
examination. Fertil. Steril. 42:281-284.

Pollock, K.H., S.R. Winterstein, C.M. Bunch, and P.D. Curtis. 1989.
Survival analysis in telemetry studies: the staggered entry
design. J. Wildl. Manage. 53: 7-15.

Ralls, K., T. Eagle, and D.B. Siniff. 1988. Movement patterns and
spatial use of California Sea otters. In D.B. Siniff and K.
Ralls (eds.) , Population Status of California Sea Otters. OCS
Study MMS 88-0021, USDI, Minerals Manage. Serv. pp. 33-63.

Robson, D.S. and D.G. Chapman. 1961. Catch curves and mortality
rates. Trans. Amer. Fish. Soc. 90:181-189.

Sabbioni, G. and H.-G. Neumann. 1990. Biomonitoring of
arylamines: hemoglobin adducts of urea and carbamate
pesticides. Carcinogenesis 11:111-115.

Sandqvist, C., A. Lukola, and M. Parvinen. 1986. Testicular
aspiration biopsy in evaluation of fertility of mink, Mustela

233 "

vison. Jnl. Reprod. Fertil. 77(2):531-536.

Schneider, K.B. 1973. Age determination of sea otters. Final
Report, Fed. Aid. Wildl. Restor. Proj., W-17-4, W-17-5.
23pp.

Siler, W. 1979. A competing-risk model for animal mortality.
Ecology 60:750-757.

Siniff, D.B. and K. Ralls. 1988. Reproduction, survival and tag
loss in California Sea otters. In D.B. Siniff and K. Ralls
(eds.) , Population Status of California Sea Otters. OCS Study
MMS 88-0021, USDI, Minerals Manage. Serv. pp. 13-32.

Snedecor, G.W. and W.G. Cochran. 1967. Statistical Methods, 6th
Ed. Iowa State University Press,. Ames.

Sokal, R.R. and F.J. Rohlf. Biometry. Second Edition. W.H.
Freeman & Co., San Francisco, CA. 859 pp.

Thomas, J.A., L.H. Cornell, B.E. Joseph, T.D. Williams, and S.
Dreischman. 1987. An implanted transponder chip used as a
tag for sea otters (Enhydra--lutris). Marine Mammal Science
3(3):271-274.

Thorud, E., O.P.F. Clausen, and T. Abyholm. 1980. Fine needle
aspiration biopsies from human testes evaluated by DNA flow
cytometry. Flow Cytometry IV, pp 175-177,
Universitetsforlaget.

Tornqvist, M., A. Kautiainen, R.N. Gatz, and L. Ehrenberg. 1988.
Hemoglobin adducts in animals exposed to gasoline and diesel
exhausts. 1. Alkenes. Jnl. Applied Toxic. 8(3):159-170.

Van Pilsum, J.F., R.J. Roon, and M.H. Koenst. Electrophoresis of
Blood Proteins. In Medical Biochemistry: Principles and
Experiments, J.F. Van Pilsum and R.J. Roon, Eds. Univ. of
Minnesota Press, 1986.

Wildt, D.E., M. Bush, C. Morton, F. Morton, and J.G. Howard. 1989.
Semen characteristics and testosterone profiles in ferrets
kept in a long-day photoperiod and the influence of HCG timing
and sperm dilution medium on pregnancy rate after laparoscopic
insemination. Jnl. Reprod. Fertil. 86(l):349-358.

Williams, T.D. and D.B. Siniff. 1983. Surgical implantation of
radiotelemetry devices in the sea otter. J. Am. Vet. Med.
Assoc. 183:1290-1291.

234

Wyrobek, A.J., L.A. Gordon, J.G. Burkart, M.W. Francis, R.W. Kapp
Jr., G. Letz, H.V. Malling, J.C. Topham, and M.D. Wharton.
1983. An evaluation of the mouse sperm morphology test and
other sperm tests in non-human mammals: A report of the U.S.
Environmental Protection Agency Gene-Tox Program. Mutat. Res.
115:1-72.

235

MARINE MAMMAL STUDY 6B

Study Title: Pre to Post Spill Comparisons of Sea Otter
Mortality in PWS Following the EVOS

INTRODUCTION

Much of the initial work to determine injury to sea otters caused
by the EVOS focused on readily observable signs of acute injury.
Efforts have since shifted toward studies to identify possible long
term effects due to acute or chronic exposure to hydrocarbons in
the environment. Changes in the characteristics of mortality (ie.
mortality rates, age and sex composition of mortality) from pre to
post spill time periods may be indicative of reduced viability of
sea otters exposed to oil or hydrocarbon residues in the
environment.

Work by Kenyon (1969) and Johnson (1987) documented patterns of
mortality for sea otter populations within areas at various stages
of reoccupation. Their findings indicate extremely low levels of
mortality for prime age otters.

The Green Island area in southwestern PWS has a well established
otter population and is within the oil spill zone. Johnson's study
provides 10 years of base line mortality data for this area as well
as 10 years of mortality data for the more recently established
populations in the non-oiled, northeastern portion of PWS. Using
the same beach survey methods as Johnson (1987) this study will
seek to determine if the overall characteristics of mortality have
changed from pre spill levels for both oiled and non-oiled
habitats.

OBJECTIVES

A. Test the hypothesis that post-spill levels of mortality
(number of carcasses per linear kilometer of beach surveyed)
are not different from pre-spill levels of mortality in PWS.

B. Test the hypothesis that the proportion of prime age carcasses
found on beaches in post-spill surveys is not significantly
different from proportions found in pre-spill beach surveys in
PWS.

C. Test the hypothesis that the proportion of female carcasses
found on beaches in post-spill surveys is not different from
proportions found in pre-spill beach surveys in PWS.

236

METHODS

Sampling Methods

For valid comparisons, beaches surveyed and methods used will be
the same as those used by Johnson (1987). Treatment beaches to be
surveyed will include those on Green Island, Little Green Island,
Channel Island, and the barrier islands northwest of Gibbon
Anchorage on Green Island. Control beaches will include'those in
the Hell's hole, Olsen Bay area of Port Gravina in the northeast
portion of PWS. These beaches will be walked once during April or
May before summer revegetation occurs which may hide old carcasses
washed high on the beach.

Skulls will be taken from carcasses and a tooth extracted for aging
(Garshelis 1984). Any fresh carcasses collected will be necropsied
as soon as possible and tissue samples for toxicology and
histopathology will be collected. Badly decomposed carcasses or
partial remains may have no evidence indicating the sex of the
individual. In these cases, if a canine tooth is present and the
carcass is that of an adult, sex may be determined by canine
diameter (Johnson 1987, Lensink 1962).

All teeth will be sectioned and prepared according to standard
procedures. Teeth will be read (aged) separately by two
experienced readers with no knowledge of where the tooth was
collected or other information on the carcass. Necropsies will be
performed and histopathology samples will be prepared and analyzed
according to standard protocols.

DATA ANALYSIS

Levels of mortality for a given year using beach survey data are
influenced by a number of variables (ie. weather and current
patterns, yearly changes in otter distribution and abundance) and
are variable (2 to 34 carcasses found in any one year on Green
Island area beaches). However multiple years data with associated
variance will provide a basis for comparing pre and post spill
mortality levels. To do so, only those beaches providing at least
five years of pre spill data will be resurveyed for comparisons.
A mean number of carcasses per kilometer of beach will be
calculated for each pre spill year in both areas of PWS and for
post spill data as they are collected.

The proportion of prime age otter carcasses will be calculated for
each year. Prime age in this study refers to those age groups with
uniformly high survival rates as measured by pre spill data. Based
on Johnson (1987), prime age are animals between 2 and 8 years old
in the Green Island area and those between 2 and 10 years in the
Port Gravina area.

The proportion of female otters will be calculated for each year.

237

Changes in these proportions could reflect changes in the
proportions of males and females in the area due to
immigration/emigration or initially high mortality of one group at
the time of the spill. Changes may also reflect differential
levels of chronic mortality between sexes due to unequal levels of
susceptibility to hydrocarbon toxins or unequal levels of exposure
to toxins because of spatial segregation.

These three variables will be analyzed separately for the two
areas. Pre spill data represent the control in this study and post
spill data represent the treatment observations. Analysis will be
a t-test using years as replicates for each dependent variable. In
the case of the first post spill year's analysis the variance will
be estimated entirely from pre-spill data.

The most sensitive indicator of abnormal change in mortality will
be the proportion of prime age carcasses found. This variable is
not influenced by many of the confounding variables associated with
the other two and a significant change in this parameter is the
most meaningful biologically.

A summary of Johnson's pre spill mortality data for the Green
Island area shows the proportion of prime age carcasses relative to
total carcasses found on beaches to range from 0.0 to 0.28 with a
9 year average of 0.12 (n = 163 carcasses, SD = 0.094). Assuming
post spill variability to be the same, a proportion of 0.32 or
greater in the first post spill year would represent a significant
increase in post spill, prime age mortality at a = 0.20. A
proportion of 0.51 would represent a significant increase at a
0.05.

BIBLIOGRAPHY

Garshelis, D. L. 1983. Age estimation of living otters. J. Wildlife
Manage. 48(2):456-463.

Johnson, A. M. 1987. Sea ot ters of Prince William Sound, Alaska
Unpublished Report, U.S. Fish and Wildlife Service, Alaska
Fish and Wildlife Research Center, Anchorage, AK.

Kenyon, K. W. 1969. The sea otter in the eastern Pacific ocean.
North Amer. Fauna 68. 352 pp.

Lensink, C. J. 1962. The history and status of sea otters in
Alaska. Ph.D. Thesis, Purdue Univ. 188 pp.

238

MARINE MAMM.AL STUDY NUMBER 6C

Study Title: A Drif t Study To Assess The Fate and Recovery of
Sea Otter Carcasses In PWS.

INTRODUCTION

Four hundred and ninety sea otter carcasses were recovered f rom PWS
during the EVOS response. Based on information from the
rehabilitation ef fort and the recovery of carcasses throughout the
spill zone, it is likely that more sea otters were killed in PWS
than in any other area affected by the oil spill. There are no
data to indicate what proportion of the sea otter carcasses from
PWS stayed within or drifted out of PWS and were lost or recovered
elsewhere.
OBJECTIVE

Determine whether simulated sea otter carcasses (floats) deployed
in PWS remain in or drift out of the Sound.

METHODS

Thirty simulated sea otter carcasses (floats) will be designed for
the study. Design of the floats is crucial because the float must
have drift characteristics similar to a sea otter carcass. Each
float will be marked with a visible tag containing the address and
phone number of the FWS should one of the floats be recovered. In
addition, each float will contain a small radio transmitter with an
external whip antenna that has an operating life 40-50 days.
Floats will be deployed by boat in sea otter habitat affected by
the oil spill in PWS. Ten floats will be deployed in PWS in three
consecutive releases. If feasible, deployment of ten simulated
otter carcasses will be concurrent with the ongoing drift study,
"An Assessment of Damage to Seabirds in PWS and the Western GOA
Resulting from the EVOS.11

Following release, the simulated carcasses will be relocated using
fixed-wing aircraft outfitted with 2 4-element yagi antennas and a
telemetry receiver. Up to 8 hours per day for 25 days will be
devoted to tracking the floats during the drift experiment. Within
PWS fixed-wing aircraft will fly parallel to the shoreline at an
elevation of 1500 ft. A systematic search pattern will be
developed for offshore areas within the Sound. Outside of the PWS
the aircraft will fly at 3,000 ft over open water following a
systematic search pattern. The aircraft periodically will search
the coastline of the KP. Relocations will be marked directly on
detailed charts or topographic maps of the study area and entered
into the computer as UTM coordinates.

Information on the recovery locations of sea otter carcasses from
Prince William Sound and the Kenai Peninsula during the oil spill

239

response will be used to estimate recovery rates.

Recovery locations f or simulated sea otter carcasses deployed in
PWS will be stratified into two groups: recoverable (on shore or
within 200 m of shore) or nonrecoverable (> 200 m off shore in the
Sound or outside of PWS) . Floats that remain offshore or that are
not found will be considered unrecoverable. The overall recovery
rate of simulated carcasses in PWS will be estimated as the
proportion of carcasses that drift into habitat in which they could
have been recovered.

BUDGET: FWS

6A 6B 6C Total

Salaries $ 389.8 5.0 0 $ 394.8
Travel 41.0 2.6 0.8 44.4
Contractual 422.3 1.0 24.7 448.0
Commodities 153.6 1.6 0.6 155.8
Equipment 53.8 0.8 7.4 62.0

Total $ 1,060.5 11.0 33.5 $1,105.0

240

MARINE MAMMAL STUDY NUMBER 7

Study Title: Assess the Fate of Sea Otters Oiled and
I Rehabilitated as a Result of the EVOS

Lead Agency: FWS

INTRODUCTION

The capturej cleaning, and care of sea otters contaminated with oil
during the EVOS oil spill has been the focus of considerable
attention and effort. During the initial weeks of the spill, the
health of many of the sea otters brought to the cleaning center in
Valdez was severely compromised by exposure to the oil, and many
died. The chronic effects of exposure to oil on otters which
survived and were released into the wild are unknown. Given that
the underlying goal of the rehabilitation program was to release
sea otters back into the wild as functioning members of their
environment, it was important that a long-term evaluation of the
process be undertaken. This information will assist in guiding
future cleaning operations for sea otters as well as aiding our
understanding of how exposure to crude oil from the EVOS affected
sea otters. Forty-five rehabilitated sea otters were implanted
with radio transmitters in summer 1989 and released shortly
thereafter in eastern PWS. Radio tracking of those individuals is
ongoing. Twelve of the instrumented sea otters are known dead;
several others are missing. Preliminary evidence suggests that
mortality of rehabilitated sea otters is higher than a sample of
animals instrumented in eastern PWS as a control.

OBJECTIVES

A. Test the hypothesis that survival of sea otters that underwent
oiling, cleaning, rehabilitation and release is not different
from that of sea otters that were not affected by the oil
spill.

B. Test the hypothesis that survival of rehabilitated sea otters
that re-enter oiled areas does not differ significantly from
that of rehabilitated sea otters that remain in oil free

areas.

C. Test the hypothesis that reproductive rate of female sea
otters that underwent oiling, cleaning and rehabilitation does
not differ significantly from that of female sea otters that
were not affected by the oil spill.

D. Document the movements of rehabilitated sea otters relative to
impacted habitat in western PWS and the KP.

241

.METHODS

Sampling Methods

Thirty-six of the instrumented sea otters were from the KP, either
from the Valdez, Seward, or Homer sea otter facilities. The
remaining nine implanted otters were from PWS. comparisons of
effects of severity of oiling as well as the effects of fresh crude
oil vs. weathered crude oil on survival of sea otters released back
into the wild were intended; however, because only nine sea otters
oiled in PWS with fresh crude oil were suitable for implantation,
analyses comparing the effects of oil type and degree will be
limited.

Forty-five rehabilitated' sea otters were instrumented prior to
release. A sample of 50 female sea otters from non-oiled areas
instrumented as part of Marine Mammal Study Number 6 (M/M) will
serve as control. Twenty-three control animals were instrumented
in eastern PWS during fall 1989. Additional animals will be
instrumented this spring as part of the control group. An existing
population of 58 radio-marked sea otters in the vicinity of the
release sites for the rehabilitated sea otters is also available
for comparison. For specifications on the transmitters and the
implant protocol see the study proposal for M/M Study 6.

Using standard sample size calculations for testing the difference
between two proportions (Snedecor and C 'ochran, 1967; p. 221), a
sample size of 45 gives a 75% chance of finding a significant
difference at alpha = 0.20, given that the population proportion
changes by .2 from an initial value of .5.

All sea otters used in this study were released back into the wild
in eastern PWS as recommended in the FWS Draft Release Strategy
Plan for Rehabilitated Sea Otters. The release sites were not
directly affected by oil from the spill and were occupied by sea
otters prior to the release. Male sea otters were released in a
male area in Nelson Bay. Females were released in a female area in
Sheep and Simpson bays. The release sites represent a short to
moderately long translocation for sea otters captured in western
PWS and along the KP.

After the initial 20 day monitoring period, the frequency of
relocation has depended upon weather, and the sex, age,
reproductive status, and whereabouts of the marked animals.
Relocations of all animals were intended to be made at least
biweekly but for many of the animals, movements have been erratic
and unpredictable, therefore they have been difficult to relocate.
During the pupping season and shortly following that period
attempts will be made to locate reproductively mature females at
least weekly although this will be impossible for some animals.
Those animals captured on the KP and that have returned to the KP

242

have only been relocated occasionally.

Attribute data for each relocation, including group size, number of
pups in the group, whether or not the focal animals had a pup, and
behavior of focal animal, are also collected. At the end of each
workday, locations of each otter are entered directly into a
computer along with the attribute data. Periodically, those data
are transferred to Anchorage, where they will be analyzed using
geoprocessing software and statistical software, including SAS. In
addition to location fixes, qualitative assessments of the health
status of each rehabilitated sea otter are being made.

Marked sea otters that have died following release are collected as
soon as possible. Carcasses that are in suitable condition are
necropsied. Tissue samples will be taken for histology and
toxicology.

Additionally, histopathological samples taken from sea otters that
died in the rehabilitation centers will be analyzed. Results will
be important for analyzing the effectiveness of current
rehabilitation techniques, for assessing the need for changes in
these techniques and for providing other information important for
future sea otter restoration efforts.

Estimates of survival rate of the rehabilitated sea otters will be
calculated for comparison with those of control animals and other
populations of sea otters within and outside of Alaska.

Survival estimates will be obtained by the product limit method and
differences in survival patterns will be tested with log-rank tests
(Pollock et al., 1989).

Reproduction and movements of rehabilitated, implanted and control
sea otters will also be examined in the proposed study. The
reproductive rate will be estimated for each population by counting
the number of females in each group (rehabilitated and controls)
observed with a pup, and dividing this value by the total number of
females observed. However, given the small sample sizes of
females, sufficient data to examine reproductive rates in a
rigorous statistical sense may not be available. Several measures
will be used in the analysis of movements including distance
between successive locations, minimum convex polygon, and distance
between extreme locations (Garshelis and Garshelis 1984, Ralls et
al. 1988). Since many of the sea otters that make up the radio
telemetry portion of this study were originally captured on the KP
and in western PWS, the release sites represented a short to
moderate translocation. The influence of translocation distance on
movements of sea otters will be examined by regressing
translocation distance on the daily rate of movement or on
dispersal distance. Dispersal distance is defined as the distance
from point of release to the location of the translocated sea
otter's first activity center at which it becomes sedentary.

243

BIBLIOGRAPHY

Garshelis, D.L. and J.A. Garshelis. 1984. Movements and
management of sea otters in Alaska. J. Wildl. Manage.
48:665-678.

Kleinbaum, D.G. and L.L. Kupper. 1978. Applied Regression
Analysis and Other Multivariable Methods. Duxbury
Press/Wadsworth Publishing Co., Belmont, California.

Pollock, K.H., S.R. Winterstein, C.M. Bunch, and P.D. Curtis. 1989.
Survival analysis in telemetry studies: the staggered
entry design. J. Wildl. Manage. 53: 7-15.

Ralls, K., T. Eagle, and D.B. Siniff. 1988. Movement
patterns and spatial use of California Sea otters. In
D.B. Siniff and K. Ralls (eds.), Population Status of
California Sea Otters. OCS Study MMS 88-00211 USDI,
Minerals Manage. Serv. pp. 33-63.

Siniff, D.B. and K. Ralls. 1988. Reproduction, survival and
tag loss in California Sea otters. In D.B. Siniff and K.
Ralls (eds.) , Population Status of C@lifornia sea otters.
OCS Study MMS 88-0021, USDI, Minerals Manage. Serv. PP.
13-32.

Snedecor, G.W. and W.G. Cochran. 1967. Statistical Methods. 6th
Edition. The Iowa State University Press, Ames, IA.

Sokal, R.R. and F.J. Rohlf. 1981. Biometry. 2nd Edition. W.H.
Freeman and Co., San Francisco.

BUDGET: FWS

Personnel $ 42.5
Travel 5.0
Contractual 87.6
Commodities 11.0
Equipment - 0.9

Total $ 147.0

244

TERRESTRIAL MAMMAL INJURY ASSESSMENT

Terrestrial mammals are an important part of the ecosystem in the
area affected by the EVOS. A wide variety of species are present,
many of which use intertidal habitats that were heavily impacted by
oil. They are important to humans for recreational viewing, sport
and subsistence hunting, and commercial and subsistence trapping.

In the 1989 damage assessment plan, 19 terrestrial mammal species
were identified as potentially being impacted by oil. Of those,
five were selected for intensive field study and nine were chosen
for general assessment only. In 1990, intensive damage assessment
studies in the field will be continued for three species: deer,
river otter, and brown bear. A literature review will also be
completed to gather information on the importance of intertidal
habitat use by black bear.

The deer study will focus on detection of lethal injury during the
spring of 1990 when deer are concentrated on beaches and,
therefore, most likely to come in contact with oil. If no
mortality attributed to oil is detected, this project will be
discontinued. The river otter and brown bear studies will explore
both lethal and sublethal injury. The river otter study includes
examination of animals found dead and assessment of oil impacts on
populations, food habits, and habitat use. The brown bear project
will examine mortalities and assess impact on reproduction and
population density

A laboratory study to determine the influence of hydrocarbons on
reproduction in ranched mink will also be conducted. It commenced
in 1989 and will end in July 1990. The work will provide
information on whether sublethal doses of hydrocarbons will
influence reproduction in mammals. Mink will provide a model for
other related terrestrial and marine mammal species.

245

TERRESTRIAL MAMMAL STUDY NUMBER I

Study Title: Assessment of the Effect of the EVOS on the Sitka
Black-tailed Deer in PWS and the Kodiak
Archipelago

Lead Agency: ADF&G

INTRODUCTION

Sitka black-tailed deer (Odocoileus, hemionus sitkensis) are the
most abundant large mammal on the islands of PWS and the Kodiak
Archipelago. ADF&G wildlife biologists estimate that there are
15,000 to 20,000 deer in PWS and up to 100,000 deer on the Kodiak
Archipelago. In addition to the intrinsic values of this
resource, it also has a substantial economic value to residents
of Alaska.

During late winter and early spring, deer in PWS and Kodiak
usually concentrate on beaches and along a relatively narrow
fringe near the coast (ADF&G 1986). Groups of over 500 deer have
been observed on some beaches. These areas commonly have a
reduced snow depth or are snow-free, and deer forage on
intertidal marine vegetation, coastal sedges, grasses, shrubs,
and herbaceous vegetation in the forest understory (ADF&G 1986).

Hinchinbrook, Montague, and Hawkins Islands contain most of the
deer habitat in PWS. Beaches on Hinchinbrook and Hawkins Islands
generally were not effected by the EVOS, whereas the northern
portion of Montague Island was lightly oiled. Deer also occur in
relatively high densities on some of the other islands in PWS
that were heavily impacted by oil. Deer are abundant throughout
the Kodiak Archipelago. Light to very light EVOS impacts were
reported along most Kodiak beaches, with heaviest concentrations
occurring on the east side of Shuyak Island and along portions of
Uyak Bay on Kodiak island.

oil has affected several types of coastal deer forage, and deer
have been observed feeding on oiled kelp. It is anticipated that
deer will be adversely affected if they consume vegetation that
has been contaminated by oil. Small to moderate amounts of crude
oil consumed by deer and other ruminants may cause, direct
mortality due to disruption of the rumen fermentation process and
aspiration of rumen fluid into the lungs (Rowe et al. 1972).
Sublethal injury also could occur, reducing animal health and
affecting reproduction.

When oil reached beaches where deer were concentrated in late
March/April 1989, snow was already melting in upland areas. Some
deer had begun their annual spring movements away from the coast
and into higher elevations. This fact, coupled with the
substantial increase in human activity on beaches soon after the

246

spill undoubtedly reduced the potential for deer exposure to oil.
However, the increased human activity probably pushed some deer
away from preferred beach feeding areas prematurely, forcing them
into areas with deeper snow. This would cause accelerated
mortality because the energy reserves of deer are at an annual
low state during late winter/early spring. Unfortunately,
quantification of such additional indirect "natural" mortality
was not possible. The winter and spring of 1989-90 may be the
best time to investigate potential impacts of oil on wintering
deer. If winter temperatures and snowpacks are within normal
limits, deer will concentrate along beaches sometime in the mid-
December to mid-February period and human activity will be far
less than it was from late March through late fall 1989.

OBJECTIVES

A. Test the hypothesis that deer on heavily oiled islands have
tissues and rumen contents that have been contaminated by
oil.

B. Test the hypothesis that deer found dead have rumen contents
in their lungs.

C. Estimate the number of dead deer per unit area on both a
heavily oiled and a non-oiled island in the Sound, if
substantial numbers of deer concentrate on oiled beaches in
the late winter of 1989-90, and there is evidence to suggest
that some of these deer are dying from oil contamination.

METHODS

A sample of live deer has been collected and examined for
hydrocarbon contamination. Deer were collected in areas near
beaches in PWS and the Kodiak Archipelago that had been affected
by oil. These collections occurred during various periods
throughout the year.

Deer were collected on Afognak Island on 7 April 1989, prior to
any reported EVOS impacts in the area. Tissue samples were
collected, wrapped in Reynolds aluminum foil and frozen for
histopathological analysis.

Deer near oiled beaches on Shuyak Island were taken on 4 May
1@89. Necropsies were conducted by a pathologist immediately
after collection and tissue samples were collected for
histopathology and hydrocarbon analysis. Additional deer on
Shuyak and from PWS were taken near oiled beaches from-* 31 May
through 15 June 1989. Gross necropsies were performed in the
field,. by wildlife biologists and tissue samples were collected
for histopathology and hydrocarbon analysis.

247

Live deer were collected in oiled areas or areas that are known
to be heavily used by deer hunters in PWS and the Kodiak
Archipelago during August and September 1989. Gross necropsies
were performed in the field by wildlife technicians and tissue
samples were collected for future histo3athology and hydrocarbon
analysis. Small amounts (approx. 2 cm ) of liver and skeletal
muscle from each animal were boiled, smelled and tasted in an
attempt to detect obvious evidence of oil contamination.

Additionally, several samples were made available for analysis
from dead deer collected by ADF&G staff on or near oiled beaches
in PWS in April, and from deer found dead and turned in by
various workers associated with the EVOS throughout the spring
and summer of 1989.

Flights will be made over selected beaches in PWS as often as
possible, but not more frequently than every two weeks during the
winter of 1989-90. If information obtained during these flights,
or observations from individuals in the field indicate that deer
are concentrating on oiled beaches, additional deer collections
will be made and searches conducted for dead deer in those areas.
If deer behavior, gross necropsy, and examination of lungs
suggest that deer are dying from oil, systematic surveys will be
conducted on a heavily oiled island and a control island of
similar size, topography, and deer density. The carcass of each
dead deer that is found will be examined in the field by a
biologist and recent mortalities will be examined by a
pathologist. Pellet group counts on each island will be done to
correct for different deer densities (Kirchhoff and Pitcher
1988a, Kirchhoff and Pitcher 1988b). If it is assumed that deer
carcasses are distributed in a "patchy fashion", a systematic
sampling scheme should be close to optimal (Snedecor and Cochran
1980) and this procedure will provide an estimate of deer
mortality per unit area on each island.

Throughout all phases of this study we will attempt to identify
potential alternative methods and strategies for restoration of
lost use, populations, or habitat if injury is identified. The
final report will include a listing of suggested ways to address
long-term restoration projects.

Tissues which were collected will be analyzed as outlined by the
EVOS Histopathological Technical Group and the Hydrocarbon
Technical Group. All statistical analyses will be performed at
an alpha level of 0. 05. Sample sizes will be used that are
adequate to detect at least 1 deer affected by oil contamination
with a given percentage of certainty, for varying proportions of
the population contaminated by hydrocarbons. These sample size
calculations are based on a binomial distribution (Mendenhall,
Schaffer and Wackerly 1981), and assume that the sample size is
very small compared to the population total. This study will
assume that at least 10% of the deer population was affected by

248

EVOS; therefore, a total sample of at least 29 deer will be
collected to be 95% certain of collecting at least one deer
affected by hydrocarbons.

BIBLIOGRAPHY

Alaska Department of Fish and Game. 1986. Alaska Habitat
Management Guide - Life Histories and Habitat
Requirements of Fish and Wildlife. Ak. Dept. Fish and
Game.' 763pp.

Kirchhoff, M.D. and K.W. Pitcher. 1988a. Evaluation of
methods for assessing deer population trends in
southeast Alaska. Fed. Aid Wildl. Res. Prog. Rep. W
-22-6. Job 2.9. 13pp.

Kirchhoff, M.D. and K.W. Pitcher. 1988b. Deer pellet-group
surveys in southeast Alaska, 1981-1987. Fed. Aid
Wildl. Res. Prog. Rep. W-22-6. Job 2.9. Objective 1.
13pp.

Mendenhall, W., R.L. Schaffer, and D.D. Wackerly. 1981.
Mathematical Statistics with Applications. Duxbury
Press. Boston, Mass. 686 pp.

Rowe, L.D., J.W. Dollahite, and B.J. Camp. 1972. Toxicity
of two crude oils and of kerosine to cattle. J. Amer.
Vet. Med. Asso. 162(2):61-66.

Snedecor, G.W. and W.G. Cochran. 1980. statistical
Methods; Seventh Edition. Iowa St. Univ. Press.
Ames, Iowa. 507 pp.

BUDGET: ADF&G

Personnel $ 60.6
Travel & per them 4.0
Contracts & Services 52.0
Supplies 8.0
Equipment 0.0

TOTAL $ 124.6

249

TERRESTRIAL MAMMAL STUDY NUMBER 2

Study Title: Review of Literature on Intertidal Habitat Use by
Black Bear

Lead Agency: ADF&G

INTRODUCTION

There is a dense population of black bear (Evarctos americanus) in
PWS. They are omnivorous, opportunistic feeders near the top of
the food chain. Black bears are known to feed in intertidal areas
and, therefore, have the potential to contact oil directly by
eating sludge washed ashore, grooming oiled hair, eating
contaminated intertidal organisms, or scavenging carcasses of
mammals and birds killed by oil offshore and deposited on beaches.

A study of the impact of the EVOS on black bear populations was
proposed in t 'he 1989 damage assessment plan. That effort proved
not feasible, given the logistical difficulties of bear capture in
the densely forested habitat of PWS. The literature search
proposed for 1990 will provide helpful background information for
evaluating the need for a revised detailed population study.

OBJECTIVE

Determine importance of intertidal habitat use by black bear to
establish the likelihood of significant impact due to beached oil.

METHODS

Black bear literature will be searched to identify and retrieve any
information on the importance of intertidal habitat use.. The final
product will include a list of citations accompanied by abstracts
of each paper and a summary that includes relevant information from
all sources.

BUDGET: ADF&G

This study will be a contract for a period March 1, 1990 - February
28, 1991 and will not exceed $10,000.

Contract $ 10.0

Total $ 10.0

250

TERRESTRIAL MAMMAL STUDY WUMBER 3

Study Title: Assessment Of The Effect Of The EVOS On River
Otters In PWS

Lead Agency: ADF&G

INTRODUCTION

River otter (Lutra canadensis) populations in PWS rely on
intertidal and subtidal environments for food. Studies of
similar coastal populations in southeastern Alaska documented
that marine fishes, crabs, and other invertebrates dominated food
habits (Larson 1983, Woolington 1984). Because critical habitat
for this species was heavily contaminated by oil, otter
populations are at risk by direct contact with oil or by
environmental changes to other habitat components. Data on
density prior to the oil spill are lacking, but river otters were
probably abundant. The goal of this study is to determine if the
EVOS will have measurable effects on these populations. The
approach is to 1) examine carcasses to determine direct effects
of oil, 2) compare pre- and post-spill river otter dietary
information from scats, 3) validate the use of a control area and
then, 4) compare population density and various biological
aspects between oiled and control study areas.

Necropsy and tissue samples obtained from otter carcasses
recovered from oiled beaches will provide information on possible
short-term impacts. Magnitude of short-term loses cannot be
measured directly because the proportion of recovered carcasses
is unknown.

This study will use parallel data collected in a control area
(Esther Passage) and an area heavily contaminated by oil (Knight
Island) to test for impacts on river otters. Radio telemetry,
rates of fecal deposition, food habit analysis, home range
determinations, and analyses of habitat selection by otters will
provide population characteristics, trends, and indexes for
comparing the two areas. Additionally, necropsy and tissue
analyses of animals collected outside of the study areas will
provide data on presence of hydrocarbons and their long-term
effects on individual animals. Results f rom the study on the
effects of hydrocarbons on captive mink (Terrestrial Mammal Study
Number 6) will provide the context for interpreting hydrocarbon
levels in river otters.

251

OBJECTIVES
Direct Effects

Al - Determine cause of death for river otter recovered from
oiled areas via necropsy and histopathological
procedures.
A2 - Test (a = 0. 05) for higher hydrocarbon levels in river
otter in oiled versus unoiled areas.

Population Change

Bi - Estimate population sizes of river otter within 10% of
the true value 95% of the time, on representative oiled
and control study areas using mark-recapture methods
and test (a = 0.05) for lower population levels in
oiled versus control areas.
B2 Estimate the rate of fecal deposition within 10% of the
true value 95% of the time for river otter. This rate
will be used as an index to population size to test (a
=' 0.05) for lower rate of deposition in oiled versus
control study areas.
B3 Test (a = 0.05) for lower survivorship of river otter
in oiled versus control study areas.

Food Habits

B4 - Test (a = 0.05) for differences in food habits of river
otters before and after the oil spill on the oiled
study area.
B5 - Test (a = 0.05) for differences in food habits of river
otters on oiled and control study areas.

Habitat Use

B6 - Test (a = 0.05) for differences in activity patterns
(foraging) of river otters between oiled and control
study areas. (limited funding and man power may not
allow data collection for this objective)
B7 - Use homerange size and use patterns to test (a 0.05)
for differences in habitat selection in river otters
between oiled and control study areas'.

METHODS

The initial impact assessment concentrated on locating two study
areas (control vs. oil impacted) with comparable numbers of
active latrine sites for mink and river otters. Each site was
given a unique name, plotted on a map and field marked for future
relocation, and a site d rawing with a rough description made in a
field notebook. Sites were cleaned of all scats and then
revisited five times between June and September 1989, to obtain

252

data on continued use. After the initial visit the number of
scats present were recorded in addition to scat collection for
later analyses.

information obtained during the 1989 initial study for impact
response was used in developing the study design for this
project. The 59 latrine sites in the control area and 57 sites
in the oiled area will be the focus of efforts to live capture
otters. Most of these sites will also be utilized to provide
scat samples for the study. With qualifications, information
obtained on otter densities, habitat selection, and population
response to oil will be available for extrapolation to other
areas of PWS. Standard operating procedures will be developed
for each segment of the long range study to insure data validity.

The following are methods for collecting data by objective.

Direct Effects.

Al- Necropsy and histopathology will be performed according to
standard procedures.
A2 Up to 20 additional animals may be collected outside the
study area to provide hydrocarbon and histological samples.
Necropsy and similar tissue analysis will continue to be
made on dead otters found throughout the entire area
impacted by the oil.

Population Change

Bi- River otters will be live trapped at latrine sites in the
control and oiled study areas. Modified Hancock live traps
and drugging boxes to hold otters, as described by Me1quist
and Hornocker (1979), will be used. Weather permitting, all
traps will be monitored daily. All traps will be equipped
with a transmitter that signals a sprung trap. Animals will
be held only as long as necessary to complete the marking
process and provide for their recovery from surgery.
Animals will then be. released at their original capture
site.

Techniques for implantation of radio transmitters will be as
described by Woolington (1984). Surgery will be done by a
licensed veterinary/biologist or project personnel
specifically trained in the technique. Each transmitter is
equipped with a "mortality model' so the fate of individual
animals can be determined.

Radioisotope implants in otters will be used to estimate
population density in the oiled and control study areas
using a mark-recapture method. Marking will be by
implantation of radio-labeled, polylactic acid (PLA) tablets
to provide a long lasting tracer that can be detected in
feces (scats) of river otter (Crabtree et al. 1989).
Recoveries of scats from latrine sites will provide the
"recaptures" for analysis. This mark-recapture technique

253

has been employed in carnivore studies (Kruuk et al. 1980),
including river otters (Knaus et al. 1983, Shirley et al.
1988).

Animals instrumented with VHF transmitters will have radio
labeled PLA tablets implanted intra-peritoneally. This
method allowed detection for over 10 months in the scats of
coyotes (Canis latrans) (Crabtree et al. 1989). A gamma
spectrometer will be used to detect and identify radio
labeled scats.

Sampling of latrine sites will provide the "recaptures" for
simple mark-recapture analysis (Seber 1982). Twenty river
otters in each study area will be uniquely marked. A closed
population model will be used, employing radio transmitters
to determine exactly how many marked animals are resident in
the study area while scats are being sampled. Mark-
recapture models for closed populations are well established
(Dennis et al., In Press; Seber 1982). Latrine sites will
be cleared of scats at the start of a sampling period, and
visited every one to two days until a predetermined number
of scats has been collected.

The distribution of marked animals is likely not to be
random, due to the necessity of focussing our capture effort
in locations of high animal abundance. Biases can result if
the recovery of scats is uneven across low and high density
areas within each main study area. A special effort will be
made to randomize the recovery of otter scats to ensure
every scat is equally likely to be collected.

B2_ Rates of fecal deposition will be used as an index to
population size in oiled and control areas. The same
latrine sites used for mark-recapture population estimates
will be used for estimating fecal deposition rates.

B3_ Estimates of survival will depend on data obtained from
otters instrumented with radio transmitters. Data will be
obtained coincidental to data gathered for objectives B1.
B6, and B7-

Food Habits

B4 & B5_ Food habits of river otter will be described from prey
remains in their feces. Such procedures have been used
successfully in past studies (Gilbert and Nancekivell 1982).
A preliminary survey of latrine sites conducted in late
April and early May 1989, located 59 latrines in the control
area and 55 latrines in the oiled area. Feces were
collected at each site and resampled four times.

Scats from river otter will be distinguished from those of
other mammals by their characteristic morphologies (Murie
1954).

254

Latrine sites will be resampled when snow free in late
spring 1990 in the same manner as those collected following
the oil spill. Thereafter, we tentatively plan to collect
feces from latrines 1-2 times/week from June through mid-,
September 1990 on both control and oiled areas.

Laboratory analysis of prey remains in feces of river otter
will follow standard procedures (Bowyer et al. 1983).

Because of differential digestibility of prey and variable
rates of passage through the gut, volumetric measures of
prey remains in mustelid feces are meaningless.
Consequently, the analysis will be confined to the
occurrence of prey items in latrines and will be expressed
in terms of percent of latrines with food items, and percent
of total food item (Bowyer et al. 1983). To assure that
subsamples from a latrine are representative of that site,
all feces from that site will be mixed and a series of
subsamples (about the volume of an individual scat) will be
drawn and analyzed separately. Sampling will continue until
the function between number of prey items and number of
samples becomes asymptotic. All latrines included in the
analysis, however, will contain at least five scats per
sampling period.

Because sample variance is unknown, it is not possible to
specify the total number of samples necessary to adequately
describe food habits at this time. Reduction in variation
of.the mean with increasing sample size (of latrines) will
be monitored for important food items to ensure that all
proportions are estimated within 0.05 of their true value
95% of the time (Kershaw 1964).

Habitat Use

B6- Activity patterns of radio equipped river otters will be
used to test for changes in the availability of prey between
oiled and control study areas. A digital recorder linked to
a radio receiver will be operated to record activity of
otters.

B7_ Data on home range and habitat selection of individuals will
be collected daily using radio-locations of telemetered
animals. Telemetry will be conducted from a small boat, and
the entire coastline of both study-areas (oiled and control)
'will be sampled each day. Because river otter are
distributed immediately along coastal areas (Larsen 1983,
Johnson 1985), telemetry "fixes" will be made over
relatively short distances, and multiple "legs" can be-used
in triangulation. Consequently, error polygons should be
small and biases from animal movements during triangulation
will be minimal. Locations determined via telemetry will be
confirmed visually whenever possible.

255

The time at which a telemetry transect starts will be
randomized each day to help minimize any bias from duel
activities of the mustelids on estimates of home range size
and habitat selection. Further, aerial telemetry will be
conducted as needed to determine locations of individuals
that cannot be located by boat. Telemetry transmitters will
be equipped with a mortality signal that will allow the
speedy recovery of dead animals.

Methods for analyzing data are detailed below for each objective.

Direct Effects

Al- A cause of death will be assigned each mink or river otter
carcass based upon a necropsy report and lab analysis of
tissue specimens. Hydrocarbon levels will be presented for
all usable samples.

A2_ A one-tailed Z test for proportions (Snedecor and Cochran,
1980) will be used to test this hypothesis.

Population Change

B1_ Analysis will follow methods described by Dennis et al., (In
Press) for sampling a closed population with replacement.
Population size and 95% confidence intervals for both
control and oil affected areas will be estimated. A one-
tailed' Z statistic will be used to determine if the
population density is lower in the oiled area versus the
control area. This test assumes that the population
estimates are normally distributed and have equal variance
(Seber 1982).

B2_ Differences in rates of scat deposition between oiled and
control study areas will be tested (a = 0.05) with a single
factor covariance analysis model (Neter et al. 1983). The
response variable will be rate of scat deposition and the
covariate will be the number of latrine sites. Main
effects will include oiling and months of study. Since a
one-tailed hypothesis is being tested with regard to the
oiling main effect, the critical region for this section of
the ANOVA table will be one-tailed. If variances are not
homogeneous, either a ranked procedure will be employed or
the data will be transformed to obtain homogeneous variance
or normality.

B3 Estimation and analysis of survival distributions for radio
marked individuals will follow standard procedures (Pollock
et al. 1989). Model assumptions include a random sample of
animals, that survival times are independent for different
animals, and that censoring mechanisms are random.

256

Food Habits

B4 and B5_ Statistical analysis will include only food
items that compose at least 10% of the diet. Comparisons of
food habits between oiled and control areas and among months
will be made with the Quade test including multiple
comparisons of food items (Conover 1980).

Habitat Use

B6- It is hypothesized that if availability of forage fishes in
the subtidal zone were reduced due to oil, otters would
spend more time foraging to obtain a diet equivalent to that
in the control area. Because study areas were selected that
contained similarly high populations of otters, it is
presumed that both otters and their food were abundant prior
to the oil spill. The oil spill may have reduced both river
otters and their prey. Consequently, the foraging
activities of otters could be expected to change with both
their population size and that of their prey.

Although this procedure will allow assessment of a reduction
in otters or a reduction in their prey, it will not detect a
simultaneous reduction in both.

Differences in activity of river otters (stratified by sex
and age class) between oiled and unoiled study areas will be
tested (a = 0.05) with a two-tailed Mann-Whitney test
(Conover 1980: 216).

B7_ The procedures of Swihart and Slade (1985a,b) will be used
to correct for auto correlation among home range locations
and to determine the time interval to achieve independence
of observations. An adequate number of relocations to
assess the seasonal home range of an individual will be
determined by obtaining an asymptotic relationship between
home range size with increasing number of relocations. once
the proper time interval and sample size have been
determined, the method of Dixon and Chapman (1980) will be
used to calculate 25%, 50%, 75% and 95% isoclines of home
range use.

Isoclines of home range use will be overlaid on detailed
maps of coastal habitats. The 95% use isocline will be
employed to determine the habitats available for a
particular animal. Proportional weighing by 25%, 50% and
75% isoclines within each habitat will determine use. Thus,
habitat use and availability will allow a determination of
habitat selection for each telemetered individual. Testing
for differences in habitat selection (rather than use)
between oiled and control areas is essential because a
difference in habitat use may occur as a result of
differential availability of habitats independent of effects
of oiling. A knowledge of habitat selection by river otters

257

is essential for extrapolating from our study areas to
effects on habitat oiled in other areas. Consequently,
habitat selection will be inferred from a significant
difference (P < 0.05) in use and availability matrices
compared simultaneously with Hotelling's T 2 statistic; a
posteriori comparisons of individual habitat types will be
accomplished using Bonferroni multiple tests (Johnson and
Wichern 1988:188). similarly, comparisons of habitat
selection in oiled and control areas will be made with a
multivariate analysis of variance (MANOVA) again using
Bonferroni multiple contrasts.

BIBLIOGRAPHY

Adorjan, A.S. and G.B. Kolenosky. 1969. A manual for the
identification of hairs of selected Ontario mammals.
Ontario Dept. Lands and Forests, Res. Rep. (Wildlife)
No. 901 64pp.

Bowyer, R.T. and K.D. Curry. 1983. Use of a roller press
to obtain cuticular impressions of guard hairs on
acetate strips. J. Mammal. 64:531-532.

Bowyer, R.T., S.A. McKenna, and M.E. Shea. 1983. Seasonal
changes in Coyote food habits as determined by fecal
analysis. Amer. Midland Nat. 109:266-273.

Bunham, K. P. and W. S. Overton. 1978. Estimation of the
size of a closed population when capture probabilities
vary among animals. Biometrika 65:625-633.

Chao, A. 1989. Estimating population size for sparse data
in capture-recapture experiments. Biometrics
45(2):427-438.

Conover, W. J. 1980. Practical nonparametric statistics.
John Wiley & Sons, New York, 493pp.

Crabtree, R.L., F.G. Burton, T.R. Garland, D.A. Cataldo, and
W.H. Rickard. 1989. Slow-release radioisotope implants
as individual markers for carnivores. J. Wildl. Manage.
53: 949-955.

Davis, D.E. and R.L. Winstead. 1980. Estimating the numbers
of wildlife populations. Pp. 221-245 in S.D.
Schemnitz, ed. Wildlife Management Techniques Manual.
Fourth Ed. The Wildl. Soc., Washington, DC.

Day, M. G. 1966. Identification of hair and feather
remains in the gut and feces of stoats and weasels. J.
Zool. (Lond.) 148:201-217.

Dennis, B., R.L. Crabtree, and E.V. Eartan. In Press.

258

Statistical methods for closed population estimation
using radioisotope tagging. J.Widl. Manage.

Dixon, K.R. and J.A. Chapman. 1980. Harmonic mean measure of
animal activity. Ecology 61:1040-1044.

Fleiss, J.L. 1973. Statistical methods for rates and
proportions. John Wiley & Sons, New York, 223pp.

Gilbert, F.F. and E.G. Nancekivell. 1982. Food habits of
mink (Mustela vision) and otter (Lutra canadensis) in
northeastern Alberta. Can. J. Zool. 60:1282-1288.

Hall, R.E. and K.R. Kelson. 1959. Mammals of North
America. Ronald Press, New York, 1983pp.

Johnson, R.A. and D.W. Wichern. 1988. Applied multivariate
statistical analysis. Prentice Hall, New Jersey,
606pp.

Kershaw, K.K. 1964. Quantitative and dynamic ecology.
Edward Arnold, London, 1983pp.

Knaus, R.M., N. Kinler, and R.G. Linscombe. 1983. Estimating
river otter populations: the feasibility of 65 Zn to
label feces. Wildl Soc. Bull. 11:375-377.
Kruuk 15 H., M. Gorman, and T. Parrish. 1980. The use of
zn for estimating populations of carnivores. Oikos
34:206-208.

Larsen, D.N. 1983. Habitats, movements, and foods of river
otters in coastal southeastern Alaska. Unpubl. M. S.
Thesis, Univ. of Alaska Fairbanks, 149pp.

Mayer, W.V. 1952. The hair of California mammals with keys
to the dorsal guard hairs of California mammals. Amer.
Midland Nat. 48:480-512.

Melquist, W.E. and M.G. Hornocker. 1979. Methods and
techniques for studying and censusing river otter
populations. Tech Report 78, Forest, Wildl. and Range
Exper. Station, University of Idaho, Moscow, Idaho,
17pp.

Morrow, J.E. -' 1979. Preliminary keys to otoliths of some
adult fishes of the Gulf of Alaska, Bering Sea, and
Beaufort Sea. NOAH Tech. Report NMFS Circ. No. 420

Murie, O.J. 1954. A field guide to animal tracks.
Houghton Mifflin, Boston, 374pp.

Neter, J., W. Wasserman, and M.H. Kutner. 1983. Applied
linear statistical methods. Richard D. Irwin,
Homewood, Illinois, 1127pp.

259

Pollock, K.H., S.R. Winterstein and M.J. Conroy. 1989.
Estimation and analysis of survival distributions for
radio-tagged animals. Biometrics 45:99-109.

Seber, G.A.F. 1982. The estimation of animal abundance and
related parameters. Macmillan, New York.

Shirley, M.G., R.G. Linscombe, N.W. Kinler, R.M. Knaus,
and V.L Wright. 1988. Population estimates of river
otters in a Louisiana coastal marshland. J. Wild.
Manage. 52: 512-515.

Snedecor, G. W.1 and W. G. Cochran. 1980. Statistical
methods, 7th ed. Iowa State University Press, Ames
Iowa 507pp.

Swihart, R.K. and N.A. Slade. 1985a. Testing for
independence of observations in animal movements.
Ecology 66:1176-1184.

Swihart, R.K. and N.A. Slade. 1985b. Influence of sampling
interval on estimates of home range size. J. Wildl.
Manage. 49:1019-1025.

White, G. C., D. R. Anderson, K. P. Burnham, and D. L. Otis.
1982. Capture-recapture and removal methods for
sampling closed populations. Los Alamos Natl. Lab.
Publ. LA-8787-NERP. 235pp.

Woolington, J.D. 1984. Habitat use and movements of river
otters at Kelp Bay, Baranof Island, Alaska. Unpubl.
M.S. Thesis, Univ. of Alaska Fairbanks, 47pp.

BUDGET: ADF&G

Personnel $ 125.4
Travel & per them 20.0
Contract 166.7
Commodities 35.5

Total $ 347.6

260

TERRESTRIAL MAMMAL STUDY NUMBER 4

Study Title: Assessment of the EVOS on Brown Bear Populations
on the Alaska Peninsula

Lead Agency: ADF&G

Cooperating Agencies: DOI, NPS, FWS

INTRODUCTION

Relatively high densities of brown bears (Ursus arctos) occur
along the 120-mile section of shoreline on the southern edge of
the Alaska Peninsula that was impacted by crude oil from the
EVOS. There has been no objective estimate of the number of
bears in the affected areas, but it is suspected that densities
along the oil-contaminated Katmai coast are higher than those
reported from other coastal brown bear populations: 1 bear/1.1
mi near Terror Lake on northern Kodiak Island (Barnes- et al.
1988) and I bear/2.0 Mi2 near Black Lake on the southern Alaska
Peninsula (Miller and Sellers 1989). These bears are an
important economic and aesthetic resource. On the Alaska
Peninsula, Alaska residents and guided non-residents harvest
about 250 bears annually, spending an estimated $2.2 million on
those hunts (ADF&G files). Thousands of visitors from around the
world come to Katmai National Park and the McNeil River State
Game Sanctuary to observe and photograph bears.

Brown bears are omnivorous, opportunistic feeders near the top of
the food chain. They may ingest oil directly by eating mousse
and tar balls washed ashore, by eating oiled plants and clams, by
scavenging oiled carcasses of animals killed offshore and
deposited on beaches, or by grooming oiled fur. Bears may also
consume animals that have been physiologically contaminated by
sublethal doses of oil. Effects of oil ingestion on individuals
could range from quick death from acute toxic effects to long-
term suppression of reproduction. Experimental work with oiled
polar bears in Canada (Oritsland, et al. 1981) indicated that two
of three animals died from organ failure after grooming. Effects
of oil contamination on bear populations could range from sharp,
immediate declines to subtle long-term reductions as chronic
effects from hydrocarbons stored in fat are expressed.

To continue the determination of impact of EVOS on individual
brown bears and the coastal Alaska peninsula brown bear
population, a study area along the coast of Katmai National Park
was selected. This project will capture and radio-collar an
additional sample of bears in oiled areas and will compare the
natural mortality rate of this sample with that of coastal
populations on Kodiak Island and near Black Lake further south on
the Alaska Peninsula which were not exposed to large amounts of
crude oil. Dead bears found incidentally and radio-collared

261

bears that die will be necropsied, tissue samples taken, and the
cause of death determined. Extent of oil ingestion and the
physiological effects will also be examined.

OBJECTIVES

A. Test the hypothesis that radio-collared brown bears in an
oil-contaminated area of the Alaska Peninsula (Katmai coast)
ingested hydrocarbons (as evidenced by the level of
hydrocarbons in fecal samples) at higher concentrations than
radio-collared bears in an area on the Peninsula that was
not contaminated (Black Lake).

B. Test the hypothesis that natural mortality rates of female
brown bears near oiled areas of the Katmai coast occurred at
a higher rate than females in other coastal brown bear
populations inhabiting non-oiled areas during a period of
three years after EVOS.

C. Test the hypothesis that some of the mortality of
brown bears near the Katmai coast can be attributed to the
physiological effects of ingesting hydrocarbons.

D. Estimate the adult brown bear population density of the
study area (approximately 150 square miles) through a
cooperative project with the NPS using a modified capture-
recapture technique (Miller et al. 1987) with the goal of
obtainihg a coefficient of variation of 0.10.

METHODS

Bears will be captured in the spring of 1990 by using a f ixed-
wing spotter aircraft to locate bears and direct a helicopter
with an immobilizing team to the site. Each bear will be
measured (skull length and width), weighed, tattooed (lips and
groin), and fitted with ear tags and a rad i o -transmitter with
mortality sensor.

Blood and fecal samples will be collected from bears captured
along the Katmai coast and near Black Lake during the spring.
Whole blood will be collected in heparinized and non-heparinized
collecting tubes. Packed cell volume and percent hemoglobin in
the blood will be determined in accordance with standard
operating procedures and serum will be frozen and sent into an
approved laboratory for analysis.

During 1990, radio-collared bears will be relocated by a fixed-
winged aircraft at scheduled two-three day intervals until over
75% of the radio-collared bears are in winter dens. One flight
per month will be scheduled during the denning period. Radio-
tracking flights should continue for two years. During 1991,
flights will be made at two week intervals while bears are active

262

and monthly during denning. A sample of at least 30 radio-
collared bears will be followed into dens each year. It is
anticipated that at least 40 bears must have functioning radio-
collars in the spring to achieve a sample of 30 in the fall. To
maintain this sample size, collaring operations will be necessary
during the spring of 1990 and possibly in 1991.

Mortality data will be collected during radio-tracking flights-.
When a dead bear is observed in the study area, gross necropsies
will be performed in the field. Data on sex, age, and probable
cause and time of death will be recorded. Tissue samples from
recent mortalities will be collected for histopathological and
hydrocarbon analysis. Annual survival, distributions, and
mortality rates will be calculated using modified Kaplan-Meier
techniques (Pollock et al. 1989). Results will be compared with
mortality rates from the Black Lake and Terror Lake (Kodiak)
study areas.

The density estimate (Miller et al. 1987) will be conducted in
the spring of 1990. Prior to the recapture portion of the
procedure, a representative sample of 50 radio-collared bears
will be required to serve as marks for the estimator. Collars
will be distributed proportionally to the estimated proportion of
bears in various reproductive categories (e.g. lone adult males,
lone adult females, subadult males, subadult females, females
with cubs -of -the-year, females with yearling or older offspring)
in the population. only independent observations of individuals
will be used in the estimator.

Data obtained from the density estimate and mortality rate
calculations will be used to estimate the total number of bears
that were killed by the effects of EVOS by comparison between
years and between the oiled area and the control area. A
subsequent population estimate, using the same methods in the
same oil-conta-minated area, will be derived in the spring of
1992. It has been reported that capture-recapture techniques,
such as the proposed density estimate procedure, tend to
underestimate the known size of big game populations (deer) by
10-20% in most instances; but, the estimators can be used to
detect population trends by comparing estimates over time from
the same area (Becker 1989). Due to suspected heterogeneity
among bear classes and lower sightability of bears, compared to
deer, we suspect that the true number of bears in the population
will be underestimated by an unknown amount somewhat greater than
the 10-20% reported in the literature.

263

A two sample, one-sided T-statistic (Snedecor & Cochran, 1980)
will be used to test the oil ingestion hypothesis (objective A).
This statistic assumes the means are normally distributed. if
necessary, transformations will be used to ensure that the
normality assumption is met.

The natural mortality rate hypothesis (Objective B) will be
tested using a log-rank test to compare the two Kaplan-Meier
survival functions (Pollack et al. 1989). This statistic assumes
that differences in survival functions are the result of a
constant shift parameter (Cox & Oakes 1988). This assumption
will be examined by cumulative hazard plots of the two
distributions and possibly involve analysis with time dependent
covariates (Cox & Oakes 1988). It is assumed that bear
populations near Black Lake and Terror Lake are more likely to be
shot by hunters, so all hunter-killed radio collared bears will
be censored (as outlined in Pollock et al. (1989) for animals
that emigrated from a study area or were otherwise lost).

Tissues will be collected and analyzed as outlined by the EVOS
Histopathological Technical Group and the Hydrocarbon Technical
Group to test the hydrocarbon mortality hypothesis (Objective C).
All statistical analyses will be performed at an alpha level of
0.05. A sample size will be used that is adequate to detect at
least one bear affected by oil contamination with a given
percentage of certainty, for varying proportions of the
population contaminated by hydrocarbons. These sample size
calculation's are based on a binomial distribution (Mendenhall,
Schaffer and Wackerly 1981) and assume that the sample size is
very small compared to the population total. For the purposes of
this study it will be assumed that at least 10% of the bear
population was affected by EVOS; therefore, a total sample size
of at least 29 bears will have to be followed by radio-telemetry
to be 95% certain of following at least one bear affected by
hydrocarbons.

The Lincoln-Peterson estimator (Overton 1971) will be used to
estimate daily adult population levels (Objective D) . The mean
of the estimates and its standard error will be used as the point
estimate and standard deviation. The assumptions of this
estimator are (White et al. 1982):

1) all radio-collars are retained;
2) all animals are correctly classified as marked or
unmarked;
3) the recaptures (sightings) of adult bears are
independent;
4) the population is geographically and demographically
closed;
5) all bears have equal capture probabilities that are
constant over time.

264

The geographic closure assumption will be met by determining the
number of radio-collared bears in the study area on a daily
basis, and assuming that the proportion of marked bears in the
area is representative of the unmarked bears. Assumption #5 can
be relaxed to: average probability of capturing a marked animal
equals the average probability of capturing an unmarked animal
(Overton 1971). If capture heterogeneity exists, which it
probably does with brown bears, then mark and recapture estimates
tend to be biased low, because the animals that are easier to
catch are over represented in the marked sample. Because of
this, calculated confidence intervals will. not retain their
statistical validity, and as a result, the precision goal was
stated in terms of the coefficient of variation.

BIBLIOGRAPHY

Barnes, V.G. Jr., R.B. Smith, and L.J. Van Daele. 1988.
Density estimates and estimated population of brown
bears on Kodiak and adjacent islands, 1987.
Unpublished report to the Kodiak Brown Bear Research
and Habitat Maintenance Trust. Anchorage, AK. '34pp.

Becker, E.F. 1989. Mark recapture estimates versus known
populations. Memo to ADF&G bear researchers, dated
Aug. 11, 1989. 13pp.

Cox, D. R. and D. Oakes. 1988. Analysis of statistical data.
Chapman and Hall. London. 201 pp.

Mendenhall, W., R.L. Schaffer, and D.D. Wackerly. 1981.
Mathematical Statistics with Applications. Duxbury
Press. Boston, Mass. 686 pp.

Miller, S.D., E.F. Becker, and W.B. Ballard. 1987. Black
and brown bear density estimates using modified
capture-recapture techniques in Alaska. Int. Conf.
Bear Res. and Manage. 7:23-35.

Miller, S.D. and R.A. Sellers. 1989. Brown bear density on
the Alaska Peninsula at Black Lake, Alaska.
Unpublished preliminary report for the National Pa rk
Service, U.S. Fish and Wildlife Service, and the
Alaska Department of Fish and Game. Anchorage, AK.
36 pp.

Oritsland, N.A., F.R. Engelhardt, F.A. Juck, R.J. Hurst, and
P.D. Watts. 1981. Effect of Crude Oil on Polar Bears.
Can. Dept. Indian Affairs & North Devel. Publ. No.
QS-8283-020-EE-Al.

Overton, W. S. 1971. Estimating the number of animals in
wildlife populations IN wildlife management techniques.

265

R. H. Giles ed. The Wildl. Soc. Washington, D. C.
Pp. 403-456.

Pollock, K.H., S.R. Winterstein, C.M. Bunck, and P.D.
Curtis. 1989. Survival analysis in telemetry studies:
The staggered entry design. J. Wildl. Manage.
53(l):7-15.

Snedecor, G.W. and W.G. Cochran. 1980. Statistical
Methods; Seventh Edition. Iowa St. Univ. Press.
Ames, Iowa. 507 pp.

White, G. C., D. R. Anderson, A. P. Burnham, and D. L. Otis.
1982. Capture-Recapture and Removal Methods for Sampling
Closed Populations. Los Alamos Natl. Lab. Publ. LA-8787-
NERP. 235 pp.

BUDGET: ADF&G

Salaries $ 44.6
Travel 4.3
Services 47.0
Commodities 9.8
Equipment 20.0

Total $ 125.7

266

TERRESTRIAL MAMMAL STUDY NUMBER 6

Study Title: Influence of Oil Hydrocarbons on Reproduction of
Mink (Mustela vision)

Lead Agency: ADF&G

INTRODUCTION

The mink is a carnivorous mammal inhabiting the margins of streams,
lakes, marshes, and marine islands throughout most of North
America. It is at the top of the food chain and thus exposed to a
wide variety of environmental contaminants. Certainly the most
dramatic ef f ect of a toxicant or pollutant is outright death of the
animal. An equally devastating effect on the animal population,
however, is for apparently healthy animals to fail to reproduce or
to produce weakened offspring. Studies with ranched mink have
documented that these animals are sensitive to many chemical and
biological compounds (Sundqvist et al.1989). Some of those known
to interfere with reproduction include heavy metals, halogenated
hydrocarbon pesticides and other aromatic, halogenated hydrocarbons
(Ringer, 1981; Sundqvist et al.1989). In the mid-1960s a decline
in reproductive performance in ranched mink (Hartsough, 1965) was
eventually traced to high polychlorinated biphenyl (PCB) content of
Great Lakes f ish -used in commercial mink diets (Aulerich et
al.1971; Aulerich et al.1973). Ranched mink fed 2-5 ppm PCBs, 1-
2.5 ppm polybrominated biphenyls, or 5-25 ppm hexachlorobenzene
suffered complete reproductive failure, significantly reduced
litter size and/or excessive kit mortality (Aulerich et al.1985;
Aulerich, and Ringer, 1979; Aulerich, and Ringer, 1977; Bleavins et
al..1984; Bleavins et al.1980; Hornshaw et al.1983; Ringer, 1981;
Ringer et al.1972; Rush et al.1983; Wren et al.1987). Particularly
dangerous are compounds like PCBs that accumulate in the
subcutaneous fat (Hornshaw et al.1983). Therefore, it is possible
that hydrocarbons in crude oil ingested by mink and other
carnivores may interfere with reproduction.

Crude oil released into the environment is immediately subjected to
a variety of weathering processes (Payne and McNabb, 1984). Within
a few weeks of an oil spill the majority of the more toxic, lower
molecular weight compounds are eliminated, primarily through
evaporation (Payne and McNabb, 1984). However, heavier distillate
products not subject to significant evaporative loss persist in the
environment and are more likely to enter the food chain in
signif icant quantities. It is the ef f ect of ingestion of this
weathered oil that will be studied.

Mink feed on fish., small mammals, frogs, aquatic insects, and
occasionally birds. These prey species live in areas impacted by
the EVOS. Therefore, it is highly probable that mink will be
exposed to oil hydrocarbons. But direct cause and effect

267

relationships such as the influence of oil hydrocarbons on
reproduction are difficult to demonstrate in a field study. The
complexity of a natural setting imposes too many uncontrolled
variables. Even if field studies measure hydrocarbons from wild
animals or show a change in the population dynamics of one or more
species, there will always be the unanswered question: Was it the
oil or some other unmeasured factor that influenced reproductive
performance? A controlled experiment in the laboratory will be
conducted to define the effects of short-term and long-term
ingestion of non-lethal amounts of weathered Prudhoe Bay crude oil
(WPBC) using ranched mink as a model species. mink will be used
for three reasons: 1) mink are known to be sensitive to
hydrocarbon pollutants, 2) mink inhabit the PWS area and are thus
at risk for exposure to oil hydrocarbons in feed, and 3) the mink
is a well-established model for laboratory research and thus there
is excellent documentation of their reproductive physiology (Enders
1952; Sundqvist, et al., 1989). Ranched mink are not domestic
animals, and their physiological response to oil ingestion in the
laboratory setting can be predicted to be no different from that of
their wild counterparts.

OBJECTIVES

A. Short-Term Ingestion of Weathered Prudhoe Bay Crude Oil

Test the hypothesis that short-term (seven day) , low-level (100
ppm) ingestion of WPBC oil during pre-estrus, diapause, gestation,
or lactation does not produce a significant (P < 0.05) difference
in the reproduction of female mink. Reproductive variables will be
the number of kits per litter, kit survival, kit growth and
maturation, and histology of adult and kit reproductive tracts.

B. Long-Term Ingestion of Weathered Prudhoe Bay Crude Oil

Test the hypothesis that continual, low-level (100 ppm) ingestion
of WPBC oil starting during pre-estrus and continuing through to
the weaning of kits does not produce a significant (P < 0.05)
difference in reproduction of female mink. Reproductive variables
will be the number of kits per litter, kit survival, kit growth and
maturation, and histology of adult and kit reproductive tracts.

METHODS

Crude oil from the Exxon Valdez will be weathered by placing it in
a flat bottomed, solvent-rinsed vessel to a depth no greater than
2 cm. The oil will be gently agitated in a fume hood for seven
days at room temperature.

All mink will be fed commercial mink ration consisting of ground
fish, chicken and beef by-products. WPBC oil will be mixed into the
feed at a rate of 100 ppm. In order to measure such small amounts
accurately, we will first dilute the WPBC in salmon oil. . The

268

resulting oiled diet will contain 100 ppm WPBC and 10 ml salmon oil
per kg commercial mink ration. The level of 100 ppm WPBC oil was
chosen because it did not alter food palatability and the oiled
food was readily consumed by the mink. In addition, this level of
contamination does not cause the mink to suffer or exhibit any
clinical symptoms.

For the short-term study, mink will be fed either 0 WPBC (control
group) or 100 ppm WPBC for seven days at a time specif ic to the
reproductive cycle (pre-estrus, diapause, pregnancy, and
lactation). A total of 70 female mink will be used. Mink (20
each) will be randomly assigned to the control and pre-estrus
groups prior to the breeding season. Successfully mated mink from
the remaining 20 animals will be randomly assigned to the diapause,
pregnancy, and lactation groups following breeding. The female
mink will be bred to untreated males. Males will be checked for
fertility by palpation of testes, demonstration of copulatory
behavior, and evidence of motile sperm in vaginal smears following
mating. Females will be allowed to rear their young to weaning
age. After weaning, adults and selected kits will be euthanized
and tissue and blood samples will be collected for histopathology,
hydrocarbon analyses and liver cytochrome P450 analyses. Estrous
behavior will be analyzed by comparing the number of females
successfully bred and the time of breeding of the 20 control
animals to the 20 pre-estrus animals. For these tests all 20
animals in each group will be included. For all other tests, only
those animals successfully bred will be included. The response
variables to be analyzed on mated females in each group include
number of kits born, number of live kits born, birth weight of
kits, survival rate of kits, and growth rate of kits.

For the long-term study twenty female mink will be fed 100 ppm WPBC
oil in their diets beginning in February, 1990 and continuing until
kits are weaned in June, 1990. The control group (n=20) from the
short-term study will also serve as a control for this study. All
mink in this long-term exposure study will be bred to untreated
males and will be allowed to rear their young to weaning age.
Adults and selected kits will be euthanized at weaning, and tissue
and blood samples will be collected for histopathology, hydrocarbon
analyses and liver cytochrome P450 analyses. The response
variables to be examined in this study are identical to those in
the short-term exposure study: mating activity, number of kits
born, number of live kits born, birth weight of kits, survival,
growth and maturation of kits.

The University of Alaska Fairbanks, (UAF), who will conduct this
research, has on file with the Office for Protection Against
Research Risks, National Institutes of Health, an "Assurance of
Compliance with Public Health Service Policy on Humane Care and Use
of Laboratory Animals". The University's Animal Facilities are
licensed by the United States Department of Agriculture and are

269

subject to twice yearly, unannounced inspections by USDA
veterinarians to ensure compliance with the Animal Welfare Act.

UAF has a full-time, staff veterinarian who supervises the
veterinary care program as outlined in the Regulations of the
Animal Welfare Act. The University also has an Institutional
Animal Care and Use Committee mandated by Public Health Service
Policy and the Animal Welfare Act. This project was reviewed and
approved, without modification, by the committee.

Because of the nature of the response variables being examined, a
number of different statistical tests will be used. comparisons
between two groups involving binomial variables, such as number of
females bred, will be tested by chi-square. Analysis of variables
involving kits as an experimental unit will be tested by analysis
of variance. When there are significant differences, treatment
groups will be compared to the control group by Dunnet's test.
Because kit birth weight and kit growth rate are expected to vary
with litter size unrelated to specific treatment effects, analysis
of covariance with litter size as the covariate will be done.
Statistical significance of differences between variables involving
the long-term exposure group and the control animals will be
determined by T-test or by chi-square as appropriate.

BIBLIOGRAPHY

Aulerich, R.J., Bursian, S.J., Breslian, W.J., Olson, B.A., and
Ringer, R.K. (1985). Toxicological manifestations of
2 , 4 , 5 , 2 4 1 , 5 - 2 , 3 , 6 , 2 1 , 3 1 , 6 1 a n d
3,4,5,31,41,51-hexachlorobiphenyl and aroclor 1254 in mink.
J.Toxicol.Environ.Health 15, 63-79.

Aulerich, R.J. and Ringer, R.K. (1977). Current status of PCB
toxicity to mink, and effect on their reproduction. Archs
environmental Contamination Toxic. 6, 279-292.

Aulerich, R.J. and Ringer, R.K. (1979). Toxic effects of dietary
polybrominated biphenyls on mink. Archs environmental
Contamination Toxic. 8, 487-489.

Aulerich, R.J., Ringer, R.K., and Iwamoto, S. (1973). Reproductive
failure and mortality in mink fed on Great Lakes fish.
J.Reprod.Fert. Suppl.19, 365-376.

Aulerich, R.J., Ringer, R.K., Seagran, H.L., and Youatt, W.G.
(1971). Effects of feeding coho salmon and other Great Lake
fish on mink reproduction. Can.J.Zool._ 49, 611-616.

Bleavins, M.R., Aulerich, R.J., and Ringer, R.K. (1980).
Polychlorinated biphenyls (Aroclors 1016 and 1242): effects on
survival and reproduction in ferrets. Archs environmental
Contamination Toxic. 9, 627-635.

270

Bleavins, M. R. , Aulerich, R. J. , and Ringer, R. K. (1984) . Ef f ects of
chronic dietary hexachlorobenzene exposure on the reproductive
performance and survivability of mink and European ferrets.
Archs environmental Contamination Toxic. 13, 357-365.

Enders, R.K. (1952). Reproduction in the mink. Proc.Am. Phil. Soc.
96, 691-755.

Hartsough, G.R. (1965). Great Lakes fish now suspect as mink food.
Am.Fur Breeder 38, 25-27.

Hornshaw, T.C. Aulerich, R.J., and Johnson, H.E. (1983). Feeding
Great Lakes fish to mink: effect on mink and accumulation and
elimination of PCB's by mink. J. Toxicol. Environ. Health 11,
933-946.

Payne, J.R. and McNabb, G.D.JR. (1984). Weathering of petroleum in
the marine environment. MTS Journal 18(3), 1-20.

Ringer, R.K. (1981). The effects of environmental contaminants on
reproduction in the mink (Mustela vison). In:" Environmental
factors in mammal reproduction" (Gilmore, D., and Cook, B.,
Eds.) pp. 232-237. University Park Press, Baltimore.

Ringer, R.K., Aulerich, R.J., and Zabik, M. (1972). Effect of
dietary polychlorinated biphenyls on growth and reproduction
in the mink. Am.Chem.Soc.Nat.Meeting 12, 149-152.

Rush, G.F., Smith, J.H., Maita, K., et al (1983). Perinatal
hexachlorobenzene toxicity in the mink. Environ. Res. 31,
116-124.

Sundqvist, C., Amador, A.G., and Bartke, A. (1989). Reproduction
and fertility in the mink (Mustela vison). J.Reprod.Fertil.
85, 413-441.

Wren, C.D., Hunter, D.B., Leatherland, J.F., and Stokes, P.M.
(1987). The effects of polychlorinated biphenyls and
methylmercury, singly and in combination on mink. II:
Reproduction and kit development. Archs environmental
Contamination Toxic. 16, 449-454.

BUDGET: ADF&G

Salaries $ 23.7
Travel 0.0
Contracts 62.3
Supplies 48.0
Equipment 0.0

Total 134.0

271

BIRD INJURY ASSESSMENT

The EVOS resulted in the death of a large number of migratory
birds, especially seabirds, waterfowl, and bald eagles. In the
months following the spill it became apparent that the vast
populations of numerous bird species that inhabit or utilize the
spill zone remained at risk to direct mortality as well as
sublethal, long-term damages.

Fourteen studies were developed and deployed during 1989 to
document damages to migratory birds. It was recognized early in
the process that it was not possible to study all the bird species
potentially affected by the oil spill nor the full scope of effects
to any species. Therefore, efforts were concentrated on studying
key species or groups of species where injury was most evident and
valid damage assessment could be determined in a reasonably cost-
effective manner.

Seven of these studies will be continued in 1990. Some studies
were not continued because it was concluded that all data pertinent
to assessing damages likely to be gathered had indeed been
gathered. Some studies, such as Bird Studies B6, B8 and B9 were
either integrated into the remaining studies or are being conducted
independent of the NRDA process.

Continuing studies have been expanded and/or modified in response
to comments from reviewers and the public. The beached bird
survey, for example, was carefully designed to provide essential
data on bird drift and sinking rates that will increase the
accuracy of total bird mortality estimates. Both the eagle and
peregrine falcon studies will provide information on loss to
breeding populations, as well as reproductive -success. The seabird
colony and seabird and waterfowl surveys will provide a means to
compare pre and post spill populations. Finally, the sea duck
study will provide important information on sublethal effects of
the spill on harlequin ducks.

272

BIRD STUDY NUMBER 1

Study Title: An Assessment of Damage to Seabirds in PWS and the
western GOA Resulting from the EVOS.

Lead Agency: FWS

INTRODUCTION

This study will assess mortality of marine birds following the EVOS
by adapting existing bird damage assessment models to estimate
total seabird mortality. The proposed field studies will examine
the general characteristics of the decomposition and disappearance
of bird carcasses under environmental conditions typical of PWS and
the GOA to determine the rate at which carcasses are lost and the
factors affecting that rate. It is not possible to simulate the
exact environmental conditions prevailing during the spill, in
particular the wind and current conditions and the presence of
large floating mats of oil.

OBJECTIVES

A. Synthesize available information on the beachcast-bird
recovery effort.

B. Determine the number of birds that died as a consequence of
oiling.

C. Determine the rate of carcass loss at sea and the time-course
of sinking, decomposition and scavenging.

D. Develop an estimate of total seabird mortality.

METHODS

The modeling approach is based on risk assessment models developed
for the Pribilof Islands, Kodiak Island, and the Southern
California Bight, and on damage assessment models developed for the
T/V Puerto Rican and T/V Apex Houston oil spills (Ford et al. 1982,
Weins et al. 1982, Ford 1984, Dobbin et al. 1986, Ford et al. 1987,
Page et al. 1990).

The carcasses of oiled seabirds may be assumed to have encountered
one of three fates:

(1) beached but not recovered,

(2) lost at sea without making landfall,

273

(3) beached and recovered.

The general approach will be to utilize data for birds which were
beached and were recovered to estimate the total number of birds
which were beached, and, with the aid of oil spill trajectory
information, to work backwards from the beaches where birds were
recovered in order to estimate the total number of birds which were
directed toward the beach but were lost at sea before being
beached.

Unless a beach is thoroughly searched at frequent intervals, some
of the beached birds will not be recovered. During the Apex
Houston oil spill in California, Page and Carter (1986) found that
only 60% of the carcasses persisted on the beach face from one day
to the next. Beached birds may not be recovered due to the
scavenging, burial, or reflotation of carcasses. A model to
estimate the actual number of carcasses deposited on a beach based
on the observed number of carcasses on the beach is presented at
Page et al. 1986. Data sufficient for this type of modeling are:

(1) The arrival rate of carcasses on a given beach as a
function of time.

(2) The frequency with which the beach was searched.

(3) The likelihood that a carcass will not be detectable as
a function of the length of time spent on the beach and
the age of the carcass when it is first cast.

some of the data types listed above will need to be derived from
information collected during the spill, and others will be
estimated as a part of the proposed field studies. The arrival of
bird carcasses on a given beach typically peaks rapidly, and tapers
of f over a period of up to a week. A typical arrival curve can
usually be constructed from more frequently searched beaches and
applied to less frequently searched beaches. Data on the
disappearance rates of beached carcasses and the distance moved by
a refloated carcass would be collected as a part of the proposed
study.

In addition to carcasses that were not recovered on searched
beaches, many carcasses were probably deposited on beaches which
were not searched at all. This effect will be accounted for by
identifying a number of beached bird recovery areas based on
coastal location and geomorphology. Within a recovery area, it is
assumed that the average lineal density of carcasses for a given
beach type in the unsampled area was the same as that in the
sampled area within each sector. Using a Geographical Information
System, the lineal extent of searched and unsearched beaches within
each sector will be determined from a digital coastline.

Even after correcting the estimate of beached birds for birds that

274

were beached but were not recovered, a large fraction of the
mortality will remain unmeasured. Carcasses that are drifting
landward may not beach because they sink or are scavenged along the
way. The process of carcass loss both at sea and on the beach f ace
is poorly understood, but is critical to the estimate of total
mortality since this appears to have been the fate of a large
proportion of the birds killed in other spills. The component of
the at-sea loss from sinking, scavenging, or other causes will be
computed by taking the number of birds (corrected for recovery on
the beach), estimating the time spent at sea, and correcting for
the loss in transit. Time spent at sea will be estimated using
trajectory data. These sources will be used to determine the
likely path followed by a drifting seabird carcass to reach a given
beach. The distribution of time spent at sea for carcasses found
on that beach will be estimated by integrating the likely path of
the oil reaching that beach with a density surface describing the
at-sea distribution of birds through which the trajectory passed.
The at-sea distribution will be estimated using historical data.
Distributional data of this type for the GOA and PWS is known to
have a high degree of variability both because of sample size and
inter-annual variation. The large scale aspects of bird
distributions, however, such as the relationship of concentrations
of birds relative to the shelf break or to colony sites can be
expected to be conservative and can be adequately described by
existing data. As part of the sensitivity analysis, the effect of
alternative at-sea distributions of birds on model results will be
tested. ongoing studies suggest that the loss of carcasses at sea
increases with time. The process of entering the near shore
environment or beaching probably accelerates the disappearance
rate, especially for older carcasses.

Some carcasses will also be carried out to sea by winds and
currents. The size of this component of the at-sea loss will be
estimated using trajectory and bird distribution data.

The effect of changing the value of a single input parameter on
model results will be examined for all single-valued parameters for
which the exact value is arguable. This will be done by running
the model with a series of different values for a given parameter
and plotting the estimated number of birds killed as a function of
the input value for that parameter. Examples of the kind of
parameters which could be analyzed in this way include the
disappearance rate of carcasses on the beach face, the proportion
of birds at risk which actually died, etc.

The effect of simultaneously using best-case and worst-case values
for all appropriate parameters will be examined in order to
generate maximum and minimum estimates of total mortality. The
effect of varying some of the inputs which are not single valued,
such as the distribution of birds at sea based on historical data,
will also be examined. This would be done by manually
constructing alternative distributions which are consistent with

275

the biology of various species, but which would lead to either
increased or decreased values of total mortality.

An alternative is to use a Monte Carlo type approach based on the
use of probability density functions for arguable parameters (Ford
et al. 1982).

Detailed information describing recovery efforts will be used to
correct the model input values for variation in level of effort.
To the extent necessary, a general description of the recovery
effort will be pieced together from records accompanying the frozen
specimens. These records indicate where the particular birds were
recovered and when.

Logbooks will be reviewed to determine the distribution of search
effort along the coastline impacted by EVOS, although in some cases
it may be necessary to estimate search effort from other sources
including interviews and returns of beachcast birds.

Dead birds collected following the spill are presently bagged and
stored in freezers in Anchorage. Seabirds die f or a variety of
reasons and may have been oiled secondarily. A sample
(approximately 10%) of recovered birds will be examined to
quantify the proportion of unoiled birds that may have been
recovered in different areas and times after the spill, and to
describe the degree of oiling and state of decomposition for a
number of birds from several locations. Some necropsy work may be
performed on unoiled or lightly-oiled birds to determine whether
oil had been ingested.

Field studies will be conducted to determine an estimate of loss
rate of oiled birds at sea. The loss of birds increases over time,
and, ultimately, would reach 100% for birds not beached. The loss
rate of birds oiled at sea can be estimated by radio-tracking of
free-drifting carcasses. The study will be conducted in two
phases, the first focusing on PWS and the second on the Alaska
Peninsula where much of the impact on seabird communities occurred.

Free-drifting carcasses will be radio-tracked to obtain data on the
rate of loss of oiled carcasses following release. The proportion
of the sample group that is deposited on land will be determined.
Carcasses will be released from known locations along different
stretches of coastline. over the duration of the study, the
position of a carcass will be monitored until it has become
beachcast or the signal is lost. Since the failure rate of the
radio-tags is very low, loss of signal indicates that a carcass has
become submerged (taking the radio with it) or has drifted out of
range. In order to distinguish between these two possibilities, a
set of "decoys" will be released as a control. In theory,
carcasses and decoys will drift in the same manner but one group
will begin to sink (the carcasses) and the other will not. By
comparison of data from the two groups, the loss rate of carcasses

276

over time can be calculated. To the limits of data, we will
correlate rate of loss with species, degree of decomposition, and
degree of oiling.

Some beached carcasses will be inspected by the boat crew for
locations that can be safely visited. Beached carcasses and decoys
will be photographed and scored for visibility and other factors
that would affect their collection during beach clean-up effort and
will be left in place and monitored from the air to determine if
they are refloated.

Each carcass and decoy will be equipped with a radio-tag that
floats upright. Each radio unit weighing approximately 15 grams,
has an effective range of about 03 nmi at 2000 feet above sea level
(ASL) flight altitude, with a battery life of 50 days. The
transmitters are placed in devices that are completely self-
righting and have neutral buoyancy. Radio-tags are activated by
removal of an external magnet. Each radio-tag has a characteristic
frequency and pulse rate.

A scanner-receiver in an aircraft equipped with paired antenna
mounted on the wings will be used to find the signal from each
transmitter. The operator can select for a particular frequency or
the instrument will automatically scan for all available
frequencies.

Determination of the precise location of a transmitter at sea is
not necessary to achieve the objectives of the -study since the
principal interest is in obtaining a frequent inventory of
carcasses and decoys present in the area. If a carcass or decoy
appears to be present on a beach, effort will be made to determine
its precise position and assess the likelihood of its recovery.
The approach used will be to f ly parallel to the beach over the
surf zone. If the transmitter is offshore, the signal will be
received by only the outboard antenna. If the transmitter is on
the beach or in the surf, the signal will be received by only the
inboard antenna. Flight altitude of the tracking aircraft over
open water will typically be 3,000 to 5,000 ft. ASL, but will
decrease to 500 ft ASL or less near the beach. On subsequent days
following the release, a search area will be identified taking into
consideration the last known location of transmitters and the
results of a trajectory model. The model to be used is the NOAA
Oil Spill Simulation Model (OSSM), using the most appropriate
current field and winds data.

Because of the great number of radio-tags deployed, a special
computer data-logging system will be used for output from the
receiver-data processor equipment.

Calculation of the coordinates of the transmitter will be
accomplished using a mathematical function describing the decrease
in signal amplitude with distance. This relationship must be

277

described for each kind of transmitter.

BIBLIOGRAPHY

Burger, A.E. 1989. Effects of the Nestucca oil spill (January
1989) on seabirds along the southwest cost of Vancouver
Island.

Report to Environment Canada. Contract No. KA601-8-4184. 26 pp
and Appendices.

Carter, H.R., G. W. Page, and R.G. Ford. 1987. The importance of
rehabilitation center date in determining the impacts of the
1986 oil spill on marine birds in central California.
Wildlife Journal 10:9-14.

Dobbin, J.A., H.E. Robertson, R.G. Ford, K.T. Briggs, and E.H.
Clark 11. 1986. Resource'damage assessment of the T/V Puerto
Rican oil spill incident. Unpubl. report, James Dobbin Assoc.
Inc.1 Alexandria, Virginia.

Fleming, T., D. Varoujean, and S. Speich. 1990. Performance and
reliability of radio-tags for Marbled Murrelets. Abstract.
Proceedings of Meetings of the Pacific Seabird Group, February
1990, Victoria, B.C., Canada.

Ford, R.G. 1984. Southern California Marine Mammal and Seabird
Risk Ahalysis. Final Report prepared for the Minerals
Management Service, U.S. Department of the Interior, by
Ecological Consulting. Contract #14-12-0001-30224. 240 pp.

Ford, R.G., J.A. Wiens, D. Heinemann, and G.L. Hunt, Jr. 1982.
Modeling the sensitivity of colonially breeding marine birds
to oil perturbation. J. Appl. Ecol. 19:1-31/

Ford, R.G., G.W. Page, and H.R. Carter. 1987. Estimating
mortality of seabirds from oil spills. Pp. 848-51. In Proc.
1987 Oil Spill Conference, American Petroleum Institute,
Washington, D.C.

Ford, R.G. and J.L. Casey. 1989. Preliminary Draft Report:
Seabird Mortality Resulting from the Nestucca oil Spill
Incident, Winter 1988-89. Prepared for Washington Department
of Wildlife. 37 pp.

Gould, P.D., D.J. Forsell, and C.J. Lensink. 1982. Pelagic
distribution and abundance of seabirds in the Gulf of Alaska
and the eastern Bering Sea. Fish and Wildlife Service, U.S.
Dept. of the Interior. FWS/OBS-82-48.

Page, G.W. and H.R. Carter (eds). 1986. Impacts of the 1986 San
Joaquin Valley crude oil spill on marine birds in central

278

California. Unpubl. report, Point Reyes Bird Observatory,
Stinson Beach, California.

Page, G.W., H.R. Carter, and R.G. Ford. 1990. Numbers of seabirds
killed or debilitated in the 1986 Apex Houston oil spill in
central California. Studies in Avian Biology. In press.

Wiens, J.A., R. G. Ford, D. Heinemann, and C. Fieber. 1982. A
statistical analysis of the distribution of seabirds in the
vicinity of Kodiak Island, Alaska. In Environmental
Assessment of the Alaska Continental Shelf. Final Reports.

Budget: FWS

Personnel $46.0
Travel 21.5
Contractual 521.1
Commodities 9.4
Equipment 0.0

Total: $598.0

279

BIRD STUDY NUMBER 2

Study Title: Surveys to Determine Distribution and Abundance of
Migratory Birds in PWS and the Northern GOA

Lead Agency: FWS

INTRODUCTION

This study will continue to examine whether the EVOS caused a
decline in the distribution and abundance of water birds in the
waters and shorelines affected by the spill. Potential injuries to
waterbirds from exposure to the EVOS include direct mortality,
changes in behavior, and decreased productivity. Surveys by small
boats and airplanes will collect information on the seasonal
distribution and abundance of waterbirds in the spill zone for
several seasons following the spill. These post-spill data will be
compared to data collected using similar methods in prior years to
determine whether the oil spill affected waterbird distribution and
abundance.

This proposal describes the boat survey and aerial survey work that
will be accomplished in the second year of this study. The study
will continue portions of Bird Studies Number 6 (Marbled Murrelets)
and Number 9 (Pigeon Guillemot), and will also record observations
of marine mammals in the study area which could prove valuable to
other ongoing studies.

OBJECTIVES
A. Aerial Surveys

1. To determine seasonal distribution and estimate relative
abundance of waterfowl and waterbirds in PWS and KP.

2. To compare relative abundance and distribution of
waterfowl and other waterbirds in oiled vs. non-oiled
areas between historical (1971) and recent (1989 and
1990) surveys. To monitor changes in the distribution and
abundance of waterfowl and other waterbirds between oiled
and non-oiled areas in PWS and KP.

3. To estimate the long- and short-term recovery rates of
waterfowl and waterbird populations impacted by the oil
spill.

B. Boat Surveys

1. To determine distribution and estimate abundance (with
95% confidence limits) of waterbirds in PWS and the
northern GOA.

2. To test the null hypothesis that estimates of waterbird

280

relative abundances, using new and comparable historic
data, are not significantly lower (a = 0.05) in oiled
than non-oiled areas in PWS and the northern GOA.

3. To estimate the long- and short-term trends of
populations that were determined in previous objectives
to be reduced by the oil spill.

4. To test the null hypothesis that the total number of
Pigeon Guillemots attending colonies at Naked Island,
PWS, following the EVOS is Dot significantly lower than
the total number attending in prior years.

5. To test the null hypothesis that the abundance index of
Marbled Murrelets on five transects on the western side
of Naked Island, PWS, following the EVOS is not
significantly lower than the abundance index on each
transect in prior years.

METHODS

A. Aerial Survey Sampling Methods

Three surveys will be conducted during the 1990-1991 oil spill
year: the spring survey during May; the fall survey during
September; and the winter survey during February. A fourth
survey (summer survey during July) may be undertaken. Survey
dates are selected to count the spring and fall populations at
or near the peak of their migration and to count winter (and
possibly summer) populations when they are most stable. Three
single-engine fixed-wing aircraft, will be used for each
survey.

The aircraft will be flown at approximately 150 feet above
water level and 200 -meters offshore, following the shoreline
as closely as possible given the aircraft's capabilities, and
maintaining an airspeed of 95 - 100 mph. All birds and marine
mammals will be observed within 200 meters from each side of
the aircraft. Date, survey beginning and stop time, wind
speed and direction, air temperature, cloud cover, ceilings
and visibility will be recorded for each survey date.

Survey weather minimums will be restricted to 1,500 feet or
greater ceilings, 10 miles horizontal visibility, surface
winds of 15 knots or less, and seas no larger than wavelets.

The entire coastline of PWS and southern KP, including
Kachemak Bay, will be surveyed during each of the three
seasonal surveys.

The visibility bias associated with counting birds from low
flying aircraft will result in an underestimate of population

281

size and bird densities, but this bias will be the same in
both oiled and non-oiled areas and remain relatively constant
between surveys. Therefore, the bias will have no effect on
comparisons between areas and years.

Estimates of waterfowl and waterbird abundance and
distribution will be done using direct comparisons appropriate
for complete shoreline survey indexes. If a significant
change in a species group or individual species (such as sea
ducks or goldeneye) is detected, then subsets of data
stratified by habitat will be analyzed to assess oil effects
in comparable oiled and non-oiled habitat for each survey.

Short- and long-term recovery rates, if there is a significant
oil effect, will be calculated using a repeated measures
ANOVA. Trends may also be compared using regression
techniques.

B. Boat Survey Sampling Methods

Surveys will be conducted from small boats manned by an
operator and two observers. observers will record all birds
and marine mammals within 100 meters on each side of the boat
within survey transects. The survey window extends
approximately 40-50 m ahead of and 100 m above the moving
boat. Date and time of survey, and environmental variables
including wind velocity and direction, air and water
temperature, weather, observation conditions, sea state, tide,
presence or absence of oil, and human activity will be
recorded for each transect.

a. PWS

A stratified random sampling design, including shoreline,
coastal/pelagic and pelagic strata, will be used to meet
Objectives 1-3. Approximately 29% of the shoreline and 25% of
coastal/pelagic and pelagic plots will be surveyed once in
March 1990 and three times (one survey each in June, July and
August) during the summer of 1990. Surveys will be conducted
jointly with Marine Mammal Study Number 6. All birds and
marine mammals within transect boundaries will be recorded.

The shoreline stratum includes all water within 200 m of any
shoreline, and will be surveyed by traveling 100 m offshore,
parallel to the coast, at 5-10 knots. The shoreline stratum
is divided into transects consistent with those used by D.
Irons during 1984-1985 surveys.

Coastal/pelagic and pelagic strata consist of plots of water
delineated by 5-minute intervals (latitude and longitude) on
NOAA charts and exclude any water within 200 m of the coast.
Coastal/pelagic and pelagic plots differ in that

282

coastal/pelagic plots include more than approximately 1 nm
(nautical mile) of shoreline, whereas pelagic plots contain
less than 1 nm of shoreline. For plots that are 5 minutes
wide (east to west), two north-south transects extending 100
m on each side of the boat and located 1 minute inside the
east and west boundaries of the plot will be steered by a
combination of compass heading and LORAN-C coordinates. For
plots that are less than 5 minutes wide due to intersection
with land, either one or two north-south transect lines will
be surveyed, depending on plot size.

b. Kodiak Island

Shoreline and pelagic transects will be surveyed off the west
and north coasts of Kodiak Island to meet Objectives 1-3. The
shoreline was divided into transects based on habitat type
following the criteria of Irons et al. (1988) and then a
simple random sample consisting of 25% of the transects was
chosen. These shoreline transects, and all pelagic transects
surveyed by Forsell and Gould (1981) will be surveyed in July
1990 and February 1991.

C. Naked Island

To meet Objective 4, the number of Pigeon Guillemots in the
Naked Island group will be determined by circumnavigating each
island in a small boat between 50 and 100 m from shore and
counting all Guillemots on land and water. Surveys will be
done during peak attendance, (between 0500 and 1000 hours),
during the first week in June, when breeding birds are in
attendance, and during the last week in July, when
non-breeding birds are in attendance as well.

Ancillary observations on Pigeon Guillemot reproduction will
be made. objectives of such observations include nesting
success, fledgling weight, the rate at which chicks are fed,
and identification of fish species fed to chicks as food.

The number of Marbled Murrelets (and other birds and marine
mammals) on five inshore transects on the western side of
Naked Island will be counted from a small boat (Zodiac) to
meet Objective 5. The transect routes will be those used by
Bird Study 9 in 1989, and in three years previous to the oil
spill to allow comparison with all previously collected data.

Estimates of waterbird abundance and variances (Objective 1)
will be made using ratio estimators and statistics appropriate
for stratified random sampling. This technique will be used
if the number of birds is positively correlated with transect
length. If the- correlation between bird numbers and transect
length is poor, simple means and variances will be calculated.

283

Differences in waterbird abundance between oiled and non-oiled
areas (Objective 2) will be tested by examining the change in
abundance on each transect between 1984 and 1989. One-tailed
t-tests or the Mann-Whitney test will be used. If areas are
stratified further based on habitat type then ANOVA will be
used. ANOVA will be used to make comparisons between pre- and
post-oil spill data with respect to oiled and non-oiled areas.

Short- and long-term recovery rates (Objective 3), if there is
a significant oil effect, will be analyzed using a repeated
measures ANOVA. Trends may also be compared using regression
techniques.

The outlier t-test will be used to test whether the number of
guillemots counted at Naked Island in 1990 is significantly
lower than the mean of mean counts in prior years (objective
4). Counts will be logarithmically transformed prior to
testing.

The outlier t-test will be used to test whether the abundance
indices of Marbled Murrelets in 1990 are significantly lower
than the mean of mean indices for prior years (Objective 5).

BIBLIOGRAPHY

Alaska Department of Environmental Conservation. 1989.
Unpublished preliminary digital maps of oil-impacted shore
based on aerial and boat surveys during early ADEC response
activities. (ARC/INFO data file].

Bowden, D.C. 1973. Review and evaluation of May waterfowl
breeding ground survey. Unpubl. Manu. Fish and Wildlife
Service. Anchorage, Alaska. 74 pages.

Butler, W. 1989. Prince William Sound and Kenai Peninsula
Migratory Bird Aerial Survey Project. Unpublished data. U.S.
Fish and Wildlife Servicef Anchorage, Alaska.

Cochran, W.G. 1977. Sampling Techniques. John Wiley and Sons,
Inc. New York, New York. 428 pages.

Conant, B., J.G. King, J.L. Trapp, and J.I. Hodges. 1988.
Estimating populations of wintering waterfowl in southeast
Alaska. U.S. Fish and Wildlife servicef Juneau, Alaska.
Unpublished Report. 16 pages.

Dwyer, T.J., P. Isleib, D.A. Davenport, and J.L. Haddock. 1975.
Marine Bird Populations in Prince William Sound Alaska. U.S.
Fish and Wildlife Service, Anchorage,, Alaska. Unpublished
Report, 21 pages.

284

Forsell, D.J. and P.J. Gould. 1981. Distribution and abundance
of marine birds and mammals wintering in the Kodiak area of
Alaska. U.S. Fish and Wildlife Service, Office of Biological
Services, Washington, D.C. FWSIOBS-81113. 81 pages.

Haddock, J.L., C. Evans, L.W. Sowl, M.E. Isleib, P. Havens,
J. Reynolds, and D. Johnson. Unpublished 1971 Prince William
Sound aerial survey data. U.S. Fish and Wildlife Service,
Anchorage, Alaska.

Hogan, M.E. and J. Murk. 1982. Seasonal distribution of marine
birds in Prince William Sound, based on aerial surveys, 1971.
U.S. Fish and Wildlife Service, Anchorage, Alaska. Unpublished
Report.

Irons, D.B., D.R. Nysewander, andJ.L. Trapp. 1988. Princewilliam
Sound sea otter distribution. U.S. Fish and Wildlife Service,
Anchorage, Alaska. Unpublished Report, 31 pages.

Irons, D.B., D.R. Nyeswander, and J.L. Trapp. ms. Prince William
Sound waterbird distributions in relation to habitat type.
U.S. Fish and Wildlife Service, Anchorage, Alaska. 24 pages.

Klosiewski, S. and L. Hotchkiss. 1990. Assessment of injury to
waterbirds from the Exxon Valdez oil spill: Surveys to
determine distribution and abundance of migratory birds in
Prince William Sound and the Northern Gulf of Alaska. Bird
Study Number 2. Preliminary Status Report. U.S. Fish and
Wildlife Service, Anchorage, Alaska.

Nishimoto, M. and B. Rice. 1987. A re-survey of seabirds and
marine mammals along the south coast of the Kenai Peninsula,
Alaska during the summer of 1986. U.S. Fish and Wildlife
Service, Alaska Maritime National Wildlife Refuge, Homer,
Alaska. Unpublished Report, 79 pages.

BUDGET

Category Aerial Survey Boat Survey Total
Budget Budget

Personnel $59.0 $207.0 $266.0
Travel 5.0 14.0 19.0
Contractual 47.0 64.0 111.0
commodities 3.0 60.0 63.0
Equipment 0.0 12.0 12.0

Total $114.0 $357.0 $471.0

285

BIRD STUDY NUMBER 3

Study Title: Population Surveys of Seabird Nesting Colonies in
PWS, the Outside Coast of the KP, Barren Islands,
and Other Nearby Colonies.

Lead Agency: FWS

INTRODUCTION

The 1989 oil spill in PWS and its drift to the west along the
Alaskan coast prompted resurvey in 1989 of seabird colonies in PWS,
the Chiswell Islands, the Barren Islands, sites along the Alaska
Peninsula, some sites near Kodiak Island, and the Semidi Islands.
Most of these colonies had been censused at least two and up to six
different years out of the last 17 years, providing some base line
for determining changes that may have occurred on the colonies
surveyed. Murres and kittiwakes on one colony site, Middleton
Island, have been censused 11 of the last 17 years.

Diving seabirds are known to be easily impacted by oil spills (King
and Sanger, 1979). In addition, these species are long-lived with
low reproductive rates, thus making any mortality of adults a
critical factor in these species' ability to recover from loss.
There are at least 320 seabird colonies that occur within the area
affected by the oil spill. They contained about 1,121,500 breeding
seabirds of which approximately 319,000 were murres (USFWS, Catalog
of Alaskan Seabird Colonies --Computer Archives 1986). Some of
these colonies are among those most visited by tourists in Alaska.
Cliff-nesting seabirds, like murres, are an important part of this
human use/tourism.

This year, the study will concentrate more on murre populations and
less on other species than it did last year. This will facilitate
increased replicate counts and improve statistical evaluation. The
breeding season censuses will occur at the Chiswell Islands, Barren
Islands, specific sites along the Alaska Peninsula and the Semidi
Islands. Egg laying is the accepted time to census murres since
their colony attendance is too variable at other times (Hatch and
Hatch, 1988; Byrd, 1989; Nysewander, 1989).

OBJECTIVES

Determine whether the numbers of selected species of breeding
colonial seabirds within the oiled area have decreased compared to
numbers previously censused at these sites. Non-oiled nesting
colonies will be surveyed as a control.

286

METHODOLOGY

Assessment of injury to murre populations is being conducted in
four general areas: 1) Chiswell Islands, 2) Barren Islands, 3)
Alaska Peninsula sites, and 4) Semidi Islands, a non-oiled control
area. Middleton Island is not part of this particular
investigation, but an ongoing study at that location will provide
numbers of murres present on those colonies for comparison with
affected areas. Changes in numbers of breeding adult seabirds will
be documented, with primary emphasis placed on murres.

Total counts are not feasible at large colonies like the Semidi and
Barren Islands; hence previously established plots will be
utilized. Two strategies will be used: 1) counts of adult murres
on plots from land-based observation points; and 2) counts from
boat-based observation vantage points where land-based observations
are not possible. These plots may also serve as a correction
factor for total counts or estimates that may be needed for
comparison with past estimates.

Specifically, the two strategies used in 1989 will be implemented
again in 1990 (Nysewander, 1989), as follows:

1) A combination of total subcolony or island counts and counts
of sample plots counted from boats will occur at colony sites
like the Barren Islands and Chiswell Islands because the
colonies are much larger, in very exposed waters, have a poor
history of censusing, and require counts from boats. Previous
sample plots were established on the basis of accessibility
and visibility.

2) Land-based plots will be continued at the Semidi Islands
because these colonies are too large for total counts, and
land plots are feasible and have been used for over 10 years.
Sample plots were previously selected on the basis of
accessibility. The Alaska Peninsula murre colonies will
require a combination of both applications since some portions
of the colonies are visible from land, but most aspects of the
colony will require boat counts.

Colonies will be recensused using standard FWS methodology for
either land-based or boat-based counts of seabirds (Byrd 1989;
Hatch and Hatch 1988 and 1989; Irons et al. 1987; Nishimoto and
Rice 1987) . Efforts will be made to complete at least three
replicate counts of colonies or plots after eggs are laid. Counts
will be conducted on three separate days between the hours of 1000
and 1600. Plot counts and photographs of plots will be used to
establish correction factors of total subcolony or island counts,
comparisons with past counts of plots, and for evaluation of future
recovery or change.

At least three observers will make the counts by binoculars from

287

the boat. Each observer counts each section of the cliff at least
two times and all counts are compared to determine if sections of
the plot were missed (differences in counts by observers cannot be
greater than 5%) and need more replicate counts.

Standard procedures and assumptions used by the FWS for colony
counts in the Alaska Maritime National Wildlife Refuge are
described by Garton 1988 and Byrd 1989. Several key assumptions
are: 1) Plots, by necessity, are not random and selection is based
on accessibility; hence this study makes the assumption that counts
within plots are representative of the way the counts varied on the
entire colony; 2) Counts of plots or entire colonies from boats
are very difficult for large colonies and replications of counts by
several observers on the same day and different days illustrate the
need to refine the accuracy and the variation recorded. This means
that even counts of entire colonies are also considered to be
indices, but this study assumes that changes in these indices
represent the changes occurring in the colony; 3) Counts of things
are very unlikely to be normally distributed and are more likely to
be skewed and clumped. This type of data requires either very
large sample sizes, the use,of a non-parametric test, or the data
needs to be transformed logarithmically and then tested by the
appropriate parametric test. This transformation normalizes the
data and is required for valid application of statistical tests on
small sample sizes (Fowler and Cohen 1986, D. Robson pers. comm.).

Standard FWS procedures mentioned prefer to compare trends between
years using numerous replicate counts where all plots are censused
each count day and these counts are replicated on successive days.
The average of daily counts on the Semidi Islands will be used to
calculate a confidence interval for the estimate as was done on the
Semidi Islands data in the past (Hatch and Hatch 1988; Hatch and
Hatch 1989). At other sites where there were fewer replicate
counts, past procedures which averaged available counts, will be
followed. The important question is whether the "after oiling"
response is outside the anticipated annual variation in colony
numbers that would be expected from past historical data without
oiling effects.

BIBLIOGRAPHY

Byrd, G.V. 1989. Seabirds in the Pribilof Islands, Alaska:
trends and monitoring methods. M.S. thesis, Univ. of Idaho,
Moscow, Idaho. 96pp.

Garton, E.O. 1988. A statistical evaluation of seabird monitoring
programs at three sites on the Alaska Maritime National
Wildlife Refuge. Univ. of Idaho, Moscow, Idaho. Unpubl. Rept.
from contract with the refuge. 15pp.

288

Fowler, J. and L. Cohen. - 1986. Statistics for Ornithologists.
British Trust for Ornithology, BTO Guide No. 22, Tring,
Hertfordshire. 175pp.

Hatch, S.A. and M.A. Hatch. 1989. Attendance patterns of common
and thick-billed murres at breeding sites: implications for
monitoring. J. Wildl. Mgmt. 53(2):483-493.

Hatch, S.A. and M.A. Hatch. 1988. Colony attendance and
population monitoring of black-legged kittiwakes on the Semidi
Islands, Alaska. Condor 90:613-620.

Irons, D.B., D.R. Nysewander, and J.L. Trapp. 1987. Changes in
colony size and reproductive success of black-legged kit-
tiwakes in Prince William Sound, Alaska, 1972-1986. U. S.
Dept. Interior, Fish and Wildl. Serv., Anchorage, Alaska,
Unpubl. Rept. 37pp.

King, J.G. and G.A. Sanger. 1979. Oil vulnerability index for
marine oriented birds. Pp. 227-239 in J.C. Bartonek and D.N.
Nettleship eds. Conservation of marine birds of northern North
America. U. S. Fish and Wildl. Serv., Washington, D.C.
319pp.

Nishimoto, M. and B. Rice. 1987. A re-survey of seabirds and
marine mammals along the south coast of the Kenai Peninsula,
Alaska during the summer of 1986. U. 'S. Fish and Wild. Serv.,
Alaska Maritime National Wildl. Refuge, Homer, Alaska.
Unpubl. Rept. 79pp.

Nysewander, D. 1989. Population surveys of seabird nesting colonies
in Prince William Sound, the outside coast of the Kenai
Peninsula, Barren Islands, and other nearby colonies.
U.S. Fish and Wildlife Serv., Alaska Maritime National
Wildlife Refuge, Homer, Alaska. Unpubl. Rept. 52pp.

U. S. Fish and Wildlife Service. 1986. Catalog of Alaskan seabird
colonies--computer archives. U.S. Fish and Wildl. Serv.,
Migratory Bird Management, Anchorage, Alaska 99503.

BUDGET: FWS

Personnel $87.5
Travel 12.3
Contractual 111.2
Commodities 24.1
Equipment 16.0

Total $251.1

289

BIRD STUDY NUMBER 4

Study Title: Assessing the Effects of the EVOS on Bald Eagles

Lead Agency: FWS

INTRODUCTION

The area af f ected by the EVOS provides year-round habitat f or
approximately 5000 adult bald eagles and seasonal habitat for an
additional estimated 2500 immatures. An unknown number of bald
eagles from breeding areas in south-central Alaska probably also
winter in the PWS.

Bald eagles are closely associated with intertidal habitats that
have been heavily impacted by the EVOS. Nearly all nests in the
spill area occur within 100 meters of the beach and eagles commonly
forage in intertidal habitats on fish and marine invertebrates.
Eagles that breed elsewhere, but spend winters in the spill area,
will also use the affected intertidal habitats for foraging.

Contamination of these intertidal habitats may result in serious
impacts to bald eagles. Effects may include direct mortality of
adults and immatures from ingestion of oil -contaminated food or
from preening of oil from feathers. Eagles that become heavily
oiled or entrapped in oil may die. Mortality of embryos can occur
when eggs are contaminated with oil carried to the nests on the
plumage of the adults. Decreases in the abundance of prey such as
herring, eulachon, salmon, or marine invertebrates may increase the
vulnerability of eagles to starvation or to disease induced by
weakened physical condition. Significant losses of breeding
adults, eggs, nestlings and non-breeding eagles are expected.

This study is designed to document the magnitude and duration of
these impacts and determine whether these impacts are a result of
oil contamination. Estimates for the number of eagles occupying
the spill area after the spill will be compared with historical
data to identify changes in the population. Occupancy and
reproduction surveys will be conducted to determine productivity
and to document differences in production between oil-affected and
non-oiled areas. Nestling and adult bald eagles from oiled and
non-oiled areas will be radio-tagged and monitored to estimate
survival ratest distribution and exposure to oiled areas, and
determine causes of mortality.

Because eagles mature slowly and are long-lived, impacts to the
population may not be readily apparent. It may require an extended
period of study to substantiate the long-term impact of oil
contamination on bald eagles.

290

OBJECTIVES

A. Estimate numbers of resident bald eagles such that the
estimate is within 10% of the actual size 95% of the time-,
determine whether changes in population size have occurred in
the oil-affected areas since 1982 and test whether the change
in number of eagles in oil-affected areas is different than
changes in non-oiled areas.

B. Test the hypothesis that productivity of bald eagles is the
same in oiled and non-oiled areas (a = 0.05).

C. Test the hypothesis that survival rates are the same for bald
eagles in oiled and non-oiled areas (a = 0.05).

D. Determine toxic and sublethal effects of oiling on eagles and
eggs.

METHODS

Population surveys (objective A). Surveys of randomly selected
plots roughly 50 square miles in size will be conducted from
Malaspina Glacier to Cape Elizabeth in early May, following
methodology discussed in Hodges et al. (1984). All shorelines in
each selected plot will be flown at an altitude of about 200 feet
and an airspeed of 90-100 knots using fixed-wing aircraft. Eagles
will be classified as either white-headed or immature. "White-
headed" eagles will include sexually mature adults and near-adults
that have predominately white heads. This survey will not directly
estimate the number of immatures, therefore we will assume that our
ability to detect all age classes is equal for birds in flight, and
a ratio of adults to immatures observed flying will be used to
estimate the number of immatures.

A parametric two-sample t-test (Steel and Torrie 1960) which does
not require equal variances will be used to test the above
hypotheses. Analysis of variance will be used for multiple
comparisons. Assumptions necessary for valid application of the t-
test will be checked (e.g., test for normality). If assumptions
are violated, we will use either an appropriate transformation or
an equivalent non-parametric test.

Productivity surveys (Obiective B) . Two surveys to determine
productivity will be conducted in the oil spill area (PWS, KP,
Kodiak/Afognak Islands and the Alaska Peninsula) and in the Copper
River Basin, an area used by eagles that may winter in the oil
spill area. The first aerial survey will be flown during mid-May
to estimate the number of adults that attempt to breed. The second
survey will be flown in mid-July to estimate the number of
successful nests and the number of young produced. Surveys will be
conducted from helicopter at an altitude of 80-200 feet at 40-60

291

kts. airspeed to determine nest status. Data collected will
include number of nests surveyed, number of nests occupied, number
of nests that successfully produce young, and number of young
produced (Postupalsky 1974).

Nests that fail will be climbed to collect dead eggs or nestlings
and to identify the cause of failure.

The hypothesis that there is no difference in the observed
production among treatment groups compared with what would be
expected if nests were assigned randomly to each of the treatment
groups will be tested using a non-parametric permutation test.

Beaches within 1/2 mile of bald eagle nest sites are representative
of bald eagle home range in coastal Alaska (USFWS, unpubl. data).
The length of shoreline within the "home range" will be measured
and the lengths of segments classified as heavily, moderately,
lightly, or unoiled will be totalled for each "home range". These
values will be used in a multiple regression to identify
relationships among the degree of oiling, nest occupancy, and
productivity parameters. Data on productivity from the Copper
River Basin will be compared with data from coastal areas.
Productivity data from southeastern Alaska will also be used for
comparative purposes.

Survival Studies (Objective C): To estimate survival rates, 60
eagles (15 adults and 15 nestlings each from oiled and non-oiled
areas) will be tagged with radio transmitters (Pollock et al.
1989). Approximately 15 adult bald eagles will also be radio-
tagged in the Copper River Basin to demonstrate whether ' eagles
breeding inland winter in areas affected by the oil spill. Weekly
aerial flights will be made to relocate the transmitters using
standard telemetry techniques (Gilmer et al. 1981) and to document
eagle numbers, distribution and mortality within the study,area.
Dead eagles will be retrieved and necropsied to determine the cause
of death.

A Z-test (Bart and Robson 1982) will be used to test for
significant differences in survival rates between eagles marked in
oiled areas and eagles marked in unoiled areas. This Z-test
requires the use of a transformation of the survival rate and
standard error to normalize its distribution and allow use of a'Z
statistic to test for differences in survival rates. Accurate
relocations of individual radio-marked eagles will allow
appropriate classification of eagles into treatment groups based on
the proportional amount of time they were located in- oiled or
unoiled areas.

Toxic and Sublethal Effects of Oiling (Oblective DL: All eagles
found dead will be collected and necropsied to determine the cause
of death, to note the extent of oiling and to look for ingested oil

292

or other signs of oil contamination. Tissue samples from the
collected specimens will be analyzed for contaminants.

Unhatched eggs collected from failed nests will be examined for oil
contamination of eggshells, egg contents, and the presence and
development of embryos.

Blood samples from free-flying birds will be collected by properly
trained personnel and analyzed to determine concentrations of
hydrocarbons and other contaminants associated with oil
contamination. Approximately equal numbers of bald eagles will be
sampled from oiled and non-oiled areas. Blood samples will also be
analyzed for standard blood chemistry profiles, which will help
identify sublethal impacts.

Significant differences in levels of contaminants and blood
characteristics between bald eagles from oiled and non-oiled areas
will be tested using a 2-sample t-test (a = 0.05). Assumptions
necessary for valid application of the t-test will be checked
(e.g., test for normality) . If assumptions are violated, either an
appropriate transformation or an equivalent non-parametric test
will be used.

BIBLIOGRAPHY

Bart, J. and D. S. Robson. 1982. Estimating survivorship when the
subjects are visited periodically. Ecology 63:1078-1090.

Bortolotti, G. R. 1984. Physical development of nestling bald
eagles with emphasis on the timing of growth events. Wilson
Bull. 96:524-542.

Carpenter, in press.

Gilmer, D. S., L. M. Cowardin, R. L. Duvall, L. M. Mechlin, C. W.
Shaiffer, and V. B. Kuechle. 1981. Procedures for the use of
aircraft in wildlife biotelemetry studies. U. S. Fish and
Wildlife Service Resource Publication 140. 19 pp.

Hodges, J. I., J. G. King, and R. Davies. 1984. Bald eagle breeding
population survey of coastal British Columbia. J. Wildl.
Manage. 48:993-998.

Kaplan, E. L. and P. Meier. 1958. Non-parametric estimation from
incomplete observations. J. Am. Stat. Assoc. 53:457-481.

Pollock, K. H., S. R. Winterstein, C. M. Bunck, and P. D. Curtis.
1989. survival analysis in telemetry studies: the staggered
entry design. J. Wildl. Manage. 53:7-15.

293

Postupalsky, S. 1974. Raptor reproductive success: some problems
with terminology. Raptor Research Report 2:21-31.

Steel, R.G.D. and J.H. Torrie. 1960. Principals and procedures in
statistics. McGraw-Hill, New York. 481 pp.

Budget: FWS

Personnel $120.0
Travel 17.0
Contractual 503.0
Commodities 25.0
Equipment 10.0

Total $675.0

294

BIRD STUDY NUMBER 5

Study Title: Impact Assessment of the EVOS On Peale's Peregrine
Falcons

Lead Agency: FWS

Cooperating Agency: ADF&G

INTRODUCTION

Peale's falcons (Falco peregrinus pealei) occur along the
southern coast of Alaska from the Aleutian Islands through
southeastern Alaska. The goal of this project is to determine
whether the EVOS has had, or will have, a measurable impact on
Peale's peregrine falcons in PWS and coastal KP.

Peale's falcon populations in Alaska have been estimated at
between 500-600 pairs (Schempf 1989, Ambrose pers. comm). An
estimated 40-60 pairs inhabit PWS and coastal KP (Janik & Schempf
1985) and another 20-30 pairs occur in the Kodiak Archipelago,
upper AP, and CI area, for a total of 60-90 pairs in coastal
habitat affected by EVOS.

Alcids, small gulls, and petrels are prime peregrine prey species
that became oiled as a result of EVOS and may be taken by fal-
cons. Oil transferred to peregrine falcons could affect in-
dividuals and the population through: 1) coating of feathers
and the resulting loss of insulation and flight capabilities; 2)
reduced reproduction due to ingestion of hydrocarbons and trace-
metals that affect the breeding physiology of adults; 3)
reduced reproduction due to transfer of oil from feathers of
incubating adults to eggs; 4) mortality of individuals due to
toxicity of oil; and 5) reduced reproduction due to reduced
prey availability when prey populations are impacted.

This project will continue to provide information on the number
of nest sites occupied by Peale's falcons and their productivity.
These data, in combination with historical data for this area,
will provide a basis to evaluate whether changes have occurred in
the distribution, abundance, and/or productivity of falcons.
Examination of secondary wing feathers taken from young, along
with prey remains and eggs collected from occupied aerie, will
provide evidence of whether crude oil was ingested or absorbed by
falcons. Analysis of wing feathers and prey remains collected
several months after the oil spill will provide information on
the bioaccumulation of trace-metals from crude oil, in marine and
terrestrial food chains.

295

OBJECTIVES

A. Test the hypothesis that nest site occupancy and
productivity are lower in the project area as a result of
EVOS than in other populations.

B. Test the hypothesis that the quantities of vanadium and
nickel in peregrine feathers are the same for birds nesting
in oiled and non-oiled areas.

C. Count and identify prey remains collected at aeries.

D. Test the hypothesis that the level of pesticide
contamination of egg clutches in the project area is less
than contamination levels reported in scientific literature
as causing reproductive failures in peregrine falcons.

METHODS

The project area will include the mainland shore and islands of
PWS from Cape Hinchinbrook along the southern coast of the KP
through Kachemak Bay.

Two surveys of the project area will be conducted. Guidelines
have been formulated to standardize survey techniques,
terminology,. and data collection. The initial survey, to
determine presence or absence of peregrines at coastal bluffs and
to collect fresh egg samples for contaminant analysis, will take
place in mid-April. A boat and helicopter will be used for
transportation. If a helicopter is used at sites with large
concentrations of cliff nesting seabirds, it will land far enough
away from bluffs to minimize disturbance. Observers will
approach on foot to survey potential nesting habitat.

The second survey via helicopter, in June, will cover the same
area, but will focus on the sites determined to be occupied by
peregrine falcons during the initial survey. Nests will be
located by observers on the ground and then visited to collect
feather samples and to band nestlings with standard aluminum
bands. Secondary feather samples from young will be collected at
aeries during the June nest survey. Prey remains and addled or
broken eggs will be collected at nest sites. During both sur-
veys, investigators will document oil on falcons and look for
bands on adults to learn where they were banded.

Twenty-five prey remains will be examined for hydrocarbon con-
tamination. Samples collected for hydrocarbon analysis will be
handled carefully. Chain-of -custody will be maintained for all
samples, and they will be stored in a secure facility at ADF&G in
Anchorage until they can be sent to an approved laboratory for
analysis.

296

Feathers grown by nestling peregrine falcons should contain
trace-elements in an array of concentrations unique to the local
ecosystem (Parrish et al. 1983). High levels of nickel and
vanadium have been associated with North Slope crude oil and
these trace-metals are bioaccumulated in marine and terrestrial
food chains (Minerals Management Service 1988). Predators at the
top of food chains, such as the peregrine falcon, may encounter
toxic levels of trace-metals because these elements are con-
centrated with each step up the food chain. Toxic quantities of
trace-metals have been implicated in population decline- of
peregrines and other raptors (Newton- 1979). Elevated levels of
nickel in the diet will produce physiological effects similar to
lead or mercury poisoning such as central nervous system disor-
ders and reduced reproductive success (Williams, pers. comm.)
Traces of these metals can be measured efficiently in birds
feathers by instrumental neutron-activation analysis (INAA)
(Wainerdi and Dubeau 1963). Feather samples from peregrines not
influenced by the oil spill from other regions of the state will
serve as controls.

Approximately 30 feather samples will be collected for trace-
metal analysis. The distal 1 cm of the fifth secondary remige
will be collected from adult and nestling peregrines for INAA as
described by Parrish et al. (1983). Feather samples will be
labeled and preserved. Chain-of -custody will be maintained for
all samples and they will be stored in a secure facility at
ADF&G in Anchorage until they can be sent to an approved laborat-
ory for INAA.

The decline of peregrine populations in North America during the
1950's through the early 1970's was linked to organochlorine
pesticides (Hickey 1969). Substantial levels of biocides have
been found in Peale's falcons in coastal British Columbia and it
has been suggested that the depressed reproductive success on
Langara Island was largely due to the effects of pollutants
(Nelson & Myres 1976). Since trace-metals may affect
reproduction in peregrines, similar to organochlorine pesticides,
a pesticide monitoring program would help identify which factors
are involved. Thus collection of fresh eggs is necessary for
pesticide analysis.

Historically, about 35 aeries are thought to be occupied each
year in the project area. The collection of 10 eggs will provide
an adequate sample without significantly impacting productivity.

Chain-of-custody will be maintained for all samples and they will
be stored in a secure facility until they can be sent to an
approved laboratory for chemical analysis as described by
Cromartie et al. (1975) and Kaiser et al. (1980).

Data analysis to achieve objective A involves a comparison of
site occupancy and productivity in the project area with other

297

peregrine populations. In order to control for yearly variation
in brood size, the mean brood size of peregrines in Norton Sound
for 1989 and 1990, will be compared to the historical data with
an analysis of variance (ANOVA) (Snedecor and Cochran 1980) using
the appropriate linear contrasts to test the hypothesis that the
mean brood sizes are equal between the historical data and Norton
Sound. Assuming the Norton Sound data do not differ from the
historical data, a two sample T-test will be used to test the
null hypothesis that mean brood size in PWS in 1990 is greater
than or equal to mean brood size in Norton Sound. The
alternative hypothesis is that mean brood size in 1990 in PWS is
less than mean brood size in Norton Sound. A similar analysis
will be done for the 1990 productivity data.

ANOVA and a two sample T-test have the following assumptions:

1) The samples are random and independent;
2) The distribution of the different means is normal; and
3) The variances of the samples are equal.

A Q-Q plot (Hoaglin et al. 1985) of the raw data will determine
whether the data is approximately normal, in which case the
Central Limit Theorum will insure that assumption two is met. If
assumption two is not met, a non-parametric test will be employed
(Conover 1980). Bartlett's statistic will be used to test
assumption three and a transformation employed, if necessary, to
meet this assumption.

Two separate Fisher's exact test (Ostle and Mensing 1982) will be
used to determine whether PWS had lower nest occupancy rates than
Norton Sound for 1989 and 1990.

Data analysis to achieve objective B involves a two sample T-test
(Snedecor and Cochran 1980) to determine whether trace-metal
concentrations are lower in the project area than outside the
project area. The null hypothesis is that nickel and vanadium
concentrations in peregrine feathers from the project area in
1990 is less than or equal to nickel and vanadium concentrations
in peregrine feathers from elsewhere in Alaska in 1990. The
alternative hypothesis is that nickel and vanadium concentrations
in peregrine feathers from the project area in 1990 were greater
than nickel and vanadium concentrations in peregrine feathers
from elsewhere in Alaska in 1990.

For objective C, if hydrocarbon prey remains are observed, an
estimate of the proportion of contaminated prey remains will be
estimated at a 95% confidence interval. The confidence intervals
require that the proportion be normally distributed. if
necessary, transformations will be used to meet this assumption
(Snedecor and Cochran 1980).

The null hypothesis contained in Objective D states that levels

298

of pesticide contamination of peregrine eggs collected in the
project area in 1990 are greater than or equal to the levels of
pesticide contamination of peregrine eggs reported in literature
as causing reproductive failures (Peakall et al. 1975). The
alternative hypothesis states that levels of pesticide contamina-
tion of peregrine eggs collected in the project area in 1990 are
less than the reported levels of pesticide contamination of
peregrine eggs associated with reproductive failures. A one-
tailed, one sample T-test (Snedecor and Cochran 1980) will be
used to test this hypothesis. This test assumes the sample was
randomly collected and the mean has a normal distribution. if
necessary, either a transformation will be used to meet the
normality assumption or the Wilcoxon Signed Ranks test (Conover
1980) will be employed to test this hypothesis.

BIBLIOGRAPHY

Conover, W. J. 1980. Practical nonparametric statistics, 2nd
ed. John Wiley & Sons, New York, N. Y. 493 pp.

Cromartie, E., W. L. Reichel, L. N. Locke, A. A. Belisle, T. E.
Kaiser, T. G. Lamont, B. M. Mulhern, R. M. Prouty, and D. M.
Swineford. 1975. Residues of organochlorine pesticides and
polychlorinated biphenyls and autopsy data for Bald Eagles,
1971-72. Pestic. Monit. J. 9: 11-14.

Hickey, J. J. (ed) 1969. Peregrine Falcon Populations: their
biology and decline. Univ. of Wisconsin Press, Madison.
596 pp.

Hoaglin, D. C., F. Mosteller, and J. W. Tukey. 1985. Exploring
data tables, trends, and shapes. John Wiley & Sons, New
York, N. Y. 527 pp.

Janik, C. A. and P. F. Schempf. 1985. Peale's peregrine falcon
(Falco. peregrinus pealei) studies in Alaska, June 12-24,
1985. Raptor Management Studies, U.S. Fish and Wildlife
Service, Juneau, Alaska. 12 pp.

Kaiser, T. E., W. L. Reichel, L. N. Locke, E. Cromartie, A. J.
Krynitsky, T. G. Lamont, B. M. Mulhern, R. M. Prouty, C. J.
Stafford, and D. M. Swineford. 1980. Organochlrine pesti-
cide, PCB, and PBB residues and necropsy data for Bald
Eagles from 29 states - 1975-77. Pestic. Monit. J. 13:
145-49.

Minerals Management Service. 1988. Draft environmental impact
statement, Outer Continental Shelf Mining Program, Norton
Sound Lease Sale. Minerals Management Service, Anchorage,
Alaska.

299

Nelson, R. W. and M. T. Myres. 1976. Declines in populations of
peregrine falcons and their seabird prey at Langara Island,
British Columbia. Condor 78: 281-293.

Newton, I. N. 1979. Population ecology of raptors, Buteo Books,
Vermillion, South Dakota. 399 pp.

Ostle, B. and R.W. Mensing. 1982. Statistics in research 3rd ed.
The Iowa State Univ. Press, Ames IA. 596pp.

Parrish, J. R., D. T. Rogers, Jr., and F. P. Ward. 1983.
Identification of natal locales of peregrine falcons (Falco
peregrinus) by trace-element analysis of feathers. Auk 100:
560-567.

Peakall, D. B., T. J. Cade, C. M. White, and J. R. Haugh. 1975.
Organochlorine residues in Alaskan peregrines. Pestic.
Monit. J. 8: 255-260.

Schempf, P. F. 1989. Raptors in Alaska. Pages 144-54 in
proceedings of the western raptor management symposium and
workshop. National Wildlife Federation, Washington, D. C.
320 pp.

Snedecor, G. W. and W. G. Cochran. 1980. Statistical methods,
7th ed. Iowa State University Press, Ames, Iowa. 507 pp.

Wainerdi, R. E. and N. P. DuBeau. 1963. Nuclear activation
analysis. Science 139: 1027-1033.

Personal Communications

Ambrose, R. E. U.S. Fish and Wildlife Service, Endangered
Species Branch, Fairbanks, Alaska.

Williams, D. Quantum Medicine, Eagle River, Alaska.

BUDGET: FWS

Personnel $ 32.2
Travel & per them 3.5
services 69.5
Commodities 2.5
Equipment 0.0

TOTAL $107.7

300

BIRD STUDY NUMBER 11

Study Title: Injury Assessment of Hydrocarbon Uptake by Sea
Ducks in PWS

Lead Agency: FWS

Cooperating Agency: ADF&G

INTRODUCTION

This study will focus on the effects of petroleum hydrocarbon
ingestion by Harlequin Ducks (Histronicus histronicus), Barrow's
Goldeneyes (Bucephala islandica), Common Goldeneyes (Bucephala
.clangula), Black Scoters (Oidemia nigra), and Surf Scoters
(Melanitta perspicillata) in PWS as a result of the EVOS. PWS is a
major wintering area for these sea duck species (Isleib and Kessel,
1973). It is also an important migration area for sea ducks in
spring and fall and a breeding site for resident Harlequin Ducks
during the summer (Hogan, 1980). Harlequin Ducks in particular,
because of their resident status and intertidal foraging habits,
are considered substantially at risk to effects of the EVOS (King
and Sanger, 1979). Goldeneyes and Surf Scoters, although
migratory, are also - at , risk because of their intertidal and
subtidal foraging habits.

The five sea duck species included in this plan are heavily
dependent on intertidal and subtidal marine invertebrates (Vermeer
and Bourne, 1982). Surf Scoters and goldeneyes utilize blue
mussels (Mytilus) and, like Harlequins, consume a wide variety of
clams, snails, and limpets (Koehle, Rothe and Dirksen, 1982;
Dzinbal and Jarvis, 1982). Bivalves, particularly blue mussels,
and small clams (Macoma) are well-known for their ability to
concentrate pollutants at high levels (Shaw et al., 1976). The
EVOS may cause severe damage to marine invertebrates that support
sea ducks throughout the year (Stekoll, Clement, and Shaw, 1980)
and bioaccumulation in the food chain may result in uptake of
petroleum hydrocarbons by sea ducks over a long period (Dzinbal and
Jarvis, 1982; Sanger and Jones, 1982). This study will determine
levels of petroleum hydrocarbon ingestion by sea ducks, and will
predict resultant physiological and life-history effects (Hall and
Coon, 1988). A predictive model may be constructed for sea duck
reproductive losses, for instance, based upon physiological ef f ects
of petroleum contamination resulting from the EVOS. Pre-oil spill
baseline data is available on petroleum contaminant levels of
Harlequin Ducks in PWS (Irons, FWS, pers. comm.).

301

OBJECTIVES

A. Continue to develop a data base describing f ood habits of the
five species of sea ducks in PWS.

B. Obtain data from other NRDA studies of petroleum hydrocarbon
levels in marine invertebrates, particularly blue mussels,
from the PWS area; Relate this data to the levels of petroleum
hydrocarbons f ound by chemical analysis of invertebrates in
gut samples from sea ducks collected in oil spill and control
areas; and Test the hypothesis (at alpha = 0. 05) that the
incidence of petroleum hydrocarbons in gut samples from
collected sea ducks is higher in the oil spill areas than in
the control areas investigated in 1989-90 (Oil Year 1).

C. Estimate by chemical analysis the petroleum hydrocarbon levels
in collected sea duck tissues and body fluids within 10% of
the actual value 95% of the time.

D. Test the hypothesis (at alpha = 0.05) that the incidence of
petroleum hydrocarbons in tissues of collected sea ducks is
significantly higher in 1989-91 in the oil spill areas than in
the control areas.

E. Estimate the ingested petroleum hydrocarbon effects on
morbidity, mortality, and reproductive potential of sea ducks.
This information may be related to other studies to identify
changes' in abundance and distribution within the affected
areas.

METHODS

This study will compare levels of petroleum hydrocarbons in
tissues of five species of ducks collected in four study areas.
The areas exposed to petroleum are western PWS and southwestern
Kodiak Island. The unexposed control sites are southeastern PWS and
southeastern Alaska, north of Juneau. Only PWS and the Juneau
control site are to be investigated in Oil Year two (March 1990-
February 1991). Tissues will be collected for evidence of either
histopathology or chemical contamination. Additional seaducks will
be collected in "clean" areas within western PWS. This is to
provide a secondary control for ducks collected at known sites for
heavy oiling of seaduck intertidal forage species.

Seaduck collection in oiled areas of PWS will be integrated with
the collection sites of blue mussels in oiled areas in order to
demonstrate that seaducks feed in contaminated areas on
contaminated prey. Ten such sites are located in western PWS. At
each site where petroleum exposure status is documented for
intertidal organisms, approximately ten Harlequin Ducks are to be
collected in the summer of 1990. These ducks will be sampled for
petroleum contamination of food items in the proventriculus, as

302

well as histopathology.

Data on petroleum hydrocarbon levels in marine invertebrates
(especially blue mussels) and the degree. of oiling at selected
sites will be acquired from the Fish/Shellfish Studies 1 and 13,
the Coastal Habitat Study, the Air/Water Studies, and Technical
Services Study Number 3.

ANOVA (Snedecor and Cochran, 1980) will be used to test the
hypothesis that incidence of petroleum hydrocarbons in gut samples
from collected sea ducks is higher in-the oil spill areas than in
the control areas.

Cumulative logit loglinear models (William and Grizzle, 1972;
Agresti, 1984) on a per species basis will be used to model the
incidence of petroleum hydrocarbons using the area in which
collected as the explanatory variable. Hypotheses concerning
differences by area in incidence of petroleum hydrocarbons will be
tested with a conditional likelihood ratio statistic for nested
models (Agresti, 1984). A Bonferroni (Snedecor and Cochran, 1980)
Z-statistic (Agresti, 1984) will be used to determine the nature of
the differences among areas if the main effect is significant.

Physiological effects will be classified as none, slight, moderate
or severe. Loglinear models (Agresti, 1984) will be used to model
the distribution of physiological classification by area by
species. A conditional likelihood ratio statistic for nested
models will be used to test the hypothesis that physiological
classification is independent of area. If area and physiological
classifications are dependent, a Bonferroni (Snedecor and Cockran,
1980) Z-statistic (Agresti, 1984) will be used to determine
differences among areas while controlling for physiological effect.

If possible, an exact test for contingency tables (Agresti, et al,
1990) with ordered responses will be used to determine whether
ducks in the oiled area were in significantly poorer physiological
condition than ducks in the control area. If it is not feasible to
perform the exact test because of unavailability of appropriate
methodology, a cumulative logit analysis (Agresti, 1984) will be
used to test this hypothesis.

Tissues will be collected for either chemical analysis (presence,
absence, or degree of petroleum residue) or histopathology.
Results will be compared to unexposed specimens from "clean"
(unexposed control) areas. Choice of materials and tissues,
handling, and discussion of results will follow published,
guidelines for interpreting residues of petroleum hydrocarbons in
wildlife tissues (Hall and Coon, 1988).

303

BIBLIOGRAPHY

Agresti, A. 1984. Analysis of ordinal categorical data.
John Wiley & Sons, New York. 287 pp.

Agresti, A., C.R. Mehta, and N.R. Patel. 1990. Exact
inference for contingency tables with ordered categories. J.
Amer. Statist. Assoc. 85: in press.

Dzinbal, K.A. and R.L. Jarvis. 1982. Coastal feeding ecology
of Harlequin Ducks in PWS, Alaska, during summer. pp. 6 - 10
in Marine birds: their feeding ecology and commercial
Fisheries relationships. Nettleship, D.A., G.A. Sanger, and
P.F. Springer, eds. Proc. Pacific Seabird Group Symp.,
Seattle, WA., 6 - 8 Jan. 1982. Can. Wildl. Serv. Spec. Publ.

Sanger, G.A. and R.D. Jones, Jr. 1982. Winter feeding
ecology and trophic relationships of Oldsquaws and White-
winged Scoters on Kachemak Bay, Alaska. pp. 20 - 28 in Marine
birds: their feeding ecology and commercial fisheries
relationships. Nettleship, D.A, G.A. Sanger, and P.F.
Springer, eds. Proc. Pacific Seabird Group Symp., Seattle,
WA., 6 - 8 Jan 1982. Can. Wildl. Serv. Spec. Publ.

Hall, R.J. and N.C. Coon. 1988. Interpreting residues of
petroleum hydrocarbons in wildlife tissues. U.S. Fish and
Wildl. Serv., Biol. Rep. 88(15). 8 pp.

Hogan, M.E. 1980. Seasonal habitat use of Port Valdez,
Alaska by marine birds. Unpublished administrative report.
U.S. Fish and Wildl. Serv., Anchorage, Ak. 25 pp.

Isleib, M.E. and B. Kessel. 1973. Birds of the North Gulf
Coast - Prince William Sound Region, Alaska. Biol. Pap. Univ.
Alaska 14.

King, J.G. and G.A. Sanger. 1979. Oil vulnerability index
for marine oriented birds. pp. 227-239 in J.C. Bartonek and
D.N. Nettleship (eds.). Conservation of marine birds in
northern North America. U.S. Fish and Wildl. Serv., Wildl.
Res. Rep. 11. Washington, D.C.

Koehl, P.S., T.C. Rothe, and D.V. Derksen. 1982. Winter
food habits of Barrow's Goldeneyes in southeast Alaska. pp.
1 - 5 in marine birds: their feeding ecology and commercial
fisheries relationships. Nettleship, D. N., G.A. Sanger, and
P.F. Springer, eds. Proc. Pacific Seabird Group Symp.,
Seattle, WA., 6-8 Jan. 1982. Can. Wildl. Serv. Spec. Publ.

Sanger, G.A. and R.D. Jones, Jr. 1982. Winter feeding
ecology and trophic relationships of Oldsquaws and White-

304

winged Scoters on Kachemak Bay, Alaska. pp. 20-28 in Marine
birds: their feeding ecology and commercial T17sheries
relationships. Nettleship, D.N., G.A. Sanger, and P.F.
Springer, eds. Proc. Pacific Seabird Group Symp., Seattle,
WA., 6-8 Jan. 1982. Can. Wildl. Serv. Spec. Publ.

Shaw, D.G., A.J. Paul, L.M. Cheek, and H.M. Feder. 1976.
Macoma balthica: an indicator of oil pollution. Mar. Poll.
Bull. 7 (2): 29-31.

Snedecor, G.W. and W. G. Cochran. 1980. Statistical
methods. Iowa State University Press. Ames, Iowa. 507 pp.

Stekoll, M.S., L.E. Clement, and D.G. Shaw. 1980. Sublethal
effects of chronic oil exposure on the intertidal clam Macoma
balthica. Mar. Biol. 57: 51-60.

Vermeer, K. and N. Bourne. 1982. The White-winged Scoter diet in
British Columbia: resource partitioning with other scoters.
pp. 30 - 38 in Marine birds: their feeding ecology and
commercial fisheries relationships. Nettleship, D.A.,G.A.
Sanger, and P.F. Springer, eds. Proc. Pacific Seabird Group
Symp., Seattle, WA., 6-8 Jan. 1982. Can. Wildl. Serv. Spec.
Publ.

Williams, O.D. and J.E. Grizzle. 1972. Analysis of contingency
tables having ordered response categories. J. Amer. Statist.
Assoc. Vol. 67: 55-63.

BUDGET: FWS

Salaries $75.0
Travel 25.0
Contracts 35.0
Supplies 5.0
Equipment 10.0

TOTAL $150.0

305

BIRD STUDY NUMBER 13

Study Title: Preliminary Survey of Passerine Birds in PWS to
Assess Impact of the EVOS.

Lead Agency: FWS

Cooperating Agency: ADF&G

INTRODUCTION

Several year-round resident passerine species are heavily dependent
upon shoreline and intertidal areas in PWS, and may have become
oiled and suffered injury. These species include the gray jay,
(Perisoreus canadensis) , Steller's jay (Cyanocitta Steller) , black-
billed magpie (Pica Pica), common raven (Corvus corax),
northwestern crow (Corvus caurinus), and others. Other migratory
passerines that use intertidal and shoreline areas also may be
similarly affected by oil contamination. These passerines include
swallows (Hirundini dae) , thrushes (Muscicap idea) , several species
of blackbirds and sparrows (Emberizidae), water pipits
(Motacillidae). These birds occur in the hundreds of thousands.
In addition to direct lethal effects of oiled plumage, birds may be
subject to sublethal effects of oil contamination, which could
affect overall health and reproductive potential. Passerine
species have major intrinsic and recreational (viewing) value.

This study is a reconnaissance survey only. It is designed to
provide preliminary information that will assist in determining
whether additional, more rigorous, studies of passerines are
needed. It was originally proposed in the 1989 damage assessment
plan to collect a wider scope of information. However, that study
was not implemented because of logistical constraints that arose
during the 1989 field season.

OBJECTIVES

A. observe, record, and report the presence or absence of
passerine species in oiled and non-oiled study sites in PWS.

B. Compare count data for 1990 between oiled and non-oiled sites.

C. Compare count data for 1990 with historical data.

METHODS

oiled and non-oiled shores will be selected for survey, with
selection cognizant of available data and logistical support.
observations of passerines will be recorded along oiled and non-

306

oiled beaches on Perry Island in PWS.

Counts and observations will be made from stationary locations and
transects that were established and surveyed in previous years.
This will allow comparison between pre-oiled years, 1982-86, and
post-oiled years 1989-90.

BUDGET: FWS

Personnel $ 3.0
Travel 0.5
Contracts 5.2
Supplies 0.8
Equipment 0.5

Total $10.0

307

ASSESSMENT OF DAMAGE TO HISTORIC PROPERTIES AND ARCHEOLOGICAL
RESOURCES

Lead Agency: USFS

Cooperating Agencies: DNR, FWS, NPS

INTRODUCTION

Holocene richness and diversity of resources resulted in the
development of the largest prehistoric populations in Alaska along
the Pacific mainland and island coasts. Kodiak Island had the
largest, most dense prehistoric population of Eskimo peoples in the
world. similar ecological abundance suggests PWS and mainland
coasts also supported major human populations. The region of oiled
beaches includes large areas where few archeological surveys have
been done. To determine damage, specific information is needed on
the location, number, and character of historic sites within the
EVOS area. This information is obtainable through intensive on-
the-ground sample surveys and direct testing.

OBJECTIVES

This study includes activities designed to identify and quantify
injury to cultural resources from a scientific standpoint and to
develop the foundation for a meaningful program to restore and
rehabilitate archeological resources. To determine the injury
caused by the spill the study will focus on the following:

A. Impacts on soil chemistry (pH, Calcium, Phosphate)

B. Impacts on soil structure and inclusions (stratigraphy;
charcoal)

C. Impacts on artifacts including petroglyphs, bone, wood,
ceramic, fiber and shell

D. Impacts on vegetative cover of sites, including new or
increased erosion on the sites

E. Occurrence of theft or vandalism on sites, including new or
increased incidences

METHODS

1. Activities will be performed in a manner consistent with the
Secretary of the Interior's Standards and Guidelines for
Archeology and Historic Preservation (48 Fed. Reg. 44716-
44740, September 29, 1983).

2. Through a literature search and in-field surveys, an estimate
of the number, type, character, and the significance of

308

historic properties in the area affected by the oil spill will
be determined.

3. Develop topologies based on site type, time period, and
location.

4. Using the topologies developed, a representative sample of
historic properties types and locations to be investigated for
impacts, will be selected. The sample will include sites in
non-oiled areas to serve as control sites.

5. Conduct archeological investigations at the selected sites and
locations.

6. Oil spill response workers and government employees will be
interviewed concerning impacts to historic properties and
archeological resources.

7. A laboratory analysis of the effects of the oil on the
physical characteristics of the soil column will be performed.
Attention will be given to its component parts to determine
changes in preservation, soil compaction, stratification, and
obscuration of the stratigraphy, as well as leaching and the
chemical breakdown of organic materials.

8. Radiocarbon age determinations and soil sample analyses for
pH, calcium, and phosphate will be performed.

9. Pre- and post-spill vandalism and erosion data will be
compiled and evaluated to establish rates and effects of
vandalism and erosion.

DISCUSSION

To assess the potential injury to historic properties along the
coast, three physical zones can be established: submerged (below
the lowest low tide), intertidal (between the lowest low and the
highest high tides), and shore margin uplands (above the highest
high tide). The greatest potential for damage exists through
direct deposition of oil in the intertidal zone. Secondary
transport into adjacent submerged areas and uplands may also injure
historic properties. Upland historic properties are also subject
to contamination from transportation of oil by wind, storm tide
inundation, migration of contaminants in ground water, oiled bird
and animal movement from the feeding/travel corridor of the
intertidal zone, and their death and decomposition on archeological
deposits. Theft of artifacts and vandalism to historic properties
and archeological resources is a potential danger in the intertidal
and upland zones. The intertidal zone contains historic properties
of great variety, numbers, and susceptibility to oil damage.
Shipwrecks, eroded/scattered artifacts, inundated stratified
archeological deposits, prehistoric rock art, prehistoric fish

309

weirs, and remnants of structures or objects deliberately placed in
the intertidal zone are among the site types known to exist. The
shore margin uplands may contain all the previously mentioned site
types, plus burials, above-ground structures, and recognizable
resource collection locations such as culturally modified trees.

In the two higher elevation zones, a major potential injury
resulting from oil contamination is interference with traditional
archeological dating techniques. Radiocarbon dating depends on
comparison of the ratio of radioactive carbon 14 to carbon 12 in
the sample being analyzed. Because petroleum contains abundant
radioactively- inert carbon from organisms dead for millions of
years, and the use of radiocarbon dating for dates up to 35,000
years ago, contamination by even a small amount of ancient carbon
is expected to result in age determinations that are significantly
older than the archeological event being dated. This would
seriously compromise radiocarbon dating as a technique for dating
human activities and paleoenvironmental events and conditions. The
potential for affecting age determination may be significant even
in areas where only a sheen exists and may be investigated in
assessing injury. In cases of oil contamination in stratified
archeological deposits, masking of the visibility and alteration of
the chemical components of the microstratigraphy may also affect
archaeologists' ability to trace strata.

Both direct and indirect injuries to historic properties may have
occurred from response and treatment activities, as well as from
increased activities in the resource areas. Further, increased
access of personnel to remote areas may have increased the
knowledge of site locations and potentially may accelerate
vandalism, theft of heritage resources, and damage to the
scientific value of the sites.

BUDGET: USFS

Personnel $ 382.0
Travel 12.0
Contracts 300.0
Equipment & Supplies 238.0

TOTAL $ 932.0

BUDGET:

Personnel $ 123.0
Travel 4.0
Contracts 96.0
Equipment & Supplies 77.0

TOTAL $ 300.00

TOTAL BOTH AGENCIES: $ 1,232.00

310

TECHNICAL SERVICES

The hydrocarbon analysis, histopathology, and mapping projects
described in this section are designed to provide high quality
technical services to studies described in other portions of the
NRDA plan. Hydrocarbon analytical services includes the
generation, archival, and retrieval of all chemical analytical
data. Histopathology involves continuing the quality assurance
program begun in 1989 and facilitating the analysis work of
approved contractors employed by individual 'studies. Mapping
includes implementing and managing a geographic information system
to record and process data collected by NRDA studies.

Appropriate information on exposure of the resource to hydrocarbon
residues from the spill is required to determine and quantify
injury. It can be demonstrated by detailed information on the
distribution and evolving chemical composition of the spilled oil
through time, in concert with analyses of petroleum hydrocarbons or
their metabolites in the tissues of organisms.

Samples of water, sediments and tissues for chemical analysis are
being collected by individual studies throughout the entire region
impacted by the EVOS. Selected samples are being analyzed by a
team of participating laboratories in accord with a centralized
QA/QC program (Appendix A) which will help ensure that all data are
of known, defensible, and verifiable quality and comparability.

Information on the incidence of oil-induced histopathological
conditions is required by many of the studies described under
Fish/Shellfish, Birds, Marine Mammals, and Terrestrial Mammals.
This information is being gathered under strict quality assurance
guidelines (Appendix B) by a group of contracted histopathologists
to ensure compatibility of results and evaluations throughout the
NRDA program.

The mapping project will continue to develop the damage assessment
geographic information system. Good progress has been made on the
collection and verification of the primary data layers.
Additionally, large scale production and transmittal of map
products has begun.

311

TECHNICAL SERVICES STUDY NUMBER I

Study Title: Hydrocarbon Analytical Support Services and
Analysis of Distribution and Weathering of Spilled
Oil

Lead Agency: USFWS, NOAA

INTRODUCTION

In order to document the exposure of resources in the PWS and GOA
ecosystems to spilled oil, NRDA projects are collecting samples
from a wide variety of environmental matrices to be analyzed for
petroleum hydrocarbons. The data resulting from the analysis of
these samples will not only be used to demonstrate an impact on
that particular resource and support the individual project, but
also to produce an integrated synthesis of the distribution of the
oil in space and time, i.e., provide information on subsurface
transport, residence time, and mass balance. Both of these uses
require that the analytical data be accurate and precise, as well
as of demonstrable quality. Analysis of the distribution of oil
requires that the data be comparable from project to project
through the entire NRDA process. The large number of samples and
the length of time involved in the NRDA process require the use of
more than one laboratory to provide analytical data. Rather than
make each project responsible for analyzing their samples, TS 1 is
responsiblefor analysis of all samples collected for hydrocarbon
chemistry. This requires the generation, archival, and retrieval
of all chemical analytical data.

To date, TS 1 has:

Developed and implemented an analytical chemistry QA/QC plan.

Developed and implemented a sample inventory and tracking
system. This system forms part of the overall database
management system for TS 1. Sample collection and quality
data and analytical and associated QA/QC data are other parts
of this system.

Contracted with the National Institute of Standards and
Technology (NIST) to develop and supply calibration standards
and control materials.

Initiated the measurement of petroleum hydrocarbons and their
metabolites in water, sediment, tissues and bile. Some
samples have already been analyzed and results comminuted to
PIS. The remaining samples in the inventory have been
assigned an analytical priority and will be released for
analysis in that order.

312

Conducted a performance audit with participating analytical
laboratories.

Initiated a synthesis of the distribution and composition of
spilled oil with TS 3.

OBJECTIVES

A. Measure petroleum hydrocarbons, hydrozarbon metabolites and
other appropriate chemical/biochemical measures of hydrocarbon
exposure in water, sediment, and biota. collected through the NRDA.

B. Prepare a QA/QC plan that establishes detailed procedures and
protocols' for sample collection, sample identification, chain of
custody, and shipping.

C. Oversee and develop a centralized QA/QC program to assist the
analytical laboratories in providing quality data and demonstrate
the accuracy, precision, and comparability of all data developed by
the program.

D. Provide technical on-site system audits of field and
laboratory data collection activities.

E . Develop and provide appropriate instrument calibration
standards and natural matrix control materials.

F. Develop an integrated synthesis of the distribution and
chemical compositionof spilled oil, as it weathers through-time,
to provide a detailed basis for final exposure assessment.

METHODS AND DATA ANALYSIS

Objectives'A-E.; @This information is provided i:n 7"Analytical
Chemistry QA/QC11 (Appendix A).

Objective F. Data will be integrated and displayed by means of
TS 3 mapping capability.

SCHEDULES AND PLANNING

Additional analytical support both in terms of number of
laboratories, and types of analyses, e.g. UV/P@are being actively
sought. At least one and perhaps two more facilities will be added
to add analytical capability.

TS I will conduct a series of training sessions on the oil spill
collection and handling of samples before the projects begin year
two field work.

A proposal to transfer bar-coding sample-tracking technology will
be considered.

313

Discussions on a sample holding time study have been initiated.

BUDGET: NOAA

Salaries $ 58.7
Travel 3.5
Contractual 832.5
Equipment 15.0
Supplies 4.5

Total $ 914.2

BUDGET: FWS

Salaries $ 62.0
Travel 5.2
Contractual 999.0
Equipment 7.0
Supplies 16.0

Total $ 1,089.2

314

TECHNICAL SERVICES STUDY NUMBER 2

Study Title: Histopathology: Examination of Abnormalities in
Tissues from Bird, Mammals, Finfish, and Shellfish
Exposed to the Spilled Oil

Lead Agency: ADF&G

INTRODUCTION

Histopathology is an important tool used in determining mechanisms
of death and sublethal effects caused by infectious agents and
toxic substances. A number of histopathological conditions are
known to result from exposure to oil. Evidence of these conditions
will be documented in tissue samples taken by individual NRDA
studies as one means of demonstrating spill-related injury.
Histopathology technical services will support that effort by
continuing the quality assurance program begun in 1989 and by
facilitating analyses by approved contractors.

OBJECTIVE

Measure the incidence of histopathological conditions and external
lesions in selected species of birds, mammals, finfish, and
shellfish collected by NRDA studies.

METHODS

Standard histological methods for collection, preservation,
processing, and interpretation, as specified in the quality
assurance program (Appendix B) , will be continued. Assistance will
be provided to NRDA studies in selecting and contracting with labs
or individuals to complete analyses.

BUDGET: ADFG

salaries $ 20.0
Travel 5.0
Contracts 70.0
supplies 5.0
Equipment - 0.0

Total $100.0

315

TECHNICAL SERVICES STUDY NUMBER 3

Study Title: Implement and Manage a Geographic Information System
(GIS) to Record and Process NRDA Data

Lead Agency: DNR and FWS

Cooperating Agency: USFS and DEC

INTRODUCTION

The purpose of Technical Services Number 3 (TS3) remains unchanged:
the group is charged with implementing and managing the geographic
information system (GIS) to record and process data collected in
NRDA studies. Good progress has been made on the collection and
verification of the primary data layers. Additionally, TS3 has
begun large scale production and transmittal of NRDA map products.

OBJECTIVES

1. Produce and disseminate requisite maps and analytical products
for participants in the natural resource damage assessment
process.

2. Create and maintain, throughout the damage assessment process,
a data base or data pertinent to the overall damage assessment
process, in a way that is accessible to all of the
participating agencies.

METHODS

Methods are essentially the same as described in the 1989 study
plan. In addition to the data layers described in the 1989 study
plan, data layers have been or will be prepared for study site
locations, sampling locations, beach segment locations and multi-
thematic atlases of pre-spill data from various sources.
Additional data layers will be added as needed by investigators and
the Trustee Council to enable geographic-based compilation of study
results and other pertinent data.

Quality control will continue to be emphasized, with review of
information in data layers against qualified data sources and full
documentation of source data and review procedures. A data backup
system which includes redundant backup and off-site storage has
been implemented and will be maintained.

316

BUDGET: DNR

Salaries $ 332.4
Travel 11.2
Contracts 69.6
Supplies 67.0
Equipment 112.0

Total $ 592.2

BUDGET: FWS

All Activities $ 200.0

Total $ 792.2

317

I
I

PART II

ECONOMICS

ECONOMIC STUDIES

The studies in this section are federal studies designed to assess
the economic value of injury to natural resources associated with
the EVOS. State studies designed to assess the economic value of
injury to natural resources resulting from the EVOS are not
discussed in this document.

The federal studies cover seven major areas: (1) commercial
fishing, (2) public land values, (3) recreation, (4) subsistence,
(5) intrinsic values, (6) research programs and (7) archaeological
resources.

318

ECONOMICS STUDY NUMBER I

Study Title: Commercial Fisheries Losses Caused by the EVOS

INTRODUCTION

This study combines the studies previously designated as Economics
Study Number 1 (Estimated Price Effects on Commercial Fisheries),
Economics Study Number 2 (Fishing Industry Costs) and Economics
Study Number 3 (Bioeconomic Models f.or Damage Assessment) with
primary emphasis on former Economics Study Number 1.

The EVOS may have resulted in substantially reduced seafood
production at, among others, Cordova, Seward, Kodiak, Kenai, and
Homer, which are some of the most important commercial fishing
ports in the United States. Both short term impacts, through
closure of certain fisheriesf and long term effects, such as
reductions in population that will not become apparent for several
years as well as through continued exposure to contaminants, may
occur. These impacts may affect both the supply of and demand for
seafood.

For example, changes in quality (both real and perceived) may have
occurred, which could adversely affect seafood markets. In the
case of several important commercial salmon fisheries, the spill
resulted in harvests being confined to "terminal" areas, thus
restricting traditional fishing patterns and timing of the harvest.

Terminal area harvests occur in close proximity to the salmon's
spawning grounds. The result can be a significant reduction in
quality, as compared to salmon harvested in more typical
circumstancef i.e., more distant from, but en route to, spawning
sites. The reduction in quality may affect the salmon's overall
marketability and/or its appropriateness and acceptability for
specific product forms. In either case, seafood consumers at every
market level incur losses.

Salmon is only one commercial species group which may have been
adversely affected. others may include Pacific halibut, Pacific
herring, sablefish, shellfish and groundfish.

OBJECTIVES

Measure the economic loss to seafood consumers caused by the EVOS.

METHODS

The species most affected by the spill must first be determined.
Conceptual models of consumer preferences and market
characteristics for certain seafood products must be created.
Furthermore, a methodology to assess statistically changes in the

319

level and quality of harvest must also be established. Next, the
data appropriate for the models must be collected, assembled and
assessed. The models will then be used to estimate the demand for
various seaf ood products, the price changes associated with the
spill, and the effects of seafood quality and quantity changes on
consumers.

BUDGET

Salaries $ 103.0
Travel 15.0
Contracts 95.0
supplies 7.0
Equipment 9.0

Total $ 229.0

320

ECONOMICS STUDY NUMBER 4

Study Title: Effects of the EVOS on the Value of Public Land

INTRODUCTION

The EVOS affected subtidal, intertidal, tidal and uplands areas on
the shores of PWS and the GOA. This study will assess the lost
market value of publicly held lands attributable to the oil spill.
It will estimate market demand for leases and sales of land in the
impacted areas, and project changes in total value of public lands.

OBJECTIVES

Determine the change in market values of public lands.

METHODS

Land appraisals are a common method of assessing the market value
of land. Appraisers usually estimate the market value of land
parcels from the selling price of similar parcels. Because no two
parcels are identical, adjustments are required to achieve
comparability. For the purposes of appraisal, market value is
generally defined as the amount in cash, or in terms reasonably
equivalent to cash, for which, in all probability, the property
would be sold by a knowledgeable owner who is willing but not
obligated to sell to a knowledgeable purchaser, who desires the
property but is not obligated to buy. Using this definition of
market valuel the effect of the oil spill on land values will be
estimated as the difference between the pre- and post- spill
selling prices.

This study will proceed in several stages. First, a conceptual
model will be developed to determine the total public value of a
land parcel and to show how land appraisals fit within that model.
Next an attempt will be made to identify instances where land
valu@ studies can provide estimates of value changes that are not
captured by other economic studies. Third, the adequacy of
appraisals as a method of assessing changes in land values due to
the spill will be evaluated.

If appraisals are warranted, they will be conducted by obtaining
data on ownership patterns in areas affected by the oil spill,
gathering data on previous oil spills and their effect on land
values, estimating the effect of the oil spill on the value of
property through use of paired-scale data, and inspecting areas
affected by the oil spill.

321

BUDGET

Salaries $ 96.0
Travel 20.0
Contracts 50.0
supplies 5.0
Equipment 9.0

Total $180.0

322

ECONOMICS STUDY NUMBER 5

Study Title: Economic Damages to Recreation

INTRODUCTION

The EVOS has impacted natural resources that support a wide range
of recreational activities including fishing, hunting, boating,
hiking, camping, and sightseeing. Because of their unique
attributes, these resources attract recreationists from throughout
the United States and other countries to PWS and the GOA coast.

The oil spill may result in economic damage to those resources,
recreational services in two principal ways: 1) some
recreationists who otherwise would have gone to the area will
choose a substitute activity and/or area, thereby potentially
suffering a loss in personal satisfaction and possibly incurring
increased costs, and 2) recreationists who visit the area may
suffer reduced satisfaction because of the oil spill's adverse
impacts on recreational services that the natural resources
otherwise would have provided. These types of losses may have been
experienced by sea kayakers, users of charterboat services,
recreational fishers, cruise ship patrons and general tourists.

While relatively few in number, sea kayakers may have been
significantly affected by the oil spill. Kayaking trips are taken
from Valdez, Kodiak, Homer, Whittier and Seward to the western
portion of PWS and the bays along the Kenai peninsula and Kodiak
Island. A typical trip involves charter boat transportation to a
site some distance from port. Most trips last more than one day
and thus include both kayaking and wilderness camping.
Southcentral Alaska includes some of the premier kayaking areas in
the world.

The potential effect of the oil spill on kayakers could take
several forms:

- beaches used for wilderness camping are oiled and unusable;
- wilderness scenery is despoiled and sense of pristine
environment is lost;
- wildlife viewing opportunities are reduced;
- areas un-oiled suffer from increased congestion;
- clean-up activities make boats for transport expensive or
impossible to charter; and
- clean-up activities spoil the wilderness nature of the
experience.

All of these potential effects may have applied during the 1989
season; some may remain for several years.

Recreational activities that use the services of charterboats and

323

other private boats for hire are typically less intense than sea
kayaking, but they are f ar more numerous. Vessels for hire and
charterboats range from the standard six passenger charterboat
called a 11sixpack" to large tour boats carrying over a hundred
passengers. All types of vessels for hire have been impacted by
cleanup activity. For brevity in this proposal, this entire group
is referred to as "charterboats." Charterboat related recreational
activities include salmon and halibut fishing, sightseeing and
viewing marine wildlife and ferrying for wilderness camping in the
PWS, KP and Kodiak areas. Charterboats go out of Valdez, Whittier,
Homer, Kodiak, Seward and the smaller villages in southcentral
Alaska.

Because access to the general area is not easy, there are
potentially substantial impacts which can be measured through a
careful study of the charter fleet. The purpose of such a study
would be to determine the reduction in the use of the PWS
environment through the charter fleet as a consequence of the oil
spill.

The level of participation in recreational fishing among the
residents of Alaska is far greater than among the residents of any
other state in the United States. Marine recreational fishing
originates in all major towns on the PWS as well as Cook Inlet,
Kodiak Island and the KP and the Alaska Peninsula. Fishing trips
are taken in several ways -- from shore, from private boats. and
from charter vessels. Because access by car from Anchorage is
relatively easy, shore fishing and private boat fishing 'on the
Kenai is quite popular. All kinds of fishing draw large numbers of
tourists to Alaska.

The previous study of@charterboats willaddress only part of the
potential recreational fishing effects. ' It is possible that the
oil spill had detrimental effects on shore and private boat
recreational fishing, as well. For example,

a) fishing trips in the potentially oiled areas may have declined
due to fear of contaminated fish and waters;
b) anglers may not have been able to find accommodations in areas
where they wanted to fish because of cleanup related
activities.,
c) -the value of particular fishing trips out of the potentially
oiled zones may have declined because sites became more
congested.

Each season, a number of cruise ships pass through PWS on their way
from Seattle or 'Juneau' to'Whittier where, they discharge their
passengers for the train trip to Anchorage.' The likelihood that.
these individuals were directly affected by the oil spill is small,
but many have canceled their trips, because of fear that the oil
spill would spoil the experience.

324

The general tourist activity sub-component of the proposal dif f ers
from the others in that it is not directed toward one specific
recreational activity. Here the goal is to determine, from
aggregate level data, the extent to which general tourist activity
in the area of the spill may have been dislocated because of clean-
up activities. There will have been losses to recreationists if
these activities were diverted away from areas thought to be
contaminated by the spill or affected by the congestion and lost
services associated with clean up. Some of the marine related part
of this damage will be captured in the investigation of the
charterboats and kayaking. However, those people who do not plan
to use boats but rather state parks or other facilities will not
have been covered.

OBJECTIVES

Develop estimates of economic injuries to recreationists.

METHODS

The study will look at the types of consumptive and nonconsumptive
recreational activities.

Sea kayaking: This study contains several stages. First, the
relevant sea kayaking population will be identified. Second, a
survey instrument which will contribute to both recreational demand
and simple contingent valuation analysis will be created. Third,
the survey instrument will be pretested. Fourth, the survey will
be administered. Fifth, the survey results will be analyzed.

Charterboat activities: Data for this study will also be collected
through a survey instrument. After development of a theoretical
framework for damage measurement, the sample frame will be defined.
A survey instrument will be designed to determine the periodic
recreational and cleanup activities undertaken by each charter
vessel, the number of recreationists served, the extent of
cancellations and the amount of time the vessel was involved in
clean up activity. Vessel owners may also be interviewed in
person. Finally, the data will be analyzed.

Recreational fishing: There is an existing model for recreational
f ishing in the KP area. This model will be investigated to
determine whether it can be usefully applied to the ef f ects of this
oil spill.

Cruise ship tours: Cruise ship firms will be contacted to
determine whether demand f or cruise ship tours to PWS were af f ected
by the oil spill. If there is evidence of substantial reductions
in demand, methods of estimating the actual losses to
recreationists will be explored.

General tourist activity: Assuming that aggregate ef f ects on

325

tourism may be accurately estimated, this study will compare those
aggregate effects with the results of the activity directed
substudies to determine whether important categories of losses have
been missed.

BUDGET

Salaries $ 229.0
Travel 27.0
Contracts 20.0
Supplies 8.0
Equipment 10.0

Total $ 294.0

326

ECONOMICS STUDY NUMBER 6

Study Title: Losses to Subsistence Households

INTRODUCTION

several communities on the shores of PWS, LCI, Kodiak Island, and
the Alaska Peninsula, and in or near the EVOS area, are highly
dependent upon noncommercial fishing, intertidal food gathering,
marine mammal hunting, and land mammal hunting for subsistence
uses. Among the small subsistence communities are Tatitlek,
Chenega Bay, English Bay, Port Graham, Ouzinkie, Port Lions, Larsen
Bay, Karluk, Akhiok, Old Harbor, and Chignik Bay. Larger
subsistence communities include Cordova, Valdez, Seldovia, and
Kodiak. Subsistence uses are defined as rural Alaska residents'
customary and traditional uses of wild, renewable resources for
direct personal or family consumption as food, shelter, fuel,
clothing, tools, or transportation; for the making and selling of
handicraft articles out of nonedible byproducts of fish and
wildlife resources taken for personal or family consumption; for
barter, or sharing for personal or family consumption; and for
customary trade. Those uses are designated as the priority public
consumptive use of wild resources.

Following the oil spill, subsistence harvests were reduced in
several communities because of health concerns. This could have
important ramifications in the economy and social order of the
communities. Potentially important economic losses to the
communities include (1) subsistence losses, (2) local inflation
affecting harvests and food procurement, (3) damage to subsistence
property and (4) loss of intrinsic value to subsistence users.

OBJECTIVES

A. Conduct a literature review and compile base line information.

B. Document extent of oil contact and clean-up on or near
historic harvest sites.

C. Document changes in subsistence use through time (i.e.,
species selection; harvest timing, quantities, areas, methods,
and efficiency; and household participation rates in harvest,
use, sharing, barter, and exchange)-

D. Document local social and economic changes that affect
subsistence use, including wage/labor patterns, income levels,
inflation rates in the -villages for goods and services,
cleanup work, outside agency demands, and industry demands.

E. Assign monetary values to losses to subsistence households.

327,

METHODS

Field observations and interviews will be used to collect
information. Changes in subsistence use and socioeconomic patterns
will be determined by conducting systematic household surveys and
interviews, and comparing these data to historic information.
Where applicable market prices and price imputation will be used to
estimate damages. For marketed goods, the cost of replacing the
goods injured by the spill will normally be the measure of economic
damage. However, the adverse effects of the spill extended beyond
marketed goods. To determine the economic damages to non market
goods and services, survey methodologies, similar to those
described in Study 7, will be employed.

BUDGET

Salaries $ 255.5
Travel 48.0
Contracts 237.5
Supplies 151.4
Equipment 192.6

Total $ 885.0

328

ECONOMICS STUDY NUMBER 7

Study Title: Loss of Intrinsic Values Due to the EVOS

INTRODUCTION

Intrinsic values include existence value, option value, and bequest
value. These values are independent of the economic values arising
from direct use of natural resources and cannot be measured by
observing use of the area affected by the EVOS. Resources with
intrinsic values include fish, birds and mammals, along with the
wilderness character, ecological integrity and/or scenic quality of
certain areas. These values are only imperfectly captured by the
prices of goods traded in markets. Accordingly, non-market methods
must be used to calculate intrinsic values. This study is designed
to use the contingent valuation method to determine the loss in
intrinsic values resulting from the oil spill.

OBJECTIVES

Determine the loss of intrinsic value of natural resources
attributable to the EVOS.

METHODS

The contingent valuation method involves use of surveys to
determine the values that people place on goods that are not traded
in markets. This study will require development of a conceptual
framework for contingent valuation survey design and analysis of
survey results. Next, research will be conducted to determine the
most accurate survey instrument for assessing intrinsic values.
This research will involve consultation with economists and survey
design experts. Substantial preliminary testing of survey formats
will be conducted among small groups of people to verify the
accuracy of the survey instrument. A nationwide survey will be
conducted using a professional survey research firm. Econometric
analysis will be used to interpret the results of the survey.

BUDGET

Salaries $ 515.0
Travel 145.0
Contracts 680.0
Supplies 295.0
Equipment 375.0

Total $2,010.0

329

ECONOMICS STUDY NUMBER 8

Study Title: Economic Damage Assessment of Research Programs
Affected by the EVOS

INTRODUCTION

The oil spill affected research programs in the vicinity of the
spill, resulting in damage to or loss of various research and
resource monitoring studies. Opportunities to study natural
resource systems in the affected area may have been lost or
diminished as a result of the spill. Research studies underway
before the Spill and conducted, permitted, cooperatively
participated in, sponsored or funded by the federal government
likely were impacted. One example is a study involving tagging of
fish that was in progress in an affected area of PWS.
Determination of the set of studies affected and the extent or
degree of damage will require careful evaluation and study.

OBJECTIVES

Assess damage to and economic loss of research investigations, and
account for the cost of resources expended in affected studies,
focusing on research-based expenditures made or committed to before
the oil spill.

METHODS

The first step in this study is to identify the universe of studies
that were underway in the affected area at the time of the spill.
The next stage requires a determination of which studies were
negatively impacted by the spill. Some of those impacts may have
been so significant that the entire study was discontinued. Other
studies may have been able to continue, but only at an increased
cost caused by the impacts of the spill. For example, sample sets
may have been destroyed or the study may have been moved to another
area. once the universe of affected research programs is
identified, this study will value the destroyed and damaged
research studies by looking first to total project costs, extra
sums expended and amounts spent on each study prior to being
impacted by the spill.

BUDGET

Salaries $ 23.5
Travel 10.0
Contracts 10.0
Supplies 7.5
Equipment 0.0

Total $ 51.0

330

ECONOMICS STUDY NUMBER 9

Study Title: Quantification of Damage to Archeological Resources

INTRODUCTION

Archeological sites along the many miles of oiled coastline and
intertidal zones may have been physically damaged by oil. Upland
sites may have been damaged by erosion caused by destruction of
site vegetation or transportation of the oil inland. Loss to
archeological resources includes direct and indirect oiling.
Determination of the number of cultural resources impacted by the
oil spill as well as the type and extent of injury to the
archeological sites has been moved to a separate science study.
The economics study is now limited to quantifying the loss to
archeological resources.

OBJECTIVES

Assess the economic damages to archeological sites.

METHODS

The archeological science study will create a data base containing
listings of the oil impacted areas and a model for the kinds of
cultural resources impacted, the degree of the impact and the
physical setting of the damaged resource. Both use and intrinsic
values of archeological resources may have been impacted.

Use Value

I . Effects on the scientific value of the archeological
resource. The magnitude of this damage depends on the
uniqueness of the affected site, the original quality of
information available at the site, the nature of the impacts,
and the willingness of the scientific community to pay for the
lost information. If the site is unique and substitute
sources of similar information do not exist, the value of the
damage may be large.

2. Loss of value as tourist and educational attractions.
unique or spectacular archeological sites have value as
tourist attractions. All significant archeological sites have
educational value as the focus of field trips and published
descriptions. Archeological information and artifacts have
value for museum interpretation and display. Oil impacts
could substantially reduce these values.

Intrinsic Value

1. Impacts on the religious, cultural or symbolic values for
native groups.

331

2. Loss of intrinsic value for the general, non-native
population.

BUDGET

Salaries $ 25.0
Travel 15.0
Contracts 0.0
Supplies 10.0
Equipment 0.0

Total $ 50.0

332

PART III

RESTORATION PLANNING

RESTORATION PLANNING PROJECT

INTRODUCTION

The Trustees recognized from the beginning that restoration of the
ecological health of areas affected by the oil spill is the
fundamental purpose f or conducting the NRDA. Initially, studies to
determine the injury to natural resources were emphasized, since
that information is basic to a determination of damages, and
finally, restoration of resources.

Since late 1989, considerable effort has gone into specific
restoration planning activities. An interagency Restoration
Planning Work Group (RPWG) was f ormed to develop and coordinate
what is envisioned to be a steadily growing level of activity
throughout this year and next. A variety of activities have
already been initiated by the RPWG and several more are proposed
to occur during 1990, as described in the following pages. In
addition, it is anticipated that restoration planning and project
activities will be expanded further in 1991 and beyond.

OBJECTIVES

The overall goal of the Restoration Planning Project is to identify
appropriate measures that can be taken to restore the ecological
health of natural resources affected by the EVOS. Among the
objectives within this overall goal are:

A. Encourage and provide for public participation and review
during the restoration planning process.

B. Identify and develop technically feasible restoration options
for natural resources and services potentially affected by the
oil spill.

C. Incorporate an "ecosystem approach" to restoration (i.e.,
broadly focus on recovery of the ecosystems, as well as
individual components).

D. Identify when active restoration measures may be warranted,
and when it may be appropriate to rely on natural recovery.

E. Identify the costs associated with implementing feasible
restoration measures, in support of the overall NRDA process.

333

DEFINITION

Restoration is a broad term that can include direct restoration,
replacement, or acquisition of resources or uses those resources
provided that are equivalent in terms of ecological or human
services.

Direct restoration refers to measures taken to restore an injured
resource, and generally equates with on-site actions. An example
would be to rehabilitate an oiled marsh ecosystem by supplementing
natural plant and animal populations after removal of the oil.

Replacement refers to the substitution of one resource for an
injured resource of the same type. An example is to use
hatchery/aquaculture techniques to establish an entirely new
fishery stock in place of one that has been severely damaged.
Replacement activities may or may not be limited to the specific
site or area where injury occurred.

Acquisition of equivalent resources means to obtain or otherwise
protect resources that are similar or related to the injured
resources in terms of ecological value, functions, or uses. An
example is to obtain or protect undamaged wildlife habitats as
alternatives to direct restoration of injured habitats. Equivalent
resources could be acquired in locations removed from the immediate
vicinity of the injured resource.

1990 RESTORATION PLANNING ACTIVITIES

Several major activities have been initiated or -are proposed under
the Restoration Planning Project in 1990. Each major activity area
is described in this section.

Public Participation

In part as a response to public comments on the 1989 NRDA Plan,
several avenues have been developed for public involvement in the
restoration planning process. TheiRPWG has conducted a public
Restoration Symposium, and held public information and scoping
meetings in several Alaskan communities directly affected by the
oil spill. Additional public meetings may be held outside Alaska
during 1990 as well. The RPWG has also begun to contact interest
groups and other organizations that have expressed an interest in
the restoration planning process, in order to gain a more direct
and detailed understanding of their concerns and suggestions. An
information flier and response form has been developed and
distributed initially in Alaskan communities in order to encourage
additional comments from residents of areas most directly affected
by the spill.' Reports generated through the Restoration Planning
Project will generally be distributed publicly. The following
paragraphs briefly describe the outcomes of the major public
activities conducted to date.

334

Restoration Symposium

A two-day public Restoration Symposium was held at the Egan Civic
and Convention Center in Anchorage, Alaska on March 25 and 26,
1990. The symposium was the first opportunity for environmental,
industry, and other interest groups and members of the general
public to present their views about the content of a restoration
plan. Formal presentations were made by more than 30 individuals.
A report documenting the presentations and comments given at the
Restoration Symposium is scheduled to be publicly distributed in
July 1990.

Community Scoping Meetings

An initial series of public information and scoping meetings was
held beginning in April 1990. The RPWG travelled to eight Alaskan
communities directly affected by the oil spill to provide an
opportunity for residents to express their views about what a
restoration plan should entail. Evening meetings were held in
Cordova, Valdez, Whittier, Seward, Kenai, Homer, Kodiak and
Anchorage. The community scoping meetings resulted in a variety
of restoration ideas being identified. Public comments received as
a result of the community scoping meetings will be documented in
the progress report scheduled for public distribution in July 1990.

Technical Workshops on Restoration

The RPWG conducted an initial three-day technical workshop on
restoration in Anchorage in early April, 1990. The workshop
provided a forum for the scientists most familiar with the effects
of the oil spill, as well as other scientists with relevant
knowledge, to focus their attention on potential restoration needs
and opportunities. A second technical workshop is planned for the
Fall 1990, and it is anticipated that more such opportunities will
occur before the conclusion of the process. One purpose of the
first technical workshop was to help identify and develop an
initial set of potentially beneficial restoration techniques that
could receive small-scale field testing during the Summer 1990. An
array of potential feasibility study projects was identified, some
of which are proposed to be initiated (see Restoration Feasibility
Studies below). The results of the workshop will be documented in
the progress report scheduled for public distribution in July 1990.

Literature Review

The first phase of a comprehensive search of worldwide literature
relevant to restoration of natural resources was initiated early in
1990. "Phase 1,11 the initial search of key computerized literature
data bases, identified several thousand potentially relevant
references, which were narrowed to approximately one thousand of
the most directly applicable citations. These references have been
screened, and the most important ones have been flagged for

335

acquisition. These references will be reviewed in detail during
the "Phase III' literature review, along with other references
identified in an expanded search. Literature review activities are
expected to continue throughout the restoration planning process.
The results of "Phase I" will be summarized in the progress report
scheduled for public distribution in July 1990. Updated results
will be presented in subsequent reports.

Feasibility Study Projects

There are relatively few existing technologies for restoration of
natural resources that can be immediately applied under Alaskan
conditions with certainty of success. For this reason, feasibility
study projects are among the most important aspects of restoration
planning. A feasibility study project may be appropriate when a
restoration idea has been developed that appears to be potentially
beneficial, but for which there is substantial uncertainty of its
success or benefit with local species or under the sub-arctic
conditions of the spill area.

The following pages present summaries for each of the initial
feasibility study projects proposed for 1990. These projects were
developed from ideas presented at the public 'symposium, the
community scoping meetings, and the technical workshop. Factors
considered in selecting 1990 studies included: the need to initiate
the particular study as soon as possible, the ability to implement
the project in a short time frame, reasonable likelihood of
success, identified public concern, relationship to other NRDA
studies, and budget priorities.

Five restoration Feasibility Studies having a total budget of
$326,400 are proposed for initial field testing in 1990. Two of
these concern restoration of intertidal resources and communities!
one addresses upland habitats used by wildlife affected by the
spill, one involves stabilization and restoration in the supratidal
zone, and one supports the potential acquisition of equivalent
resources through review of land status, uses, and plans.

Three restoration technical support projects with a budget of
$236,500 are planned. The first will institute a formal peer
review process for restoration project results and planning. The
second will compile shoreline status information from both response
and NRDA sources to support selection of sites and habitats for
future feasibility studies and restoration projects. The third
technical support project will fund development of detailed
proposals for feasibility studies to be considered for
implementation in 1991.

336

BUDGET

Restoration symposium $ 50.0
Community scoping meetings 40.0
Technical workshops 200.0
Literature collection/review 90.0
Feasibility study projects 562.9
Report preparation/publication 150.0
Salaries 600.0
Travel 70.0

TOTAL $1,762.9

Lead agencies: EPA, ADF&G

Cooperating Agencies: DNR, DEC, DOA, DOI, DOC

337

RESTORATION TECHNICAL SUPPORT PROJECT NUMBER 1

Project Title: Peer Reviewer Process for Restoration Feasibility
Studies

Lead Agency: RPWG

Cooperating Agencies: DOJ, DOL

INTRODUCTION

The initial feasibility study projects to be conducted during the
1990 field season were developed with the assistance of many of the
scientists involved in the NRDA process, after considering comments
received at the technical workshop and a series of public meetings
held in Spring 1990 in Alaska. Due to the limited time available
before projects need to be in the field, an additional more formal
round of peer review is not possible. This technical support
project is designed to incorporate formal peer review in the
design, implementation, and evaluation of 1991 and future
feasibility studies. It will also provide for detailed review of
1990 feasibility study results.

OBJECTIVE

Implement a peer reviewer process to assure the scientific quality
of feasibility studies and restoration projects.

METHODS

Peer reviewers may include experts -already involved in the NRDA
process, experts involved in the technical workshops on
restoration, or other selected individuals. Peer reviewers would
review and comment on feasibility study proposals (including
overall design and detailed study plans) and results. The budget
for 1990 is based on the services of 10 expert reviewers for five
days each, plus expenses. It is anticipated that this technical
support project will expand in 1991, as additional feasibility
studies are initiated and as results from 1990 feasibility study
projects become available.

338

BUDGET: DOJ, DOL

Salaries: $ 0.0
Travel: 0.0
Contractual Services: 70.0
supplies: 5.0
Equipment: 0.0

TOTAL: $75.0

339

RESTORATION TECHNICAL SUPPORT PROJECT NUMBER 2

Project Title: Assessment of Beach Segment Survey Data

Lead Agency: DNR

Cooperating Agencies: DEC, ADF&G, USFS, NPS, EPA

INTRODUCTION

There is a large volume of beach-survey information obtained
through response activities (e.g., the fall and spring surveys) and
NRDA studies (e.g., CH 1) . All of these data are being integrated
into a standard NRDA data base. This information is being reviewed
and summarized with respect to restoration planning needs and will
complement and support Restoration Feasibility Study Number 5 (RF
5). Together, this information will help identify potential sites
at which (a) hands-on restoration projects may be carried out, and
(b) equivalent resources may be acquired. Additionally, it should
prove valuable in providing further information for analytical
purposes in the development of the restoration planning matrix.

OBJECTIVES

A. obtain and translate to maps, pertinent beach survey
information that is important for feasibility studies and
restoration projects.

B. Analyze possible trends in information for applicability to
restoration feasibility studies.

C. Create a data base for future reference use in restoration
projects.

Relationships with Other Studies:

This project relates directly to RF 5 and provides data of
fundamental importance to the entire Restoration Planning Project.

METHODS

Research and map, using standard cartographic and G.I.S.
techniques, all available information from the Fall 1989, Spring
1990, and Fall 1990 walk-a-thon and shoreline assessment team
surveys. Combined with RF 5, this will provide further support in
the selection process for specific restoration sites and habitats.
It may also prove advantageous for documenting natural recovery
processes that may be occurring. Care will be taken to not
duplicate existing data bases and maps. The need is to integrate
new information and summarize it in a form helpful to the
Restoration Planning Project. This project will essentially add a

340

"restoration layer" to the existing NRDA data base.
BUDGET: DNR

Salaries $ 16.0
Travel 0.0
Contractual Services 5.0
supplies 4.0
Equipment - 0.0
TOTAL 25.0

341

RESTORATION TECHNICAL SUPPORT PROJECT NUMBER 3

Project Title: Development of Potential Feasibility Studies for
1991

Lead Agency: ADF&G and EPA

Cooperating Agencies: DNR, DEC, DOA, DOI, DOC

INTRODUCTION

A variety of potential restoration feasibility studies need to be
undertaken before recommendations can be made in the Restoration
Plan. Due to funding and timing constraints in 1990, it was
possible to carry out only a limited number of such studies in
the current season. There is much that can and needs to be done,
however, to develop the substance of feasibility study proposals
for possible implementation in 1991. A number of specific areas
have been identified for development of study plans. These
include (A) Monitoring "Natural" Recoveries, (B) Pink Salmon
Stock Identification, (C) Herring Stock Identification/Spawning
Site Inventory, (D) Artificial Reefs for Fish and Shellfish, (E)
Alternative Recreation Sites and Facilities, (F) Historic Sites
and Artifacts, and (G) Availability of Forage Fish. In addition,
as new information becomes available through the NRDA process,
public comments, and technical consultations, the RPWG expects to
identify additional restoration ideas and areas of concern for
which feasibility studies may be appropriate.

Objectives:

A. To identify restoration ideas and areas of concern for which
feasibility studies may be necessary and appropriate.

B. To develop feasibility study plans and proposals which may
be considered for implementation in 1991 and beyond.

Relationships with Other Studies:

This project relates directly to Restoration Technical Services
Project Number 1, implementation of a peer reviewer process, as
well as the entire NRDA and Restoration Planning Project.

Methods:

Based on public comments, NRDA results, and consultations with
technical experts, the RPWG anticipates that candidate restora-
tion projects will be identified on an on-going basis. In order
to fully evaluate some of these suggestions, it will be necessary
to carry out feasibility studies. The RPWG then needs to convene
ad hoc committees consisting of combinations of agency personnel,
peer reviewers, and outside experts to more fully develop the

342

study plans and proposals. Support is needed to convene meet-
ings, particularly involvin4 travel by outside experts. In some
cases, site visits will be eeded to examine particular problem
areas related to the oil spill or successful restoration projects
which have been implementedlelsewhere.

BUDGET:

salaries $ 5.0
Travel 77.5
Contractual Services 40.0
Supplies 11.0
Equipment 3.0

TOTAL 136.5

343

RESTORATION FEASIBILITY STUDY NUMBER I

Study Title: Re-establishment of Fucus in Rocky Intertidal
Ecosystems

Lead Agency: EPA

Cooperating Agency: USFS

INTRODUCTION

Qualitative evidence indicates that rockweed, the marine alga,
Fucus, was damaged by both the spilled oil and the cleanup
effort. Fucus is a critical structural component of the inter-
tidal habitat in the oil-spill area, and it serves as an impor-
tant spawning substrate for herring. Re-establishment of this
species will increase the rate of recovery of other associated
biotic communities.

There may be a substantial delay in natural recovery of areas
where populations were reduced over large areas (100-1000 m of
shoreline), because dispersal of seeds is limited (< I m in most
circumstances). Drift plants may increase this distance, but the
importance of this mode is unknown.

The reproductive and life history of Fucus is well known, and
techniques for collection of seed are well established. In
southern parts of the range plants are fertile year round, so the
timing of the application of seeds may be relatively unimportant
in the establishment of the plant. The specific life history
cycle of the plant in PWS and the GOA is not known. It is
expected, however, that the plants will be fertile for at least
most of the spring and summer.

Objectives:

A. Document the extent and magnitude of recruitment of Fucus in
areas subjected to alternative cleaning technologies.

B. Determine the feasibility of re-establishing Fucus in dam-
aged areas.

C. Develop and demonstrate potential large scale seeding tech-
niques to re-establish fucus.

D. Demonstrate the efficacy of seeding versus transplanting
Fucus.

E. Identify the costs of implementing a full-scale Fucus resto-
ration project.

344

Relationships with Other Studies:

This study is fundamental to bringing an ecosystem approach to
the restoration program. It relates directly to RF 2, re-estab-
lishing critical intertidal fauna, and to various NRDA studies,
particularly Coastal Habitat Study Number 1.

Methods

The study plan has two parts: (1) laboratory experiments that
develop techniques for obtaining large quantities of embryos
suitable for use in reseeding, and (2) field experiments to test
the effectiveness of embryo reseeding and transplanting in
habitats that experienced varying degrees of oiling and cleaning.

Laboratory experiments will be conducted to determine embryo
attachment strength over time. Since the seeds must remain in
suspension, experiments will also be conducted to assure their
viability in culture media for at least two weeks. Although
techniques for obtaining Fucus embryos are simple and well known,
these techniques will be modified and tested for the production
and handling of the large numbers of embryos that would be
necessary for a full-scale reseeding project.

Field tests will then be conducted with various "seeding" proce-
dures (e.g., dispersal of embryos, dispersal of embryos, and
transplants of fertile adults). All three methods will be tested
in one control and one habitat that was disturbed by oil and
subsequently cleaned. Dispersal of embryos will then be tested
in habitats with different combinations of oil and cleanup
techniques (e.g., bioremediated, hot water wash). The experimen-
tal design will use three replicates of each habitat type, three
replicates of each procedure, and three replicates of controls to
measure natural settlement. Variables to be measured include
height of Fucus plants, numbers of plants, and percent
vegetative cover. Maps prepared by the Damage Assessment
Geoprocessing Group will be used to identify potential study
sites. In the initial project, primary study sites will be in or
near Herring Bay, PWS.

BUDGET: EPA

Salaries $ 2.0
Travel 11.0
Contractual Services 135.0
Supplies 2.0
Equipment 0.0

TOTAL 150.0

345

RESTORATION FEASIBILITY STUDY NUMBER 2

Study Title: Re-establishment of Critical Fauna in Rocky
Intertidal Ecosystems

Lead Agency: USFS

Cooperating Agency: EPA

INTRODUCTION

Intertidal ecosystems on rocky shores, including both f auna and
flora, were seriously affected by the oil spill and cleanup
activities. Initial results suggest that certain key faunal
species, such as grazers and predators, that are likely to
structure these intertidal communities, were moderately to heavily
affected. Natural restoration processes in these communities will
be limited by recolonization rates of these key species, which in
some cases are known to be quite low. Re-establishment of Fucus
alone may therefore not be sufficient to ensure a return to pre-
spill conditions on ecologically meaningful time scales. Before a
restoration plan is proposed, we should demonstrate the feasibility
of enhancing the rate of recovery of the intertidal community by
the re-establishment of key grazers and predators. If the natural
recoveries of Fucus and intertidal fauna can be augmented by
restoration projects, it will be of fundamental benefit to the
marine ecosystem.

OBJECTIVES

A. Compare rates of recovery of rocky intertidal communities with
and without key faunal species and combinations of species.

B. Demonstrate the feasibility of restoring rocky intertidal
communities by enhancing colonization by key faunal species.

C. Determine the costs of implementing a full-scale restoration
project to re-establish key faunal species in rocky intertidal
ecosystems.

Relationships with Other Studies:

This study will be carried out in conjunction with the Fucus study,
R/F 1, and it is related to several other NRDA studies,
particularly CH 1.

METHODS

Based on results of NRDA studies, limpets have been identified as
important grazers that were harmed by the oil spill in rocky
intertidal ecosystems. Predators, such as Nucella and
Leptasterius, also could be important in structuring these

346

intertidal communities. Rates of recovery of intertidal areas with
and without key species and combinations of species will be
compared. Grazer, predator, and grazer-predator exclusion and
enhancement plots will be established in habitats that experienced
differing degrees of oiling or were subjected to different cleanup
techniques (e.g., bioremediated, hot-water high-pressure cleaned) .
A key aspect of the study will be demonstrating the feasibility of
enhancing colonization by key species.

BUDGET: USFS

Salaries $ 0.0
Travel 5.0
Contractual Services 65.0
Supplies 2.0
Equipment 3.0

TOTAL 75.0

347

RESTORATION FEASIBILITY STUDY NUMBER 3

Study Title: Identification of Potential Sites for
Stabilization and Restoration with Beach Wildrye

Lead Agency: DNR

Cooperating Agencies: USFS

INTRODUCTION

The EVOS and associated cleanup efforts have affected supratidal
beach ecosystems, of which a key component is the native grassl
beach wildrye (Elymus mollis). The supratidal beach wildrye
plant community is extremely important in the prevention of
erosion in the coastal environment. Erosion can lead to the
destabilization and degradation of cultural and recreational
sites as well as of wildlife habitats (e.g., for ground-nesting
birds). There are well established techniques for restoring rye
grasses and other plants on coastal dune systems, including at
some sites in Alaska. It is necessary, however, to first
identify sites at which damage has occurred and restoration
efforts appear to be feasible, and it is also necessary to
establish the cost of a full-scale restoration project in the
EVOS area.

OBJECTIVES

A. Determine the distribution and areal extent of supratidal
sites at which beach wildrye restoration efforts will be
needed and feasible.

B. Identify potential sites for pilot projects to re-establish
supratidal stands of beach wildrye.

C. Determine the costs of implementing a full-scale project to
restore supratidal stands of beach wildrye.

Relationships with Other Studies:

This feasibility study addresses a key component in supratidal
beach ecosystems. It relates directly to other feasibility
studies and potential restoration projects in the areas of
cultural, recreational, and avian resources.

METHODS

Beach segment survey data, aerial photographs, on-site
inspections, and other sources of coastline status data will be
used for a preliminary identification of sites where stands of
beach wildrye have been injured and erosion is occurring or may
occur as a result. Based on these preliminary results,

348

individual sites will be visited and evaluated for their
potential as sites at which beach wildrye restoration techniques
may be developed and tested. The on-ground activities will
include documenting the size, type, and extent of damage and the
depth of oil, if present, in the substrate. This study will
enable development and evaluation of a proposal for a full-scale
feasibility study of restoration methods in subsequent years.

BUDGET: DNR

Salaries $ 14.4
Travel 5.6
Contractual Services 5.0
Supplies 3.1
Equipment 0.0

TOTAL $ 28.1

349

RESTORATION FEASIBILITY STUDY NUMBER 4

Study Title: Identification of Upland Habitats Used by Wildlife
Affected by the EVOS

Lead Agency: FWS

Cooperating Agency: ADF&G

INTRODUCTION

A variety of marine birds, waterfowl,'and other bird and mam-
malian species were killed by the spill or injured by contami-
nation of their prey and habitats. Many of these wildlife
species are dependent on aquatic or intertidal habitats for such
activities as feeding and resting, but they use upland habitats
in forests, along streams, or above tree line to fulfill other
life-history requirements (e.g., nesting, shelter). Through the
public scoping process and technical workshop, many people have
suggested that protection of upland wildlife habitats from
further degradation may be an important way to help wildlife
recover from the effects of EVOS. To explore this potential, it
is necessary to learn more about the specific upland habitats
upon which these species depend and how they use them. While
such a feasibility study would be a large and complex under-
taking, an initial study that primarily focuses on the marbled
murrelet (Brachyrumphus marmoratus) and the harlequin duck
(Histrionicus histrionicus) will be conducted in 1990. The
results of this study will provide a basis for developing and
evaluating a broader feasibility study proposal that will more
fully explore the ecological relationship between marine-depen-
dent wildlife and upland habitats.

OBJECTIVES

Objectives A-C specifically apply to both harlequin ducks and to
marbled murrelets, the primary subjects of the 1990 study:

A. Develop and test methods for establishing the presence of
breeding birds.

B. Develop and test methods for locating nest sites.

C. Identify and characterize nest habitats and sites.

D. Define the parameters of and develop a proposal for a full-
scale upland habitat feasibility study for marine birds,
waterfowl, and other species.

E. Determine the costs of implementing a full-scale restoration
project concerning upland habitats used by marine-dependent
wildlife.

350

Relationships with Other Studies:

This study relates directly to the results and field work of Bird
Studies 2 and 11 and RF 5.

METHODS

Marbled murrelet: Naked Island in PWS will be the primary study
site. The presence of breeding murrelets will be recorded by a
stationary observer at dawn, at which times murrelets fly to
inland nest sites. Murrelet altitude, behavioral, and other data
will be recorded for each bird observed. Sites with high mur-
relet activity will be identified and then searched for nests.
The efficacy of the dawn detection technique will be evaluated.

Harlequin duck: Streams in PWS will be selected for investigation
based upon reported concentrations of ducks, survey data from
NRDA projects, and interviews with knowledgeable field personnel.
Once streams are identified as having a high potential for
harlequin nests, there will be intensive ground searches for
nests. As nests are located, the nest sites and habitats will be
characterized by such parameters as distance from the stream and
coast, topography, and vegetative cover.

BUDGET: FWS, ADF&G

Salaries $13.3
Travel 1.0
Contractual Services 3.0
Supplies 2.5
Equipment 3.5

Total 23.3

351

RESTORATION FEASIBILITY STUDY NUMBER 5

Study Title: Land Status, Uses, and Management Plans in Relation
to Natural Resources and Services

Lead Agency: DNR

Cooperating Agencies: USFS, NPS, ADF&G

INTRODUCTION

Through the restoration scoping process members of the public have
suggested a wide variety of projects to acquire equivalent
resources. Examples are the acquisition of timber or development
rights, conservation easements, recreational and cultural sites,
inholdings within state and federal protected areas, and buffer
strips along streams and coasts. In addition, scientists
participating in the technical workshop found that in some cases
habitat protection projects would be the best means of providing
for the long-term restoration of injured wildlife resources. In
order to begin to identify and evaluate potential restoration
projects of this type, it is necessary to summarize existing
information about the land status., uses, and management plans for
both privately and publicly owned lands. This initial effort will
focus on the oil-spill area and adjacent lands and will also serve
to identify potential sites for other types of restoration
projects.

OBJECTIVES

A. Summarize and map the land status and ownership, land-use
designations, and existing and proposed uses of tidelands and
related uplands.

B. Summarize and map the extent and degree of oiling and coastal
morphology as necessary for restoration planning purposes.

C. Summarize and map natural resources and services, including
vegetation, fish and wildlife populations, habitats, and
sensitive areas, recreation, and commercial forestry.

Relationships with Other Studies:

These data are fundamental to the entire Restoration Planning
Project and especially to those feasibility studies and potential
restoration projects that concern the acquisition of equivalent
resources.

METHODS

The DNR, through the NRDA Study TS 1, has compiled much of the
necessary data on their computerized G.I.S. Additional resource

352

and land use information is available in state and federal
management plans and resource inventories and from the Alaska
Coastal Management Program. The RPWG and technical advisors will
be consulted to define the specific area and information needs,
which will then be obtained from the various existing data bases.
After determining the most feasible means and best resolution to
portray the information, it will be summarized, produced, and
distributed, primarily in map form.

BUDGET: DNR

Salaries $ 34.0
Travel 1.0
Contracts 5.0
supplies 10.0
Equipment 0.0

Total $ 50.0

353

PART IV

BUDGET

Budget Summary for the Exxon Valdez Oil Spill Damage Assessment - 1990
Budgets are in 1000's of Dollars

Budgets are costs for projects from 3-1-90 through 2-28-91

Study
Category Number Title Agency Budget

Coastal CH1 Comprehensive Assessment ADF&G 156.7
Habitat USFS 9,113.0

Air/Water A/W2 Injury to Subtidal ADF&G 333.5
NOAA 466.8

A/W3 Hydrocarbons in Water DEC 47.5
NOAA 472.5

A/W6 Oil Fate and Toxicity NOAA 870.0

Fisheries FIS1 Salmon Spawning Area Injury ADF&G 391.5

FIS2 Egg and Preemergent ADF&G 302.8
Fry Sampling
F/S3. Coded-Wire Tagging ADF&G 1,990.0

F/S4 Early Marine Salmon Injury ADF&G 150.0
NOAA 400.0

F/S5 Dolly Varden Injury ADF&G 290.0

F/S7a Salmon Spawning Area Injury, LCI ADF&G 117.6

F/S7b Salmon Spawning Area Injury, ADF&G 460.3
Kodiak & Chignik

354

Budget Summary for the Exxon Valdez Oil Spill Damage Assessment Program
Budgets are in 1000's of Dollars

Budgets are costs for projects from 3-1-90 through 2-28-91

Study
Category Number Title Agency Budget

Fisheries F/S8a Egg & Preemergent Fry
Sampling, LCI ADF&G 71.0

F/S8b Egg & Preemergent Fry
Sampling, Kodiak & Chignik ADF&G 149.3

F/S11 Herring Injury ADF&G 558.4

F/S13 Clam Injury ADF&G 229.2

F/S15 Spot Shrimp Injury ADF&G 65.0

F/S17 Rockfish Injury ADF&G 109.4

FIS18 Trawl Assessment in PWS NOAA 186.1

F/S22 Crab Injury, Outside PWS NOAA 110.0

F/S24 Trawl Assessment,
Outside PWS NOAA 450.0

F27 Sockeye Salmon
Overescapement ADF&G 392.0

F28 Run Reconstruction ADF&G 175.1

355

Budget Summary for the 'Exxon Valdez Oil Spill Damage Assessment Program
Budgets are in 1000's of Dollars

Budgets are costs for projects from 3-1-90 through 2-28-91

Study
Category Number Title Agency Budget

F30 Data Base Management ADF&G 120.0

Marine mmi Humpback Whale NOAA 92.0
Mammals
MM2 Killer Whale NOAA 255.8

MM4 Sea Lion NOAA 171.2

MM5 Harbor Seal NOAA 159.3

MM6a sea otter Injury FWS 1,060.5

MM6b Sea Otter Mortality
comparisons FWS 11.0

MM6c Sea Otter Drift Study FWS 33.5

MM7 Sea Otter Rehabilitation FWS 147.0

Terrestrial TM1 Injury to Sitka Black-
Mammals Tail Dear ADF&G 124.6

TM2 Injury to Black Bear ADF&G 10.0

TM3 Injury to River Otter
ADF&G 347.6

356

Budget Summary for the Exxon Valdez,Oil Spill Damage Assessment Program
Budgets are in 1000's of Dollars

Budgets are costs for projects from 3-.1-90 through 2-28-91

Study
Category Number "Title Agency Budget

TM4 Injury to Brown Bear ADF&G 125.7

TM6 Reproduction of Mink ADF&G 134.0

Birds B1 Beach Bird Survey FWS 598.0

B2 censuses & Seasonal
Distribution FWS 471.0

B3 Seabird Colony Surveys FWS 251.1

B4 Bald Eagles FWS 675.0

B5 Peale's Peregrine Falcons FWS 107.7

B11 Sea,Ducks. FWS 150.0

B13 Passerines FWS 10.0

Technical TS1. Hydrocarbon Analysis NOAA 914.2
Services FWS 1,089.2

TS2 Histopathology ADF&G 100.0

357

Budget Summary for the Exxon Valdez Oil Spill Damage Assessment Program
Budgets are in 1000's of Dollars

Budgets are costs for projects from 3-31-90 through 2-28-91

Study
Category Number Title Agency Budget

TS3 GIS DNR 592.2
FWS 200.0

ARCH1 Archeology USFS 932.0
DNR 300.0

Restoration RP1 Restoration Planning' ALL 1,762.9
Planning

overhead State of Alaska ADF&G 1,745.0

Dept. of Agriculture USFS 1,245.0

Dept. of Interior FWS 500.0

Dept. of Commerce NOAA 953.8.

Environmental Protection
Agency EPA 44.1

Discontinued Studies
Completion All Agencies 140.0

3.58

Budget Summary for the Exxon Valdez Oil Spill Damage Assessment Program
Budgets are in 1000's of Dollars

Budgets are costs for projects from 3-1-90 through 2-28-91

Study
Category Management Entity Agency Budget

Economics ECON1 commercial Fisheries Losses ALL FED 229.0

ECON4 Public Land Value Effects ALL FED 180.0

ECON5 Recreational U.ses Damage ALL FED 294.0

ECON6 Subsistence Losses ALL FED 885.0

ECON7 Intrinsic Value Loss ALL FED 2,010.0

ECONS Research Program Damage ALL FED 51.0

ECON9 Archeological Resource Damage ALL FED 50.0

TOTALS $37,330.2

359

Trustee Budget Summary for the Exxon Valdez Oil Spill Damage Assessment Program
Budgets are in 1000's of Dollars

Budgets are costs for projects from 3-1-90 through 2-28-91

Trustee Budget

State of Alaska $ 10,504.9
Department of Agriculture 11,545.4
Department of the Interior 5,559.4
Department of Commerce 5,757.0
Environmental Protection Agency 264.5
All Federal 3,699.0

TOTALS $ 37,330.2

360

APPENDICES

APPENDIX A

STATE/FEDERAL DAMAGE ASSESSMENT PLAN

ANALYTICAL CHEMISTRY

QUALITY ASSURANCE (QA)/QUALITY CONTROL (QC)

TABLE OF CONTENTS

1. QUALITY ASSURANCE FOR ANALYTICAL CHEMISTRY

1.1 Study-Specific AQ Plans (QAP)
1.2 Technical System Audits
1.3 Standards and Quality Control Materials
1.4 Analytical Performance Evaluations
1.5 Data Reporting and Deliverables

2. MINIMUM REQUIREMENTS: SAMPLING AND SAMPLING EQUIPMENT

2.1 Sampling Identification and Labeling
2.2 Field Chain of Custody

3. MINIMUM REQUIREMENTS: ANALYSIS

4. MINIMUM REQUIREMENTS: REPORTING AND DATA DELIVERABLES

Appendix A (continued)

This document describes the Quality Assurance for the analytical
chemistry portions of the Exxon Valdez Damage Assessment Process.
it is to be used in conjunction with the Analytical Chemistry
Quality Assurance Programs of the Trustee Agencies. It describes
only those minimum requirements necessary to validate the data
generated by analytical chemistry laboratories. Quality assurance
requirements for other types of measurements are not addressed.
For instructions in meeting the requirements described in @his
document, please consult "Collection and Handling of Samples,"
which was prepared by the Analytical Chemistry Group for use in
training field personnel or the following Agency representatives:

Carol-Ann Manen, National Oceanic and Atmospheric Administration,
(907) 789-6014.

Everett Robinson-Wilson, U.S. Fish and Wildlife Service,
(907) 786-3493.

Rolly Grabbe, Alaska Department of Environmental Conservation,
(907) 364-2155.

John Moore, U.S. Fish and Wildlife Service, (301) 497-0524.

Appendix A (continued)

1. Quality Assurance for Analytical Chemistry

Each Trustee agency through their individual standard documented
QA programs and guidances shall ensure that all data generated by
or for that agency and their contractors, in support of the Exxon
Valdez Damage Assessment, are of known, defensible, and verifiable
quality.

These documented QA programs and guidances include but are not
limited to:

NOAA National Status and Trends Program, Mussel Watch Phase
4 Work/QA Project Plan
Quality Assurance of Chemical Analyses Performed Under
Contract With the USFWS
EPA SW-846, Chpt. 1, QA/QC Requirements
EPA Guidelines and Specification for Preparing Quality
Assurance Project Plans, QAMS-005
EPA Handbook for Sampling and Sample Preservation of Water and
Wastewater

In addition, an interagency team of leading scientists from the
Trustee agencies and the Environmental Protection Agency, hereafter
referred to as the Analytical Chemistry Group (ACG), shall develop
and oversee a centralized program which will demonstrate the
quality and comparability of the chemical data obtained by the
Trustee agencies.

The major components of this centralized QA program will be:

1. Development of study-specific analytical chemistry QA plans.

2. Technical on-site system audits of field and laboratory data
collection activities.

3. Development and provision of appropriate instrument
calibration standards and control materials.

4. Laboratory performance evaluations by means of intercomparison
exercises.

5. Review of data deliverables and all supportive documentation
to evaluate data quality.

Appendix A (continued)

1.1 Study-Specific Quality Assurance Plans

Prior to the initiation of each study, the study manager must
prepare and submit a study-specific analytical chemistry QAP to the
ACG for review and concurrence. This plan shall specify each
study's goals, sampling procedures, analytical procedures, and all
quality control measures and acceptance criteria associated with
those procedures.

The QAP must be study-specific, however any documented QA guidance
and/or appropriate Standard operating Procedures (SOP's) used by
the Trustee agencies may form the basis of individual study QA
plans.

A Quality Assurance Plan must address the following:

Title Page - Includes the signatures of the individuals
responsible for the project and ACG concurrence.

Proiect Description and Sampling Objectives - Briefly
describes the what, where, and why of the project.

Data Needs - Describes what elements, compounds, classes
of compounds, and/or physical data are required. Must
describe the desired detection limitst precision and
accuracy of the data for the study.

Sampling and Labelling Procedures - Describes sample
collection, including field QC and preservation. Estimates
the number and kind of samples to be collected. Minimum
requirements for sample collection are described in
Section 2.

Chain of CustodV - Describes Chain-of-Custody and
documentation procedures. Minimum requirements are
described in Section 2.

Analytical Procedures - References or describes in detail
proposed method(s).

Internal Quality Control - Describes type and frequency of
internal quality control. Minimum requirements are
described in Section 3.

Calibration Procedures and Frequency - Describes the
methods and frequency for calibrating field and laboratory
instruments. These must be specified in SOP's.

Appendix A (continued)

Data Verification Describes the data verification in SOP
form and includes; (1) the methods used to identify and
treat outliers, and (2) the data flow from generation of
raw data through storage of verified results.

Data Deliverables - Specifies reporting needs additional
to the minimum requirements described in Section 4.

Technical System and Performance Audits - Specifies'field
or intra-laboratory audits planned by the responsible
Agency.

1.2 Technical System Audits

On-site system audits may be performed without prior notification
by the ACG after consultation with the responsible agency.

1.3 Standards and Ouality Control Materials

The National Institute of Standards and Technology (NIST) will
develop and provide appropriate standards and quality control
materials.

1.4 Analytical Performance Evaluations

Prior to the initiation of work, each analytical laboratory will
be required to demonstrate its capability. This will be
accomplished by providing laboratory documentation on the
performance of the proposed methods and through the analysis of an
accuracy based material. The results of this analysis must be
within +/- 15% of the value of each analyte or measurement
parameter.

Any changes in analytical methodology from that proposed in the
original QA plan shall be validated under agency procedures and
documented to the ACG.

A series of three intercomparison exercises, utilizing the blind
analysis of gravimetrically prepared materials, extracts of
environmental matrices (tissue, sediment and water) or the matrices
themselves, will be conducted annually. Participation in these
exercised is mandatory. Materials will be prepared by, and data

Appendix A (continued)

returned to the NIST for statistical analysis. The NIST will
report to the chairperson of the ACG. Unacceptable perf ormance
will result in the discarding of the associated data.

The ACG will review and provide written reports on the results of
intercomparison studies to the Management Team.

1.5 Data Reporting and Deliverables

Data deliverables will be reviewed by the generating Agency to
verify the quality and useability of the data. A QC report on each
data set will be provided to the ACG for review.

All data and associated documentation will be held in a secure
place under chain-of -custody procedures until the Trustees indicate
otherwise.

2. Minimum Requirements: Sampling and Sampling Equipment

Sample collection activities must be described in SOP's.
References to existing documents are acceptable.

The method of collection should not alter the samples.

Sample collection and storage devices shall not alter the sample.

Samples shall be held in a secure place under appropriate
conditions and under chain-of-custody until the Trustees indicate
otherwise.

2.1 Sampling Identification and Labelling

An SOP will be in place for each study which describes procedures
for the unique identification of each sample. A sample tag or
label will be attached to the sample container. A waterproof
(indelible) marker must be used on the tag or label. Included on
the tag are the sample identification number, the location of the
collection site, the date 'of collection and signature of the
collector.

The information above will also be recorded in a field notebook
along with other pertinent information about the collection and
signed by the collecting scientist.

Appendix A (continued)

2.2 Field Chain-of-Custody

The field sampler will be personally responsible for the care and
custody of the samples collected until they are transferred to
another responsible party.

Samples will be accompanied by a chain-of-custody record or field
sample data record. When samples are transferred from one
individual's custody to another's, the individuals relinquishing
and receiving will sign, date and note the time on the record.

Shipping containers will be custody-sealed for shipment. Whenever
samples are split, a separate chain-of -custody record will be
prepared for those samples and marked to indicate with whom the
samples are being split.

Samples shall be maintained in a manner that preserves their
chemical integrity from collection through final analysis.

Sample shipper will arrange for sample receipt.

After analysis, any remaining sample and all sample tags, labels
and containers shall be held under chain-of -custody procedure until
the Trustees indicate otherwise.

3. Minimum Requirements: Analysis

The applicable methodology must be referenced or described in
detail in the SOP's for each measurement parameter.

Method limits of detection must be calculated by matrix and
analyte.

Control of the analytical method in terms of accuracy and precision
must be demonstrated.

Calibration must be verified at the end of each analysis sequence.

Samples must be quantified within the demonstrated linear working
range for each analyte.

Standard curves must be established with at least 3 points besides
0.

Field blanks, procedural blanks, reference materials, replicates
and analyte recovery samples must be run at a minimum frequency of
5 percent each per sample matrix batch.

Appendix A (continued)

A minimum list of the petroleum hydrocarbon compounds which are to
be considered for identification and quantification in water,
tissue and sediment include the volatiles, i.e., benzene, toluene,
xylene and the polynuclear aromatic and aliphatic hydrocarbons
listed below:

Naphthalene n-dodecane
2-Methylnaphthalene n-tridecane
1-Methylnaphthalene n-tetradecane
Biphenyl n-pentadecane
2,6-Dimethylnaphthalene n-hexadecane
Acenaphthylene n-heptadecane
Acenaphthene pristane
2,3,5-Trimethylnaphthalene n-octadecane
Fluorene phytane
Phenanthrene n-nonadecane
Anthracene n-eicosane
1-Methylphenanthrene
Fluoranthene
Pyrene
Benz(a)anthracene
Chrysene
Benzo(b)fluoranthene
Benzo(k)fluoranthene
Benzo(a)pyrene Benzo(e)pyrene
Indeno(1,2,3-c,d)pyrene Perylene
Dibenz(a,h)anthracene
Benzo(g,h,i)perylene

4. Minimum Requirements: Reporting and Data Deliverables.

Measurement resultst including negative results, as if three
figures were significant must be reported.

Results of quality control samples analyzed in conjunction with the
study samples -must be reported.

Documentation demonstrating analytical control of precision and
accuracy on an analyte and matrix specific basis must be reported.

APPENDIX B

EXXON VALDEZ OIL SPILL DAMAGE ASSESSMENT PLAN
HISTOPATHOLOGY GUIDELINES

Histopathology is an important tool used in determining mechanisms
of death and sublethal effects caused by infectious agents and
toxic substances. A definitive diagnosis often does not result
from histological examination, but can give strong support to other
positive measurements. Tissues deteriorate (autolyze) rapidly
after an animal dies; therefore, to be of value, any samples taken
for histological evaluation as part of the damage assessment of the
Exxon Valdez oil spill must be collected, preserved, and processed
under strict guidelines.

Sample Collection and Preservation Protocols

Standard protocols for necropsy and preservation of tissue samples
for histopathology shall be used throughout the oil spill
assessment studies. Different protocols have been designed to
accommodate the different groups of animals to be encountered in
the assessment studies. Necropsy procedures have been established
and provided to study managers under separate cover for a variety
of different animal groups including finfish, bivalve mollusks,
brachyuran and crab-like anomurans (i.e., king crabs) , shrimp,
marine and terrestrial mammals, and migratory and nonmigratory
waterfowl.

Paired sampling of animals from oiled versus non-oiled sites will
be done for comparative purposes. Histopathological sampling
should be done during any observed acute episodes of mortality or
morbidity to determine the cause of death or abnormality. These
types of samples are the most valuable in assessing acute toxicity
affects and will be the most likely samples collected for birds and
mammals due to their high visibility in the impacted areas.
Because of the low visibility of fish and shellfish, many histology
samples will consist of random collections in impacted and control
areas with little prior obvious indication of morbidity or
mortality.

Any histological processing of samples collected from apparently
normal shellfish will be performed after results of parallel
hydrocarbon sampling are known; i.e., positive hydrocarbon results
may merit further histopathology studies. This would not be
advisable for fish and other higher animals that possess an active
mixed function oxidase (MFO) liver enzyme system which could
metabolize hydrocarbons to other compounds providing negative
hydrocarbon resultsf while potentially still exhibiting
toxicological lesions. Analyses of enzyme function may show an
activated MFO system in exposed fish and higher animals.
Consequently, histology and hydrocarbon samples, as well as other
appropriate samples, such as liver and bile, will be taken from the

same animal when possible for analyses of metabolites and enzyme
function. If certain fish and shellfish are too few or small,
subsampling other animals from the same site at the same time will
be necessary.

Processing and Interpretation Protocols

Histopathology assessment of birds and mammals will be done
primarily on tissues from clinically affected animals using
established criteria of cellular degenerative and necrotic changes
recognized by a board certified veterinary pathologist.

Histopathological analysis of finfish and shellfish tissues will
include the criteria above as well as indices established in the
Amoco Cadiz oil spill studies (Haensly, et al., 1982; Berthou, et
al., 1987) to allow some quantification of potentially subtle
degenerative changes in tissue histology of otherwise clinically
normal animals. Briefly, these indices include mean concentration
of mucus cells per mm.2 of gill lamellae (fish); mean concentration
of mucus cells per mm of epidermis in 10 f ields (f ish) ; mean
concentration of macrophage centers per mm of liver; mean
concentration of hepatocellular vacuolation due to fatty
degeneration (f ish) ; a mean and total tissue necrosis index
(invertebrates) ; histological gonadal index (invertebrates) ; and
differences in prevalences and intensities of incidental lesions
caused by infectious agents (fish and invertebrates).

Ouality Assurance in Field Collection of Samples and in
Interpretation of Results

Field Collection:

Veterinary personnel trained in sample taking will be utilized for
onsite necropsies of birds and mammals in order to ensure adequate
quality control and standardized sample collection. The same high
standards will be attainable in fish and invertebrates in that
sample collection will be done by trained finfish and shellfish
biologists. A fish pathologist and technician are available to
train field personnel and assist in necropsy and preservation of
finfish and shellfish samples at collection sites.

Finfish and shellfish samples can be coordinated through an ADF&G
fish pathologist, Fisheries Rehabilitation, Enhancement and
Development Division.

Interpretation of Results:

Quality control of all processed work will require independent
blind reading of subsampled histology slides by two different
laboratories. Tissues with known lesions will be included
periodically in groups of tissue samples for blind reading and
determination of competency in interpretation.

Chain of Custody Guidelines

Due to the evidentiary nature of sample collecting investigations,
the possession of samples will be traceable from the time the
samples are collected until they are introduced as evidence in
legal proceedings. To maintain and document sample possession,
chain of custody procedures will be followed.

The field sampler will be personally responsible for the care and
custody of the samples collected until they are transferred. All
samples will be accompanied by a chain of custody record and will
be custody-sealed. This procedure includes use of a custody seal
such that the only access to the package is breaking the seal.
When samples are transferred from one individual's custody to
another's, the individuals relinquishing and receiving will sign,
date, and note the time on the record. This record documents the
transfer of custody of samples from the sampler to another person
and, ultimately, to a specified analytical laboratory.

Shipping containers will also be custody-sealed for shipment. The
seal shall be signed before the sample is shipped. The chain of
custody record will be dated and signed to indicate transfer. The
original record will accompany the shipment and a copy will be
retained by the sample collector. Whenever samples are split, a
separate chain of custody record will be prepared for those samples
and marked to indicate with whom the samples are being split. If
samples are being sent by common carrier, copies of all bills of
lading or air bills must be retained as part of the permanent
documentation.

Subcontracting for Histological Work

Subcontracting work for histopathology processing and
interpretation will be under the control of an interagency team
referred to as the Histology Technical Group which will determine
if potential contractors are qualified to do the work.
Qualifications for mammal and avian samples will require a board
certified veterinary pathologist. Finfish and shellfish work will
require individuals with a demonstrated publication record in the
field of histopathology.

References

Bell, T.A., and V.V. Lightner. 1988. A Handbook of Normal/Penaeid
Shrimp Histology. The World Aquaculture Society, Baton Rouge,
LA.

Berthou, F., G. Balouet, G. Bodennec, and M. Marchand. 1987. The
occurrence of hydrocarbons and histophatological abnormalities
in oysters for seven years following the wreck of the Amoco
Cadiz in Brittany (France). Mar. Environ. Res. 23:103-133.

CERCLA. 1988. Natural Resource Damage Assessments. 53 Federal
Regulation 5166 and 9769.

Haensly, W.E., J.M. Neff, J.R. Sharp, A.C. Morris, M.F. Bedgood,
and P.D. Boem. 1982. Histopathology of Pleuronectes platessa
L. from Aber Wrac1h and Aber Benoit, Brittany, France: long-
term effects of the Amoco Cadiz crude oil spill. J. Fish Dis.
5:365-391.

Sparks, A.K. 1985. Synopsis of Invertebrate Pathology Excluding
Insects. Elsevier Publ., New York.

APPENDIX C

GLOSSARY OF TERMS, ACRONYMS

ADF&G Alaska Department of Fish and Game
AFK Armin F. Koernig Fish Hatchery
AHs Aromatic Hydrocarbons
AHH Aryl Hydrocarbon Hydroxylase
ANOVA Analysis of variance
A/W Air/Water
AWL Age, Weight, Length
CERCLA Comprehensive Environmental Response, Compensation and
Liability Act
C/H Coastal Habitat
CI Cook Inlet
CIK Cook Inlet/Kenai
CTD Conductivity/temperature/depth
CWA Clean Water Act
CWT Coded wire tag
DEC Alaska Department of Environmental Conservation
DNR Alaska Department of Natural Resources
DOA Department of Agriculture
DOC Department of Commerce
DOI Department of the Interior
DOJ Department of Justice
DBMS Database Management System
EPA Environmental Protection Agency
EIS Economic Study
EVOS Exxon Valdez Oil Spill
FRED Fisheries Rehabilitation, Enhancement and Development
Division, ADF&G
FIS Fish/Shellfish
FWS U.S. Fish and Wildlife Service
GC-MS Gas chromatography-mass spectrometry
GOA Gulf of Alaska
KAP Kodiak Archipelago/Alaska Peninsula
KP Kenai Peninsula
LCI Lower Cook Inlet
Mro Mix@d function oxiaase
MLLW Mean lower low water
M/M Marine Mammal
NIOSH National Institute of Occupational Safety and Health
NMFS National Marine Fisheries Service
NOAA National Oceanic and Atmospheric Administration
NPH Naphthalene
NPS National Park Service

APPENDIX C

GLOSSARY OF TERMS, ACRONYMS

NRDA Natural Resource Damage Assessment
NSO Nitrogen-sulphur-oxygen
PED Potential egg deposition
PHN Phenanthrene
PI Principal Investigator(s)
PWS Prince William Sound
PWSAC Prince William Sound Aquaculture
QA/QC Quality Assurance/Quality Control
RPWG Restoration Planning Work Group
SCAT Shoreline Cleanup Advisory Team
SSAT Spring Shoreline Assessment Team
T/M Terrestrial Mammals
T/S Technical Services
USFS United States Forest Service
VFDA Valdez Fisheries Development Association

2
U.S. GPO * 1990 794-976

; fluffilimm,
3 6668 00003 7541