[Federal Register Volume 61, Number 89 (Tuesday, May 7, 1996)]
[Rules and Regulations]
[Pages 20560-20614]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 96-10747]




[[Page 20559]]


_______________________________________________________________________

Part II





Department of the Interior





_______________________________________________________________________



Office of the Secretary



_______________________________________________________________________



43 CFR Part 11



Natural Resource Damage Assessments: Type A Procedures; Final Rule

  Federal Register / Vol. 61, No. 89 / Tuesday, May 7, 1996 / Rules and 
Regulations  

[[Page 20560]]



DEPARTMENT OF THE INTERIOR

Office of the Secretary

43 CFR Part 11

RIN 1090-AA21 & 1090-AA23


Natural Resource Damage Assessments--Type A Procedures

AGENCY: Department of the Interior.

ACTION: Final rule.

-----------------------------------------------------------------------

SUMMARY: This final rule amends the regulations for assessing natural 
resource damages under the Comprehensive Environmental Response, 
Compensation, and Liability Act. Federal, State, and Indian tribe 
natural resource trustees may use these regulations to obtain 
compensation from potentially responsible parties for natural resource 
injuries resulting from hazardous substance releases. Trustees obtain a 
rebuttable presumption in litigation for damages, up to $100,000, 
calculated in accordance with this rule. The rule does not change the 
overall administrative process for conducting assessments but simply 
revises an existing ``type A'' procedure for assessing natural resource 
damages in coastal and marine environments and establishes a new type A 
procedure for the Great Lakes.

EFFECTIVE DATE: The effective date of this final rule is June 6, 1996. 
The incorporation by reference of certain documents listed in this rule 
was approved by the Director of the Federal Register and is effective 
June 6, 1996.

FOR FURTHER INFORMATION CONTACT: Mary Morton at (202) 208-3302 (for 
questions about the rule language) or David Rosenberger at (202) 208-
3811 (for questions about the computer models). Interested parties may 
obtain copies of the computer models and supporting documentation free 
of charge from the Department through July 31, 1996, and thereafter for 
a fee from the National Technical Information Service, 5285 Port Royal 
Road, Springfield, VA 22161, ph: (703) 487-4650. The models are also on 
the Internet at http://www.usgs.gov/doi/oepc/oepchome.html.

SUPPLEMENTARY INFORMATION: This preamble is organized as follows:

I. Background
    A. Statutory Provisions
    B. History of this Rulemaking
    C. Oil Pollution Act Regulations
II. Relationship of Today's Final Rule to the Existing Regulations
    A. Preassessment Phase
    B. Assessment Plan Phase
    C. Assessment Phase
    D. Post-Assessment Phase
III. Nature of Type A Procedures
IV. Workings of the NRDAM/CME and NRDAM/GLE
    A. Overview
    B. Data Inputs and Modifications
    C. Geographic Information System
    D. Submodels
V. Use of the NRDAM/CME and NRDAM/GLE in Other Contexts
VI. Summary of Major Changes from the Proposed Rules
    A. Rule Language
    B. NRDAM/CME and NRDAM/GLE
VII. Response to Comments
    A. General Comments
    B. Technical Documents
    C. Selection of Assessment Procedures
    D. User-Supplied Information
    E. Physical Fates
    F. Species Distribution and Abundance
    G. Toxicity and Mortality
    H. Loss of Production
    I. Catch and Bag Losses
    J. Habitat Restoration
    K. Assimilative Capacity Restoration
    L. Restocking
    M. Consideration of Costs and Benefits of Active Restoration
    N. Damages for Fishing and Hunting Losses
    O. Damages for Lost Wildlife Viewing
    P. Damages for Beach and Boating Closures
    Q. Judicial Review and the Rebuttable Presumption

I. Background

A. Statutory Provisions

    The Department of the Interior (the Department) is amending the 
regulations for assessing natural resource damages under the 
Comprehensive Environmental Response, Compensation, and Liability Act, 
as amended (42 U.S.C. 9601 et seq.) (CERCLA). CERCLA provides that 
certain categories of persons, known as potentially responsible parties 
(PRPs), are liable for natural resource damages resulting from a 
release of a hazardous substance. CERCLA sec. 107(a). Natural resource 
damages are monetary compensation for injury to, destruction of, or 
loss of natural resources. CERCLA sec. 107(a)(4)(C).
    Only those Federal, State, and Indian tribe officials designated as 
natural resource trustees may recover natural resource damages. CERCLA 
defines ``State'' to include:

    The District of Columbia, the Commonwealth of Puerto Rico, Guam, 
American Samoa, the United States Virgin Islands, the Commonwealth 
of the Northern Marianas, and any other territory or possession over 
which the United States has jurisdiction. CERCLA sec. 101(27).

    Trustees must use all sums they recover in compensation for natural 
resource injuries to restore, rehabilitate, replace, or acquire the 
equivalent of the injured natural resources. CERCLA sec. 107(f)(1). 
Trustee officials may also recover the reasonable costs of assessing 
natural resource damages. Natural resource damages are distinct from 
response costs. Response costs are the costs of actions taken under the 
National Contingency Plan (40 CFR part 300) to remove threats to human 
health and the environment caused by hazardous substance releases. 
Today's final rule addresses only the assessment of natural resource 
damages and is not intended for use in connection with response-related 
activities, such as setting cleanup priorities.
    CERCLA requires the President to promulgate regulations for the 
assessment of natural resource damages resulting from hazardous 
substance releases. CERCLA sec. 301(c). The President delegated the 
responsibility for promulgating these regulations to the Department. 
E.O. 12316, as amended by E.O. 12580. The regulations must identify the 
``best available'' procedures for assessing natural resource damages. 
CERCLA sec. 301(c)(2). CERCLA requires that the natural resource damage 
assessment regulations include two types of assessment procedures. 
``Type A'' procedures are ``standard procedures for simplified 
assessments requiring minimal field observation.'' CERCLA sec. 
301(c)(2)(A). ``Type B'' procedures are ``alternative protocols for 
conducting assessments in individual cases.'' CERCLA sec. 301(c)(2)(B). 
Federal and State trustees who perform assessments in accordance with 
these regulations receive a rebuttable presumption in court. CERCLA 
sec. 107(f)(2)(C). The Department must review the regulations, and 
revise them as appropriate, every two years. CERCLA sec. 301(c)(3).

B. History of this Rulemaking

    On March 20, 1987, the Department published a final rule 
establishing a type A procedure for coastal and marine environments 
that incorporated a computer model, known as the Natural Resource 
Damage Assessment Model for Coastal and Marine Environments (NRDAM/
CME). 52 FR 9041. The Department indicated that it would consider 
developing additional type A procedures as experience was gained with 
the type A procedure for coastal and marine environments. Id. at 9057. 
On June 2, 1988, the Department published an advance notice of proposed 
rulemaking soliciting comment on the development of a type A procedure 
for Great Lakes environments that would incorporate a computer model 
called the Natural Resource Damage Assessment Model for Great Lakes 
Environments (NRDAM/GLE). 53 FR 20143. A few months later,

[[Page 20561]]

the Department published an advance notice of proposed rulemaking 
announcing the commencement of the statutorily required biennial review 
of the type A procedure for coastal and marine environments. 54 FR 5093 
(Feb. 1, 1989).
    On July 14, 1989, the U.S. Court of Appeals for the District of 
Columbia Circuit issued two decisions that affected these two pending 
type A rulemakings. The Department had issued type B procedures on 
August 1, 1986. 51 FR 27674. State, industry, and environmental group 
petitioners challenged those procedures in State of Ohio v. United 
States Department of the Interior (Ohio v. Interior), 880 F.2d 432 
(D.C. Cir. 1989). The court in Ohio v. Interior upheld various aspects 
of the type B procedures but ordered the Department to revise the type 
B procedures to reflect the statutory preference for using restoration 
costs as the measure of natural resource damages. The court used the 
term ``restoration costs'' to encompass the cost of restoring, 
rehabilitating, replacing, and/or acquiring the equivalent of the 
injured natural resources. The court also ordered the Department to 
revise the type B procedures to allow for the recovery of all reliably 
calculated values lost to the public as a result of the injury to 
natural resources.
    State, industry, and environmental group petitioners also 
challenged the original type A procedure for coastal and marine 
environments in State of Colorado v. United States Department of the 
Interior (Colorado v. Interior), 880 F.2d 481 (D.C. Cir. 1989). The 
court in Colorado v. Interior upheld the Department's sequential 
approach to developing type A procedures but urged the Department to 
develop additional type A procedures to address as many different cases 
as possible. The court also remanded the type A procedure for coastal 
and marine environments, based on the reasoning in the Ohio v. Interior 
decision, to permit the Department to allow for the calculation of 
restoration costs. The original type A procedure for coastal and marine 
environments calculated damages based solely on certain lost public 
uses of the injured resources.
    On September 22, 1989, the Department published an advance notice 
of proposed rulemaking stating that it would revise the type A 
procedure for coastal and marine environments in compliance with Ohio 
v. Interior and Colorado v. Interior during the ongoing biennial 
review. 54 FR 39013. The Department also announced that it would modify 
the development of the type A procedure for Great Lakes environments to 
conform with Ohio v. Interior and Colorado v. Interior. 54 FR 39015 
(Sept. 22, 1989).
    The Department published a notice of proposed rulemaking for the 
type A procedure for Great Lakes environments on August 8, 1994. 59 FR 
40319. The August 8, 1994, Federal Register notice also contained two 
proposed amendments to the natural resource damage assessment 
regulations that would affect all type A procedures. The Department 
proposed to revise the conditions under which both type A and type B 
procedures could be used in the same assessment, and to make explicit 
the scope of judicial review of assessments performed using type A 
procedures. The Department later extended the comment period on the 
August 8, 1994, proposed rule through February 6, 1995. 59 FR 54877 
(Nov. 2, 1994).
    On December 8, 1994, the Department issued a notice of proposed 
rulemaking for the modified type A procedure for coastal and marine 
environments. 59 FR 63300. On February 7, 1995, the Department extended 
the comment periods on both the proposed Great Lakes type A rule and 
the proposed coastal and marine type A rule through July 6, 1995. 60 FR 
7155 and 7156. The Department noted that, in light of the similarities 
between the two proposed rules, it would consider the public comments 
on the two rules concurrently. Id. at 7156 and 7157. Today's final rule 
covers both the type A procedure for coastal and marine environments 
and the type A procedure for Great Lakes environments.

C. Oil Pollution Act Regulations

    Originally, trustees could use the Department's regulations to 
assess natural resource damages resulting from either a hazardous 
substance release under CERCLA or an oil or hazardous substance 
discharge into navigable waters under the Clean Water Act (33 U.S.C. 
1251 et seq.). However, the Oil Pollution Act of 1990 (OPA) amended the 
natural resource damage provisions of the Clean Water Act. See 33 
U.S.C. 1321, 2702(b)(2), and 2706(a). OPA authorized the National 
Oceanic and Atmospheric Administration (NOAA) to develop new natural 
resource damage assessment regulations for assessing natural resource 
damages resulting from discharges, or threats of discharges, of oil 
into navigable waters that, once final, would supersede the provisions 
of the Department's regulations addressing oil. 33 U.S.C 2706(e)(1) and 
2751(b). NOAA published a final OPA rule on January 5, 1996. 61 FR 439.
    The Department began developing the type A procedures before the 
enactment of OPA and, thus, originally included both hazardous 
substances and oil in the NRDAM/CME and NRDAM/GLE algorithms and 
databases. The Department has worked closely with NOAA during the 
development of the type A procedures. During its rulemaking, NOAA 
indicated it would allow use of the Department's type A procedures 
under the OPA regulations. See 59 FR 1062, 1124-25 (Jan. 7, 1994); and 
60 FR 39803, 39831 (Aug. 3, 1995).
    NOAA's final rule states that trustees may use ``[m]odel-based 
procedures, including type A procedures identified in 43 CFR part 11, 
subpart D,'' provided that any such procedure meets the following 
conditions:

    (1) The procedure must be capable of providing assessment 
information of use in determining the type and scale of restoration 
appropriate for a particular injury;
    (2) The additional cost of a more complex procedure must be 
reasonably related to the expected increase in the quantity and/or 
quality of relevant information provided by the more complex 
procedure; and
    (3) The procedure must be reliable and valid for the particular 
incident. 61 FR at 503 (15 CFR 990.27).

    Therefore, the Department has retained components relating to oil 
in the final versions of the NRDAM/CME and NRDAM/GLE, while recognizing 
that these components are without any direct regulatory effect. The 
Department is also providing responses to comments it received on the 
oil-related components of the type A models. However, the Department 
wishes to emphasize that its regulations do not govern the assessment 
of natural resource damages for oil discharges under OPA. Trustees who 
wish to use the type A procedures and obtain a rebuttable presumption 
for assessments of oil discharges must follow the process established 
by NOAA's regulations.
    Further, some of the language in the CERCLA rule varies from that 
in the OPA rule. For example, today's final rule incorporates the 
existing definition of ``reasonable cost'' at 43 CFR 11.14, from which 
the definition in the OPA rule differs. See 61 FR at 504 (15 CFR 
990.30). Section 11.35(b) of today's final rule, which requires 
trustees to conduct type B procedures if the PRPs advance the 
reasonable costs of using such procedures, differs from the OPA rule 
conditions governing PRP requests for alternative assessment 
procedures. See 61 FR at 501 (15 CFR 990.14(b)(6)). Also, Sec. 11.44(f) 
of today's final rule

[[Page 20562]]

provides that if the models calculate damages in excess of $100,000, 
then trustees who wish to obtain a rebuttable presumption must either: 
(1) limit the portion of their claim calculated with the type A 
procedure to $100,000; or (2) compute all damages using type B 
procedures. The OPA rule, on the other hand, contains no dollar cut-off 
for use of specific procedures. Because use of the type A procedures 
for oil discharges is governed by the OPA rule, the Department defers 
to NOAA on how such differences are to be resolved when the NRDAM/CME 
and NRDAM/GLE are used for assessments of oil discharges.

II. Relationship of Today's Final Rule to the Existing Regulations

    The existing regulations establish an administrative process for 
conducting assessments. See 43 CFR part 11. The administrative process 
covers all the steps trustees need to follow if they wish to obtain a 
rebuttable presumption in litigation of their claim. However, trustees 
have the authority to settle their damage claims at any time during the 
administrative process and the Department continues to encourage 
trustees and PRPs to pursue settlement. Furthermore, trustees are not 
required to follow the regulations. If, however, trustees and PRPs fail 
to reach a settlement and the case is litigated, trustees will only 
obtain a rebuttable presumption if they performed their assessment in 
accordance with the regulations.
    The same general administrative process applies regardless of 
whether type A or type B procedures are used. The process has four 
phases: Preassessment, Assessment Plan, Assessment, and Post-
Assessment. During the Assessment Phase, trustees use type A and/or 
type B procedures to perform the technical work needed for the actual 
determination of damages.
    Today's final rule does not change this overall administrative 
process. The rule simply revises the type A procedures available for 
use during the Assessment Phase and modifies the standards for using 
both type A and type B procedures for the same release.

A. Preassessment Phase

    Today's final rule does not affect the Preassessment Phase. The 
Preassessment Phase consists of the activities that precede the actual 
assessment. For example, upon detecting or receiving notification of a 
release, trustees decide, based on a number of criteria, whether 
further assessment actions are warranted. Trustees document this 
decision in the Preassessment Screen Determination. For more 
information on the Preassessment Phase, see subpart B of 43 CFR part 
11.

B. Assessment Plan Phase

    If trustees determine that additional assessment work is warranted, 
they begin the Assessment Plan Phase. The Assessment Plan Phase 
includes the preparation of a written Assessment Plan describing the 
procedures trustees intend to use to determine damages. The trustees 
must make the draft Assessment Plan available for public review and 
comment.
    The regulations provide two types of assessment procedures: type A 
and type B. Type A procedures, such as those contained in today's final 
rule, are simplified procedures requiring minimal field observation. 
Type B procedures involve more detailed field studies. The Assessment 
Plan documents whether trustees plan to use a type A procedure, type B 
procedures, or both. Today's final rule revises the standards that 
trustees must follow when selecting assessment procedures .
    Section 11.34 of today's final rule identifies several conditions 
that must be met before trustees can use a type A procedure and obtain 
a rebuttable presumption. If the conditions are not met, then trustees 
who elect to follow the regulations must use type B procedures to 
assess all damages. If the conditions are met, then trustees must 
decide whether to use a type A procedure, type B procedures, or both. 
This decision is based on whether the benefits of the increased 
accuracy provided by type B procedures would offset the anticipated 
additional cost of using type B procedures, and whether the anticipated 
damages would exceed the anticipated cost of using type B procedures.
    Trustees may use both type A and type B procedures for the same 
release if: (1) The type B procedures are cost-effective and can be 
performed at a reasonable cost; (2) the type B procedures are used only 
to determine damages for injuries or economic values of a type not 
addressed by the type A procedure; and (3) there is no double recovery. 
Section 11.36 of the final rule lists the categories of damages that 
are included in the type A models and for which trustees may not 
conduct supplemental type B studies. Trustees must document in the 
Assessment Plan how they intend to prevent double recovery when they 
use both type A and type B procedures.
    Today's final rule also maintains the requirement that trustees use 
type B procedures, even if they determine that use of a type A 
procedure would be appropriate, whenever a PRP submits a written 
request and justification for use of type B procedures and advances all 
reasonable costs of using type B procedures within a time frame 
acceptable to the trustees.
    For more information on the Assessment Plan Phase, see Secs. 11.30 
through 11.37 of today's final rule and subpart C of 43 CFR part 11.

C. Assessment Phase

    During the Assessment Phase, trustees conduct the work described in 
the Assessment Plan. The work consists of three steps: Injury 
Determination; Quantification; and Damage Determination. In Injury 
Determination, trustees determine whether any natural resources have 
been injured. If trustees determine that resources have been injured, 
they proceed to Quantification, in which they quantify the resulting 
change in baseline conditions. ``Baseline'' conditions are the 
conditions that would have existed had the release not occurred. 
Finally, in Damage Determination, trustees calculate the monetary 
compensation to be sought as damages for the natural resource injuries. 
Damages include two components: (1) The cost of restoring, 
rehabilitating, replacing, and/or acquiring the equivalent of the 
injured natural resources; and (2) the economic value lost by the 
public pending recovery of the resources (compensable value).
    When trustees use type B procedures, they perform Injury 
Determination, Quantification, and Damage Determination through 
laboratory and field studies. The regulations provide a range of 
alternative type B scientific and economic methodologies for conducting 
such studies. For more information on use of type B procedures during 
the Assessment Phase, see subpart E of 43 CFR part 11.
    When trustees use a type A procedure, they perform Injury 
Determination, Quantification, and Damage Determination through a 
computer model. Today's type A procedure for coastal and marine 
environments incorporates Version 2.4 of the NRDAM/CME. Today's type A 
procedure for Great Lakes environments incorporates Version 1.4 of the 
NRDAM/GLE.
    Trustees must supply a number of data inputs to operate the NRDAM/
CME and the NRDAM/GLE. The rule also requires trustees to modify 
certain data contained in the models if they have more reliable 
information. Section 11.41 and Appendices II and III of the final rule 
describe the required data inputs and modifications. After trustees 
supply

[[Page 20563]]

the data inputs and modifications, the models themselves perform the 
remaining calculations necessary to establish if there has been an 
injury, quantify the extent of injury, select appropriate restoration 
actions, and value economic losses. With the availability of these 
computer models, trustees will now be able to pursue compensation for 
cases in which the cost of detailed type B studies is prohibitive.
    Trustees may not implement type B procedures until after the public 
review period on the Assessment Plan. However, today's final rule 
provides that trustees who use a type A procedure must perform a 
preliminary application of the model before issuing the Assessment Plan 
and then include the data inputs and the results of the preliminary 
application in the publicly reviewed Plan. This requirement should 
provide PRPs and other members of the public with a more meaningful 
opportunity for comment. Performance of a preliminary application of 
the models will also allow trustees to determine if type B procedures 
are warranted in light of a new cap on the damages that can be claimed 
through use of a type A procedure.
    The rule now provides that if the preliminary application indicates 
damages in excess of $100,000, then trustees who wish to obtain a 
rebuttable presumption must decide whether to: (1) limit the portion of 
their claim calculated with the type A procedure to $100,000; or (2) 
compute all damages using type B procedures. The $100,000 limit applies 
only to damages calculated by a type A procedure and does not limit 
damages calculated through supplemental type B studies. This dollar 
cut-off is based on the fairness of allowing trustees to receive a 
rebuttable presumption for damages calculated by the NRDAM/CME or 
NRDAM/GLE given the current level of experience with these models. The 
cut-off is not based on reliability. The Department believes the NRDAM/
CME and NRDAM/GLE are capable of generating reliable damage estimates 
at levels above $100,000. Therefore, although trustees cannot use the 
models and obtain a rebuttable presumption above $100,000, the 
Department believes the models are appropriate for use in other 
contexts, such as settlement negotiations and litigation without the 
rebuttable presumption.
    After the close of the comment period on the Assessment Plan, 
trustees must carefully review and substantively respond to all 
comments they receive and must decide whether to continue using the 
type A procedure. If they do decide to continue using the type A 
procedure, they must make any necessary revisions to the user inputs, 
and perform a final application of the model.
    For more information on the Assessment Phase, see Secs. 11.40 
through 11.44 of the final rule. For more information on how the NRDAM/
CME and the NRDAM/GLE perform Injury Determination, Quantification, and 
Damage Determination, see Section IV of this preamble.

D. Post-Assessment Phase

    Once the Assessment Phase is completed, trustees enter the Post-
Assessment Phase. Today's final rule does not substantively modify the 
Post-Assessment Phase.
    During the Post-Assessment Phase, trustees prepare a Report of 
Assessment detailing the results of the Assessment Phase. When trustees 
use a type A procedure, the Report will include the printed output of 
the final model application. If a trustee is aware of reliable evidence 
that a private party has recovered damages for commercial harvests lost 
as a result of the release, the trustee must eliminate from the claim 
any damages for such lost harvests included in the lost economic rent 
calculated by the model. If a trustee is aware of reliable evidence 
that the model application covers resources beyond his or her trustee 
jurisdiction, the trustee must either: (1) have the other trustees who 
do have jurisdiction over those resources join in the type A 
assessment; or (2) eliminate any damages for those resources from the 
claim.
    Trustees present the Report of Assessment to the PRPs along with a 
demand for damages and assessment costs. If a PRP does not agree to pay 
within 60 days, the trustees may file suit. Federal and State trustees 
receive a rebuttable presumption of correctness if they performed their 
assessments in accordance with the Preassessment Phase, Assessment Plan 
Phase, Assessment Phase, and Post-Assessment Phase requirements set 
forth in the regulations. Once a court awards damages or the trustees 
and PRPs have reached a settlement, trustees establish an account to 
hold the recovered damages pending preparation of a Restoration Plan 
describing how they intend to use the funds.
    When trustees use a type A procedure, they are not restricted to 
implementing the general restoration methods used by the model to 
calculate the restoration cost component of the damage claim. Instead, 
trustees have the discretion to spend recovered sums on other actions 
to restore, rehabilitate, replace, and/or acquire the equivalent of the 
injured resources.
    Also, existing 43 CFR 11.93(d), which was not a subject of this 
rulemaking, provides that trustees may apply several type A recoveries 
to a single Restoration Plan, so long as the Plan is intended to 
address the same or similar injuries as those identified in each 
application of the type A procedure.
    For more information on the Post-Assessment Phase, see subpart F of 
43 CFR part 11.

III. Nature of Type A Procedures

    The Department believes it is important that trustees, PRPs, and 
the public clearly understand what the type A procedures are, as well 
as what they are not, intended to provide. The NRDAM/CME and the NRDAM/
GLE are sophisticated computer models. These models incorporate a 
significant level of site-specific detail about actual physical and 
biological conditions in the geographic areas they encompass. The 
language and legislative history of CERCLA suggest that Congress 
envisioned type A procedures as look-up tables based on dollars per 
gallon or unit of affected area. See, e.g., S. Rep. No. 96-848, 96th 
Cong. 2d Sess. 86 (July 11, 1980). In requiring the development of two 
types of assessment procedures--one simplified and the other more 
complex and site-specific--Congress made a policy choice that trustees 
be provided with a simplified, inexpensive mechanism for obtaining 
recoveries in smaller cases. By envisioning a mechanism such as a look-
up table, Congress obviously recognized that trustees who use type A 
procedures should not be required to develop--or be prejudiced for not 
developing--the same degree of site-specific accuracy as might be 
achieved using more expensive type B procedures. Nevertheless, in order 
to increase accuracy, the Department has developed computer models that 
enable the consideration of site-specific factors. For example, the 
NRDAM/CME and NRDAM/GLE take into account physical variations among 
geographic areas, differences in the toxicity and physical 
characteristics of hazardous substances, seasonal and temperature 
effects, and differences in the biological productivity of the spill 
site. The Department believes that when applied correctly using 
reliable input data, the NRDAM/CME and NRDAM/GLE are powerful, reliable 
tools for assessing the injuries and compensable values they address.
    However, as sophisticated and reliable as they are, the NRDAM/CME 
and NRDAM/GLE do not, and were

[[Page 20564]]

never intended to, constitute automated type B procedures. The NRDAM/
CME and NRDAM/GLE are, after all, only models of selected aspects of 
reality and, like all models, they are incapable of precisely capturing 
reality in every case. Modeling always necessitates some simplifying 
assumptions, and the modeling of something as complex as the effects of 
hazardous substance spills on natural resources necessitates numerous 
simplifying assumptions.
    Section 11.34 of the final rule identifies a number of assumptions 
the Department made during the development of the NRDAM/CME and NRDAM/
GLE. If these assumptions are not reasonable in a particular case, 
trustees may not use the models and obtain a rebuttable presumption. 
But even when these assumptions are reasonable, the models' damage 
estimates will differ from the damages that type B procedures would 
produce. However, Congress explicitly authorized the development of 
simplified type A procedures that required less field work than type B 
procedures and then explicitly granted a rebuttable presumption to 
assessments performed using these type A procedures just as it granted 
a rebuttable presumption to assessments performed using type B 
procedures. Finally, the Department has retained in today's final rule 
the safety valve that always allows PRPs to require trustees to use 
type B procedures rather than a type A procedure if they advance all 
reasonable costs of using such type B procedures within an acceptable 
time frame.
    The standard for evaluating the results of the NRDAM/CME or the 
NRDAM/GLE in a particular case is not whether the model projections 
conform precisely to field observations. Rather, the standard is 
whether the overall damage figure calculated by the models is fair and 
reasonable in light of the feasibility and cost of developing more 
specific information using type B procedures. For example, if a spill 
occurs in an area where biological conditions are relatively uniform 
over a wide area, the fact that the NRDAM/CME or NRDAM/GLE project that 
the surface trajectory would turn to the right when in fact it turned 
to the left is not necessarily adequate grounds to reject wholesale the 
results of the model.

IV. Workings of the NRDAM/CME and NRDAM/GLE

A. Overview

    The NRDAM/CME and the NRDAM/GLE consist of integrated submodels and 
databases that calculate natural resource damages based on certain 
types of estimated restoration costs and compensable values. The NRDAM/
CME and the NRDAM/GLE are complex computer models; however, their use 
is not restricted to computer specialists.
    The NRDAM/CME was developed under contract to the Department by 
Applied Science Associates, Inc., A.T. Kearney, Inc., and Hagler Bailly 
Consulting, Inc. The NRDAM/GLE was developed under contract to the 
Department by Applied Science Associates, Inc., and Hagler Bailly 
Consulting, Inc.
    ``CERCLA Type A Natural Resource Damage Assessment Model for 
Coastal and Marine Environments Technical Documentation,'' dated April 
1996 (the NRDAM/CME technical document) describes the NRDAM/CME. Volume 
I of the NRDAM/CME technical document discusses the content and 
derivation of the NRDAM/CME submodels and databases. Volume II is a 
user's manual. Volume III is a compilation of the chemical and 
environmental databases used by the NRDAM/CME. Volume IV contains the 
biological databases on the species life histories, species abundances, 
and trophic-level production rates used by the NRDAM/CME. Volume V is a 
compilation of the compensable values and restoration costs used by the 
NRDAM/CME. Volume VI is a listing of the active source code for the 
NRDAM/CME.
    ``CERCLA Type A Natural Resource Damage Assessment Model for Great 
Lakes Environments Technical Documentation,'' dated April 1996 (the 
NRDAM/GLE technical document) describes the NRDAM/GLE. Volume I of the 
NRDAM/GLE technical document discusses the content and derivation of 
the NRDAM/GLE submodels and databases. Volume II is a user's manual. 
Volume III is a compilation of all the databases used by the NRDAM/GLE. 
Volume IV is a listing of the active source code for the NRDAM/GLE.
    Today's final rule incorporates by reference the NRDAM/CME, the 
NRDAM/CME technical document, the NRDAM/GLE, and the NRDAM/GLE 
technical document. Anyone can obtain computer diskettes containing the 
models and technical documents from the National Technical Information 
Service for a fee. The technical documents supplied on diskette are 
formatted in WordPerfect 5.1. Some databases are formatted in 
QuatroPro. Hard-bound copies of the technical documents are 
also available. Also, to facilitate prompt distribution of the models, 
the Department will be providing diskettes of the models and technical 
documents free of charge until July 31, 1996.
    The models have a menu-driven graphic display to assist users. The 
minimum computer configuration required to use the models is:
     IBM-compatible personal computer (PC) using MS-
DOS 3.3 or higher;
     80386 processor or better with math co-processor;
     1.4 megabyte 3.5 inch floppy disk drive;
     4 megabytes of RAM with 540 kilobytes available;
     Hard disk with 75 megabytes of available space;
     VGA monitor; and
     Microsoft-compatible mouse and mouse driver 
software. For further information on installation of the models, see 
Section 2, Volume II of the NRDAM/CME and NRDAM/GLE technical 
documents.

B. Data Inputs and Modifications

    The models' databases include most of the data used by the models 
to determine injury and damages. However, the final rule requires 
trustees to provide certain data inputs. The rule also requires 
trustees to modify certain data contained in the models if they have 
more reliable information. The required data inputs and modifications 
are described in Sec. 11.41 and Appendices II and III.
    Trustees may have direct knowledge of some of the required data 
inputs. Additional information may be available from the On-Scene 
Coordinator (OSC), who is responsible for managing response actions 
following a release. The U.S. Coast Guard will normally be the OSC for 
releases in coastal or marine environments or the Great Lakes. However, 
trustees remain responsible for ensuring that all data inputs are 
reliable.

C. Geographic Information System

    The models incorporate a geographic information system (GIS) that 
supplies geographically distributed information to the submodels. The 
submodels divide space into series of rectangular grids. In the NRDAM/
CME, each grid contains 10,000 cells (100  x  100). In the NRDAM/GLE, 
each grid contains 2,500 cells (50  x  50). The size of a specific grid 
and, therefore, the interior cells, varies based on the physical 
geometry of and the availability of natural resource information about 
the particular geographic area. For example, the GIS uses smaller grids 
for nearshore areas than for offshore areas. The models assign a 
habitat type to each grid cell. The GIS draws the necessary

[[Page 20565]]

environmental and biotic data from the appropriate databases. The 
models assume that conditions are uniform throughout a particular grid 
cell.
    For further information about the GIS and grid system, see Section 
2, Volume I of the NRDAM/CME technical document; and Section 3.15, 
Volume I of the NRDAM/GLE technical document.

D. Submodels

    Both models include four linked submodels: a physical fates 
submodel, a biological effects submodel, a restoration submodel, and a 
compensable value submodel. The NRDAM/GLE also has a hydrodynamics 
submodel.
1. Physical Fates Submodel
    The physical fates submodel estimates the distribution of the 
released substance on the water surface, along shorelines, in the water 
column, and in sediments over time. The submodel uses an array of 
computational ``particles'' to represent the released substance. A 
variable fraction of the released substance is associated with each 
particle. The submodel tracks the distribution of the particles in both 
time and space as they move across a three-dimensional gridded 
environment.
    Modeled wind and current effects drive the movement of the 
particles on the water surface and in the water column. In the NRDAM/
GLE, the hydrodynamics submodel simulates the wind-driven currents 
occurring in the water column. In the NRDAM/CME, the physical fates 
submodel simulates wind-driven currents in the upper water column and 
employs user-supplied data inputs on background and tidal currents to 
simulate movement in the upper and lower water column.
    Drawing data about the physical and chemical properties of the 
released substance from the chemical and toxicological database, the 
submodel continues simulating the transport and fate of the substance 
until all environmental exposure levels are below a specified 
concentration (the acute toxicity threshold). The acute toxicity 
threshold serves as a switch to turn off the physical fates submodel 
and activate the biological effects submodel. The submodel creates a 
time-series file of surface slick coverage, shoreline coverage, and 
substance concentration levels in the water column and in bottom 
sediments that is used by the biological effects submodel.
    For further information on the physical fates submodel, see Section 
3, Volume I of the NRDAM/CME and NRDAM/GLE technical documents. For 
further information on the chemical and toxicological database, see 
Section 7, Volume I, and Section 2, Volume III of the NRDAM/CME and 
NRDAM/GLE technical documents.
2. Biological Effects Submodel
    The biological effects submodel determines whether certain types of 
natural resource injuries have resulted from the release and, if so, 
quantifies those injuries. The biological effects submodel determines 
and quantifies the following types of injury: (1) Direct mortality 
resulting from short-term exposure to the released substance; (2) 
direct loss of production resulting from short-term exposure to the 
released substance; (3) indirect mortality resulting from food web 
losses; and (4) indirect loss of production resulting from food web 
losses. The biological database supplies data on habitat type and 
species biomass to the biological effects submodel.
    The biological effects submodel determines direct mortality of 
fish, shellfish, and wildlife and direct loss of production for plants 
and invertebrates by calculating exposure of different species to the 
released substance. When performing these calculations, the biological 
effects submodel uses the time series data generated by the physical 
fates submodel concerning the distribution and concentration of the 
released substance.
    The biological effects submodel determines direct mortality of fish 
and shellfish through use of an array of computational ``particles'' 
that move through the gridded environment. Each particle represents a 
portion of the fish or shellfish populations potentially exposed to the 
release. Each time a particle enters an area with dissolved water or 
sediment concentrations of the spilled substance, the submodel 
calculates the percentage mortality of the fish or shellfish population 
represented by the particle. These calculations continue until 
concentrations of the released substance fall below acute toxicity 
thresholds.
    The biological effects submodel uses similar procedures to 
determine direct mortality of birds and mammals. However, the submodel 
only determines direct mortality of birds and mammals when the released 
substance forms a surface slick.
    The biological effects submodel determines direct mortality of fish 
and shellfish eggs and larvae through use of particle arrays that move 
with the currents, as biologically appropriate. For plants and 
invertebrates, the submodel determines direct loss of production based 
on the assumption that such biota are uniformly distributed throughout 
a particular habitat type within the model grids rather than through 
use of particle arrays.
    Once the biological effects submodel determines direct mortality 
and direct loss of production , the submodel then calculates indirect 
mortality and indirect loss of production for fish, shellfish, and 
wildlife resulting from reductions in food resources. The submodel uses 
a generalized food web model to determine the effect that direct loss 
of plant production, invertebrates, and noncommercial fish and mammals 
have on higher trophic-level fish, shellfish, and wildlife.
    After determining injuries from both direct exposure and food web 
losses, the biological effects submodel quantifies those injuries both 
in terms of lost populations over time and, in the case of fish, 
shellfish, and wildlife, fishing and hunting losses. The submodel also 
computes fishing and hunting losses resulting from closures. The 
compensable value submodel uses this information to determine 
compensable value.
    For further information on the biological effects submodel, see 
Section 4, Volume I of the NRDAM/CME and NRDAM/GLE technical documents. 
For further information on the biological database, see Section 6, 
Volume I, and Volume IV of the NRDAM/CME technical document; and 
Section 8, Volume I, and Section 3, Volume III of the NRDAM/GLE 
technical document.
3. Restoration Submodel
    The restoration submodel estimates the cost, if any, of restoring 
the injured resources. The submodel first evaluates possible habitat 
restoration and restocking actions. The submodel analyzes the costs and 
benefits of any possible habitat restoration and restocking actions to 
determine whether these forms of active restoration or natural recovery 
should be assumed for purposes of the models' damage calculations. In 
some cases, the submodel also determines the cost of restoring lost 
assimilative capacity. The active restoration costs, if any, computed 
by the restoration submodel comprise one component of the damage 
figure; the other component, compensable value, is calculated by the 
compensable value submodel.
    For certain types of habitats, the restoration submodel evaluates 
habitat restoration action. The submodel identifies those habitats for 
which human intervention may potentially facilitate recovery. For each 
such habitat in each affected area, the restoration submodel evaluates 
the effect that a

[[Page 20566]]

particular active restoration alternative would have on the compensable 
value calculated by the model. If the relevant active habitat 
restoration alternative would result in a lower total compensable value 
for a particular grid cell than reliance upon natural recovery, then 
the restoration submodel computes the cost of performing that 
alternative for that grid cell. The restoration cost database supplies 
information on unit restoration costs to the restoration submodel. The 
biological effects and compensable value submodels supply information 
to the restoration submodel concerning the extent of injury and 
compensable value with and without active habitat restoration. If the 
active habitat restoration alternative would not result in a lower 
total compensable value than reliance upon natural recovery, then the 
restoration submodel does not compute any habitat restoration costs.
    The restoration submodel evaluates the following types of active 
habitat restoration alternatives against natural recovery:

    For open water sediments: dredging and refilling with clean 
material (shallow water); or capping (deep water);
    For wetlands, macroalgal beds, and seagrass beds: replacement of 
contaminated substrate and replanting (if sediments are toxic); or 
replanting (if sediments are not toxic but mortality has occurred);
    For invertebrate reefs (coral and mollusk): replacement of 
contaminated substrate and reseeding (if sediments are toxic); or 
reseeding (if sediments are not toxic but mortality has occurred); 
and
    For shorelines in coastal or marine environments: washing of 
sand and gravel; replacement of mud; and chemical washing of rocky 
shoreline.

    The restoration submodel then considers restocking of fish and 
wildlife. If stocks of the same age as the injured fish and wildlife 
are available through captive breeding programs, then the submodel 
computes the cost of restocking those species after the habitat has 
recovered, either through natural recovery or active habitat 
restoration. The restoration cost submodel supplies data on the 
availability and cost of stocks to the restoration submodel.
    If the relevant active habitat restoration alternative would reduce 
compensable value or if restocking is possible, then the submodel 
performs a cost-benefit test of these forms of active restoration. The 
submodel compares the total costs of active habitat restoration and 
restocking against the measured benefits of such restoration (i.e., 
compensable value assuming natural recovery minus compensable value 
assuming active habitat restoration and restocking). If the costs 
exceed ten times the measured benefits, then the submodel assumes, for 
purposes of generating a damage figure, that natural recovery, rather 
than active restoration, will be used to reestablish baseline 
conditions. If the costs do not exceed the measured benefits by ten 
times, then the submodel assumes, for purposes of generating a damage 
figure, that habitat restoration and restocking actions will be 
implemented.
    Finally, for releases that generate a damage figure related to 
mortality and loss of productivity, the restoration submodel also 
calculates the cost of restoring the water's baseline ability to absorb 
pollutants (assimilative capacity). In the case of such releases, the 
restoration submodel determines the amount of the released substance 
that would remain in the environment after environmental exposure 
levels are below acute toxicity thresholds and after any habitat 
restoration actions are completed. The submodel then computes the cost 
of removing a contaminant mass with toxicity equivalent to the 
remaining non-acutely toxic dispersed mass of the released substance 
from other identified contaminated sites. When determining the amount 
of contaminant mass to remove, the submodel adjusts for the relative 
degradability of that contaminant compared to that of the spilled 
substance. The restoration cost database supplies data on unit costs to 
the restoration submodel.
    The restoration submodel sums the costs of any selected types of 
active restoration. The models combine this figure with the compensable 
value figure computed by the compensable value submodel to form the 
final damage figure.
    For further information on the restoration submodel, see Section 5, 
Volume I of the NRDAM/CME and NRDAM/GLE technical documents. For 
further information on the restoration cost database, see Sections 5, 
12, and 13, Volume I, and Sections 5 through 7, Volume V of the NRDAM/
CME technical document; and Section 9, Volume I and Section 5, Volume 
III of the NRDAM/GLE technical document.
4. Compensable Value Submodel
    Compensable value, as computed by the compensable value submodel, 
is the sum of certain economic use values lost to the public pending 
the reestablishment of baseline conditions through either natural 
recovery or active restoration, as determined by the restoration 
submodel. Only public losses are included in compensable value.
    The submodel computes the following types of compensable values:

    Lost economic rent for lost commercial harvests resulting from 
any closures specified by the authorized official and/or from 
population losses;
    Lost recreational harvests resulting from any closures specified 
by the trustee and/or from population losses;
    In the NRDAM/CME, lost wildlife viewing, resulting from 
population losses, by residents of the States bordering the 
provinces in which the population losses occurred;
    In the NRDAM/GLE, lost wildlife viewing, resulting from 
population losses, by residents of local areas bordering the 
provinces in which the population losses occurred; Lost beach 
visitation due to closure; and
    In the NRDAM/GLE, lost boating due to closure.

    The submodel calculates compensable value for lost economic rent by 
multiplying the total lost harvest of the species, as computed by the 
biological effects submodel, by the commercial price per unit of 
harvest, as supplied by the compensable value database. The rule 
provides that if a trustee is aware of reliable evidence that a private 
party has recovered damages for commercial harvests lost as a result of 
the release, the trustee must eliminate from the claim any damages for 
such lost harvests included in the lost economic rent calculated by the 
model.
    The submodel calculates compensable value for lost recreational 
harvests by multiplying the total lost recreational harvest of the 
species, as computed by the biological effects submodel, by the 
marginal value of harvesting an additional animal, as supplied by the 
compensable value database. The submodel computes damages only for 
harvests lost due to populations losses or closures. The submodel does 
not compute damages for lost quality of recreational fishing unrelated 
to lost harvests or for lost trips due to de facto closures.
    The compensable value submodel computes compensable value for a 
specific range of lost wildlife viewing. First, the submodel only 
calculates wildlife viewing damages resulting from population losses 
and does not address damages resulting from closures. Second, the 
submodel only calculates losses incurred by certain segments of the 
wildlife viewing public. The models divide geographic areas into 
provinces. The NRDAM/CME computes lost wildlife viewing only for 
residents of States bordering the provinces in which the population 
loss occurred. The NRDAM/GLE computes lost wildlife viewing only for 
residents of local areas bordering the provinces in which the 
population loss occurred. The submodel calculates damages by 
multiplying the

[[Page 20567]]

number of viewing trips affected by the release by the per-animal 
marginal viewing value for the animals killed.
    The compensable value submodel computes compensable value for lost 
beach visitation only if trustees specify that there has been a closure 
of a beach. The submodel does not calculate damages for lost quality of 
beach visitation or for lost beach visitation due to de facto closures. 
If a closure is specified, the compensable value submodel calculates 
compensable value by multiplying the length of beach closed per day and 
the number of days closed, as supplied by trustees, by the per-day 
value of trips to the closed length. The compensable value database 
supplies data on the per-unit value of lost beach visitation.
    The NRDAM/GLE computes compensable value for lost boating only if 
trustees specify that there has been a closure of a boating area. The 
model does not calculate damages for lost quality of boating or for 
lost boating trips due to de facto closures. If a closure is specified, 
the compensable value submodel calculates compensable value by 
multiplying the geographic area closed per day and the number of days 
closed, as supplied by trustees, by the per-day value of trips to the 
closed area. The compensable value database supplies data on the per-
unit value of lost boating. The NRDAM/CME does not compute compensable 
value for lost boating.
    The per-unit values in the compensable value database are stated in 
1991 dollars for the NRDAM/CME and 1990 dollars for the NRDAM/GLE. The 
compensable value submodel uses the Gross National Product Implicit 
Price Deflator, as supplied by trustees, to adjust per-unit values to 
current dollars. The compensable value submodel discounts the value of 
future losses using a three percent discount rate.
    After applying the Gross National Product Implicit Price Deflator 
and the discount rate, the compensable value submodel sums the lost 
values to calculate a compensable value figure. This figure is added to 
the restoration costs, if any, computed by the restoration submodel to 
form the final damage figure calculated by the models.
    The rule provides that if a trustee is aware of reliable evidence 
that the model application covers resources beyond his or her 
jurisdiction, the trustee must either: (1) Have the other trustees who 
do have jurisdiction over those resources join in the type A 
assessment; or (2) eliminate any damages for those resources from the 
claim. Further, the rule provides that if the model output indicates 
damages in excess of $100,000, then trustees who wish to obtain a 
rebuttable presumption must either: (1) Limit the portion of their 
claim calculated with the type A procedure to $100,000; or (2) compute 
all damages using type B procedures.
    For further information on the compensable value submodel, see 
Sections 8 through 11, Volume I of the NRDAM/CME technical document; 
and Section 6, Volume I of the NRDAM/GLE technical document. For 
further information on the compensable value database, see Sections 8 
through 11, Volume I, and Sections 1 through 4, Volume V of the NRDAM/
CME technical document; and Section 6, Volume I, and Section 4, Volume 
III of the NRDAM/GLE technical document.

V. Use of the NRDAM/CME and NRDAM/GLE in Other Contexts

    The Department is issuing today's final rule in compliance with the 
statutory requirement to develop procedures for conducting simplified 
assessments that are entitled to a rebuttable presumption. The 
standards in today's final rule apply only when trustees use the type A 
models to develop a damage figure and intend to obtain a rebuttable 
presumption for that figure in litigation. Trustees who use the models 
in other contexts, such as settlement negotiations or litigation 
without the benefit of the rebuttable presumption, are not subject to 
the rule standards. In these other contexts, trustees are free to make 
modifications to the model databases beyond those permitted under the 
rule and to use some, but not all, of the components of the models.
    For example, trustees may wish to use the models to develop a 
benchmark damage figure for settlement negotiations but may have more 
up-to-date or more site-specific information on recreational fishing 
values. In that case, trustees may choose to apply the models using 
modified recreational fishing values, notwithstanding the rule 
provisions concerning modification of the model databases. In other 
situations, trustees may choose to rely on the models' predictions of 
injury but perform their own analyses of restoration alternatives and 
compensable values. Trustees may also choose to rely on the models' 
damage calculations for some resources but for other resources 
substitute their own damage calculations for other resources covered by 
the models. The Department believes that although use of the type A 
models in these ways would not be covered by today's rule and, 
therefore, would not be entitled to a rebuttable presumption, such use 
can produce reliable damage estimates if done properly.

VI. Summary of Major Changes from the Proposed Rules

    The Department has made numerous changes in the rule language and 
models based on the comments received. The Department discusses its 
rationale for these changes in Section VII of this preamble.

A. Rule Language

    The Department has made several major substantive changes to the 
proposed rule language. With regard to the applicability of the type A 
procedures, the Department has modified the conditions that must be met 
before a trustee can use a type A procedure to obtain a rebuttable 
presumption and has eliminated the provision that would have required 
trustees to use the type A procedures in some circumstances. Instead of 
delineating ``primary'' and ``secondary'' conditions for use as the 
proposed rule did, the final rule now provides that if the conditions 
for use of the models listed in Sec. 11.34 are met, then trustees 
decide whether to use type A or type B procedures based on an 
evaluation of the averaged data and simplifying assumptions listed in 
the NRDAM/CME and NRDAM/GLE technical documents. The Department has 
also more clearly delineated the conditions under which trustees can 
use type B procedures to supplement a type A procedure and the process 
for doing so.
    With regard to operation of the NRDAM/CME and the NRDAM/GLE, the 
rule now allows trustees to modify the habitat designations in the 
models and still obtain a rebuttable presumption. The rule also 
requires trustees to perform a preliminary application of the NRDAM/CME 
or NRDAM/GLE and make the results available for public review before 
performing a final application and presenting a demand to the PRP.
    The final rule contains three new provisions that require trustees 
in some cases to adjust the damage figure calculated by the models 
before presenting a demand. First, the rule now provides that if 
trustees are presented with evidence that private parties have obtained 
recoveries for lost commercial harvests, they must eliminate any 
damages for such lost harvests included in the lost economic rent 
calculated by the model. Second, the rule provides that if a trustee is 
aware of reliable evidence that the model application covers resources 
beyond his or her trustee jurisdiction, the trustee must

[[Page 20568]]

either: (1) Have the other trustees who do have jurisdiction over those 
resources join in the type A assessment; or (2) eliminate any damages 
for those resources from the type A damage calculation. Third, the rule 
provides that if the model output indicates damages in excess of 
$100,000, then trustees who wish to obtain a rebuttable presumption 
must either: (1) Limit the portion of their claim calculated with the 
type A procedure to $100,000; or (2) compute all damages using type B 
procedures.
    The Department has also eliminated the proposed clarification of 
the scope of review of a type A assessment in a natural resource damage 
case.
    Finally, as part of its regulatory reform efforts, the Department 
has rewritten the final rule in plain English. The Department believes 
this revision has made the rule significantly clearer and easier to 
read.
    The following is a section-by-section analysis of the final rule:

Subpart A--Introduction

Section 11.15  What Damages May a Trustee Recover?

    The Department has rewritten the heading of this section to make it 
easier to understand. The final rule language revising subsection 
(a)(1) is unchanged from the August 1994 proposed rule. The final rule 
eliminates the separate subsections referring to type A procedures, 
type B procedures, or a combination of type A and B procedures in the 
same assessment. Sections 11.34 through 11.36 include the criteria and 
standards for selecting type A procedures, type B procedures, or a 
combination, making additional detail in this introductory section 
unnecessary.

Section 11.18  Incorporation by Reference

    The final rule slightly revises and updates the proposed rule 
language incorporating by reference the NRDAM/CME technical document, 
and adding language incorporating by reference the NRDAM/GLE technical 
document.

Section 11.19  Information Collection

    The final rule retains the December 1994 proposed rule language to 
remove and reserve this section.

Subpart C--Assessment Plan Phase

Section 11.30  What Does the Authorized Official do if an Assessment is 
Warranted?

    The final rule makes several revisions to this section that were 
not included in the proposed rules, but which are necessary to conform 
to other provisions in today's final rule. Existing subsection (a), 
which applied to both type A and type B procedures, did not authorize 
performance of any assessment methodologies until after the period of 
public review and comment for the Assessment Plan. Section 11.42 of 
today's final rule requires trustees to perform a preliminary 
application of the NRDAM/CME or NRDAM/GLE before releasing the 
Assessment Plan for public review and comment. Trustees who use type B 
procedures, however, must still make the Assessment Plan available for 
public review and comment before performing any of the procedures 
contained in the Plan. See Sec. 11.32(c) of today's final rule. The 
Department has revised the heading of the section and the language of 
subsection (a) to make them easier to understand and to make this 
conforming change. The Department has also modified subsection 
(c)(1)(vi) to make a necessary conforming change cross-referencing 
other rule provisions.

Section 11.31  What Does the Assessment Plan Include?

    The final rule revises the heading and rule language to make the 
section easier to understand. Subsection (a)(1) adopts as final the 
language in the August 1994 proposed rule.
    The Department has revised subsection (b) from the August 1994 
proposed rule to make it clear that the Assessment Plan must include a 
detailed explanation of how the trustee's decision to use a type A 
procedure, type B procedures, or a combination, satisfies the 
decisional standards contained in the rule.
    Subsection (c) clarifies and corrects existing rule language, which 
was garbled in 1988. Compare 53 FR 5174 (Feb. 22, 1988) with 51 FR at 
27731. Although this language was not in the proposed rules, it is a 
nonsubstantive change. Subsection (c)(1) has been modified to make a 
necessary conforming change cross-referencing redesignated Sec. 11.37.
    Subsection (d) revises the existing rule language to make it easier 
to understand. Subpart D contains the requirements concerning 
identification and documentation of information, and therefore it is 
unnecessary to repeat them in subsection (d).

Section 11.32  How Does the Authorized Official Develop the Assessment 
Plan?

    The Department has revised the heading of this section to make it 
easier to understand.
    The final rule revises subsection (c) to make it easier to 
understand and to make the same necessary conforming change described 
in the discussion of Sec. 11.30.
    The final rule language revising subsection (f) is slightly 
reworded, but substantively the same as, the language in the August 
1994 proposed rule. As explained in the August 1994 notice of proposed 
rulemaking, this provision clarifies that the confirmation of exposure 
requirement applies to type B, but not type A, procedures. Original 
Secs. 11.34(a)(1), 11.31(c)(1), and 11.33(b)(4) already established 
this distinction. Today's final rule language merely makes the rule 
easier to understand.

Section 11.33  What Types of Assessment Procedures Are Available?

    Today's final rule revises Sec. 11.33 to limit this section to 
providing a brief description of the difference between type A and type 
B procedures.

Section 11.34  When May the Authorized Official Use a Type A Procedure?

    New Sec. 11.34 combines and revises changes that were proposed for 
Sec. 11.33 in the August 1994 and December 1994 proposed rules. This 
section now states the threshold conditions that must be present before 
a trustee may use a type A procedure, many of which were included among 
the ``primary'' conditions in the proposed rules.

Section 11.35  How Does the Authorized Official Decide Whether to Use 
Type A or Type B Procedures?

    New Sec. 11.35 further revises changes that were proposed for 
Sec. 11.33. The section provides decisional criteria for the 
determination whether to use type A or type B procedures, assuming that 
the conditions in Sec. 11.34 are met. The final rule language requires 
trustees to base the decision whether to use type A or type B 
procedures on an evaluation of the data and assumptions in the type A 
procedures, as described in the NRDAM/CME and NRDAM/GLE technical 
documents. These assumptions include many of the ``secondary 
conditions'' contained in the proposed rules.

Section 11.36  May the Authorized Official Use Both Type A and Type B 
Procedures for the Same Release?

    New Sec. 11.36 provides standards for when trustees may use both a 
type A procedures and type B procedures for the same release. The 
August and December 1994 proposed revisions to Sec. 11.33 included 
similar modifications.

[[Page 20569]]

Today's final rule language provides clearer, more specific criteria, 
and specifically identifies the categories of injury and compensable 
value addressed by the type A procedures.
    Subsection (d) addresses the issue of which type B procedures must 
be followed when a trustee decides to combine a type A and type B 
procedures in a single assessment.

Section 11.37  Must the Authorized Official Confirm Exposure Before 
Implementing the Assessment Plan?

    The Department has revised the heading of this section (formerly 
Sec. 11.34) and has modified subsection (a) from the proposed rule to 
make it easier to read. Subsection (a) clarifies the intent of the 
existing rule that the confirmation of exposure requirement applies 
only to type B procedures. Although former Sec. 11.34(a) did not 
expressly distinguish between type B and type A procedures, former 
Secs. 11.31(c)(1) and 11.33(b)(4) limited the confirmation of exposure 
requirement to type B procedures.

Subpart D--Type A Procedures

Section 11.40  What Are Type A Procedures?

    The Department has revised the heading of this section and the 
language of subsection (a) to make them easier to read, to add 
references to the type A procedures for Great Lakes environments, to 
provide additional information about both type A procedures, and to 
incorporate the requirement that a trustee must follow the procedures 
in Secs. 11.41 through 11.44 when using either of the two type A 
procedures. Today's final rule provides a more detailed description of 
type A procedures than was contained in the August 1994 proposed 
revision to Sec. 11.40.

Section 11.41  What Data Must the Authorized Official Supply?

    This section identifies the data inputs and modifications that the 
trustee must supply to use the NRDAM/CME or NRDAM/GLE. Today's final 
rule modifies and simplifies proposed Sec. 11.42 (c) and (d) in the 
August 1994 proposed rule, and proposed revisions to Sec. 11.41 in the 
December 1994 proposed rule. The final rule language for Sec. 11.41 is 
considerably shorter than that in the proposed rules, because the 
format for data inputs and modifications is now contained in two new 
appendices to the rule. The final rule now requires trustees to make 
certain modifications to the model databases, including the habitat 
designations, if they have reliable evidence that the databases are 
incorrect.

Section 11.42  How Does the Authorized Official Apply the NRDAM/CME and 
NRDAM/GLE?

    This section contains a new procedure requiring trustees to perform 
a preliminary application of the NRDAM/CME or NRDAM/GLE as part of the 
process for deciding whether to use a type A procedure. If the trustee 
decides to continue with a type A procedure, then the data inputs, 
modifications, and results of the preliminary application become part 
of the Assessment Plan.

Section 11.43  Can Interested Parties Review the Results of the 
Preliminary Application?

    This section requires trustees who decide to continue with a type A 
procedure to develop an Assessment Plan, which must include the data 
inputs, modifications, and results of the preliminary application. The 
trustee must make the Assessment Plan available for public review and 
comment.

Section 11.44  What Does the Authorized Official do After the Close of 
the Comment Period?

    Subsections (a) through (c) of this section state the procedural 
and substantive requirements following public comment on the Assessment 
Plan, which include performing a final application of the NRDAM/CME or 
NRDAM/GLE and preparing a Report of Assessment. Subsection (d) includes 
specific criteria to preclude double recovery for economic rent for 
lost commercial harvests if a private party has already recovered for 
the same damages. Subsection (e) resolves a potential problem arising 
when trustees have not agreed in advance to use a type A procedure 
jointly. Subsection (f) limits the damages that may be recovered by 
trustees who use the NRDAM/CME or NRDAM/GLE and intend to obtain a 
rebuttable presumption.

Subpart E--Type B Procedures

Section 11.73  Quantification Phase-Resource Recoverability Analysis

    The Department has revised subsection (a) to make a necessary 
conforming change to cross-reference redesignated Sec. 11.35 (now 
Sec. 11.38).

Subpart F--Post-Assessment Phase

Section 11.90  What Documentation Must the Authorized Official Prepare 
After Completing the Assessment?

    The Department has revised the final rule from the August 1994 
proposed rule to make the heading and rule language simpler and easier 
to understand. The substantive effect of this provision is the same as 
existing Sec. 11.90.

Section 11.91  How Does the Authorized Official Seek Recovery of the 
Assessed Damages From the Potentially Responsible Party?

    Today's final rule revises the heading of the section and the first 
sentence of subsection (a) to make the rule language simpler and easier 
to understand. The substantive effect of this provision is the same as 
existing Sec. 11.91.

Appendices

    The Department has added two new appendices to the rule. These 
appendices specify the format for data inputs and modifications for the 
NRDAM/CME and NRDAM/GLE.
B. NRDAM/CME and NRDAM/GLE
    The Department has made several major substantive changes to the 
NRDAM/CME and NRDAM/GLE computer code and databases. The Department has 
revised the chemical database for both the NRDAM/CME and NRDAM/GLE to 
incorporate an additional 24 oils and petroleum products. The 
Environment Canada publication, ``A Catalogue of Crude Oil and Oil 
Product Properties,'' and NOAA's ADIOS (Automated Data Inquiry for Oil 
Spills) database provided the principal sources of information for 
revision of the databases. The Department also deleted the following 
hazardous substances from the database: pure metals, nontoxic 
substances, and substances for which the toxicity threshold was less 
than the water solubility. The Department deleted a total of 31 
hazardous substances from the NRDAM/CME database and 32 hazardous 
substances from the NRDAM/GLE database.
    The Department has included an additive toxicity model for oil and 
petroleum products in the biological effects submodel to address the 
additive toxicity of the multiple substances in oil and petroleum 
products. The additive toxicity model also addresses the effects of oil 
weathering.
    The Department has updated the wildlife viewing values contained in 
both the NRDAM/CME and NRDAM/GLE economic databases based on

[[Page 20570]]

recent information available from the 1994 addendum to the 1991 
National Survey of Fishing, Hunting, and Wildlife-Associated Recreation 
developed by the U.S. Fish and Wildlife Service (FWS). Also, in the 
NRDAM/CME, the Department revised the wildlife viewing values to 
reflect the total population of the respective coastal states.
    In the NRDAM/CME, the Department has modified the habitat grids to 
provide a finer scale resolution. The Department changed the scale from 
a 50 x 50 grid to a 100 x 100 grid. The Department has also upgraded 
the Microsoft compiler to allow for use of 32-bit processing 
and additional random access memory (RAM).
    The Department has revised the east coast wetland habitats 
represented in the NRDAM/CME grids for provinces 11, 12, and 13 (New 
York and New Jersey) to incorporate more site-specific data provided by 
commenters. See Section 3.4, Volume III of the NRDAM/CME technical 
document.
    In the NRDAM/CME, the Department has substantially revised wildlife 
abundance data for provinces 40 through 51 (west coast and the Gulf of 
Alaska) based on additional information and data provided by public 
commenters.
    The Department has added a habitat editor to the NRDAM/GLE user 
interface consistent with that provided in the proposed NRDAM/CME.
    The Department has included intertidal seagrass as an additional 
habitat type in the NRDAM/CME. The intertidal seagrass habitat includes 
those common habitats for tropical seagrass and eelgrass.
    The Department has disaggregated the model output files for the 
injury and damage calculations resulting from direct kills versus food 
web and habitat losses, and from commercial versus recreational fishing 
losses.
    The Department has revised the active habitat restoration 
alternatives evaluated for structured habitats (i.e., wetlands, 
seagrass beds, macroalgal beds, and invertebrate reefs) to include not 
only sediment replacement with replanting but also replanting alone.
    The Department has eliminated the calculation of compensable value 
for lost boating and subsistence losses from the NRDAM/CME.
    The Department has revised the restoration submodel to include a 
cost-benefit test for determining whether the measured benefits of 
active habitat restoration and restocking, as compared to natural 
recovery, are worth the additional costs.
    Finally, the Department has revised the calculation of assimilative 
capacity restoration costs to correct for the degradation rate of the 
spilled substance and to limit the calculation of assimilative capacity 
restoration costs to cases where biological injury has occurred and 
produces compensable value.

VII. Response to Comments

    The Department received numerous public comments on the proposed 
type A procedures. The Department and NOAA also asked several 
independent technical reviewers to examine the proposed NRDAM/CME. The 
Department made the comments of these independent technical reviewers 
available to the public and included them in the administrative record 
for this rulemaking. See 60 FR 28773 (June 2, 1995). The Department 
provides responses to both the public comments and the comments of the 
independent technical reviewers below.
    In addition to the issues discussed below, commenters addressed a 
number of issues beyond the scope of this rulemaking. The Department 
explicitly limited this rulemaking to four issues: the revision of the 
existing type A procedure for coastal and marine environments; the 
development of a new type A procedure for Great Lakes environments; the 
conditions for combined use of type A and type B procedures; and the 
scope of judicial review of assessments performed using type A 
procedures. See 59 FR at 40319-20, 63300, and 63302. Nevertheless, some 
commenters raised additional issues, including: whether trustees should 
be allowed to pool natural resource damage recoveries to implement 
regional restoration plans; the permissibility of using type A and type 
B procedures for the same release; and whether lost economic rent and 
the cost of restoring lost assimilative capacity are legally 
permissible categories of damages. The Department has not evaluated, 
and is not providing substantive responses to, comments on these issues 
in this rulemaking.
    Section 11.93(d) of the existing regulations, which was promulgated 
in 1987, allows pooling of multiple type A recoveries to implement a 
single restoration plan, so long as the plan is intended to address the 
same or similar injuries as those identified in each application of the 
type A procedure. See 52 FR at 9100. The Department neither reproposed, 
revisited, nor solicited comment on Sec. 11.93(d) and merely cited it 
in the preambles to the proposed rules by way of background. 59 FR at 
40324 and 63305.
    Section 11.15(a)(1)(iii) of the original type A rule, which was 
promulgated in 1987, established that trustees could use both type A 
and type B procedures for the same release under certain circumstances. 
See 52 FR at 9095. The Department did not repropose, revisit, or 
solicit comment on whether CERCLA allows trustees to combine type A and 
type B procedures. The only issue raised and addressed in this 
rulemaking was whether the Department should expand the authorization 
for combined use of type A and type B procedures.
    Finally, the Department did not repropose, revisit, or solicit 
comment on its long-standing positions on the recoverability of damages 
for lost economic rent and lost assimilative capacity. Both the 
original type B rule and the original type A rule explicitly allowed 
for the recovery of lost economic rent. See 43 CFR 11.83(c)(1); 51 FR 
at 27749; and 52 FR at 9047. The Department has recognized the loss of 
assimilative capacity as a legitimate category of natural resource 
damages since the promulgation of the original type B procedures in 
1986. 51 FR at 27716; see also 59 FR at 14273. The Department has begun 
a biennial review of the type B procedures and will be considering the 
issues of lost economic rent and lost assimilative capacity in that 
context. See 59 FR 62749 (Oct. 19, 1994).

A. General Comments

    Comment: Some commenters supported the concept of a reliable, 
accurate, automated damage assessment procedure that would eliminate 
the need for expensive tailor-made studies. However, other commenters 
objected to the calculation of damages through what they considered to 
be abstract application of theoretical, generic models. Some of these 
commenters thought that many of the calculations of the NRDAM/CME and 
NRDAM/GLE were based on unsubstantiated assumptions.
    A number of commenters, including some of the independent technical 
reviewers, questioned the Department's use of ``grand averages'' to 
extrapolate data for a specific species, substance, or location, to 
different species, substances, and locations. Commenters were 
particularly concerned about the extrapolation of economic values made 
in the compensable value submodel. For example, commenters noted that 
some of the studies used to value recreational fishing in the NRDAM/CME 
were based on freshwater fishing and commercial fishing. Commenters 
also stated that many of the studies used outdated data and outdated or 
unreliable

[[Page 20571]]

methodologies. For example, commenters noted that recreational hunting 
values were derived from a 20-year old contingent valuation study. Some 
commenters suggested specific criteria that they thought should be met 
when performing benefits transfer (i.e., the extrapolation of economic 
values derived from studies of one situation to another situation).
    Response: CERCLA requires that type A procedures involve ``minimal 
field observation'' and authorizes type A procedures to be based on 
``units of discharge or units of affected area.'' CERCLA sec. 
301(c)(2)(A). The Senate Report that accompanied the predecessor bill 
to CERCLA provides the following indication of Congress' intent:

    Natural resource damage assessments based on this type of 
regulation [type A] should require as little fieldwork as possible, 
and rely on a combination of habitat values, tables of values for 
individual species, and previously conducted surveys and laboratory 
studies, related to units of discharge or units of affected area. S. 
Rep. No. 96-848 at 86.

    This language indicates that Congress envisioned the development of 
type A procedures that do not require the performance of any new 
studies but instead use existing studies to provide generalized values 
that can be applied in specific cases. Inherent in the concept of 
developing unit values from existing studies is the notion of making 
assumptions in the absence of empirical data and applying average 
values across a range of nonidentical items. Therefore, the Department 
believes that CERCLA authorizes it to make appropriate extrapolations 
from existing data.
    The science of natural resource damage assessment is still 
evolving. The universe of relevant studies is still very small for many 
crucial aspects of damage assessment. Existing data are particularly 
limited as to the effects of small spills. Even when addressing the 
limited range of scenarios covered by the NRDAM/CME and the NRDAM/GLE, 
the Department faced significant challenges in bridging data gaps. 
Although Congress did authorize the Department to make extrapolations 
from existing data, the Department recognizes that any such 
extrapolations must be reasonable. Thus, when developing the models, 
the Department tried to make use of the most reliable information 
available based on extensive reviews of published and unpublished 
information and data; make only those assumptions that are necessary; 
ensure that any assumptions that are made are reasonable; and identify 
clearly all assumptions that were required for the development of 
simplified procedures.
    With regard to the compensable value submodel, the Department did 
apply specific criteria during its selection of studies to use for 
benefits transfer. The Department used only studies that: (1) Were 
based on an extensive literature review and consultations with relevant 
governmental agencies; (2) reasonably represented the natural resource 
and public use under investigation; (3) contributed to a reasonable 
representation of the different regions included in the models; (4) 
were conducted by a recognized university-associated researcher or 
established consulting firm; and (5) used appropriate valuation 
methodologies. The Department believes that these criteria adequately 
address all the concerns that the commenters' suggested criteria are 
intended to address. The first three criteria assure that the resources 
considered in the selected studies are as similar as possible to the 
resources to be valued in the models. The fourth criterion assures that 
the selected studies are scientifically sound. The fifth criterion 
assures that the selected studies use appropriate valuation 
methodologies.
    Comment: One commenter suggested that the Department had developed 
the models by selecting values from a few studies while ignoring 
others. The commenter argued that the Department had failed to provide 
adequate justification for the values it selected.
    Response: The Department conducted extensive searches for available 
information. Some data the Department identified were not used because 
better or more applicable data were available. However, none of the 
identified data was ignored. The Department believes that the NRDAM/CME 
and NRDAM/GLE technical documents adequately explain and justify the 
values in the models.
    Comment: Some commenters thought that the proposed type A models 
were so technically flawed that they did not meet the statutory 
standard of ``best available procedures'' and, therefore, trustees 
should not obtain a rebuttable presumption if they use the models. 
These commenters urged the Department to abandon the models noting that 
Colorado v. Interior does not require or authorize the Department to 
issue a model that is unreliable. One commenter acknowledged that the 
proposed revised NRDAM/CME appeared to be an improvement over the 
original NRDAM/CME Version 1.2 issued in 1987. However, the commenter 
thought the proposed revised model still contained too many flaws to 
accomplish its intended purpose. Another commenter stated that the 
damage figures produced by the models are nothing more than sheer 
speculation and are not legally sufficient due to the compounding of 
errors, uncertainties, biases, and overestimates.
    Response: As discussed in more detail below, the Department has 
carefully reviewed all comments it received on the proposed models and 
rule language. Based on this review, the Department has made numerous 
modifications to the models and the rule language. Where the Department 
concluded that no changes were needed, the Department has explained its 
reasoning. The Department believes that the final type A models, as 
revised in response to comments, are best available procedures when 
used in accordance with the standards and process set forth in today's 
final rule. The models, with their state-of-the-art modeling and 
extensive databases, represent a significant advancement beyond the 
original NRDAM/CME issued in 1987. The final type A procedures provide 
for reliable, cost-effective, simplified assessments that are entitled 
to a rebuttable presumption.
    Comment: Several commenters thought the Department had been overly 
ambitious in attempting to develop models like the NRDAM/GLE and the 
NRDAM/CME. Specifically, these commenters stated that the biological 
effects submodel attempted to perform a task that is beyond the current 
state of ecological modeling. The commenters contended that state-of-
the-art ecological modeling is not yet capable of producing accurate 
quantitative determinations and is primarily useful only for making 
qualitative predictions. The commenters also thought that the multiple 
iterative calculations performed by the biological effects submodel did 
not alleviate the problem but simply amounted to averaging of nonsense.
    Response: The Department agrees that ecological models should 
generally be used only for qualitative predictions. However, the 
biological effects submodel in the NRDAM/GLE and the NRDAM/CME is not a 
true ecological model in the sense suggested by commenters. Ecological 
models evaluate the changes in ecosystem structure and function 
resulting from disturbances. The biological effects submodel, on the 
other hand, is a toxicological effects model. The biological effects 
submodel simply calculates acute mortality and lost production and 
projects these injuries forward as biota not present or used in future 
years. The submodel need not, and does not attempt to, address the 
higher-order ecological

[[Page 20572]]

changes in the structure and functions of biological systems as true 
ecological models do.
    The Department believes that the NRDAM/CME and NRDAM/GLE are 
reasonable tools for assessing the injuries and compensable values that 
they address and do not generate ``nonsense.'' Further, the use of 
iterative calculations is designed to, and does, enhance the 
reliability of damage estimates in particular cases. The biological 
effects submodel uses several randomized algorithms for processes, such 
as swimming by fish, that are considered random at the relevant spatial 
and temporal scales. For each spill modeled, the submodel performs 
multiple iterative runs and then selects the mean result. This approach 
is a generally accepted method of modeling the most probable biological 
effects for events that have an element of randomness.
    Comment: Some commenters thought the proposed models were 
fundamentally flawed because they used overly simplistic simulations of 
movement of biota within a population. The commenters stated that these 
simulations could not be improved because of the lack of basic data on 
population movement.
    Response: The Department believes that the NRDAM/CME and the NRDAM/
GLE use the best available procedure for simulating the movement of 
biota and that this procedure is reliable for the purposes of a 
simplified damage assessment. The Department acknowledges that the 
directed movement of biota is not well understood quantitatively. 
However, at the smallest scale, there is a random component to the 
movements of animals within the habitats they occupy, and the NRDAM/CME 
and NRDAM/GLE can and do model this component. The models do not 
simulate within-season, between-habitat movements, except where 
currents carry organisms across boundaries. However, the seasonal and 
habitat-specific abundances included in the database do account for 
inter-habitat movement between seasons.
    Comment: Some commenters, including some of the independent 
technical reviewers, thought that the Department should validate the 
models against real-world data and perform sensitivity analyses. A few 
commenters also thought the Department should calibrate the models.
    Response: The Department has conducted extensive sensitivity 
studies of both the NRDAM/CME and NRDAM/GLE. It is difficult to conduct 
conclusive validation studies of the models due to the extreme lack of 
data on the natural resource effects of small spills. In fact, although 
more data exist for large spills, even those data are limited. 
Nonetheless, the Department has used the data that are available to 
conduct validation studies of the NRDAM/CME physical fates and 
biological effects submodels and believes that these studies suggest 
that the submodels provide reasonable estimates of the actual physical 
fates and biological effects of spills. Even less data exist for spills 
in the Great Lakes than for spills in coastal and marine environments. 
However, since the NRDAM/GLE contains the same algorithms as the NRDAM/
CME, the Department believes the results of the validation studies of 
the NRDAM/CME also support the NRDAM/GLE.
    Because of the cost involved in performing site-specific type B 
studies, trustees have rarely pursued damage claims for minor releases. 
Therefore, virtually no data exist with which to validate the 
restoration and compensable value submodels or determine the need for 
calibrating the damage estimates produced by the models. In the absence 
of such data, the Department has relied primarily on careful reviews of 
the accuracy and reasonableness of the data and algorithms used in the 
models. The Department believes that these reviews of the scientific 
underpinnings of the models provide adequate support for the 
reliability of the damage estimates produced by the models.
    The Department further believes that the models are consistent with 
congressional intent underlying the directive to produce procedures for 
simplified assessments. The models are best available simplified 
procedures. They produce reliable, fair, and reasonable results when 
used for their intended purpose. The Department has clearly identified 
the capabilities and limitations of the models and has allowed trustees 
to select between type A and type B procedures based on specified 
criteria. Finally, the Department has retained the provision allowing 
PRPs to require trustees to use type B procedures if they advance the 
reasonable cost of using such procedures within an acceptable time 
frame.
    Comment: One commenter stated that the NRDAM/GLE should be peer 
reviewed in an open forum prior to promulgation.
    Response: The Department believes that the NRDAM/GLE has been 
adequately reviewed. The proposed model was made available for public 
review and comment for eleven months. Also, the review of the proposed 
NRDAM/CME by independent technical reviewers was directly relevant for 
the NRDAM/GLE because the NRDAM/GLE incorporates the same basic 
modeling as the NRDAM/CME.
    Comment: Some of the independent technical reviewers claimed that 
the proposed NRDAM/CME underestimated damages. In support of this 
claim, these reviewers noted that when used to calculate damages for 
certain actual releases, the model generated damage figures that were 
usually at least an order of magnitude less than the figure for which 
the parties settled.
    Response: The Department believes that when the conditions set 
forth in Sec. 11.34 are met, the models will generate reasonable and 
appropriate damage figures for the injuries and losses these simplified 
procedures address. The Department does not believe that historical 
settlements provide an accurate or meaningful standard against which to 
judge the reliability of damage figures generated by the NRDAM/GLE and 
the NRDAM/CME. Although real-life case data on physical fates and 
biological effects can, in some instances, provide useful comparisons 
when evaluating the physical fates and biological effects submodels, 
bottom-line settlement figures may differ from model damage figures for 
a number of reasons that have nothing to do with reliability.
    First, because of the cost involved in performing site-specific 
type B studies, trustees have rarely pursued damage claims for minor 
releases. Therefore, historical natural resource damage settlements 
usually involve large spills. The type A models were designed for minor 
releases and are based on various assumptions that often are not 
reasonable in the case of large spills. Therefore, the restoration and 
compensable value submodels would not have been applicable to the cases 
in which natural resource damage settlements have been reached.
    Second, it is difficult to determine the appropriate user inputs 
for some of the actual cases, many of which are several years old. For 
example, user-supplied information on beach, and fisheries closures can 
significantly affect the total damage figure, yet data on the actual 
extent of such closures are in some cases no longer available.
    Third, the models do not purport to capture all, or even most, of 
the ``real world'' or ``actual'' damages that could be determined if 
the costs of a full on-site assessment were not a consideration. 
Instead, the models use averaged values to calculate a specific subset 
of the damages resulting from a release. When used for the minor

[[Page 20573]]

releases for which they are intended, the models yield reliable and 
appropriate damage figures that are calculated at a reasonable cost. 
Past natural resource damage settlement agreements have generally 
identified a single damage figure that is not broken down by component. 
In fact, most settlement agreements to date have not even listed which 
types of injuries and losses the agreement is intended to address. 
Therefore, it is usually impossible to determine if the model is even 
calculating the same type of damages as those covered by the 
settlement, let alone whether the calculation produces a damage figure 
that matches the settlement figure. The larger--and more complicated--
the release, the greater the likelihood of a divergence between the 
type A damage figures and the more site-specific damages that might be 
calculated using type B procedures. The fact that such divergence 
occurs, and even at times might appear ``extreme,'' does not suggest 
unreliability or an inappropriate ``underestimation'' of damages by the 
type A models. Rather, it only serves to illustrate the limited 
function these procedures are intended to serve, and the reason they 
are designed to be used for minor releases, for which the costs of type 
B procedures cannot be justified when compared to the anticipated level 
of damages.
    Finally, settlements are the result of negotiation. The negotiation 
process usually begins before either party has completed its assessment 
work. Settlement negotiations are influenced by both parties' 
perception of several factors extraneous to the assessment process. 
These factors include: the transaction costs associated with delaying 
settlement or terminating negotiations and litigating the case; the 
strength of the liability portion of the case; the PRP's financial 
condition; and the trustee's ability to fund a complete assessment. In 
light of the influence of these factors in settlement negotiations and 
the other difficulties in comparing settlement figures against model 
calculations, the Department does not believe that variances between 
model damage figures and historical settlements indicates anything 
about the reliability of the models, when used as intended.
    Comment: One of the independent technical reviewers questioned why 
the damages calculated by the proposed NRDAM/CME do not agree with 
those calculated by the original NRDAM/CME for the same spill.
    Response: The new NRDAM/CME differs significantly from the 1987 
version of the model due to modifications made in compliance with the 
Colorado v. Interior remand as well as modeling and database 
improvements made as a result of the biennial review. Among the most 
significant differences, the original model assumed a generic study 
area defined by the user with uniform depth, habitat, and environmental 
conditions. Today's final NRDAM/CME allows for geographic resolution of 
multiple habitats, depths, coastline, shore type, currents, ice cover 
and other environmental condition. The new NRDAM/CME contains much 
larger biological and economic databases, resolving many more species 
categories and geographic regions. Also, the new NRDAM/CME contains a 
restoration submodel and restoration cost database. The Department 
believes that these and other changes have resulted in significant 
improvements in the reliability of the calculations of the model.
    Comment: One commenter stated that the models were unreliable 
because NOAA used them to develop proposed OPA compensation formulas 
that generated unrealistic damage figures.
    Response: The Department does not believe that damage figures 
produced by NOAA's proposed OPA compensation formulas are relevant to 
the evaluation of either the proposed or final versions of the NRDAM/
GLE and NRDAM/CME. On January 7, 1994, NOAA proposed compensation 
formulas for determining natural resource damages under OPA. 59 FR at 
1176-77. These formulas were based on early developmental drafts of the 
NRDAM/GLE and the NRDAM/CME that the Department made available to NOAA 
in 1991. The Department has extensively modified both the NRDAM/CME and 
NRDAM/GLE since 1991. For example, the Department has revised the 
algorithms contained in the physical fates and biological effects 
submodels; expanded and updated the biological databases; and revised 
the chemical and economic databases. Section VI.B of this preamble 
identifies other major changes that the Department made to the NRDAM/
CME and NRDAM/GLE as a result of public comments.
    Comment: One commenter stated that the proposed NRDAM/CME 
dramatically underestimated damages as compared to the compensation 
table developed by the State of Washington under its natural resource 
damage laws. The commenter expressed concern that PRPs may use the 
NRDAM/CME to seek reductions in the State compensation table.
    Response: The Department does not believe it is appropriate or 
relevant to compare the results of type A model runs against the 
figures in Washington's compensation table, because the type A models 
and the State table are based on different approaches to damage 
assessment. The Washington table establishes a pre-set, per-gallon 
scale of damages. The type A models, on the other hand, estimate the 
actual effects of the release and then generate a site-specific damage 
figure based on the cost of restoring injured resources plus selected 
public economic values lost pending recovery.
    With regard to PRPs' potential use of the type A models to 
undermine the Washington table, the Department would like to emphasize 
that the type A models were developed specifically for use under 
Federal law. State or tribal simplified procedures may take into 
account costs, economic values, or other considerations not reflected 
in the type A models. As such, the damages produced by the type A 
models are not an appropriate point of comparison for evaluating State 
or tribal procedures. The type A models in no way preempt State or 
tribal procedures that are authorized under and designed to enforce 
non-Federal laws.
    Comment: Several commenters questioned the disparity between the 
levels of sophistication of different components of the models. Some of 
the independent technical reviewers noted that the compensable value 
submodel, unlike the relatively complex physical fates and biological 
effects submodels, essentially amounted to a look-up table. These 
reviewers thought that the Department should develop a more dynamic 
economics model. Other commenters thought that significant disparities 
in complexity existed even within the physical fates and biological 
effects submodels.
    Response: The Department has attempted to incorporate the best 
available procedures for modeling all components of the type A models. 
The Department acknowledges that the levels of intricacy vary 
throughout the models. These variances reflect the differing degrees of 
current technology and scientific knowledge. Economic science has not 
progressed to the point where there are general models of recreational 
demand that can be readily applied to specific recreational activities 
at specific locations. This is in distinct contrast to the biological 
and physical sciences. The physical fates and biological effects 
submodels are based on parameterizations of known and generally 
accepted models of physical and biological processes.
    Comment: One of the independent technical reviewers stated that the 
models incorporate some biases that will result in underestimates of 
damages and other biases that will result in

[[Page 20574]]

overestimates. The technical reviewer suggested that the models provide 
a range of damage estimates that reflect consistent use of conservative 
assumptions on one end and consistent use of liberal assumptions at the 
other end. Another independent technical reviewer suggested that the 
models be modified to perform an uncertainty analysis for each run.
    Response: The Department believes it has adequately and 
appropriately addressed the potential for bias in the NRDAM/CME and 
NRDAM/GLE. The type A procedures are principally designed to establish 
a process for trustees to follow if they wish to pursue a natural 
resource damage claim and obtain a rebuttable presumption in court. In 
a suit for damages, trustees will need to identify a specific claim. 
Therefore, the Department has developed type A models that generate a 
single damage figure rather than a range of possibilities.
    Moreover, where commenters, or the Department itself, identified 
specific potential biases in the proposed models, the Department 
modified the models to correct for such biases to the extent possible. 
Where the Department could not eliminate the potential for bias, it 
identified the simplifying assumptions made in the models that produce 
that potential. As discussed further below, those assumptions that 
could result in significant overestimates of damages if they are not 
reasonable in a particular case are listed in Sec. 11.34 as conditions 
that must be met if the trustees expect to obtain a rebuttable 
presumption. Those assumptions that are not likely to result in 
significant overestimates of damages if they are not reasonable in a 
particular case, and, in fact, may result in underestimates, are 
explicitly identified in Section 1, Volume I of the NRDAM/CME and 
NRDAM/GLE technical documents. Section 11.35(a) provides that if a type 
A procedure is applicable, trustees must determine whether to use type 
A or type B procedures based on an evaluation of those model 
assumptions.
    As discussed in Section III of this preamble, the type A models are 
neither expected nor intended to produce damage estimates that 
``match'' the results of more complex site-specific assessment 
procedures. Therefore, the Department has concluded that a traditional 
uncertainty analysis is not needed.
    Comment: Several commenters thought the scope and complexity of the 
proposed NRDAM/GLE and NRDAM/CME were too great. A few commenters 
thought the models were so complex and difficult to use that operating 
them was beyond the ability of untrained users. One commenter thought 
the technical documents should clearly state the required user 
qualifications. Several commenters, including some of the independent 
technical reviewers, suggested improved user interfaces. Some of the 
independent technical reviewers thought that additional user guidance 
was needed; one suggested that the Department develop an animated 
tutorial.
    Response: While the Department acknowledges that the NRDAM/CME and 
NRDAM/GLE are functionally very complex, it does not believe that they 
require an undue level of expertise to operate. Users must simply be 
able to: (1) Understand the conditions for use in Sec. 11.34; (2) 
evaluate the models' simplifying assumptions listed in of Section 1, 
Volume I the technical documents; (3) evaluate the averaged data 
included in the models as described in Volumes III through IV of the 
NRDAM/CME technical document and Volume III of the NRDAM/GLE technical 
document; and (4) enter correctly the required user-supplied data as 
described in Appendices II and III of the rule. Users who meet these 
standards will obtain reliable results regardless of whether they have 
a full understanding of all the models' components.
    As discussed further below, the Department has revised the 
regulatory conditions for use of the models to clarify a number of 
points of confusion. Section 1, Volume I of the NRDAM/CME and NRDAM/GLE 
technical documents now contains a clearer, simpler discussion of all 
the major model assumptions of which users should be aware when 
determining whether to use type A or type B procedures. The Department 
has also rewritten the regulatory discussion of the user-supplied 
information and moved that discussion into appendices in an attempt to 
make it easier to read. Volume II of the technical documents includes a 
revised discussion of how to develop and input the user-supplied data. 
Finally, the models provide a graphic user interface that has been 
revised to further simplify the task of the user. While additional 
guidance might be helpful and may be developed in the future, the 
Department believes that the current level of guidance is adequate to 
allow non-expert users to operate the model correctly.
    Comment: Some of the independent technical reviewers questioned why 
the user interface was not consistent with Windows software.
    Response: The Department chose to develop the user interface as a 
stand-alone product that would not require licensing a copyrighted 
product such as Windows software.
    Comment: A few commenters complained about the speed of the 
proposed models. Some commenters called upon the Department to upgrade 
the computer platform required to run the models. The commenters 
thought that such an upgrade would enable users to complete model runs 
in hours rather than days and would allow the models to use more 
detailed databases, thus increasing accuracy.
    Response: In developing the type A models, the Department had to 
strike a balance between the desire for the speed afforded by high-
powered computer equipment and the need to ensure that any type A 
procedure developed is readily accessible to a wide array of potential 
users. The Department believes it has struck the appropriate balance in 
the PC environment.
    It is evident from even a cursory review of the technical documents 
that the models are very complex and perform millions of individual 
calculations during a run. The Department has made every effort to 
optimize the models for speed without compromising their accuracy or 
applicability. Obviously, there continue to be advances in PC 
technology. For the development of the type A models, it was necessary 
for the Department to settle on a widely-available computer platform 
and finalize the rule. While more recent technological developments 
will allow these models to run faster on improved computer platforms, 
the Department decided that maintaining the models for use on 386 PCs 
would not compromise their function or purpose and would keep them 
readily accessible to potential users.
    Model run times are affected by the complexity of the spill (e.g., 
amount spilled, duration of the spill, and degradation rate of the 
spilled substance) as much as the computer platform utilized. 
Nonetheless, for minor spills, most runs are executed in a matter of 
minutes rather than hours or days even on a 386 PC. The models will 
take significantly less time to run on a 486 PC or a Pentium 
PC, but the user is not precluded from using an older model of 
computer.
    Since the issuance of the proposed rule, the NRDAM/CME has been 
moved to a 32-bit FORTRAN compiler. This move allowed the 
Department to subdivide the habitat grids by a factor of four and 
increase the number of computational particles used to represent 
spilled material and biota. These changes should improve the accuracy 
of the model. The area

[[Page 20575]]

modeled in the NRDAM/GLE is much smaller than that modeled in the 
NRDAM/CME. Therefore, the Department concluded that these changes were 
not needed in the NRDAM/GLE to increase speed or accuracy.
    Comment: One of the independent technical reviewers stated that 
when he attempted to replicate test runs on the proposed NRDAM/CME he 
obtained different results.
    Response: Users will obtain identical results if, but only if, they 
use identical inputs. The Department designed the type A models so that 
they will produce identical results, regardless of the make or model of 
PC used, if the user-supplied inputs are identical. To accomplish this 
result, the Department built a table of random numbers into the models' 
code rather than have the models use the random number generating 
features of the microprocessor.
    Comment: Several commenters, including some of the independent 
technical reviewers, suggested that the Department include additional 
categories of damages in the type A models. Commenters recommended that 
the Department add the following losses to the models: sublethal 
biological effects; chronic biological effects; wetland losses; nonuse 
losses (i.e., economic values that are not dependent on use of a 
resource, such as the value of knowing a resource exists); de facto 
beach, boating, and fisheries closures; reductions in the quality of 
boating and beach recreation in the absence of closures; reductions in 
the quality of recreational fishing unrelated to mortality or closures; 
and ecosystem functional losses such as reductions in filtration, 
mineral recycling, and decomposition. These commenters expressed 
concern that if the models are not expanded to cover additional losses, 
then type A assessments will consistently underestimate damages. They 
noted that Ohio v. Interior and Colorado v. Interior instructed the 
Department to allow for the recovery of all reliably calculated losses. 
Commenters also thought that, in light of the cost of type B 
procedures, it was disingenuous of the Department to state that 
trustees could simply use type B procedures to calculate damages for 
losses not included in the models.
    Response: The Department has attempted to include in the models all 
categories of loss and injury for which adequate, reliable information 
exists in a format that enables the calculation of damages for the wide 
range of substances, resources, and geographic areas covered by the 
models. The Department acknowledges that the type A models do not 
address all potential losses and injuries that might result from a 
release and that, in some cases, losses not included in the models may 
be significant. The Department further acknowledges that Ohio v. 
Interior and Colorado v. Interior instructed the Department to allow 
for the recovery of all reliably calculated values. The issue, then, is 
reliability. The exclusion of certain categories of injury and loss 
from the models was based on the Department's evaluation of whether 
there was adequate reliable information to support their inclusion.
    For example, the Department has considered the comments suggesting 
the addition of nonuse losses, but continues to believe that the 
addition of such values is not feasible at this time. As discussed in 
the proposed NRDAM/CME technical document, most studies of nonuse 
values do not report marginal nonuse values that would be required for 
the type A models as they are presently designed. See Section 8.5.2, 
Volume I of the proposed NRDAM/CME technical document. Furthermore, 
these studies have tended to focus on the nonuse values of threatened 
or endangered species. As a consequence, the bulk of available studies 
are not directly applicable to the estimation of nonuse values that 
would be lost as a result of the small spills addressed by the type A 
models.
    Furthermore, the final rule explicitly provides that where trustees 
expect losses that are not addressed by the models, they may consider 
using type B procedures in addition to a type A procedure, provided 
that type B procedures are cost-effective, can be performed at a 
reasonable cost, and do not result in double recovery. The Department 
recognizes that type B procedures are likely to be significantly more 
costly than type A procedures and, in some cases, trustees may not be 
able to perform type B procedures and still satisfy the rule's 
reasonable cost standard. Nevertheless, the Department does not believe 
that the cost of performing type B procedures justifies the inclusion 
in the models of losses for which there is an inadequate basis to 
determine damages. During future biennial reviews, the Department will 
reevaluate whether additional information has become available that 
supports expansion of the categories of losses and injuries included in 
the models.
    Comment: One of the independent technical reviewers stated that 
additional detail should only be added to the models if it influences 
the final damage figure.
    Response: The ultimate purpose of all the calculations made by the 
type A models is the determination of a reliable damage figure. 
Therefore, while reviewing the comments and deciding which changes to 
make to the models, the Department has focused on whether the suggested 
changes would significantly improve the reliability of the final damage 
figure.
    Comment: One commenter suggested that trustees be allowed to use 
simplified procedures developed by States and receive a rebuttable 
presumption under the CERCLA regulations. Another commenter requested 
that the Department develop compensation tables for commonly released 
hazardous substances.
    Response: Some simplified State or tribal procedures may well be 
appropriate for use under CERCLA. However, only a handful of coastal 
States have developed such procedures. Further, these State procedures 
have been developed under State laws, which may establish somewhat 
different objectives and standards than CERCLA. The Department believes 
it would need to evaluate carefully any particular State or tribal 
procedure to determine its consistency with CERCLA's regulatory mandate 
before allowing it to be used and accorded a rebuttable presumption 
under these regulations. Therefore, the Department decided it was more 
appropriate to develop its own simplified procedures for the coastal 
and marine and Great Lakes environments.
    The primary advantage of compensation tables appears to be their 
ease of use. The Department believes that the NRDAM/CME and NRDAM/GLE 
are simple enough to operate that compensation tables are not 
necessary. Further, the Department believes that the models will 
provide a level of site-specific accuracy beyond that which a 
compensation table could offer.
    Although the Department has decided not to incorporate compensation 
tables or simplified State or tribal procedures in this rulemaking, the 
Department has begun to evaluate the need for, and feasibility of, 
additional type A procedures. See 60 FR 24604 (May 9, 1995). The 
Department will further consider the use of simplified State procedures 
and the development of compensation tables in that context.

B. Technical Documents

    Comment: Some commenters stated that the Department had failed to 
provide adequate documentation explaining how the proposed models 
operated and why the Department made the choices it did when developing

[[Page 20576]]

different components of the proposed models. One commenter stated that 
the scope and complexity of the models were too great and suggested 
that a revised program be developed and accompanied by a simplified 
synopsis of the technical assumptions and formulas presented in a 
format more amenable to comment. Commenters cited case law requiring 
agencies to provide a complete explanation and defense of models used 
in the development of regulations. The commenters noted that the 
Department's obligation to provide a full discussion of the type A 
models was even greater because the models are used to determine 
monetary liability of particular parties.
    Response: The Department acknowledges its duty to provide an 
adequate explanation and justification of the models and to provide the 
public with a meaningful opportunity to review and comment on the 
proposed models. The Department believes it has fulfilled this duty.
    The proposed models were accompanied by lengthy and detailed 
technical documents describing the content, workings, and development 
of the models. The proposed NRDAM/CME technical document exceeded 2,400 
pages in length; the proposed NRDAM/GLE technical document was almost 
1,500 pages in length. Also, the preambles to the proposed rules 
provided a roadmap to the technical documents, highlighting areas of 
potential concern and identifying where various issues were discussed 
in the technical documents. The Department made the proposed models and 
technical documents available on diskette free of charge to anyone who 
requested them.
    To assist commenters in reviewing the models, the Department 
equipped the proposed models with a user interface that included pull-
down menus, ``help'' screens, and graphic displays of the physical 
environments and user-generated runs of the physical fates submodel. 
The Department also incorporated pertinent calculations from the 
physical fates, biological effects, restoration, and compensable value 
submodels into the printed model output to enable reviewers to evaluate 
the reliability of the models for incident-specific model applications.
    The Department notes that the goal of developing models that 
calculate compensatory damages for spills throughout the Great Lakes 
and coastal and marine environments has necessitated a relatively high 
level of complexity in modeling. The Department recognizes that with 
models as complex as the NRDAM/CME and NRDAM/GLE some reviewers will 
always want more information on specific elements while others will be 
overwhelmed as the documentation becomes more extensive. Although the 
Department never deliberately omitted any discussion it thought would 
be of interest to reviewers, the Department did recognize that 
providing too much information can be just as problematic as providing 
too little. The Department has tried to be sensitive to the risk that 
important information can become buried in a mountain of detail.
    The Department extended the public comment period on the proposed 
NRDAM/CME once and on the proposed NRDAM/GLE twice. The total comment 
periods were seven months for the proposed NRDAM/CME and eleven months 
for the proposed NRDAM/GLE. Those reviewers left with questions after 
reviewing the models and technical documents were free to contact 
Departmental staff at any time during the comment period.
    Finally, the Department has provided additional discussion of 
specific model aspects in the final versions of the NRDAM/CME and 
NRDAM/GLE technical documents and in this preamble as a result of 
specific public comments.
    Comment: Several commenters, including some of the independent 
technical reviewers, said that the technical documents were either 
unclear or difficult to use. Others noted confusing table captions and 
headings, inconsistencies, incorrect citations, and typographical 
errors. One commenter suggested that major assumptions for each 
submodel be placed in bold print at the beginning of each section. One 
commenter recommended that the technical documents be amended to give 
examples of when the models might underestimate or overestimate 
damages.
    Response: The Department has reviewed and revised the NRDAM/CME and 
NRDAM/GLE technical documents to further clarify algorithms, 
assumptions, and data sources. The Department has also checked the 
documents for consistency, particularly with regard to terminology and 
has fixed the noted typographical errors and incorrect citations. 
Section 1, Volume I of the NRDAM/CME and NRDAM/GLE technical documents 
now more clearly identifies all the major assumptions of which trustees 
should be aware when deciding whether to use the models and describes 
the likely results if the assumptions are not reasonable in a 
particular case. Further, the discussion of each submodel in Volume I 
of the technical documents now starts with a list of the assumptions 
relevant to that submodel.

C. Selection of Assessment Procedures

    Comment: The Department received numerous comments on the proposed 
conditions for use of the type A models. The proposed rules identified 
a set of primary conditions and a set of secondary conditions. Under 
the proposed rules, if any primary condition were not met, trustees 
would not have been allowed to use the type A procedure. If all primary 
and all secondary conditions were met, trustees would have been 
required to use the type A procedure for all damages. If all primary 
conditions but only some secondary conditions were met, trustees could 
have used a combination of type A and type B procedures.
    Some commenters thought the proposed rules were overly prescriptive 
in dictating which type of assessment procedures trustees may use. 
These commenters argued that trustees should have greater discretion to 
determine which procedures, type A, type B, or a combination, are 
appropriate in a particular case. Commenters expressed concern that the 
conditions regarding use of the type A procedures were vaguely defined 
and would invite confrontation and litigation if they were imposed as 
requirements. These commenters supported expansion of the authority to 
use type A and type B procedures in combination, but thought the 
proposed rules still did not provide adequate flexibility. These 
commenters also stated that the type A procedures were particularly 
useful when used with selective site-specific studies of impacts not 
addressed in the type A models.
    Other commenters, including one of the independent technical 
reviewers, thought that the proposed rules gave trustees too much 
discretion in selecting assessment procedures. Some of these commenters 
thought that the conditions regarding use of the type A procedures 
should be clearer and stricter. One of the independent technical 
reviewers suggested that the Department recharacterize the assumptions 
made by the models as limits of applicability. One commenter argued 
that trustees be required to use a type A procedure unless they provide 
scientific justification for using type B procedures. On the other 
hand, some commenters expressed concern that the proposed rules would 
allow excessive use of the type A procedures and suggested making the 
primary conditions more restrictive.
    Several commenters objected to the proposed provision allowing 
combined

[[Page 20577]]

use of type A and type B procedures. The commenters argued that 
Congress intended the type A and type B procedures to be mutually 
exclusive. These commenters also thought that combined use of type A 
and type B procedures would pose significant risks of double recovery 
of damages and that the proposed rules failed to provide any guidance 
on how to prevent such double recovery. One commenter stated that 
combined use of type A and type B procedures was inconsistent with the 
``average'' values justification for simplified procedures, since type 
B procedures would be used to offset type A underestimates without any 
corresponding offset of type A overestimates. Another commenter 
expressed concern that if allowed to supplement type A assessments, 
trustees would spend enormous sums assessing nonuse values for small 
releases even though such releases are unlikely to produce any 
meaningful nonuse losses. Some commenters stated that if the final rule 
allowed use of type B procedures to supplement a type A assessment, 
then such use should be limited to resources not included in the type A 
procedure.
    Response: The type A models are powerful tools for completing 
assessments and beginning restoration as quickly and cost-effectively 
as possible. The Department has sought to balance the utility of making 
these tools available in the widest possible range of cases against the 
potential dangers that they may produce unreliable results when 
stretched beyond their limits or that they may result in double 
recovery when inappropriately combined with type B procedures.
    The Department has carefully reexamined both the proposed 
conditions regarding use of the models as well as the additional major 
simplifying assumptions incorporated into the models and described in 
the technical documents. The Department has concluded that the 
conditions for use of the models should recognize two different 
categories of assumptions built into the models. The first category 
encompasses those assumptions that could result in significant 
overestimates of damages if they are not reasonable in a particular 
case. The second category encompasses those assumptions that are not 
likely to result in significant overestimates of damages if they are 
not reasonable in a particular case and that may well result in 
underestimates.
    The Department believes it is inappropriate to grant a rebuttable 
presumption to an assessment performed using the NRDAM/CME or NRDAM/GLE 
if one of the assumptions in the first category is not reasonable in 
the particular case. If an assumption in the second category is not 
reasonable in a particular case, it may be appropriate for trustees to 
use type B procedures to ensure that the public receives full 
compensation for its losses. However, the Department believes trustees 
in those cases should have the option of using the type A models when 
the costs of type B procedures are not reasonable. The appropriateness 
of the models in these cases will depend on site-specific factors. The 
Department has concluded that it is more appropriate to allow trustees 
to analyze these factors in the context of a particular case than to 
establish inflexible, overly rigid standards.
    Therefore, the Department has identified all the major model 
assumptions and for each one determined into which of the two 
categories they fall. Those assumptions in the first category are 
identified in Sec. 11.34 of the final rule as conditions that must be 
met if trustees intend to use the NRDAM/CME or NRDAM/GLE and obtain a 
rebuttable presumption. These assumptions include most of the primary 
conditions in the proposed rules.
    The Department has identified the assumptions in the second 
category and listed them, along with the other assumptions, in Section 
1, Volume I of the NRDAM/CME and NRDAM/GLE technical documents. These 
assumptions include many of the secondary conditions in the proposed 
rule. Section 11.35(a) provides that if the conditions for use of a 
type A procedure are met, the trustee must decide whether to use that 
procedure or use type B procedures by weighing the difficulty of 
collecting site-specific data against the suitability of these 
additional assumptions as well as of the averaged data described in 
Volumes III through IV of the NRDAM/CME technical document, and in 
Volume III of the NRDAM/GLE technical document.
    The Department has eliminated the proposed provision that would 
have required trustees to use a type A procedure in some cases. That 
requirement was originally motivated out of concern over potential 
misuse of unnecessarily expensive and time-consuming type B procedures. 
59 FR at 40322. Although the models are cost-effective, reliable tools 
where applicable, the Department has concluded that trustees should not 
be prevented from conducting site-specific work if they can do so at a 
reasonable cost and if the additional costs of performing type B 
procedures are warranted in light of the degree of additional precision 
and accuracy that such procedures will provide.
    The issue of the legal permissibility of allowing trustees to use 
both type A and type B procedures for the same release is one that the 
Department decided and resolved in 1987 and is beyond the scope of this 
rulemaking. Today's final rule merely expands the use of supplemental 
type B studies beyond resources not addressed by the type A procedure 
to include compensable values and injuries of a type not addressed by 
the type A procedure.
    The Department acknowledges that combined use of type A and type B 
procedures can, in some instances, pose potential double counting 
problems. However, trustees should not be forced to choose between 
forgoing compensation for a public loss not addressed by the type A 
model on the one hand and funding a full-scale, time-consuming, labor-
intensive type B assessment of all injuries on the other hand. Instead, 
the potential problems with combined use of type A and type B 
procedures should be addressed through limitations designed to protect 
against double recovery.
    The final rule provides that trustees who use a type A procedure 
may perform additional type B studies only for injuries or compensable 
values of a type not addressed by the type A procedure. The secondary 
conditions in the proposed rules have been recast to identify 
explicitly the injuries and compensable values that are addressed in 
the type A models and, therefore, may not be supplemented with type B 
procedures.
    Given the vast range of potential scenarios, it is infeasible to 
develop a single, uniform formula for preventing double recovery. 
Instead, Sec. 11.15(d) of the existing regulations prohibits double 
recovery of damages. Also, Sec. 11.36(a)(2) of today's final rule 
provides that trustees may only perform supplemental type B procedures 
if such procedures will not result in double recovery. Further, 
Sec. 11.36(c) requires trustees to provide an explanation in the 
Assessment Plan of how they intend to avoid any double recovery in the 
case of combined use of type A and type B procedures. PRPs and the 
public will have an opportunity to review the trustees' strategy for 
preventing double recovery when the Assessment Plan is made available 
for public comment.
    The Department agrees with the comment that the type A procedures 
can be particularly useful when combined with selective studies of 
impacts not addressed by the models. The Department would like to 
ensure that

[[Page 20578]]

where combined use of type A and type B procedures is warranted, 
trustees are freed from conducting duplicative assessment procedures. 
Therefore, the Department has modified the final rule to clarify that 
when using type B procedures for compensable values that are not 
included in a type A procedure, but that result from injuries addressed 
by the type A procedure, trustees need not conduct injury determination 
and quantification all over again using type B procedures. Instead, 
trustees may rely on the injury projections of the type A model and 
simply use one of the type B valuation methodologies authorized by 
Sec. 11.83 (a) and (c) to compute compensable value.
    With regard to the concern about unwarranted type B studies of 
nonuse values, aside from the implausibility of the scenario suggested 
by the commenter, the Department notes that calculation of nonuse 
values using type B procedures is under examination in a separate 
rulemaking. See 59 FR 23097 (May 4, 1994). Therefore, this rulemaking 
need not address this issue.
    Finally, the Department believes it is appropriate to revise the 
existing rule to allow supplemental use of type B procedures beyond 
resources not addressed in the type A models. The public can experience 
significant and distinct losses associated with the same resource. Ohio 
v. Interior emphasized that the regulations should allow for the 
recovery of all reliably calculated lost values. 432 F.2d at 464. The 
Department sees no reason to impose an arbitrary distinction between 
losses associated with different resources and losses associated with 
the same resource so long as there is no double recovery.
    Comment: One commenter suggested that trustees be allowed to use 
supplemental type B procedures to determine damages for habitats that 
are not accurately represented in the models.
    Response: In cases where the models assign an incorrect habitat 
designation for a specific area, trustees have the ability to correct 
that designation and would not need to conduct supplemental type B 
studies. In cases where releases affect habitats beyond the models' 
level of spatial detail, trustees may perform supplemental type B 
studies so long as such studies do not address injuries or compensable 
values in the categories listed in Sec. 11.36(b) of the final rule. The 
Department does not believe it is appropriate to expand this authority 
to conduct supplemental type B studies and still obtain a rebuttable 
presumption. When such small habitats are affected, the models will 
nonetheless determine injury and damages for the geographic area in 
which those habitats are located. If a trustee were to use one of the 
models and then conduct supplemental type B studies of such a habitat, 
the trustees would need to adjust the type A damage figure to eliminate 
any damages calculated for the area over which the habitat is located. 
The Department has concluded that in the context of a simplified 
assessment, trustees who wish to obtain a rebuttable presumption should 
be limited to conducting type B studies for the purposes of addressing 
additional injuries and compensable values that are not included in the 
model rather than substituting for damages already calculated by the 
model.
    Comment: A number of commenters thought that trustees should be 
prohibited from using type A procedures unless all interested trustees 
agree to a single joint assessment. These commenters stated that such a 
provision was necessary to avoid the problems of double recovery and 
improper allocation of damages among trustees. These commenters thought 
that these problems were more significant for type A assessments than 
for type B assessments because the type A models provide less detail 
than type B procedures on the type and location of injured resources 
and the damages associated with those resources.
    Response: The Department acknowledges that the type A models pose a 
unique problem when trustees do not act jointly. The type A models 
generate a total damage figure for all affected resources. Therefore, 
if a trustee acts independently and applies a type A model, the total 
damage figure generated by the model might include damages for 
resources that are not under that trustee's jurisdiction.
    To address this problem, Sec. 11.42 now requires a trustee to 
perform a preliminary application of the model before making the draft 
Assessment Plan available for public review and comment. The trustee 
must include a summary of the model application in the draft Assessment 
Plan and make available a copy of the model output. The output of the 
model does in fact identify the type and location of injured resources. 
Section 11.31(a)(2) of the existing regulations requires trustees to 
include in the Assessment Plan a statement of authority for asserting 
trusteeship for those resources addressed in the Plan. Therefore, PRPs 
and other interested members of the public will have an opportunity to 
comment on whether any of the injured resources identified in the model 
output are beyond the scope of the trustee's jurisdiction.
    Also, Sec. 11.44(e) provides that if a trustee is aware of reliable 
evidence that a type A application covers resources beyond his or her 
trustee jurisdiction, the trustee must either: (1) Have the other 
trustees who do have jurisdiction over those resources join in the type 
A assessment; or (2) eliminate any damages for those resources from the 
claim for damages.
    Furthermore, the Department strongly encourages trustees to work 
together to ensure that natural resource damage assessments remain 
focused on restoring the injured resources rather than debating over 
which trustee has jurisdiction over them. As noted by some of the 
commenters, Sec. 11.32(a)(1) of the existing regulations requires a 
trustee to notify all other interested trustees before beginning an 
assessment and encourages all trustees to cooperate and coordinate. 
Also, Sec. 11.15(d) of the existing regulations prohibits double 
recovery of damages.
    The issue of inter-trustee coordination extends beyond this 
rulemaking to the overall administrative process for conducting all 
assessments. The potential for overlapping claims exists whenever 
trustees conduct separate assessments, regardless of whether type A or 
type B procedures are used. The Department has initiated a biennial 
review of the administrative process for conducting assessments. The 
Department will be further examining the issue of inter-trustee 
coordination during that review. 59 FR at 52752.
    Comment: A few commenters stated that PRPs should be ensured a 
meaningful opportunity to participate in the selection of assessment 
procedures. These commenters requested that PRPs be given a chance to 
review trustees' assumptions and reasoning. Commenters also expressed 
support for cooperative trustee-PRP assessments.
    Response: The Department agrees that PRPs should have an 
opportunity to participate in selection of assessment procedures. 
Section 11.32(a)(2)(iii)(A) of the existing regulations already 
requires trustees to invite PRPs to participate in the development of 
the type and scope of the assessment as well as the performance of the 
assessment procedures. Today's final rule does not change that 
requirement. Section 11.32(c) requires trustees to make their 
Assessment Plans available for public review and comment. The proposed 
rule required trustees to include in their Assessment Plans 
documentation of their decision whether to use a type A procedure, type 
B procedures, or both. Section 11.31(b) of today's final rule

[[Page 20579]]

now makes more explicit the trustees' duty to provide a detailed 
explanation of their rationale for using a type A procedure, type B 
procedures, or both. Also, Sec. 11.35(d) now clarifies that trustees 
may change their decisions about the types of procedures they use based 
on public comments.
    Comment: Many commenters addressed specific proposed conditions for 
use of the models. Some commenters questioned the condition regarding 
whether the data in the models reasonably represented the spatial and 
temporal distribution of affected biological resources. One commenter 
suggested that this condition was inconsistent with the habitat editor. 
Another commenter requested clarification of the term ``reasonably 
represented.'' This commenter expressed concern that the condition 
seemed to require trustees to collect baseline data, which would defeat 
the intent of requiring minimal field observation in type A procedures.
    Response: The Department has reexamined this proposed condition 
regarding use of the models. The condition addressed two different 
model assumptions. First, the condition addressed the assumption that 
the release did not affect any small but important environments beyond 
the level of spatial detail of the model. Second, the condition 
addressed the assumption that species biomass is averaged spatially and 
temporally. The Department has concluded that if the first assumption 
is not reasonable in a particular case, then the model will most likely 
underestimate, rather than overestimate, damages. Therefore, the 
Department has eliminated this assumption from the conditions for use 
listed in Sec. 11.34 of the final rule. Instead, the Department has 
identified the assumption in the NRDAM/CME and NRDAM/GLE technical 
documents as one of the factors for trustees to consider when deciding 
whether to use type A or type B procedures, once they have established 
that the conditions set forth in Sec. 11.34 are met. See Section 1, 
Volume I of the NRDAM/CME and NRDAM/GLE technical documents. The 
Department has clarified the second assumption, concerning species 
biomass, and included it in Sec. 11.34 as a condition that must be met 
if trustees intend to use the models and obtain a rebuttable 
presumption.
    The habitat editor does not conflict with either of these 
assumptions. The final rule allows trustees to change the habitat 
designation for an entire existing grid cell. However, the rule does 
not allow trustees to redraw the boundaries of the grid cells or modify 
the species biomass for a particular habitat. Even with correct habitat 
designation, edited or through the built-in designation, the models may 
not reflect small habitats or populations with densities that differ 
from the seasonal average.
    The Department acknowledges the confusion generated by the term 
``reasonably represented.'' The term was not intended to require 
trustees to conduct field surveys to collect baseline data. Instead, it 
was designed to address cases where information already existed about 
baseline conditions and such pre-existing information differed 
significantly from the data in the model. Section 11.34(e) now simply 
provides that a trustee may not use the models if he or she is aware of 
reliable evidence that, for species expected to represent a significant 
portion of the claim, the species biomass is significantly lower than 
the species biomass assigned by the models.
    Comment: A few commenters noted that the models may significantly 
underestimate damages when the released substance causes chronic or 
sublethal effects, when sensitive habitats or life stages are affected, 
when animals aggregate for feeding or reproduction, or when long-term 
effects, such as reproductive impairment or changes in food web 
structure, are expected.
    Response: The Department acknowledges that the type A models may 
not accurately calculate total damages in the situations identified by 
the commenters. However, the Department has included provisions in the 
final rule to address these situations. Section 11.35(a) provides that 
if a type A procedure is applicable, trustees must determine whether to 
use type A or type B procedures based on an evaluation of the model 
assumptions listed in Section 1, Volume I of the NRDAM/CME and NRDAM/
GLE technical documents. One of the listed assumptions is that there 
are no affected environments beyond the spatial detail of the models. 
This assumption will alert trustees to the potential for 
underestimating damages where sensitive habitats are affected. Another 
listed assumption is that species biomass is averaged spatially and 
temporally. This assumption will alert trustees to the potential for 
underestimating damages when animals aggregate. Finally, the rule 
explicitly identifies the injuries addressed by the type A models; 
therefore, trustees will have notice that they will need to perform 
supplemental type B procedures if they wish to address chronic or 
sublethal biological injuries.
    Comment: Some commenters thought that trustees should be allowed to 
use the models only if the release is a single event. These commenters 
expressed concern that in the absence of such a requirement, trustees 
could use a type A procedure to assess one release in a multi-release 
incident and use type B procedures to assess the other releases. The 
commenters thought that such a practice would result in double counting 
because some of the injuries predicted for one release would already be 
accounted for in the assessment of another release. These commenters 
also thought that the rule should be rewritten to clarify that the type 
A models can be applied only to releases of a single substance. The 
commenters noted that without this change, similar double counting 
problems could arise from multiple applications of the models.
    Response: The Department has concluded that the model assumption 
that the release is a single event need not be made a condition for use 
of the models. Instead, Section 1, Volume I of the technical documents 
notes that the models assume that each spill is an independent, short-
term event.
    Section 11.15(d) of the existing regulations already prohibits 
double recovery of damages. In the case of a multi-release incident, if 
trustees choose to use a type A model for one release and then conduct 
type B studies for the other releases, they will be required to ensure 
that the type B procedures do not result in double recovery. The 
Department acknowledges that in some cases it may be difficult for 
trustees to satisfy this requirement. However, the Department believes 
that in those multi-release cases where trustees can tailor their type 
B studies to address only the effects of the releases not assessed by 
the type A model, they should have the opportunity to do so.
    With regard to releases of multiple substances, the rule now 
provides that trustees must select and assess only one of the 
substances that was released. See Appendices II and III. This 
requirement will eliminate double counting problems. In fact, toxicity 
of mixtures has been found to be additive or synergistic in aquatic 
environments for a wide variety of substances. See Section 4, Volume I 
of the NRDAM/CME technical document. Thus, this requirement may 
actually result in underestimates of damages. However, the Department 
believes that in cases of mixtures when the cost of using type B 
procedures is not reasonable, trustees should have the option of using 
a type A procedure rather than forgoing all compensation.
    Comment: A few commenters addressed the use of the models for

[[Page 20580]]

substances not specifically identified in the database. Some commenters 
supported giving trustees flexibility to use the models for such 
substances provided that they identified a proxy that was included in 
the database and documented the reasons why the use of that proxy was 
appropriate. Other commenters expressed concern that allowing use of 
proxies would add a significant range of discretion given the number of 
different physical and chemical attributes that must be considered when 
identifying a proxy. Some commenters, including one of the independent 
technical reviewers, suggested that the Department expand the oil 
database.
    Response: The Department has concluded that allowing the use of 
proxies for hazardous substances without significant guidance on 
selection of such proxies would raise serious concerns about the 
uniformity and reliability of the type A model results. Moreover, 
developing guidance on selection of proxies would be impractical given 
the extremely wide range of hazardous substances and the diversity of 
their relevant attributes. Therefore, for chemical releases, trustees 
may only use the models if the released material is one of the specific 
chemicals listed in the database.
    As discussed in Section I.C. of this preamble, use of the models 
for oil discharges is governed by NOAA's OPA rule rather than by 
today's final rule. However, the Department notes that it has expanded 
the database to include 33 types of oils that cover a broad range of 
chemical and physical characteristics.
    Comment: Some commenters stated that trustees should be allowed to 
use the type A models and obtain a rebuttable presumption for releases 
that did not originate in, but later migrated into, a coastal or marine 
or Great Lakes environment. These commenters argued that the type A 
models could accommodate such releases. On the other hand, one 
commenter thought that such use should not be allowed because there are 
no data on conditions outside the boundaries of the type A models.
    Response: The Department has concluded that the type A models can 
produce reliable damage figures for releases that do not originate in, 
but do migrate within, the boundaries of the models provided that the 
user supplies appropriate data inputs. So long as the user supplies 
data inputs that reflect conditions at the point that the substance 
enters the model boundaries, the models are just as capable of 
computing reliable damages as they would be if the release had actually 
started at that point. In such cases, the models will start their 
simulations at the point that the released substance enters water 
within a geographic region represented in the models. The only 
potential problem is that the models will not account for the effects 
produced before the release entered the model boundaries, a 
consideration that may support use of type B procedures in some cases. 
However, the Department believes trustees should have the option of 
using the models to assess such releases when the cost of performing 
type B procedures to develop a more complete damage figure is not 
reasonable.
    Therefore, the rule allows trustees to use the models for releases 
that occur outside the boundaries of the models so long as the user-
supplied inputs appropriately reflect conditions at the point that the 
substance entered such waters rather than the point of the original 
release. Appendices II and III specify that when using the models for 
releases that originate on land or outside the model databases, 
trustees must adjust the data inputs.
    Comment: One commenter said that the definition of ``minor'' was 
vague, but supported the Department's discussion of it and the proposal 
to allow trustees discretion to define ``minor'' on a case-by-case 
basis. Other commenters, including one of the independent technical 
reviewers, thought that the Department should define ``minor.'' One 
commenter suggested that the rule require trustees to justify their 
determination of whether a release is minor with scientific 
documentation. Another comment recommended the Department define 
``minor'' based on spill size, prediction of affected area, or 
resulting damage estimates.
    Response: In light of Congressional intent to restrict use of type 
A procedures to minor releases and after considerable analysis and 
deliberation, the Department has decided to impose a specific dollar 
cut-off for use of the models to obtain a rebuttable presumption. The 
final rule provides that if the model output indicates damages in 
excess of $100,000, then trustees who wish to obtain a rebuttable 
presumption must either: (1) limit the portion of their claim 
calculated with the type A procedure to $100,000; or (2) compute all 
damages using type B procedures. The Department believes this provision 
establishes an appropriate standard of fairness for allowing trustees 
to receive a rebuttable presumption for damages calculated by the 
NRDAM/CME or NRDAM/GLE given the current level of experience with these 
models.
    The language and legislative history of CERCLA indicate that 
Congress intended the type A procedures as a tool for obtaining a 
rebuttable presumption in cases of minor releases. Thus, the Department 
included a provision in the proposed rules that prohibited trustees 
from using the models to obtain a rebuttable presumption unless the 
release was minor. The proposed rule provided no definition of 
``minor,'' and the Department indicated in the preamble to the proposed 
rules that it had been unable to develop a uniform standard for all 
substances and areas encompassed by the models. See 59 FR at 40330 and 
63313. However, after reviewing the comments, the Department has 
concluded that given the significance of this term and the fact that 
type A procedures were intended as simplified procedures requiring 
limited analysis by trustees, it is appropriate to provide clear 
guidance.
    The Department evaluated a number of different approaches to 
defining ``minor.'' First, the Department reviewed the language and 
legislative history of CERCLA. The Senate Report that accompanied the 
predecessor bill to CERCLA states:

    [A] simplified type of regulation is necessary to effectively 
deal with damage assessment in most ``minor'' releases of hazardous 
materials * * *. The other type of regulations [type B] would be 
employed in large or unusually damaging releases and would be used 
to guide the site-specific damage assessment. S. Rep. No. 96-848 at 
86.

However, nothing in the legislative history indicates what Congress 
meant by ``minor.''
    Next, the Department considered basing the definition on the 
technical limitations of the NRDAM/CME and the NRDAM/GLE for modeling 
large or highly toxic spills. However, sensitivity analyses of the 
models failed to reveal any clear stages at which the model assumptions 
became invalid.
    The Department then considered relying upon existing standards 
developed in other contexts of environmental law. The U.S. Coast Guard 
has developed a volume-based system for classifying oil spills for 
purposes of spill response. See 40 CFR 300.5 (minor discharge of oil is 
one of less than 1,000 gallons to inland waters or 10,000 gallons to 
coastal waters). The U.S. Environmental Protection Agency (EPA) is 
responsible for developing a parallel rating system for hazardous 
substance spill response. While EPA has developed a qualitative system, 
this system does not provide the type of clear, quantitative limits 
that the Department believes are needed in this context. See 40 CFR 
300.5 (minor release of hazardous substance is one that poses minimal 
threat to public

[[Page 20581]]

health or welfare of the U.S. or the environment).
    Next, the Department considered basing a definition of minor on the 
point at which type B procedures can no longer be performed at a 
reasonable cost. However, because trustees have rarely pursued damage 
claims for smaller spills, the Department was unable to develop 
reliable estimates of the cost of conducting type B procedures in such 
cases.
    Therefore, the Department was left to make this policy decision 
about the upper limit on applicability of type A procedures without the 
benefit of clear empirical standards or legal precedents. The 
Department has chosen to base this limit on its sense of when it is no 
longer ``fair'' to allow trustees to obtain a rebuttable presumption 
using the NRDAM/CME or NRDAM/GLE as opposed to performing type B 
procedures. The Department believes that, given the current level of 
experience with these models, $100,000 represents a reasonable cut-off 
for their use. As more experience is gained with these models, the 
Department will reconsider this cut-off in future biennial reviews. 
Further, because this regulatory cut-off is based on considerations of 
fairness rather than the inherent reliability of the models, the 
Department wishes to emphasize that although use of the models to 
calculate damages above $100,000 is not entitled to a rebuttable 
presumption, such use may nonetheless be appropriate in other contexts, 
such as settlement negotiations or litigation without the benefit of 
the rebuttable presumption.
    Finally, the Department recognizes that in some instances the 
models may project damages in excess of $100,000, yet it may not be 
reasonable to perform type B procedures. The Department believes that 
trustees should be allowed the option of claiming damages up to 
$100,000 in such cases instead of forgoing all compensation. Therefore, 
the Department has eliminated the proposed rule condition that type A 
procedures only be used for minor releases and instead imposed a cap on 
the level of damages that trustees can claim through use of a type A 
procedure and still obtain a rebuttable presumption.
    Comment: Several commenters thought that the proposed condition 
requiring uniform subsurface currents would render the NRDAM/CME 
inapplicable to all spills in Massachusetts, New Hampshire, and Maine.
    Response: The Department acknowledges that the condition regarding 
subsurface currents may limit the applicability of the NRDAM/CME in 
some circumstances but notes that it does not render the model 
inapplicable to all locations where subsurface currents are not 
uniform. The models use vertically averaged currents and assume that 
the speed and direction of horizontal transport is uniform over depth 
at a specific latitude and longitude. The models do include randomized 
motion in the vertical dimension, but not directed motion. The 
vertically averaged current is essentially a current that provides the 
correct net transport averaged vertically. If the transport of the 
released substance cannot be reasonably represented by a vertically 
averaged current, then the NRDAM/CME's projections may not be reliable. 
For example, substances with high densities, such as sulfuric acid, may 
sink rapidly through the water column so that the principal mass is 
transported in the direction of the subsurface current. However, in 
many cases, such as when a substance remains at or near the surface or 
sinks slowly, subsurface currents will not affect the fate of the 
spilled substance. In these cases the model can reliably predict 
damages. Therefore, the rule allows trustees to use the NRDAM/CME, even 
if subsurface currents are not uniform, so long as they are not 
expected to significantly affect the level and extent of injuries.

D. User-Supplied Information

    Comment: Several commenters suggested changes to the proposed 
models and rules that would require trustees to confirm injury. These 
commenters asserted that the proposed models merely assume injury and 
that the proposed rules inappropriately failed to require field 
verification of this assumption. The commenters noted that CERCLA 
limits recovery to damages that ``result from'' a release. These 
commenters argued that this limitation requires that trustees conduct 
field studies that prove that an injury actually occurred and that it 
was caused by the release in question.
    A number of commenters thought that Congress intended traditional 
tort law standards of causation to apply to natural resource damage 
cases and cited case law in support of this position. Some commenters 
noted that Ohio v. Interior rejected a challenge to the Department's 
strict acceptance criteria for determining injury under the type B 
procedures and upheld the Department's interpretation that CERCLA 
adopted traditional causation standards. 880 F.2d at 471. The 
commenters stated that CERCLA does not create a different standard of 
proof of causation when type A procedures, as opposed to type B 
procedures, are used.
    Several commenters observed that CERCLA calls for type A procedures 
that involve ``minimal'' rather than ``no'' field observation. The 
commenters thought that the Department was engaging in sheer 
speculation when it asserted in the August 8, 1994, notice of proposed 
rulemaking that requiring confirmation of injury in type A assessments 
would be unduly burdensome. Finally, the commenters stated that none of 
the steps that trustees must take before applying a type A model, 
including the Preassessment Screen Determination, satisfy the required 
standard of causation.
    Response: The type A models do not ``assume'' that injury occurs. 
Using both the information provided by the trustees and the biological 
and environmental information about the spill site contained in the 
model databases, the models perform millions of calculations to 
determine whether or not the release has caused an injury. The models 
project the distribution of the released substance over space and time, 
track the changing toxicity of the substance over that space and time, 
and simulate the movements of biota throughout the area around the 
release. The models only conclude that injury has occurred if biota are 
exposed to the released substance at concentrations and durations that 
exceed acute toxicity thresholds. If such thresholds have not been 
exceeded, the models conclude that there has been no injury. The models 
can and have projected that no injury resulted from particular 
releases. In such cases, the models determine that damages equal zero.
    The issue is not whether the Department is attempting to excuse 
trustees from their legal requirement to prove causation; the 
Department agrees that trustees must demonstrate that injury resulted 
from a release. Rather, the issue is whether trustees should be allowed 
to use the type A models to make that demonstration. The Department 
believes that it is appropriate to allow trustees to use a type A model 
to demonstrate injury without on-site verification of the model 
projections.
    There is a tension between the statutory provision requiring 
trustees to demonstrate that injury ``resulted from'' a release and the 
provision requiring the development of simplified assessment procedures 
that involve ``minimal field observation.'' As noted in the cases cited 
by commenters, the requirement that trustees demonstrate that injury 
resulted from the release indicates that

[[Page 20582]]

Congress intended that natural resource damage liability be 
compensatory. PRPs are to be held liable not just because they are 
responsible for a release but because they are responsible for a 
release that caused an adverse effect. On the other hand, by requiring 
the development of type A procedures, Congress recognized that 
assessment work can be expensive and time-consuming. In the case of 
minor releases, it is often not cost-effective or feasible to conduct 
more than minimal field observations. The Department has struggled to 
resolve the tension between these two statutory requirements by 
developing type A procedures that rely on computer models to predict 
actual site-specific effects to the maximum extent practicable but do 
not require on-site verification of the models' injury predictions.
    Although the regulations do not require trustees to verify the 
injury projections of the type A models, they do require trustees who 
use a type A model to make several determinations that require field 
observations. Before trustees can proceed with a type A procedure, they 
must perform a Preassessment Screen Determination in which they 
establish if assessment work is warranted. Existing Sec. 11.23(e) 
requires trustees to make a preliminary determination that the 
following conditions have been met:

    A release of a hazardous substance has occurred; Natural 
resources under their trusteeship have been or are likely to have 
been adversely affected;
    The quantity and concentration of the released substance is 
sufficient to potentially cause injury;
    Data sufficient to pursue an assessment can be obtained at a 
reasonable cost; and
    Response actions will not sufficiently remedy the injury.

    If the trustees determine that these conditions are met, they must 
then demonstrate that the released substance entered the water, which 
is the only pathway considered by the models. Trustees must develop 
information on ambient environmental conditions at the site of the 
release and on the extent of response actions. Also, trustees must 
evaluate whether potentially affected resources are addressed by the 
model.
    The Department agrees that these determinations in and of 
themselves do not establish causation. They are not intended to do so; 
causation is demonstrated by the models. However, these determinations 
do fulfill the standard of minimal field observations. The Department 
notes that Congress chose to use the word ``observations'' rather than 
``studies,'' ``surveys,'' or ``analyses.'' Therefore, the Department 
interprets ``minimal field observations'' to be information that is 
readily or routinely collected following a release. The Department 
rejects the argument that the language of CERCLA requires on-site 
verification of causation, since observations would often be inadequate 
to determine causation. For example, the released substance may sink or 
disperse into the water column, precluding visual methods for 
documenting pathways and exposure. Furthermore, the potentially exposed 
resources may be difficult and costly to sample (e.g., endangered 
species, marine mammals, or subtidal organisms). In other instances, 
the persistence of the substance, or the remote location of the 
release, may prevent trustee scientists from reaching the spill site in 
a timely manner to conduct assessment work.
    Moreover, today's final rule does not establish a new standard for 
proof of causation. The regulations did not require trustees to verify 
the injury projections of the original NRDAM/CME. The existing 
regulations require trustees to confirm exposure before implementing 
type B studies, but not type A procedures. Today's final rule does 
include language in Sec. 11.31(c)(1) clarifying that the confirmation 
of exposure requirement only applies to type B procedures. The original 
type A rule contained the same substantive provision, only worded 
differently. See 52 FR at 9064. This final rule simply rewords and 
relocates the existing provision.
    Even when trustees use type B procedures, there are some 
circumstances under which the existing regulations allow them to use 
models to determine injury to surface water, groundwater, and air. See 
43 CFR 11.64(b)(6), 11.64(c)(8), and 11.64 (d)(2). The existing 
regulations also allow trustees who perform type B procedures to use 
models to demonstrate that a groundwater or air pathway exists between 
the site of the release and injured biological resources. See 43 CFR 
11.63(c)(5)(ii)(C) and 11.63(d)(4). Demonstration of a pathway is an 
integral part of establishing causation.
    The Department acknowledges that the existing regulations do not 
explicitly allow trustees who use type B procedures to demonstrate 
biological injury based on models alone. However, as Ohio v. Interior 
recognized, the language and legislative history of CERCLA are 
ambiguous as to the standard of proof of causation. 880 F.2d at 470. 
Nothing in the statutory language prohibits the Department from 
establishing different standards of proof of causation under type A and 
type B procedures, and the Department believes that in light of the 
statutory description of type A procedures, different standards are 
warranted. In fact, as discussed above, the legislative history of 
CERCLA suggests that the Department would have been justified in 
developing a look-up table or compensation formula as a type 
Aprocedure. Instead, the type A models use both site-specific 
information provided by the trustees and biological and environmental 
information about the spill site contained in the databases to 
approximate more precisely the actual effects of the release.
    Finally, the Department notes that the final rule has been revised 
to require trustees to perform a preliminary application of the model 
and make the results available for public comment before presenting a 
damage claim. Therefore, PRPs will have an opportunity to evaluate the 
injury projections. They can then decide whether they have information 
that indicates that the projections are wrong and that the user inputs 
need to be modified or that type B procedures should be used.
    Comment: The Department received several comments on the wind 
inputs. Some commenters expressed concern about the adequacy of the 
proposed rule language allowing trustees to supply one set of wind data 
for a 30-day period. These commenters noted that wind data were 
critical to correct functioning of the models and requested that 
trustees be required to supply actual hourly data. One commenter noted 
that such data are readily available from the National Climatic Data 
Center. These commenters also thought that actual wind data should be 
supplied for the entire duration of the model application because of 
gross oversimplifications in the data supplied by the models when the 
user-supplied wind data run out. One of the independent technical 
reviewers suggested that users be allowed to enter wind data in their 
choice of units.
    Response: The Department agrees that one set of wind data for an 
entire 30-day period may not be adequate in all cases. Trustees are 
free to supply hourly wind data; however, the Department also believes 
that requiring trustees to do so in all cases would be onerous. 
Therefore, the Department has modified the final rule to require 
trustees to supply data on prevailing wind conditions for each day of 
the 30-day period. Recognizing that the type A procedures were intended 
to provide simplified procedures requiring minimal analysis by 
trustees, the Department has concluded that it is

[[Page 20583]]

appropriate to establish a uniform time frame for entry of wind data. 
Wind speed and direction are most relevant to the simulation of the 
surface trajectory. Released substances will generally only float for a 
few days or weeks before they sink or go ashore. Therefore, the 
Department believes that 30 days is an appropriate time frame for the 
vast majority of releases to which the type A models will be applied. 
Users are free to supply more than 30 days worth of wind data if they 
choose but are not required to do so. If a simulation continues past 30 
days and the user has only supplied 30 days worth of data, the models 
will supply climatological wind data. With regard to the units of 
measurement for wind data, the Department notes that users are allowed 
to enter wind data measured either in knots or in meters per second. 
The Department believes this provides users with appropriate 
flexibility.
    Comment: Several commenters addressed the currents inputs to the 
proposed NRDAM/CME. Some commenters, including a few of the independent 
technical reviewers, found it difficult to enter currents data and 
suggested that default values be made available. On the other hand, one 
commenter thought that the tool for supplying currents data in the 
proposed NRDAM/CME was already too simplistic. One of the independent 
technical reviewers noted a ``bug'' in the program.
    Response: Currents have a profound impact on the physical fate of 
spilled substances and are highly variable. Provision by the model of a 
single set of default values for currents would adversely affect the 
reliability of the model. Therefore, the Department believes it is 
appropriate to require users to supply some level of site-specific data 
on currents. However, the Department is also committed to ensuring that 
the NRDAM/CME remain accessible to a wide range of potential users and, 
thus, recognizes the need to avoid excessively complicated user inputs. 
The Department has revised the currents entry tool to make it easier to 
use and to correct the ``bug.'' The Department has also provided 
additional guidance on developing and entering currents files in Volume 
II of the NRDAM/CME technical document.
    Comment: Some commenters objected to the proposed provision 
allowing trustees to decide whether or not to have the models consider 
ice cover. These commenters stated that the rule should require 
trustees to have the models consider ice cover if ice is present during 
or after the release.
    Response: The Department believes that the ice model contained in 
the NRDAM/CME and NRDAM/GLE reasonably reflects average ice conditions. 
However, the Department has modified the final rule to provide that 
when trustees have reliable evidence that ice was not present at the 
spill site, they must disable the ice modeling function.
    Comment: A few commenters addressed the data inputs for response 
actions. One commenter thought that the models should take into account 
the effects of spill prevention and containment measures required under 
OPA and other laws. Some commenters stated that trustees should only be 
required to supply the volume of the released substance that was 
removed during the first 24 hours after the release. These commenters 
noted that the types of acute effects considered by the models should 
have occurred within that first 24-hour period.
    Some commenters thought that limiting the data inputs to the first 
24 hours would also alleviate the problem with the NRDAM/CME identified 
in the December 8, 1994, preamble. In that preamble, the Department 
noted that there may be cases where the proposed NRDAM/CME would not 
subtract the full volume removed even though users had provided full 
and accurate information about removal actions. This problem arose 
because the proposed model required users to specify the location and 
time frame of the removal. The proposed model then subtracted the mass 
removed at the time and location specified by the user. If the user 
specified that mass was removed from a location before the time that 
the model projected the substance would reach that location, then the 
model was unable to subtract any removed mass. See 59 FR at 63307-08. 
Another commenter expressed concern about this problem and suggested 
modifying the NRDAM/CME so that in such a situation the volume of the 
released substance actually removed would be subtracted from the 
nearest location where the model predicted that an equivalent volume of 
the substance could be found at that time.
    Response: The final rule and models do account for the effects of 
successful spill prevention and containment measures. The models 
calculate damages only for the volume of substance that entered the 
water, was not removed, and caused injury. Therefore, any material 
prevented from entering the water as a result of voluntary or mandatory 
spill prevention or containment measures would not be considered by the 
models.
    With regard to entry of data on removal that occurred more than 24 
hours after the initial release, the final rule requires trustees to 
specify the time of the removal. Therefore, even for injuries that do 
occur within the first 24 hours, entry of the total volume removed 
would not result in an underestimate of damages because the models will 
take into consideration that the removal did not occur entirely within 
the first 24 hours. The Department assumes that the commenter's concern 
is that the models will underestimate damages if they subtract the 
entire volume removed because only the removal during the first 24 
hours would reduce the likelihood of acute injury. However, not all 
direct effects considered by the models will occur within the first 24 
hours after the release.
    To address the problem identified in the December 1994 preamble, 
the Department has modified the final rule language addressing the 
required data input for response actions. Appendix II now states that 
when developing the data input on response actions, trustees must 
specify a geographic area that encompasses the entire surface water and 
shoreline area over which the spilled substance was likely to have 
spread. This requirement should ensure that the NRDAM/CME will subtract 
the full volume of spilled material that was removed during response.
    Comment: Numerous commenters addressed whether users should be 
allowed to modify the model databases. One commenter suggested that the 
models would be too labor-intensive if trustees were expected to edit 
numerous databases. However, most commenters supported retaining the 
habitat editor in the final version of the models, noting the 
prevalence of default values in the habitat database. Some commenters, 
including one of the independent technical reviewers, stated that the 
Department should continuously update the model databases to ensure 
that they reflect current information. Other commenters expressed doubt 
that the Department could conduct such updates in a timely manner. 
These commenters thought that model accuracy would be increased if 
trustees were allowed to edit not only the habitat designations but 
also other data such as species biomass, baseline fishing mortality 
rates, commercial fish prices, and restoration costs. These commenters 
noted that PRPs would be protected from potential misuse of the editing 
feature by trustees because the input data would be subject to 
challenge. Some commenters recommended various mechanisms for allowing 
users

[[Page 20584]]

to substitute more precise or up-to-date site-specific information. For 
example, one commenter suggested creating an administrative process for 
obtaining variances from the model parameters or data.
    Response: The Department acknowledges that allowing users to revise 
the models' databases would enable fine-tuning to better reflect site-
specific conditions. On the other hand, Congress specifically mandated 
the development of type A procedures to simplify assessments and 
minimize fieldwork. The more trustees are expected to edit the model 
databases, the less the type A procedures fulfill this mandate and the 
closer such procedures approach the data-gathering requirements of type 
B procedures.
    Therefore, the Department has decided to allow trustees to modify, 
under some circumstances, the models' default values for water 
temperature, total suspended sediment concentrations, mean settling 
velocity of suspended solids, air temperature, and habitat type. 
However, trustees may not make any additional modifications to the 
databases if they intend to obtain a rebuttable presumption.
    The proposed rules included provisions allowing trustees to supply 
site-specific values for water temperature, total suspended sediment 
concentrations, mean settling velocity of suspended solids, and air 
temperature. The Department continues to believe that these parameters 
are highly variable and can profoundly affect the physical fate of 
released substances. Therefore, trustees should be allowed to change 
the default values for these parameters if they have more accurate 
data.
    The Department has also concluded that retention of the habitat 
editor is appropriate given the importance of habitat designations to 
the total damage figure and the prevalence of default habitat 
designations in the models. Also, despite specific solicitations in 
both notices of proposed rulemaking, relatively few commenters supplied 
revised habitat information to the Department and some State commenters 
specifically stated that they had been unable to review the habitat 
designations.
    The Department does not believe that the habitat editor requires 
excessive effort to operate. First, trustees are not required to edit 
habitat designations. Second, while the task could be substantial if a 
single user attempted to edit the large regions covered by the models, 
it is not so labor intensive for the area actually affected by any 
single spill. The Department also notes that trustees often perform 
habitat mapping as part of pre-spill planning, which should further 
expedite habitat editing. Once edited, the revised habitat designations 
may be saved within the models and used again for future model 
applications. However, trustees who save such redesignations would 
still need to justify those redesignations in any future Assessment 
Plans.
    With regard to other model databases, the Department believes that 
allowing additional editing would undermine Congress' intent for 
developing type A procedures. Allowing other edits would require users 
to make additional conforming changes that would complicate use of the 
models. For example, changing the fisheries biomass or parameters would 
require recalculation of egg and larval abundance using the model 
equations. Changing wildlife abundances would require recalculation of 
lost wildlife viewing values.
    The final rule no longer requires use of the type A models. 
Therefore, when trustees have, or are provided with, evidence that the 
model databases are inaccurate for a specific incident, they are free 
to use type B procedures, provided they can do so at a reasonable cost. 
When trustees or PRPs already have site-specific information indicating 
that model data are inaccurate, the cost and effort associated with 
conducting type B procedures, and in turn the need for a type A 
procedure, should be reduced.
    The Department does not have the resources available to support an 
administrative process for reviewing petitions for variances from the 
computer model parameters or databases. However, the Department is 
statutorily required to review and update the models every two years 
and, thus, will be incorporating more up-to-date information as it 
becomes available. Also, the Department notes that the models already 
update the compensable value and restoration cost databases to account 
for inflationary effects through application of the implicit price 
deflator. Finally, the Department notes that while the results of model 
runs made with customized changes beyond those identified in the rule 
would not receive the rebuttable presumption for damage claims made 
under CERCLA, they may nonetheless be reliable and useful in other 
contexts, such as settlement negotiations or litigation without the 
benefit of the rebuttable presumption.
    Comment: A few commenters addressed the implicit price deflator 
data input. In response to the Department's solicitation of comment on 
whether the proposed rules should be modified to require trustees to 
supply the implicit price deflator for the Gross Domestic Product (GDP) 
instead of that for the Gross National Product (GNP), one commenter 
indicated that use of the implicit price deflator for the GNP should be 
retained. Another commenter stated that the Department should change 
the base year from 1987 to 1992.
    Response: The Department has decided to retain the GNP implicit 
price deflator as the index with which to adjust past dollar amounts to 
current dollar equivalents. Because GNP refers to income that is 
available to U.S. residents, it is appropriate for analyses that are 
related to the use of that income, such as expenditures for 
environmental restoration. GDP, on the other hand, refers to income 
that is derived from production within the U.S., regardless of whether 
that income is available to U.S. residents or accrues to non-U.S. 
residents. For more information regarding GNP and GDP, readers are 
referred to the August 1991 issue of the Survey of Current Business 
available from the Bureau of Economic Analysis, U.S. Department of 
Commerce. Also, the Department has updated the index numbers in the 
models to accommodate the change in base year from 1987 to 1992.
    Comment: One of the independent technical reviewers questioned the 
user inputs to the proposed NRDAM/CME concerning boat closures. The 
technical reviewer thought that it might be difficult for trustees to 
estimate the number of boats affected by the closure.
    Response: As discussed in Section VII.P of this preamble, the 
Department has eliminated the calculation of damages for lost boating 
from the NRDAM/CME.
    Comment: A few commenters responded to the Department's 
solicitation of comment on whether users should be allowed to supply a 
site-specific discount rate and, if so, how they should determine the 
correct rate. These commenters stated that the models should use a 
fixed discount rate but that the fixed rate should be three percent 
rather than the seven percent rate included in the proposed models.
    Response: The Department believes that the appropriate discount 
rate is the consumer's rate of time preference for natural resource 
services. This is the rate at which individuals are willing to trade 
natural resource services today for similar natural resource services 
in the future. The Department further believes that the real 
(inflation-adjusted) rate of return on U.S. Treasury bills is a 
reasonable proxy for this rate of time preference. An analysis of real 
rates of return on U.S. Treasury bills reveals an

[[Page 20585]]

indicated annual discount rate of three percent. Therefore, the 
Department agrees with the commenters that future lost use values 
should be discounted at a three percent rate. This discount rate has 
been incorporated in the models.
    The Department believes that use of a fixed discount rate is 
appropriate in the context of a simplified procedure. Further, the 
Department believes it is unlikely that the rate will change 
significantly over the next two years. During the biennial review, the 
Department will reexamine this issue.

E. Physical Fates

    Comment: Several commenters, including some of the independent 
technical reviewers, thought the physical fates submodel was well 
developed and well tested. Another commenter stated that the proposed 
models produced spill trajectories that resembled actual spill events. 
Conversely, one commenter experienced difficulty with the models' 
trajectory component noting that, during some trial runs, the spill did 
not move.
    Response: The Department acknowledges and appreciates the 
supportive comments concerning the physical fates submodel. Physical 
fates modeling is a technical discipline that has received extensive 
study. In reply to the trajectory difficulty, the Department notes that 
the model's spill trajectory is dependent upon the user's entry of wind 
and current data. Spilled material does not move during simulations if 
the user does not supply wind and current data.
    Comment: Several commenters, including some of the independent 
technical reviewers, were disappointed by coarse resolution of most 
habitat grids in the NRDAM/CME. These commenters complained that the 
limited resolution results in the loss of consideration of critical 
shoreline habitats.
    Response: The Department has improved the resolution in the NRDAM/
CME by a factor of four through use of a new compiler and additional 
memory.
    Comment: Several commenters claimed the proposed models used 
inadequate and nonrepresentative data on the physical and chemical 
properties of hazardous substances and oil and provided references to 
additional data sources. The commenters stated that the chemical and 
physical data on gasoline and diesel oil, particularly the sulfur 
content in diesel oil and vapor pressures in gasoline, do not 
accurately represent the products in use today. One of the independent 
technical reviewers stated that the relative toxicities used in the 
models for No. 6, No. 2, and crude oils do not agree with experiment 
results reported in the literature. Several commenters also concluded 
the Department inappropriately applied the same degradation rates to 
the nine oils included in the models (except for No. 2 diesel oil). 
They could not substantiate the model degradation rates, and concluded 
they were invalid. One of the independent technical reviewers suggested 
that the Department update the parameters for oils. Some commenters 
also thought that the models relied on overly simplistic and outdated 
modeling techniques.
    Response: The Department found the physical and chemical parameters 
included in the models for hazardous substances to be in agreement with 
commenters' independent literature search, except with regard to the 
partition coefficient for epichlorohydrin. After reviewing the 
literature, the Department has decided to substitute the proposed 
parameter for this substance.
    The Department has also increased the number of oils and petroleum 
products included in the models to a total of 33 and has revised the 
physical and chemical parameters for oils based on the most recent 
literature. The degradation rates for oils in the models apply to 
acutely toxic low molecular weight components. Thus, they are 
consistent for all oils. The Department reviewed the accuracy of these 
data and documented all sources in the technical documents when 
revising the oil database. For further information, see Section 2, 
Volume III of the NRDAM/CME and NRDAM/GLE technical documents.
    The Department does not agree that the modeling techniques used are 
either outdated or overly simplistic. The Department has never intended 
the physical fates submodel, or any of the other submodels, to provide 
a comprehensive treatment of all known physical, chemical, and 
biological processes occurring in aquatic environments, nor is such 
treatment necessary for the limited purposes of the type A procedures. 
Instead, the modeling techniques employed are intended to reasonably 
approximate the most relevant processes pertaining to the fates and 
effects of spills that occur in aquatic environments based on readily 
available user input data.
    Comment: One commenter thought that the models overstated wildlife 
injuries by inappropriately treating chemicals less dense than water as 
slicks. The commenter also stated that the proposed models' predictions 
of oil and chemical slicks failed to account for many of the physical 
processes that affect the size of the surface slicks.
    Response: The models do treat all chemicals that are lighter than 
water as ``slicks.'' However, chemicals that are very soluble, such as 
ethanol and ammonia, will very quickly mix into the water column. Thus, 
chemicals that are highly soluble do not remain on the surface long 
enough to have any direct effect on wildlife at the water surface.
    The Department disagrees that the models do not adequately account 
for the physical processes affecting surface slicks. The major 
processes affecting the size of surface slicks are spreading, 
evaporation, and entrainment, all of which are simulated in the models. 
For further discussion, see Sections 3.3 through 3.5, Volume I of the 
NRDAM/CME and NRDAM/GLE technical documents.
    Comment: One commenter referred the Department to the Ohio spill 
database as an additional source of information.
    Response: The Department took this database into consideration when 
determining the types of spills to use for the sensitivity analysis of 
the NRDAM/GLE.
    Comment: One commenter questioned whether the use of varying grid 
sizes makes the NRDAM/GLE less accurate for spills in areas of large 
grid size.
    Response: The Department designed the grid sizes to represent the 
Great Lakes habitats at a resolution required by the local spatial 
variability. Areas with more spatial variability have smaller grids and 
higher resolution. The areas with large grid sizes are the open lakes 
where fine detail is not necessary. Thus, the NRDAM/GLE is not less 
accurate in the areas where grids are largest.
    Comment: One commenter questioned why the Department did not 
incorporate the U.S. Army Lake Survey grid in the NRDAM/GLE.
    Response: Development of the NRDAM/GLE grid required the 
representation of spatially varying habitats and depths as well as the 
contours of the connecting channels within a regular rectangular grid 
system. The U.S. Army Lake Survey grid does not consistently meet this 
requirement. Moreover, the Department has no reason to believe that use 
of the U.S. Army Lake Survey grid would significantly improve the 
reliability of the NRDAM/GLE.
    Comment: A few commenters, including one of the independent 
technical reviewers, thought that the Department should use three-
dimensional hydrodynamics models. Several commenters thought the

[[Page 20586]]

proposed type A models unrealistically assumed that currents were 
uniform with depth; unjustifiably failed to incorporate three-
dimensional currents modeling; and inappropriately failed to account 
for the effects of wind-driven turbulence mixing processes that 
increase mixing as high winds make the water surface rough. One 
commenter stated that a three-dimensional hydrodynamics model was 
needed to account for the effects of seiches (i.e., occasional and 
sudden oscillations of the water of lakes, bays, or estuaries caused by 
wind or changes in barometric pressure).
    Response: The NRDAM/CME and NRDAM/GLE use a three-dimensional 
transport model. The transport model assumes that currents are uniform 
vertically at each horizontally defined location. If the Department did 
include a three-dimensional current dynamics model in the NRDAM/CME and 
NRDAM/GLE, then users would need the assistance of expert hydrodynamic 
modelers to operate the type A models. For example, three-dimensional 
current dynamics models are dependent on physical forces at the 
boundaries of the modeled area, which the user would be required to 
enter. The Department believes that imposing such complex modeling 
requirements would conflict with the statutory directive to develop 
simplified assessment procedures.
    The models utilize vertically averaged currents with the assumption 
that horizontal transport is uniform in speed and direction over depth 
at a given location in horizontal space (i.e., at a given latitude and 
longitude). One of the conditions for use of the models is that 
subsurface currents either are not expected to significantly affect the 
level and extent of injuries or are reasonably uniform with depth in 
the area of the spill. The models also include randomized motion in the 
vertical dimension, but not directed motion in the vertical. Thus, the 
currents carrying spilled material must be representable in a manner 
consistent with this assumption of the models.
    The vertically averaged current is essentially a current that 
provides the correct net transport, averaged vertically. If the 
transport of spilled material cannot be reasonably represented by a 
vertically averaged current, the condition for use would not be met and 
the model would not be applicable. However in many cases, three-
dimensional representation of current dynamics would not significantly 
change the damages calculated by the models. For example, for 
substances of low density, such as toluene, that remain at or near the 
surface, the present current dynamics model adequately addresses 
physical fates.
    The Department recognizes that seiches occur in the Great Lakes. 
The Department has not, however, included such processes in the NRDAM/
GLE because these processes are not a significant transport process for 
determining the physical fate of spilled substance in the Great Lakes. 
Further, any change in the location of a shoreline brought about by a 
seiche is likely to be small and should not have a significant effect 
on the injuries calculated by the model.
    Comment: One commenter thought that the models should not include 
liquid asphalt. The commenter noted that liquid asphalt hardens and 
sinks to the bottom quickly, and can be completely removed by dredging.
    Response: The Department agrees that the models are not designed to 
estimate damages for substances such as liquid asphalt. The models 
assume that oils and petroleum products float initially, although they 
may subsequently entrain, adsorb to particles, and sink. Liquid asphalt 
does not act in this manner and, thus, has not been included in the 
databases.
    Comment: One commenter criticized the NRDAM/CME's treatment of ice, 
stating that the model inappropriately assumes that ice is always a 
solid mass.
    Response: The model does not assume that ice is always a solid 
mass. For a discussion of how the models treat ice, see Section 3.11, 
Volume I of the NRDAM/CME and NRDAM/GLE technical documents.
    Comment: Some commenters thought a smooth function relating 
requisite thickness to spillet diameter would be more realistic than 
the step function proposed in the models.
    Response: Not enough quantitative information is available to 
develop a smooth function. Available data are, in fact, in the form of 
a step function. Further, given the available data and the steepness of 
the relationship between mortality and dose in the pertinent range, the 
form of the function would not significantly affect the reliability of 
the model calculations.
    Comment: One of the independent technical reviewers noted that the 
proposed NRDAM/CME treated wetlands as water cells and questioned why 
they were not oiled as other shorelines.
    Response: Wetlands are either fringing or extensive in the model. 
``Fringing'' wetlands are those which form narrow wetlands along 
shorelines. ``Extensive'' wetlands, on the other hand, are those 
sufficiently large to encompass a majority of the grid cell. For 
fringing wetlands, oiling occurs in the same manner as for other 
shorelines. For extensive wetlands, slicks keep moving across the area, 
as they do in water, but they can oil wetland biota in the same way as 
they do in fringing wetlands. Oil may accumulate in sediments in all 
cells, by partitioning onto suspended sediments and sinking.

F. Species Distribution and Abundance

    Comment: Several commenters thought the default habitat 
designations in the NRDAM/CME generally provided an inadequate 
representation of coastal and marine habitats. Others recommended 
specific changes such as including habitat data available in existing 
or upcoming studies. One commenter provided specific habitat data for 
New York and New Jersey. Another commenter requested that the 
Department use information that would be available from the upcoming 
Texas Natural Resource Inventory.
    Response: As stated in the notices of proposed rulemaking, the 
Department recognized the shortcomings of the default habitat 
designations and specifically solicited comment on those designations. 
See 59 FR at 40330 and 63314. The notices also provided technical 
instructions on transmitting information to the Department about 
suggested changes to the default habitat designations. The Department 
has reviewed and revised the habitat designations in the NRDAM/CME 
based on technical data provided to the Department for the coastal and 
marine waters of New York and New Jersey. The Department did not 
receive other specific recommendations for changing the default habitat 
designations within the format requested.
    The Department recognizes that additional habitat data are 
continually becoming available. However, the Department needed to 
finalize the data in the models and chose not to delay issuance of the 
models to incorporate recent or upcoming studies, such as the Texas 
inventory. During future biennial reviews, the Department will update 
the habitat designations to reflect newer information. Meanwhile, 
today's final rule allows trustees to change the default habitat 
designations. As further studies provide better data, trustees may 
substitute such data for the default values included in the models 
using the habitat editor function.
    Comment: One commenter stated that shoreline types in the NRDAM/GLE 
should include a cohesive (clay) component. The commenter also thought 
that rocky shoreline should be

[[Page 20587]]

changed to either rocky bluff or cobble beach.
    Response: The Department does not believe that changes to the 
shoreline types are necessary. The suggested changes would have little 
impact on the reliability of the damage figure since biological 
injuries are not calculated based on shoreline designations. The models 
use shoreline type to approximate the oil retention on shorelines. 
Further, the distinct holding capacities for these shoreline types are 
unavailable but could be expected to fall within the holding capacities 
of the shoreline types already represented within the NRDAM/GLE.
    Comment: One commenter provided maps of habitat types for Michigan 
and suggested that wetlands in the connecting waterways be hand-edited.
    Response: Although the Department appreciates the effort provided 
by the commenter, the grid scale of the information was of a much 
smaller resolution than that contained in the NRDAM/GLE and could not 
be directly applied to revisions of the habitat grids. The Department 
believes the NRDAM/GLE habitat designations are consistent with the 
maps provided, were such information consolidated at the larger NRDAM/
GLE grid scale. As a result, the default habitat designations were not 
revised in the NRDAM/GLE, and the Department has not hand-edited the 
habitat maps of the connecting waterways.
    Comment: Some commenters, including one of the independent 
technical reviewers, thought that the models should map the location of 
critical habitats, such as bird colonies or rare communities and 
plants.
    Response: The Department does not believe it is appropriate, in the 
context of developing simplified type A procedures, to attempt to map 
all critical habitats and rare communities throughout the entire 
geographic region covered by the models. The models were developed 
based on an assumed average abundance of biota by habitat within a 
biological province. The Department believes that if a release is 
expected to affect a critical habitat or rare community that is not 
adequately represented by the models, then use of type B procedures 
should be considered.
    Comment: A number of commenters, including some of the independent 
technical reviewers, identified and suggested the Department fill data 
gaps in fish abundance. Several commenters believed the fish abundance 
data in the models were based on commercial catch data that were not 
generated with sound scientific methods and are known to have low 
accuracy. The commenters stated that the abundance data were 
inconsistent with catch data provided by Federal and State fisheries 
agencies and, thus, did not account for variability in fish 
populations. The commenters questioned the level of care and effort 
that had gone into estimating total lake fish stock data in the NRDAM/
GLE.
    Response: The Department notes that fishery statistics are 
collected to fulfill a variety of different research and management 
needs. For the purposes of these models, fish biomass abundance was 
required. However, such data were not uniformly available for all 
species and all geographic regions covered by the models. As a result, 
the Department drew upon available data from State and Federal fishery 
management and research organizations and extrapolated where needed to 
fill gaps. The criteria used for the selection and use of available 
data are outlined in Section 6.3, Volume I of the NRDAM/CME technical 
document. For all stocks where the National Marine Fisheries Service 
(NMFS) or State agencies have performed a stock assessment, these stock 
sizes were used. The Department used biomass surveys if stock 
assessments were not available. The Department used catch data only if 
these other data sources were unavailable. Thus, the data are 
consistent with that collected by Federal and State agencies and 
represent the best available data for each species included. Since the 
most valuable species in terms of total catch are also the most 
studied, data are likely to be more accurate for valuable species. 
Damages resulting from less significant species in the catch are 
typically insignificant.
    Where the Department had to use catch data to estimate biomass, it 
used both commercial and recreational catch data compiled by NMFS, the 
Federal agency charged with assessing and regulating fisheries stock. 
These catch data are the best available source of information. The 
Department used commercial catch data as the sole source of information 
only where there was not a significant recreational fishery.
    The Department did attempt to account for variability in fish 
populations. The areal extent of a species and its seasonal movements 
are based on life-history information for that species. Some species do 
not in fact move seasonally. Those that do are indicated in the 
database. In some cases, catch from large areas is represented in 
smaller areas if the life history warrants. In other cases, the data 
are not supportive of regional specificity. Also, some stocks do vary 
annually to a significant degree. For stocks where data were available, 
averages for the most recent three years were used. The models are 
designed to represent an average year.
    Considerable care and effort went into estimating total lake fish 
stock data in the NRDAM/GLE, as documented in Section 3, Volume III of 
the NRDAM/GLE technical document. The Department considered both the 
data available and the species behavior in terms of habitats utilized 
by season.
    Comment: Some commenters, including a few of the independent 
technical reviewers, questioned the wildlife abundance data sources 
used for determining hunting and trapping losses. One commenter thought 
that some abundance data from one area were used inappropriately to 
represent abundances in other areas. For example, the commenter noted 
that the study by Onuf (1987) was used inappropriately to extrapolate 
bird abundances over the entire west coast; and that the study by 
Breuggeman (1989) was used inappropriately to extrapolate marine mammal 
abundances to both the California coast and to Galveston Bay. Another 
commenter questioned the references to Bellrose (1980), specifically as 
applied to trumpeter swans in Prince William Sound.
    Response: The Department has obtained additional sources of bird 
abundance data for the west coast and has incorporated those data into 
the wildlife abundance database of the NRDAM/CME. Also, the Department 
has revised the abundance data for Galveston Bay to include more recent 
data on dolphins. The Department has included additional osprey data 
and updated the eagle and harbor seal data. Further, the Department has 
deleted data for mysticetes (baleen whales) for provinces 4, 5, and 7. 
The Department applied data in Bellrose (1980) only where no other data 
were available. The Department has now replaced most references to 
Bellrose (1980) with more recent data. In particular, the Department 
has updated data for Prince William Sound based on a 1990 to 1995 
waterbird survey. For further information, see Section 5, Volume IV of 
the NRDAM/CME technical document.
    Comment: One of the independent technical reviewers thought that 
the definition of subtidal wetland was unclear and recommended that 
intertidal seagrass be added as a habitat type.
    Response: Subtidal wetlands are the subtidal shallow waters in and 
around extensive wetlands. The Department has amended the NRDAM/CME 
technical document to include this more precise

[[Page 20588]]

definition of subtidal wetlands. See Table 4.4, Volume I, of the NRDAM/
CME technical document. The Department has also included intertidal 
seagrass as an additional habitat type.
    Comment: A number of commenters, including one of the independent 
technical reviewers, thought that use of province-wide abundance 
figures was inappropriate given the sizes of the provinces. An 
independent technical reviewer suggested that provinces be subdivided. 
Another commenter stated that grid-specific wildlife abundances should 
be used where such data are available.
    For west coast wildlife densities, one commenter noted that the 
NRDAM/CME's provinces are not consistent with spatial strata within 
which seasonal densities of a species are relatively uniform. The 
commenter noted that this can result in wildlife densities that are 
orders of magnitude too high in some cases or too low in others.
    Another commenter pointed out that the estimates of wildlife 
seasonal densities for the west coast were uniformly high and the 
numbers generated by extrapolation of density may exceed entire world 
population for many species. The commenter provided recent survey 
density data for birds and mammals. One of the independent technical 
reviewers suggested that the Department update data on west coast bird 
abundances.
    Response: The wildlife data for the west coast and Gulf of Alaska 
were completely revised with more recent and actual survey data 
provided by one commenter. Wildlife data for other provinces were also 
updated with more recent information. These corrections have eliminated 
the extrapolation errors noted. The Department believes the revisions 
made to the database have sufficiently addressed the possible need for 
subdividing province-wide abundances. For further information, see 
Section 5, Volume IV of the NRDAM/CME technical document.
    Comment: One commenter thought the way east coast province-wide 
wildlife abundances were distributed to the available individual 
habitats within NRDAM/CME grids resulted in an underrepresentation of 
wildlife abundances within the habitat grids. The commenter noted that 
individual habitat grids cannot hold enough habitat area to add up to 
the assumed provincial totals. The commenter suggested adding a 
multiplication factor to correct for the underrepresentation of habitat 
area, and resulting proportionate underrepresentation of wildlife 
abundances.
    Response: The habitat grid will, by the fact that it is 
rectangular, never precisely represent actual shore length, which is 
curved. However, the Department notes that the NRDAM/CME's grid 
resolution has been increased by a factor of four and the habitat data 
for the area in question (provinces 11 through 13) have been revised 
based on information provided to the Department. These revisions have 
significantly improved the precision of the shore length estimate. The 
Department considered the commenter's suggestion of using a 
multiplication factor to correct shore width. However, such a 
multiplier would affect the manner in which the physical fates submodel 
addresses the oiling of shorelines and induce additional error into the 
calculations performed by the model. Thus, the multiplication factor 
method was not employed.
    Comment: A few commenters, including some of the independent 
technical reviewers, suggested that monthly rather than seasonal 
averages be used where such data are available. Commenters noted that 
use of monthly averages was particularly important for migratory 
species.
    Response: Adequate data do not exist at this time for most species 
to incorporate monthly averages. As more data become available, the 
Department will consider incorporating monthly averages in the models 
during future biennial reviews.
    Comment: One commenter questioned the reliability of the abundance 
data in the proposed NRDAM/CME and offered different data from State 
wildlife agencies and available literature. The commenter recommended 
the following changes to the databases: modification of data for pigeon 
guillemots and kingfishers in Puget Sound and the Straits of Juan de 
Fuca (province 51) based on 1994 census data by Washington State; 
deletion of data on puffins along the New Hampshire coast; modification 
of data on bald eagles and osprey in Maryland based on Maryland's 1994 
census; modification of data on pelicans and bald eagles in Florida 
based on the Florida Natural Resource Department's estimates; and 
addition of data on Loggerhead and Kemp's Ridley turtles for the Gulf 
of Mexico and Galveston Bay.
    Response: The Department acknowledges that some errors occurred 
during the compilation of the wildlife abundance database, and 
additional information and data sources have been identified as a 
result of the public's review of these models. In fact, the Department 
specifically sought the assistance of the public on the wildlife 
abundance data. See 59 FR at 40330 and 63314. The Department has made a 
number of changes to the databases as a result of the public comments 
and now believes that the abundance data are more reliable.
    In response to the specific data comparisons made by the commenter, 
the Department has updated the data for province 51 and revised seabird 
abundances for the northeast. Puffins are no longer included in 
province 2, but are present in province 3, offshore New Hampshire and 
Maine. The State of Maryland's 1994 census data are more recent and 
have been used. The Florida Natural Resource Department's 1994 
estimates for pelican and bald eagles are more recent and have been 
used. The Department has not included the data for Loggerhead and 
Kemp's Ridley turtles in the offshore Gulf of Mexico since these 
species are not found in waters greater than 200 meters deep (NOAA 
1985). Also, no data documenting abundances of these turtles in 
Galveston Bay are available. For further information, see Section 5, 
Volume IV of the NRDAM/CME technical document.
    Comment: A commenter thought the wildlife abundance data were 
generally biased.
    Response: The Department believes that the data used in the models 
represent the best available information collected by independent 
scientists and government agencies.
    Comment: One commenter noted that the assumption of even 
distribution and random movement of biota may not be true. For example, 
the commenter noted that fish eggs and larvae are not randomly 
distributed.
    Response: Although some biological populations may not be evenly 
distributed and may not move randomly across large areas encompassing 
multiple habitat types, the Department believes that, for purposes of 
these models, it is reasonable to assume that populations are randomly 
distributed within a single habitat type.
    Comment: Several commenters noted that the proposed wildlife 
mortality model assumed that a wildlife species redistributes itself 
uniformly over its habitat each and every day. The commenters thought 
this assumption was not justifiable and would tend to overstate 
wildlife mortality.
    Response: The Department believes that, for purposes of these 
models, it is reasonable to assume that a wildlife species 
redistributes itself uniformly over its habitat each day. Habitat area 
within the models is of a size on the order of an individual's actual 
home range. Each individual does tend to

[[Page 20589]]

cover its home range each day, either for feeding or territorial 
purposes. Thus, populations do redistribute themselves daily.
    Comment: A commenter noted that the models' assumption that the 
density of a species in its designated habitat is constant throughout a 
given province could result in overestimates of actual mortalities for 
those provinces with multiple grids.
    Response: The Department does not believe that this assumption 
leads to overestimates for provinces with multiple grids. The 
biological database assigns the densities of species on all appropriate 
designated habitats regardless of the number of grids contained in the 
province. However, losses occurring in one grid are not distributed 
across grid boundaries and, thus would not overestimate losses.
    Comment: A commenter thought the NRDAM/GLE technical document 
should specify the time frame duration for all assumptions about 
species density.
    Response: The model assumes that species densities are uniform by 
season. The abundance tables clearly specify that density figures are 
provided on a seasonal basis. See Tables III.3.17 through III.3.27 and 
III.3.40 through III.3.50, Volume III of the NRDAM/GLE technical 
document.
    Comment: A commenter noted that several of the groups of marine 
birds and mammals used in the NRDAM/CME could be eliminated, based on 
documentation of relative lack of vulnerability to oil spills.
    Response: The Department acknowledges that some groups of marine 
birds and mammals are not as sensitive as others to the effects of oil 
spills. However, the model also evaluates indirect effects (e.g., via 
the food web) for both oil and chemical spills, which could be 
significant in certain scenarios.
    Comment: One commenter stated that the species contained in the 
NRDAM/CME should be limited to those for which reliable abundance data 
exist, or those that are threatened or endangered.
    Response: The NRDAM/CME is limited to those species for which 
reliable abundance estimates were available. Section 6, Volume I of the 
NRDAM/CME technical document explains the criteria the Department used 
to establish the estimates. The Department included threatened or 
endangered species where data were available. However, in light of the 
limited range of injuries and compensable values considered by the 
models, if injuries to these species are significant, trustees should 
consider using type B procedures instead of a type A procedure.
    Comment: One commenter addressed species abundance in New York 
Harbor. The commenter was unable to match the species abundance data 
for New York Harbor with that of the Erwin and Korschger (1979) 
reference cited. The commenter also stated that inaccuracies in 
abundance data resulted from a failure to adequately identify habitat 
types.
    Response: The Department has rechecked the data and made revisions, 
as appropriate, using the Erwin and Korschger (1979) source and other 
more recent sources. The results of the retabulated data are contained 
in Section 5, Volume IV of the NRDAM/CME technical document.
    The wildlife abundance data contained in the databases are 
province-wide abundances and, thus, are independent of the habitat 
grids and the habitat types assigned to the grid cells. Table 6.4, 
Volume I of the NRDAM/CME technical document contains the areas of 
habitat used for all calculations of wildlife abundances. Further, the 
habitat grids for New York Harbor have been revised based on comments 
submitted by the States of New York and New Jersey.
    Comment: Another commenter stated that Great Lakes wildlife 
abundances should be based on a more thorough review of the literature 
and thought that Burt (1976) was an inappropriate source of data. The 
commenter recommended that the models incorporate site-specific data 
wherever they are available instead of applying average values for 
several provinces.
    Response: The Department has incorporated site-specific information 
into each of the lake provinces. The Department used Burt (1976) only 
if more province-specific data were not available. Public commenters 
supplied no additional data on wildlife abundances in the Great Lakes. 
Further, the Department conducted a thorough search for published and 
unpublished data and located no additional sources of data. When 
abundances were highly variable among lake province-specific sources of 
equal validity, the Department averaged available data to reduce the 
error associated with the estimates. Professional judgment, based on 
life history information for the species in question, was used to 
determine how available data would be applied.
    Comment: Some commenters disagreed with the statement in the 
proposed NRDAM/CME technical document that oysters are one of the most 
important species in southern Maine and New Hampshire.
    Response: The commenter appears to have misinterpreted the data 
contained in the proposed technical document. The Department 
acknowledges that NMFS does not list commercial oyster fisheries in 
northern New England; however, there is a small recreational fishery 
for oysters. Thus, they are not commercially important but do have 
recreational significance.
    Comment: One commenter asserted that shrimp is by far the most 
significant catch in South Carolina. The commenter also thought that 
roughtail rays, orange filefish, and scad were overly abundant in the 
database relative to shrimp.
    Response: The Department could not find any Federal or State stock 
assessments or biomass surveys for shrimp, roughtail rays, orange 
filefish, or scad in South Carolina. Therefore, the Department based 
the abundance data for these species on NMFS commercial catch 
statistics. The NMFS catch data do not support the statement that 
shrimp are the most significant catch in South Carolina. Shrimp are a 
significant part of the inshore catch for many of the reporting areas, 
but are not common in offshore areas. The database includes all species 
for which NMFS catch data were available. Roughtail rays do form a 
large percentage of the catch by weight; however, they are not as 
significant economically as shrimp and, thus, are not as highly valued 
in the model. Orange filefish and scad are not major portions of the 
catch and the model accurately reflects that.
    Comment: One commenter questioned why the NRDAM/CME did not include 
lobster in New York Harbor.
    Response: The Department could not find any Federal or State stock 
assessments or biomass surveys of lobster in New York Harbor; 
therefore, the Department relied on NMFS commercial catch statistics. 
The NMFS commercial catch statistics do not show any catch for lobster 
in New York Harbor.
    Comment: One commenter sought clarification of several aspects of 
the discussion of young-of-the-year modeling in the NRDAM/CME technical 
document. The commenter questioned the meaning of the term ``stable 
distribution;'' the connection between young-of-the-year and the adult 
stocks; and the meaning of the term ``monthly mean,'' since no monthly 
mean abundances are present in the young-of-the-year database.
    Response: The term ``stable'' means constant in time. The 
Department has clarified the NRDAM/CME technical document on this 
point. See Section

[[Page 20590]]

4.3.2, Volume I of the NRDAM/CME technical document. The Department has 
also expanded the NRDAM/CME technical document to explain the 
connection between adult stock and young-of-the-year. See Section 6.5, 
Volume I of the NRDAM/CME technical document. The Department has 
modified the derivation of young-of-the-year abundance estimates so 
that such abundances are estimated on a daily basis for the first year 
of life and averaged for each month. See Section 6.5, Volume I, and 
Section 3, Volume IV of the NRDAM/CME technical document.
    Comment: One commenter questioned how the NRDAM/CME could determine 
the young-of-the-year surviving the first year of life under 
equilibrium conditions without knowing the first year natural mortality 
rate.
    Response: The NRDAM/CME calculates the number of one-year-old 
individuals needed to replace the fished stock, assuming equilibrium 
populations. Thus, the actual numbers of eggs and larvae are not 
calculated or needed. The model is calculating the percentage of one-
year-old animals lost because of the spill.
    Comment: Given that the abundance of a species may be seasonal, one 
commenter questioned how it was possible that the natural mortality 
rate and the fishing mortality rate are constant for members of a 
species group within and across years, as the equation for young-of-
the-year requires.
    Response: The natural mortality rates and the fishing mortality 
rates apply to the species population (stock) regardless of location 
and abundance. Stock abundances apply on a province-wide basis. The 
abundances in a specific province may vary by season due to migration 
of stock in and out of different provinces. Therefore, there is no 
inconsistency between using a constant mortality rate over time and 
using abundance figures that vary by season. For further discussion, 
see Section 6.5, Volume I of the NRDAM/CME technical document.
    Comment: One commenter thought the young-of-the-year database in 
the NRDAM/CME was incomplete, noting that there are several species 
groups in the adult database that are absent in the young-of-the-year 
database.
    Response: Certain species, such as anadromous fish, do not spawn in 
marine habitats. Thus, there may be adults present in a given province 
without young-of-the-year present.
    Comment: One commenter questioned whether forage fish production 
was nonexistent outside of structured habitats as assumed in the 
database for California provinces 40 through 47.
    Response: The NRDAM/CME does not assume that forage fish production 
is nonexistent outside of structured habitats. The food web model 
includes forage fish, in particular planktivorous forage fish, in open 
water habitats.
    Comment: One commenter questioned the inclusion of goldfish as 
representative herbivorous forage fish in wetlands in the proposed 
NRDAM/GLE technical document.
    Response: The Department used goldfish in the table to which the 
commenter refers simply as an example, because it is one of the few 
truly herbivorous fish. Most forage fish feeding on the bottom are 
omnivorous.
    Comment: One commenter thought the proposed NRDAM/CME technical 
document was unclear how zooplankton production was determined since no 
zooplankton production values were presented in the biological 
database.
    Response: The NRDAM/CME calculates zooplankton production based on 
a percentage of phytoplankton production. The technical document has 
been clarified. See Section 4.4, Volume I of the NRDAM/CME technical 
document.
    Comment: A commenter noted that the rates of production for 
planktivorous forage fish were not provided in the NRDAM/CME biological 
database and questioned how primary production for these fish was 
calculated.
    Response: The NRDAM/CME calculates planktivorous forage fish 
production based on a percentage of zooplankton production. The 
technical document has been clarified. See Section 4.4, Volume I of the 
NRDAM/CME technical document.
    Comment: One of the independent technical reviewers questioned why 
several bait fish species were not included in the models.
    Response: The Department did not include such species because it 
could find no quantitative data for such species.

G. Toxicity and Mortality

    Comment: Some of the independent technical reviewers stated that 
the biological effects submodel was logical and well conceived for 
assessing the effects of minor spills. However, other commenters 
asserted that the submodel suffered from an overall lack of supporting 
data and questioned the methods used for calculation of mortality and 
toxicity data. One commenter noted that bioassay studies measuring the 
lethality of oil and petroleum mixtures directly were preferable to the 
oil toxicity assumptions used in the models and thought such 
information must surely be available. Another commenter asserted that 
laboratory bioassays overestimate metals toxicity and that the 
Department should consult EPA's Water Effects Ratio studies on binding 
capacity in natural waters. A few commenters thought that the models' 
treatment of metals speciation was overly simplistic. Another commenter 
maintained that the proposed toxicity values were too high compared 
with EPA water quality criteria. Further, the commenter suggested the 
Department refer to EPA sediment quality criteria for benthic 
organisms.
    Response: The Department believes that the biological effects 
submodel incorporates the best available mortality and toxicity data. 
The Department has not been able to locate bioassay data on the 
toxicity of hydrocarbon products that have been developed under 
carefully controlled conditions, with constant aromatic concentrations 
in the water. Also, oil and petroleum products are highly variable in 
their percent composition and bioassay results. Very few studies have 
addressed oil toxicity and, therefore, the Department has not used 
direct bioassay data on oils in the models. See Section 4.2.3, Volume I 
of the NRDAM/CME and NRDAM/GLE technical documents.
    Laboratory bioassays more closely correspond to dissolved metal 
concentration toxicity than total metal concentration in the water. The 
models estimate dissolved metal concentrations, and these are the 
concentrations assumed to be causing the injuries. Thus, the models 
address metal speciation to the extent possible, without incorporating 
a complex speciation model. Further, even such a complex speciation 
model could not be coupled to appropriate toxicity data at the present 
time, as is well noted in EPA and other literature.
    EPA's water quality criteria would not be an appropriate basis for 
the models' toxicity calculations. The toxicity data in the models are 
mean values for acute response. Water quality criteria are designed to 
be lower than any concentration found to have either an acute or 
chronic response for even the most sensitive species. Thus, the water 
quality criteria should be lower than the models' toxicity data. EPA 
sediment quality criteria are evaluated in Section 4.2.1, Volume I of 
the NRDAM/CME and NRDAM/GLE technical documents.
    Comment: One commenter provided a highly technical review of the 
proposed toxicity model. The commenter noted that it was the commonly 
accepted toxicity model currently used in

[[Page 20591]]

environmental toxicology. The commenter encouraged the Department to 
apply chemical- and species-specific values to the alpha and gamma 
terms in the toxicity algorithm.
    Response: The Department appreciates the in-depth review of the 
toxicity model provided by the commenter; however, insufficient data 
exist at this time to make the recommended changes. To the extent 
possible, the Department has used gamma values that are chemical-
specific and alpha values that vary by class of chemicals. See Section 
4.2.1, Volume I of the NRDAM/CME and NRDAM/GLE technical documents.
    Comment: One commenter questioned the accuracy of the toxicity 
calculations of the proposed NRDAM/GLE, particularly with regard to 
releases of metals. The commenter provided information on a series of 
test cases in the Niagara River that the commenter ran using the 
proposed NRDAM/GLE.
    Response: Based on the information provided, the Department 
believes the test cases run by the commenter may not provide an 
appropriate basis for evaluating the NRDAM/GLE. The NRDAM/GLE was not 
designed to address multiple releases from various sources over a 
number of years. It appears that the commenter may have run the cases 
with the accumulated mass of contaminants as if that mass had resulted 
from single event spills of short duration. Also, the model was not 
designed to evaluate long-term chronic exposures to hazardous 
substances.
    The Department does agree, however, that pure metals are not 
correctly modeled by the NRDAM/CME and NRDAM/GLE. The toxicity data for 
these metals are based on bioassay studies that measured the dissolved 
metal ion concentrations in water. Such toxicity data are not 
representative of the chemical state of the metal that would occur 
under natural environmental conditions. As a result, the Department has 
deleted all pure metals from the chemical databases.
    Comment: Some commenters, including one of the independent 
technical reviewers, expressed concern about the models' failure to 
account for the additive toxicity of aromatics in oils.
    Response: The Department has modified the NRDAM/CME and NRDAM/GLE 
to include an additive toxicity model that accounts for the combined 
lethality of similar acting aromatics in oils. The Department used the 
information identified by one of the commenters to construct this 
additive toxicity model. See Section 4.2, Volume I of the NRDAM/CME and 
NRDAM/GLE technical documents for the specific algorithm used and 
further clarification of the changes made.
    Comment: Several commenters noted examples, primarily for metals, 
where the proposed NRDAM/CME used toxicity thresholds that were lower 
than naturally occurring water solubilities. The commenters suggested 
these errors were due to the Department's inappropriate use of 
freshwater toxicity data in saltwater environments.
    Response: Comparison of the salt- and freshwater databases for 
those chemicals where data existed showed no significant differences in 
toxicity values given the variability of such data. However, the 
Department has deleted those chemicals where a difference would be 
expected (i.e., those making up salinity such as sodium).
    The toxic thresholds in the models are only used as switches to end 
the calculations of the physical fates submodel. When the physical 
fates submodel determines that concentrations are below this level in 
all locations, it stops running. The threshold is the concentration 
that would cause one percent mortality at 30 degrees Celsius after 96 
or more hours of exposure for the most sensitive species group. The 
biological effects submodel calculates the actual mortality for each 
species group based on duration of exposure and temperature. The 
proposed technical documents contained an incorrect list of the toxic 
thresholds actually used by the models. The documents have been 
corrected. See Table III.2.1, Volume III of the NRDAM/CME and NRDAM/GLE 
technical documents.
    The toxicity data in the database, including the threshold values, 
apply to dissolved concentrations, not total concentrations. For 
chemicals that are highly partitioned, such as metals and nonpolar 
organics, the dissolved concentrations will be a small fraction of the 
total. The physical fates submodel partitions chemicals in both the 
water column and the sediments. Only dissolved chemical in the water 
column or in the sediment pore water causes toxicity in the models. 
Thus, the toxicity values should not be compared to total 
concentrations in water, but rather to the dissolved portion only. This 
accounts for the discrepancies perceived by the commenter when 
comparing thresholds to total metal background concentrations. The 
Department has eliminated all chemicals with a solubility below the 
toxic threshold.
    Comment: One of the independent technical reviewers thought that 
the use of toxicity values based on acute toxicity laboratory tests 
could underestimate toxic effects, because in acute tests involving up 
to 96 hours of exposure, animals are typically not fed. Thus effects 
result from water-borne exposure only. The technical reviewer concluded 
that the models may be inappropriate for spills of hydrophobic organic 
compounds where most of the exposure would be through contaminated 
food.
    Response: The Department agrees that the models may not fully 
capture the effects of a spill that contaminates food sources. The 
Department has revised the technical documents to clarify this point. 
See Section 4.2.1, Volume I of the NRDAM/CME and NRDAM/GLE technical 
documents.
    Comment: One of the independent technical reviewers thought that 
the models should account for the incremental effects of spills on 
existing levels of contamination.
    Response: The Department did consider inclusion of background 
contamination data in the models. However, data sources were 
insufficient to include such information for the entire area covered by 
the models. As a practical matter, background contamination present at 
the spill site before a spill would lower the threshold for effects by 
the spill. Thus, not including background contamination in the models 
is likely to underestimate injuries. However, the Department believes 
that in cases where background contamination is significant but the 
cost of using type B procedures is not reasonable, trustees should 
still have the option of using a type A procedure.
    Comment: Commenters suggested that rather than use average exposure 
concentrations in the plume, the models should evaluate mortalities 
that accrue from the actual cumulative exposure. One of the independent 
technical reviewers thought that it was not meaningful to estimate 
fractions of animals killed and recommended that the models round up 
fractional mortalities to total animals killed.
    Response: Inadequate data currently exist to estimate effects of 
cumulative exposure. Further, the computational complexities and the 
potential size of the internal, intermediate data files are effectively 
beyond the capacity of currently available PCs. Instead, the biological 
effects submodel performs multiple iterative runs, using different 
randomized algorithms, and then averages the results of these runs to 
avoid anomalous model outputs and increase reliability. See Section 
4.3.1, Volume I of the NRDAM/CME and NRDAM/GLE technical documents.
    The model is intended to generate valid mortality estimates for 
portions of

[[Page 20592]]

populations rather than discrete deaths of individual animals. 
Therefore, fractional mortality figures are appropriate.
    Comment: One commenter thought the portions of the NRDAM/CME and 
NRDAM/GLE technical documents pertaining to the use of particles to 
represent biological populations were unclear. The commenter sought 
clarification of how the models operate when the particles hit a 
physical boundary within the models. The commenter also sought 
clarification of the fate of particles and their exposure history 
during a change of seasons. The commenter believed there should be 
exposure memory for particles representing those species whose density 
is constant across seasons. The commenter further supported the use of 
multiple iterative runs to minimize the variability error caused by 
using a finite number of particles to represent a population and by 
limiting particle movements on a daily basis.
    Response: Particles may be transported out of a grid at the 
downstream edge and ``created'' as previously unexposed particles at 
the upstream edge. Particles intersecting land are reflected back into 
the water. At the change in seasons, the models assume that new 
individuals are present (at pre-spill abundances) and do not carry over 
the exposure history from the past season. The past season's injuries 
are tabulated and the exposure history for the new season is set at 
zero. Therefore, in a case where exposure extends across the seasonal 
boundary, the time of exposure would be underestimated. The technical 
documents have been clarified on these points. See Section 4.3, Volume 
I of the NRDAM/CME and NRDAM/GLE technical documents. The Department 
agrees that multiple runs of the same scenario will give a better 
prediction of damages than a single run. Thus the models' internal 
procedures average multiple runs to arrive at a damage figure.
    Comment: Several commenters generally criticized the proposed 
models' wildlife mortality calculations. Some commenters maintained 
that the models fail to adequately distinguish among effects of 
different types of compounds and questioned the application of wildlife 
mortality probabilities to substances other than crude oil since the 
only supporting data sources were for crude oil.
    Response: The Department considers the wildlife mortality model to 
be reasonable, scientifically justified and consistent with experience 
in actual spill events. The probabilities of wildlife mortality used by 
the models are based on data obtained from observations of real spills. 
Some modifications have been made to the probability values based on 
more recent information.
    The Department acknowledges that the supporting data sources for 
wildlife mortality do address crude oil. However, the Department 
believes that the models adequately account for the differences between 
crude oil and other petroleum products by calculating wildlife exposure 
dose based on the oil thickness and slick size. The mortality threshold 
is based on the exposure dose that is sufficient to cause an observable 
effect in experimental studies. If the exposure level exceeds the 
threshold, then wildlife mortality is assumed. For example, petroleum 
products that spread to sheen quickly, entrain, and/or evaporate have 
much less effect on wildlife in the models than thick, long-lasting 
oils, such as crude oil. Thus, the models account for differences among 
hydrocarbon-based oils and products through the physical fates submodel 
and the exposure algorithm. This algorithm is further explained in 
Section 4.3.4, Volume I of the NRDAM/CME and NRDAM/GLE technical 
documents.
    Comment: Some commenters contended that no data were available to 
support mortality rates for raptors. Commenters also questioned the 
extrapolation of sea otter mortality probabilities to polar bears.
    Response: The Department considers the eagle mortality rates to be 
appropriate for raptors in general. Mortality of eagles and other 
raptors as calculated by the models generally results from contact with 
slicks in shallow waters and along shorelines, and both eagles and 
other raptors occupy such areas to similar degrees. The eagle mortality 
rates used in the models are also supported by evidence from the Exxon 
Valdez spill. Further, osprey behave very much like eagles. Thus, the 
Department believes it is reasonable to use the same mortality rate for 
eagles and osprey.
    The Department recognizes that no explicit data are available on 
the probability of polar bears dying from spills. However, the 
Department believes that it is appropriate to use the same probability 
for all furbearers because the mechanisms of exposure and toxicity, 
namely ingestion of oil through grooming, are similar.
    Comment: One commenter thought that the NRDAM/GLE should use a 90 
to 95 percent mortality rate for ducks contacting heavy oil, as the 
proposed NRDAM/CME did.
    Response: The Department has modified the NRDAM/GLE to include duck 
mortality rates that are consistent with those in the NRDAM/CME. Based 
on recent information, the Department has incorporated a 99 percent 
mortality rate in both models for those ducks that are exposed to a 
spill over the threshold dose. See Section 4.3.4, Volume I of the 
NRDAM/GLE technical document.
    Comment: Commenters noted an apparent lack of correspondence 
between numbers of marine birds the models estimated to be killed, and 
numbers killed based on actual wildlife recoveries and detailed damage 
assessments. One commenter believed that the mortality counts for birds 
impacted by oil spills in Florida appeared low in almost all cases. 
Commenters suggested an alternative hindcast model for wildlife 
mortality estimation that was used for the T/V Puerto Rican, Apex 
Houston, Nestucca, and Exxon Valdez oil spills.
    Response: The Department believes that the apparently low 
mortalities observed by the commenter were due to the low abundance 
data for certain bird species contained in the proposed biological 
database for the Florida coast. The Department has revised the wildlife 
database for the Florida coast based on additional information provided 
by commenters. Model runs conducted with the revised wildlife abundance 
database no longer reveal large discrepancies. See Section 5, Volume IV 
of the NRDAM/CME technical document.
    Also, the Department has revised the probability of mortality for 
aerial divers using hindcasts, as suggested. See Section 4.3.4, Volume 
I of the NRDAM/CME technical document.
    Comment: Several commenters, including one of the independent 
technical reviewers, generally questioned the oil mortality 
probabilities and suggested that they be calibrated to data from actual 
spills, noting that the wildlife mortality probabilities in the 
proposed models were inconsistent with experience in the Exxon Valdez 
spill.
    Response: The mortality probabilities included in the models are 
based in part on data from actual spills. Very little data exist on the 
natural resource effects of small spills of the type addressed by the 
type A models. Therefore, it is impossible to determine the need for 
calibration. Nonetheless, the Department did consider data on physical 
fates and biological effects collected after the Exxon Valdez spill 
when evaluating the NRDAM/CME's mortality predictions. Bird and marine 
mammal injuries estimated by the

[[Page 20593]]

model provide reasonable agreement with estimated kills caused by the 
spill.
    Also, the commenters may not have correctly interpreted how the 
model calculates wildlife mortality. The commenters inferred that the 
wildlife mortality probabilities are multiplied times all animals at 
risk, and include populations for all of the northern Gulf of Alaska in 
the total population at risk. Instead, the model only multiplies the 
probabilities times the animals actually encountering oil and receiving 
a dose above the threshold value. Thus, the population ``at risk'' in 
this sense is orders of magnitude lower than the commenter's suggested 
value.
    Comment: Several commenters asserted that the models overestimate 
wildlife injuries by failing to account for weathering of oil, the 
effect of temperature, and the fact that light products like gasoline 
and other floating chemicals would be readily washed from the coats of 
furbearers.
    Response: The physical fates submodel does account for weathering. 
Evaporation, degradation, and entrainment reduce the area and thickness 
of slicks, which in turn reduces the frequency with which wildlife will 
be exposed to oil doses large enough to induce effects. See Section 
3.5, Volume I of the NRDAM/CME and NRDAM/GLE technical documents. For 
example, extremely volatile compounds evaporate so quickly that their 
surface slicks have essentially no effect on wildlife.
    The Department does not think that temperature is likely to have a 
significant effect on mortality rates. While one effect of oiling is a 
decrease in thermal conductance of fur and feathers, it is generally 
thought that the predominant toxic effects result from ingestion of oil 
during grooming. See Section 4.3.4, Volume I of the NRDAM/CME technical 
document. Also, animals are adapted to the climate in which they live. 
Thus, a tropical species suffers from hypothermia at a higher 
temperature than subpolar species.
    Finally, the light products and chemicals contained in crude oil 
are widely recognized to be the more toxic components. The light 
products and chemicals are hydrophobic, and so would not be washed from 
fur by seawater. Given the same dose, in terms of mass of hydrocarbons, 
these products are expected to have similar effects to crude oils.
    Comment: One commenter thought that the proposed 10 percent 
mortality rate for terrestrial mammals was too high.
    Response: The Department agrees that the proposed mortality rate of 
10 percent was too high and has reduced it to 0.1 percent. The 
Department considers this revised rate to be reasonable as compared to 
the mortality rate for seals. The NRDAM/CME assumes a seal mortality 
rate of 1.0 percent based on hindcast projections using Exxon Valdez 
data. Seals continuously inhabit open waters and shorelines, whereas 
terrestrial wildlife inhabit shorelines only a portion of the time. 
Therefore, terrestrial mammals would be expected to have lower rates of 
encountering spills and, thus, lower mortality rates.
    Comment: Commenters suggested the models grossly overestimate fish 
mortality from oil spills by overestimating dissolved hydrocarbon 
concentrations. These commenters particularly took issue with the 
Department's assumption that total aromatic hydrocarbon content of a 
spill remains the same by percentage before and after the spill. These 
commenters cited studies showing that weathering results in little of 
the hydrocarbon content entering the water column.
    Response: The Department does not think that the models 
overestimate fish mortality resulting from oil spills. Aromatics in 
oils are known to cause the most acute toxicity. See Section 4.2.3, 
Volume I of the NRDAM/CME and NRDAM/GLE technical documents. The 
physical fates submodel separately tracks dissolved aromatics of two 
molecular weight size classes: (1) monoaromatic benzenes and (2) 
diaromatic compounds. These aromatics do volatilize rapidly in the 
models. Thus, relatively little of these aromatics end up in the water 
column and cause toxicity. See Section 3.5, Volume I of the NRDAM/CME 
and NRDAM/GLE technical documents.
    Also, the models do not assume that the percent composition of 
total aromatic hydrocarbons remains constant in the spilled substance. 
The Department has clarified the text of the technical documents on 
this point. See Section 4.2.3, Volume I of the NRDAM/CME and NRDAM/GLE 
technical documents.
    Comment: Commenters asserted that it is too simplistic to use a 
mean LC50 (the concentration at which 50 percent of test organisms die 
within a defined time period) for the whole taxonomic class of fish or 
for all plants, algae, or angiosperms. Further, commenters maintained 
that the Department failed to provide information sufficient to 
evaluate the statistical relationships involved.
    Response: The NRDAM/CME and NRDAM/GLE technical documents contain 
detailed descriptions of the process the Department used to derive 
average toxicity parameters. Researchers have found that the LC50 for 
one species is a reasonable predictor of toxicity for other species 
within the same family and that, in many cases, cross-family 
correlations are also significant. See Sections 4 and 7, Volume I of 
the NRDAM/CME and NRDAM/GLE technical documents. Also, the 
administrative record for this rulemaking includes additional material 
on the derivation of average toxicity parameters.
    Available information is insufficient to disaggregate further the 
toxicity values used in the models. Disaggregation of the few data 
available would increase the error in the model result because of the 
uncertainties associated with individual data points. For plants, 
algae, or angiosperms, available data in EPA's AQUIRE database support 
the use of a mean LC50 value. There are insufficient data to quantify 
differing values by plant group. The Department does not believe that 
use of more specific plant values would significantly improve the 
reliability of the model damage figure because the model damage figures 
are not sensitive to the value assumed, within the range of observed 
data for a given chemical.
    Comment: A few commenters generally criticized the Department's 
approach to assessing fish mortality from toxicity data using 
statistically averaged values, asserting that information necessary to 
review the adequacy of such an approach was not included in the 
proposed technical documents. One of the commenters noted apparent 
inconsistencies between the values used in the models and the values in 
the AQUIRE database. Another commenter asserted that unless the 
Department could demonstrate that the oil toxicity algorithms in the 
revised NRDAM/CME were more accurate than those in the original NRDAM/
CME issued in 1987, the Department should return to the original 
algorithms.
    Response: The Department believes it has provided sufficient 
information and opportunity to review the approaches used to calculate 
fish mortality and to derive the toxicity data contained in the 
databases. Section 4, Volume I of the NRDAM/CME and NRDAM/GLE technical 
documents explains how the models calculate fish mortality. Section 2, 
Volume III of the NRDAM/CME and NRDAM/GLE technical documents provides 
toxicity values for each substance contained in the chemical databases 
and the source of information used to derive those values. Due to the 
volume of material, the Department has

[[Page 20594]]

not included in the technical documents all of the raw data and 
statistical analyses that were compiled for each toxicity value. 
However, the sources of the raw data, such as AQUIRE, are available to 
the public. Also the methods used to derive the statistically averaged 
toxicity values are consistent with those commonly used in aquatic 
toxicology. The toxicity values used in the model are based on all 
literature in the AQUIRE database as of November 1991. The AQUIRE 
database has not been updated since that time. The commenter appears to 
have reviewed only a limited range of data in the AQUIRE database.
    Finally, the Department believes that the revised NRDAM/CME is an 
improvement over the original NRDAM/CME and more accurately calculates 
oil toxicity. The technical documents explain all oil toxicity values 
used in the models. See Section 4.2.3, Volume I, and Section 2, Volume 
III of the NRDAM/CME and NRDAM/GLE technical documents. The toxicity 
parameters are for dissolved aromatic hydrocarbons of less than 200 
molecular weight, which is what the models calculate as the toxic 
material. Literature estimates are for whole oil, total petroleum 
hydrocarbons, or water soluble fractions, and thus are inapplicable to 
and higher than those for the dissolved low molecular weight fraction.
    Comment: One commenter noted that the susceptibility of developing 
fish eggs and larvae to toxic substances changes over time and sought 
clarification whether the models account for seasonal variations in 
toxicity.
    Response: The Department recognizes that the susceptibility of fish 
eggs and larvae to toxic substances may change with age. However, the 
Department does not think that sufficient research data have been 
compiled to quantify a change in toxicity by age of eggs and larvae for 
all the species groups and chemical substances contained in the 
chemical database. Therefore, the models have not been revised to 
account for this potential effect.
    Comment: Some commenters, including one of the independent 
technical reviewers, criticized the models for failing to account for 
the effects of avoidance of a spill by fish.
    Response: The Department acknowledges that some portion of fish 
populations may avoid spills of some types of chemicals. However, the 
Department was unable to identify adequate quantitative data on this 
phenomenon to include it in the models at this time.
    Comment: One of the independent technical reviewers thought the 
NRDAM/CME should account for the fact that fish and macroinvertebrates 
may be exposed to intertidal contamination when the tide is in.
    Response: The model does account for tidal inundation in its 
calculation of the water column plume. Contamination in the water may 
move into intertidal areas when water is present over them. When the 
tide goes out, the plume is transported out as well.
    Comment: One of the independent technical reviewers thought the 
fish swimming speeds used in the proposed models were extremely low.
    Response: The Department believes that the swimming speeds 
incorporated in the models are appropriate for the time step involved. 
The models use these speeds as distance moved in a single direction in 
an individual time step. Direction is randomized, so that after many 
time steps, motion is random. The rates are low because of the time 
step used. Otherwise, there is too much migration of fish in the 
models. For further discussion, see Section 4.3.1, Volume I of the 
NRDAM/CME and NRDAM/GLE technical documents.
    Comment: One commenter questioned the accuracy of the statement in 
the proposed NRDAM/GLE technical document that short duration 
disturbances of biota are not evaluated in light of the fact that the 
model specifically focuses on acute injury.
    Response: The statement in the technical document to which the 
commenter refers simply addresses the model assumption that seasonal 
biological abundances do not change for reasons other than the spill. 
See Section 4.1.2, Volume I of the NRDAM/GLE technical document.
    Comment: One commenter stated that the Department had made an 
unsubstantiated assumption that, for wildlife, death of parent animals 
will necessarily lead to death of immature animals. However, other 
commenters stated that, according to the equations in the proposed 
technical documents, the models fail to account for lost future harvest 
of young that are killed as a result of the death of their parents. A 
few commenters thought the models assumed that the spill occurred at 
the time of fledging and that the time of fledging is constant for all 
species. Some commenters also thought that the hatching and fledging 
times presented in the database were excessive.
    Response: The models assume that if adult birds or mammals are 
killed while their young are dependent upon them, then the young will 
be lost as well. See Section 4.5.2, Volume I of the NRDAM/CME and 
NRDAM/GLE technical documents. The models do not assume that the spill 
occurs at the time of fledging or that the time of fledging is constant 
for all species. The wildlife life history parameters used by the 
models include the age (in months) at which young are fledged or 
weaned, not the month of the year in which they are fledged or weaned. 
The fledging and weaning ages in the database are the ages at which 
young become independent. See Section 4, Volume IV of the NRDAM/CME 
technical document, and Section 3.6, Volume III of the NRDAM/GLE 
technical document.
    A substantial volume of literature addresses the parental care of 
young birds and mammals and the inability of those young to survive 
without parents. See Section 4, Volume IV of the NRDAM/CME technical 
document, and Section 3.8.4, Volume III of the NRDAM/GLE technical 
document. The Department believes it has provided ample documentation 
to substantiate that the death of parent animals does result in the 
death of young that are dependent on them.

H. Loss of Production

    Comment: One of the independent technical reviewers stated that 
direct oiling can kill seagrass and thereby reduce habitat function. 
The technical reviewer suggested that the models account for the effect 
of oiling on submerged macrophytes such as seagrass and kelp.
    Response: The NRDAM/CME calculates the sublethal loss of production 
and the acute lethal effects to subtidal seagrasses and other submerged 
macrophytes exposed to concentrations of dissolved oil in the water. 
However, evidence of oil coating and smothering macrophytes that are 
under water or in floating beds, does not appear to be available. The 
Department does not consider smothering of subtidal seagrass to be 
likely given that oil slicks float on the water surface and wave action 
would flush the oil from floating beds. Therefore, the Department has 
not included the coating and smothering of subtidal macrophytes as an 
injury calculated by the model.
    Intertidal seagrass habitats have been added to the NRDAM/CME as a 
new habitat type. The Department recognizes that intertidal seagrass 
can be coated by oil and smothered. Research studies have shown 
saltmarsh plant mortalities to occur from exposure to oil thicknesses 
of 14 millimeters. As a result, oiling of intertidal seagrass beds over 
a threshold thickness of 14

[[Page 20595]]

millimeters is considered lethal to seagrass, which is consistent with 
the model's treatment of oiling effects on other intertidal habitat 
types. See Section 4.3.3.2, Volume I of the NRDAM/CME technical 
document.
    Comment: One commenter thought that the proposed method of 
estimating post-spill recovery of macrophyte primary production was 
adequate for the purposes of the NRDAM/CME. However, the commenter 
suggested that the Department consider adjusting the sigmoidal function 
used to estimate the rate of biomass production so that the maximum 
rate of production occurs at one-half of the pre-spill biomass level.
    Response: The Department acknowledges the limitations of available 
data to specifically define the shape of the recovery curve for 
macrophyte production. However, the model results are not sensitive to 
the specific shape of the recovery curve; therefore, further effort to 
refine the curve is not likely to result in significantly improved 
reliability.
    Comment: A few commenters, including one of the independent 
technical reviewers, thought the models should consider the effects of 
compensatory growth (i.e., enhanced production due to the removal of 
inter- and intra-species density-dependent growth-limiting factors). 
These commenters stated that the models would overestimate production 
losses unless compensatory growth were considered. The commenter also 
suggested that, for lower trophic levels, the models should use the 
upper 75th percentile of the available LC50 or EC50 (i.e., the 
concentration at which growth is reduced by 50 percent) values rather 
than the mean of the LC50 and EC50 values.
    Response: The Department acknowledges that compensatory growth is 
not addressed within the models. However, the Department does not 
believe that production losses are overestimated as a result. For the 
small spills for which the models were designed, density-dependent 
growth and survival effects are likely to be insignificant. Further, 
density-dependent effects of large changes in fish population sizes 
have been difficult to quantify because of large natural variations and 
other ecological forces. The known disturbances that enhance 
productivity in ecosystems are natural events to which the ecosystem 
adapts rather than hazardous substance spills.
    The selection of only the upper 75th percentile of the LC50 and 
EC50 data would arbitrarily exclude toxicity data that may be 
representative of all sensitive and nonsensitive organisms in the 
field. Instead, the use of only the higher LC50 data would likely be 
representative of only the direct mortality to nonsensitive organisms.
    Comment: One of the independent technical reviewers recommended 
that the Department adopt a habitat-based approach to calculating the 
loss of production. Under this approach, each grid would have an 
associated biomass per unit area, and if that grid were oiled, an 
assumed percent of yield would be lost.
    Response: The models do use a habitat-based approach in the 
calculation of lost production. Each contiguous grid cell of the same 
habitat type within the grid has an associated biomass per unit area. 
The models calculate a percentage loss of biomass for the defined 
habitat type based on concentration and time of exposure. Biomass 
losses are summed and multiplied times the fishing mortality rate 
within the habitat type to calculate the lost yield. See Sections 4.3 
and 6.2, Volume I of the NRDAM/CME technical document, and Sections 4.3 
and 8.2.2, Volume I of the NRDAM/GLE technical document.
    Comment: Several commenters stated that the models rely on 
assumptions about ecological and exposure processes that are 
inconsistent with actual environmental conditions and could result in 
inflated damage estimates. Some commenters thought that the models 
inappropriately assume that ecological communities are evenly 
distributed both temporally and spatially. Several commenters disagreed 
with the model assumption that all demersal young-of-the-year are 
attached to the bottom and thought that this assumption would 
overestimate injury. Some commenters stated that the assumption that 
the depth of bioturbation is always 10 centimeters would also lead to 
overestimates of injuries. One of the independent technical reviewers 
thought that the models should account for vertical migration.
    Response: Ecosystems have the resilience to adapt to a number of 
unpredictable disturbances. Further, most small spills do not appear to 
significantly alter ecosystem structure or the temporal distribution of 
populations. The purpose of the models is to evaluate effects of spills 
small enough that they do not significantly alter ecosystem structure 
and dynamics. The Department believes it is reasonable to assume stable 
temporal distributions and that such an assumption is not inconsistent 
with actual environmental conditions during small spills. Therefore, 
the models continue to assume temporal distribution is constant by 
season for fish and shellfish and by month for young-of-the-year.
    The models do not assume even spatial distributions of ecological 
communities. The abundance data contained in the biological database 
were developed assuming evenly distributed species abundances within a 
particular habitat and biological province occupied by the species, and 
within a given season or month. The species abundance data are assigned 
to the models' biological computational particles. Abundances of young-
of-the year change monthly between habitats, provinces, and portions of 
the water column (pelagic, demersal or benthic), as appropriate to the 
species' life history. Therefore, the particles do not represent 
abundances that are evenly distributed spatially.
    The Department acknowledges that a small portion of demersal young-
of-the-year are potentially carried by currents. However, for the 
purposes of these models, the Department believes it is reasonable to 
assume that all demersal young-of-the-year are attached to the bottom. 
See Section 4.3.2, Volume I of the NRDAM/CME and NRDAM/GLE technical 
documents. If any demersal young-of-the-year did in fact drift with the 
current, they would likely be more exposed to the released substance, 
which would also be moving with the current. Therefore, if the 
assumption that all demersal young-of-the-year are attached to the 
bottom is not reasonable in a particular case, the models are likely to 
underestimate rather than overestimate damages.
    Also, simulation of vertical migration would require substantial 
additional modeling. The Department does not believe that adding such 
complexity would substantially improve the reliability of the final 
damage figure.
    In some contaminated, and thus ecologically stressed, habitats, the 
depth of bioturbation may be less than 10 centimeters. However, for 
most regions covered by the models, the Department believes it is 
reasonable to assume a depth of 10 centimeters, since most estuarine 
and marine sediments are relatively uncontaminated. Contrary to the 
commenters' assertion that this assumption may cause an overestimate in 
pore water concentrations in areas where bioturbation is small, the 
dilution of contaminants over the assumed depth of 10 centimeters could 
underestimate pore water concentrations and injuries.
    Comment: Some of the commenters, including one of the independent 
technical reviewers, thought the food web model was crude. For example, 
the independent technical reviewer

[[Page 20596]]

questioned the assumption that all biota are equivalent as food 
sources. Another commenter stated that the model provided only a rough 
approximation of upper trophic-level production losses attributable to 
spill-related reductions in primary productivity.
    Response: The Department acknowledges that the food web model is 
generalized for all aquatic habitats, but believes it is reasonable for 
the purposes of the NRDAM/GLE and NRDAM/CME. The food web model 
provides a reliable approximation of spill-related upper trophic-level 
production losses due to lost primary productivity. The Department 
believes that development of a more sophisticated, geographic-specific 
food web model would not significantly improve the reliability of the 
final damage figure.
    Comment: One of the independent technical reviewers thought that 
the consumption rate parameters contained in the food web model should 
vary seasonally.
    Response: Seasonal changes in lower trophic-level production rates 
and in temperature already partially provide a seasonal effect in the 
models. The Department does not believe that further refinement would 
significantly improve the reliability of the final damage figure.

I. Catch and Bag Losses

    Comment: Several commenters addressed the methods used to translate 
fish mortality and wildlife mortality into reductions in catch and bag. 
Commenters noted that a correct translation was needed because 
mortality is not a sufficient basis for damages if it does not result 
in a quantifiable reduction in the services provided by the resources. 
Some commenters thought the proposed single-species approach to 
modeling fishery complexes and ecosystems was simplistic but adequate 
for the purposes of the type A models. Other commenters argued that 
predictions of catch and bag losses were wholly unreliable and resulted 
in inflated estimates of service losses.
    Some commenters stated that the models sometimes predict a catch 
loss that exceeds the prediction of total mortality. These commenters 
asserted that inaccurately high fishing rates in the proposed NRDAM/GLE 
lead to the estimation that 50 to 70 percent of some fish species would 
be caught in a single year. Commenters complained that lost wildlife 
bag predictions failed to account for regional differences. Some 
commenters thought that failure to consider hunting regulations would 
inappropriately result in the calculation of losses even when hunting 
is prohibited. Commenters also criticized the use of uniform annual 
hunting mortality rates that do not assume any particular underlying 
daily patterns that would produce such rates.
    Response: The type A models calculate lost catch based on a 
standard fisheries model. See Section 4.5.1, Volume I of the NRDAM/CME 
and NRDAM/GLE technical documents. Fishing mortality is an 
instantaneous rate, the coefficient of an exponential, first-order 
``decay'' curve. The nature of this equation is that an instantaneous 
fishing mortality rate of one corresponds to an annual harvest of 63 
percent of the standing stock abundance present at the beginning of the 
year. The models account for growth of animals over the year they are 
harvested. Thus, the harvest may exceed the standing stock biomass 
present at any given time if the growth rate of the species is high. An 
annual harvest rate of 63 percent of standing stock abundance at the 
beginning of the year is not equivalent to a 63 percent catch of all 
the fish species stock cumulatively available throughout the year. The 
fishing mortality rates assumed in the database are based on best 
available fisheries statistics estimated by NMFS and State fisheries 
management agencies.
    The models apply hunting mortality rates to a population, i.e., a 
stock or group of interbreeding animals. See Section 4.5.3, Volume I of 
the NRDAM/CME and NRDAM/GLE technical documents. The hunting mortality 
rate in the models is simply the probability of being hunted 
successfully at some time and place over one year of life. Therefore, 
the models need not consider exactly where or when the animals are 
actually taken during a given year or whether hunting seasons are open 
or closed at the spill site. The Department derived the hunting 
mortality rates from tagging studies. A considerable amount of 
literature shows that birds and mammals have similar hunting mortality 
rates per animal (or per 100 animals) throughout North American 
populations due to migratory behavior and biological limits on 
productivity. See Section 4, Volume IV of the NRDAM/CME technical 
document. Therefore, the models use per-animal rates that are 
justifiably constant in time and space.
    Comment: Some commenters thought it was inconsistent for the models 
to assume that a fish species may be present or absent in certain 
seasons while also assuming that fishing and natural mortality rates 
are constant and act continuously on the population throughout the 
year.
    Response: The models include fishing and natural mortality rates 
that apply to stocks, which may move seasonally. The rates apply to the 
entire population of fish within the provinces occupied by a given 
stock no matter whether they are present or absent from particular 
locations at different times of the year.
    Comment: One commenter thought that the fish biomass figures used 
in the calculation of predation rates should reflect the entire stock, 
not just the exploitable stock biomass.
    Response: The models do, in fact, use the total stock biomass in 
these calculations.
    Comment: A few commenters noted that the yield model assumes that 
the biomass of each fish species is uniformly distributed over the 
entire province in which it is found, while the figures in the database 
reflect the actual distribution of the stock. The commenters further 
noted that if a species did not extend throughout the entire province, 
then this assumption would lead to an overestimate of loss of yield.
    Response: The Department chose province boundaries to minimize this 
source of error in the models. Additionally, there are three sub-areas 
(habitats) within each province with unique fishery biomasses. See 
Section 6.3, Volume I of the NRDAM/CME technical document.
    Comment: A few commenters raised questions about the interplay 
between the user-supplied information on closures and the models' 
spill-related mortality predictions. The commenters noted that the 
models do not adjust fishing mortality rates when there is a closure. 
Commenters suggested that users be allowed to make such an adjustment 
to avoid underestimating recovery periods.
    Response: The Department believes the assumption of a constant 
fishing mortality rate is reasonable for purposes of calculating 
recovery periods after minor spills. Only very large changes in fishing 
would be measurable in the population over the long term. For minor 
releases, it is unlikely that extended closures will occur.
    Comment: One commenter questioned the yield formulas in the 
proposed NRDAM/CME technical document and offered alternative formulas. 
The commenter noted that the proposed equations appeared to be missing 
the final year's contribution to yield.
    Response: The Department has reviewed the alternative formulas 
offered by the commenter and has concluded that they are more precise. 
The Department has modified the final versions of the models 
accordingly. The

[[Page 20597]]

models continue to calculate losses through the last year after the 
spill effects where killed individuals would have lived and died 
naturally. The Department has clarified the language of the technical 
documents on this point. See Section 4.5.1, Volume I of the NRDAM/CME 
and NRDAM/GLE technical documents.

J. Habitat Restoration

    Comment: Several commenters thought that the proposed methodology 
for calculating habitat restoration costs was so flawed that the 
Department should eliminate such damages from the models. These 
commenters identified four major flaws in the methodology that, taken 
together, render it invalid: (1) The models' acute toxic threshold is 
unrealistically low, leading the models to overestimate the size of the 
areas needing restoration; (2) the models often overestimate the time 
necessary for natural recovery; (3) the models select specific physical 
restoration measures without regard to their feasibility; and (4) the 
Department overestimated the fixed costs for small spills.
    Response: As discussed in more detail below, the Department 
believes that the methodology for calculating habitat restoration costs 
is reliable. Therefore, the Department has retained the methodology in 
the final versions of the NRDAM/CME and the NRDAM/GLE subject to the 
revisions described in Section VI.B of this preamble.
    Comment: Several commenters said the Department had failed to 
consider adequately the feasibility or the cost-effectiveness of the 
specific restoration actions included in the models. For example, 
commenters argued that sediment capping in deep water is not a proven 
technique and that the cost of such capping is highly site-specific. 
Therefore, these commenters thought that deep-water capping should be 
eliminated from the models. Commenters also thought that the Department 
had failed to substantiate its claim that the types of habitat 
restoration actions considered by the models are in fact the most cost-
effective. Some commenters, including one of the independent technical 
reviewers, thought the Department had unduly limited the range of 
restoration actions the models consider, and should have included off-
site restoration, partial rehabilitation, and other mitigating actions.
    Response: The Department has concluded, based on an extensive 
literature review, that all restoration options considered by the 
models are technically feasible. For additional discussion see Sections 
5 and 12, Volume I of the NRDAM/CME technical document, and Sections 5 
and 9, Volume I of the NRDAM/GLE technical document.
    Sediment capping in deep water is technically possible. The models 
calculate damages based on the cost of capping only if toxicity in 
sediments of an entire grid cell is so persistent that recovery 
following capping would be faster than natural recovery. The Department 
believes that in such cases, sediment capping is appropriate. The 
Department has attempted to account for site-specific factors relating 
to offshore capping by including cost figures that take into 
consideration the different distances that equipment and sediments must 
be transported.
    The Department also carefully evaluated a wide range of possible 
habitat restoration actions, including replacement of the affected 
resources with other resources, acquisition of equivalent resources, 
natural recovery, and other technologies. Based on this evaluation, the 
Department believes that it has identified the most cost-effective 
types of habitat restoration that can reasonably be included in models 
of this type.
    With regard to off-site restoration, the Department acknowledges 
that after recovering damages through the use of a type A procedure, 
trustees may well decide that the recoveries are best spent on off-site 
actions, such as the purchase and enhancement of nearby property to 
provide equivalent habitat to that lost. However, the availability of 
nearby land of the same habitat type as that injured, the cost of any 
available land, and the need for and feasibility of any actions to make 
the land equivalent in quality to that lost are all highly site-
specific factors. The Department does not believe that adequate data 
are currently available to include off-site restoration among the list 
of restoration actions evaluated by the models. With regard to partial 
rehabilitation, the Department has improved the resolution of the 
NRDAM/CME by a factor of four, which allows the restoration submodel to 
consider restoration of much smaller geographic areas.
    Comment: One commenter noted that techniques for reestablishing 
freshwater macrophyte beds of wild celery have been well demonstrated 
and should be separately included in the NRDAM/GLE.
    Response: The NRDAM/GLE does evaluate actions to restore wild 
celery beds but does so through consideration of a single type of 
restoration action for all aquatic bed habitats. The Department does 
not think the differences between the techniques for restoring various 
types of aquatic beds necessitate the development of distinct per-unit 
restoration costs for each type.
    Comment: Several commenters said that the models grossly 
overestimate the size of the areas to be restored because they use 
unrealistically low toxicity thresholds. To illustrate this point, 
commenters noted that the models assign a toxicity threshold to cupric 
chloride that is less than 1/40th its normal concentration in 
saltwater. These commenters argued that a spill of cupric chloride 
would have to spread over a huge area before it would dissipate below 
the toxic threshold, which would then necessitate restoration over a 
similarly huge area.
    Response: The toxicity thresholds included in the models merely 
serve as switches to end a model run. When the physical fates submodel 
determines that concentrations of the released substance are below the 
threshold level for that substance in all locations, it stops running. 
The threshold is the concentration that would cause one percent 
mortality at 30 degrees Celsius after 96 or more hours of exposure in 
the most sensitive species group. When calculating habitat restoration 
costs, the restoration submodel will examine all areas over which the 
spill has spread but will only calculate the cost of active restoration 
if it would result in lower compensable value than natural recovery. 
Compensable value is generated only when there is mortality or loss of 
production. The biological effects submodel calculates mortality and 
loss of production not on the basis of the toxicity thresholds but 
rather on the basis of mean LC50 and EC50 values. Therefore, the 
toxicity thresholds do not determine the extent of habitat restoration. 
Further, the models will only include the cost of active restoration in 
the final damage figure if such active restoration passes the cost-
benefit test discussed below.
    Comment: There were several comments concerning the models' 
predictions of recovery times. One of the independent technical 
reviewers suggested that the recovery times be modified. Another 
commenter noted that recovery times are very uncertain and that 
relatively small adjustments can have significant effects on estimates 
of total losses. Several commenters said the proposed restoration 
submodel generally overestimated the time required for natural 
recovery. Some commenters stated that the NRDAM/CME appeared to 
incorporate recovery times for seagrass beds and coral reefs that 
exceeded the literature values listed in the proposed NRDAM/CME 
technical document. Other commenters

[[Page 20598]]

questioned the lack of data for saltmarsh wetlands and mudflats and 
criticized the Department's use of identical recovery times for 
mudflats and sandy beaches.
    Response: The Department believes that the scientific literature 
supports the recovery times for seagrass beds, coral reefs, saltmarsh 
wetlands, and mudflats contained in the NRDAM/CME. Further, the 
recovery periods included in the NRDAM/CME for seagrass beds and coral 
reef are consistent with the literature cited in the technical 
document. See Section 4.3.3.2, Volume I of the NRDAM/CME technical 
document.
    Comment: One commenter questioned the Department's assumption that 
reproduction of fish and wildlife species resumes normal levels as soon 
as toxicity is no longer present.
    Response: The Department recognizes that substances may cause 
sublethal or chronic injuries that affect reproduction after 
concentrations have dropped below acutely toxic levels. However, the 
models do not attempt to address the effects of such sublethal or 
chronic injuries. If trustees believe such injuries are likely to be 
significant, they should consider conducting type B studies instead of, 
or in addition to, a type A procedure.
    Comment: There were a few comments about the inclusion of fixed 
restoration costs in the models. Some commenters, including some of the 
independent technical reviewers, thought that the fixed cost figure was 
too low; other commenters thought it was too high. One of the 
independent technical reviewers thought different fixed costs should be 
applied depending on the type of habitat affected.
    Response: Trustees who use a type A procedure will have to develop 
a restoration plan once they obtain compensation for the natural 
resource injuries. See 43 CFR 11.93(a). The proposed models included a 
fixed cost of $18,300 to cover restoration planning costs. The 
Department recognizes that the extent of the restoration planning costs 
for each particular case may vary dramatically. Such costs depend on 
whether the trustees intend to implement the restoration actions chosen 
by the model or develop other restoration actions. The costs also vary 
depending on the complexity of the selected restoration actions. In 
light of the highly site-specific nature of restoration planning costs, 
the Department has chosen to eliminate them from the models.
    Nevertheless, Sec. 11.15(a)(3) of the regulations, which is not 
changed by this rulemaking, allows trustees to recover the reasonable 
and necessary costs of an assessment. These costs include 
``[a]dministrative costs and expenses necessary for, and incidental to 
* * * restoration, rehabilitation, replacement, and/or acquisition of 
equivalent resources planning.'' 43 CFR 11.15(a)(3)(ii). Therefore, 
trustees who use a type A procedure and wish to recover restoration 
planning costs may develop their own estimates of such costs and 
include them as assessment costs in the demand presented to the PRPs 
under revised Sec. 11.91(a).
    Comment: One commenter was generally concerned about the quality of 
the Department's information-gathering efforts on habitat restoration 
techniques, noting the lack of current references in the technical 
documents. The independent technical reviewers recommended that the 
Department update the per-unit restoration cost data included in the 
models.
    Response: NOAA conducted an extensive literature search on natural 
resource restoration while developing guidance documents in connection 
with its natural resource damage assessment rulemaking under OPA. The 
Department has updated the per-unit restoration costs included in the 
models based on information compiled through this NOAA effort as well 
as other recent information. Also, as discussed below, the Department 
has revised the habitat restoration actions evaluated for structured 
habitats (i.e., wetlands, seagrass, macroalgal, coral, mollusk, and 
reef).
    Comment: A few commenters thought that the costs of upland disposal 
of sediments in the proposed NRDAM/GLE inappropriately failed to factor 
in long-term operation and maintenance.
    Response: The model includes the per-unit costs that a commercial 
facility would charge to accept sediment for disposal. These costs are 
one-time costs charged by the commercial facility and, thus, should 
include the facility's anticipated long-term operation and maintenance 
costs. Trustees would not incur any additional long-term costs.
    Comment: Some of the independent technical reviewers thought that 
the NRDAM/CME incorrectly assumed that seagrass replanting and oyster 
reef seeding would entail destruction and subsequent reestablishment of 
the entire habitat. These technical reviewers also thought this 
assumption would produce a bias against invoking habitat restoration 
because habitat restoration of this nature will generally result in 
compensable value in excess of that which would occur under natural 
recovery.
    Response: The Department has reevaluated the habitat restoration 
actions included for seagrass beds, invertebrate reefs, and other 
structured habitats and decided that in certain circumstances it would 
be technically feasible and more cost-effective to perform restoration 
actions that are less invasive than replacement of substrate. 
Therefore, the Department has revised the NRDAM/CME to include two 
potential restoration actions for such habitats. Where the sediments 
are sufficiently contaminated, the restoration submodel evaluates 
substrate replacement or capping followed by replanting or reseeding of 
the vegetation or invertebrate structure. Where sediments are not 
contaminated but mortality of the structural habitat has occurred, the 
submodel evaluates replanting or reseeding alone. The submodel 
evaluates each affected grid cell. See Section 12.2, Volume I of the 
NRDAM/CME technical document.
    Comment: One commenter thought that the Department had 
inappropriately based freshwater wetland restoration alternatives for 
hazardous substance releases on oil-related experiences.
    Response: The Department did not base freshwater wetland 
restoration for hazardous substance releases on data relating to oil 
spills. In fact, data on freshwater wetland restoration arises almost 
entirely out of non-oil experiences. See Section 5.7, Volume I of the 
NRDAM/GLE technical document.
    Comment: One commenter noted that removal and replacement of soils 
and vegetation is subject to failure and, even if ``successful,'' 
results in different habitat.
    Response: The Department agrees that removal of substrate and 
replanting can result in failures and the emergence of different 
habitats. The Department developed recovery rates and per-unit 
restoration costs that account for such risks.
    Comment: One commenter objected to the inclusion of washing and 
steam cleaning as restoration alternatives, noting that such actions 
caused injury during the Exxon Valdez response.
    Response: The Department has retained washing and steam cleaning as 
a potential shoreline habitat restoration alternative in the NRDAM/CME. 
The Department agrees that washing and steam cleaning cause injury but 
believes that inclusion of such alternatives in the model is 
appropriate subject to the decision criteria used by the models. When 
the model evaluates these alternatives, it assumes that the actions 
will destroy the habitat. The model will only select these alternatives 
if,

[[Page 20599]]

notwithstanding their deleterious effects, they nonetheless result in 
lower compensable value than natural recovery. In coastal and marine 
environments, the tidal flux creates a distinct shoreline habitat that 
is not present in the Great Lakes. Therefore, the NRDAM/GLE does not 
include shoreline washing and steaming.
    Comment: One commenter disagreed with the statement in the proposed 
NRDAM/GLE technical document that fish production is not negatively 
affected by dredging.
    Response: The Department acknowledges, and the models recognize, 
that dredging of a habitat reduces egg and larval fish production in 
the dredged habitat to zero initially, and then production follows 
vegetative and benthic recovery. The models simply assume that adult 
fish production is unaffected.

K. Assimilative Capacity Restoration

    Comment: Several commenters argued that assimilative capacity 
restoration costs are not legally recoverable. A number of commenters 
thought that inclusion of damages for lost assimilative capacity was 
overly speculative. These commenters stated that the presence of a 
spilled substance only causes a meaningful reduction in assimilative 
capacity if the resource will be required to assimilate more of a 
similar substance in the same area before the spilled substance 
degrades. The commenters argued that the type A models merely assume 
that assimilative capacity has been reduced. Some of these commenters 
thought that the assumption of an actual reduction in assimilative 
capacity is particularly troubling in the case of minor spills. The 
commenters noted that if type B procedures are used, trustees would be 
required to demonstrate an actual reduction in assimilative capacity. 
One commenter noted that Ohio v. Interior had stated that a procedure 
that permitted unduly speculative assessments would not constitute a 
best available procedure under CERCLA. 889 F.2d at 462.
    Response: The issue of whether lost assimilative capacity is a 
legally permissible category of damages was decided and resolved by the 
Department in 1986 and is beyond the scope of this rulemaking. The only 
issues that the Department is considering in this rulemaking are: 
whether the type A models adequately demonstrate a loss of assimilative 
capacity; and, if so, whether the models accurately compute the costs 
for restoring that loss. The Department believes that the assimilative 
capacity restoration costs computed by the final NRDAM/CME and the 
NRDAM/GLE are based on demonstrable, rather than speculative, losses.
    Assimilative capacity is an ecosystem's ability to repair itself by 
digesting, degrading, transforming, absorbing, or otherwise eliminating 
the pollutants placed in it. The Department recognizes that there are 
contrasting views of the nature of assimilative capacity. The purist 
position is that no materials placed in aquatic environments will ever 
simply disappear. Some substances cannot be broken down and will 
reenter and recycle through the ecosystem even if they have no 
detrimental effects. Other substances may be completely digested in the 
ecosystem and transformed into harmless or naturally occurring 
elements. This digestion, however, will consume some of the ecosystem's 
resources (e.g., dissolved oxygen) at the expense of natural processes 
and components of the system. Thus, following this purist approach, all 
additions to a water body will change it to a greater or lesser degree, 
and the only way to ensure restoration of the equilibrium in the 
ecosystem is to eliminate or remove all introduced material.
    Recognizing, however, that absolute removal of a discrete spilled 
substance can be impractical, if not impossible, some experts have 
adopted a more pragmatic approach. Under this approach, the 
assimilative capacity of a water body is usually viewed in relation to 
some water quality standard or level of service. In other words, 
assimilative capacity is the ability of a water body to absorb a 
particular pollutant up to the point where certain detrimental effects 
are realized.
    The Department carefully considered both the purist and pragmatic 
approaches to assimilative capacity. Based on this consideration, the 
Department included in the proposed type A models a methodology for 
computing assimilative capacity restoration. After reviewing the public 
comments, the Department has modified that methodology to ensure that 
the models more accurately quantify assimilative capacity losses. The 
Department believes that the modified methodology is appropriate for 
inclusion in the models.
    The Department does not believe that recovery of damages for lost 
assimilative capacity using the type A models is speculative. Trustees 
are authorized to recover damages to restore injured resources to their 
baseline conditions. Baseline is measured in terms of the services that 
the injured resources would have provided in the absence of the 
release. See 43 CFR 11.70(a). The assimilation of pollutants is a real 
service provided by natural resources and is well-founded in scientific 
literature. Assimilative capacity will be reduced whenever a release 
occurs. Releases use some of the assimilative capacity of aquatic 
environments and so long as the pollutants remain in the environment, 
some portion of assimilative capacity--a service provided by the 
natural resource--is lost. The Department does not agree that reduction 
in assimilative capacity is dependent on a subsequent release of the 
exact same substance in the exact same area. A release reduces 
assimilative capacity regardless of subsequent spill events.
    Nevertheless, the Department recognizes that there are practical 
limitations on measuring and addressing assimilative capacity loss, 
particularly in the context of a standardized procedure for minor 
releases. The issue, then, is not whether there is a reduction in 
assimilative capacity but, rather, the extent of the reduction and the 
type of actions that are appropriate to restore the lost assimilative 
capacity. As discussed above, assimilative capacity can be seen as the 
ability to absorb pollutants up to a threshold where detrimental 
effects occur. The type A models focus on injury to biological 
resources. Therefore, the threshold for meaningful loss of assimilative 
capacity now built into the models is mortality or loss of production 
resulting in compensable value. Releases that generate compensable 
value related to mortality or loss of production have, by definition, 
exceeded the assimilative capacity of the ecosystem. In the case of 
such releases, the models estimate the cost of restoring assimilative 
capacity. The cost is based on the removal of a mass equivalent in 
toxicity to the amount of the spilled substance that remains after 
concentrations have fallen below acute toxicity thresholds and after 
any habitat restoration actions are completed. Such removal will return 
the aquatic system to a state that is functionally equivalent to its 
baseline condition. When there has been no acute mortality or loss of 
productivity, the models do not calculate any assimilative capacity 
restoration costs.
    Comment: Some commenters asserted that the models fail to take into 
account how spatial, temporal, and chemical factors affect the 
assimilative capacity and function of a resource.
    Response: The Department has modified the models to account more 
fully for the factors affecting

[[Page 20600]]

assimilative capacity. The extent of assimilative capacity reduction 
depends on how long a substance remains in the environment. The 
proposed versions of the type A models did not consider degradation 
rates when computing lost assimilative capacity. The Department has 
modified the models to correct for the degradation rate of the released 
substance relative to the degradation rates of the contaminants found 
in the sediment at the sites at which dredging is presumed to occur. 
See Section 5.4.4, Volume I of the NRDAM/CME and NRDAM/GLE technical 
documents. The Department believes that this modification, along with 
the models' methodology for ensuring that the removed mass is 
equivalent in toxicity to the released substance, adequately account 
for the different factors affecting assimilative capacity.
    Comment: Several commenters objected to basing assimilative 
capacity restoration costs on the cost of dredging projects outside the 
area affected by the release. The commenters complained that there was 
no relationship between these dredging projects and the injured 
resources. One commenter asserted that inclusion of these damages was 
motivated by a desire to circumvent the normal appropriations process 
for funding dredging projects unrelated to the spill. Commenters, 
including one of the independent technical reviewers, observed that at 
dredging sites with heavy contamination, less material would have to be 
dredged than at sites with lower levels of contamination. These 
commenters noted that this method generates dramatically different 
figures for different geographic areas and that such differences were 
unfair because they were unrelated to the spill. A few commenters 
stated that since the Department seems to believe that assimilative 
capacity is affected over extremely large areas, the models should base 
damages on the lowest dredging cost for any of the provinces.
    Some commenters, including one of the independent technical 
reviewers, also argued that it would be more cost-effective to 
calculate damages based on prevention of wastewater discharges from 
point sources, such as publicly owned treatment works (POTWs). Another 
commenter suggested that the rule allow PRPs the option of determining 
and implementing more cost-effective methods of restoring lost 
assimilative capacity rather than paying the damages calculated by the 
models.
    Response: The Department agrees that the method for restoring lost 
assimilative capacity should be reasonably related to the actual 
release and cost-effective. The Department believes that the 
methodology included in the final type A models meets both these 
standards. Biota are potentially exposed to the released substance 
throughout an entire biological province; therefore, the Department 
evaluated potential dredge sites on a province-wide basis. Within each 
province, the Department focused on National Status and Trends (NST) 
sites and International Joint Commission (IJC) areas of concern because 
there are considerable data available on the sites and because they are 
heavily contaminated. The higher the toxic mass per volume at a site 
is, the cheaper is the cost of dredging sediment equal in toxicity to 
the remainder of the released substance. Therefore, for each province, 
the Department determined which NTS site or IJC area of concern had the 
highest toxic mass per volume and used that site to develop 
assimilative capacity restoration costs. See Section 13, Volume I of 
the NRDAM/CME technical document, and Section 9.4, Volume I of the 
NRDAM/GLE technical document.
    Had the Department restricted dredging to areas closer to the site 
of the release, the costs would likely have increased significantly, 
because the levels of contamination in those areas would be lower than 
at NST sites and IJC areas of concern. On the other hand, had the 
Department considered NST sites and IJC areas of concern well beyond 
the boundaries of the province in which the release occurred, the 
dredging would be less clearly related to the actual release.
    The Department considered using the cost of preventing discharges 
from point sources, such as POTWs, as a possible basis for assimilative 
capacity restoration costs. However, the Department was unable to 
locate adequate data on point source discharges for toxic chemicals. 
Commenters presented no additional information that would enable the 
Department to develop assimilative capacity restoration costs based on 
reducing point source discharges.
    With regard to allowing PRPs to develop alternative methods of 
restoring lost assimilative capacity, the objective of the type A 
models is to provide a sum certain on an inexpensive, expedited basis. 
Allowing PRPs to develop alternative restoration methods, providing 
trustees with an appropriate opportunity to evaluate the feasibility 
and adequacy of the PRPs' proposal, and giving the public a chance to 
review the proposal could undermine this objective. Where the models 
predict significant assimilative capacity restoration costs, PRPs who 
believe that they could restore lost assimilative capacity in a more 
cost-effective manner than that predicted by the type A models would 
have the option of funding type B procedures or pursuing an appropriate 
settlement with the trustees.
    Comment: One commenter thought that loss of assimilative capacity 
was a legitimate basis for recovery but stated that damages for the 
loss should be characterized as compensable value rather than 
restoration costs.
    Response: The Department believes that the damages for lost 
assimilative capacity are correctly categorized as restoration costs 
rather than compensable value. Under the existing regulations, 
compensable value is the economic value that the public loses until the 
injured resources recover. See 43 CFR 11.83(c)(1). The type A models 
calculate damages for lost assimilative capacity based not on economic 
value lost to the public but rather the cost of restoring baseline 
services.
    Comment: One commenter noted that the proposed NRDAM/CME in some 
cases predicts that more of the substance will remain in the 
environment than was spilled in the first place.
    Response: This result was caused by a coding error that the 
Department has corrected in the final version of the NRDAM/CME.
    Comment: Commenters noted that, contrary to assertions in the 
August 8, 1994, and the December 8, 1994, notices of proposed 
rulemaking, the models do compute habitat restoration costs and 
assimilative capacity restoration costs for the same release.
    Response: The models compute assimilative capacity restoration 
costs for any toxic mass that remains in the environment either because 
no habitat restoration action is taken or because habitat restoration 
does not fully remove the toxic mass. The preamble and technical 
documents have been clarified. See Section 5.4.4, Volume I of the 
NRDAM/CME and NRDAM/GLE technical documents.
    Comment: One commenter noted that there are now 43, not 42, IJC 
areas of concern in the Great Lakes.
    Response: At the time that the proposed NRDAM/GLE was being 
developed, there were only 42 IJC areas of concern. As more IJC areas 
of concern and NST sites are identified, the Department will consider 
updating the models in future biennial reviews. However, it was not 
feasible for the Department to revise the NRDAM/GLE to account for this 
one additional site within the available time frame.
    Comment: One commenter questioned the use of standardized sediment 
LC50s

[[Page 20601]]

based solely on bulk sediment concentrations.
    Response: The LC50s used in the models are for pore water 
concentrations, not bulk sediment concentrations.
    Comment: One commenter stated that the proposed NRDAM/GLE generated 
inappropriately high assimilative capacity restoration costs for 
releases of metals. The commenter thought these inappropriate costs 
resulted from underestimating the rate at which metals attenuate in the 
Niagara River and Lake Ontario. The commenter stated that metals fall 
to the bottom, are buried, and pose no toxicity threat.
    Response: As discussed in Section VII.G of this preamble, the 
Department has eliminated all pure metals from the chemical database. 
For metal compounds, the Department believes that the NRDAM/GLE 
adequately accounts for attenuation. The model calculates the fate of 
the released substance by partitioning the dissolved fraction from the 
particulate fraction in both the water and the sediments. The LC50s 
used in the model for assimilative capacity calculations are limited to 
those for dissolved chemicals.

L. Restocking

    Comment: Several commenters questioned the reasonableness of the 
methodology for invoking and calculating restocking costs. A few 
commenters thought that the proposed NRDAM/CME should not include 
restocking costs for species, such as dolphins, polar bears, eagles, 
and alligators, that have never been and are unlikely to be restocked 
after a spill. These commenters stated that including restocking costs 
for such species was particularly troubling in light of the weakness of 
the model's underlying estimates of mortality for these species. 
Further, the commenters thought the Department's restocking cost 
estimates for captive breeding programs, in some cases, came from 
unpublished sources and lacked real-world precedent. The commenters 
noted, for example, that there was no reason to assume that osprey and 
raptors cost the same to restock as eagles. Finally, another commenter 
considered the salmon restocking costs included in the proposed NRDAM/
CME to be unrealistically high.
    Response: The Department agrees that it is highly unlikely that 
trustees would restock polar bears and, thus, has deleted polar bear 
restocking costs from the NRDAM/CME. The Department has limited the 
rest of the restocking component to those species that are actually 
available from hatcheries or commercial suppliers and used the actual 
market prices of acquiring such species. The Department believes that 
this approach adequately ensures that the restoration costs in the 
model are realistic and reasonable. The Department has added Table 12.7 
to Volume I of the NRDAM/CME technical document to present the data 
obtained from the International Animal Exchange, Inc.
    Bald eagles and other raptors have been restored by hand-rearing 
hatchlings in a number of States. The models use the same costs for all 
raptors because the activities and effort required are similar for all 
species. The eagle restocking cost may be considered a general figure 
for all raptors.
    The Department has revised the calculations for fish restocking 
costs in the NRDAM/CME. The proposed NRDAM/CME used an average size for 
all salmon species. The final model uses species-specific parameters 
for salmon in each of the three applicable provinces. See Table 12.6, 
Volume I of the NRDAM/CME technical document.
    Comment: One commenter questioned the assumed source of restocked 
animals, noting that simply moving animals from one location in the 
wild to another would still leave the public with a net loss.
    Response: The Department believes that translocation of animals in 
the wild may, under some circumstances, be an appropriate restoration 
action. However, for purposes of the type A procedures, the Department 
chose to consider only restocking of captive-bred animals, the cost of 
which is generally lower than the cost of translocation in the wild. 
See Table 12.7, Volume I of the NRDAM/CME technical document.
    Comment: One commenter thought that it would be more appropriate to 
consider predator control than restocking for canvasback and redhead 
ducks in the Great Lakes.
    Response: After recovering damages through the use of a type A 
procedure, trustees may well decide that the recoveries are better 
spent on predator control rather than restocking. However, the 
feasibility, effectiveness, and methods of predator control are highly 
site-specific. The Department does not believe that adequate data are 
currently available to include predator control among the list of 
restoration actions evaluated by the models.
    Comment: One of the independent technical reviewers thought the 
restocking scenario described in the proposed NRDAM/CME technical 
document would result in a net loss of fish and wildlife.
    Response: The scenario does not result in a net loss of fish and 
wildlife because the model calculates a loss and allows only one-to-one 
replacement, with correction for restocking survival, of missing 
individuals. The Department has clarified the technical document. See 
Section 5.4.3, Volume I of the NRDAM/CME technical document.
    Comment: One of the independent technical reviewers noted that the 
proposed NRDAM/CME technical document stated that only fish that would 
be caught are restocked in the model. The technical reviewer thought 
that all fish killed should be restocked, not just those ultimately 
caught.
    Response: The models assume that all fish and shellfish killed are 
restocked, if stocks are available. The technical document has been 
clarified. See Section 5.4.3, Volume I of the NRDAM/CME technical 
document.
    Comment: One of the independent technical reviewers thought that 
the models should grant habitat restoration priority over restocking.
    Response: The models decide whether to invoke habitat restoration 
independently from the decision whether to restock. Therefore, 
prioritization is not necessary.

M. Consideration of Costs and Benefits of Active Restoration

    Comment: Numerous commenters, including some of the independent 
technical reviewers, criticized the models for failing to consider 
whether the various active restoration alternatives were warranted in 
light of the relationship between the benefits of those actions and the 
costs of the alternatives. Some commenters offered examples where the 
models computed restoration costs that were millions of dollars or 
hundreds of times greater than estimated compensable values. Many 
commenters thought that the models should incorporate a decision rule 
to screen out restoration actions that would impose grossly 
disproportionate costs.
    Response: As it indicated in the March 25, 1994, type B rulemaking, 
the Department believes that the relationship between costs and 
benefits is an important factor in selecting an appropriate restoration 
action. See 59 FR at 14271. The Department acknowledges that the 
proposed rules and models did not explicitly address this factor. After 
careful consideration, the Department has revised the models to perform 
a cost-benefit analysis of habitat restoration and restocking actions.
    If the relevant active habitat restoration alternative would reduce

[[Page 20602]]

compensable value or if restocking is possible, then the submodel 
performs a cost-benefit test of these forms of active restoration. The 
submodel compares the total costs of active habitat restoration and 
restocking against the measured benefits of such restoration (i.e., 
compensable value assuming natural recovery minus compensable value 
assuming active restoration). If the costs exceed ten times the 
measured benefits, then the submodel assumes, for purposes of 
generating a damage figure, that natural recovery, rather than active 
restoration, will be used to reestablish baseline conditions. If the 
costs do not exceed the measured benefits by ten times, then the 
submodel assumes, for purposes of generating a damage figure, that 
habitat restoration and restocking will be implemented.
    The Department determined in the March 25, 1994, rulemaking that 
although cost-benefit considerations are an important factor in 
selecting an appropriate restoration action when a trustee uses type B 
procedures, the exact determination of how to evaluate this factor 
should be resolved on a case-by-case basis. Therefore, 43 CFR 11.82(d) 
lists the relationship between costs and benefits as one of several 
factors that trustees must evaluate before selecting a restoration 
action when using type B procedures. The Department continues to 
believe that this is the most appropriate approach in the type B 
context. However, in the type A context, where a model, rather than the 
trustees, determine the range and type of restoration actions on which 
to base the damage claim, and where the intent is to minimize the level 
of analysis that trustees must conduct, the Department believes it is 
appropriate to impose a bright-line standard.
    The Department has concluded that the evaluation of the costs and 
benefits of active habitat restoration and restocking versus natural 
recovery should focus on incremental costs and benefits. Therefore, the 
models compare the total costs of active habitat restoration and 
restocking against compensable value assuming natural recovery minus 
compensable value assuming active restoration.
    When determining what an appropriate cost-benefit ratio would be in 
the type A context, the Department considered the dicta in Ohio v. 
Interior suggesting three-to-one as a possible ratio. See 880 F.2d at 
443-44, n. 7. However, the NRDAM/CME and NRDAM/GLE quantify only a very 
narrow range of compensable values. Thus, the Department does not 
believe that a three-to-one ratio is appropriate. Therefore, the 
Department was left to make this policy determination without the 
benefit of clear empirical standards or legal precedents. The 
Department has selected a ratio of ten-to-one based on its sense of 
fairness and reasonableness.
    Although the models do impose a uniform standard, the Department 
continues to believe that a truly ``correct'' cost-benefit ratio 
depends on site-specific factors. The standard contained in today's 
models is not intended to suggest that a similar ratio is appropriate 
in a type B context but rather has been included in recognition of the 
unique nature of the type A procedures. Even when trustees use type A 
procedures, if the ratio is exceeded, they may nevertheless conclude 
that compensable values assuming natural recovery as determined by the 
models will not provide adequate funding for necessary restoration 
actions. In such cases, trustees are free to calculate damages using 
type B procedures.
    Finally, the Department has chosen not to apply the cost-benefit 
test to assimilative capacity restoration. The Department believes that 
assimilative capacity does have an economic value. However, the 
Department is unaware of any economic study that calculates the 
consumer surplus or economic rent associated with assimilative 
capacity. Accordingly, the Department has not included assimilative 
capacity in the cost-benefit test since its inclusion does not affect 
the calculation of compensable values.

N. Damages for Fishing and Hunting Losses

    Comment: Some of the independent technical reviewers noted that 
closures to recreational fishing represent a change in access rather 
than in catch rate. These technical reviewers thought that random 
utility models (RUMs) should be used to value both the changes in catch 
rates and in access, noting that RUMs are designed to capture 
substitution across sites.
    Response: The Department acknowledges that recently evolved 
techniques for resource valuation could potentially improve the 
calculation of damages in the models. However, inclusion of such 
techniques would require considerable additional work, including the 
development of a RUM describing recreational choices across the broad 
geographic regions covered by the models and a database containing the 
parameters required as inputs to the RUM. The Department has concluded 
that such additional work is not feasible at this time. The Department 
may reconsider this issue in future biennial reviews of the models.
    Comment: Commenters objected to the calculation of damages for lost 
recreational fishing on the grounds that the Department had failed to 
link the injury to a reduction in services by using trip values rather 
than marginal (per fish) values. Other commenters, including some of 
the independent technical reviewers, noted that all species are 
assigned to a single mode of recreational fishing. They stated that 
data are available on percentage caught by each mode, and recommended 
these data be included in the models to weight the recreational fishing 
values.
    Response: Though not fully explained in the proposed NRDAM/CME and 
NRDAM/GLE technical documents, the Department established the link 
between the injury and a reduction in services by calculating average 
trip values (dollars of trip value per kilogram of fish caught) and 
then adjusting these values by applying ratios of average trip values 
to marginal values (additional trip value per fish caught) that were 
obtained from studies that have compared these values. The Department 
has revised the technical documents. See Section 9.3.4, Volume I of the 
NRDAM/CME technical document, and Section 6.3, Volume I of the NRDAM/
GLE technical document.
    The Department acknowledges that additional data on percentage of 
fish caught by different modes of fishing are becoming available. 
However, due to the fact that the majority of species in a particular 
area tend to be caught by a dominant mode of fishing and that there are 
not always major differences between values in the various modes of 
fishing, the Department does not believe that weighting recreational 
fishing values using data on percentages caught by each fishing mode 
would significantly improve the reliability of the final damage figure. 
As additional data become available, the Department will reconsider 
this issue in future biennial reviews.
    Comment: One commenter noted that, in an attempt to maximize total 
utility, anglers tend to reduce the number of fishing trips they take 
as the marginal utility from fishing falls in response to a reduction 
in the catch rate. This commenter argued that the models should, 
therefore, account for reductions in participation in addition to 
reductions in catch. The commenter suggested that the models estimate 
reductions in participation for spills that exceed some given threshold 
volume.
    Response: The Department believes that, while participation in 
fishing activities may be affected by moderate or large-scale spills, 
small spills are

[[Page 20603]]

likely to affect primarily the quality of such activities. 
Specifically, the Department believes that the principal effect of 
small spills on recreational anglers is a reduction in catch rather 
than a reduction in fishing trips. The commenter appears to agree with 
this position by suggesting that the models account for reductions in 
participation for spills over a given threshold size. However, where 
participation is reduced, trustees may conduct supplemental type B 
studies.
    Comment: One commenter thought that the models inappropriately 
assumed that species that lacked catch data had no recreational value.
    Response: The Department recognizes that there may be species for 
which catch data are not available that nevertheless have recreational 
value. However, the only fish-related compensable value that the models 
compute is for lost harvests. Without catch data, the models cannot 
determine lost harvests and, thus, cannot compute compensable value. If 
an injured fish species has values unrelated to harvest, then trustees 
may conduct supplemental type B studies to capture those values.
    Comment: A few commenters, including some of the independent 
technical reviewers, thought that the lost recreational fishing and 
hunting values used in the models were outdated.
    Response: The Department has not updated the recreational fishing 
and hunting values in the models. The Department will revisit this 
issue during the next biennial review.
    Comment: One of the independent technical reviewers questioned why 
the hunting values included in the proposed models did not account for 
the effects of changes in effort.
    Response: The Department is not aware of any evidence that effort 
changes in cases of minor spills.
    Comment: Some of the independent technical reviewers recommended 
that the Department include habitat equivalency analysis (HEA) in the 
models.
    Response: HEA is a method of determining damages for interim losses 
that does not require the explicit calculation of the economic value 
lost by the public. Instead, HEA bases damages on the cost of obtaining 
or creating additional acreage that would provide the same habitat 
services as that lost pending recovery of the injured resources. While 
HEA has merit, the Department has elected not to pursue its inclusion 
in the type A models at this time. The use of HEA to compute 
compensation for interim losses is an issue that extends beyond this 
rulemaking. HEA is currently not listed as one of the type B 
methodologies for calculating compensable value, although the 
regulations do allow use of additional unlisted methodologies if they 
meet certain criteria. The Department is conducting a biennial review 
of the type B methodologies and will be examining the use of HEA in 
that context. The Department believes that inclusion of HEA in a type A 
procedure should await the resolution of the biennial review of the 
type B procedures. Also, the availability and cost of obtaining habitat 
equivalent to that injured is highly site-specific. The Department 
currently does not have adequate data to incorporate HEA into the 
models.
    Comment: Commenters objected to the method used to value lost 
subsistence fishing. The commenters stated that the proposed NRDAM/CME 
does not clearly define subsistence anglers and inappropriately assumes 
that the full value of the subsistence resource is lost without 
considering substitutes. In addition, the commenters argued that the 
proposed NRDAM/CME measures subsistence loss by the gross cost of an 
alternate food supply (rather than considering net subsistence losses) 
and adjusts the costs to account for supposed differences in protein 
between store-bought and wild-harvested fish based on a study of birds, 
not fish. Several commenters thought that inclusion of damages for lost 
subsistence fishing was legally impermissible. On the other hand, one 
of the independent technical reviewers noted that subsistence hunting, 
not only subsistence fishing, is significant in Alaska and should be 
included in the model.
    Response: The Department acknowledges that the proposed NRDAM/CME 
failed to clearly define subsistence anglers, failed to consider 
substitutes, and inappropriately measured subsistence loss as the gross 
cost of an alternate food supply. All of these shortcoming raise 
significant questions about the reliability of the model's calculation 
of compensable value for subsistence loss. Data are unavailable at this 
time to correct these problems. Therefore, the Department has decided 
to delete subsistence losses from the NRDAM/CME. Because the Department 
has decided to delete subsistence losses for technical reasons, it is 
not necessary for the Department to address the legal permissibility of 
trustees recovering natural resource damages for subsistence losses.
    Comment: Some commenters, including some of the independent 
technical reviewers, objected to the inclusion in the models of damages 
for loss of commercially harvested fish and furbearers. These 
commenters rejected the explanation that inclusion of damages for such 
losses was designed to compensate the public for lost economic rent. 
The commenters noted that the government does not in fact charge rent 
for commercial harvests and concluded that the public, therefore, does 
not incur any loss of economic rent. Commenters argued that the models 
were not capturing the public's lost economic rent but rather were 
inappropriately calculating the commercial users' private losses, which 
are not recoverable as natural resource damages under CERCLA. The 
commenters cited Satsky v. Paramount Communications, Inc. for the 
proposition that trustees may only bring claims ``for injuries to 
interests which all citizens hold in common.'' 7 F.3d 1464, 1470 (10th 
Cir. 1993). The commenters argued that private economic interests, such 
as commercial losses, are not interests that all citizens hold in 
common. Citing various floor debates on the Superfund Amendments and 
Reauthorization Act, the commenters argued that Congress intended the 
double recovery prohibition to bar natural resource damage claims for 
losses that are subject to private recovery. Commenters also noted that 
if damages for such losses are retained in the models, then serious 
double recovery problems arise because commercial users will assert 
overlapping claims.
    Response: The type A models include damages for lost commercial 
harvests in order to capture lost economic rent. The issue of whether 
lost economic rent is a legally permissible category of damages was 
decided and resolved by the Department in 1986 and is beyond the scope 
of this rulemaking. The only issues that the Department is considering 
in this rulemaking are: whether economic rent is, in fact, generated by 
commercially harvested species; if so, whether the type A models 
correctly calculate any loss of that economic rent resulting from 
releases covered by the models; and whether the rule adequately 
protects against double recovery.
    In the preamble to the original type B rule, economic rent is 
defined as ``the excess of total earnings of a producer of a good or 
service over the payment required to induce that producer to supply the 
same quantity currently being supplied.'' 51 FR at 27691. In other 
words, economic rent for commercially harvested resources is the fee 
that commercial harvesters could pay to the government and still find 
harvesting economically feasible.

[[Page 20604]]

    Commercial harvesters invest capital in equipment (e.g., gear, 
traps, and boats). This capital could have been liquidated and put to 
another use, such as investment in a bank. Therefore, commercial 
harvesting is worthwhile only if the harvester receives a price for the 
harvest that both covers labor and fuel costs as well as provides a 
reasonable return on capital. To the extent that the harvester receives 
a price that exceeds costs plus a reasonable return on capital, the 
government could charge a fee and the harvester would continue to 
engage in harvesting. Thus economic rent is generated.
    The Department believes that economic rent is being generated by 
commercially harvested fish and wildlife. The one situation in which 
economic rent is clearly eliminated is when natural resources are 
exploited to the point that all profits, including economic rent, have 
been competed away. This situation arises when commercial harvests are 
not regulated and, thus, harvesters have free and unlimited access. 
Economic theory predicts that in such cases harvesters will ignore both 
the value of the resources for future use as well as the costs of 
crowding. Fisher, A.C., Resource and Environmental Economics, New York: 
Cambridge Univ. Press, 1981.
    One way of preventing this situation and generating economic rent, 
is to charge fees. Currently, resource managers do not generally charge 
fees except to cover administrative costs of processing permits. See, 
e.g., Magnuson Act, 16 U.S.C. 1854(d). However, resource managers do 
regulate commercial harvesting through limits on the gear that may be 
used, limits on the length of the harvest season, catch restrictions, 
tradeable permits for limited entry or individual catch quotas, and 
other programs. For example, NMFS has recently established a pilot 
program to buy back fishing permits to restore stocks of cod, haddock, 
and flounder in the Atlantic Ocean. These programs are designed to 
protect the resources for future use. The Department believes that 
these programs do in fact curb overexploitation of stocks and thus 
prevent profits (including economic rent) from being driven down to 
zero. For example, Alaska salmon and herring fisheries are regulated 
with tradeable entry permits. The aggregate value in 1988 of all 
permits in the salmon, herring, and herring roe fisheries was $925 
million.
    Market prices are largely set on a national or international basis. 
Therefore, in the case of minor spills, it is unlikely that market 
prices will change. Also, total biomass effects should be limited for 
minor spills, thus it is unlikely that there will be long-term effects 
on the catchability of resources. Finally, the Department believes it 
is reasonable to assume for minor spills that the same number of 
commercial harvesters will continue to expend the same amount of 
economic resources to conduct harvests and that the markets for the 
necessary labor and capital inputs to commercial fishing are 
competitive. Therefore, the type A models compute damages based on the 
harvesters' forgone revenue, which is the market price the harvesters 
could have received at the time the resources would have been 
harvested. The Department believes that this figure will capture lost 
economic rent.
    With regard to potential double recovery, the Department 
acknowledges that in some cases commercial harvesters may bring private 
causes of action that include economic rent. As noted above, resource 
managers generally do not charge fees to capture economic rent; 
therefore, when commercial harvesters sell their harvests they obtain a 
profit that includes economic rent. Under OPA, commercial harvesters 
have a specific private cause of action. CERCLA does not grant 
commercial harvesters a private cause of action. However, some 
commercial harvesters, such as commercial fishermen in coastal waters, 
probably do have private causes of action for hazardous substance-
related injuries under State law or common law. When commercial 
harvesters bring a claim for lost profit it will most likely include 
economic rent. Therefore, if trustees also bring a claim for lost 
economic rent there will be a potential double recovery problem.
    The governmental regulation of commercial fish and wildlife harvest 
implies a public concern that these resources be managed in order to 
sustain their contribution to economic productivity. The public's value 
for commercial harvesting is further reflected in express policy 
statements that the government is committed to promoting resources' 
contribution to economic productivity. See, e.g., Fish and Wildlife Act 
of 1956, 16 U.S.C. 742a. Therefore, when there are reductions of 
commercial harvests, the public suffers a loss.
    If commercial harvesters can and do bring a private cause of 
action, then the harvesters may be fully compensated and the public's 
interest in promoting commercial harvests may be satisfied. If 
harvesters are compensated for full social losses, then trustees should 
not recover separate damages for lost economic rent. However, in some 
cases commercial harvesters may not have a private cause of action or 
their recoveries may be subject to geographic or temporal limitations. 
For example, commercial harvesters may be limited to recovering damages 
incurred during the period of formal closure, or incurred in the area 
closed or in the area in which fish were directly exposed to the 
released substance. See, e.g., Golnoy Barge Co. v. M/T Shinoussa, Civ. 
No. H-90-2414, 1993 WL 735038 (S.D. Tex. Aug. 17, 1993). In other 
cases, commercial harvesters may choose not to bring private causes of 
action. If commercial harvesters do not obtain direct full 
compensation, then the public's interest is not satisfied and the 
trustees may bring a claim for lost economic rent.
    For minor releases where damages may be relatively low and data 
establishing injury and causation may be difficult to obtain, the 
Department believes that it is unlikely that commercial harvesters will 
go to the expense and trouble to pursue a legal claim. Therefore, the 
Department has retained the calculation of lost economic rent in the 
models. However, to prevent double recovery, Sec. 11.44(d) provides 
that if the trustee is aware of reliable evidence that a private party 
has recovered damages for commercial harvests lost as a result of the 
release, the trustee must eliminate from his or her claim any damages 
for such lost harvest that are included in the lost economic rent 
calculated by the model. When the Assessment Plan is made available for 
public review and comment, PRPs and commercial harvesters will have an 
opportunity to alert trustees to any private actions for lost 
commercial harvests.
    Comment: A few commenters, including some of the independent 
technical reviewers, thought that the prices used to calculate damages 
for lost commercial harvests were invalid because they did not account 
for seasonal and regional variations. Commenters also stated that the 
data used were from 1984 through 1988 and should be updated.
    Response: The models account for seasonal and regional differences 
in commercial prices to the extent possible given available data. In 
the Great Lakes, pelt prices were available on a State-by-State basis. 
The Department used these prices to develop average prices for each of 
the Great Lakes. In the coastal areas, pelt prices were available only 
on a regional basis.
    At the time the Department performed the bulk of the work on the 
commercial fisheries component, 1984 through 1988

[[Page 20605]]

was the most recent 5-year period for which final statistics were 
available for the offshore zone in which the catch occurred (rather 
than ports where harvest was taken). The Department decided to average 
the figures over a five-year period to eliminate short-term 
variability. The Department believes that the commercial fishing 
statistics currently incorporated in the models are reliable. However, 
as additional data become available, the Department will consider 
updating the models during future biennial reviews.

O. Damages for Lost Wildlife Viewing

    Comment: One commenter supported the proposed approach for 
calculating damages for lost wildlife viewing as reliable and 
reasonable. However, most commenters criticized the approach. 
Commenters noted that the Department itself appeared to have serious 
reservations about the approach. A number of commenters thought that 
the approach would consistently underestimate damages; others thought 
that damages would be overestimated. Some commenters stated that the 
Department's methodology for calculating wildlife viewing losses was so 
unreliable that such losses should be deleted from the models.
    Several commenters thought it was incorrect to assume that a given 
reduction in wildlife population would produce a comparable reduction 
in the wildlife seen. Thus, the commenters concluded that the 
Department was unable to link the injury to a reduction in services. 
The commenters further stated that the calculation used inappropriate 
assumptions and studies to calculate the value of a wildlife viewing 
trip on a per-animal basis.
    A few commenters thought the values appear arbitrary since they 
vary so widely by province. One commenter suggested that the wide 
variations in province values be eliminated either through the use of 
uniform average values or through the deletion of extreme values. On 
the other hand, some commenters thought the models did not adequately 
account for the variable characteristics of affected sites.
    Response: In the notices of proposed rulemaking, the Department 
solicited comment on a number of aspects of the proposed methodology 
for determining damages for lost wildlife viewing. 59 FR at 40328-29 
and 63311-12. Based on its careful consideration of the comments it 
received, as well as its own reexamination of the models, the 
Department has modified the methodology and believes that it is sound.
    Wildlife viewing is one of the most significant direct use services 
that wildlife provide to humans. Inclusion of lost wildlife viewing 
damages in the type A models necessitated that the Department draw 
conclusions about the relationship between changes in wildlife 
populations and changes in the perceptions of wildlife by viewing 
participants. There is no known empirical research that indicates how 
participants perceive changes in wildlife populations. The proposed 
models assumed that a given percentage reduction in wildlife 
populations results in the same percentage reduction in the wildlife 
seen by participants.
    The Department has determined that percentage reductions in 
wildlife viewing perception could be lower or higher than the 
percentage reduction in wildlife populations. For example, a wildlife 
population may consist of 100 animals yet the public may only perceive, 
on average, 50 of those animals. If a spill were to kill 50 percent of 
the population, then the public's perception of the population would be 
reduced by 20 percent if the 50 killed animals included only 10 of the 
animals the public normally sees. However, if the 50 killed individuals 
included 30 of the animals the public normally sees, then the change in 
perception would be 60 percent. Therefore, in the absence of empirical 
evidence on the subject, the Department has decided to assume that 
percentage reductions in wildlife viewing perception are equal to 
percentage reductions in wildlife populations.
    When determining how a given reduction in the number of animals 
viewed affects the value the public derives from viewing, the 
Department relied on the only available studies that identified a 
marginal value for wildlife viewing (Loomis et al. (1989) and Cooper 
and Loomis (1991)). The Department believes that these studies are 
reliable. The Department acknowledges that one of the studies dealt 
with trips that were not taken primarily for the purpose of wildlife 
viewing. However, contrary to the assertions of some commenters, the 
studies specifically examined how changes in the number of animals seen 
affected the value of the trips. Therefore, the wildlife viewing values 
used in the type A models do not reflect any other non-viewing aspects 
of recreational trips.
    The biological effects submodel quantifies wildlife mortality in 
terms of the number of animals killed. Therefore, the Department needed 
to develop per-animal viewing values. The Department developed such 
per-animal values by using Loomis et al. (1989) and Cooper and Loomis 
(1991) to establish the relationship between changes in wildlife seen 
and changes in value and by using FWS' National Survey of Fishing, 
Hunting, and Wildlife-Associated Recreation to establish total viewing 
values for particular species. The Department first estimated the total 
value of wildlife viewing at ocean- or lake-side for an entire State. 
Then the Department allocated this total value among species and 
wildlife individuals within species that reside along the State's 
ocean- or lake-side. For further discussion of this methodology, see 
Section 8.4, Volume I of the NRDAM/CME technical document, and Section 
6.4.1, Volume I of the NRDAM/GLE technical document.
    The Department recognizes that the per-animal viewing values 
assigned by the models for some species vary widely by province. 
However, the Department has concluded that these variances are not 
errors that need to be corrected but, rather, reflect actual and 
relevant regional differences. The value that the public derives from 
viewing a particular animal in a particular area depends on how many 
people engage in wildlife viewing in that area and how many animals of 
that type there are to view in that area. Therefore, the models 
appropriately contain relatively low per-animal values for species that 
are abundant and for areas with low participation in wildlife viewing. 
Conversely, the models contain higher per-animal values for less 
abundant species and for areas with higher wildlife viewing 
participation.
    Not only are the regional variances in per-animal values 
appropriate, but also, such variances do not lead to unrealistic 
differences in the damage figures calculated by the models for 
particular releases. When determining damages for a particular release, 
the models calculate damages for lost wildlife viewing based on the 
probability that the release will kill wildlife. The probability that 
an animal will come into contact with, and be killed by, a release is 
directly related to the wildlife abundance in the area affected by the 
release. In areas with low wildlife abundance, minor releases would 
have a low probability of killing wildlife. Therefore, although the 
per-animal values in these areas may be significantly higher than in 
areas of high abundance, the actual damage figure may not be.
    Comment: Some commenters contended that the methodology for 
calculating viewing damages relies on assumptions that overstate the 
number of wildlife viewing trips to affected

[[Page 20606]]

areas. Some commenters also claimed that the Department had used 
unreliable survey data to derive per-animal viewing values. Several 
commenters thought the models used outdated data.
    Response: The Department has revised the wildlife viewing damage 
component of the models to reflect more recent data. When developing 
the proposed models, the Department used FWS' 1985 National Survey of 
Fishing, Hunting, and Wildlife-Associated Recreation. This survey was 
updated in 1991. The Department has revised the final models to reflect 
this update. FWS has conducted wildlife-related recreation surveys 
roughly every 5 years since 1955. The results of these surveys are 
widely disseminated and are relied upon both by governments and by 
private individuals for a range of purposes. The Department believes it 
is appropriate to use these surveys to determine participation rates 
for wildlife viewing as well as the value of wildlife viewing trips.
    Comment: A few commenters, including some of the independent 
technical reviewers, suggested that the proposed models be expanded to 
account for viewing losses incurred by individuals other than residents 
of coastal counties. One commenter stated that in Michigan, regional 
populations are not concentrated either in lake shore counties or at 
substantial distances inland and suggested that the NRDAM/GLE 
incorporate lost wildlife viewing on a two-county deep basis. One of 
the independent technical reviewers noted that the absence of damages 
for lost viewing by out-of-State tourists was a significant problem in 
Alaska.
    Response: The Department has revised the NRDAM/CME to account for 
impacts on the populations of coastal States rather than just coastal 
counties. See Section 8.4, Volume I of the NRDAM/CME technical 
document. Wildlife viewing damage calculations in the NRDAM/CME rely on 
State-level estimates of participation rates for wildlife viewing at 
the ocean. The Department believes that it is more appropriate to apply 
these State-level participation rates to State populations than to 
county populations in the NRDAM/CME.
    The Department has not applied the participation rates in the final 
NRDAM/GLE to the entire State population due to the existence of 
alternative recreation sites in other parts of the Great Lakes States. 
Instead, the Department has generally limited the participation rate to 
the populations of lake-side counties in the NRDAM/GLE. In a few 
instances involving very narrow counties, the Department has gone 
beyond the lake-side county to include residents of areas within an 
average one-way trip distance from the provinces where injury occurred. 
Statewide average participation rates were used in the NRDAM/GLE due to 
a lack of reliable data at the county level.
    The Department recognizes that out-of State visitors may experience 
viewing losses; however, adequate data are not currently available to 
incorporate such losses into models. As more data become available, the 
Department may reconsider this issue in future biennial reviews. 
Meanwhile, the rule allows trustees to conduct supplemental type B 
studies to assess damages for such losses.
    Comment: Some commenters, including a few of the independent 
technical reviewers, suggested including viewing losses incurred by 
individuals under the age of 18. One commenter noted that the 
Department had relied on wildlife participation survey data supplied by 
individuals 16 years or older but used population data that only 
included individuals 18 years or older.
    Response: The Department has revised the models to include viewing 
losses experienced by individuals 16 years of age and older.
    Comment: Some commenters, including a few of the independent 
technical reviewers, thought that spills could affect not only the 
quality of the wildlife viewing trips but also the number of trips 
taken. These commenters thought the models should account for this 
effect.
    Response: Adequate data are not currently available to incorporate 
damages for lost participation in wildlife viewing. Further, while 
participation in wildlife viewing may be affected by moderate or large-
scale releases, minor releases, for which the type A models are 
designed, are likely to affect primarily the quality of such activity. 
As more data become available, the Department may reconsider this issue 
in future biennial reviews.
    Comment: A few commenters thought the models did not account for 
all species that had viewing value.
    Response: The Department believes that the species addressed by the 
models account for the vast majority of viewing values.
    Comment: One commenter noted that the deaths of migratory species 
could result in viewing losses in locations beyond the spill site and 
thought that the models should account for such losses.
    Response: The Department acknowledges that releases can result in 
off-site viewing losses; however, adequate data are not currently 
available to incorporate such losses into models. As more data become 
available, the Department may reconsider this issue in future biennial 
reviews. Meanwhile, the rule allows trustees to conduct supplemental 
type B studies to assess damages for such losses.
    Comment: Some commenters asked whether the models considered the 
effect of closures on wildlife viewing losses.
    Response: The models do not address the effect of closures on 
wildlife viewing. The models only compute mortality-related wildlife 
viewing losses.

P. Damages for Beach and Boating Closures

    Comment: The Department received several comments on the proposed 
NRDAM/CME's methodology for calculating damages for beach closures. A 
few commenters thought that the Department had made unsubstantiated and 
questionable assumptions in estimating beach visitation. Some 
commenters, including one of the independent technical reviewers, 
claimed that the Department had failed to account adequately for 
differences in visitations and values associated with different beach 
types and different seasons. One independent technical reviewer stated 
that beach visitation is highly variable within provinces, and that 
more localized visitation data could be mapped in a GIS from survey 
data.
    Response: The Department could not locate adequate data to develop 
separate per-person values for visiting Federal beaches versus non-
Federal beaches. However, the Department believes the NRDAM/CME 
adequately accounts for differences between the two types of beaches 
through the use of different visitation figures. A sample of monthly 
visitation data was taken from representative beaches in each economic 
province. This sampling was designed to account for differences in 
visitation that exist between economic provinces. The Department also 
collected separate visitation data for National Seashores and other 
public beaches. The user is required to enter the type of beach 
affected (National Seashore or other public beach) as an input to the 
NRDAM/CME. Further, before the compensable value submodel applies an 
average beach use value to the estimated loss in beach use, the 
submodel adjusts the value to account for seasonal variations. The 
NRDAM/CME uses seasonal variations in the duration of beach visits as a 
reasonable approximation of the seasonal variations in beach use 
values. Data on the duration of beach visits were available

[[Page 20607]]

for National Seashores, but not for other public beaches. Therefore, 
the Department applied the National Seashore duration data to other 
public beaches. For further discussion, see Section 10, Volume I of the 
NRDAM/CME technical document.
    Comment: Several commenters raised concerns about valuation of 
beach closures in the proposed NRDAM/GLE. Some commenters criticized 
the use of studies of saltwater beaches to value Great Lakes beach use. 
A few commenters stated that the studies the Department used to value 
Great Lakes beach use were outdated. Other commenters complained that 
the NRDAM/GLE used the same values for Federal and non-Federal beaches.
    Response: In the NRDAM/GLE, the per-day value of beach recreation 
is based on the average of several values reported for general beach 
recreation. The Department acknowledges that basing beach-related 
recreation values on studies carried out exclusively at Great Lakes 
locations would have been ideal. However, no such studies were 
available. Intuitively, beach recreation on the Great Lakes combines 
some aspects of seashore beach recreation with some aspects of 
freshwater beach recreation. Therefore, the Department used a range of 
studies, including both studies of saltwater beach recreation as well 
as studies of freshwater beach recreation.
    The Department believes that the data used are still the best 
available. If additional data on the value of beach-related recreation 
in the Great Lakes become available, the Department will consider 
incorporating such data in the NRDAM/GLE during future biennial 
reviews.
    Data limitations required that the Department make assumptions 
concerning the distribution of monthly trips to non-Federal lake 
shores. The NRDAM/GLE assumes that the distribution is identical to the 
distribution of monthly trips to National Lake shores. See Section 
6.4.2, Volume I of the NRDAM/GLE technical document.
    The Department could not locate adequate data to develop separate 
per-person values for visiting Federal beaches versus non-Federal 
beaches. However, the Department believes the NRDAM/GLE adequately 
accounts for differences between the two types of beaches because the 
per-person values translate into different per-meter values for Federal 
and non-Federal beaches.
    Comment: Some of the independent technical reviewers were confused 
by the discussion in the proposed technical documents on consideration 
of substitutes when valuing beaches. They recommended that values from 
appropriate studies be selected based on an understanding of the 
welfare economics of substitutes. Specifically, they expressed concern 
over the Department's use of studies of loss of beach use over large 
regions. The technical reviewers thought that such studies would not 
account for individuals' ability to substitute one beach for another 
within a given region and, thus, would not correctly capture the value 
of particular beaches.
    Response: The Department did use studies that account for the 
effect of substitution. See Section 10.3.2, Volume I of the NRDAM/CME 
technical document.
    Comment: One of the independent technical reviewers suggested that 
losses on private beaches should be included.
    Response: CERCLA authorizes trustees to pursue claims only for 
injuries to public resources, namely those resources ``belonging to, 
managed by, held in trust by, appertaining to or otherwise controlled 
by'' the United States, a State, or an Indian tribe. CERCLA sec. 
101(16). Private beaches will generally not constitute public 
resources. Therefore, the rule only allows trustees to use the type A 
models for public beaches.
    Comment: One of the independent technical reviewers stated that the 
NRDAM/CME should be modified to allow for the recovery of damages for 
lost use of rocky shoreline.
    Response: The Department was unable to identify sufficient data on 
the relative value of recreation on different shoreline types in 
coastal and marine environments to permit distinctions between rocky 
and sandy shorelines. Therefore, the models compute damages for lost 
use of all types of closed Federal or State beaches; however, the 
models assign the same values regardless of habitat type. The 
Department believes that this approach is reasonable in light of the 
geographic breadth of the studies used to develop the beach values in 
the models. Further, although the August 1994 proposed rule required 
trustees who used the NRDAM/GLE to specify whether a closed beach was 
rocky or sandy, the model itself did not assign different values based 
on that designation. Therefore, the Department has revised the final 
rule to remove the requirement.
    Comment: Several commenters thought the models used unreliable 
boating area coefficients that were likely biased upward. The 
commenters complained about the use of freshwater boating studies to 
determine a boating value for coastal and marine waters in the proposed 
NRDAM/CME. The commenters stated that marine boating is different from 
freshwater boating. Some commenters, including some of the independent 
technical reviewers, thought that the coastal regions of the United 
States were so heterogenous that applying one boating value to the 
entire coastline of the United States was not appropriate.
    Further, the commenters stated that the proposed NRDAM/GLE 
inappropriately assumed that all boat trips were evenly distributed 
along the Great Lakes shoreline, which would result in an overestimate 
of damages. These commenters also thought the studies used to derive 
boating value in the NRDAM/GLE were unpublished and outdated, and 
inappropriate for the transfer of values.
    Response: The Department was unable to establish the similarity of 
marine boating to freshwater boating for the NRDAM/CME. Therefore, the 
Department has eliminated the calculation of lost boating values from 
the NRDAM/CME.
    As to the NRDAM/GLE, the density of recreational boating is higher 
near ports. Thus, due to the greater likelihood of spills occurring 
near port facilities, the Department believes the even allocation of 
boating trips used in the model could tend to underestimate the number 
of boating trips that would be affected by a small spill rather than 
overestimate the value.
    The Department does not agree that the models calculate only 
upwardly biased boating damages. In the NRDAM/GLE, lost trips are 
estimated using a formula that distributes the estimated number of 
trips across the total area of water covered by the near shore 
forecast. In some areas this may result in underestimates of the number 
of affected boating trips. For example, areas near major ports or 
marina facilities are likely to have a higher density of boating trips 
than is assumed in the NRDAM/GLE. Nevertheless, the Department believes 
that in those areas when type B procedures cannot be performed at a 
reasonable cost, trustees should have the option of using the NRDAM/
GLE.

Q. Judicial Review and the Rebuttable Presumption

    Comment: Some commenters supported proposed Sec. 11.91(c)(2), and 
the Department's related statements regarding which elements of a type 
A damage assessment can be challenged following application of the 
NRDAM/CME or NRDAM/GLE in a specific case, and which aspects of the 
rule (which

[[Page 20608]]

incorporates the models) must be challenged within 90 days of 
promulgation, as provided by section 113(a) of CERCLA. Proposed 
Sec. 11.91(c)(2) states that judicial review of a type A damage 
assessment shall be limited to the trustee's decision to use the type A 
procedure and the incident-specific data supplied by the trustee. It 
further states that the decision to use the type A procedure and the 
incident-specific data collected by the trustee receive CERCLA's 
rebuttable presumption. The commenters supporting this proposed 
provision suggested that it would reduce the potential for litigation 
and attendant delays to restoration, although they conditioned their 
support on how the Department responded to their other comments.
    Numerous other commenters strongly objected to proposed 
Sec. 11.91(c)(2), as beyond the Department's authority and as contrary 
to the two provisions in CERCLA interpreted and applied in proposed 
Sec. 11.91(c)(2). These commenters asserted that proposed 
Sec. 11.91(c)(2) is outside of the rulemaking authority provided by 
section 301(c)(1) of CERCLA, and therefore beyond the Department's 
legal authority. These commenters stated that while the Department has 
authority to promulgate regulations for assessments, it does not have 
authority to prescribe what effect those assessments will have in 
future judicial proceedings.
    These commenters also asserted that the Department's proposed rule 
language that judicial review of a type A assessment in a specific case 
would be limited to the trustee's decision to use the type A procedure 
and the incident-specific data supplied by the trustee is contrary to 
sections 113 and 107(f)(2)(C) of CERCLA and the Administrative 
Procedure Act (5 U.S.C. 551 et seq.). These commenters stated that 
because CERCLA provides a rebuttable presumption to assessments 
performed in accordance with the regulations, CERCLA allows a PRP to 
introduce any relevant evidence that may rebut the presumption. In the 
view of these commenters, proposed Sec. 11.91(c)(2) would deny PRPs any 
meaningful opportunity to rebut the rebuttable presumption, effectively 
rendering the presumption irrebuttable. These commenters interpreted 
section 113(a) of CERCLA as only barring future judicial review of 
whether the regulations are valid and meet statutory requirements. 
Section 113(a) does not, according to these commenters, preclude 
judicial consideration of all relevant evidence in an assessment not 
performed at the time the regulations are reviewed--e.g., evidence in a 
particular case that the computer model grossly overstates the number 
of affected fish and wildlife. Some commenters suggested that 
application of proposed Sec. 11.91(c)(2) would raise constitutional due 
process questions.
    Finally, some of these commenters objected to the provisions in 
proposed Sec. 11.91(c)(2) that the rebuttable presumption applies to 
the trustee's decision to use the type A procedure and the data inputs. 
These commenters stated that CERCLA's rebuttable presumption is granted 
only to the final results of an assessment, not to intermediate steps 
such as the decision to use the type A procedures and the data input 
selections. These commenters also asserted that the Department had 
provided no criteria for ensuring the accuracy and reliability of site-
specific data inputs, and that absent such regulatory criteria, the 
inputs would not be entitled to the rebuttable presumption. The 
commenters cited the absence of a rebuttable presumption for the 
statement of trusteeship as an example where the Department concluded 
that the absence of guidance resulted in the absence of a rebuttable 
presumption.
    Response: The Department has given this issue careful 
consideration, and decided not to include proposed Sec. 11.91(c)(2) in 
today's final rule. Whether or not a court ultimately would uphold the 
authority of the Department to promulgate proposed Sec. 11.91(c)(2), 
and the interpretation of CERCLA reflected in it, the Department has 
concluded that there is sufficient uncertainty about the precise effect 
of sections 107(f)(2)(C) and 113(a) of CERCLA, as applied in a specific 
case, to warrant leaving proposed Sec. 11.91(c)(2) out of the rule at 
this time.
    The Department recognizes that there are numerous scenarios giving 
rise to natural resource damage claims. Whether or not proposed 
Sec. 11.91(c)(2) as applied would be overly broad could depend on the 
specific circumstances of a given case. Aside from whether the 
commenters objecting to this provision are legally correct, the 
questions raised have convinced the Department that this issue is best 
addressed outside the rulemaking context, at least until additional 
experience has been gained through case-specific application of these 
simplified procedures. Therefore, the Department has concluded for now 
that the precise delineation of CERCLA's preclusive review and 
rebuttable presumption provisions to type A damage assessments is best 
left to specific cases. As a result, the Department need not address 
whether it has the legal authority under section 301(c)(2) to 
promulgate proposed Sec. 11.91(c)(2), nor whether that section 
accurately interprets CERCLA.
    The Department does note that some of the comments interpreting 
CERCLA's rebuttable presumption provision appear largely to read 
CERCLA's preclusive review provision out of the statute. Undoubtedly, 
Congress intended that PRPs have a meaningful opportunity to rebut 
specific aspects of a trustee's case, but not if it is a ``matter with 
respect to which review could have been obtained'' in a challenge to 
the regulations. CERCLA sec. 113(a). In the case of the type A rules, 
the validity of the regulations themselves is closely related to the 
content and workings of the incorporated computer models. The 
Department believes that section 113(a) of CERCLA requires that 
challenges to aspects of the rule that are clearly discernible from the 
rule language, the models, the incorporated technical documentation, 
and this Federal Register notice be brought within 90 days after 
promulgation. Sections 107(f)(2)(A) and 113(a) of CERCLA should be read 
in harmony with one another.
    Also, contrary to the view of some commenters, the statutory 
language of CERCLA does not limit the effect of the rebuttable 
presumption to the ``final results'' of an assessment. Rather, section 
107(f)(2)(C) of CERCLA provides that ``[a]ny determination or 
assessment of damages * * * made * * * in accordance with the 
regulations * * * shall have the force and effect of a rebuttable 
presumption.'' Although the Department has decided not to promulgate 
regulatory language delineating the scope of the rebuttable 
presumption, the statute appears to be worded more broadly than was 
recognized by these commenters. Furthermore, the rule does provide 
procedures requiring trustees to include data inputs in the Assessment 
Plan, which is subject to public review and comment. As such, the rule 
does contain procedural criteria to ensure the accuracy and reliability 
of data inputs through a public review and comment process. The 
commenters also failed to recognize a primary reason for the Department 
declining to afford the rebuttable presumption to the statement of 
trusteeship. Unlike a statement of trusteeship, which is in essence a 
legally-founded assertion, data inputs are factual in nature and can be 
checked, reviewed, and verified through the public comment process.

[[Page 20609]]

National Environmental Policy Act, Regulatory Flexibility Act, 
Paperwork Reduction Act, and Executive Orders 12866, 12630, 12778, and 
12612

    The Department has determined that this rule does not constitute a 
major Federal action significantly affecting the quality of the human 
environment. Therefore, the Department has not prepared any further 
analysis pursuant to section 102(2)(C) of the National Environmental 
Policy Act (43 U.S.C. 4332(2)(C)).
    The Department certifies that this rule will not have a significant 
economic effect on a substantial number of small entities under the 
Regulatory Flexibility Act (5 U.S.C. 601 et seq.). The rule provides 
technical procedural guidance for the assessment of damages to natural 
resources. It does not directly impose any additional cost. As the rule 
applies to natural resource trustees, it is not expected to have an 
effect on a substantial number of small entities.
    This rule does not contain information collection requirements that 
require approval by the Office of Management and Budget (OMB) under the 
Paperwork Reduction Act (44 U.S.C. 3501 et seq.).
    OMB has reviewed this rule under Executive Order 12866. This rule 
does not have takings implications under Executive Order 12630. The 
Department has certified to OMB that this rule meets the applicable 
standards provided in Sections 2(a) and 2(b)(2) of Executive Order 
12778. This rule does not have federalism implications under Executive 
Order 12612.

List of Subjects in 43 CFR Part 11

    Coastal zone, Environmental protection, Fish, Hazardous substances, 
Incorporation by reference, Indian lands, Marine resources, National 
forests, National parks, Natural resources, Public lands, Recreation 
areas, Sea shores, Wildlife, Wildlife refuges.

    For the reasons set out in the preamble, Title 43, Subtitle A of 
the Code of Federal Regulations is amended as follows:

PART 11--NATURAL RESOURCE DAMAGE ASSESSMENTS

    1. The authority citation for Part 11 continues to read as follows:

    Authority: 42 U.S.C. 9651(c), as amended.

Subpart A--Introduction

    2. Section 11.15 is amended by revising the heading and paragraph 
(a)(1) to read as follows:


Sec. 11.15  What damages may a trustee recover?

    (a) * * *
    (1) Damages as determined in accordance with this part and 
calculated based on injuries occurring from the onset of the release 
through the recovery period, less any mitigation of those injuries by 
response actions taken or anticipated, plus any increase in injuries 
that are reasonably unavoidable as a result of response actions taken 
or anticipated;
* * * * *
    3. Section 11.18 is amended by revising paragraph (a)(4) and adding 
a new paragraph (a)(5) to read as follows:


Sec. 11.18  Incorporation by reference.

    (a) * * *
    (4) The CERCLA Type A Natural Resource Damage Assessment Model for 
Coastal and Marine Environments, Technical Documentation, Volumes I-VI, 
dated April 1996, prepared for the U.S. Department of the Interior by 
Applied Science Associates, Inc., A.T. Kearney, Inc., and Hagler Bailly 
Consulting, Inc. (NRDAM/CME technical document). Interested parties may 
obtain a copy of this document from the National Technical Information 
Service, 5285 Port Royal Road, Springfield, VA 22161; PB96-501788; ph: 
(703) 487-4650. Sections 11.34 (a) (b) and (e), 11.35(a), 11.36(b), 
11.40(a), and 11.42(a), and Appendix II refer to this document.
    (5) The CERCLA Type A Natural Resource Damage Assessment Model for 
Great Lakes Environments, Technical Documentation, Volumes I-IV, dated 
April 1996, prepared for the U.S. Department of the Interior by Applied 
Science Associates, Inc., and Hagler Bailly Consulting, Inc. (NRDAM/GLE 
technical document). Interested parties may obtain a copy of this 
document from the National Technical Information Service, 5285 Port 
Royal Road, Springfield, VA 22161; PB96-501770; ph: (703) 487-4650. 
Sections 11.34 (a) (b) and (e), 11.35(a), 11.36(b), 11.40(a), and 
11.42(a), and Appendix III refer to this document.
* * * * *
    4. Section 11.19 is removed and reserved.

Subpart C--Assessment Plan Phase

    5. Section 11.30 is amended by revising the heading and paragraphs 
(a) and (c)(1)(vi) to read as follows:


Sec. 11.30  What does the authorized official do if an assessment is 
warranted?

    (a) If the authorized official determines during the Preassessment 
Phase that an assessment is warranted, the authorized official must 
develop a plan for the assessment of natural resource damages.
* * * * *
    (c) * * *
    (1) * * *
    (vi) Any other Assessment Plan costs for activities authorized by 
Secs. 11.30 through 11.38.
* * * * *
    6. Section 11.31 is amended by revising the heading and paragraphs 
(a)(1), (b), (c) introductory text, (c)(1), and (d) to read as follows:


Sec. 11.31  What does the Assessment Plan include?

    (a) General content and level of detail. (1) The Assessment Plan 
must identify and document the use of all of the type A and/or type B 
procedures that will be performed.
* * * * *
    (b) Identification of types of assessment procedures. The 
Assessment Plan must identify whether the authorized official plans to 
use a type A procedure, type B procedures, or a combination. Sections 
11.34 through 11.36 contain standards for deciding which types of 
procedures to use. The Assessment Plan must include a detailed 
discussion of how these standards are met.
    (c) Specific requirements for type B procedures. If the authorized 
official plans to use type B procedures, the Assessment Plan must also 
include the following:
    (1) The results of the confirmation of exposure performed under 
Sec. 11.37;
* * * * *
    (d) Specific requirements for type A procedures. If the authorized 
official plans to use a type A procedure, the Assessment Plan must also 
contain the information described in subpart D.
    7. Section 11.32 is amended by revising the heading and paragraph 
(c)(1) and adding a new paragraph (f)(3) to read as follows:


Sec. 11.32  How does the authorized official develop the Assessment 
Plan?

* * * * *
    (c) Public involvement in the Assessment Plan. (1) The authorized 
official must make the Assessment Plan available for review by any 
identified potentially responsible parties, other natural resource 
trustees, other affected Federal or State agencies or Indian tribes, 
and any other interested member of the public for a period of at least 
30 calendar days, with reasonable extensions granted as appropriate. 
The authorized official may not perform any

[[Page 20610]]

type B procedures described in the Assessment Plan until after this 
review period.
* * * * *
    (f) * * *
    (3) Paragraphs (f)(1) and (f)(2) of this section do not apply to 
the use of a type A procedure.
    8. Section 11.33 is revised to read as follows:


Sec. 11.33  What types of assessment procedures are available?

    There are two types of assessment procedures:
    (a) Type A procedures are simplified procedures that require 
minimal field observation. Subpart D describes the type A procedures. 
There are two type A procedures: a procedure for coastal or marine 
environments, which incorporates the Natural Resource Damage Assessment 
Model for Coastal and Marine Environments, Version 2.4 (NRDAM/CME); and 
a procedure for Great Lakes environments, which incorporates the 
Natural Resource Damage Assessment Model for Great Lakes Environments, 
Version 1.4 (NRDAM/GLE).
    (b) Type B procedures require more extensive field observation than 
the type A procedures. Subpart E describes the type B procedures.
    9. Sections 11.34 and 11.35 are redesignated as Secs. 11.37 and 
11.38 and new Secs. 11.34 through 11.36 are added to read as follows:


Sec. 11.34  When may the authorized official use a type A procedure?

    The authorized official may use a type A procedure only if:
    (a) The released substance entered an area covered by the NRDAM/CME 
or NRDAM/GLE. Section 3.4, Volume III of the NRDAM/CME technical 
document (incorporated by reference, see Sec. 11.18) identifies the 
areas that the NRDAM/CME covers. Section 6.2, Volume III of the NRDAM/
GLE technical document (incorporated by reference, see Sec. 11.18) 
describes the areas that the NRDAM/GLE covers;
    (b) The NRDAM/CME or NRDAM/GLE cover the released substance. Table 
7.1, Volume I of the NRDAM/CME technical document lists the substances 
that the NRDAM/CME covers. Table 7.1, Volume I of the NRDAM/GLE 
technical document lists the substances that the NRDAM/GLE covers;
    (c) The released substance entered water at or near the surface;
    (d) At the time of the release, winds did not vary spatially over 
the area affected by the release in a way that would significantly 
affect the level or extent of injuries;
    (e) The authorized official is not aware of any reliable evidence 
that, for species that are likely to represent a significant portion of 
the claim, the species biomass is significantly lower than the species 
biomass assigned by the NRDAM/CME or the NRDAM/GLE Tables IV.2.1 
through IV.2.115 and IV.5.1 through IV.5.77, Volume III of the NRDAM/
CME technical document list the species biomasses in the NRDAM/CME. 
Tables III.3.17 through III.3.27 and III.3.40 through III.3.50, Volume 
III of the NRDAM/GLE technical document list the species biomasses in 
the NRDAM/GLE ; and
    (f) Subsurface currents either: are not expected to significantly 
affect the level or extent of injuries; or are reasonably uniform with 
depth over the water column in the area affected by the release.


Sec. 11.35  How does the authorized official decide whether to use type 
A or type B procedures?

    (a) If the authorized official determines under Sec. 11.34 that a 
type A procedure is available, the authorized official must then decide 
whether to use that procedure or use type B procedures. The authorized 
official must make this decision by weighing the difficulty of 
collecting site-specific data against the suitability of the averaged 
data and simplifying assumptions in the type A procedure for the 
release being assessed. The authorized official may use type B 
procedures if they can be performed at a reasonable cost and if the 
increase in accuracy provided by those procedures outweighs the 
increase in assessment costs. Section 1, Volume I of the NRDAM/CME 
technical document (incorporated by reference, see Sec. 11.18) lists 
the simplifying assumptions made in the NRDAM/CME. Volumes III through 
IV of the NRDAM/CME technical document list the data in the NRDAM/CME. 
Section 1, Volume I of the NRDAM/GLE technical document (incorporated 
by reference, see Sec. 11.18) lists the simplifying assumptions made in 
the NRDAM/GLE. Volume III of the NRDAM/GLE technical document lists the 
data in the NRDAM/GLE.
    (b) The authorized official must use type B procedures rather than 
a type A procedure whenever a potentially responsible party:
    (1) Submits a written request for use of type B procedures along 
with documentation of the reasons supporting the request; and
    (2) Advances all reasonable costs of using type B procedures within 
a time frame acceptable to the authorized official.
    (c) If there is no available type A procedure, the authorized 
official must use type B procedures to calculate all damages.
    (d) Except as provided in paragraph (b) of this section, the 
authorized official may change the type of procedure used in light of 
comments received on the Assessment Plan. [See Sec. 11.32(e)(2) to 
determine if the authorized official must provide for additional public 
review.] However, if the authorized official decides to use type B 
procedures in lieu of a type A procedure, and cannot confirm exposure 
under Sec. 11.37, the authorized official may not then use a type A 
procedure.


Sec. 11.36  May the authorized official use both type A and type B 
procedures for the same release?

    (a) The authorized official may use both a type A procedure and 
type B procedures for the same release if:
    (1) The type B procedures are cost-effective and can be performed 
at a reasonable cost;
    (2) There is no double recovery; and
    (3) The type B procedures are used only to determine damages for 
injuries or compensable values that do not fall into the categories 
addressed by the type A procedure. [Sections 11.14(v) and 11.62 define 
``injury.'' Section 11.83(c)(1) defines ``compensable value.'']
    (b) The type A procedures address the following categories of 
injury and compensable value:
    (1) Direct mortality of species covered by the NRDAM/CME or NRDAM/
GLE resulting from short-term exposure to the released substance. 
Volume IV of the NRDAM/CME technical document (incorporated by 
reference, see Sec. 11.18) lists the species that the NRDAM/CME covers. 
Section 3, Volume III of the NRDAM/GLE technical document (incorporated 
by reference, see Sec. 11.18) lists the species that the NRDAM/GLE 
covers;
    (2) Direct loss of production of species covered by the NRDAM/CME 
or NRDAM/GLE resulting from short-term exposure to the released 
substance;
    (3) Indirect mortality of species covered by the NRDAM/CME or 
NRDAM/GLE resulting from disruption of the food web by direct mortality 
or direct loss of production;
    (4) Indirect loss of production of species covered by the NRDAM/CME 
or NRDAM/GLE resulting from disruption of the food web by direct 
mortality or direct loss of production;
    (5) Lost assimilative capacity of water column and sediments;
    (6) Lost economic rent for lost commercial harvests resulting from 
any

[[Page 20611]]

closures specified by the authorized official and/or from population 
losses;
    (7) Lost recreational harvests resulting from any closures 
specified by the authorized official and/or from population losses;
    (8) For the type A procedure for coastal and marine environments, 
lost wildlife viewing, resulting from population losses, by residents 
of the States bordering the provinces in which the population losses 
occurred. [A province is one of the geographic areas delineated in 
Table 6.1, Volume I of the NRDAM/CME technical document.] For the type 
A procedure for Great Lakes environments, lost wildlife viewing, 
resulting from population losses, by residents of local areas bordering 
the provinces in which the population losses occurred. [A province is 
one of the geographic areas delineated in Table 8.1, Volume I of the 
NRDAM/GLE technical document.];
    (9) Lost beach visitation due to closure; and
    (10) For the type A procedure for Great Lakes environments, lost 
boating due to closure.
    (c) If the authorized official uses both type A and type B 
procedures, he or she must explain in the Assessment Plan how he or she 
intends to prevent double recovery.
    (d) When the authorized official uses type B procedures for 
injuries not addressed in a type A procedure, he or she must follow all 
of subpart E (which contains standards for determining and quantifying 
injury as well as determining damages), Sec.  11.31(c) (which addresses 
content of the Assessment Plan), and Sec. 11.37 (which addresses 
confirmation of exposure). When the authorized official uses type B 
procedures for compensable values that are not included in a type A 
procedure but that result from injuries that are addressed in the type 
A procedure, he or she need not follow all of subpart E, Sec. 11.31(c), 
and Sec. 11.37. Instead, the authorized official may rely on the injury 
predictions of the type A procedure and simply use the valuation 
methodologies authorized by Sec. 11.83(c) to calculate compensable 
value. When using valuation methodologies, the authorized official must 
comply with Sec. 11.84.
    10. Newly designated Sec. 11.37 is amended by revising the heading 
and paragraph (a) to read as follows:


Sec. 11.37  Must the authorized official confirm exposure before 
implementing the Assessment Plan?

    (a) Before including any type B methodologies in the Assessment 
Plan, the authorized official must confirm that at least one of the 
natural resources identified as potentially injured in the 
preassessment screen has in fact been exposed to the released 
substance.
* * * * *
    11. The heading of subpart D is revised to read as follows:

Subpart D--Type A Procedures

    12. Section 11.40 is amended by revising the heading and paragraph 
(a), removing paragraph (b), removing the heading from paragraph (c), 
and redesignating paragraph (c) as paragraph (b) to read as follows:


Sec. 11.40  What are type A procedures?

    (a) A type A procedure is a standardized methodology for performing 
Injury Determination, Quantification, and Damage Determination that 
requires minimal field observation. There are two type A procedures: 
the type A procedure for coastal and marine environments; and the type 
A procedure for Great Lakes environments. The type A procedure for 
coastal and marine environments incorporates a computer model called 
the Natural Resource Damage Assessment Model for Coastal and Marine 
Environments Version 2.4 (NRDAM/CME). The NRDAM/CME technical document 
(incorporated by reference, see Sec. 11.18) includes and explains the 
NRDAM/CME. The type A procedure for Great Lakes environments 
incorporates a computer model called the Natural Resource Damage 
Assessment Model for Great Lakes Environments Version 1.4 (NRDAM/GLE). 
The NRDAM/GLE technical document (incorporated by reference, see 
Sec. 11.18) includes and explains the NRDAM/GLE. The authorized 
official must follow Secs. 11.41 through 11.44 when using the type A 
procedures.
    (b) * * *
    13. Section 11.41 is revised to read as follows:


Sec. 11.41  What data must the authorized official supply?

    (a) The NRDAM/CME and the NRDAM/GLE require several data inputs to 
operate. The authorized official must develop the following data 
inputs:
    (1) The identity of the released substance;
    (2) The mass or volume of the identified substance that was 
released;
    (3) The duration of the release;
    (4) The time of the release;
    (5) The location of the release;
    (6) The wind conditions;
    (7) The extent of response actions;
    (8) The extent of any closures;
    (9) The implicit price deflator; and
    (10) For the NRDAM/CME, the condition of the currents and tides.
    (b) The authorized official must change the data in the NRDAM/CME 
and the NRDAM/GLE for the following parameters if he or she is aware of 
more accurate data:
    (1) Air temperature;
    (2) Water temperature at the surface;
    (3) Total suspended sediment concentration;
    (4) Mean settling velocity of suspended solids; and
    (5) Habitat type.
    (c)(1) If the release occurred in Alaska and the authorized 
official is not aware of any reliable evidence that ice was absent from 
the site of the release, then he or she must turn on the ice modeling 
function. Otherwise, the authorized official must leave the ice 
modeling function off.
    (2) If the release occurred in the Great Lakes and the authorized 
official is aware of reliable evidence that ice was absent from the 
site of the release, then he or she must turn off the ice modeling 
function.
    (d) The authorized official must develop the data inputs and 
modifications and include them in the Assessment Plan in the format 
specified in Appendix II (for the NRDAM/CME) or Appendix III (for the 
NRDAM/GLE).
    14. New Secs. 11.42 through 11.44 are added to subpart D to read as 
follows:


Sec. 11.42  How does the authorized official apply the NRDAM/CME or 
NRDAM/GLE?

    (a) The authorized official must perform a preliminary application 
of the NRDAM/CME or NRDAM/GLE with the data inputs and modifications 
developed under Sec. 11.41. Volume II of the NRDAM/CME technical 
document (incorporated by reference, see Sec. 11.18) describes how to 
apply the NRDAM/CME. Volume II of the NRDAM/GLE technical document 
(incorporated by reference, see Sec. 11.18) describes how to apply the 
NRDAM/GLE. For cases involving releases of two or more substances or a 
release of a mixture of substances, the authorized official may only 
apply the NRDAM/CME or NRDAM/GLE once using only one of the substances.
    (b) If the preliminary application of the NRDAM/CME or NRDAM/GLE 
indicates damages in excess of $100,000, then the authorized official 
must decide whether to:
    (1) Limit the portion of his or her claim calculated with the type 
A procedure to $100,000; or
    (2) Compute all damages using type B procedures.

[[Page 20612]]

Sec. 11.43  Can interested parties review the results of the 
preliminary application?

    After completing the preliminary application of the NRDAM/CME or 
NRDAM/GLE, if the authorized official decides to continue with the type 
A procedure, he or she must issue an Assessment Plan for public comment 
as described in Sec. 11.32. The Assessment Plan must include the 
information described in Sec. 11.31, the data inputs and modifications 
developed under Sec. 11.41, and a summary of the results of the 
preliminary application. The Assessment Plan must also identify a 
contact from whom a complete copy of the printout of the preliminary 
application can be obtained.


Sec. 11.44  What does the authorized official do after the close of the 
comment period?

    (a) The authorized official must carefully review all comments 
received on the Assessment Plan, provide substantive responses to all 
comments, and modify the Plan as appropriate. [See Sec. 11.32(e)(2) to 
determine if the authorized official must provide for additional public 
review.]
    (b) If, after reviewing the public comments, the authorized 
official decides to continue with the type A procedure, he or she must 
then perform a final application of the NRDAM/CME or NRDAM/GLE, using 
final data inputs and modifications based on Sec. 11.41 and any 
reliable information received during the public review and comment 
period.
    (c) After completing the final application of the NRDAM/CME or 
NRDAM/GLE, the authorized official must prepare a Report of Assessment. 
The Report of Assessment must include the printed output from the final 
application as well as the Preassessment Screen Determination and the 
Assessment Plan.
    (d) If the authorized official is aware of reliable evidence that a 
private party has recovered damages for commercial harvests lost as a 
result of the release, the authorized official must eliminate from the 
claim any damages for such lost harvests that are included in the lost 
economic rent calculated by the NRDAM/CME or NRDAM/GLE.
    (e) If the authorized official is aware of reliable evidence that 
the NRDAM/CME or NRDAM/GLE application covers resources beyond his or 
her trustee jurisdiction, the authorized official must either:
    (1) Have the other authorized official(s) who do have trustee 
jurisdiction over those resources join in the type A assessment; or
    (2) Eliminate any damages for those resources from the claim for 
damages.
    (f) If the final application of the NRDAM/CME or NRDAM/GLE, 
adjusted as needed under paragraphs (d) and (e), calculates damages in 
excess of $100,000, then the authorized official must limit the portion 
of his or her claim calculated with the type A procedure to $100,000.
    (g) After preparing the Report of Assessment, the authorized 
official must follow the steps described in subpart F.
    15. The heading of subpart E is revised to read as follows:

Subpart E--Type B Procedures

    16. Section 11.73 is amended by revising the second sentence of 
paragraph (a) to read as follows:


Sec. 11.73   Quantification phase-resource recoverability analysis

    (a) * * * The time estimated for recovery or any lesser period of 
time as determined in the Assessment Plan must be used as the recovery 
period for purposes of Sec. 11.38 and the Damage Determination phase, 
Secs. 11.80 through 11.84.
* * * * *

Subpart F--Post-Assessment Phase

    17. Section 11.90 is amended by revising the heading, paragraphs 
(a) and (b), and the first sentence of paragraph (c) as follows:


Sec. 11.90   What documentation must the authorized official prepare 
after completing the assessment?

    (a) At the conclusion of an assessment, the authorized official 
must prepare a Report of Assessment that consists of the Preassessment 
Screen Determination, the Assessment Plan, and the information 
specified in paragraphs (b) and (c) of this section as applicable.
    (b) When the authorized official has used a type A procedure, the 
Report of Assessment must include the information specified in subpart 
D.
    (c) When the authorized official has used type B procedures, the 
Report of Assessment must include all documentation supporting the 
determinations required in the Injury Determination phase, the 
Quantification phase, and the Damage Determination phase, and 
specifically including the test results of any and all methodologies 
performed in these phases. * * *
    18. Section 11.91 is amended by revising the heading and by 
removing the first sentence of paragraph (a) and inserting three new 
sentences in its place to read as follows:


Sec. 11.91   How does the authorized official seek recovery of the 
assessed damages from the potentially responsible party?

    (a) At the conclusion of the assessment, the authorized official 
must present to the potentially responsible party a demand in writing 
for the damages determined in accordance with this part and the 
reasonable cost of the assessment. [See Sec. 11.92(b) to determine how 
the authorized official must adjust damages if he or she plans to place 
recovered funds in a non-interest-bearing account.] The authorized 
official must deliver the demand in a manner that establishes the date 
of receipt. * * *
* * * * *
    19. New Appendices II and III are added to read as follows:

Appendix II to Part 11--Format for Data Inputs and Modifications to the 
NRDAM/CME

    This appendix specifies the format for data inputs and 
modifications to the NRDAM/CME under Sec. 11.41. Consult the back of 
this appendix for definitions.

Starting Point for the NRDAM/CME

    The NRDAM/CME begins its calculations at the point that the 
released substance entered water in an area represented by its 
geographic database. Any water within the geographic boundaries of 
the NRDAM/CME is a ``coastal or marine environment.'' The authorized 
official must determine all data inputs and modifications as of the 
time and location that the released substance entered a coastal or 
marine environment. In the case of a release that began in water in 
an area within the boundaries of the NRDAM/CME, this point will be 
the same as the point of the release. However, for releases that 
begin on land or that begin outside the boundaries of the NRDAM/CME, 
this point will not be the point of the release but rather the point 
at which the released substance migrates into a coastal or marine 
environment.

Required Data Inputs

    Documentation of the source of the data inputs; and

Identity of Substance

    For release of single substance:
    Name of the substance that entered a coastal or marine 
environment as it appears in Table 7.1, Volume I of the NRDAM/CME 
technical document (incorporated by reference, see Sec. 11.18).
    For releases of two or more substances or a release of a mixture 
of two or more substances:
    Name of only one of the substances that entered a coastal or 
marine environment as it appears in Table 7.1, Volume I of the 
NRDAM/CME technical document.

Mass or Volume

    For release of single substance:
    Mass or volume of identified substance that entered a coastal or 
marine environment

[[Page 20613]]

stated in tonnes, barrels, gallons, liters, pounds, or kilograms.
    For releases of two or more substances or a release of a mixture 
of two or more substances:
    Mass or volume of the one identified substance (rather than 
total mass) that entered a coastal or marine environment stated in 
tonnes, barrels, gallons, liters, pounds, or kilograms.

Duration

    Length of time over which the identified substance entered a 
coastal or marine environment stated in hours.

Time

    Year, month, day, and hour when the identified substance first 
entered a coastal or marine environment.

Location

    Latitude and longitude, stated in degrees and decimal minutes, 
where the identified substance entered a coastal or marine 
environment.

Winds

    At least one set of data on prevailing wind conditions for each 
day of the 30-day period beginning 24 hours before the identified 
substance entered a coastal or marine environment. Each set must 
include:
    Wind velocity stated in knots or meters per second; and
    Corresponding wind direction stated in the degree angle of the 
wind's origin.

[One possible source of information is the National Climatic Data 
Center, Asheville, NC (703) 271-4800.]

Response Actions

    If removed from water surface:
    A rectangular geographic area encompassing the surface water 
area over which the released substance was likely to have spread, 
stated in terms of the northern- and southern-most latitude, and the 
eastern- and western-most longitude;
    One or more time frames for removal stated in terms of the 
number of days and hours after the identified substance entered a 
coastal or marine environment that removal began and ended; and
    For each time frame, volume of the identified substance removed 
from the water surface (not the total volume of contaminated water 
or sediments removed) stated in barrels, gallons, or cubic meters.
    If removed from shoreline:
    A rectangular geographic area encompassing the shoreline area 
over which the released substance was likely to have spread, stated 
in terms of the northern- and southern-most latitude, and the 
eastern- and western-most longitude;
    One or more time frames for removal stated in terms of the 
number of days and hours after the identified substance entered a 
coastal or marine environment that removal began and ended; and
    For each time frame, volume of the identified substance removed 
(not the total volume of contaminated water or sediments removed) 
stated in barrels, gallons, or cubic meters.

Closures

    Documentation that the closure was ordered by an appropriate 
agency as a result of the release;
    Province(s) in which closure occurred; and
    For beaches:
    Whether the beach was Federal or State (including municipal or 
county);
    Number of days of closure stated by calendar month; and
    Length of shoreline closed, stated in kilometers, for each month 
in which closure occurred.
    For fisheries and shellfish harvest areas:
    Whether area closed was seaward open water, landward open water, 
or structured;
    Number of days of closure; and
    Area closed stated in square kilometers.
    For furbearer hunting or trapping areas and waterfowl hunting 
areas:
    Number of days of closure; and
    Area closed stated in square kilometers.

Implicit Price Deflator

    Quarterly implicit price deflator for the Gross National Product 
(base year 1992) for the quarter in which the identified substance 
entered a coastal or marine environment. [See the Survey of Current 
Business, published by the U.S. Department of Commerce/Bureau of 
Economic Analysis, 1441 L Street, NW, Washington, D.C., 20230, (202) 
606-9900.]

Currents

    For a rectangular geographic area encompassing the area affected 
by the release stated in terms of the northern- and southern-most 
latitude, and the eastern- and western-most longitude:
    At least one set of data concerning background (mean) current 
consisting of--
    An east-west (U) velocity stated in centimeters per second or 
knots;
    A north-south (V) velocity stated in centimeters per second or 
knots; and
    Latitude and longitude of the origin of the U and V velocity 
components.
    At least one set of data concerning tidal current at time of 
flood stage (i.e., rising tide) consisting of--
    An east-west (U) velocity stated in centimeters per second or 
knots;
    A north-south (V) velocity stated in centimeters per second or 
knots; and
    Latitude and longitude of the origin of the U and V velocity 
components.

[Possible sources of information are: the National Ocean Service, 
U.S. Department of Commerce, Riverdale, MD (310) 436-6990; and the 
Eldridge Tide and Pilot Book, Robert Eldridge White Publisher, 
Boston, MA (617) 742-3045.]

Tides

    Hour of high tide on the day that the identified substance 
entered a coastal or marine environment;
    Tidal range at point that the identified substance entered a 
coastal or marine environment stated in meters; and
    Whether the tide in the area affected by the release is diurnal 
(i.e., completes one full cycle every day) or semi-diurnal (i.e., 
completes two full cycles every day).

Modifications to the NRDAM/CME Databases (if Any)

    Documentation of the source of the modification; and
    For air temperature:
    Air temperature, stated in degrees Celsius, assigned by the 
NRDAM/CME at the point that the identified substance entered a 
coastal or marine environment (see Table III.3.2, Volume III of the 
NRDAM/CME technical document); and
    Substitute air temperature stated in degrees Celsius.
    For water temperature at the surface:
    Water temperature at the surface, stated in degrees Celsius, 
assigned by the NRDAM/CME at the point that the identified substance 
entered a coastal or marine environment (see Table III.3.3, Volume 
III of the NRDAM/CME technical document); and
    Substitute water temperature stated in degrees Celsius.
    For total suspended sediment concentration:
    Total suspended sediment concentration, stated in milligrams per 
liter, assigned by the NRDAM/CME at the point that the identified 
substance entered a coastal or marine environment (see Section 3, 
Volume I of the NRDAM/CME technical document); and
    Substitute suspended sediment concentration stated in milligrams 
per liter.
    For mean settling velocity of suspended solids:
    Mean settling velocity of suspended sediments, stated in meters 
per day, assigned by the NRDAM/CME at the point that the identified 
substance entered a coastal or marine environment (see Section 3, 
Volume I of the NRDAM/CME technical document); and
    Substitute suspended sediment concentration stated in milligrams 
per liter.
    For habitat type:
    Latitude and longitude bounds of area for which the habitat type 
is being modified;
    Habitat type assigned by the NRDAM/CME (see Section 3.4, Volume 
III of the NRDAM/CME technical document); and
    Substitute habitat type.
    For releases in Alaska, if the authorized official leaves the 
ice modeling function off, he or she must provide documentation that 
ice was absent at the site of the release.

Definitions

    Background (mean) current--net long-term current flow (i.e., one 
direction only), attributable to forces such as winds, river flow, 
water density, and tides, that remains when all the oscillatory 
(tidal) components have been removed either mathematically or by 
measurement techniques.
    Landward open water--a body of water that does not contain 
vegetation (e.g., wetland, seagrass, or kelp) or invertebrate reef 
(e.g., coral reef) and is classified as ``landward'' in Table 6.2, 
Volume I of the NRDAM/CME technical document.
    Province--one of the geographic areas delineated in Table 6.1, 
Volume I of the NRDAM/CME technical document.
    Seaward open water--a body of water that does not contain 
vegetation (e.g., wetlands, seagrass, or kelp) or invertebrate reef 
(e.g., coral reef) and is classified as ``seaward'' in

[[Page 20614]]

Table 6.2, Volume I of the NRDAM/CME technical document.
    Structured--in an area that contains vegetation (e.g., wetlands, 
seagrass, or kelp) or invertebrate reef (e.g., coral reef).
    Tidal current--currents caused by alternating rise and fall of 
the sea level due to the gravitational forces between the earth, 
moon, and sun.
    Tidal range--difference between the highest and lowest height of 
the tide.

Appendix III to Part 11--Format for Data Inputs and Modifications 
to the NRDAM/GLE

    This appendix specifies the format for data inputs and 
modifications to the NRDAM/GLE under Sec. 11.41. Consult the back of 
this appendix for definitions.

Point of Analysis

    The NRDAM/GLE begins its calculations at the point that the 
released substance entered water in an area represented by its 
geographic database. Any water within the geographic boundaries of 
the NRDAM/GLE is a ``Great Lakes environment.'' The authorized 
official must determine all data inputs and modifications as of the 
time and location that the released substance entered a Great Lakes 
environment. In the case of a release that began in water in an area 
within the boundaries of the NRDAM/GLE, this point will be the same 
as the point of the release. However, for releases that begin on 
land or that begin outside the boundaries of the NRDAM/GLE, this 
point will not be the point of the release but rather the point at 
which the released substance migrates into a Great Lakes 
environment.

Required Data Inputs

    Documentation of source of data inputs; and

Identity of Substance

    For release of single substance:
    Name of the released substance that entered a Great Lakes 
environment as it appears in Table 7.1, Volume I of the NRDAM/GLE 
technical document (incorporated by reference, see Sec. 11.18).
    For releases of two or more substances or a release of a mixture 
of two or more substances:
    Name of only one of the released substances that entered a Great 
Lakes environment as it appears in Table 7.1, Volume I of the NRDAM/
GLE technical document.

Mass or Volume

    For releases of single substance:
    Mass or volume of identified substance that entered a Great 
Lakes environment stated in tonnes, barrels, gallons, liters, 
pounds, or kilograms.
    For releases of two or more substances or a release of a mixture 
of two or more substances:
    Mass or volume of the one identified substance (rather than 
total mass) that entered a Great Lakes environment stated in tonnes, 
barrels, gallons, liters, pounds, or kilograms.

Duration

    Length of time over which the identified substance entered a 
Great Lakes environment stated in hours.

Time

    Year, month, day, and hour when the identified substance first 
entered a Great Lakes environment.

Location

    Latitude and longitude, stated in degrees and decimal minutes, 
where the identified substance entered a Great Lakes environment.

Winds

    At least one set of data on prevailing wind conditions for each 
day of the 30-day period beginning 24 hours before the identified 
substance entered a Great Lakes environment. Each set must include:
    Wind velocity stated in knots or meters per second; and 
Corresponding wind direction stated in the degree angle of the 
wind's origin.

[One possible source of information is the National Climatic Data 
Center, Asheville, NC (703) 271-4800.]

Response Actions

    Percentage of identified substance removed from water surface, 
bottom sediments, and shoreline; and
    For each medium cleaned (water surface, bottom sediments, or 
shoreline), the number of days after the identified substance 
entered a Great Lakes environment that removal began and ended.

Closures

    Documentation that the closure was ordered by an appropriate 
agency as a result of the release; and
    For boating areas:
    Number of weekend days of closure stated by calendar month;
    Number of weekday days of closure stated by calendar month; and
    Area closed stated in square kilometers.
    For beaches:
    Whether the beach was Federal or State (including municipal or 
county);
    Number of days of closure stated by calendar month; and
    Length of shoreline closed stated in meters.
    For fisheries:
    Whether area closed was an offshore, nearshore, or wetland 
fishery;
    Number of days of closure; and
    Area closed stated in square kilometers.
    For furbearer hunting or trapping areas and waterfowl hunting 
areas:
    Number of days of closure; and
    Area closed stated in square kilometers.

Implicit Price Deflator

    Quarterly implicit price deflator for the Gross National Product 
(base year 1992) for the quarter in which the identified substance 
entered a Great Lakes environment. [See the Survey of Current 
Business, published by the U.S. Department of Commerce/Bureau of 
Economic Analysis, 1441 L Street, NW, Washington, D.C., 20230, (202) 
606-9900.]

Modifications to the NRDAM/GLE Databases (if Any)

    Documentation of the source of the modifications; and
    For air temperature:
    Air temperature, stated in degrees Celsius, assigned by the 
NRDAM/GLE at the point that the identified substance entered a Great 
Lakes environment (see Table III.6.1, Volume III of the NRDAM/GLE 
technical document); and
    Substitute air temperature stated in degrees Celsius.
    For water temperature at the surface:
    Water temperature at the surface, stated in degrees Celsius, 
assigned by the NRDAM/GLE at the point that the identified substance 
entered a Great Lakes environment (see Table III.6.2.6, Volume III 
of the NRDAM/GLE technical document); and
    Substitute water temperature stated in degrees Celsius.
    For total suspended sediment concentration:
    Total suspended sediment concentration, stated in milligrams per 
liter, assigned by the NRDAM/GLE at the point that the identified 
substance entered a Great Lakes environment (see Section 3, Volume I 
of the NRDAM/GLE technical document); and
    Substitute suspended sediment concentration stated in milligrams 
per liter.
    For mean settling velocity of suspended solids:
    Mean settling velocity of suspended sediments, stated in meters 
per day, assigned by the NRDAM/GLE at the point that the identified 
substance entered a Great Lakes environment (see Section 3, Volume I 
of the NRDAM/GLE technical document); and
    Substitute suspended sediment concentration stated in milligrams 
per liter.
    For habitat type:
    Latitude and longitude bounds of area for which the habitat type 
is being modified;
    Habitat type assigned by the NRDAM/GLE (see Section 6.2, Volume 
III of the NRDAM/GLE technical document); and
    Substitute habitat type.
    If the authorized official turns off the ice modeling function, 
then he or she must provide documentation that ice was absent from 
the site of the release.

Definitions

    Nearshore fishery--fishery in an open water area that is less 
than 30 feet in depth or is in a connecting channel.
    Offshore fishery--fishery in an open water area that is 30 feet 
or more in depth.
    Wetland fishery--fishery that is not in an open water area.

    Dated: April 25, 1996.
Bonnie R. Cohen,
Assistant Secretary--Policy, Management, and Budget.
[FR Doc. 96-10747 Filed 5-6-96; 8:45 am]
BILLING CODE 4310-RG-P