Modeling cumulative impacts and the carrying capacity of small tidal creeks and inlets

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

Task 19 FINAL PRODUCT VMRC
FY 1993 Cam. Impacts Carrying Capacity Creeks A
Inlets

Mode g Cumulative Impacts
and the
Carrying Capactiy
of
Small Tidal Creeks and Inlets

Bruce J. Neilson
Carl Hershner
Megan Greiner

Virginia Institute of Marine Science
School of Marine Science
College of William and Mary
Gloucester Point, Virginia 23062

This report was funded, in part by the Department of Environmental
Quality's Coastal Resources Management Program through Grant
#NA47OZO287-01 of the National Oceanic and Atmospheric Administration,
(9W-)

0
,ffice of Ocean and Coastal Resource Management, under the Coastal Zone
Management Act of 1972, as amended.

Modeling Cumulative Impacts and the Carrying Capacity of

Small Tidal Creeks and Inlets

INTRODUCTION

Estimation of water quality impacts associated with use changes on land or water is
primarily dependent on knowledge of what will be added to the water as a consequence of the
change in use. Understanding of the mixing and transport of substances added to water bodies is
reasonably advanced, with numerous mathematical models available to generate assessments of
dilution and dispersion. The accuracy of model output is constrained, however, by the accuracy of
the estimates of what is being added to the water (loadings). Frequently, loading of pollutants is
unknown and extremely difficult to measure.

When pollutant loading occurs as a result of point source discharge, measurement of

amounts added can be determined if concentrations and flow rates are known. While not a trivial

problem, it is usually possible to mount an outfall sampling program sufficient to generate the
required estimate. The problem is much more intractable when dealing with nonpoint source
pollution. Addition of nutrients, sediments, biological and chemical oxygen demand, fecal
coliforms and other substances to water bodies from surrounding lands is generally variable in
both time and space. Estimating a time and space averaged loading is a difficult undertaking
requiring extensive long term sampling of all the various delivery pathways. Few studies have
undertaken the comprehensive sampling necessary to generate such numbers.

Even when estimates of pollutant loadings are developed, application of the estimates to

areas outside of the original sampling site entails an enormous number of assumptions about
modes and rates of delivery, assumptions which are rarely documented or tested. As a
consequence, prediction of water quality impacts, and particularly cumulative impacts, remains a
speculative undertaking. Model output can be no better than the quality of the data used as input.

'Me purpose of this project was to identify pollutant loading values which might be used
as input for a series of water quality models applied to small tidal creeks and inlets in Virginia's
coastal plain. The intent was to identify values from literature sources which might be used in
application of the models, absent better or more site specific information. Estimates of biological
oxygen demand, chemical oxygen demand and fecal coliform loadings were of specific interest.

FINDINGS

Numerous studies have sought to determine pollutant loads associated with different land
uses through intensive field work and long term sampling ( eg. Clesceri et al., 1986; Beulac et al.
1982; Sonzogni, 1980). The majority of these works measure pollutant input through surface
runoff following rainfall events. Runoff export coefficients are calculated as average annual
pollutant loads per unit area. There are several limitations to braod application of these values.
While studies have suggested that land use is the most important factor controlling pollutant loads
in runoff waters (eg. Rast et al. 1983, Whipple et al. 1978), the net input of pollutants is
understood to be a result of rainfall intensity and frequency, soil type, watershed slopes, and small
scale patterns of land cover/use. Transport through groundwater pathways is also understood to
be locally important and highly variable. Incorporation of all of these considerations in water
quality models, while an objective of ongoing work, is not yet practical. Current models, and
particularly those developed in the related work for this project, rely primarily on measures of
land use and a generalized runoff coefficient.

Uttormark et al. (1974) concluded from a survey of literature that there is little

justification for the delineation of land useage beyond categories of urban, agriculture, forest and
pasture. Thus, most studies report values for these classes of land uses. A survey of the
literature indicates loading values can vary widely, generally over one to two order of magnitude
for any single pollutant and land use type.

This report provides a summary of values for both runoff coefficients (TABLE 1) and
storm water runoff load estimates (TABLE 2) for total suspended solids, biological oxygen
demand, coliform levels, chemical oxygen demand, total nitrogen and total phosphorus.

SPECIAL NOTE FOR FECAL COLIFORMS

Estimation of fecal coliform loadings is particularly important in efforts to evaluate projects
proposing marina construction. At the present time, the number of boat slips available in a marina
is used as a predictor of probable pathogen concentrations in surrounding waters. The purpose of
this determination is to set shellfish harvest controls, hopefully preventing the harvest and/or
marketing of shellfish which might contain unacceptably high concentrations of pathogens. The
assumption that slip number is an adequate predictor of pathogen concentrations is generally
viewed as grossly oversimplified, but it has served as a practical solution absent more
sophisticated methods of prediction.

One of the goals of this project was generation of mathernatic models which might
enhance the ability of managers to assess probable pathogen loading around marinas. The models
basically estimate dispersion of fecal coliforms (used as an indicator of pathogen levels) based on
circulation patterns. Model output is clearly dependent on the initial loading assumption. This
number is extremely difficult to estimate with any accuracy.

In 1985, the United States Environmental Protection Agency developed a Coastal Marinas
Assessment Handbook. The purpose of the handbook was to provide guidance in the design and

evaluation of marina projects. As part of that undertaking, the EPA contractor conducted an
extensive review of available information in an effort to determine the fecal coliform loading
attributable to each boat in a marina. The handbook referenced work done by Carstea et al.
1975 in identifying the assumptions necessary to develop an estimate. These assumptions include:
- average persons per boat is three;
- average per capita discharges of coliform bacteria and BOD are 2 billion and 75.6 g
respectively;
- half of the people on board contribute fecal material in 24 hours;
- coliform bacteria populations do not increase;
- a boat in use spends one hour in the marina;
- 25 to 40 percent of boats present are in use and evenly distributed.

The difficulties associated with use of a number based on all these assumptions is self-evident.

The estimation of fecal coliform loading per boat in a marina is further complicated by the
increased use of on-board marine sanitation devices. An even greater uncertainty derives from the
potential influx of fecal coliforms from non-point sources. Schima et al. 1994 investigated the
relationships between fecal coliform levels at 2,614 sampling stations and their landscape
positions. They found a basic pattern of increasing fecal coliform bacteria densities with distance
upstream in tidal creeks and inlets. They refered to this as a "land mass" effect, or simply the
amount of land within a fixed radius of sample locations. Regression analysis of the MPN
concentrations at sampling points along the Eastern Shore of Virginia indicated the following
variables (in order of decreasing importance) were significant in explaining changes in the sampled

MPN values:

- surface area of water in a 400 rn raduius around the sampling point

- season

- tide stage during sampling
- rainfall amounts in the 2 days prior to sampling
- surface area of urban land in a 400 m radius around the sampling point

- near shore groundwater hydraulic gradient
- slinity
- proximity to nearest shoreline
- near shore soil permeability

- near shore runoff events

- surface area of agricultural land in a 400 m radius around the sampling point
- near shore Darcy velocity

- water temperature
One conclusion to be drawn from all of this is that while fecal coliform inputs from nonpoint
sources on land may be very important, they are also very variable and difficult to predict, even on

the basis of land use.

TABLE 1

LOADING RATES BY LAND USE (kg/ha/yr)

LANDUSE TSS BOD COLIFORM SOURCE

residential 420 35 - Wanielista, 1978

11-487 - - Bannerman et al., 1984

360-390 - Marsalek, 1978

620-2,300 - - Sonzogni, 1980
- 30-50 - Loehr, 1974

- 25,621- Ellis, 1986
- - 82,500(mpn/g)

commercial 840 87 Wanielista, 1978

957 - Bannerman et al., 1984

360 - Maralek, 1978
50-830 - Sonzogni, 1980
- - 36,900(mpn/g) Ellis, 1986

agriculture mean 450 18 - Wanielista, 1978
range 180-4,200 4-31 -

pasture mean 343 11.5 - Wanielista, 1978
range 10-840 6-17 -

forest mean 85 5 - Wanielista, 1978
range 15-132 2-7 -

TABLE I (continued)

LOADING RATES BY LANDUSE (kg/ha/yr)

LANDUSE TN TP SOURCE
residential 5.0-7.3 Sonzogni, 1980
9-11.2 1.6-3.4 Marsalek, 1978

5.4-18.0 1.00-2.47 EPA, 1983

6.6 1.8 Wanielista,1978
- 1.2-8.0 Whipple et al., 1978

commercial 1.9-11 - Sonzogni, 1980

11.2 3.4 Marsalek, 1978

16.3 2.22 EPA, 1983

14.5 2.7 Wanielista, 1978

pasture mean 6.2 0.5 Wanielista, 1978
range 2.0-12.0 0.1-2.1

4.94 0.74 Beulac & Reckhow, 1982

- 0.34-0.56 Mackiernan, 1985

agriculture 8.89 2.22 Beulac & Reckhow, 1982
mean 26.0 1.05 Wanielista, 1978

range 15.0-37.0 0.18-1.62
1.68-5.6 Mackiernan, 1985

0.06-2.9 Loehr,1974

TABLE 1 (continued)

LOADING RATES BY LANDUSE (kg/ha/yr)

LANDUSE TN TP SOURCE

forest mean 3.0 0.10 Wanielista, 1978

range 2.0-5.1 0.01-0.86
2.47 0.25 Beulac & Reckhow, 1982

- 0.06-0.11 Mackiernan, 1985

0.03-0.9 Loehr, 1974

TABLE2

STORM WATER RUNOFF ESTIMATES (mg/1)

POLLUTANT SOURCE

TSS 141-224 EPA, 1983

1,401-2,909 Wanielista, 1978

urban 227 Carstea et al., 1975

Durham, NC mean 1,440 Colston, 1974

range 194-8,620
agricultural 90-5,000 Dombush et al., 1974

watershed 180-6,000

cultivated 1,021
pasture 38

BOD urban area 12-160 Loehr,1974

17 Carstea et al., 1975
Cincinnati, OH mean 19 Weibel et al., 1964

range 2-84
agricultural 7 Loehr,1974
watershed 5-30 Dornbush et al., 1974

3-15

COLIFORM 1,000-

21,000 MPN/100ml EPA, 1983

> 2,000 MPN/100ml Olivieri et al,. 1977
Washington, DC 76,100 Wanielista, 1978
Durham, NC mean 23,000 Colston, 1974
range 100-200,000

TABLE 2 (continued)

STORM WATER RUNOFF ESTIMATES (mg/1)

POLLUTANT SOURCE
COD Durham, NC mean 170 Colston, 1974
range 20-1042
Cincinatti, OH mean 99 Weibel et al., 1964
range 20-610
agricultural 50-360 Dornbush et al., 1974

watershed 70-780

pasture 49

TN 5.6-7.1 EPA, 1983

urban area mean 3.1 Weibel et al., 1966

range 0.3-75
3.1 Carstea et al., 1975

forested 0.3-1.8 Loehr, 1974
agriculture 9.0 Loehr, 1974

TP 0.4-0.5 EPA, 1983
agriculture 0.04-2.4 Dornbrush et al., 1974

forested 0.01-0.11 Loehr, 1974

agriculture 0.02-1.7 Loehr, 1974
urban area 0.2-1.1 Loehr, 1974

1.1 Carstea et al., 1975

REFERENCES

Bannerman, R., K. Baun, M. bohm. P.E. Hughes and D.A. Graczyk. 1984. Evaluation of
Urban Nonpoint Source Pollution Management in Milwaukee County, Wisconsin.
Report No. P1384-114164. Chicago, 111. EPA Region V.

Beulac, M.N. and K.H. Reckhow. 1982. An examination of landuse nutrient export
relationships. Water Resources Bulletin. 18:1013-1023.

Carstea, D., A. Binder, R. Strieter, L. Boberschmidt, L. Thomas and J. Golden. 1975.
Guidelines for the environmental impact assessment of small structures and related activities in
coastal bodies of water. Prepared by MITRE Corporation for U. S. Army Corps of Engineers,

New York District.

Clesceri, N.L. S.J. Curran, and R.I. Sedlak. 1986. Nutrient Loads to Wisconsin Lakes:

Part I: Nitrogen and Phosphorus export coefficients. Water Res. Bull. 22(6):983-1000.

Colston, N.V. Jr. Characterization and treatment of urban land runoff. Environ. Protection

Technol. Series. EPA 670/2-74-096. 1974.

D'Elia, F. 1987. Nutrient Enrichment of the Chesapeake Bay: Too much of a good thing?

Env. 29:6-33.

Dombush, J.N., J.R. Anderson, L.L. Harms. 1974. Quantification of pollutants in agricultural
runoff. EPA-660/2-74-005. US EPA. Washington, D.C.

Ellis, J.B. 1986. Pollutional aspects of urban runoff. pp. 1-38 in H.C. Torna, J. Marsalek,
and M.Desbordes, fd- Urban Runoff Pollution. Springer-Verlag, NY, Inc.

Loehr, R.C. 1974. Characteristics and comparative magnitude of non-point sources.
Journal WPCF 46(8):1849-1872.

Mackiernan, G.B., D.A. Flemer, Nehlsen, W. V.K. Tippie and R.B. Biggs. 1985. Chesapeake
Bay: A profile of environmental change. EPA, Annapolis, MD. Chesapeake Bay Program.

Marsalek, J. 1978. Pollution Due to Urban Runoff: Unit Loads and Abatement Measures.
Windsor, Canada: Pollution by Land Use Activities Reference Group of the

International Joint Commission.

Nagnien, R.E., R.M. Summers and K.G. Sellner. 1992. External Nutrient Sources, Internal
Nutrient Pools, and Phytoplankton Production in Chesapeake Bay. Estuaries
15(4):497-516.

Olivieri, V., C. Kruse, K. Kawata, and J. Smith. 1977. Microorganisms in Urban Stormwater.
Report No. EPA 600/2-77-087. Cincinnati, OH: EPA Municipal Environmental
Research Laboratory.

Overton, D.E. and M.E. Meadows. 1976. Stormwater Modeling. NY Academic Press.

Rast, W. and G.F. Lee. 1983. Nutrient Loading Estimates for Lakes. J.Env. Engineering. ASCE
109(2):502-517.

Schima, F.J., W.G. Reay, D.L. Gallagher, G. M. Simmons, Jr., J.L. WaIdon and K.K. Reay.
1994. Groundwater transport of fecal coliform bacteria to open coastal waters of Virginia's
coastal plain: A GIS approach. Final Report to Virginia's Coastal Resource

Sonzogni, W.C. 1980. Pollution from land runoff. Environ. Sci. and Tech. 14:148-153.

United States Environmental Protection Agency. 1985. Coastal Marina Assessment Handbook.
USEPA 904/6-85-132. Region 4, Atlanta, Georgia.

United States Environmental Protection Agency. 1983. Results of the Nationwide Urban
Runoff Program: Volume 1. USEPA, Washington, D.C.

Uttormark, P.D., J.D. Chapin and K.M. Green. 1974. Estimating Nutrient Loading of Lakes
from non-point sources. US EPA 600/3-74-020. US EPA. Corvallis, OR.

Wanielista, M.P. 1978 Stormwater Management: Quantity and Quality. Ann Arbor

Science Publishers.

Weibel, S.R. et al. 1964. Urban land runoff as a factor in stream pollution. J. WPCF.
36(7):914-924.

Weibel, S.R. et al. 1966. Pesticides and Other Contaminants from Raingall and Runoff as

Observed in Ohio. Jour. Amer. Water Works Assn. 58:1075-1084.

Whipple, W. Jr. J.V. Hunter, and S.L. Yu. 1978. Runoff pollution from multiple family housing.

Water Res. Bull. 14:288-301.

Coastal Screening Model
User Manual

Version 0.95

Developed for:
Virguilia Marine Resource Commission

Virginia Polytechnic Institute and State University
Division of Envirorunental Engineering and Sciences
Department of Civil Engineering

Mary Ann Parcher
Research Assistant

Dr. Daniel Gallagher
Associate Professor

June 20, 1995

This brochure was reprinted by the Department of Environmental Quality's Coastal Resource
Management Program through Grant #NA370ZO360-01 of the National Oceanic and Atmospheric
Admini tration, Office of Ocean and Coastal Resource Management, under the Coastal Zone
Management Act of 1972, as amended.
C I PC FI/ The views expressed herein are those of the authors and do not necessarily reflect the
views of NOAA or any of its subagencies.

Table of Contents
1. Introduction ................................................................................................. 1-1
1. 1 Model Overview ....................................................................................................... 1-1
1.2 Notation Review ....................................................................................................... 1-2
1.3 System Requirements ................................................................................................ 1-4
1.4 Installation Procedures .............................................................................................. 1-4
1.5 Getting Started .......................................................................................................... 1-4
2. Model Selection Advisor ............................................................................ 2-1
2.1 Model Types ............................................................................................................. 2-1
2. 1.1 Land Use Models ........................................................................................................... 2-2
2.1.2 Water Quality Models .................................................................................................... 2-2
2.2 Operating the Advisor ............................................................................................... 2-2
2.3 Water Quality Parameters .......................................................................................... 2-3
2.3.1 Dimensions .................................................................................................................... 2-4
2.3.2 Time .............................................................................................................................. 2-4
2.3.3 Loadings ........................................................................................................................ 2-4
2.3.4 Tidal .............................................................................................................................. 2-4
2.3.5 Water Quality Parameters ............................................................................................... 2-4
2.4 Model Selection ........................................................................................................ 2-5
3. Watershed Model ....................................................................................... 3-1
3.1 Modeling Approach ....... 7** . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .3-1
3.1.1 Septic Systems ............................................................................................................... 3-3
3.2 Model Format ........................................................................................................... 3-3
3.3 Program Options ........................................................................................................ 3-4
3.3.1 Hydrologic Unit Data .............................................................................0 ....................... 3-4
3.3.2 Climate Data .................................................................................................................. 3-5
3.4 Initial Conditions ....................................................................................................... 3-6
3.5 Land Use Information ................................................................................................ 3-8
3.5.1 Runoff Curve Numbers .................................................................................................. 3-9
3.5.2 Universal Soil Loss Equation Factors ............................................................................ 3-13
3.6 Nutrient Input ......................................................................................................... 3-20
3.7 Point Sources .......................................................................................................... 3-21
3.8 Evapotranspiration Conditions ................................................................................ 3-22
3.9 Septic Systems ........................................................................................................ 3-23
3. 10 Model Results ....................................................................................................... 3-24
3. 10.1 Tabular Output ...........................................0 .............................................................. 3-24
3.10.2 Graphical Output ........................................................................................................ 3-25
4. Marina Water Quality Model ...................................................................... 4-1
4.1 Model Format .................................................................0............................. 0 ........... 4-2
4.2 Parameters ................................................................................................................ 4-3)
4.2.1 Hydraulic Pa6meters ..................................................................................................... 4-33
4.2.2 Contaminant Parameters ..........0 ....................................................................................... @@3
4.2.3 Region to be Modeled ................................................................0 .................................... 4-3
4.3 Options ............................................... 0 ...................................................................... 4-4
4.4 Model Results ........................................................................................................... 4-4

4.4.1 Tabular Output ................................................................................................................
4.4.2 Transects ....................................................................................................................... 44
5. Tidal Prism Model ...................................................................................... 5-1
5.1 Modeling Approach .................................................................................................. 5-2
5.2 Model Format ........................................................................................................... 5-3
5.3 Parameters ................................................................................................................ 5-4
5.4 Stream Geometry ...................................................................................................... 5-5
5.5 Segmentation ............................................................................................................ 5-6
5.6 Loading Data ............................................................................................................ 5-6
5.7 Table Results ............................................................................................................ 5-7
6. Finite Section Model .................................................................................. 6-1
6.1 Modeling Approach ................................................................................................... 6-1
6.2 Model Format ........................................................................................................... 6-2
6.3 Options ...........................................................................................................A ........ 6-3
6.3.1 Water Quality Variables ................................................................................................. 6-3
6.3.2 Boundary Conditions ...................................................................................................... 6-4
6.3.3 Differencing Options ...................................................................................................... 6-4
6.4 Input Data ................................................................................................................. 6-5
6.5 Model Results ........................................................................................................... 6-5
6.5.1 Graph Output ................................................................................................................. 6-6
6.5.2 Table Output .................................................................................................................. 6-6
7. Spill Model ............................................................................................ * .... 7-1
7.1 Model Format ........................................................................................................... 7-1
7.2 Parameters ................................................................................................................ 7-2
7.3 Model Results ........................................................................................................... 7-2
7.3.1 Graph Results ................................................................................................................ 7-3
7.3.2 Table Results ................................................................................................................. 7-3
7.4 Calibration ................................................................................................................ 7 - ")
8. Utility Models ............................................................................................ 8-1
9. References ................................................................................................. 9-1

List of Tables

Table 3-1. Weather File Locations and Record Lengths .................................................................. 3-6
Table 3-2. Sediment Delivery Ratio Based on Watershed Size ........................................................ 3-7
Table 3-3. Runoff Curve Numbers for Cultivated Agricultural Land ............................................. 3-10
Table 3-4. Runoff Curve Numbers for Other Rural Land .............................................................. 3-11
Table 3-5. Runoff Curve Numbers for Urban Areas ...................................................................... 3-12
Table 3-6. Description of Soil Hydrologic Groups ........................................................................ 3-13
Table 3-7. Values of Soil Erodibility Factor (K) in t1a, .................................................................. 3-14
Table 3-8. Values for the Topographic Factor (LS) ...................................................................... 3-16
Table 3-9. Values of C for Cropland, Pasture, and Woodland ....................................................... 3-16
Table 3-10. C Factor Values and Slope-Length (LS) Limits for Construction Sites ....................... 3-17
Table 3-11. Values of P for Agricultural Lands ............................................................................ 3-18
Table 3-12. Values of P for Construction Sites ............................................................................. 3-18
Table 3-13. Estimated Values for Land Uses in the Hydrologic Unit Database ............................... 3-19
Table 3-14. Dissolved Nutrients in Agricultural Runoff ................................................................ 3-21
Table 3-15. Nutrient Accumulation Rates for Northern Virginia Urban Areas ............................... 3-21

Table of Figures

Figure 2- 1. Model selection advisor parameter options .................................................................... 2-3
Figure 3-1. Conceptual model of a watershed, adapted from Haith et al., 1992 ................................ 3-1
Figure 3-2. Watershed sediment delivery ratio, adapted from Vanon@ 1975 ..................................... 3-7
Figure 5-1. Elevation view of a hypothetical river illustrating the volumes of water exchanges between
adjacent. segments during a complete tidal cycle . ...................................................................... 5-4
Figure 6-1. Finite Section Con on . ................................................................................. 6-1
Figure 6-2. Components of mass balance equation, adapted from Thomann and Mueller (1987) . ..... 6-2

1. Introduction
Significant demand exists for housing and marina development along tidal creeks and inlets around the
Chesapeake Bay. These changes in land use have the potential to adversely impact the water quality of
Virginia's waterways. Existing water quality models can predict these water quality changes, but
existing models require extensive field data, model calibration, and expertise to use. The Virginia
Marine Resource Commission (VMRC) issues permits fbr most coastal development and does not have
the necessary resources to apply such models on a routine basis. 'Me Coastal Screening (CS) Model
was developed to provide the VMRC with a tool for a preliminary evaluation of development
applications and to identify those projects with relatively insignificant water quality impacts.

As the name implies, the CS Model is a screening-level model. The unavailability of measured field
values requires that the model contain, many assumptions to reduce the amount of required input. As
model complexity and site-specific information is reduced, the uncertainty of the predicted values
increases. Therefore, the goal of the CS Model is not to predict a specific water quality concintration,
but only to identify those projects that can be permitted without further in-depth analysis. A proposed
project passing the screening model would indicate no major impacts with a fair degree of uncertainty.
Projects with greater impacts or with high degrees of uncertainty would require additional analysis and
modeling effirt.

1.1 Model Overview
The Coastal Screem*ng Model is a PC-based computer model that operates in the Windows
environment Ile components of the CS Model are several water quality models, a land use model, a
few utility codes, and a model selection advisor.

The model selection advisor is designed to aid the user in determining which model is most applicable
for the current assessment. The advisor contains a checklist of various water quality parameters and
model options that the user can select for each modeling effort. Based on the options selected by the
user, the advisor determines the appropriate model(s). A description of model characteristics and
parameters and a brief discussion of required inputs is provided for each model to further aid the user
in the selection process. A complete discussion of the model selection advisor and all its components is
provided in Chapter 2.

The Watershed Model is a lumped parameter, watershed-scale land use model. It is based on empirical
equations for runoff, erosion, and sediment yields. The model will estimate strem flow, runoff,
sediment yield, total and dissolved nitrogen, and total and dissolved phosphorus at the watershed outlet.
It can be used to estimate potential pollutant loads entering coastal waters as a result of land use
changes. A comprehensive discussion of the Watershed Model and its required inputs is provided in
Chapter 3.

The Manna Water Quality Model is a steady-state, two-dimensional model that can be used to estimate
coliform bacteria concentration, carbonaceous biochemical oxygen demand (CBOD), and nitrogenous
biochemical oxygen demand (NBOD). It is based on three analytical mixing and transport models that
describe the physical processes in wide channels, narrow channels, or sermi-enclosed bays. A
comprehensive discussion of the Marina Water Quality Model and its required inputs is provided in
Chapter 4.

The Tidal Prism Model is a dynamic, one-dimensional, segmented water quality model that considers
the tidal effects in estuaries and rivers. It can be used to estimate coliform bacteria concentration,
CBOD, or other contaminants subject to a first order decay. Each contaminant must be simulated
separately. A comprehensive discussion of the Tidal Prism Model and its required inputs is provided in
Chapter 5.

The Finite Section Model is a steady-state, one-dimensionaL segmented water quality model that does
not consider tidal effects m. the stream channel. It can be used to estimate coliform bacteria
concentration, DO, CBOD, and NBOD from point or known nonpoint sources. A comprehensive
discussion of the Finite Section Model and its required inputs is provided in Chapter 6.

The Spill Model is a dynamic, one-dimensional water quality model that can be used to estimate the
concentration of a pollutant "spilled" into a river or creek. The contaminant is assumed to be
completely Truxed vertically and axially and decrease over tune through first order decay. A detailed
discussion of the Spill Model and its required inputs is provided in Chapter 7.

The CS Model also includes several utility models to aid the user with quick, simple calculations.
Currently, there are utility models for determining dispersion coefficient estimation, DO saturation, and
Manning's open channel flow. A complete discussion of these models is provided in Chapter 8.

1.2 Notation Review
The following conventions are used throughout the user manual.

� Keyboard strokes, menu choices, command buttons, and anything you are asked to type are printed
in boldface.

� Keyboard strokes are surrounded by < >. For example, the "Enter" key is denoted as <Enter>.

� Command buttons are surrounded by For example, a "Close" command button is denoted as
[Close].

� The terms "click on" and "select" refer to a specified mouse action. When you are told to click on
or select an item, locate the mouse cursor over the selected item and press the left mouse button. If
you prefer to use the keyboard instead of the mouse, you can navigate the model by using <Alt>
and the underlined letter of the desired menu or command option.

This manual uses standard Windows terminology. For users who are not familiar with the Windows
environment, some common elements of a window are described below.

Title Bar: Ile title bar displays the window@ title. Ile window title displayed in the CS Model
corresponds to the currently selected modeling option, however, when a component model is minimi@
the title bar returns to the general Coastal Screening Model title.

Menu Bar: - The menu bar at the top of each window displays the commands available for that
window. The main menu bar of the CS Model provides access to the individual models. When a
component model is selected, its menu bar becomes active. Menu bars for individual models provide
file and printer access.

1-2

Minimize Button: This is the button with the down arrow on the night side of the title bar. When this
button is selected, the window is reduced to its icon. Ile window can be maximized (returned to its
original size) by double-clicking on the icon. Ile CS Model and all its component models have
mmunize buttons and can be reduced to an icon.

Dialog Boxes: A dialog box requests or provides information, but does not have a menu bar. For
example, a dialog box will appear when the File/Open menu option is selected in any of the models.

List Boxes: A fist box provides a fist of available choices from which you must make a selection. To
select an item from the fist, move the mouse pointer to the desired item and click the left mouse button,
or access the list boxes by pressing the <Tab> key until the desired list box gains the focus and then
use the arrow keys until the desired item is highlighted. When space is limited, a "drop-down" list box
may be used. A drop-down list box displays only one entry. To access all of the available items, click
on the down arrow button next to the displayed fist item. A list box will appear beneath the originally
displayed item. Items in a drop-down fig box are selected in the same manner as a regular list box.

Scroll Bars: Some dialog and list boxes may contain more information than can be viewed in the
allocated area, so scroll bars will be added to allow the user to access all of the information. Scroll
bars contain an up and a down arrow as well as a scroll button. To scroll information line by line,
click repeatedly on one of the scroll arrows until the desired information comes into view. To page
through the fist quickly, click the scroll bar above or below the scroll button, or drag the scroll button
up or down the scroll bar.

Command Buttons: A command button performs a command or action when chosen by a user. When
selected, a command button not only carries out the appropriate action, but appears as if it's being
pushed in and released.

Option Buttons: Option buttons allow the user to select one option from the group. An option button
is activated by clicking on it or its description.

Check Boxes: Check boxes allow the user to decide whether or not some action should be taken. A
check box-is activated by clicking on it or its description. If a check box is selected, then its value is
true and the action it describes will occur. Unlike option buttons, check boxes operate independently of
each other and more than one can be selected within a group.

Text Boxes: A text box displays text that can be edited by the user. To change information in a text
box, click on the text box and t)W the new information. Text boxes can also be accessed by pressing
the <Tab> key until the desired text box gains the focus. In the CS Model, text boxes are used to
allow user input for model parameter values and can usually be distinguished by their white
background.

Spin Buttons: Spin buttons are a set of up and down arrows associated with a text box. Clicking on
either of the spin buttons will incrementally change the value in the text box.

Notebook: The component models utilize a notebook format to enter data and display model results.
Each page of the notebook can be accessed by clicking on the corresponding tab or pressing <Alt>
plus the underlined letter of the tab header.

1.3 System Requirements
You will need an EBM or EBM-compatible personal computer and a monitor that are capable of
running MicroSoft Windows (version 3.1 or higher). 'Me minimum system requirements include:

0 A 386-33 CPU
0 4 mega bytes (MB) of random access memory (RAM)
0 5 M]3 of fi-ee hard disk space
0 A VGA monitor with 16 colors

A 386-based system is technically sufficient, but depending on which component model is selected, the
processing time could be quite slow, particularly since the models are designed to run in an interactive
rather than a batch mode. The recommended system requirements are:

0 A 486-33 DX or 486-66 DX CPU
0 8 M13 of RAM
0 5 MB of fi-ee hard disk space
0 A VGA monitor with 256 colors

� monitor operating in 256 color mode increases the clarity of map and picture displays.

� mouse or other type of t-acking device is not strictly required by the program, but will greatly
improve the ease with which the models can be accessed and used.

1.4 Installation Procedures
'Me current setup process requires two steps. 'Me first install the CS model itself. Insert Disk I into
your disk drive. From the Fide menu of the Program Manager or File Manager, choose Run. Type <a:
setup (or b: if you are using the b drive). The set-up routine will prompt you for the path where you
want to install the program files. (Follow the set-up instructions on the screen). 'rhe second step
installs the ancillary data files and maps. Insert the Ancillary Disk 1 and run the setup.bat file. You
wiI1 be asked for the path where you insta lied the CS model.

1.5 Getting Started
To access the Coastal Screening Model, click on the Coastal icon from your Windows program
manager. The title bar and menu bar for the Coastal Screening Model will appear. The model
selection advisor or any of the component models can be accessed from Models on the menu bar.

1-4

2. Model Selection Advisor
The model selection advisor is designed to aid the user in determining which model is most applicable
for the current assessment Ile advisor divides the component models into two types: (1) water quality
models and (2) land use models. Since this version of the CS Model contains only one land use model,
the model selection advisor is most appropriate for selecting among the water quality models. If a land
use model is desired, the advisor can be used to view a brief description of the Watershed Model or to
activate the model, however, the Watershed Model can be activated directly from the menu bar.

The advisor contains a checklist of various water quality parameters and model options that the user
can select for each modeling effort. Based on the options selected by the user, the advisor determines
the appropriate model(s). If only one model fits the selected criteria, a screen containing a brief model
description appears. The user then has the option to perform another search or activate the selected
model. If the advisor selects more than one model, a model choice screen will appear, and the user can
view a description for each of the selected models. The user then has the option to perform another
search or activate one of the selected models.

Section 2.1 contains a detailed discussion of modeling and types of models. Section 2.2 discusses how
to access and navigate through the advisor. A description of all the water quality parameter options
available within the advisor is provided in Section 2.3. Criteria for final model selection am discussed
in Section 2.4.

2.1 Model Types
The CS Model is composed of both a land use model and several water quality models. These model
types are explained in greater detail in the following subsections. While all of these models are
available, none of them are linked together. Each component model operates seperately.

Models are a conceptualization of how the "real" world operates. No model completely reflects reality,
but models can be useful improving our understanding of natural systems and in assessing
anthropogenic impacts on these systems. Models can be defined in a vari ety of ways.

Empirical models are commonly termed "black box" models. These models use equations that are
derived from measured data, but do not attempt to describe all of the physical properties that are
involved in assessing the data. For example, fide charts have been developed based on the measured
tidal cycles. The use of these values does not require a complete understanding of how tides are
generated. It just important to know under what conditions the charts are valid. Empirical equations
explain the measured data for the given conditions. As conditions vary from the ones applied during
the equation development, the validity of the empirical equation is reduced. With greater number of
measured values and testing conditions, the applicability of the equations increases.

Many models use equations that attempt to reflect the processes involved. The factors in the equations
are based on real world processes. The advective-dispersive transport equation applied in many water
quality models is an example of a process model.

2-1

2.1.1 Land Use Models
Land use models evaluate the loadings of some constituent from the upland areas to a water body.
Typically modeled parameters include sediment and nutrients. Most land use models examine the
loadings from an entire watershed, are a commonly termed watershed models.

Watershed models can be either lumped or distributed parameter models. A lumped parameter model
will assess the loadings from each land use without considering the spacial variations or stream
routing. Each designaW land use is assumed to be uniform and have constant soil and cover
properties. A distributed parameter model allows for spatial variation. Typically, the watershed is
divided into a grid of cells, where each cell has defined land use, topography, and soil properties. ne
overland flow is routed from cell to cell or to a defined stream channel. A distributed parameter model
provides loadings throughout the watershed, not just at the watershed oudeL It also allows for
upstream land uses to impact downstream land use.

2.1.2 Water Quality Models
Water quality models evaluate the water quaI4 constituents in the water column of a stream or river.
Ile complexity of the model will depend on the number of stream dimensions that are being modeled,
whether temporal variations are being considered, and whether the solution is analytical or numerical.

Most water quality models are based on the principle of the conservation of mass. All system outputs
must equal system inputs. T'he mass balance calculation is performed for specified volumes of water
over a given time period. Material balances typically involve dissolved or suspend substances, such as
organic carbon, nitrogen, phosphorus, suspended sediment, and dissolved oxygen. This principle can
be applied to any substance whose transformation kinetics am known (McCutcheon and French, ?).
2.2 Operating the Advisor
To access the model selection advisor, select Models/Advisor from the main menu. The advisor
prompts you to select the type of model desired.

To select a land use model, click on the corresponding option button or enter <Alt L>. A screen
displaying the description of the Watershed. Model will appear. The model can be activated by
selecting [Run Model]. If you do not wish to run the model at this time, choose [Close] to return to
the main advisor screen.

To select a water quality model, click on the corresponding option button or enter <Alt W>. A screen
displaying all the parameter options will appear (Figure 2- 1). Ile various model options are grouped
by parameter type. Each of these parameter options is explained in Section 2.3. Select the desired
options by clicking on the appropriate option buttons or check boxes. The parameter option groups
can also be accessed by using the <Tab> key. Once the focus is on the desired parameter option
group, individual options can be selected using the up and down arrow keys.

2-2

Model Selection Advisor
. . . ........... .... .....

.. ........ .
pa, M-M

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

. .........

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

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

---------- -----
... . ...............

........ . .......
11 ... .......

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

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

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

.................
..... . ....... ... ...
X, EX@--@:i
.............

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

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

......................
.......... . ...
..............-
........................
Rim . . . . . . . .
. ..... ....
. . . . . . . . . .
. . . . . . . . . . . . . .

Figure 2-1. Model selection advisor parameter options.

After all the desired parameter options are selected, choose the [Start Advisor] button by either
clicking on it or pressing <Alt A>. The advisor will determine which models best match the desired
criteria. If only one water quality model matches the selected parameter options, a screen will appear
that provides a brief description of the model. You can either choose [Run Model] to activate the
selected model or [Close] to return to the parameter options screen. If more than one model meets the
desired criteria, a model choice screen will appear. The names of the water quality models that best
match the selected options are displayed on command buttons. To learn more about a particular model
or to activate a model, click on the command button displaying its name or press <Alt> and the
underlined letter in its name. A screen will appear that provides a brief description of the model.
Choose (Run Model] to activate the selected model, [Return] to return to the model choice screen to
evaluate other model options, or [Close] to return to parameter options screen.

2.3 Water Quality Parameters
The parameter options for water quality models are divided into five categories. Each of these
categories and the possible options are discussed in detail in the following subsections. Several of the
categories contain an (Unimportant] option. If this option is selected in any of the categories, then
that category is not used by the advisor during the model selection process. T'he water quality
parameters option group does not contain an [Unimportant] option. It is assumed that the user will
always be trying to model for at least one of these parameters.

2 - _3

To select the desired water quality parameters simple click on the appropriate option button or check
box. Ile various frames housing the options can be accessed using the <Tab> key. Once the desired
fi-arne has gained the focus, use the arrow keys to select the desired option(s).

2.3.1 Dimensions
Water quality models are classified based on the importance of longitudinal, laterad, and vertical
variations in the water quality constitutents. As more aspects of the stream are considered, the
complexity of the equations required to describe the variations increase. While water quality models
can be zero-, one-, two-, or three-dimensional, the component models included with the CS Model are
only one- or two-dimensional.

A one-dimensional model considers only the longitudinal variations in describing stream water qualiy.
It is assumed that "mixing processes will provide complete mixing both laterally and vertically, such
that the concentrations gradients are only along the axis of the river" (Thomann and Mueller, 1987).

A two-dimensional model will consider the water quality variations in two directions. Most two-
dimensional water quality models describe variations in the longitudinal and lateral direction and
assume complete vertical integration.

The selection screen includes a branching one dimensional model option which will be implemented in
future versions.

2.3.2 Time
This option addresses how water quality models address temporal variations. A model is classified as
being either steady-state or dynamic. A steady-state model provides an analysis for a specific set of
input conditions for some point or period of time (Basta and Moreau, 1982). A steady-state model can
also consider two or more time periods that are seperated by some number of time periods if there is no
analysis of these intervening time periods. In contrast, a dynamic model will examine successive time
periods, where the inputs from the first time period will affect the inputs into the second time period.
The analysis is much more complex than for models utilizing steady-state conditions.

2.3.3 Loadings
Point sources provide a constazit input to the stream from a specified location along the channel.
Typical point sources include wastewater treatment plants and processing plants. Nonpoint source
discharges are generated by overland flow and do not originate from one fixed location. Nopoint
source loadings are typically calculated by the use of a watershed model. However, some models can
consider a known nonpoint source load as an additional input.

2.3.4 Tidal
This option allows the user to decide if the effects of tidal variations should be explicitly considered.
Most water quality models address average tidal responses in the dispersion coefficient.

2.3.5 Water Quality Parameters
The water quality parameters are the constituents that one of the component water quality models will
be evaluating. Except for salinity, these parameters are typically selected as indicators of stream health
and water contamination. Salinity is typically modeled to establish dispersion coefficients.

2-4

Carbonaceous Biochemical Oxygen Demand (CBOD)@. CBOD is an indicator of organic pollution
measured in terms of the oxygen demand that can develop as the organics are degraded. Units are
commonly in mg/L.

Nitrogenous Oxygen Demand (NBOD): NBOD is the equivalent measure of the organic nitrogen
and ammoma, that will consume oxygen as they are converted to mtrite and nitrate. Units are
commonly in mg/L.

Dissolved Oxygen (DO): A direct measure of the amount of oxygery dissolved within the water. Units
are commonly in mg/L.

Coliforms: Colifbnns are bacteria in the Enterobacteriaceae fiiniily and are commonly use as an
indicator of fecal contamination. The coliforin group includes Escherichia coli, Enterobacter
aerogenes, and KkbsielZa pneumoniae. These organisms make up approximately 10 percent of the
intestmal microorganisms found in humans and other ammals. Coliforms are used as an indicator
species because they lose viability in water at slower rates than most of the major intestinal bacterial
pathogens (Prescott et aL, 1990). Water quality and potability is commonly evaluated by testing for
the presence of coliforni bacteria. Units are usually measured in the number of organisms/L.

Salinity: Salinity is a measurement of the salt concentrations in the water column. It is commonly
used to calibrate a finite section model for dispersion. It is trwted as a conservative material. Usually,
the only loads for salinity arise at the mouth of the estuary or river.

2.4 Model Selection
The model selection advisor counts the total number of options selected by the user and compares this
value to the number of these options that apply to each of the water quality models. The model that
matches the greatest number of options is selected. Under this scheme, each option has equal value.
This assumption may not reflect the needs of the user. The user may want one of the criteria to be
more significant than the others and this is not possible with the current equal weighting scheme. If
there is a critical option, the user can run the advisor with only that option selected and the other
parameters set to [Unimportant].

The model selection advisor can be a useful tool for model selection, especially when the user is first
becoming fluniliar with the various component models, however, there is no substitute for good
professional judgment. The user should become familiar with the characteristics, assumptions, and
limitations of all the component models to be able to accurately estimate input parameters and to assess
the model results.

2-5

3. Watershed Model
The Watershed Model is based on the Generalized Watershed Loading Functions model developed by
Haith et al. (1992) at Cornell University. The model predicts monthly and annual sediment, nitrogen,
and phosphorus loadings from complex watersheds. The Soil Conservation Service (SCS) Curve
Number Equation and the Universal Soil Loss Equation are used to predict runoff and erosion,
respectively. Individual land uses, point sources, and septic systems are evaluated. The Watershed
model is a lumped parameter model, and therefore, land use is considered uniform with respect to soil
and cover.

3.1 Modeling Approach
'Me Watershed model features a simple, daily time step hydrologic budget as depicted in Figure 3-1.
The daily unsaturated and shallow saturated zone water budget for day t in cm is

Precipitation
I I I
Evapotranspiration

Inf, tmtion

Unsatura
Zone Runoff

Percolation
Soil
Surface

Stream Channel

Figure 3-1. Conceptual model of a watershed, adapted from Haith et al, 1992.

3-1

U t+1 = Ut + Rt + Mt - Qt - Et - PCt

S t+1 = St + PCt - Gt - Dt

where Ut = unsaturated zone soil moisture
St = shallow saturated zone soil moisture
Rt = rainfall
Mt = snowmelt
Qt = watershed runoff Qxt, for all land uses k
Et = evapotranspiration
PCt = percolation into the shallow saturated zone
Gt = groundwater flow into the stream
Dt = deep seepage flow into the deep saturated zone

Values for rainfall and snowmelt are provided by the included weather files. Watershed
runoff is calculated by the SCS Curve Number Equation as

Qkt = (Rt + Mt - 0.2DSkt)2
__________________

Rt + Mt + 0.8DSkr

where DSkt is the storage detention parameter. A more detailed discussion of curve numbers and how
either (CVtPEt) or (Ut + Rt + Mt - Qt), where CVt is a cover coefficient and PEt is the potential
evapotranspiration. Hamon (1961) described the potential evapotranspiration as

PEt = 0.021H2tet
__________

Tt + 273

where Ht is the number of daylight hours per day during the month containing day t, et is the saturated
water vapor pressure in millibars on day t, and Tt is the temperature on day t in degree C. When Tt is
less than or equal to 0, then PE is set to zero. Bosen (1960) approximated teh saturate vapor pressure as

et = 33.8639 [(0.00738Tt + 0.8072)8 -0.000019(1.8Tt + 48) + 0.001316], for Tt greater than or = to 0

Percolation occurs when unsaturated zone water exceeds the available soil water capacity (U*) and is
the maximum of either zero or (Ut + Rt + Mt - Qt - Et - U*).

The groundwater discharge and deep seepage are modeled as a simple linear reservoir as described by
Haan (1972) by the following equations:

Gt = r St

Dt = s St

where r is the groundwater recession coefficient and s is the seepage coefficient. These coefficients are
described in Section 3.4.

3-2

3.1.1 Septic Systems
The septic system option in the Watershed Model is based on research presented by Mandel (1993).
The model will calculate the dissolved nutrient loads from septic system under four conditions:
nomt-A short-circuited, ponded, and direct discharge. These loads are determined from the per capita
daily effluent loads and the monthly population being served by septic system in each condition.

Normal Systems: These septic systems conform with guidehnes and standards suggested by the
USEPA and are in compliance with state regulations. The effluent from these systems infiltrates into
the shallow saturated zone. Nitrogen in the effluent is considered to be either removed by plant uptake
or converted to nitrate and transported to the stream by ground water discharge. While normal systems
contribute to dissolved nitrogen, the model does not consider phosphorus inputs. It is assumed that any
phosphorus from the effluent becomes adsorbed to soil particles, and hence, does not leave the system.
Since normal septic systems are operating properly, they are generally located away from streams.
This means that as the dissolved nitrogen from the effluent move through the shallow saturated zone, it
is diluted by other groundwater sources. Thus, the Watershed Model makes the monthly dissolved
nitrogen load proportional to the groundwater discharge to the stream.

Short-Circuited Systems: These septic systems have been placed in an inappropriate location, and
the effluent is not being properly treated by the septic field. These systems are located too close to
surface waters, and therefore, negligible adsorption of phosphorus occurs. The only mechanism for
nutrient removal is plant uptake. The Watershed Model computes dissolved loads for both nitrogen
and phosphorus.

Ponded Systems: Ponded septic systems are the result of hydraulic overload or hydraulic failure.
The effluent from these systems "ponds" over the adsorption field. If the temperature is below
fiwzing, than the ponded effluent is assumed to fi-eeze and nutrients can accumulate. Monthly nutrient
loads from these systems consider whether or not nutrients have accumulated due to frozen conditions.

Direct Discharge: These systems are illegal and discharge effluent directly to surface waters. The
nutrient loads from these systems are based solely on the per capita tank effluent and the population
served. Removal by plant uptake is not considered. These systems contribute to dissolved loads of
both nitrogen and phosphorus.

3.2 Model Format
The Watershed Model utilizes a notebook format to enter data and display model results. The first eight
notebook pages provide screens for input data and the last two notebook pages are used to present
tabular and graphical model results, respectively. Each page of the notebook can be accessed by
clicking on the corresponding tab or pressing <Alt> plus the underlined letter of the tab header.
Comprehensive discussions regarding model input and parameter estimation are provided in Section 3.3
to Section 3.9. Each of these sections corresponds to an individual notebook page in the Watershed
Model. Simulation results and model output are discussed in Section 3. 10.

For each notebook page used to enter data, there is an example data set. This data set can be used as a
tutorial for the Watershed Model. To access this data click on the [Example Data] buttons on each
notebook page. There is data available to meet all model options. The example weather file must also
be selected from the weather file list box. These data can be used to become familiar with how to
navigate through the model and it provides an example of typical input and model results.

3-3

The first notebook page, Project Info, serves as the Watershed Model's introductory screen. This
notebook page can be used for record keeping purposes. There are text boxes available to enter in the
project title, project number, date, and the names of the individuals involved with the model run. This
notebook page may be omitted if desired.

"Me menu bar on the Watershed Model provides the user with file access, editing capability, and help
information. Selecting FiJe will display a drop-down menu list with the following items: New, Open,
Save, Save As, Print, and Exit. New deletes the current data set. A message box will ask for
verification prior to deleting current values. Open calls up the Open File dialog box to select an
existing data set. Save will save the current data set using the existing file name. If a file name has not
been defined, the Save As File dialog box will appear prompting the user to enter a file name. Save As
opens the Save As File dialog box to allow the current data set to be saved with a new file name.
Selecting Print will call up the Print File dialog box. The user can select to print the input data and/or
any of the tabular output displays. Note that graphs are not printed from this menu item. The graphs
are printed directly from the Graphical Output notebook page using the [Print Graph] command
button. Eidt closes the Watershed Model and returns the user to the main CS Model window. If the
current data set has not been saved, it will be lost. Selecting Edit will display a drop-down menu fist
with Cut, Copy, and Paste. These menu items can be used to edit input data. Cut will remove the
selected text and move it to the windows clipboard. Copy will place a duplicate of the selected text on
the clipboard, but not remove the original text. Paste can be used to put text placed on the clipboard
back onto one of the input screens. Help will provide information pertaining to the Watershed Model.

3.3 Program Options
rhe Watershed Model will simulate stream flow, sediment yields, nutrient loads, and septic system
inputs. Click on the desired model option or press the <Tab> key to gain access to the option group
and use the arrow keys to select the desired model option. The required input parameters will vary
depending on the model option selected. Any notebook pages that contain unnecessary input
information for the selected model option Will have "greyed' tabs. These tabs are disabled, which
means that the user no longer has access to these pages. This feature aids the user 'in determining what
information is required to run the model.

3.3.1 Hydrologic Unit Data
Regardless of which model option is selected, land use data must be entered. If the land uses of the
watershed being modeled are not known, land use values can be retrieved from the Hydrologic Unit
Database (HUD). Hydrologic units define watershed boundaries and the HUD contains land use
information for each hydrologic unit in Virginia. The database information was provided by the
Virginia Department of Conservation and Recreation, Division of Soil and Water Conservation.
Acreage values were supplied for 13 land uses: cropland, hay, orchard, idle hawiland, Agricultural
Stabilization and Conservation Service (ASCS) set aside, Conservation Reserve Program (CRP) land,
pastureland, forest, urban residential, urban industrial, urban other, waterbodies, and Christmas trees.
This data was restructured to provide the VMRC with more manageable land use categories and to
meet the input requirements of the Watershed Model. Data for waterbodies were discarded because
they are not applicable to the Watershed Model. Acreage values for orchards and Christmas trees were
combined into one land use category since these two land uses would have similar model *input values.
Farmland that is enrolled in the CRP or has been designated as an ASCS set aside is not in production.
These land uses are likely to have model input values which are similar to idle farmland, and therefore-

3-4

their acreage values were combined with the acreage valu Ies for idle flirmland in each hydrologic unit.
This restructuring of the database information yielded nine land use categories in the HUD. Upon
completion of the database restructuring, the acreage values for these nine land uses were converted to
hectares to meet the model input requirements.

To access the HUD, select the [Yes] option under "Use Hydrologic Unit data?". A frame will appear
that contains a drop-down list box for selecting the appropriate hydrologic unit If the appropriate
hydrologic unit is not known, it can be determined by using the map viewer. The map viewer list box
contains the names of all the counties in tidewater Virginia. Select the name of the county where the
watershed currently being modeled is located and choose the [View] command button. A picture box
containing a map of the selected county will be displayed. The map defines all of the hydrologic units
within the county. Various parts of the map can be viewed by'selecting one of the command buttons
beneath the picture box or using the horizontal and vertical scroll bars. Once the appropriate
hydrologic unit has been determined, select that value from the hydrologic unit list box and choose
(Close] to exit the map viewer. The types of land uses and their corresponding areas will automatically
be entered into the Land Uses notebook page.

Because land use information changes over time, the Virginia Department of Conservation and
Recreation, Division of Soil and Water Conservation should be periodically contacted to obtain the
most recent data.

3.3.2 Climate Data
The Watershed Model requires temperature and precipitation data. This data is provided for nine
locations in Virginia@ Table 3-1 contains a detailed description of the location where the climate data
was collected and can be used to select the most appropriate weather file for the area being modeled. A
tenth weather file provides climate data that corresponds to the example data seL Use the scroll bar to
view all the available weather files and select the desired weather file by clicking on it.

The weather files are organized by month. The first entry is the number of days in the month and
subsequent entries are daily temperature ( *Q and precipitation (cm) values. The weather files are
arranged to correspond to the assumptions of the model. "Both the groundwater and sediment portions
of the model require that simulated years begin at a time when soil moisture conditions are known and
runoff events have flushed the watershed of the previous year's accumulated sediment" (Haith et al.,
1992). In Virginia, this corresponds to early spring, and therefore, the weather files provided with the
model are arranged in April to March weather years.

The climate data was provided by the USGS and converted to the format required by the model. Daily
maximum and minimum temperatures were averaged and then converted from Fahrenheit to Celsius.
Daily precipitation values were converted from inches to centimeters. Some of the weather files
contained data gaps. For a missing daily temperature measurement, the temperatures from the day
prior to and after the missing measurement were averaged. For a missing daily precipitation
measurement, the value was set to zero. If there were numerous missing measurement points or
consecutive missing measurement points, then the data for the entire year was discarded.

After a weather file has been selected, the length (number of years) of the climate record will be
displayed in the simulation time text box. The simulation length cannot exceed this value, but a shorter
simulation time may be entered. To enter a value in the text box, move the mouse cursor to the text

3-5

box and click the left mouse button to set the focus to the text box and then type the desired simulation
length.

Table 3-1. Weather File Locations and Record Lengths

Station
Location No. Lat Long Dates
Blackstone FAA Airport 773 N37:05:00 W077:57:00 April 1949 - March 1972
(Nottoway County) (23 years)
Corbin 2009 N38:12:00 W077:22:00 April 1959 to March 1992
(Caroline County) (33 years)
Louisa 5050 N38:02:00 W078:00:00 April 1949 to March 1992
(Louisa County) (minus 1968, incomplete)
(42 years)
Nassawadox 5931 N37:28:00 W075:52:00 April 1957 to March 1976
(Northampton cEintj) (19 years)
Onley 1 S 6362 N37:41:00 W075:43:00 April 1930 to March 1955
(Accomack County) (25 years)
Richmond WSO Airport 7201 N37:30:00 W077:20:00 April 1942 to March 1992
(Henrico County) (minus 1951, incomplete)
(42 years)
Wallops Island WSSF 8849 N37:56:00 W075:28:00 April 1967 to March 1980
(Accomack County) (13 years)
Warsaw 2 N 8894 N37:59:00 W076:46:00 April 1951 to March 1992
(Richmond County) (I =*Us 1970, incomplete)
1(40 years)
Washington WB; Chantilly 8903 N38:57:00 W077:27:00 April 1963 to March 1992
(Loudon County) (29 years)

3.4 Initial Conditions
The values for the initial conditions can be entered directly into each text box by clicking on it and then
entering the desired value from the keyboard. The text boxes may also be accessed using the <Tab>
key. As each text box becomes active, its corresponding label win turn blue.

Unsaturated Available Water Capacity (U*): 'Me available unsaturated zone soil moisture capacity
is used to estimate percolation to the groundwater, which occurs when the unsaturated zone water
exceeds the available soil water capacity (Haith et al., 1992). "In principle, U* is equivalent to a mean
watershed maximum rooting depth multiplied by a mean volumetric soil available water capacity. The
latter also requires determination of a mean unsaturated zone depth, and this is impractical for most
watershed studies. A default value of 10 cin can be assumed for pervious areas, corresponding to a
100 cm rooting depth and a 0. 1 cm/cm volumetric available water capacity" (Haith et al., 1992).
Selecting the [Defaultj button will assign a value of 10 cm to U*.

Sediment Delivery Ratio: Ile sediment delivery ratio is the ratio between the amount of sediment
yield and the gross erosion in a watershed (Gottschalk, 1964).

3-6

Table 3-2 shows the effect of drainage basin size on the sediment delivery ratio. Figure 3-2 provides a
commonly used area-based relationship from Vanoni (1975) that can be used to estimate the sediment
delivery ratio based on the area of the watershed being modeled. Since the sediment delivery ratio is
very site specific, selecting the [Default] button will not provide a value for this parameter.

Table 3-2. Sediment Delivery Ratio Based on Watershed Size
Drainage Area (lan) Sediment Delivery Ratio (percent)
0.1 53.0
0.5 39.0
1.0 35.0
5.0 27.0
10.0 24.0
50.0 15.0
100.0 13.0
200.0 11.0
500.0 8.5
26,000 4.9
Sow=. Robiuson, 1979.

Watershed Sediment Delivery Ratio

I lit I I III

it
M @.- I I
0.1

1 17%- 1 1 It

0.01
1 10 100 1000 10000

Watershed Area (kmj

Figure 3-2. Watershed sediment delivery ratio, adapted from Vanoni, 1975.

Recession Coefficient (r): The baseflow recession coefficient is used to determine ground water
discharge to stream flow. Standard hydrograph separation techniques can be used to estimate the
recession coefficient from stream flow records.

3-7

Recession coefficients are measured for a number of hydrographs and an average value is used for the
simulations. Typical values range from 0.0 1 to 0.2. The [Defaulti button supplies an arbitrary value
Of 0. 1.

Seepage Coefficient (s): ne rate constant for deep seepage loss is used during the calculation of the
groundwater discharge to stream flow. The seepage coefficient is multiplied by the daily shallow
saturated zone soil moisture to determine the amount entering into the deep saturated zone, and thereby
leaving the watershed. There are no standard techniques available fbr estimating the seepage
coefficient. If the coefficient cannot be determined by calibration, then a conservative approach is to
assume that all precipitation exits the watershed by either evapotranspiration or stream flow, and
therefore deep seepage equals zero. Selecting the [Default] button will set s to zero.

Unsaturated Storage - The initial unsaturated soil zone moisture in cm.
Saturated Storage - The initial shallow saturated soil zone moisture in cm.
Snowmelt Water - The amount of snow melt water in cin at the beginning of the simulation.
Antecedent Rain and Melt - The amount of ram and snow melt for the five previous days.

'Me preceding four parameters are difficult to characterize, but they will not affect simulation results
beyond the first several months. To ameliorate this problem, assign arbitrary values to these initial
conditions and discard the results from the first year of the simulation. A common approach is to
assign the value of the unsaturated available water capacity (U*) to the unsaturated storage and zero to
the remaining variables. The [Defaultl button will assign a value of zero to all the following variables,
except for unsaturated storage, which is assigned 10 cm.

Selecting the [Example Data] button will provide initial conditions data for the example watershed.
To provide reasonable output data, this data should be used in conjunction with the example data
provided for the other input parameters and the example weather file.

The (Cancel] button will clear all values from the initial conditions text boxes. If this button is
selected, a message box will appear asking for verification prior to deleting current values.

3.5 Land Use Information
A minimum of one land use with a non-zero area (in hectares) must be specified prior to beginning any
simulation. A land use without a corresponding area value- is not allowed. To model land use changes
over time, a run must be performed with the initial land use cover and then a second run must be
performed with the modified land uses.

Land uses are divided into two categories: rural and urban. This separation is required because of
model formulation. The Watershed Model uses a modification of the Universal Sod Loss Equation
(USLE) to determine erosion from rural sources. The USLE was not developed for urban land uses,
and is therefore not applied to these areas. In addition, the model will calculate both a dissolved and a
solid phase nutrient load from rural sources, but urban nutrient loads are modeled as being entirely a
solid phase.

To enter land use values into either the Rural Land Use or the Urban Land Use spreadsheet simply
click on the desired cell. "Me spreadsheets can also be accessed using the <Tab> key. Once the
desired spreadsheet.has gained the focus, use the arrow keys to move from cell to cell. For each land

3-8

use in the watershed, enter its name and area in hectares. : If the land use name you wish to enter is two
words, enter the name without a hyphen. When the model graphs the results, it will use the first two
letters from each word in the name for the graph label if the name consists of two discrete words. For
example, the land use "Impervious Residential" would have a graph label of "Im-Re". but a land use
designated as "Impervious-Residential" would have a graph label of "Imper".

Rows my be removed or added to either the Rural or the Urban Land Uses spreadsheets by clicldng on
the (Delete Row] or [Insert Row] buttons. When either of these command buttons are selected, the
active raw in the spreadsheet will be altered accordingly.

Selecting the (Example Data] button will provide land use data fbr the example watershed. To provide
reasonable output data, this land use data should be used in conjunction with the example data
provided for the other input parameters and the example weather file.

The [Cancel] button will clear aU values from both the Rural and the Urban Land Uses spreadsheets.
If this button is selected, a message box will appear asidng for verification prior to deleting all cell
values.

3.5 *1 Runoff Curve Numbers
For each rural and urban land use entered in the spreadsheet, a runoff curve number (CN) must be
assigned. Curve numbers represent a relationship between precipitation and runoff volume. Ile Soil
Conservation Service developed a method to estimate excess rain volume (runoff) based on the
precipitation volume and the volume of total storage. The storage parameter (DS) in centimeters is
obtained from the equation
DS = 2540 - 25A
CN
Curve numbers exist for three antecedent moisture conditions: (1) CNlk is for below-average (dry)
moisture conditions; (2) CN2k is for average moisture conditions; and (3) CN3,, is for above-average
(wet) moisture conditions. The Watershed Model requires CN2k to be input. The model computes the
values for CN I k and CN3k from CN2k. During simulations, the model will evaluate the current
moisture conditions and supply the appropriate curve number. Table 3-3 to Table 3 )-5 contain
suggested curve numbers for average antecedent moisture conditions (CN20 for a variety of land uses
and are based on soil hydrologic groups. A description of the four soil hydrologic groups, for both
undisturbed and disturbed soils, is provided in Table 3-6. Disturbed soils are characterized by a major
alteration of the soil profile, as would occur from construction or development.

3-9

Table 3-3. Runoff Curve Numbers for Cultivated Agricultural Land
Hydrologic Soil Hydroi Zic Group
Land Use/Cover' Conditionb A B C I D
Fallow Bare Soil N/A 77 86 91 94
Crop Residue Cover Poor 76 85 90 93
Good 74 83 88 90
Row Crops SR Poor 72 81 88 91
Good 67 78 85 89
SR + CR Poor 71 80 87 90
Good 64 75 82 85
C Poor 70 79 84 89
Good 65 75 82 86
C+CR Poor 69 78 83 87
Good 64 74 8 r 85
Poor 66 74 so 82
Good 62 71 788 81
C&T + CR Poor 65 73 79 81
Good 61 70 77 80
Small Grains SR Poor 65 76 14 11
Good 63 75 83 87-
SR+CR Poor 64 75 83 86
Good 60 72 80 84
C Poor 63 74 82 85
Good 61 73 81 84
C+CR Poor 62 733 81 84
Good 60 72 80 83
C&T Poor 61 72 79 82
Good 59 70 78 81
C&T + CR Poor 60 71 78 81
Good 58 69 77 80
Close-seeded or SR Poor 66 77 85 89
Broadcast Legumes Good 58 72 81 85
or Rotation Meadow C Poor 64 75 83 85
Good 1 55 1 69 78 83
C&T Poor 63 73 80 83
Good .51 67 76 f 80
Source. Soil Conservation Service, 1986.
' CR = Crop Residue; SR = Straight Row, C = Contoured-, C&T = Contoured and Terraced
bHydrologic condition is based on a combination of factors that affect infiltration and runoff, including: (1) density and
canopy of vegetative areas, (2) amount of year-round cover, (3) amount of close-seeded legumes in rotations, (4) percent
of residue cover on the land surflice (good @: 20%), and (5) degree of surface roughness.

3-10

Table 3-4. Runoff Curve Numbers for Other Rural Land

Hydrologic So' Hydrol *c Gro P
Land Use/Cover Condition A B C D
Pasture, grassland or range - continuous forage Poor 68 79 86 89
for grazine Fair 49 69 79 84
Good 39 61 74 80
Meadow - continuous grass, protected from N/A 30 58 71 78
grazing,_ generally mowed for hay I
Brush - brush/weeds/grass mixture with brush the Poor 48 67 77 83
major elemen? Fair 35 56 70 77
Good 30 48 65 73
Woods/grass combination (orchard or tree fim)c Poor 57 73 82 16
Fair 43 65 76 82
Good 32 58 7Z 79
Wood? Poor 45 66 77 3
Fair 36 60 73 79
77
Good 1 30 55 70
Farmsteads N/A 59 1 74 82 86
Source: Soil Conservation So-vice, 1986.
Poor. <50% ground cover or heavily grazed with no mulch; Fair. 50-701/6 ground cover and not heavily grazed; Good:
>75% ground cover and lightly or only occasionally grazed.
b poor
. <50% ground cover, Fair. 50-70% ground cover, Good: >75% ground cover.
Estimated as 50% woods and 50% pasture.
d Poor. forest litter, small trees and brush are destroyed by heavy grazing or regular burning; Fair. woods are grazed but
not burned and some forest litter covers the soils-, Good: woods are protected from grazing and litter and brush
adequately cover the soil.

Table 3-5. Runoff Curve Numbers for Urban Areas

Soil Hydrolozic Gro v
Land Use/Cover A B C D
Open space (lawns, parks, golf courses, cemeteries, etc.):
Poor condition (grass cover < 50%) 68 79 86 89
Fair condition (grass cover 50-75%) 49 69 79 84
Good condition (grass cover> 75 %) 39 61 74 80
Paved parldng lots, driveways, roofs, etc. 98 98 98 98
Streets and roads:
Paved with curbs and storm sewers 98 98 98 98
Paved with open ditches 83 89 92 93
Gravel 76 85 89 91
Dirt 72 82 87 _j 89
Average Imperviousness" (0/0) -
Residential average lot size,
0.05 ha (1/8 acre) 65 77 85 90 92
0. 10 ha (1/4 acre) 38 61 75 83 87
0. 15 ha (1/3 acre) 30 57 72 81 86
0.20 ha (1/2 acre) 25 54 70 80 85
0.4 ha (I acre) 20 51 68 79__, 8@
Commercial and business 85 (@@) 89 92 94 95
Industrial districts 72 81 88 91 93
Source: Novotny and Olem, 1994 and the Soil Conservation Service, 1986.
' The remaining pervious areas (lawns) are considered to be in good pasture condition for these curve numbers.
b Curve numbers are computed assummg the runoff from the house and driveway is directed toward the street with a
minimum of roof water directed to lawns where additional infiltration could occur.

3-12

Table 3-6. Description of Soil Hydrologic Groups

Soil
-Group Description
Undisturbed Soils
A Low runoff potential and high infiltration rates even when thoroughly wetted. Chiefly deep,
well to excessivelv drained sands or gravels. High rate of water transmission (>0.75 cm/hr).
B Moderate infiltration rates when thoroughly wetted. Chiefly moderately deep to deep,
moderately well to well drained soils with moderately fine to moderately coarse textures.
IModerate rate of water tz-ansmission (0.40-0.75 cnVhr).
C Low infiltration rates when thoroughly wetted. Chiefly soils with a laver that impedes
downward movement of water, or soils with moderately fine to fine texture. Low rate of
water transmission (0. 15-0,40 cm/hr).
D High runoff potential. Very low infilti-ation rates when thoroughly wetted. Chiefly clay soils
with a high swelling potential, soils with a permanent high water table, soils with a claypan
or clay layer at or near the surface, or shallow soils over nearly impervious material. Very
Ilow rate of water transmission (0-0. 15 cnvhr).
Disturbed Soils
A Sand. loamy sand. sandy loam
B Silt loam, loam
C Sandy clal, loam
D IClav loam siltv clay loam, sandy'clay, silty clay, clay
Source: Soil Conservation Service, 1986.

3.5.2 Universal Soil Loss Equation Factors
The Universal Soil Loss Equation (USLE) is the most widely accepted estimator of sod loss caused by
upland erosion (Novotny and Chesters, 198 1; Schwab and Frevert, 1985; Wischmeier and Smith,
1978). The USLE is an empirical equation that was developed through statistical analysis of more
than 40 years of measured soil losses from many small, experimental plot studies. The Agricultural
Research Service of the USDA established soil erosion experiment stations at different geographic
locations throughout the country that represented a wide range of sod and climatic conditions. The
USLE was developed by Wait Wischmeier and Dwight Smith from the dam collected at these
experiment stations and other available information from a network of state and federal research units
(Browning, 1979; Sandels, 1986). The USLE was originally developed to provide a reliable means of
selecting adequate erosion control practices for farm fields and construction areas. More recently, the
USLE has been used to predict sediment losses for pollution control programs or as an erosion
estimator in land use models.

Variables influencing upland erosion are: climate, soil properties, vegetation, topography, and human
activities (Novotny and Chesters, 198 1). The USLE accounts for these erosion variables by estimating
average annual soil loss by the following six major fiactors:
A = RKLSCP

where.
A = average annual soil loss mi tons per acre (t/a),

R = raimUl and ninoff factor, which is the number of rainfall erosion index units plus a
factor for runoff from snowmelt or applied water where such runoff is significant,
K = soil-erodibility fiLctor *in t1a, which is the average soil loss per unit of erosion index for a
soil in cultivated continuous fallow with a slope length of 72.6 ft and slope of 9 percent,
L = slope-length factor, which is the ratio of soil loss from the field slope length to that from
a 72.6 ft length under identical conditions,
S = slope-steepness flictor, which is the ratio of soil loss from the field slope gradient to that
from a 9 percent slope under otherwise identical conditions,
C = cover and management flictor, which is the ratio of soil loss for given conditions to that
from cultivated continuous fallow, and
P = conservation practice factor, which is the ratio of soil loss for a given practice to that for
straight row firming up and down the slope.

Ile Watershed Model uses a modification of the USLE to calculate erosion from rural sources. Since
the USLE is an empirical equation which was derived from data for agricultural and rural land uses,
applying it to urban land areas is not appropriate. For each rural land use, an "erosion product" must
be specified. This is the product of KLSCP. The Watershed Model uses this erosion product to
calculate the erosion from each source area k on day t in Mg as
Xkt = 0. 132 RE, Xk (LS)k Ck PkI ARk
where, ARk is the area of source area k and RF, is the rainffl erosivity on day t. Unlike all it's other
inputs, the Watershed Model expects that the value of K will be in English units. 'Me model converts
the erosion product to metric units with the flictor 0. 132.

The Watershed Model estimates rainfall erosivity by the empirical equation developed by Richardson et
al. (1983) as
REt = 64.6 at &1-81
where, at is the rainfall erosivity coefficient and R, is the rainfall on day t. Ile rainfall erosivity
coefficient is *input on the Evapotranspiration Conditions notebook page and the model uses the values
in the weather file for rainfall values.

Estimating the USLE Factors

A brief description of each of the factors and a means to estimate them is provided. This is only an
abbreviated overview of usm*g the USLE. A comprehensive discussion of the equation's derivation,
fiLctors and their estimation, applications, and limitations can be obtained in Wischmeier and Smith
(1978).

Soil Erodibility Factor (K-): This is a measure of the potential erodibility of the soil. It is based on
soil properties. Values for K have been experimentally determined. Table 3 )-7 contains values of K
based on soil texturt and organic matter content. Representative values of K for most soils types and
texture classes can be obtained from SCS offices. Site-specific values of K can also be calculated
using soil erodibility nomograph techniques.

Table 3-7. Values of Soil Erodibility Factor (K) in t/a

Organic Matter Content %)
Texture < 0.5 2 4
Sand 0.05 0.03 0.02
Fine sand 0.16 0.14 0.10
Very fine sand 0.42 0.36 0.28
Loamy sand 0.12 0.10 0.08
Loamy fine sand 0.24 0.20 0.16
Loamy very fine sand 0.44 0.38 0.30
Sandy loam 0.27 0.24 0.19
Fine sandy loam 0.35 0.30 0.24
Very fine sandy loam 0.47 0.41 0.33)
Loam 0.38 0.34 0.29
Silt loam 0.48 0.42 0.33
Silt 0.60 0.52 0.42
Sandy clay loam 0.27 0.25 0.21
Clay loam 0.28 0.25 0.21
Silty clay loam 0.37 0.32 0.26
Sandy clay 0.14 0.13 0.12
Silty clay 0.25 0.23) 0.19
Clay 0.13-0.29
Sou=: Stewart et al., 1975.

Topographic Factor (LS): The effects of slope-length and slope-steepness have been research
separately and represent individual factors in the USLE, but they are commonly combined into a single
topographic factor for application purposes. Table 3-8 provides LS values for some common slope
lengths and percent slopes. Other combinations of length and gradient can be determined by
interpolating between the values in Table 3-8. The values in Table 3-8 were calculated by
LS = (V72.6)- (65.41 sin29 + 4.56 sinO + 0.065)
where, I is the slope length in feet, 0 is the angle of slope, and in is the slope-length exponent based on
percent slope. For slopes of 5 percent or greater, m equals 0.5. For slopes of 3.5 to 4.5 percent, m
equals 0.4. On slopes of I to 3 percent, in equals 0.3, and on uniform gradients of less than I percent,
rn equals 0.2

Cover and Management Factor (C): Also called the cropping management factor and the vegetative
cover factor, C "estimates the effect of ground cover conditions, soil conditions, and general
management practices on erosion rate" (Novotny and Chesters, 198 1). 'Me C Factor reflects the
amount of protection against raindrop impact and the subsequent soil particle displacement. C is a
dimensionless factor with values ranging between 0 and 1. A cover and management factor of I
represents continuous fallow tilled up and down the slope. This condition would potentially yield the
greatest amount of erosion. As the value of C approaches zero, the vegetative cover or management
efforts have reduced the potential for erosion. Table 3 )-9 provides some common values for C for
agricultural land uses. The cover and management factor can also be applied to construction sites.
Table 3-10 contains C Factor values and slope-length limits for construction sites. Factor C is usually
provided in terms of an average annual value for a particular combination of crop systems and
management (Wischmeier and Smith, 1978).

Table 3-8. Values for the Topographic Factor (LS)
Percent Sl ve Leneth (feet)
Slope 25 50 75 100 150 200 300 400 1 500 600 800 1000
0.2 .060 .069 .075 .080 .086 .092 .099 .105 .110 .114 .121 .126
o.5 .073 .083 .090 .096 .104 .110 .119 .126 .132 .137 .145 .152
0.8 .086 .098 .107 .113 .123 .130 .141 .149 .156 .162 .171 .179
2 .133 .163 .185 .201 .227 .248 .280 .305 .326 .344 .376 .402
3 .190 .233 .264 .287 .325 .354 .400 .437 .466 .492 .536 .573
4 .230 .303 .357 .400 .471 .528 .621 .697 .762 .820 .920 1.01
5 .268 .379 .464 .536 .656 .758 .928 1.07 1.20 1.31 1.52 1.69
6 .336 .476 .583 .673 .824 .952 1.17 1.35 1.50 1.65 1.90 2.13
8 .496 .701 .959 .992 1.21 1.41 1.72 1.98 2.22 2.43 2.81 3.14
10 .685 .968 1.19 1.37 1.68 1.94 2.37 2.74 3.06 3.36 3.87 4.33
12 .903 1.28 1.56 1.80 2.21 2.55 3.13 3.61 4.04 4.42 5.1-1 5.71
14 1.15 1.62 1.99 2.30 2.81 3.25 3.98 4.59 5.13 5.62 6.49 7.26
16 1.42 2.01 2.46 2.84 3.48 4.01 4.92 5.68 6.35 6.95 8.03 8.98
18 1.72 2.43 2.97 3.43 4.21 3.86 5.95 6.87 7.68 9.41 9.71 10.9
20 1 2.04 1 2.88 1 3.53 1 4.08 1 5.00 1 5.77 1 7.07 1 8.16 1 9.12 1 10.0 111.5 1 12.9
Source: Wischmeier and Smith, 1979.

Table 3-9. Values of C for Cropland, Pasture, and Woodland

Land Use C
Continuous fallow tilled up and down slope 1.0
Shortly after seeding or harvesting 0.3-0.8
For crops during main part of growing season
Corn 0.1 -0.3
Wheat 0.05-0.15
Cotton 0.4
Soybeans 0.2-0.3
Meadow 0.01-0.02
For permanent pasture, idle land, unmanaged woodland
Ground cover 95 - 100% as grass 0.003
Ground cover 95 - 100% as weeds 0.01
Ground cover 80% as grass 0.01
Ground cover 80% as weeds 0.04
Ground cover 60% as grziss 0.04
Ground cover 60% as weeds 0.09
For managed woodland
Tree canopy of 75 -100% 0.001
Tree canopy of 40 - 75% 0.002 - 0.004
Tree canopy of 20 - 40% 0.003-0.01
Source: Novotnv and Chesters, 1981

Table 3-10. C Factor Values and Slope-Length (LS) Limits for Construction Sites

Application
Mulch Type (tonnes/ha) Slope (%) C LS
No mulch or seeding All 1.0
Straw or hay tied down by anchoring 2.25 <5 0.2 60
and UwJdng equipment used on slope 2.25 6-10 0.2 30
3.4 <5 0.12 90
3.4 6-10 0.12 45
4.5 <5 0.06 100
4.5 6-10 0.06 60
4.5 11-15 0.07 45
4.5 16-20 0.11 30
4.5 21-25 0.14 23
Crushed stone 300 <15 0.05 60
300 16-20 0.05 45
300 21-33 0.05 30
540 <20 0.02 90
540 21-35 0.02 60
Wood chips 15 <15 0.08 23
15 16-20 0.08 15
27 <15 0.05 45
27 16-20 0.05 23
56 <15 0.02 60
56 16-20 0.02 45
56 21-331 0. 2 30
Asphalt emulsion 12 m'/ha 0.03
C - During first 6 C - After the 6th
weeks of gro week of growth
Temporary seeding with grain or fitst-
growing grass with:
No mulch 0.70 0.10
Straw 2.25 0.20 0.07
3.4 0.12 0.05
Stone 300 0.05 0.05
540 0.02 0.02
Wood chips 15 0.08 0.05
27 0.05 0.02
;.6 0.02 0.02
Sod 0.01 0.01
Source: Novotny and Olern, 1994.

C onservation Practice Factor (P): The conservation practice factor accounts for management
activities which slows runoff water and thus reduces sediment transport capacity, thereby retaining
detached soil particles near their sources. This includes contouring, compacting, strip cropping, and
establishing sediment basins. P is a dimensionless factor with values ranging between 0 and 1. The P
Factor has a value 'of I when no management activities are applied. Table 3-11 and Table 3-12
provide values for the conservation practice factor.

Table 3-11. Values of P for Agricultural Lands

Strip Croppina and Terr 9
Slope (percent) Contouring Alternate Meadows Closegrown
0-2.0 0.6 0.3 0.45
2.1-7.0 0.5 0.25 0.40
7.1-12.0 0.6 0.30 0.45
12.1-18.0 0.8 0.40 0.60
18.1-24.0 0.9 0.45 0.70
> 24 1.0- 1.0 1.0 j
Source: Novotny and Olem, 1994.

Table 3-12. Values of F for Construction Sites

Erosion Control Practice P
Surface Condition with No Cover
Compact, smooth, scraped with buffdozer or scraper up and down hiR 1.30
Same as above, except raked with bulldozer or scraper up and down hifl 1.20
Compact, smooth, scraped with bulldozer or scraper across slope 1.20
Same as above, except raked with bulldozer or scraper across slope 0.90
Loose as a disked plow layer 1.00
Rough irregular surface, equipment tracks in all directions 0.90
Loose with rough surface > 0.3 in depth 0.80
Loose with smooth surface < 0.3 in depth 0.90
Structures
Small sediment basins - 0.09 ha basintha 0.50
Small sediment basins - 0. 13 ha basin/ha 0.130
Downstream sediment basin with chemical flocculants 0.10
Downstrearn sediment basin without chemical flocculants 0.20
Erosion control structures - normal rate usage 0.50
Erosion control structures - high rate usage 0.40
Strip building 0.75
Source: Novotny and Olem, 1994.

Estimation of KLSCP for the Hydrologic Unit Database: The following is a rough estimate of
USLE factor values to use for the land uses in the HUD. These values are for the tidewater area of
Virginia and are based on the following assumptions:
� Fairly permeable sandy to sandy loamy soils
� Fairly level terraMi with a percent slope of 5 percent

� A slope length of 500 feet
� No support practices are being applied
When evaluating a potential source of poflution, the worst case scenario is often used to ensure that any
water quality impacts would be detected. This rationale was applied during the selection of the slope-
length factor and the conservation practice flictor. As slope length increases, the potential for soil
detachment and transport also increases, therefore a slope length of 500 feet was selected as a
conservative value. By using a large value for the slope length, a greater potential for erosion exists.
Although Table 3-8 contains data extrapolated to 1000 feet, these values have not been validated with
field measurements, and could increase the uncertainty of model calculations. A conservation practice
flictor of I was selected, since this implies that nothing is being done to minimize erosion on any of the
land uses.

Selection of the soil erodibility flictor is somewhat arbitrary. This factor should be estimated based on
the average of known sod types for each land use. A constant value was selected for all land,uses and
is based on general property values of the Coastal Plain area.

A percent slope value of 5 percent was selected as an average value for the tidewater area. As slopes
steepen, soil loss increases much more rapidly than runoff and a large slope value could have
tremendous impact on model results. There are certainly slopes 'in the tidewater area that are steeper
than 5 percent, but to use the maximum value for this ffictor would present a very unlikely scenario for
a watershed.

Values selected for the cover and management factor are average values for each land use and have
been selected from the literature values.

The estimates contained in Table 3-13 are an educated best guess, and could be inaccurate if the area
being modeled does not match the assumptions. It is always best to use the most site-specific
information whenever possible.

Table 3-13. Estimated Values for Land Uses in the Hydrologic Unit Database

Land Use K LS C P KLSCP
Cropland 0.24 1.20 0.35 1.0 0.1008
Forest 0.24 1.20 0.005 1.0 0.0014
Hay 0.24 1.20 0.02 1.0 0.0058
Idle Farmland 0.24 1.20 0.1 1.0 0.0288
Orchards/Christmas Trees 0.24 1.20 0.1 1.0 0.0288
Pasture 0.24 1.20 0.08 1.0 0.0230

Limitations

The USLE estimates gross erosion from sheet and rill erosional processes. It does riot consider gully or
channel erosion, which can be a significant source of erosion in some watersheds. Also. the USLE does
not address soil loss by wind erosion.

3-19

"Me USLE is designed to predict longtimc-average sail losses for specified conditions" (Wischmeier
and Smith, 1978). It is not recommended for trying to predict soil losses from a specific storm event.

The USLE estirriates only erosion potential. It does not estimate transport or deposition, and therefore
it does not provide a direct estimate of sediment yield. To estimate the sediment yield, the USLE must
be multiplied by a delivery ratio for the watershed.

"Soil losses computed with the USLE are best available estimates, not absolutes. They will generally
be most accurate for medium-textured soils, slope lengths of less than 400 feet, gradients of 3 to IS
percent, and consistent cropping and management systems that have been represented in the erosion
plot studies. The farther these limits are exceeded, the greater will be the probability of significant
extrapolation error" (Wischmeier and Smith, 1978).

3.6 Nutrient Input
The Nutrient Info notebook page contains two spreadsheets: one for rural land uses and one f6r urban
land uses. The land use category names entered into the land uses spreadsheets on the Land Uses
notebook page will be automatically transferred to the nutrient spreadsheets. For the rural land uses,
the dissolved nitrogen and phosphorus concentrations in mg/L must be entered. Table 3-14 contains
flow weighted nutrient concentrations measured by Dornbush et al. (1974) for several agricultural land
uses. For urban land uses, the nutrient build-up in kg/ha-day must be entered. Table 3-15 provides
nutrient accumulation rates for urban areas around northern Virginia as measured by Kuo et al. (1988).

'Me Watershed Model requires groundwater concentration values for nitrogen and phosphorus in mg/L.
These values should be computed from area weighted averages for land use type. According to Reay
(1994), typical ground water nitrogen concentrations as N03-N are < I ing/L for forest, 2 mg/L for
pasture and residential areas, and 10 mg/L for agriculture. Typical phosphorus concentrations are
0. 15 mg/L for all mentioned land uses.

Soil nutrient levels must be entered in mg/kg for both nitrogen and phosphorus. These values should be
computed from area weighted averages for land use type.

Ile values for the groundwater and soil nutrient parameters may be entered by selecting the
appropriate text box with the mouse or using the <Tab> key until the desired text box gains the focus.
As each text box gains the focus, it's label will turn blue for easy identification.

Selecting the [Example Data] button will provide nutrient concentration and loading data for the
example watershed. To provide reasonable output data, this data should be used in conjunction with
the example data provided for the other input parameters and the example weather file.

'Me [Cancel] button will clear all values from the nutrient information text boxes and spreadsheets. If
this button is selected, a message box will appear asking for verification prior to deleting current
values.

Table 3-14. Dissolved Nutrients in Agricultural Runoff

Land Use Nitrogen (mg/L) Phosphorus (mg/L)
Fallow 2.6 0.10
Corn 2.9 0.26
Small grains 1.8 0.30
Bay 2.8 0.15
Pasture 3.0 0.25
Source: Donibush et al., 1974.

Table 3-15. Nutrient Accumulation Rates for Northern Virginia Urban Areas

Land Use Total Nitrogen ftlha-day) Total Phosphorus ft/ha-day)
Impervious Surfaces
Single family residential
Low density (uniwha < 1.2) 0.045 0.0045
Medium density (units/ha 2:1.2) 0.090 0.0112
Townhouses and apartments 0.090 0.0112
High rise residential 0.056 0.0067
Institutional 0.056 0.0067
Industrial 0.101 0.0112
Suburban shopping center 0.056 0.0067
Central business district 0.101 0.0112
Pervious Surfaces
Single flunily residential
Low density (units/ha < 1.2) 0.012 0.0016
Medium density (units/ha @: 1.2) 0.022 0.0039
Townhouses and apartments 0.045 0.0078
High rise residential 0.012 0.0019
Institutional 0.012 0.0019
Industrial 0.012 0.0019
Suburban shopping center 0.012 0.0019
Central business district 0.012 0.0019
Source: Kuo et al., 1988.

3.7 Point Sources
The Watershed Model can account for nitrogen and phosphorus loadings from continuous point source
discharges. If there are known continuous point source discharges into the watershed, they should be
modeled to provide the most accurate picture of watershed nutrient loadings.

To enter nutrient loads into the Point Sources spreadsheet, click on the spreadsheet or use the <Tab>
key until the spreadsheet gains the focus. The monthly loadings should be entered in kilograms.

Selecting the [Example Data] button will provide point source data for the example watershed. To
provide reasonable output dam this data should be used in conjunction with the example data provided
for the other input parameters and the example weather file.

The [Cancel] button will clear all values from the point sources spreadsheet. If this button is selected,
a message box Will appear asking for verification prior to deleting current values.
3.8 Evapotranspiration Conditions
The parameters on the Evapotranspiration Conditions notebook page are used to determine the water
balance in the Watershed Model. Default values for all of the monthly ET parameters have been
provided. They are based on Virginia's geographic location and typical grow mig season.

Evapotranspiration Cover Coefficient (CV,): The amount of daily evapotranspiration is determined
by multiplying the potential ET by the cover coefficient. This parameter can be difficult to es@rimate,
but Haith et al. (1992) has developed a simplified procedure:
1 . Cover Coefficients should vary between 0 and 1, in principle.
2. Cover coefficients will approach their maximum value when plants have developed fiffl foliage.
3. Because evapotranspiration measures both transpiration and evaporation of soil water, the
lower limit for cover coefficients will be greater than zero. This lower limit essentially
represents a situation without any plant cover.
4. The protection of soil by impervious surfaces prevents evapotranspiration.
Cover coefficients for forests reach minimum values of 0.2 to 0.3 as leaf area indices approach zero.
Similarly, cover coefficients for farmland with annual crops can fall to 0.3 prior to planting and after
harvesting. Cover coefficients for perennial crops and conifers tend towards 1. 0. "11iis suggests that
monthly cover coefficients can be given the value 0.3 when fbliage is absent and 1.0 otherwise" (Haith
et al., 1992). For urban area , a constant value of 1.0 can be assigned to pervious surfaces and 0.0 to
impervious surfaces. The monthly values for the cover coefficient should reflect the area weighted
average for the various land use types in the watershed. Assuming that the amount of impervious
surfaces is small in relation to the watershed extent, default values have been selected based on
seasonal foliage development in Virginia. The values can be displayed by selecting the [Default]
button. If the watershed being modeled contains substantial urban development or unusual conditions,
the default cover coefficients might not be applicable.

Daylight Hours: This value represents the mean daylight hours for each month. This parameter is
fairly constant for most of Virginia. The mean daylight hours for 38' latitude are used as default
values (USEPA, 1984). Selecting the [Default] button will display these values in the
Evapotranspiration spreadsheet.

Growing Season: The monthly value is classified as either dormant or growing and is represented by
either a 0 or a 1, respectively. The growing season parameter determines the breakpoints between
antecedent moisture conditions. For months in the dormant season, AM 1 = 1. 3 cm and AM2 = 2. 8 cm.
For months in the growing seasons, AM I = 3.6 cm and AM2 = 5.3 cm. The growing season for
Virginia is considered to be April to October. Selecting the (Default] button will display the growing
season values 'in the Evapotranspiration spreadsheet.

Rainfall Erosivity Coefficient (a,): The rainfall erosivity coefficient is used to deterrruine rainfall
erosivity and contributes to the calculation of erosion by the USLE. 'ne coefficient varies with season
and aeoaraphic location. but can be estimated by using the methods developed by Seiker et al. ( 1990).
According to the rainfall erosivity zones defined by Wischmeler and Smith (1978), all of tidewater
Virginia is located in Zone 30. For this zone. Selker et al. has defined the cool season (October to

3 -2 2

March) value to be 0. 12 and the warm season (April to September) value to be 0.30. Selecting the
[Default] button Will display these values in the Evapotranspiration spreadsheet.

Selecting the (Example Data] button will provide evapotranspiration data for the example watershed.
To provide reasonable output data, this data should be used in conjunction with the example data
provided for the other input parameters and the example weather file.

The [Cancel] button will clear all values from the evapotranspiration spreadsheet- If this button is
selected, a message box will appear asking for verification prior to deleting current values.

3.9 Septic Systems
The per capita nutrient load must be specified for both nitrogen and phosphorus. The USEPA (1980)
reports loadings of 10.4 g/day of total nitrogen and 3.5 g/day of phosphorus. Loading values can also
be determined from representative septic tank wastewater flow and effluent concentrations. The
USEPA (1980) indicates that a typical on-site wastewater disposal system will discharge 170 L/day
per person. Mean nitrogen concentrations in septic tank effluent were measured as 73 mg1L, while
mean phosphorus concentrations were 14 mg/L when phosphate detergents were being used and only
7.9 mg/L when non-phosphate detergents were used (Alhajar et al., 1989). These values yield a per
capita septic tank effluent of 12.0 gtday for nitrogen, 2.5 g/day for phosphorus if phosphate detergent
is used, and 1.5 g/day if non-phosphate detergent is used.

The per capita growing season nutrient uptake in g/day must also be specified. This refers to the
uptake of nutrients from septic tank effluent by the ground cover (usually grasses) over the septic field.
There has not been much research in this area. Haith et al. (1992) provided estimates of 1.6 g/day for
nitrogen and 0.4 g/day for phosphorus, but these are speculative values. The most conservative
approach is to assume that plant uptake is minimal and enter a value of zero. If predicted nutrient
loadings from septic systems are within acceptable limits when nutrient uptake by plants is not
considered, then the septic systems are probably not contributing nonpoint nutrient loadings.

The values for these parameters may be entered by selecting the appropriate text box with the mouse or
using the <Tab> key until the desired text box gains the focus. As each text box gains the focus, it's
label will turn blue for easy identification.

The number of individuals served per month by the four different conditions of septic systems should
be entered into the Septic Systems spreadsheet. These four conditions are described in Section 3. 1.
'Me number of individuals served by each system can be obtained by perfbrming surveys or by
contacting local public health officials. To input data, click on the desired spreadsheet cell or use the
<Tab> key until the spreadsheet gains the focus, and then use the arrow keys to move around in the
spreadsheet.

Selecting the [Example Data] button will provide septic system data for the example watershed. To
provide reasonable output data, this data should be used 'in conjunction with the example data provided
for the other input parameters and the example weather file.

The [Cancel] button will clear all values from the septic system text boxes and spreadsheet. If this
button is selected, a message box will appear asking for verification prior to deleting current values,

3.10 Model Resuft
Simulation results from the Watershed Model can be examined as either tabular or graphical output.
To view the tabular results, click on the Tabular Output notebook tab or press <Alt T>. A compete
description of this notebook page is provided in subsection 3.10.1. To view the results in a graphical
format, click on the Graphical Output notebook tab or press <Alt G>. A complete description of this
notebook page is provided in subsection 3.10.2. Selecting the Graphical Output notebook page will
run the Watershed Model. If simulation results have not been previously calculated or if any of the
input parameters have been altered since the simulation results were generated, the Watershed Model
will begin a new'simulation. The time required to complete the simulation will depend on the number
of years being modeled and the processing power of your computer. The mouse pointer will change to
an hourglass while the simulation is being processed.

3.10.1 Tabular Output
The simulation results can be viewed in the following five ways:
I . Summary by Month
2. Summary by Source Oand use category)
3. Annual Results
4. Annual Results by Source (land use category)
5. Monthly Results
Selecting Summary by Month win provide the average monthly values for precipitation, ET,
groundwater flow, stream flow, runoff@ erosion, sediment yield, dissolved and total nitrogen, and
dissolved and total phosphorus. These are the average values for the entire simulation period. The
average annual value for each parameter is also provided in this summary.

Selecting Summary by Source will provide average parameter values for each land use category.
Results include runoff, erosion per hectare, dissolved and total nitrogen, and dissolved and total
phosphorus. The area of each land use is also provided. In addition to the designated land uses,
dissolved nutrient loadings are also provide for the groundwater discharge, as well as point sources and
septic systems if either has been modeled.

Annual Results provides the average annual values for precipitation, ET, groundwater flow, stream
flow, runoff, erosion, sediment yield, dissolved and total nitrogen, and dissolved and total phosphorus
for each year in the simulation period.

Annual Results by Source provides the same information as Summary by Source, but instead of values
averaged over the entire simulation period, the values are presented for each year of the simulation.
For large simulation periods, output for the entire simulation may not be desired. To give the user the
flexibility to view only selected output years, an input dialog box appears when this display option is
selected. Enter the desired beginning and ending years of the output you wish to view and then [Close]
the input box.

Selecting Monthly Results provides the same information as Summary by Month, but instead of
monthly values averaged over the entire simulation period, the monthly values are presented for each
year of the simulation. For large simulation periods, output for the entire simulation may not be
desired. To ive the user the flexibility to view only selected output Years. an input dialog box appears
91

when this display option is selected. Enter the desired beginning and ending years of the output you
wish to view and then [Close] the input box.

The default option is Summary by Month. To select a different option, use the mouse or press the
<Tab> key. After the desired display option has been selected, click on (Show Output]. If simulation
results have not been previously calculated or if any of the input parameters have been altered since the
simulation results were generated, the Watershed Model will begin a new simulation. Ile time
required to complete the simulation will depend on the number of years being modeled and the
processing power of your computer. The mouse pointer will change to an hourglass while the
simulation is being processed. 'Me results will be displayed in a noninteractive spreadsheet,

Any of the simulation results spreadsheets can be printed by selecting File/Print from the menu bar. A
Print dialog box will appear with check boxes fbr all the output options. Select as many output options
as desired by using the mouse or by pressing <Alt> and the underscored letter in the option name.
Once all desired options have been selected, choose the [OK@ command button. This will send the
selected files to the designated windows printer. Tle output spreadsheets will print in color if a color
printer is designated. Select the [Cancel] button when the print option is no longer desired.

3.10.2 Graphical Output
Tliere are four types of graphs:
1. Annual Mean by Source (land use category)
2. Monthly Mean
3. Annual
4. Monthly
When you select the desired graph option, a parameter option box will appear with the applicable
parameter types. When a parameter type is selected, the corresponding graph will be displayed. This
graph can be sent to the printer by clicking on the [Print Graph] button. Annual Mean by Source and
Monthly Mean graphs are bar graphs. Annual and Monthly graphs are line graphs. When the Monthly
graph option is selected, an input dialog box will appear requesting the yearly range over which to
graph the data. Since data for all months is displayed for each specified year, the Monthly graph
option is more applicable for viewing simulations with short time periods or smaller portions of
simulations with large time periods.

4. Marina Water Quality Model
'Me Mmina, Water Quality (MWQ) Model is an analytical model for steady-state, two-dimensional
contaminant transport from a continuous shoreline source. The model is based on the work of Hamrick
and Neilson (1989). A marina was designated as the continuous shoreline contaminant source. The
model is applicable to marinas located along theL shoreline of a wide channel. Hamrick and Neilson
define a wide channel as being typically wider than I 00m and having "measurable net fi=h water
discharge in addition to tidal driven flow". The MWQ Model considers advection, dispersion, and first
order decay.

Two analytical solutions are included in the MWQ Model: an infinite channel solution and a finite
channel solution. T"hese solutions are designated, respectively, by selecting either the neglect channel
end effects or the consider channel end effects option on the Options notebook page. Both solutions
assume constant water depth, decay coefficient, longitudinal and trarisverse dispersion coefficients, and
longitudinal velocity. The solutions are based on depth and tidal cycle average conditions. Dierefore,
the contaminant concentration is constant with depth. In addition, the MWQ Model assumes that the
tidal range is significantly less than the water depth. TIle contaminant is assumed to be uniformly
mixed over the water depth at time T,
T_ = 120 h
qM
where, h is the mean water depth and q. is the maximum tidal velocity magnitude (Hamrick and
Neilson, 1989). If Tz is less than or equal to the inverse of the decay coefficient and to the semi-
diurnal tidal period, than the vertical uniformity condition is satisfied.

The infinite channel solution is based on the assumption that any contaminant will decay before it
reaches the boundaries. Advective transport is limited to the longitudinal direction. This solution is
C = M - exp[ U__7 K.[
-7hVY _K, D
,D, 14 -0

where, C = depth and tidal cycle averaged contaminant concentration
M = mass of the contaminant
D,, = longitudinal dispersion
Dy = transverse dispersion
u = tidally averaged mean discharge velocity
K,i = the decay coefficient
x = the current value along the channel
y = the current value across the channel
Y, = the modified Bessel function of the second kind of order zero
B = channel width

The finite channel solution is for a stream with a no flow (closed) upstream boundary and a
downstream boundary, open to another water body. The downstream boundary is infinitely diluted.
Because of the no flow upstream boundary, longitudinal advective transport is assumed to be
negligible. This solution is

C=- M .0 00 (- ly K. +@&()@,B)2 +IYK. X+2j4)2+@&(Y+2iB)2
- Z Z (x+21, +jr,)'
*JD@y (i-j- @ @D. Dy FD. DY

where, L, is the upstream boundary location and Ld is the downstream boundary location.

The major weakness of the MWQ Model is the use of idealized stream geometry, topography, and flow
fields. This allows for the use of analytical solutions, but site-specific data cannot be used, even if it is
known.

4.1 Model Format
The Mum Water Quality Model utilizes a notebook format to enter data and display model results.
'Me first three notebook pages provide screens fbr input data and the last two notebook pages are used
to present tabular and graphical model results. Each page of the notebook can be accessed by clicking
on the corresponding tab or pressing <Alt> plus the underlined letter of the tab header. Discussions
regarding model input and parameter estimation are provided in Sections 4.2 and 4.3. Both of these
sections corresponds to an individual notebook page in the MWQ Model. Simulation results and model
output are discussed in Section 4.4.

The fir-st notebook page, Project Info, serves as the MWQ Model's introductory screen. This notebook
page can be used for record keeping purposes. There are text boxes available to enter in the project
title, date, and the names of the individuals involved with the model run. This notebook page may be
omitted if desired.

On the Parameters notebook page, there is an example data set. This data set can be used as a tutorial
for the MWQ Model. To access this data click on the [Test Data Set] button. This data can be used to
become fluniliar with "how to navigate through the model and it provides an example of typical input
and model results. There is data available to rneet all model options.

The menu bar on the MWQ Model provides the user with file access, editing capability, and help
information. Selecting File will display a drop-down menu list with the following items: New, Open,
Save, Save As, Print, and Exit. New deletes the current data set. A message box will ask for
verification prior to deleting current values. Open calls up the Open File dialog box to select an
existing data set. Save Will save the current data set using the existing file name. If a file name has not
been defined, the Save As File dialog box will appear prompting the user to enter a file name. Save As
opens the Save As File dialog box to allow the current data set to be saved with a new file name.
Selecting Print will send the current model data and results to the default Windows printer. Exit closes
the MWQ Model and returns the user to the main CS Model window. If the current data set has not
been saved, it will be lost. Selecting Edit will display a drop-down menu fist with Cut, Copy, and
Paste. These menu items can be used to edit input data. Cut will remove the selected text and move it
to the windows clipboard. Copy will place a duplicate of the selected text on the clipboard, but not
remove the oniginal text. Paste can be used to put text placed on the clipboard back onto one of the
input screens. Help will provide information pertaining to the MiWQ Model.

4-2

4.2 Parameters
'Me hydraulic and contaminant parameters and the region to be modeled must de defined- Ile values
for these parameters can be entered directly into each text box by clicking on it and then entering the
desired value from the keyboard. Ile text boxes may also be accessed using the <Tab> key. As each
text box becomes active, its corresponding label will turn blue. Clicking on the label fbr the various
input parameters will display a pop-up message box with help information.

The (Cancel] button wiH clear aU values from the parameter text boxes. If this button is selected, a
message box wifl appear asking for verification prior to deleting current values.

4.2.1 Hydraulic Parameters
Mean Water Depth: T"his is the water depth averaged over the area to be modeled. The depth is
assumed to be constant.

Maxiinum Tidal Velocity: The magnitude of maximum tidal velocity. Units are meters per second.
Mean Discharge Velocity: Tidally averaged mean discharge velocity. A conservative estimate can
be obtained by setting this value to zero.
Channel Width: The width of the channel is assumed to be constant over the region being modeled.
This value should be entered in meters.

Tidal Period: 'Me tidal period is usually 12.4 hours.
Distance Downstream to Open End: This is the distance downstream to the open end of the channel,
and it must be greater than or equal to the downstream distance to be modeled.
Distance Upstream to Closed End: This is the distance upstream to the closed end of the channel,
and it must be gre:ater than or equal to. the upstream distance to be modeled.

4.2.2 Contaminant Parameters
The MWQ Model will predict results for coliforms, CBOD, and NBOD. Subsection 2.3 ).5 contains a
detailed discussion of these parameters. Check boxes allow for the selection of the contaminants to
model. More thari one contaminant can be modeled simultaneously. Both a loading and a first order
decay rate must be specified for each parameter that will be modeled. Loadings for all parameters are
considered to be continuous and steady-state, and the contaminant is introduced into the water body at
a point along the shoreline. Required units for loading of CBOD and NBOD are kg/day. Coliform
loadings are in organisms per second. Decay rates for all parameters should be per day.

4.2.3 Region to be Modeled
Distance DownstreArn: Distance downstrearn from the contaminant source being modeled.
Distance Upstream: Distance upstream from the contaminant source being modeled.

Width Across Channel: The distance across the channel to be modeled. This distance must be less
than or equal to the actual channel width.
Display Length: This value will control the interval used to display the distance aiong the channel
the tabular display of the MWQ Model's results on the Tabular Output notebook page. Units are Uil

meters. A value can be entered directly into the text box or the spin buttons can be used to select the
desired interval.

Display Width: This value will control the interval used to display the distance across the channel in
the tabular display of the MWQ Model's results on the Tabular Output notebook page. Units are in
meters. A value can be entered directly irito the text box or the spin buttons cm be used to select the
desired interval.

4.3 Options
Selecting the Options notebook page will provide a preliminary assessment of the initial data. 'Me
longitudinal and transverse dispersion coefficients are calculated and the model assumptions are
evaluated. If you do not want to use the calculated dispersion coefficients, select the check box and
enter the desired values in the user selected text boxes. Model assumptions are evaluated based on the
chosen selection option to consider or neglect channel end effects. If the displayed values are much
greater than 1, thari the given assumption is valid. If any of the assumptions are not valid, different
parameter values should be entered prior to starting a simulation.

4.4 Model Results
Selecting the Tabular Output notebook page will run the MWQ Model. The mouse pointer will change
to an hourglass while the simulation is being processed. Results from the MWQ Model can be
exan-dned as either concentrations plots or 'in a tabular format. To run the model and view the tabular
results, click on the Tabular Output notebook tab or press <Alt T>. A compete description of this
notebook page is provided in Subsection 4.4. 1. To view the results in a graphical format, click on the
Transects notebook tab. A complete description of this notebook page is provided in Subsection 4.4.2.

4.4.1 Tabular Output
"Me MWQ Model displays the results in a spreadsheet format. Ile distance along the channel, as
specified on the Parameter notebook page, is displayed 'in the first column. The values are for the
specified interval in both the upstream and downstream direction. The first row of the spreadsheet
displays the distance across the channel, as specified on the Parameter notebook page. The rest of the
matrix contains the values for the modeled water quality variable. There is a set of options buttons to
the right of the spreadsheet that can be used to change which water quality parameter is being
displayed. All values are displayed in exponential notation.

4.4.2 Transects
For each of the water quality variables, the longitudinal concentration can viewed for the transects
across the channel. The distance upstream and downstream from the discharge point is set on the x-
axis and the parameter concentration is placed along the v-axis. To change which transect is being
displayed, click on one of the scroll arrows. 'Me new transect value will be displayedin. the text box.
The transect value will correspond to the display width value selected on the Parameter notebook page.
Once the desired timisect value is selected, click on the [Redraw] button to refresh the graphical
display.

5. Tidal Prism Model
The Tidal Prism Model was developed by Jason Luettinger as a part of his master's degree research.
The model description contained in this section is provided from his thesis with his permission.

The Tidal Prism Model is a one-dimensi6nal tidal flushing model capable of predicting the longitudinal
distribution of contaminant concentrations at high tide intervals. The algorithm was first proposed by
Ketchum in 1951 and later modified by Kuo and Neilson in 1988. The model differs from traditional
finite segment models in that the segmentation of a water body is based upon the geometry and
hydrodynamic flow characteristics of the water body rather than a manual segmentation process.
Instead of solving a large set of simultaneous equations as is necessary for typical finite segment
models, the Tidal Prism Model successively solves one mass balance equation progressively from the
mouth to the upstream boundary segment for each tidal cycle increment. Concentrations are initialized
after each tidal cycle such that the model uses the previous tidal cycle concentrations as the initial
conditions for the next time step. This iterative process continues until a specified time period has
elapsed or the system reaches a steady-state condition.

Ketchum's original Tidal Prism Model was proven to be fiindamentally sound in theory, but some
minor flaws were shown to exist in later revisions of the model. Ketchum's model based the
segmentation of a water body on the fact that the upstream river flow was a non-zero parameter.
Segmentation was proposed to begin at the upstream limit of tidal current reversal and continue
downstream to the mouth of the water body. Because a non-zero flow was required in Ketchum's
proposed model, it was not applicable to tidal embayments or dammed tidal rivers were an inflow may
not always exist. Ketchum demonstrated that the Tidal Prism Model could accurately predict salinity
distributions in three very different estuaries including the Raritan River and Bay, Alberni Inlet- and
Great Pond (Ketchum, 195 1).

Three major revisions have been proposed to Ketchum's original model. Dver and Taylor (1973) first
corrected a fundamental error in the model's mass balance equation and also proposed a fitting
parameter associated with the mixing. This unnamed parameter is referred to 'in later articles as the
44return flow factor" or the "returning ratio", and was identified as a relatively important parameter in
fitting the model (Sanford et al., 1992). Wood (1978) fiirther expanded this concept by proposing an
alternative to the "inter-segment exchange" in terms of an average dispersion coefficient. This
dispersion coefficient was included to account for the relative mixing which occurs between the
neighboring segments of a water body. Wood's proposed model was calibrated to Ketchum's data on
the Raritan Estuary for comparison of fit. It was shown that there was no single dispersion index
which could provide an accurate representation of the salinity profile, but that a combination of low
dispersion near the freshwater source and high dispersion near the seaward boundary produced a fairly
accurate fit.

Wood's model combined the basic ideas proposed by Ketchum (195 1) and Dyer and Taylor (1973),
but still required that the flow be a non-zero parameter for segmentation of the water body. Kuo and
Neilson (1988) later expanded the model such that it became applicable to cases where the water body
is branched and/or the freshwater discharge is negligibly small. Kuo and Neilson proposed a
flindamental change in the segmentation of the water body 'in order to allow for the zero flow case. In
this version of the model, the segmentation begins at the mouth of the water body and continues
upstream until a cutoff point is reached. This cutoff may be a physical barrier such as a dam or
embayment, or a hydraulic barrier such as the point where current reversal stops and pure advecuive

flow begins in a tidal river. Kuo and Neilson's final modification of the Tidal Prism Model is capable
of simulating point and non-point discharges and can handle both conservative and non-conservative
substances. The application of the model is relatively simple as it requires a minimal amount of
physical data including depth, surface area and tidal range.

Ile Tidal Prism Model has been demonstrated to be successful in predicting water quality in many
small coastal embayments in Virginia, including the Lynnhaven Bay system on the lower Chesapeake
Bay (Kuo and Neilson, 1988).

5.1 Modeling Approach
As the surface of the river rises during flood tide, this increase in water volume must be accounted for
from some source. In a tidal river, this increase in volume is accounted for from both the landward
river flow and the seaward flooding tide. This increase in water volume is referred to as the "tidal
prism" of the river. For every given flow volume in a river, there exists an imaginary boundary in the
river where the entire volume change dunng the rise in tide can be equated to the volume of ri@er flow
during that same time period. In other words, there exists a boundary where the prism volume is
exactly equal to the river flow (P(x) = R(x)). The entire river landward of this boundary, where P(x)
R(x), can be considered a purely advective river because there is no flow reversal, and therefore no
influence from the flooding waters from downstream. Ile portion of the river downstream of this
boundary experiences flow reversal and tidal mixing. This part of the river is truly under tidal
influence, and therefore experiences much different hydraulic mixing conditions. The Tidal Prism
Model was developed to model this portion of the river. Note that when there is negligible flow, the
entire river is influenced by this type of tidal mixing.

The feature that makes the Tidal Prism Model unique when compared to other finite section tidal
models is in the process by which the water body is segmented. During flood tide, large amounts of sea
water pass through the mouth of the water body and nuix with the inland fresh water. Because this
water mixes with the fi-eshwater during high tide, a portion of the contaminant is flushed out during ebb
tide. This process is known as "tidal flushing". Rather than hypothetically dividing the estuary into
finite segments, the tidal prism model assures complete mixing within each segment throughout a tidal
cycle by segmenting the water body according to its physical characteristics, i.e. tidal oscillation,
upstream flow and topography.

Segmentation begins at the mouth of the water body. The segment length is defined as the distance a
particle of water may travel upstream during one flood tide. Longitudinal segmentation is achieved by
continuing upstream from the mouth placing segment U-ansects at lengths according to the flooding
distance. Because the segment length is always equal to or less than the distance water will travel
during a flooding cycle, complete mixing is assured throughout the water body. The Tidal Prism
Model is a mathematical simulation of this tidal flushing process. 'Me model is capable of predicting
the longitudinal distribution of contaminants within a body of water and is therefore well suited for
long coastal embayments or tidal rivers (Kuo and Neilson, 1988). The Tidal Prism Model is
considered a dynamic model because it is able to predict the distribution of contammants at any given
point *in time, and is therefore not restricted to the steady-state condition.

The Tidal Prism Mass Balance Equation

The Tidal Prism Model mass balance equation is based upon the exchange of mass between segments
over an entire ebb to flood tide cycle. This equation is soived from the mouth (segment 1) toward the

j _'@

landward segments, successively solving for the high tide concentration in each segment. Ile equation
requires only the upstream and downstream boundary concentrations and the initial concentrations
existing in the water body. The equation begins at cycle 1, using the initial concentration fields that are
entered for the first time period, and solves successively upstr=m for the final concentrations in each
segment. These calculated concentrations in each segment then become the new starting point for the
second cycle and the process continues. ne cycling process continues at intervals equal to one tidal
cycle until the desired period of time has elapsed or a steady-state condition is reached. Figure 5-1 is
an elevation view of a hypothetical water body. This figure illustrates the exchanges which take place
between each segment. During flood tide, a volume of water equal to the prism volume minus the river
flow (P(n)-R(n)) moves into the landward segments from the downstream segments. Because the
model assumes that complete mixing takes place within each segment, this water that has flooded from
the downstream segments mixes completely with the water in the adjacent landward segments. During
ebb tide, a volume of water equal to the prism volume plus the river flow (P(n)+R(n)) moves seaward
from each segment. This exchange of volume between the ebb and flood tides creates what is re&rred
to as tidal flushing. Mass is slowly transported out of the water body by the constant dilutioq of the
sea water which floods into the coastal embayments or tidal rivers, mixes with the ambient
"contarninated" water, and is once again removed during ebb tide.

The returning ratio or "alpha" is incorporated into the flood tide transport portion of the mass balance
equation. This parameter accounts for the volume of water which moves out of a segment during ebb
tide and then returns at the same concentration in the following flood tide. Alpha therefore represents
the relative mixing which occurs between segments. Low values of alpha signify less water exchange
and therefore less mixing and vise versa. Sanfbrd et al. (1992) discuss *in detail the different
hydrodynamic properties of a water body which influence the value of this returning ratio at the
seaward boundary. In general, it was shown that the differences in current direction and magnitude at
the intersection of the embayment with the receiving water will influence the amount of contaminant
that is "returned" to the embayment upon flood tide.

5.2 Model Format
The Tidal Prism Model utilizes a notebook format to enter data and display model results. The first
five notebook pages provide screens for 'input data and the last notebook page is used to present tabular
and graphical model results. Each page of the notebook can be accessed by clicking on the
corresponding tab or pressing <Alt> plus the underlined letter of the tab header. Comprehensive
discussions regarding model input and parameter estimation are provided in Section 5.3 to Section 5.6.
Each of these sections corresponds to an individual notebook page *in the Tidal Prism Model.
Simulation results and model output are discussed in Section 5.7.

For each notebook page used to enter data, there is an example data set. This data set can be used as a
tutorial for the Tidal Prism Model. To access this data click on the [Example Data] buttons on each
notebook page. There is data available to meet all model options. These data can be used to become
fluniliar with how to navigate through the model and it provides an example of typical input and model
results.

The first notebook page, Project Info. serves as the Tidal Prism Model's 'introductory screen. This
notebook page can be used for record keeping purposes. There are text boxes available to enter in the
project title. date, and the names of the individuals involved with the model run. This notebook page
may be omitted if desired.

5 - 3

P(N) R(N) n+1 n-I MHW 2
r I
I p(n) I MLW

F(D) + R(n) Op P(n-l)+R(n-1)
V(N) I I V(D) V(2) V(I)
P(n) - R(n) 41 P(B- 1) - FL(n- 1)

Figure 5-1. Elevation view of a hypothetical river illustrating the volumes of water excifanges
between adjacent segments during a complete tidal cycle.

The menu bar on the Tidal Prism Model provides the user with file access, editing capability, and help
information. Selecting File Will display a drop-down menu list with the following items: New, Open,
Save, Save As, Print, and Exit. New deletes the current dam set. A message box will ask for
verification prior to deleting current values. Open calls up the Open File dialog box to select an
existing data set. Save will save the current data set using the existing file name. If a file name has not
been defined, the Save As File dialog box will appear prompting the user to enter a file name. Save As
opens the Save As File dialog box toallow the current data set to be saved with a new file name.
Selecting Print will send the current model data and results, including the graph to the default
Windows printer. Exit closes the Tidal Prism Model and returns the user to the main CS Model
window. If the current data set has not been saved, it will be lost. Selecting Edit will display a drop-
down menu fist with Cut, Copy, and Paste. These menu items can be used to edit input data. Cut will
remove the selected text and move it to the windows clipboard. Copy will place a duplicate ofthe
selected text on the clipboard, but not remove the original text. Paste can be used to put text placed on
the clipboard back onto one of the input screens. Help will provide information pertaining to the Tidal
Prism Model. .

5.3 Parameters
The Parameters notebook page prompts the user to enter a series of stream parameters necessary to run
the Tidal Prism Model. The screen is divided into two portions: (1) the upper portion, which consists
of a series of text boxes that are used for numeric input data entered from the keyboard, and (2) the
bottom portion, which consists oftwo radio option buttons that allow the user to choose the duration of
time over which stream concentrations will be calculated.

The following numeric stream parameters must be entered in the upper text box portion of this
notebook page.

Number of Stream Divisions: In order to simulate the changing cross-sections of a body of water, the
Tidal Prism Model allows the user to divide the water body M*to a number of divisions. Each division
will consist of a 2iven len2th of the water body with an approximately constant cross-section and flow

volume. This text box prompts the user to enter the number of divisions in the water body being
studied.

Upstream River Flow: The upstream river flow parameter is considered the flow volume in m/day
which enters the UPSErcarn boundary segment of the water body.

Ambient River Concentration: The background concentration of the contaminant of interest present
in the upstream river flow is considered the ambient river concentration. The user is prompted to enter
this concentration in mg/L.

Mouth Concentration: The model requires that a mouth boundary concentration be entered at the
downstream limit of the water body. For calculation purposes, this concentration will remain constant
with time. The user should include stream divisions far enough downstream to allow for this constant
boundary condition.

Delta X: Ile delta X parameter refers to the calculation interval that is used for segmenting the water
body. The model will iterate upstream from the mouth at intervals equal to delta X in this calculation.
Small delta X values correspond to a more accurate segmentation of the water body, but also require
more computer time for the calculation. A delta X of I meter is recommended.

Minimum Segment Length: The minimum segment length is the rruiru*murn distance between segment
transects; that is allowed before the segmentation routine ceases. Because the segment lengths decrease
as one moves towards the upstream boundary of a water body, the upstrearri limit of the model's
segmentation will directly depend upon the minimum segment length in most cases.

The bottom portion of the parameters screen consists of the Desired Stream Concentration panel. In
this panel the user is given the choice of having the model compute concentrations after a finite number
of tidal cycles or at a steady-state condition. A steady-state condition is assumed to be reached when
the segment concentrations change less than 0. 000 1 percent between tidal cycles. The steady-state
option will usually require slightly more computer time due to the increased number of tidal cycles
which must be run through before this condition is reached. If the alternative option is chosen, the user
may enter the desired number of tidal cycles with the spin button at the bottom of the screen. The Tidal
Prism Model will calculate final concentrations after any number of tidal cycles have passed. This
option is useful when the user desires information about the time varying changes which may occur in a
stream after a spill or a slug discharge of finite duration has occurred. In this case, the number of tidal
cycles can be increased at a set increment between runs and the dispersion of the plume can be
visualized over time.

5.4 Stream Geometry
This screen allows the user to enter the stream geometry cross-section and flow data for each division
of the water body. Six variables are required to describe the cross-section of a stream, which is
estimated as a trapezoidal polygon. These include the followmig:

Slope I and Slope 2: These two slope variables refer to the "run to rise" ratio of the stream bank. In
other words, the side slopes are the ratio of the distance traveled horizontally divided by the distance
traveled vertically from the water surface to the bottom the channel. These variables are entered as
single. decimal values.

Base Width: The base width is the width across the "bottom" of the channel.

High Depth: ne high depth is the mean depth of the stream division at peak high tide.
Low Depth: The low depth is the mean depth of the stream division at low tide.
Length: The length refers to the length of the individual stream division.

In addition to the stream geometry parameters, the user may also enter a value for the additional flow
into a division. This additional flow is assumed to begin at the upstream limit of the division and is
added to the total upstream flow 'in the river. This variable was incorporated into the model to account
for a tributary flow or a significant discharge into the water body. A significant flow is considered to
be one in which the river flow is significantly increased downstream of the point. When the river flow
is significantly increased, this change in flow must be included in the calculation of the segment
transect locations for an accurate representation of the real world condition.

5.5 Segmentation
Ile segmentation notebook page consists of a noninteractive spreadsheet display of the Tidal Prism
Model's segmentation of the water body. The segmentation spreadsheet displays seven parameters
associated with each segment. These seven parameters 'include the following:

Segment Number: Numbering begins at the mouth and continues upstream.
Location: This value is the distance from the mouth (considered to be at 0 meters) to the
upstream transect of the segment.
Length: This is the length between the upstream and downstream transects of a segment.
Low Volume: This is the volume of the segment at low tide.
High Volume: This is the volume of the segment at high tide.
Prism Volume: The prism volume of a segment is the total volume of the tidal prism upstream of
a segment.
River Flow: The river flow is the total volume of water entering a segment from upstream.
5.6 Loading Data
The user may enter all relevant discharge loading parameters in the Loading Data notebook page.
Mass and volume data are entered into the Discharge Parameters spreadsheet. A loading source is
represented by both the volume per day that is discharged and the mass of the contaminant that is
carried within that volume. The user may enter a discharge volume (m'/day) and the mass of
contaminant discharged (gra/day) at any given point in the water body. Segment numbers and locations
are given in the first two columns to aid the user *in locating the point of discharge.

At the bottom of the screen, two panels have been included which define the loading characteristics and
the value of the returning ratio to be used in the calculation of final concentrations. The user may
define any given type of contaminant discharge. For instarice, a continuous discharge may be chosen
for a treatment plant flow or an industrial waste in which discharge is constant over time. For
circumstances where flow is not constant over time, such as runoff events or a one time slug discharge,
the finite discharge option button may be chosen. With this option. the user may indicate the duration

5 -6

of the discharge using the spin button located at the bottom of the screen. For example, a spill of some
contaminant lasting two days would be represented by a finite discharge of a four tidal cycle duration.

The returning ratio can be defined as a constant value throughout the water body, or may be varied
linearly from the upstream boundary to the mouth. The linear gradient alpha value is represented by
the ratio of the total number of segments minus the segment number divided by the total number of
segments. This simple linear formula will vary alpha from zero at the upstream boundary segment
(total number of segments = segment number), to a value close to one at the first segment (segment
number two). For situations where the user desires to define alpha as a constant value, the spin button
at the bottom of the panel may be used to increment alpha between zero and one.

5.7 Table Results
The Table Results notebook page is divided into two halves. On the left hand side of the screen is the
noninteractive concentration output spreadsheet. This spreadsheet displays the tabular results of the
concentration routine. Concentrations are given in mg/L for each of the water body's segmeiits. On
the right half of the screen is the graphical output of the concentration data. This graph consists of a
plot of the segment concentrations versus their corresponding distance from the mouth. For each model
run, the concentration at half of the total cycles are plotted in addition to the final concentration in each
segment. Ile plot of the concentr 'ations at half of the total cycles was included to give the user an idea
of the general trend that the concentrations were moving in. For example, in an instantaneous spill one
would expect that the contaminant plume would disperse over time due to the dilution effects in the
water body. In this case, the concentration at half of the total cycles would be higher dm the final
concentration in each segment. Alternatively, for cases where the discharge is continuous, the final
concentrations would be higher 'in each segment due to the constant input of the contaminant. This
difference would continue until a steady-state condition is reached at which time the two plots would
overlap. exactly. The graphical output therefore gives the user information about both the final
concentrations in each segment and the general trend of the concentrations in the system.

5-7

6. Finite Section Model
The Finite Section Model (FINSEC) is a one-dimensional, steady-state water quality model that can be
used to dezerm@ne the concentrations of water quality variables in a river. It is based on dividing the
river or estuary into a number of approximately homogeneous regions or sections where the
concentration gradient is not significant within each reach. A mass balance is constructed around each
section for each water quality variable to be modeled. The model does not consider tidal effects.

6.1 Modeling Approach
The river is divided into segments of varying lengths, where the segment length is selected based on the
principle concentration gradients of the substance being modeled. If segments of equal length are
applied, the area of each segment will typically increase as the river be ins to widen near the mouth or
9
downstream boundary. The average river depth would also typically increase. An assumption in the
FINSEC Model is that velocity gradients and dispersion will provide complete mixing both literally
(across the river) and vertically such that the concentrations gradients are only along the axis of the
river (T'homann and Mueller, 1987). Figure 3-2 illustrates a possible segmentation of a river reach.
Segments are numbered from I to n.

The concentration of substance S in any segment i is calculated using a basic mass balance equation.
Thomann and Mueller (1987) have outlined the following components of the mass balance equation:
1. transport of S due to advective flow
2. mass transport due to tidal dispersion and density nuixing
3. loss of mass due to decay, and
4. any external sources or sinks of S.
This concept is graphically displayed in Figure 6-2. The subscript i refers to the value in segment i.
The subscripts i- 1, i and i, i+ I refer to the value at the interface between segment i and the upstream
and downstream segments, respectively. The FINSEC Model solves for the concentrations of
substance S at these interfaces. There are several different methods that can be applied to solve the

waste
Tributary Loading

Upstream
Inflow Ocean or
1 2 1 3 Bay

Trib
UP

I I f I I I I I

Figure 6-1. Finite Section Conceptualization.

Segment i

Advective Advective
Flow, QI-1, I Volume, Vi Flow, QI + 1, I
[St]
Decay, Kt
Dispersion, EI-t, I Dispersion, EI + 1, I

Sources or
Sinks, Wi
Figure 6-2. Components of mass balance equation, adapted from Thomann and Mueller
(1987).

interface concentration: central, backward, or length dependent differencing options. These options are
discussed in greater detail in subsection 6.3.3.

Applying a backward differencing option, the numerical approximation of the mass balance equation
can be written as
dSi +
Vi __ = Qi-1,iSi-1 - Qi,i+1Si + Ei-1,i (Si-1 - Si) + Ei+1,i (Si+1 - Si) - KiViSi _ Wi

dt
The FINSEC Model utilizes the steady-state condition, which assumes that all inputs, flow, exchanges,
and reaction rates are constant over time. This reduces the numerical mass balance equation to a
simple linear algebraic equation because Vi dSi There will an algebraic equation for each
___ = 0.

dt
defined segment. This set of simultaneous algebraic equations is solved using a matrix inversion.

6.2 Model Format
The FINSEC Model utilizes a notebook format to enter data and display model results. Ile first three
notebook pages provide screens for input data and the last two notebook pages are used to present
tabular and graphical model results, respectively. Each page of the notebook can be accessed by
clicking on the corresponding tab or pressing <Alt> plus the underlined letter of the tab header.
Discussions regarding model input and parameter estimation are provided in Sections 6.3 and 6.4.
Both of these sections corresponds to an individual notebook page in the FINSEC Model. Simulation
results and model output are discussed in Sections 6.5 and 6.6.

The first notebook page, Project Info, serves as the FINSEC Model's introductory screen. This
notebook page can be used for record keeping purposes. There are text boxes available to enter in the
project title, date, and the names of the individuals involved With the model run. This notebook page
may be omitted if desired.

6-2

On the Input Data notebook page, there is an example data set. This data set can be used as a tutorial
for the FINSEC Model. To access this data click on the [Example Data] button. This data can be
used to become familiar with how to navigate through the model and it provides an example of typical
input and model results. There is data available to meet all model options.

The menu bar on the Finite Section Model provides the user with file access, editing capability, and
help information. Selecting File will display a drop-down menu list with the following items: New,
Open, Save, Save As, Print, and Exit. New deletes the current data set. A message box will ask for
verification prior to deleting current values. Open calls up the Open File dialog box to select an
existing data set. Save will save the current data set us mig the existing file name. If a file name has not
been defined, the Save As File dialog box will appear prompting the user to enter a file name. Save As
opens the Save As File dialog box to allow the current data set to be saved with a new file name.
Selecting Print will send the current model data and results to the default Windows printer. Note that
graphs are not printed from this menu item. The graphs are printed directly from the Graph notebook
page using the [Print Graph] command button. Exit closes the FINSEC Model and returns the user to
the main CS Model window. If the current data set has not been saved, it will be lost. Selecting Edit
will display a drop-down menu list with Cut, Copy, and Paste. These menu items can be used to edit
input data. Cut will remove the selected text and move it to the windows clipboard. Copy will place a
duplicate of the selected text on the clipboard, but not remove the original text. Paste can be used to
put text placed on the clipboard back onto one of the input screens. Help will provide information
pertaining to the FINSEC Model.

6.3 Options
The Options notebook page is used to set the basic options for the model. These are the water quality
variables, boundary conditions, and differencing options. Each of these options is explained in the
following subsections.

6.3.1 Water Quality Variables
The concentrations of the water quality variables *in the river are what the FINSEC Model is being used
to determine. The model will currently estimate concentrations for five water quality variables. A
brief description of each variable is provided below, but a more detailed description is available in
Subsection 2.3.5 The default variable is salinity. Any or all of these variables may be selected for a
model simulation.

Carbonaceous Biochemical Oxygen Demand (CBOD): CBOD is an 'indicator of organic pollution
measured in terms of the oxygen demand that can develop as the organics are degraded. Units are
mg/L.

Nitrogenous Oxygen Demand (NBOD): NBOD is the equivalent measure of the organic nitrogen
and ammonia that will consume oxygen as they are converted to nitnite and nitrate. Units are mg/L.

Dissolved Oxygen (DO): A direct measure of the amount of oxv en dissolved within the water
el .9
column. Units are mg/L.

Coliforms: Coliforms are bacten'aL'ri the Enterobacteriaceae family and are commoniv used as an
indicator of fecal contamination. Units are organisms/L.

6-3

Salinity: Salinity is commonly used to calibrate a finite section model for dispersion. It is treated as a
conservative material. Usually, the only loads for salinity arise at the mouth of the estuary or river.

6.3.2 Boundary Conditions
The FINSEC Model simulates a river or a portion of a river reach. The simulated area will have
boundaries on both the upstream and downstream ends. As with all models based on differential
equations, boundary conditions must be applied to obtain a solution. There are two type of boundary
conditions available in the FINSEC Model: fixed or linear gradient.

A fixed boundary condition requires a known concentration at the boundary. This concentration does
not vary regardless of any changes in loading or kinetics that occur within the modeled reaches. If this
option is selected, you will need to specify the boundary concentration for section 0 in the predicted
column of the Input Data spreadsheet.

A linear gradient boundary condition assumes that the trend in the two adjacent intenior sections, is
linear and can be extrapolated out to the boundary section. The boundary concentration is a function
of what occurs within the modeled region.

Boundary conditions must be applied to both the upstream and downstream sections. However, it is
perfectly allowable to mix boundary condition types. For example, a gradient upstream boundary and
a fixed downstream boundary is permitted.

If fixed boundaries are appropriate, they are generally easier to work with than gradient boundary
conditions. Gradient boundaries are more difficult to apply, but are more appropriate where little is
known about the boundary concentrations. However, gradient boundaries should not be used when
there are significant loadings to sections near the boundary because the trend will no longer be linear
and extrapolation to the boundary is not valid.

6.3.3 Differencing Options
The FINSEC Model solves the mass balance differential equation numerically by applying a finite
difference scheme. The concentration of substance S at the interface between segment i and segment i-
I must be determined for all segments. Differencing options provide the choice of how to solve for the
concentrations at the interfaces. There are thr= options: central, backward, and length dependent.

In almost all cases, the central difference option is the most appropriate. This assumes that the
concentration of S at the interface is equally influenced by the concentrations in both segment i and
segment i-I. This option is applicable when dispersive transport predominates. It is the most accurate
numerically, and exhibits the least numerical dispersion, however, it may yield negative predicted
concentrations if the section lengths are too large. To ensure that all predicted concentrations will be
positive, all segment lengths should be less than or equal to 2EN.

The backward difference option is most appropriate when advection predominates dispersion. This
option assumes that the concentration of S at the interface is completely dependent on the upstream
concentration 'in segment 1-1. It is easier to program and guarantees that the predicted concentrations
will be positive regardless of the section lengths. However, it is subject to high numerical dispersion.
Numerical dispersion is an increase 'in the dispersion of substance S as a result of the discretization
process. The concentration gradient of S is more spread out than the anaiytical solution would predict.

6-4

The length dependent approach is most appropriate when section lengths differ significantly. This
option will interpolate the concentration of S at the interface using similar triangles to compare slopes.
It is included for more advanced users.

6.4 Input Data
The Input Data notebook page is where the necessary hydraulic and chemical properties are entered for
the FINSEC Model. The appearance of this notebook page will vary depending on the modeling
options selected. The user should first set the number of sections to be modeled, including the two
boundary sections. This number can be entered by either clicking on the text box and typing the
number from the keyboard or by clicking on the spin buttons until the desired number appears. The
data entry spreadsheet itself consists of one row for each section that will be modeled. 'Me boundary
sections are highlighted in blue. If the number of sections is increased, blank rows are added to the
bottom of the spreadsheet. If the number of rows is decreased, the bottom row is removed from the
spreadsheet and any data in that row are lost.

The number of columns requiring data input will vary depending on which water quality variables are
selected. The length, cross-sectional area, flow, and dispersion coefficient for each segment must
always be entered. If the substance being modeled degrades in the natural environment than a decay
coefficient should be entered. Any known loadings of the modeled substances should also be entered
for each applicable segment.

The river reach being modeled can be divided into any number of sections. As the number of sections
increase, the number of simultaneous equations to solve increases and simulation time will also
increase. Segment length should be based on stream geometry, points of discharge, and the selected
differencing option. Thomann and Mueller (1987) state that a section length of 1-2 miles (1.6-3.2 k:m)
will generally provide a good representation of the actual river. Segment lengths should be smaller
around waste loading (discharge) points. If larger segments are used at discharge points than dilution
could produce predicted concentrations that would be less than an observed value. The selected
differencing option could affect either the positivity of predicted values or the amount of numerical
dispersion generated by the model. If the central differencing option is selected, numerical dispersion in
not a problem, but positivity could be. Refer to Subsection 6.3.3. If the backward differencing option
is selected dm the effects of numerical dispersion should be considered when selecting segment length.
Smaller segments will yield less numerical dispersion.

If dissolved oxygen is being modeled, entry boxes for the temperature and salinity are shown. These
are used to calculate the dissolved oxygen saturation concentration. For the Coastal Screening Model,
sea level is always used for the elevation. If salinity is being modeled, the saturation calculation uses
the modeled salinity concentration in each section and ignores the mean value.

6.5 Model Results
Selecting the Graph or Tables notebook page Will run the FINSEC Model. The mouse pointer will
change to an hourglass while the simulation is being processed. Results from the FINSEC Model can
be examined as either concentrations plots or in a tabular format. To view tile results 'in a graphical
forriiat, click on the Graph notebook tab or press <Alt G>. A complete description of this notebook
page is provided in Subsection 6.5. 1. To view the tabular results, click on the Tables notebook tab or
press <Alt T>. A compete description of this notebook page is providedmi Subsection 6.5.2.

6_5

6.5.1 Graph Output
The results from the FINSEC Model can be viewed as concentration or loading plots versus distance.
To display the model results, select the desired water quality van'ab!e ::nd statistic option. The graphs
can be sent to a printer by clicking on the [Print Graph] button or pressing <Alt P>.

6.5.2 Table Output
Selecting the Tables notebook page displays a nomnteractive spreadsheet containing the model results.
Ile row number corresponds to the segment number. The number of columns will vary depending on
the number of water quality variables selected. The input values for flow and dispersion in each
segment and for the water quality variables are reiterated. The volume of each segment has been
calculated from the segment length and cross-sectional area and is displayed in the spreadsheet. The
predicted concentrations in each segment for each modeled water quality variable is also displayed in
the spreadsheet.

6-6

7. Spill Model
Fhe Spill Model solves the advective diflusive equation for an instantaneous spill of a material 'into a
6ver or a stream. Instantaneous is the name applied to the assumption that the material is completely
mixed vertically and axailly immediately following the spill. The model provides an analytical
solution. The equation is established for one dimension, and therefore, the concentration of the material
is onlv a function of the distance downstream and the time elapsed since the spill occurred. The initial
pulse of material will move downstream with time because of advection. As time increases, the pulse
will spread out because of dispersion and the total mass of the material will decrease through first order
decay.

The Spill Model is based on the following mass balance equation:
as 19S E as, k
at dX 0 2 X

where s = solvent concentration
t = elapsed time since spill occurred
x = distance downstream from spill location
E = dispersion coefficient
u = average velocity
k = first order decay rate

7.1 Model Format
The Spill Model utilizes a notebook format to enter data and display model results. The first two
notebook pages provide screens for input data, the next two notebook pages present tabular and
graphical model results, and the final notebook page allows for model calibration. Each page of the
notebook can be accessed by clicking on the corresponding tab or pressing <Alt> plus the underlined
letter of the tab header. Section 7.2 provides information regarding model input and parameter
estimation. Simulation results and model output are discussed in Section 7.3. Model calibration is
explained in Section 7.4.

An example data set is included with the Spill Model and can be used as a tutorial. To access this data
click on the [Example Data] buttons on the Parameter notebook page. These data can be used to
become familiar with how to navigate through the model and it provides an example of typical input
and model results.

The first notebook page, Project Info, serves as the Spill Model's introductory screen. This notebook
page can be used for record keeping purposes. There are text boxes available to enter in the project
title, date, and the names of the individuals involved with the model run. This notebook page may be
on-dtted if desired.

The menu bar on the Spill Model provides the user with file access, editing capability, and help
information. Selecting File will display a drop-down menu list With the following items: New, Open.
Save, Save As, Print, and Exit. New deletes the current data set. A message box will ask for
verification prior to deleting current values. Open calls up the Open File dialog box to select an

7-1

existing data set. Save will save the current data set using the existing file name. If a file name has not
been defined, the Save As File dialog box will appear prompting the user to enter a file name. Save As
opens the Save As File dialog box to allow the current data set to be saved with a new file name.
Selecting Print will send the current model data and results, including the graph to the default
Windows printer. Exit closes ttie Spill Model and returns the user to the main CS Model window. If
the current data set has not been saved, it will be lost. Selecting Edit will display a drop-down menu
list with Cut, Copy, and Paste. These menu items can be used to edit input data. Cut will remove the
selected text and move it to the windows clipboard. Copy will place a duplicate of the selected text on
the clipboard, but not remove the original text. Paste can be used to put text placed on the clipboard
back onto one of the input screens. Help Will provide information pertaining to the Spill Model.

7.2 Parameters
This notebook page allows the user to enter the required model data and decide how the simulation
results will be displayed. The values for these parameters can be entered directly into each text box by
clicking on it and then entering the desired value from the keyboard. The text boxes may also be
accessed using the <Tab> key. As each text box in the Model Data fi-ame becomes active, its
corresponding label will turn blue. Clicking on the label for the various input parameters will display a
pop-up message box with help information.

Flow: Ile average flow of the river in m'/second.
Dispersion: 'Me dispersion coefficient of the river allows for tidal effects and diffusion from
concentration gradients. Units are square kilometers per day. A typical dispersion coefficient is
0.03 - 0.3 krn2/day.
Area: This is the average cross-sectional area of the river in square meters.
Spill Mass: The total mass of the contaminant spilled *into the river in kilograms.
Decay Rate: 'Me first order decay rate of the contaminant deternunes the rate at which the contaminant
disappears. The higher the decay rate, the faster the contaminant will disappear. Units are l/day.

The display results fi-ame allows the user to specify how the model will display the simulation results.
The beginning and ending distances are from the point of the spill. Negative (upstream) distances are
permissible. The upstream concentration values could be important in systems where dispersion
predominates flow. The distance interval is entered in the text box beneath the Every header. The
beginning and ending time in days must also be specified. Time zero is when the spill occurred.
Selection of tirne should reflect the decay rate of the contaminant.

The [Clear] button will delete all values frorn the parameter text boxes. If this button is selected, a
message box will appear asking for verification prior to removing current values.

7.3 Model Results
Selecting either the Graph Results or the Table Results notebook page will begin the model simulation.
The mouse pointer will change to an hourglass while the simulation is being processed. Results from
the Spill Model can be examined as either concentrations plots or ul a tabular format. To view the
results 'in a graphical forniat, click on the Graph Results notebook tab or press <Alt G>. A complete
description of this notebook page is provided in Subsection 7.3. 1. To view the tabular results, click on

the Table Results notebook tab or press <Alt T>. A compete description of this notebook page is
provided in Subsection 7.3.2.

7.3.1 Graph Results
The predicted concentrations can be viewed versus distance or time. The concentration plot versus
distance is automatically displayed when the Graph Results notebook page is activated. There will be a
concentration plot for each specified time period. Each of these plots is in a different color. To change
to the concentration versus time plots, select the time option button below the plot area. The graph will
change accordingly. There will be a concentration plot for each specified distance interval.

7.3.2 Table Results
The model results are presented ii a spreadsheet format. The first column is distance along the river
from the point of the contaminant spill. The second column is the time from when the contaminant was
spilled. The time series will reflect the values selected on the Parameter notebook page. Tb@ time
series will be repeated for each distance interval. Third column contains the contaminant concentration
in mg/L at the specified distance and time.

7.4 Calibration
This notebook page allows the user to calibrate the Spill Model to observed data if it is available. To
enter the observed values, click on the [Observed Data] button. A new form will fill the notebook
page. Enter the observed concentrations in the spreadsheet and give their appropriate time and distance
from the spill. If needed, an entire row can be deleted from the spreadsheet by clicking on the [Delete
Row] button. The (Example Data] button will provide a sample data set. The [Clear] button will
delete all values from the spreadsheet. If this button is selected, a message box will appear asking for
verification prior to removing current values. When you have entered all observed values, select the
[Close] button to return to Calibration notebook page.

Ile Spill Model Calibration allows the user to change one of the input parameters and quickly assess
the results through the concentration plots, residual plots, or the calibration statistics. To change one
or more parameters, click on the appropriate spin button or enter the desired value directly into the
appropriate text box. To evaluate the effect that the new parameter value has on the predicted
concentrations, select the (Run Model] option button. The graphical display and calibration statistics
will be automatically updated.

If the predictions option is selected in the plot type frame, then the user can view the concentrations
plots as either points or curves. There are also option buttons for selecting either distance or time as
the x-axis label. The default plot is predicted contaminant concentration versus distance. When the
residuals option is selected, the other two option frames disappear. The residual plot of the observed
versus the predicted concentrations will appear in the graph window. Viewing the residuals can help
improve the calibration effort if the residuals indicate a trend.

The calibration statistics are shown in the shaded box to the fight of the parameter text boxes. The
Spill Model supplies both the sum of the squared deviations (SSR) and the sum of the absolute
deviations (NUD).

After a calibration run has been perforined, if the exact predicted values or residuals are desired, select
the [Observed Data] button to view the spreadsheet. The predicted values and the residuals will be in
the fourth and fifth colunins, respectively.

8. Utility Models
The utility models option has not been completely developed for this version of the CS Model. Utility
models are designed to aid the user with quick, simple calculations. Currently, there is a utility model
for estimating dispersion coefficients, for determining DO saturation, and for Manning's open channel
flow.

9. References
Alhajar, B.J., J.H. Harkin, and G. Chesters. 1989. Detergent formula and characteristics of
wastewater in spetic tanks. Journal of the Water l'odurion Control Federation, 61(5): 605-613.

Basta, DJ, and Moreau. 1982. Introduction to analyzing natural systems. Analyzing Natural
Systems, Chapter 2, D.J. Basta and B.T. Bower, (eds.), Resources for the Future, Washington, D.C.

Bosen, J.F. 1960. A formula for approximation of saturated vapor pressure over water. Monthly
Weather Reviews, 88(8): 275-276.

Browning, G.M. 1979. Development for and of the Universal Soil Loss Equation. Universal Soil
Loss Equation: Past, Present, and Future. SSSA Special Publication Number 8. Soil Science Society
of America, Madison, Wisconsin.

Dombush, J.N., J.R. Anderson, and L.L. Harms. 1974. Quantification ofPollutants in Agricultural
Runoff. EPA-660/2-74-005. U.S. Environmental Protection Agency, Washington, D.C.

Dyer, K.R., and P.A. Taylor. 1973. A simple, segmented, prism model of tidal mixing in well-mixed
estuaries. Estuarine and Coastal Marine Science, 1: 411418.

Gottschak L. C. 1964. Sedimentation. In: V. T. Chow (ed.). Handbook ofApplied Hydrolou.
MacGraw-Hill, New York, NY.

Haan, C.T. 1972. A water yield model for small watersheds. Water Resources Research, 8(l): 58-69.

Haith, D.A., R. Mandel, and R. S. Wu. 1992. Generalized Watershed Loading Functions, Version
2 0, User's Manual. Department of Agnicultural and Biological Engineering, Cornell University,
Ithaca, New York.

Hamon, W.R_ 196 1. Estimating potential evapotranspiration. Proceedings of the American Society of
Civil Engineers, Journal of the Hydraulics Division, 87(HY3): 107-120.

Hamrick, I.M., and B.J. Neilson. 1989. Determination ofMarina Buffer Zones Using Simple Mixing
and Transport Models. Virginia Institute of Marine Science.

Ketchum, B.H. 195 1. The exchange of fresh and salt waters in tidal estuaries. Journal ofMarine
Research, 10(l): 18-37.

Kuo, A.Y., and B.J. Neilson. 1988. A modified tidal prism model for water quality in small coastal
embayments. Water Science and Technology, 20(6): 133-142.

Kuo, C.Y., K.A. Cave, and G.V. Loganathan. 1988. Planning urban best management practices.
Water Resources Bulletin, 24(l): 125 - 13 2.

Luettinger, J.C. 1995. A Tidal Pr ism Anai.ilsis oj"rhe Soluble Copper Mixing Zone around the
Occoquan Water Treatment Plant. Master's thesis submitted to the faculty of Virginia Polytechnic
Institute and State Universitv.

9-i

Mandel, R. 1993. The Impact of Septic Systems on Surface Water Quality. M.S. dissertation
subn-dtted to the School of Civil and Environmental Engineering, Cornell University, Ithaca, N.Y.

McCutcheon, S.C., and R.H. French. ??. Water Quality Modeling, Volume P Transport and Surface
Exchange in Rivers. CRC Press, Boca Raton, Florida.

Novotny, V, and G. Chesters. 198 1. Handbook ofNonpoint Pollution. Van Nostrand Reinhold
Company, New York, N.Y.

Novotny, V. and H. Olem. 1994. Water Quality Prevention, Identification, and Management of
Diffuse Pollution. Van Nostrand Reinhold Company, New York, N.Y.

Prescott, L.M., J.P. Harley, and D,A. Klein. 1990. Microbiology. Wm. C. Brown Publishers

Reay, W. 1994. Personal communication. Virigina Polytechnic Institute and State University.

Richardson, C.W., G.R. Foster, and D.A. Wright. 1983. Estimation of erosion index from daily
rainf@l amount. Transactions of theAmerican Society ofAgncultural Engineers, 26(l): 153-157,160.

Robinson, A.R. 1979. Sediment Yield as a function of upstream erosion. Universal Soil Loss
Equation: Past, Present, and Future. SSSA Special Publication Number 8. Soil Science Society of
America, Madison, Wisconsin.

Sandels, Mark. 1986. Stud@dng Soil Erosion with the Universal Soil Loss Equation. Oxford
Polytechnic, Discussion paper in Geography No. 24.

Sandford, L.P., W.C. Boicourt, and S.R. Rives. 1992. Model for estimating tidal flushing of small
embayments. Journal of Waterway, Port, Coastal and Ocean Engineering, 118: 635-654. ASCE.

Schwab, G.O., and R.K. Frevert. 1985. Elementary Soil and Water Engineering (3rd edition). John
Wiley & Sons, Inc.

Selker, J.S., D.A. Haith, and J.E. Reynolds. 1990. Calibration and testing of daily rainfall erosivity
model. Transactions of the American Society ofAgricultural Engineers, 33(5): 1612-1618.

Soil Conservation Service. 1986. Technical Release No. 55 (2nd edition). U.S. Department of
Agriculture, Washington, D.C.

Stewart, B.A., D.A. Woolhiser, W.H. Wischmeler, J.H. Caro, and M.H. Frere. 1975. Control of
Water Pollutionfrom Cropland. Volume P A Manualfor Guideline Development. EPA-600/2-75-
026a. U.S. Environmental Protection Agency, Washington, D.C.

Thomann, R.V., and J.A. Mueller. 1987. Principles ofSurface Water Quality Modeling and
Control. HarperCollins Publishers, Inc.

USEPA. 1980. Design Manual: Onsite Wastewater Treatment and Disposal Systems. EPA-625/1 -
80-012. U.S. Environmental Protection Agency, Washington, D.C.

9-2

NOAA,COASTAL SERVICES CT LIBRARY
@R
3 6668 14112740 9

USEPA. 1984. User Manualfor the Pesiicide Root Zone Model (PkZM). EPA600/3-84-109. U.S.
Environmental Protection Agency.

Vanoni, V.A. (ed.). 1975. Sediment Engineering. American Society of Civil Engineers, New York,
NY.

Wischmeier, W.H., and D.D. Srruith. 1978. Predicting Rainfall Erosion Losses: A Guide to
Conservation Planning. Agricultural Handbook 537. U.S. Department of Agriculture, Washington,
D.C.

Wood, T. 1977. A modification of existing simple segmented tidal prism models of mixing in
estuaries. Estuarine and Coastal Marine Science, 8: 339-347.

9-3