i
I
  i1

[I
                    '  O
                                               903R83003
                                    CHESAPEAKE BAY BASIN MODEL
                                          A Final  Report

                                           prepared  for
     U.S. Environmental Protection Agency
            Chesapeake Bay  Program
          2083 West Street - Suite 5G
          Annapolis,  Maryland   21401
 I
1  i

 I
                 Prepared by

Northern Virginia Planning District Commission
          7630 Little River Turnpike
          Annandale, Virginia  22003
                                          January 1983

-------
 I
 I
 I1
 I
 I
 I
 I
 I
 •
•
          CHESAPEAKE BAY BASIN MODEL
                A Final Report


                 prepared for
     U.S. Environmental Protection Agency
            Chesapeake Bay Program
          2083  West  Street  -  Suite  5G
          Annapolis, Maryland  21401
I
I
I
_                                         Prepared by
Northern Virginia Planning District Commission
          7630 Little River Turnpike
          Annandale, Virginia  22003
                 January 1983

-------

-------
I
I
f
I
I
I
I
I
I
I
I
I
I
I
                              TABLE OF CONTENTS
CHAPTER

I.    INTRODUCTION 	     1
      Background 	     1
      Scope of Report	     4

II.   MODELING FRAMEWORK 	     5
      Introduction 	     5
      Description of Framework 	     5
         Description of Model Software .....  	     9

III.  BASIN MODEL NETWORK	    11
      Introduction 	    11
      Sub-basin Network	    11
     .Meteorologic Dataset 	    21
      Idealized Channel Network	    23
      Discharge and Diversion Datasets 	    27

IV.   PROCESSES REPRESENTED BY BASIN MODEL 	    29
      Hydrologic Submodel	    29
      Nonpoint Pollution Loading Submodel	    29
      Receiving Water Submodel 	    32

V.    CALIBRATION/VERIFICATION RESULTS 	    39
      Introduction 	    39
      Hydrology Submodel Calibration/Verification	    39
         Methodology	    39
         Results	    44
      Calibration/Verification of Nonpoint Pollution
         Loading Submodel	    56
         Introduction	    56
         Data Reduction/Management Requirements for
            Model Calibration	    59
         EPA/CBP Test Watershed Model Calibration:
            Site Selection	    62
         EPA/CBP Test Watershed Model Calibration:
            Hydrology Results	    64
         EPA/CBP Test Watershed Model Calibration:
            Nonpoint Pollution Results 	    64
         Urban Test Watershed Model Calibration:
            Northern Virginia and Metropolitan Washington
            NURP Studies	    80
         Verification of Calibrated Nonpoint  Pollution
            Loading Factors	    81

-------
I
I
f
I
I
I
I
I
I
CHAPTER

         Determination of Representative NFS Loading
            Factors	    89
         Dry Weather Flow Concentrations .........  	    93
         Scale-Up to River Basin Models	    95
         Precipitation Loadings on Bay Surface  Area.  .  .  .  .	    98
         Recommendations for Future Test Watershed  Studies  	    98
      Receiving Water Submodel Calibration/Verification	100
         Methodology	100
         Fall Line Loading Results	103
         Results Based upon Upstream and Fall Line  Gages	163

VI.   FRAMEWORK FOR MANAGEMENT STUDIES	188
      Introduction	188
      Periods Selected for Management Studies	188
         Analyses of Long-Term Point/Nonpoint Source
            Impacts	188
         Analyses of "Worst Case" Point Source  Impacts  	   190
         Analyses of "Worst Case" Nonpoint  Source Impacts	190
         Futher Evaluations of Selected Periods.  .......  	   196
      Intrepretation of Model Output 	   196
      Framework for Model Production Runs	197

REFERENCES	200

APPENDIX A	'	A-l

APPENDIX B	B-l

-------
I
I
f
I
I
I
I
I
I
I
I
I
I
I
                               ACKNOWLEDGMENTS

      The work described herein was funded through a Cooperative Agreement
(No. CR807816-01)  with the U.S. Environmental Protection Agency's Chesapeake
Bay Program.  We gratefully acknowledge the assistance of EPA staff who
participated in the project under the direction of Mr. Thomas B. DeMoss,
Deputy Director.  The project is currently being directed by Ms. Virginia
Tippie.  Other EPA staff members with whom we worked were Ms. Bess Gillelan,
Mr. Joe Macknis, and Mr. Jim Smullen.

      The work of the Northern Virginia Planning District Commission was
performed under the overall direction of Regional Resources Division
Director, John P.  Hartigan, P.E.   Mr. Hartigan is the primary author of this
report.  Day to day project management was under the direction of Dr.
Elizabeth Southerland.  Senior level personnel involved in the modeling
activities were Mr. Hugo Bonuccelli, Mr. Alan Cavacas, Mr. John Friedman,
Mr. Thomas Quasebarth, Ms. Karen Roffe, Mr. Todd Scott, and Mr. James
White.  Senior level personnel were assisted by engineer interns Ms. Cynthia
Birch, Mr. James Eakin, Ms. Susan Lees, Ms. Mary Jo Rimkus, and Mr. Mark
Taylor.  Their help during data reduction and model calibration tasks was
invaluable.  Special thanks go to Ms. Carol Erbe who typed this manuscript
and Mrs. Betty Fox who prepared some of the tables.  Mr. Steve Kostyal is
responsible for the graphics and artwork.
                                              111

-------
I
I
f
I
I
I
I
I
I
I
I
I
I
I
                             CHESAPEAKE  BAY  MODEL

                                  INTRODUCTION
Background

    The Chesapeake Bay,  located  in eastern Maryland  and  Virginia,  is one of
the largest and most economically important  of  the  850 estuaries  in the
United States.  In order to determine  the sources of eutrophication problems
in the upper Bay and major tidal tributaries and to  formulate  appropriate
control strategies, the U.S. EPA Chesapeake  Bay Program  funded development
of the following three computer  models to represent  the  fluvial and
estuarine sections of the Bay  system:

    A.   Basin Model;  For the major river basics  (e.g., Susquehanna,
         Potomac, and James)  tributary to the Bay, this  model  simulates
         streamflow and  the transport  of point  source and  nonpoint pollution
         loadings in the free-flowing  streams upstream of  the  fall lines.
         For coastal watersheds  (e.g.,  Chester, Choptank,  and  York river
         basins) which are drained by  major  estuaries and  the  Bay, this
         model simulates the freshwater streamflows  and  pollutant  loadings
         delivered to tidal waters.

    B.   Major Tidal Tributary Model;   This  model serves as  the interface
         between the Basin Model and the Bay Model by simulating pollutant
         transport through the Potomac, James,  Rappahannock, York, Patuxent,
         and Chester estuaries.  This  model  is  applied to  the  portion of the
         estuary extending from  head of tide downstream  to the last bridge
         crossing upstream from  the Bay, an  area where assumptions of
         one-dimensional transport are  applied.  The use of  this interface
         model keeps computer costs from becoming prohibitive  and  increases
         the utility of the Bay  model package by allowing  the  more complex
         Main Bay Model  (i.e., two-dimensional  model) to be  restricted to
         the lower reaches of the major estuaries and the  Main Bay, where
         more detailed hydrodynamic simulations are  required.

    C.   Main Bay Model;  This two-dimensional, depth-averaged model
         simulates Baywide water quality impacts of  pollutant  loadings
         delivered to the Bay by the Basin Model and the Major Tributary
         Estuary Model.

A flow chart which outlines the  relationships among  the  three  models of the
Chesapeake Bay system is shown in Figure 1.

    The computer model of the Chesapeake Bay system  is designed to simulate
a history of water quality conditions for specified  streamflows and
pollutant loadings.  The major purpose  of a  water quality  computer model is
to enable the user to look beyond the measured  data  and establish
cause-effect relationships that  explain water quality levels at various

-------
                         METEOROLOGIC
                             INPUT
                              T
WATERSHED

MODEL
   / POINT  \
   \SOURCES/
 PLUVIAL
 RECEIVING
 WATER MODEL
                              POLLUTANT
                              TRANSPORT,
                               DECAY, a
                             TRANSFORMATIO
TIDAL
TRIBUTARY

8  MAIN  BAY

MODEL
                HYDRO-
               DYNAMICS
RECEIVING
 WATER
QUALITY
Figure 1.
                Flow Chart  Showing Models

                Included  in Chesapeake Bay

                Model Package
                        -2-

-------
I
I
             locations  in the  system.   In a complex system such as the Chesapeake Bay and
             its  major  tributaries,  it is difficult to develop causal relationships from
             statistical analyses of water quality monitoring data.  This difficulty can
             be attributed to  the limited duration of synoptic records available for the
             majority of the system,  particularly for high flow periods characterized by
             significant nonpoint pollution loadings.  However, even if relatively long
             water  quality records were available, the establishment of cause-effect
             relationships from analyses of water quality monitoring data would be very
             difficult.   Development of causal relationships would be limited by the
             inability  of the  monitoring agency to maintain a constant land use pattern
             and  wastewater treatment plant loadings during a monitoring period which
             would  have to be  long enough to cover a wide range of hydrologic
             conditions.  By subjecting the current land use pattern and wastewater
             treatment  plant loadings to a full range of potential hydrologic conditions
             (e.g.,  severe drought;  wetter than average summer season),  a computer model
             of the Chesapeake Bay system can be used to extend short-term water quality
             records and to examine  the most important processes responsible for water
             quality levels in various sections of the Bay.  Since all the major
             processes  responsible for pollutant discharges and transport are represented
             by the computer model of the Bay system, it can be used to quantify the
             sources of pollutant concentrations measured at a water quality monitoring
             station.   For example,  a computer model of the major river basins tributary
             to the Chesapeake Bay can be used to determine the point and nonpoint source
             contributions to  pollutant loadings monitored at the fall lines.  Likewise,
             a computer model  of the Chesapeake Bay proper can be used to compare the
             water  quality impacts of river basin pollution loadings and pollutant
             contributions from other sources such as Bay bottom sediments and the
             ocean.  More important,  a computer model of the Bay system can be used to
             evaluate the water quality impacts of alternate land use patterns and
             wastewater treatment plant discharges under the same long-term hydrologic
             conditions.

                 All three computer  models representing the Bay system can be operated
             for  several months at a time to determine the frequency (i.e., number of
             days)  of adverse  water  quality in the Bay as well as minimum and maximum
             concentrations during different hydrologic conditions.  By operating the
             three  models in series  to simulate the entire Bay system,  management
             agencies can for  the first time evaluate the Baywide impacts of regional
             water  quality management strategies in terms of the frequency of violations
             of water quality  criteria/standards for different beneficial uses (e.g.,
             fisheries  habitat,  recreation).   For example,  the models can be operated
             with existing and revised regional wastewater treatment standards to
             evaluate the impacts of regional water quality management programs on the
             Chesapeake Bay.   Locational differences in seasonal pollutant delivery by
             point  and  nonpoint sources can be examined with the models  to identify
             sections of the tributary river basins (e.g.,  Coastal Plain watersheds vs.
             Piedmont watersheds)  where pollution controls promise the greatest benefit
             in terms of Chesapeake  Bay water quality.  Since it represents the
             hydrologic cycle  in the tributary area of approximately 64,000 sq mi, the
             package of computer models can also be used to quantify Baywide water
                                                -3-

-------
quality  impacts of water  resources management  strategies  such as increased
water  supply diversions or major  flood  control impoundments which can have a
significant effect on  seasonal  freshwater  inflow  to  Chesapeake Bay.   In
short/ the Bay model package  described  herein  is  a  state-of-the-art  planning
tool which will allow  state and regional management  agencies to relate
upstream water resources  management  decisions  to  Chesapeake Bay water
quality.

    At the outset, it  should  be emphasized that a computer  model is  only a
tool to  assist with the difficult task  of  comparing  and evaluating various
alternatives for water quality  management.  The computer  model can neither
answer all questions nor  solve  all problems related  to water quality
management planning.   It  provides input to the decision-making process, but
it  should not be confused with  the process by  which  the management decisions
are made.  The value of the computer model is  its capability to quickly and
efficiently analyze the systemwide impacts of  management  alternatives and
provide  insights which are not  available from  measured data alone.

Scope  of Report

    This report covers the calibration/verification  and selected
applications of the Basin Model.  Chapter  II of this report covers the
modeling framework, including a summary of the temporal and spatial
dimensions of the modeling problem and  a discussion  of the  selected  computer
software.  In order to provide  management  agencies with an  indication of how
major  sections of river basins  were  approximated  with the Basin Model,
Chapter  III describes  the sub-basin  and channel network.  Chapter IV
provides a description of hydrologic, hydraulic,  and water  quality processes
represented by the Basin  Model.  Chapter V 'Summarizes
'calibration/verification  results. Chapter VI  outlines the  types of
management studies conducted  with the Basin model.
                                    -4-

-------
                                  CHAPTER II

                              MODELING FRAMEWORK
Introduction

    A modeling framework includes the identification of  the  water  quality
problem to be addressed by the modeling  study,  the  theoretical  structure  of
the computer model software, and guidelines for delineating  the  network of
elements (e.g., sub-basins, reaches) used to describe  the prototype  water
resources system.  Development of the modeling  framework for the Bay model
package required considerations of  such  factors as  temporal  and  spatial
scales of the water quality problem, the availability  of field data  for
model calibration/verification, and the  available computer budget.

    Of these factors, the temporal and spatial dimensions of the modeling
problem figured most prominently in software selection and the configuration
of networks for these models.  There are two aspects to  the  determination of
time and space scales:  (a) the temporal and spatial extent  of the water
quality problem (For example, in the case of time scales, is the water
quality problem short-term or long-term  in nature?  An example of  a
short-term problem is the hour-to-hour change in dissolved oxygen, while  an
example of a long-term problem is several consecutive  weeks  with dissolved
oxygen less than 5.0 mg/1.  In the case of space scales, is  the  water
quality problem localized or regional in scope?); and  (b) the temporal and
spatial interval of model calculations (For example, do water quality
calculations have to be made every hour  in order to study the condition of
the system?  Are very detailed spatial grids required  to represent the
watershed or receiving water?)

    This chapter describes the modeling framework for  each computer  model
and presents a summary of the model software selected  to fit the framework.

Description of Framework

    The principal functions of the Basin Model are to  use meteorologic
records to calculate the streamflow and nonpoint pollution loadings  in the
Bay's approximately 64,000 sq mi drainage area  (see Figure 2) and  to
simulate the transport of point source and nonpoint pollution to the Bay's
estuarine system.   In other words,  the water quality problem to  be addressed
is the transport of streamflow and pollutant loadings  to the Bay and its
tidal tributaries, rather than localized receiving water problems  (e.g.,
dissolved oxygen sag) within the fluvial river system.

    The temporal dimensions of the River Basin modeling framework are as
follows:

    a.   Since localized water quality problems in the river basins  are not
         the principal focus of the modeling study, simulations  of daily  or
                                   -5-

-------
                                       SUSQUEHANNA
                                           RIVER
                                           BASIN
                             POTOMAC
                               RIVER
                               BASIN
Figure  2.  Map of Chesapeake Bay Basin Showing Three Major  River Basins
                              -6-

-------
I
I
f
I
I
I
I
I
 I
 I
 I
 I
 I
 I
     weekly pollutant loadings delivered to the Bay's estuarine  system
     are more important than simulations of hour-to-hour changes in
     water quality within the tributary rivers.

b.   Surface runoff and nonpoint pollution loadings are generated  and
     delivered to the river system over relatively short time periods
     (e.g., generally several hours) during and following a  rainstorm.
     Pollutant transport through the river basins to the Chesapeake
     Bay's estuarine system may require several days, during which time
     pollutant degradation can occur.  In other words, one would not
     expect all point source and nonpoint pollution loadings to  reach
     the Bay due to physical, chemical, and biological processes which
     are operating during channel transport.  Therefore, the Basin Model
     must be capable of accounting for pollutant degradation enroute to
     the Bay, although detailed simulations of localized water quality
     problems due to pollutant decay and transformations are not
     required.

c.   The Basin Model should be capable of producing a long-term  record
     of streamflow and pollutant loadings which reach the Bay from a
     long-term record of precipitation and evaporation.  This will
     permit analyses of the frequency of water quality criteria
     violations in the Bay system.

d.   In order to properly simulate river basin hydrology, soil erosion,
     and nonpoint pollution loadings, the Basin Model should be  capable
     of accounting for long-term changes in watershed state variables
     (e.g., soil moisture, seasonal variations in vegetative cover, soil
     disturbance due to cropland tillage, and pollutant loading  factors).

e.   Since runoff and soil loss equations in most computer models
     require short-term rainfall intensity data in order to accurately
     calculate the amount of rainfall which does not infiltrate  into the
     soil profile and soil erosion due to the kinetic energy of
     raindrops, precipitation records should be input to the Basin Model
     at intervals of one hour or less.

f.   Since localized water quality impacts in the river basins are not
     the principal focus of the modeling study, idealized channel
     reaches with relatively high travel times (e.g., 1-3 days)   and
     lengthy computations time-steps (e.g., 12 hrs)  for flow routing and
     water quality processes can be used to ensure that computer costs
     for long-term simulations do not become prohibitive.  If local
     water quality problems were of interest, (e.g.,  dissolved oxygen
     sag in certain segments of a river), shorter channel travel times
     and computational time-steps would be required and the computer
     costs for model operation would be significantly higher.

g.   April 1st through October 31st is the critical period for studies
     of eutrophication management in the Bay system.
                                                -7-

-------
r
                   The spatial dimensions of the modeling framework  are  as  follows:

                   a.   The number and size of elements  (e.g.,  sub-basins,  channels)  in the
                        model network should reflect  the fact  that the Basin Model  is
                        primarily intended for the calculation of nonpoint  pollution
                        loadings on the fluvial channel  system and pollutant transport  to
                        the Bay's estuarine system, rather  than localized water  quality
                        problems.

                   b.   So long as a sufficiently short  rainfall intensity  interval  (e.g.,
                        1-hr) is used for model input, relatively large  sub-basins  (e.g.,
                        1,000-2,000 sq mi) can be used to accurately simulate  nonpoint
                        pollution loadings and point  and nonpoint pollution transport to
                        the Bay.  The factors limiting sub-basin size are homogeneity of
                        hydrologic characteristics (e.g., soils, geology) and  similarities
                        in rainfall characteristics (i.e.,  clusters  of raingages are used
                        to develop a single rainfall  record for each sub-basin).

                   c.   The Basin Model's channel reaches can  be relatively long (e.g.,
                        30-50 mi depending upon slope) so long as the network  provides  a
                        reasonable representation of  transport from  the  river  basin  to  the
                        Bay's estuarine system and also  maintains computational  stability.
                        In general, a single channel  reach  can be assigned  to  each
                        sub-basin and still satisfy this criterion.

                   d.   Since localized receiving water  quality problems are not addressed
                        by the modeling study, a one-dimensional model assuming  complete
                        mixing conditions in each channel reach should suffice for most
                        receiving waters  in the major river basins.  For a  one-dimensional
                        representation, channel system geometry is represented by a  linear
                        network of volume elements (i.e., reaches),  each of which is
                        assumed to be completely mixed so that there is  no  water quality
                        variation laterally or with depth.   Water quality variation  occurs
                        longitudinally (in the x-direction)  as flows are transported out of
                        one volume element into the next.

                   e.   In order to minimize computer costs for model executions, the
                        series of major reservoirs (i.e., Lake Aldred and Lake Clarke)
                        located near the mouth of the Susquehanna River  may be aggregated
                        and represented as a single impoundment which has similar pollutant
                        trap efficiencies.

                   In summary, the major components of the  Basin Model framework are as
                follows:

                   o    Capability to simulate pollutant decay and transformations during
                        channel transport which requires a  time scale on the order  of
                        several days
                                                   -3-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
    o    Capability to simulate long-term records of  streamflow  and
         pollutant loadings reaching the Bay  in order to permit  frequency
         analyses

    o    Capability to simulate continuous changes  in river  basin  soil
         moisture and seasonal changes in other important  land surface
         features

    o    Use of long-term rainfall records with intensity  data at  relatively
         short intervals

    o    Use of a level of detail in sub-basin and  channel reach network
         which reflects the focus of the Basin Model  on the  calculation of
         nonpoint pollution loading and transport to  the Bay's estuarine
         system:  relatively large sub-basins (e.g.,  1,000-2,000 sq  mi) and
         relatively long channel reaches (e.g., 30-50 mi)

    o    Use of one-dimensional representation for  fluvial channel system

    o    Reliance on relatively high channel  travel times  (e.g., 1-3 days)
         and relatively long computational times (e.g., 0.5  day)

    o    Where feasible, aggregation of reservoirs  in series  into  a  single
         idealized impoundment

    o    The total number of sub-basin/channel reach  elements should not
         exceed about 100, to ensure that long-term simulations  are
         affordable

    Description of Model Software

    The software (I)  selected for the Basin Model is  a version of  the U.S.
Environmental Protection Agency's HSPF model  (2) , which is currently being
disseminated by EPA as the state-of-the-art nonpoint  pollution simulation
model.  The Model falls into the classification of  "continuous simulation"
watershed engineering models (3) because it calculates a continuous  record
of soil moisture levels, streamflow, nonpoint pollution loadings,  and
receiving water quality from time series input data consisting of  continuous
meteorologic records.  Therefore, the selected Basin  Model can be  used to
produce the long-term streamflow and water quality  records required  for
frequency analyses.  The watershed components of the  selected model  are
capable of providing a reasonable simulation of a sub-basin  on the order of
1,000 to 2,000 sq mi so long as a representative record of mean  hourly
rainfall is used as input.  Likewise, the receiving water  components of the
selected model can provide an acceptable one-dimensional representation of
pollutant transport and decay in relatively long channel reaches with travel
times on the order of 1-3 days for high flow conditions.

    The selected Basin Model consists of three submodels:
                                               -9-

-------
    o    Hydrologic submodel:  Based upon a modified version of the Stanford
         Watershed Model (_4) , this submodel is operated with hourly rainfall
         records and daily evaporation records.  It calculates the amount of
         rainfall converted to runoff, a continuous record of soil moisture,
         and subsurface recharge of stream channels.  The hydrologic
         submodel was calibrated/verified with daily streamflow records  from
         U.S.G.S. streamgages on the major tributaries and main stems of
         river basins.

    o    Nonpoint pollution loading submodel:  This submodel, which is a
         version of the U.S. Environmental Protection Agency's NFS model
         (5) , calculates pollutant loadings on a stream channel from
         rainfall intensity records (i.e., to simulate soil erosion due  to
         raindrop detachment) and the hydrologic submodel's output records
         of surface runoff and subsurface flows.  Nonpoint pollution loading
         factors were calibrated with monitoring data from the
         EPA/Chesapeake Bay Program (EPA/CBP) test watershed study (_§) and
         other recent nonpoint pollution loading studies  <2'^'2'i2' and  were
         verified with monitoring data from the EPA/CBP fall line monitoring
         study (11).

    o    Receiving water submodel:  This submodel (e.g.,  see receiving water
       •  algorithms in 2) routes strearaflow and pollutant loadings through
         the river and lake system.  The major physical,  chemical, and
         biological processes which cause pollutant decay and
         transformations in the Chesapeake Bay Basin are  included.  Input
         includes the sum of runoff and subsurface flows  output by the
         hydrologic submodel, the sum of nonpoint pollution loadings output
         by the nonpoint pollution loading model, and point source pollution
         loadings.  This submodel also accounts for major water supply
         diversions that reduce flows in the river basin  channel system.
         Rate coefficients  (e.g., BOD decay rate) and parameters for other
         water quality processes were calibrated with monitoring data at
         U.S.G.S. monitoring stations on river basin main stems.

    The selected software has been used for several modeling studies in  the
metropolitan Washington, D.C. region during the past five years.  It has
been applied most extensively to the 580 sq mi Occoquan River Basin of
Northern Virginia for studies ranging from nonpoint pollution management
assessments  (IQ_,12_, 13) to evaluations of advanced wastewater treatment  (AWT)
needs (_L4,JL5) .  The model has also been interfaced with estuary models for
studies of AWT standards for embayments of the Potomac Estuary  (16).
Versions of the model have  also been used by the Maryland-National Capital
Park and Planning Commission as a basis for flood management and nonpoint
pollution management plans  for several watersheds in Montgomery County,
Maryland (i^ilB) and  by  the Metropolitan Washington Council of Governments
for nonpoint pollution modeling studies in the Maryland suburbs of
Washington, D.C.  (19).
                                   -10-

-------
I
I
f
I
I
I
I
I
I
V
 I
 I
 I
 I
 I
                                  CHAPTER III

                              BASIN MODEL NETWORK
Introduction

    As shown in Figure 3, the Basin Model network for the major  river basins
consists of 35 sub-basins on the order of 1,000-2,000 sq mi, 28  idealized
channels with lengths ranging from 21 to 184 mi, and six reservoirs.  In
addition, 29 Coastal Plain watersheds are included in the Basin  Model
network to provide detailed representations of direct inflows to the major
tidal tributaries and embayments of the Main Bay.  The sub-basin network for
Coastal Plain watersheds is shown in Figure 4.

    Each sub-basin is characterized by relatively homogeneous hydrologic
characteristics, while each channel or reservoir- accepts inflows from a
single sub-basin.  This network is more generalized than networks used for
previous regional studies in the respective river basins, since  pollutant
transport to the Chesapeake Bay is of greater interest than localized water
quality problems.

Sub-basin Network

    The methodology for delineating the sub-basin network is based upon an
analysis of geographic variations in the following characteristics, which
are listed in the assumed order of consideration:

    1.   Physiographic province
    2.   Topography
    3.   Hydrologic soil group
    4.   Total water holding capacity of soil

1:500,000 scale map overlays were developed for each of the aforementioned
datasets for purposes of hydrologic segment delineation.  In the case of the
physiographic province dataset, the actual boundaries of each province are
shown on the overlay.  A map of the physiographic provinces in the
Chesapeake Bay Basin is shown in Figure 5.  In the case of datasets 2, 3,
and 4, the overlays display the predominant or average characteristic in a
100 sq mi grid cell network which is used to manage and aggregate basinwide
physical features data (i.e., each cell has dimensions of 10 mi  x 10 mi).
The 100 sq mi grid size was selected as a reasonable level of detail for a
64,000 sq mi drainage area.  The author's previous modeling experiences in
the Northern Virginia region suggest that localized variations in physical
features tend to have a relatively insignificant effect on the development
of lumped parameter model datasets for sub-basins on the order of a few
hundred square miles.  Therefore,  100 sq mi grids should provide an adequate
indication of geographic variations in predominant or average
characteristics within the Chesapeake Bay basin.  Confidence in  the
representativeness of the 100 sq mi grid cell datasets was gained through
                                               -11-

-------
          CHESAPEAKE  BAY  BASIN
             MODEL  SUB-BASINS
             ABOVE  FALL LINE
       10
       I
  100
. L I
      Mi I «i

    LEGEND
	SUB-BASIN BOUNDARY
	CHANNEL
                  Figure 3
                   -12-

-------
I
I
f
I
I
I
I
I
I
I
I
I
I
CHESAPEAKE  BAY  BASIN

   MODEL  SUB-BASINS
    BELOW  FALL LINE
                       Bonenria
            LEGEND

            -  SUB-BASIN BOUNDARY
                Elizabeth
                          Figure 4


                            -13-

-------
                         //                   *2v
                         't APPALACHIAN  ^ -f  \%
                         'RIDGE a  VALLEYS '     ^\
                                       ICOASTAL
                                       i  PLAIN
                                    i. ^.
Figure 5.   Physiographic Provinces  in Chesapeake Say Basin
                               -14-

-------
sensitivity studies based upon County soils maps and 1:24,000 scale
topographic maps for Howard and Kent Counties in Maryland and the Occoquan
River Basin in Northern Virginia (20) .

    The results of the sensitivity studies for Howard and Kent counties were
as follows:  (a) average soils characteristics derived by overlaying a
Statewide soil association map with a 100 sq mi grid cell network agreed
very closely with average characteristics derived by planimetering a County
soil association map; and (b)  average slopes derived by overlaying the 100
sq mi grid cell network on a 1:250,000 scale topographic map (100-ft and
50-ft contour intervals)  agreed reasonably well with average slopes derived
by overlaying a 25 sq mi grid cell network on a 1:24,000 scale topographic
map (20-ft contour intervals).  In the case of the sensitivity studies for
the 580 sq mi Occoquan River Basin, a five-year simulation of weekly runoff
volumes (cu ft/week)  and nonpoint pollution loads (Ibs/week) based upon a
100-sq mi grid cell dataset of soils and topographic characteristics
correlated very well with a five-year simulation based upon a 23-ac grid
cell dataset (21)  that had been previously derived for a detailed modeling
study (12).  The results of these sensitivity studies tend to support the
selected data reduction methodologies for soils and slope datasets.

    Due to the size of the Chesapeake Bay basin, the assessment of soils
characteristics were based upon "soil associations."  This decision was
based on previous modeling experiences which indicated that localized
variations in soils characteristics, as indicated by a "soil series" dataset
for example,  tend to have a relatively insignificant effect on the
development of lumped parameter values for very large sub-basins.  ,Soil
association characterizations were carried out on a state-by-state basis.
The distribution of the Chesapeake Bay basin's 135 soil associations among
the five states in the drainage area is as follows:  Virginia: 54; Maryland:
27; West Virginia: 12; Pennsylvania: 34; and New York: 8.  Each soil
association consists of one-to-four soil series which are listed in order of
dominance.  Composite characteristics for each association are typically
developed by weighting the characteristics of the individual soil series
according to the fraction of the association that is typically attributed to
the series.  Unfortunately, statewide data on the fraction of each
association that is attributed to each series is available only for the
State of Pennsylvania and the Northern Virginia region.  The only source of
such data in other sections of the basin are the soil surveys prepared by
the individual counties in each state.  Based on a review of typical soil
series distributions attributed to Pennsylvania and Northern Virginia soil
associations, the following relative distributions were assumed for use in
analyzing soil association charactertistics throughout the Chesapeake Bay
basin:
    Soil Series Order
	Total Number of Soil Series  in Association	
Two Soil Series   Three Soil Series   Four Soil Series
    1st Series
    2nd Series
    3rd Series
    4th Series
     60%
     40%
60%
30%
10%
50%
30%
10%
10%
                                   -15-

-------
The "soil series order" refers to the hierarchy implicit in the soil
association name, with the first series always being the most predominant
and subsequent series exhibiting less predominance depending upon  the order
of appearance.  These assumed distributions were used to weight the
hydrologic characteristics of each series to derive a composite
characteristic for each soil association.  The results of the sensitivity
studies of Howard and Kent Counties and the Occoquan River Basin support the
use of these distributions (20).

    Average values of the following characteristics were derived for each
soil association by applying the assumed distributions to the soil series
data presented in State SCS-5 reports:  permeability; hydrologic soil group;
total water holding capacity; soil texture; soil depth; and erodibility
factor (K).  Using the 100 sq mi grid cell network, 1:500,000 scale map
overlays showing the average value for the predominant soil association  in
each grid cell were developed for each of these characteristics.

    The first step in the delineation of homogeneous hydrologic segments was
to overlay the physiographic province map with the 100 sq mi grid  map of
average slopes..  Areas with relatively uniform slopes within physiographic
provinces were selected as the first segment approximation.  The next step
was to overlay this intermediate segment network with the grid map.s of
hydrologic soil group and total water holding capacity.  More detailed
segments were derived from these last two overlays since they delineated
areas with relatively similar infiltration rates and soil moisture storage
capacities.  Particular attention was given to defining more detailed
segments in the Coastal Plain where surface runoff and nonpoint pollution
loads have a greater chance of reaching the Bay's estuarine system due to
the relatively short travel times.  This analysis of basinwide physical
features  resulted in the delineation of a preliminary hydrologic unit
network consisting of 23 hydrologic segments.

    This preliminary segment network was further refined by analyzing areal
variations in rainfall patterns within each segment.  National Weather
Service  (NWS) tapes with hourly/daily raingage records for the period
1966-1978 were used for this analysis.*  A total of 93 raingages were
included  in this study.  A generalized map of this raingage network is shown
in Figure 6.  The distribution of these hourly and daily raingages among the
major physiographic provinces  is as follows:
*NWS rainfall tapes were only available  through  1978  at  the  time  this  study
was  initiated.  Since budget constraints prevented  the acquisition  of  tapes
of rainfall data for the full period of  record at each gage,  the  period
1966-1978 was selected for  use  in calibration/verification and production
runs because it included a  good mix of wet, dry, and  average  years.  As
indicated in a later section, analyses of  streamflow  records  for  the  30-year
period  (1949-1978) covered  by NWS hourly raingages  indicate  that  1966
through 1978 is an excellent surrogate for the longer period  of  record.
                                   -16-

-------
I
I
f
I
I
I
I
I
I
I
I
I
I
I
      90
     MiltS
LEGEND

Hourly Raingages
Daily Raingages
                              Figure 6.  Map Showing  Distribution  of

                                         Hourly and Daily  Raingages
                                               -17-

-------
                                        NO. OF           NO. OF
       PHYSIOGRAPHIC PROVINCE        HOURLY GAGES      DAILY GAGES

     Appalachian Plateau                  23                0
     Appalachian Ridge & Valley           31                3
     Piedmont                             14                5
     Coastal Plain                        12              	5_
                                TOTAL:    80               13

Statistics such as mean annual volume, standard deviation, and coefficient
of variation were calculated for each raingage and compared with surrounding
gages to identify groups of gages that have similar characteristics.
Raingage groupings were further refined through intercorrelation analyses
based upon daily rainfall totals.  Basinwide isohyetal maps and Thiessen
polygons constructed for the final raingage groupings  (e.g., Figure 7) were
used to further subdivide seven hydrologic segments, resulting in the final
network of 30 segments shown in Figure 8.  Maps showing the NWS raingages
assigned to each hydrologic segment are shown in Appendix A.  As may be seen
from a comparison with the major physiographic province map (Figure 5), many
hydrologic segment boundaries correspond to the physiographic boundaries as
might be expected for a generalized network representing a 64,000 sq mi
drainage area.  Segments 1-10 represent Coastal Plain province areas,
segments 11-15 represent Piedmont province areas, segments 16 and 17
represent the Blue Ridge and Great Valley province, segments 18-24 represent
Appalachian Ridge and Valley province areas, and segments 25-30 represent
Appalachian Plateau province areas.  Average soils characteristics and
topography data used in the hydrologic submodel were tabulated for each
segment by weighting the values stored in the 100 sq mi grid cell dataset.

    Since the hydrologic segments often overlapped river basin boundaries, a
network of sub-basins was delineated to represent each river basin.  In
order to maximize homogeneity, the sub-basins were sized to ensure that the
majority of the drainage area was located within a single hydrologic
segment.  In addition, an effort was made to maintain a bankfull channel
travel time on the order of 24-72 hrs in establishing the outflow of each
sub-basin.  The resulting sub-basin network for the major river basins
represented by Basin Model is shown in Figure 3.  The hydrologic
characteristics assigned to the segments traversed by each sub-basin were
weighted for input to the hydrologic submodel.

    In light of the size of the study area, sophisticated remote-sensing
techniques offered the only feasible method for defining land cover data for
each sub-basin.  Existing land use summaries for each sub-basin are based
upon interpretations of LANDSAT satellite images from the period 1977-1979
with state-of-the-art software available at the Goddard Space Flight Center
in Greenbelt, Maryland.  The majority of the LANDSAT scenes used in this
study are based upon 1979 conditions.  The following land use
classifications were available from the LANDSAT dataset for all sub—basins:

-------
I
I
f
I
I
I
I
I
 I
 I
 I
 I
 I
 I
180470


o
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
X SEGMENT NUMBER
xxxxxx GAGE NUMBER
t
1
MAP
NOT TO SCALE
                                 Figure 7.   Sample Raingage Grouping

-------
Figure a.  Map of Hydrologic Segments in Basin Model
                                  -20-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
                   o    Forest and  idle land
                   o    Pasture land
                   o    High tillage cropland
                   o    Low tillage cropland
                   o    Urban land

In the major metropolitan areas in  the approximately  64,000  sq mi drainage
area, urban land is broken down into residential and  commercial  categories.
A total of 15 LANDSAT scenes were required to cover the entire Chesapeake
Bay basin.  For each quadrant of a LANDSAT scene, "ground-truth" sites
identified on available low altitude aerial photographs were used to
interpret statistical groupings of ground cover signatures.  Based  upon  the
ground-truth checks, each ground cover signature was  assigned to a  land  use
classification and the total area in each classification was tabulated by
sub-basin for model input.  Based upon follow-up accuracy checks in areas
which were not used for ground-truth sites and comparisons of river basin
data with published data, it is felt that the sub-basin LANDSAT  database
provides a very reasonable representation of existing land use in the
Chesapeake.  A summary of the LANDSAT analysis and results is presented  in
Appendix B.

Meteorologic Dataset

    To produce a single hourly rainfall record for each hydrologic  segment,
for the period 1966-1978, the Thiessen polygon method was used.  Special
software was designed to produce an hourly rainfall record which was based
on a representative area-weighted daily rainfall volume, preserved  hourly
rainfall intensities, and compensated for missing records.  The  excellent
hydrologic calibration results described later in this report indicate that
the segment rainfall records generated in this manner provide a  reasonable
representation of total rainfall volumes and distributions within the
relatively large river basins.

    Other daily meteorologic data (e.g., evaporation, air temperature, solar
radiation, wind speed)  required for the hydrologic and water quality
submodels were derived for the eight meteorologic regions shown  in  Figure 9,
in comparison with the thirty precipitation regions which were designated.
Eight regions for non-precipitation data were felt to be adequate because
the gages for these records are fewer in number than  the rainfall gages,
areal variations in these meteorologic indicators tend to be easier to
characterize, and hydrologic reponses in the river basins tend to be more
sensitive to month-to-month fluctuations in these data in comparison with
the day-to-day fluctuations in rainfall which are so  important.

    For the hydrologic sub-model,  the predominant hydrologic segment and
meteorologic region determine the hourly rainfall dataset and general
meteorologic dataset, respectively,  which are assigned to each sub-basin.
                                                -21-

-------
CHESAPEAKE  BAY BASIN
METEOROLOGIC  REGIONS

-------
I
I
l'
I
I
I
I
I
I
 I
 I
 I
 I
 I
 9
Idealized Channel Network

    A single idealized trapezoidal channel with constant cross-sectional
geometry is assigned to each sub-basin.  To facilitate the development of
idealized cross-sections, all channels are terminated at U.S.G.S.
streamgaging stations, with the exception of tributary channels which are
extended to their confluence points, where data on channel geometry  is
available from the table  (Form 9-207) used to construct the  stage-discharge
relationship for the gage.  As indicated above, idealized channel length was
established in conjunction with the determination of sub-basin size  in order
to maintain bankfull travel times on the order of 1 to 3 days for each
channel reach.  Flood plain slopes for each idealized channel were
determined from 1:24,000  scale topographic maps for the sub-basin.  A sketch
of the Basin Model's channel reach system is shown in Figure 3.  As may be
seen, the idealized channel system is restricted to the main stems of the
major river basins, since the focus of Basin Model applications is the
transport of pollutant loadings to the Chesapeake Bay rather than localized
receiving water problems.  It should also be noted that channels are
generally not included in the Basin Model to transport streamflows and
pollutant loads from Coastal Plain sub-basins to the tidal waters that drain
these areas.  In limiting the idealized channel system to the main stems of
major river basins, it has been assumed for purposes of this pollutant
transport study that the  time lag and pollutant transformations achieved by
channel storage in minor  tributaries and small Coastal Plain watersheds is
relatively insignificant  and can be neglected.  It should be noted that such
an assumption may not be  appropriate for modeling studies focussing on local
flooding problems or local receiving water quality impacts.  Such an
assumption would be particularly inappropriate for local point source
management studies since  the tributaries which must assimilate the waste
load during low flow periods would not be included in the model.

    To calculate the slope of each idealized channel, the Basin Model relies
upon channel length and the channel bottom elevation at the  upstream and
downstream end.  Since the idealized channel reaches begin and end at
U.S.G.S. streamgaging stations, the channel invert elevations reported for
each streamgage were assigned to the respective channels.  Downstream invert
elevations for tributary channels were extrapolated from the next two
upstream U.S.G.S.  streamgaging stations.

    Data on channel roughness coefficients was collected from State and
regional agencies which had performed local studies in the Chesapeake Bay
basin.  An average value was derived for each idealized channel reach based
upon the arithmetic means of roughness coefficient values at several
representative cross-sections.

    For purposes of pollutant transport studies with the Basin Model, the
assumption of a uniform change in cross-section geometry between selected
U.S.G.S. gaging stations provides a reasonable approximation of average
cross-sectional area and hydraulic radius within the idealized channel
network.  Since the Basin Model is not intended for assessments of localized
                                                -23-

-------
O
s
W
z
w
2
Z
             8

             03
             HH
             C£
             *
             u
             a,
             I
o
o
0
o
o
o
p~
CN
O
o
•a1
to
O
in
cn
cn
p-
CN
H
p-
CN
in
r-l
o
CO
o
r-l
o
cn
o
r-l
m
r-
0
o
p-
o
o
VD
CN
O
0
CO
CN
r-l
in
VO
in
0
p-
CN
cn
p-
CM
P-
P-
co
p-
CN
iH
T
CO
r-l
0
m
co
o
in
p~
o
o
o
o
o
co
in
o
o
rH
cn
i—i
o
r-l
rH
0
m
rH
rH
m
cn
p-
CN
cn
cn
o
cn
0
m
rH
o
o
o
m
o
o
o
in
o
o
r-l
O
CN
O
en
i—i
o
to
CO
o
r-l
CN
in
TJ>
co
o
CO
.
o
o
o
m
o
o
0
in
0
o
rH
CO
rH
cn
CN
P-
to
o
CO
CN
CO
m
rH
o
£
o
cn
o
m
p-
o
o
rH
O
0
OJ
p-
o
0
rH
in
rH
r-
o
0
in
cn
VO
o
m
co
,-N
co
CO
o
p-
.
o
o
r-
o
o
0
o
o
o
in
o
o
CO
CO
rH
o
co
CO
0
in
in
in
1-1
*r
o
in
en
rH
CN
T
O
CO
cn
o
p>
CO
o
o
rH
m
o
o
0
in
o
0
in
to
i— i
cn
CN
cn
0
CN
O
o
cn
CN
VO
CO
o
to
0
in
o
rH
<— 1
cn
o
CO
(N
O
O
m
o
o
cn
o
o
cn
rH
to
to
-in
m
m
(N
T
t—
CTl
in
H
rH
P-
o
p-
CN
0
0
i-H
«
O
ro
p-
O
O
ro
O
0
rH
o
0
in
cn
o
ro
p~
O
o
CN
CO
CN
CO
rH
m
CO
en
rH
rH
rH
O
rH
rH
cn
O
o
rH
CO
O
O
O
O
CO
CN
0
0
•V
rH
m
00
in
CO
rH
rH
CN
to
o
0
CN
0
0
cn

-------
I
I
f
I
I
I
 I
 I
 I
 I
 I
 I
 I




[continued)
ORK PARAMETERS
3
rH JT?
 ^
W >—
a
in
co
0
o
n
o
0
rH
O
0
n
in
in
r~
r-
cn
yj
i-H
in



CN

rH
in
en
CM

in
CO
o
o
m
o
o
rH
O
o
CO
in
o
10
en
U3
r-l
in



*r

0
CN
rH
CN

O CO O O CN
O O O O O
O 0 0 0 0
O O 00 CO kO
o o o o o
o o o o o
rH *3* rH ^* C^
U3 U? U3 ^O O
O O O O CN
O O O O O
en en y? en *o
rH rH
en CN vo co *3*
rr [•» "T CN rH
in en CN y3 CN
in n ^r co en
I-H n m r- en
rH rr rH m
P"" ^3 ^5* ro in
m o en rH in
co o us ^j* r**
en CN T CN rH



en r^ co T co

v^ fn ^* o^ *fj*
rH CO ^* tp CO
in m in ^* en
rH

co a*
cr> r^~
O 0
O 0
0 0
in *?
o o
0 0
rH CN
O T
O O
O O
en . in
rH
m co
Cn rH
rH
CN O
•-H en
m
O CN
co m
en in




VO 
r-l •& CN
CN in en CN
rH P- •V
rH TJ* ^3 r**
in CN o f
co rH en rH




•ST rH -3- U3

I-H in co o
•*)• co rH en
rH rH


33
EH ^
O  •
2 -H
Cd £
CN
en
CO
I—
r-
CN
                                                                                                     en

                                                                                                      0)
                                                                                                      3
                                                                                                      CTl
                                                                                                      o
                                                                         w
                                                                         "3

g

0
rH
03
B

m m o o
r~ r~ ca CN
rH rH rH CN




•K
w
Or
>H
EH


co co co co




4k

K
CJ
OS

o o m o
v-o r** r~* co
rH rH *H rH



0
O
CN






CO





o
en
rH



o
CN
CN






CO





O
O
CN



O
CN
CN






CO





o
rH
CN



1
1
1






co





o
CN
CN

O

1 0 1 0 I
I -cr i us i
1 CN 1 CN 1






CO CO CO K CO


N


o in o o o
en en ^r in vo
CN CN CN CN CN

J-l
01
£

O
UH
C
u O
'o
S > 3*1
(0 ^
d) (i) as
lH W CJ
4J <1) Q5
CO (A
n n ••

CO OS EH
O
* Z
                                                                          -25-

-------
4J
 C
 0
 o
jQ
 to
EH
       a
       EH
       w
       ss
as
s
w

z
<

 •
CN CN T
in


o o 01
• • •
CN T f~
rH rH O
rH rH

O O O
r^ co o^
CN CN CN


OS CO CO

m o o
VD r^ oo
CN CN CN

0
in
0
o
o
01
o
t
o

VO
o
o
•
o

o
•
CN
rH

0
rH
cn
CN

CO
^r
01



CO
co
en






o
•^
**o
rH


0

en
01


1
1
1


CO

o
en
CN

m
CN
rH
O
^
in
0

o

u?
o
o
t
o

o
.
^*
^

CO
01
rH


01
en




CN
in
en
rH





^O
rH
r^
01


T
•
oo
o
""

o
rH
01


CO

o
o
OI

in
CN
rH
0
T
in
0

o

CO
o
o
•
o

CO
•
o
rH

t—
CN
O
CN

r-
CN
o



f-
rH
p»»






CN
in
en
rH


co
.
01
rH


O
CN
01


OS

o
rH
01

in
CN
rH
0
•q-
in
o

o

vo
o
o
•
o

^*
•
en
rH

\o
VO
CN


r-
o
CN



OI
CO
yo






r-
rH
p~



31
•
CN



1
1
1


CO

O
CN
OI

O
O
rH
0
in
in
o

o

01
VO
o
,
o

in
•
^
CN

r-
rH
0
rH

r^
rH
O
rH


in
01
in
•H





en
rH
rH
OI


0

CN
rH


O
T
01


OS

O
01
ol

O
O
rH
O
in
in
o

d

o
rH
O
,
O

CN
•
VO


o
o
01


in
01




rH
Ol
rH






in
01
in
rH


O
•
rH
CN


r
i
i


CO

o
*4*
01

01


 0)


 •3

 cn
•H

b


 C
                                                                                                                 Q
                                                                                                                 W
                                                                                                                 to
                                                                                                                 m

                                                                                                                 u
                                                                                                                 o
                                                                                                                 14-1

                                                                                                             M  o
                                                                                                             •H  rj

                                                                                                          e  i-
                                                                                                          (T3  M
                                                                                                          a)  a;  23
                                                                                                          ^  co  cj
                                                                                                         -ui  0)  OS
                                                                                                         co  a

                                                                                                          ii   ii   ••
                                                                                                                 (M
                                                                                                         co  a;  EH
                                                                  -26-

-------
I
I
I'
I
I
I
I
I
I
 I
 I
 I
 I
 I
flooding problems or local receiving water quality impacts, fluctuations  in
channel geometry between the U.S.G.S. gaging station nodes can generally  be
ignored without introducing significant error into the  simulations of
pollutant travel time within each idealized reach.  Likewise, local
variations in the channel roughness coefficient can be  neglected for the
same reason so long as the value assigned to each idealized reach provides
an adequate approximation of average conditions between the upstream and
downstream ends of each channel.

    A summary of the characteristics of Basin Model channel reaches is
presented in Table 1.  Reading from left to right, the  entries in Table 1
are as follows:  Reach number; type of channel (i.e., either stream or
reservoir); the downstream reach number (i.e., "TRIB TO"); reach length;
upstream invert elevation (ft above m.s.l.); downstream invert elevation;
bottom width of incised channel; top width of incised channel; depth of
incised channel; average slope (ft/ft)  of floodplain; Mannings n for
channel; and Mannings n for floodplain.

    The following major reservoirs are represented by the Basin Model as
single-layer lakes:  Lake Anna (York River Basin); Lake Chesdin (Appomattox
River Basin); Raystown Reservoir (Juniata River Basin); the two Patuxent
River reservoirs; the Lake Aldred/Lake Clarke reservoir system (Susquehanna
River Basin); and Conowingo Reservoir (Susquehanna River Basin).   In the
case of the two Patuxent reservoirs and the Lake Aldred/Lake Clarke system,
the two reservoirs in series are combined into a single idealized
impoundment with appropriate aggregate characteristics.  For the Conowingo
hydroelectric reservoir, a separate operating rule computer program (22) is
used to calculate daily spills from simulated daily streamflows entering the
reservoir.  The procedure for simulating flows and pollutant transport
through Conowingo Reservoir is as follows:  (1)  execution of the Basin Model
is stopped immediately upstream of Conowingo Reservoir  and simulated daily
streamflows are stored for input to the Conowingo operating rule program;
(2) the operating rule program calculates daily Conowingo spills for the
simulation period; and (3)  the calculated daily spills  are input to the
Basin Model's representation of Conowingo Reservoir,  and the model is
operated with the simulated flows and loads entering the impoundment.

Discharge and Diversion Datasets

    Industrial and municipal wastewater treatment plants located upstream of
the fall line are included in the Basin Model,  while discharges below the
fall line are included in either the Tidal Tributary Model or the Main Bay
model.   Only treatment plants with existing discharges  in excess of 0.5 MGD
are explicitly represented,  under the assumption that smaller discharges
represent a negligible fraction of the pollutant load delivered to the Bay's
estuarine system.   Since both point and nonpoint source loadings from a
sub-basin are evenly distributed along the length of an idealized channel,
wastewater treatment plant flows and loadings are aggregated into a single
dataset for each sub-basin.   Wastewater treatment plant flows and treatment
levels are based upon a survey of recent NPDES  monitoring records.   Since
                                                 -27-

-------
most NPDES records do not include data on nitrogen and phosphorus
discharges, effluent levels are assigned to each plant to reflect the
reported treatment level.  Constant average daily flows and effluent levels
are assumed for each treatment plant for each day of the simulation period.

    Major water supply diversions are also represented by the Basin Model.
Since the model evenly distributes the diversions along an idealized reach
or reservoir, average daily withdrawals are summed by reach for Basin Model
input.
                                     -28-

-------
I
I
f
I
I
I
I
1
I
 I
 I
 I
 I
                                  CHAPTER IV

                     PROCESSES REPRESENTED BY BASIN MODEL
Hydrologic Submodel

    Based upon a modified version  (1,5) of the Stanford Watershed Model  (_4) ,
which was developed in 1960, this  submodel can be operated  with  5-minute,
15-minute, or hourly rainfall records and daily evaporation records.  A  flow
chart summarizing the hydrologic processes represented by the  submodel is
shown in Figure 10.  It calculates the amount of rainfall converted  to
runoff, a continuous record of soil moisture during and after  rainstorms,
and subsurface recharge of stream channels.  For hydrologic simulations  of
the 63 sub-basins in the Chesapeake Bay drainage area, continuous records of
hourly rainfall (i.e., 24 values per day) were used.  For hydrologic
simulations of the small test watersheds used to develop nonpoint pollution
loading factors, a continuous record of either 5-minute (i.e.  288 values per
day) or 15-minute (i.e., 96 values per day) data were used  because of the
shorter times-of-concentration.

    During storm periods, rainfall is distributed among surface  runoff and
soil moisture storage compartments based upon adjusted infiltration  rates
and the nominal storage capacities assigned to different sections of the
soil profile.  Between rainstorms, water storage in soil moisture zones  is
depleted by mechanisms such as evapotranspiration and subsurface recharge of
streams, thereby freeing up soil moisture storage capacity  for rainfall
inputs from the next storm.

    The most sensitive parameters in the submodel are the infiltration rate
and the soil moisture storage capacities (i.e., lower zone  and upper zone).
The model's infiltration rates are based upon sub-basin soils
characteristics such as hydrologic soil group, permeability, total water
holding capacity and depth to restrictive layer.  Both the  infiltration  rate
and soil moisture storage capacities are estimated from sub-basin data and
refined by calibrating the model with observed streamflow records.
Parameters governing subsurface recharge of streams are generally estimated
from analyses of observed hydrographs and refined during calibration.

Nonpoint Pollution Loading Submodel

    The nonpoint pollution loading submodel is a slightly modified version
(1,9)  of the U.S.  Environmental Protection Agency's NPS model  (5) .  It
operates on rainfall intensity records and on the hydrologic submodel's
output of surface runoff and subsurface flow records.  All  constituents  are
treated conservatively (i.e.,  no degradation or transformations)  during
overland flow transport to a stream channel,  whereupon instream processes
are applied through the receiving water submodel.  Nonpoint pollution
loading processes represented by the submodel are summarized in Figure 11.
 r
                                                 -29-

-------
                                      CJ
                                      o
                                      •a
                                      s.
                                      Cn
                                      0
                                      1-1
                                      0
                                      V4
                                      •o
                                      §
-30-

-------
I
I
I
I
1
I
I
I
I
r
o
3
2
                       ui
                       ce
1
1

-------
    For cropland, the model assumes that sediment generation  and washoff
(i.e., soil loss) are the driving forces for loadings of  all  pollutants.
Cropland loadings of sediment, which are calculated from  rainfall  records
with a soil loss algorithm related to the Universal Soil  Loss Equation  (23),
are assigned sediment "potency factors"  (i.e., ratio of pollutant  mass  to
sediment mass)  to calculate loadings of other pollutants.  The
representation of cropland areas is enhanced by several model features  such
as the capability to assign monthly vegetative cover percentages that
represent seasonal variations in exposed ground cover resulting from crop
growth and harvest and the capability to simulate soil disturbance on
user-specified dates to account for tillage practices.  For urban  and
pasture land uses, nonpoint pollution washoff algorithms  relate the washoff
of accumulated pollutant loads to the simulated runoff rate in each
time-step.  Accumulated pollutant loads at the start of a rainstorm are
calculated from the "daily pollutant accumulation rates"  (Ibs/ac/day)
assigned to each land use classification to represent the buildup  of
pollutants on the land surface and in the atmosphere (i.e., air pollution  ..
between rainstorms).  For the forestland category, pollutant  loading
calculations are based upon soil loss/potency factors as  well as daily
pollutant accumulations, with the former more prominent during periods  of
low leaf cover (i.e., fall and winter) and the latter more prominent during
periods of high leaf cover (i.e., spring and summer).

    Nonpoint pollution loading factors such as sediment potency factors and
daily pollutant accumulation rates have been developed for the Chesapeake
Bay Basin from model calibration studies with EPA/CBP test watershed data
(jj) and with other monitoring studies (_7,_8,_9, 1£) , as described in  a later
section of this report.  Given the state-of-the-art of nonpoint pollution
loading models, loading factors such as  sediment potency  factors and daily
pollutant accumulation rates are probably best viewed as  empirical factors
which can provide a reasonable approximation of a land use's  nonpoint
pollution loading potential, much like the C coefficient  in the "rational
formula" is viewed as an empirical factor that relates rainfall intensity  to
peak runoff.  The model has the capability to use monthly variations in
pollutant loading factors.  This feature permits a representation  of
variations in the pollutant loading potential of cropland areas due to  such
factors as fertilizer/manure applications, crop growth, crop  harvest, etc.
Subsurface flow loadings based upon user-specified concentrations  are added
to hourly runoff pollution loadings and delivered to the  outlet of each
sub-basin.

Receiving Water Submodel

    The hydrologic and nonpoint pollution loading submodels are used to
calculate hourly  runoff, subsurface flow, and pollutant loadings delivered
to a stream channel or reservoir by the  tributary sub-basin.  The  receiving
water submodel (_lr.2/.24.) combines the hourly streamflow and pollutant
loadings from the sub-basin models with daily point source loadings,
subtracts out water supply diversions, and calculates daily pollutant
transport and concentrations throughout  the stream and reservoir system.
                                     -32-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
While all pollutant loading calculations in the nonpoint pollution  loading
submodel assume no pollutant decay or transformation/  the  receiving water
model simulates the major physical, chemical/ and biological processes  that
change the magnitude and form of pollutants being transported downstream.
The one-dimensional receiving water model is operated  on an hourly
computation interval with the streamflow and pollutant loading  records
produced by the hydrologic and nonpoint pollution loading  submodels as  well
as daily records of solar radiation/ cloud cover/ maximum/minimum daily air
temperature/ dewpoint temperature, average wind velocity,  and
precipitation/evaporation.  Streamflow transport is handled with a  form of
kinematic wave routing, while pollutant transport out  of a given channel
reach into a downstream channel reach is based upon advection (i.e.,
transport of a pollutant by movement of the parcel of  water containing  it) .
Since the travel time through any reach is significantly greater than the
one-hour computational interval, plug flow conditions  and  negligible
dispersion are assumed for pollutant transport calculations.

    Included among the water quality processes calculated  on a one-hour
interval for each channel reach or reservoir are:  heat transfer processes
used to calculate water temperature which affects a range  of water  quality
processes; BOD changes due to decay, sedimentation, sediment releases and
decomposition of dead algae; nitrification (i.e., conversion of ammonia and
nitrite to nitrate) when sufficient dissolved oxygen is available;
denitrif ication under oxygenless conditions; sedimentation of organic
nitrogen and organic phosphorus; transformation of nitrogen and phosphorus
from organic to inorganic forms and back again; phytoplankton growth,
respiration and sinking; sediment oxygen demand and bottom sediment releases
of BOD, nitrogen, and phosphorus; and dissolved oxygen changes due  to BOD
decay, nitrification, reaeration, phytoplankton photosynthesis and  sediment
oxygen demand.  The receiving water submodel relies upon standard sanitary
engineering equations to represent the water quality processes summarized
above.  For most BOD, phosphorus, and nitrogen decay mechanisms, the
standard first-order differential equations are used.

    The first-order differential equations assume that the amount of
constituent decayed in each time interval is a function of: (1)  the amount
of constituent present at the start of the time interval; -and (2) the rate
constant for the reaction being considered.  The differential equation  can
be integrated to yield the following equation for water quality processes
involving pollutant decay (i.e., reduction in concentration due to  physical,
chemical or biological mechanisms) during each time interval:
                      C = C e
              ,         to
             where
                      Ct = concentration at time t (i.e.,  at end of time interval)

                      C0 = concentration at time tQ (i.e.,  at start of time interval)

                      KQ = decay rate constant for water quality process (hr"-"-)

                      (t-tn)  =  time interval for calculation
                                                -33-

-------
 Decay  processes represented by this equation in the Basin Model include the
 following:   nitrification (i.e.,  conversion of ammonia N to nitrite N and
 ultimately  to nitrate N) ;  hydrolysis of organic N (i.e., conversion of
 organic N to ammonia N) ;  conversion of organic P to inorganic P;  and
 oxidation of carbonaceous BOD reaeration.

     For phytoplankton (i.e.,  chlorophyll-_a) , both growth and decay are
 simulated:
  ,         to
 where
          KG = growth rate factor representing increase in phytoplankton
               concentration due to uptake of nitrogen and phosphorus (hr  )

          KD = decay rate factor representing reduction in phytoplankton
               concentration due to respiration and zooplankton feeding
               (hr'1)

 Following the calculation of the net change in phytoplankton concentration,
 the model calculates algal death with rate factors that reflect the
 availability of nutrients in the water column.

     The representation of algal growth, which is based upon Michaelis-Menton
 kinetics with a limiting condition (e.g., light, temperature, phosphorus, or
. nitrogen)  assigned to each computation interval, merits some elaboration.
 Given the scale of the modeling problem,  chlorophyll-_a simulations are used
 to approximate the aggregate effects of three processes that probably would
 be treated separately for a smaller-scale receiving water modeling study:

     1.   Uptake and removal of N and P through growth and settling of
          free-floating phytoplankton:  free-floating chlorophyll-_a is
          subject to removal by sedimentation while being advected from one
          channel reach to the next;

     2.   Uptake and removal of N and P through growth of benthic algae:
          uptake and removal by attached algae and rooted aquatic plants
          differs from free-floating phytoplankton processes in that the
          former typically remains within its channel reach while the latter
          is advected downstream; and

     3.   Removal of suspended inorganic P and inorganic N through
          sedimentation:  since the Basin Model does not provide an explicit
          representation of phosphorus sorption/desorption reactions with
          suspended sediment,* chlorophyll-^ growth assists with the
          approximation of inorganic P and N removal.
 *Such a model would not only have to represent sorption/desorption kinetics,
 but it would also have to provide a more rigorous representation of sediment
 transport mechanisms than is available in most models used for management
 studies today.
                                     -34-

-------
I
I
l'
I
I
I
I
I
I
 I
 1
 I
 I
The representation of benthic algae and sedimentation of  inorganic N  and  P
must be approximated through free-floating phytoplankton  simulations,  since
this is the only type of phytoplankton that the Basin Model  is capable of
representing.  Since upstream concentrations are not of interest  for  the
Basin Model study {i.e., the Model is not calibrated to represent localized
receiving water quality problems), the tendency of  this procedure to
overestimate chlorophyll-^ concentrations in the upstream sections of  the
river basins is not of serious consequence.  What is critical is  that  the
Basin Model must provide a reasonable representation of pollutant transport
through the main stem river system to the fall line.

    This approach to chlorophyll-^ simulation is quite reasonable for
calculations of benthic algae growth, since other models  (2.t25) use the same
equations for both free-floating phytoplankton and  attached
phytoplankton/rooted aquatic plants.  However, care must  be  taken with such
a representation to ensure that the simulated advection of free-floating
phytoplankton does not result in a significant overestimate  of organic P  and
N delivery to the fall line.  Comparisons of simulated and recorded N  and P
concentrations at upstream gages during dry weather flow  periods were  used
to identify growth rate parameters that satisfied this criterion.

    As for the inclusion of inorganic P and'N adsorption  to  sediments  in  the
calculation of chlorophyll-_a, this Michaelis-Menton approximation parallels
first-order decay terms in other models (26), and probably provides a  more
deterministic representation since it accounts for  adsorption (i.e.,
conversion to settleable organic form) and sedimentation  through  a two-step
process.  The representation of these processes is most important for
simulations of nonpoint pollution transport during  high flow periods,  since
adsorption of point source ortho-P loadings onto suspended sediment during
dry weather periods is likely to be a relatively insignificant instream
process due to the relatively low TSS concentrations in comparison with
wet-weather periods.  In the prototype river basins, high percentages  of  the
inorganic P and N loadings delivered to a stream channel  and adsorbed  to
suspended sediment are likely to be delivered through unimpounded channel
systems to the fall line during the well-distributed storms  that produce  the
overwhelming majority of seasonal and annual fall line loadings.  The
relatively high transport rates for well-distributed storms  can be
attributed to such factors as:  (a)  much greater significance of  sediment
fines in washload (i.e., rather than coarse grained particles in bedload)  in
any sorption-desorption reactions for inorganic P and N;  and (b)  relatively
high fall line delivery ratios for sediment fines due to  turbulent flow
conditions and relatively high channel velocities throughout the  river basin
channel system.   Given these assumptions, the Basin Model's  treatment  of
inorganic P and N transport should be a very good approximation of the
prototype for well-distributed storms, in that both the prototype and  the
model should exhibit relatively high fall line delivery ratios.  For  storms
which are not uniformly distributed throughout the  river  basin, the model
relies upon chlorophyll-a growth and settling under quiescent conditions  in
downstream channel reaches to achieve the required  reductions in the
delivery of upstream inorganic P and N loads to the fall  line (i.e., NPS
                                                -35-

-------
loads of inorganic P and N produced by a localized storm upstream of the
fall line are reduced through algal growth in a downstream reach where
detention times are consideratly higher and flow conditions are relatively
quiescent).   Thus, comparisons of simulated and recorded loads to check on
this mechanism for inorganic P and N removal focussed on rainstorms which
were not uniformly distributed throughout the river basin.

    Additional details on the Model's approximation of phytoplankton growth
are presented below:

    A.   Representation of Phytoplankton;  To quantitatively describe the
         dynamic behavior of phytoplankton populations, assumptions must, be
         made concerning their behavior.  In the Chesapeake Bay Basin Model,
         the entire phytoplankton population is considered as one species,
         and the mean behavior of the population is described through a
         series of generalized equations.  While such an approach obscures
         the behavior of individual species, the overall effect of the
         phytoplankton population on water quality can be modeled with
         reasonable accuracy.  Observed and simulated phytoplankton
         concentrations are based upon the concentration of "chlorophyll-_a,"
         which is a widely accepted measure of the overall phytoplankton
         population.  The mean phytoplankton composition simulated by the
         model is defined by the carbon:nitrogen:phosphorus ratio, the
         percent of algae dry weight as carbon, and the ratio of
         chlorophyll-ja to algal phosphorus.  Based upon a review of typical
         literature values (2,25_,26] , the representation of phytoplankton
         growth in each channel and reservoir reach is based upon a C:N:P
         ratio of 106:16:1 (i.e., Redfield ratio), 49% of algae dry weight
         as carbon, and 0.6 ug/L chlorophyll-_a per 1.0 ug/L of algal
         phosphorus.

    B.   Determination of Limiting Conditions for Phytoplankton Growth;  The
         growth of phytoplankton is dependent upon temperature, plant
         nutrient concentrations, light, minerals, and vitamins.  In most
         waters, the necessary concentrations of minerals and vitamins are
         available for phytoplankton growth; consequently, the variables
         which generally determine phytoplankton growth rate are
         temperature, plant nutrients, and light.  During each hourly time
         step in which the water temperature will support algal growth, the
         Model compares the phytoplankton growth rates associated with each
         of these parameters and assigns the smallest of the growth rates to
         phytoplankton in the upper layer of the Reservoir.

         1.   Temperature Control:  The Model relies upon a linear smoothing
              function to adjust algal growth rates to account for
              temperature effects.  The smoothing function is typically
              based upon a low temperature threshold which triggers the
              start of growth and a high temperature threshold for optimal
              growth, with the smoothing factor ranging from 0.0 at the low
              threshold to 1.0 at the high threshold.  For the
                                    -36-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
 I
representation of the channel system in the Chesapeake Bay
drainage area, the temperature endpoints of the smoothing
functions were set at levels that permitted the necessary
levels of phytoplankton growth in all four seasons.

Phosphorus-Limited Growth;  Phytoplankton are dependent upon
uptake of orthophosphorus to provide the continual supply of
phosphorus necessary for ordinary cellular metabolism and
reproductive processes.  In phosphorus limited situations, the
resultant growth rate has been shown to be dependent not only
on the concentration of phosphate ions, but on the nitrate
concentration as well.  In the model, the phosphorus-limited
growth rate factor is computed in accordance with
Michaelis-Menton kinetics:

CROP = VMAXP *     P04    *     N03
               CMMP + PO4   CMMNP-+ NO3

where CROP is the algal growth rate factor (hr  )  under
phosphorus-limiting conditions, VMAXP is the maximum growth
rate (hr  )  under phosphorus-limiting conditions, CMMP and
CMMNP are Michaelis-Menton constants, PO4 is the  •
orthophosphorus concentration in mg/1 as P and NO3 is the
nitrate-N concentration in mg/1 as N.  VMAXP is the most
sensitive parameter for phosphorus-limited growth
calculations.  VMAXP values for each reach were based.upon
typical literature values (26) which were refined during model
calibration to produce acceptable mean chlorophyll-^
concentrations in each river basin.  In general, the river
systems with the highest wastewater discharges exhibited the
highest VMAXP values, and vice versa.  River values ranged
from 0.16 hr   in the James River system, to approximately
0.25 hr~^ in the Potomac River system, to approximately 0.30
in the Susquehanna and Patuxent river systems.  After
accounting for the fact that the Model does not permit algal
growth during periods of darkness (i.e., 9-12 hrs per day
depending upon the season),  the calibrated VMAXP values for
river systems are equivalent to 1.9-2.4 day"  for the James
River,  3.2-3.5 day"1 for the Potomac River and 3.6-4.2
day"  for the Patuxent River.  These maximum growth rates
are within the range reported by Baca and Arnett (1976)  and
are typically less than default values recommended for
continuous simulation models (2,25^)  which rely upon
non-multiplicative Michaelis-Menton formulations.   However,
they do fall outside the range of several other literature
values for free-floating phytoplankton.   It should be
emphasized,  as indicated above, that these algal growth
representations are intended to account for both free-floating
and benthic algae as well as adsorption/sedimentation
mechanisms for reducing inorganic P and N during instream
 I
                      -37-

-------
     transport.   Therefore,  it is not unusual that the calibrated
     VMAXP values are in excess of some literature values, because
     the  Basin Model represents two different types of
     phytoplankton and overall partitioning between organic and
     inorganic nutrient fractions whereas the literature values
     typically represent only free-floating phytoplankton.

     Nitrogen-Limited Growth;  Nitrogen is essential to
     phytoplankton for assimilation of cellular proteins and
     enzymes.  The sum of ammonia-N and nitrate-N concentrations
     represent the pool of inorganic nitrogen available to support
     algal growth.  The ratio of ammonia-N uptake to nitrate-N
     uptake is based upon an ammonia preference factor specified by
     the  Model user.  In nitrogen-limited situations,  the model
     uses a Michaelis-Menton expression to determine the growth
     rate factor as a function of the total inorganic  nitrogen
     concentration:

     GRON = VMAXN *     (N03 + NH3)
                    CMMN + (NO3 + NH3)

     where GRON is the algal growth rate factor (hr  )  under
     nitrogen-limiting conditions, VMAXN is the maximum growth rate
     (hr~^)  under nitrogen limiting conditions, and CMNN is the
     Michaelis-Menton constant.  Simulations of nitrogen-limited
     conditions typically occurred during extreme low flow periods
     in reaches with significant wastewater discharges, where
     ambient inorganic N:inorganic P ratios were less than 10.0.
     Based upon calibration studies for low flow periods, maximum
     growth rates (VMAXN) which were typically similar to or
     slightly higher than VMAXP values were derived.

4.    Light-Limited Growth;  During each hourly time step, the Model
     computes the amount of radiation available to phytoplankton in
     the upper layer of the idealized reservoir by applying
     computed light adsorption rates to the amount of radiation
     entering the water.  The Model then applies the
     Michaelis-Menton equation for light-limited growth of
     phytoplankton:

     GROL = VMAXLT *     LIGHT
                     CMML + LIGHT

     where GROL is the phytoplankton growth rate factor (hr"1)
     under less than optimal light conditions, LIGHT is the light
     intensity available for algal growth in langleys per minute,
     and CMML is the Michaelis-Menton constant for light-limited
     growth.   As previously indicated, the Model does not permit
     growth to occur during night-time hours when solar radiation
     is set to zero.  In general, VMAXLT values were set at levels
     that restricted light-limiting conditions in river reaches to
     night-time hours.
                          -38-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 9
                                   CHAPTER V

                       CALIBRATION/VERIFICATION RESULTS
Introduction

    Before the Basin Model could be used to evaluate management strategies,
it had to be "fine-tuned" or calibrated to ensure that the model accurately
represents the prototype river basins.  Following calibration, the model had
to be verified by operating it for conditions which are different from the
calibration conditions.  Three levels of calibration were required for the
Basin Models  Basinwide hydrology calibration; nonpoint pollution loading
factor calibration for single land use watersheds; and receiving water
calibration to set subsurface flow concentrations and rate coefficients for
water quality processes.

Hydrology Submodel Calibration/Verification

    Methodology.  As indicated in Table 2, a total of 14 streamgage records
were used for hydrology calibration.  The locations of streamgages used for
calibration/verification are shown in Figure 12.  The period April 1971
through October 1976 was used for model calibration, since this period
included a good mixture of relatively wet, dry, and average years.  Model
verification was based upon the periods April 1966 through June 1970 and
November 1976 through December 1978, which were generally somewhat drier
than the calibration period.  During calibration, the models were operated
with meteorologic records for the entire 5.75-yr period and a single set of
parameter values.  Based on comparisons of simulated and observed
streamflows, the most sensitive parameter values were iteratively adjusted
to establish the final parameter sets.  After acceptable agreement was
achieved on a seasonal and annual basis, simulated and observed daily
streamflows were compared for each storm event to set hydrograph shape
factors and for dry weather flow periods to set baseflow recession
constants.  Following calibration, the models were operated for the 6.4 year
verification period without any adjustment to the calibrated parameter
values to determine how well the models represented conditions different
from the calibration period.

    Since only 3% of the Chesapeake Bay Basin is currently covered by urban
development, the calibration activities focused on soils parameters that
determine an undeveloped area's hydrologic characteristics.  Due to the
distribution of raingages and streamgages, it was not possible to calibrate
the Basin Model for every major watershed in the 64,000 sq mi drainage
area.  Therefore, one objective of hydrology calibration was to derive
relationships between river basin physical features and model parameter
values which could be applied to major watersheds that could not be
calibrated separately.  In other words, rather than indiscriminately
adjusting the model's parameter values to produce the best possible
comparisons between simulated and recorded streamflows at each streamgage,
                                                -39-

-------
                                  Table  2
     USG3 Gages  Used for Calibration/Verification of Basin Model
                                               3SAI^AG2 AREA
             "J 3 G 5
                         G A
                                                             CALISRATTON/ViSiriCATICtt
— 1
KEY
A
3 .

C
D

£

r
G

a
• i

j

K

r"
M
M

0
?

2

R
S
T



:!UMSEH
02035000
01563000

01613000
0163100

01636500

0161300
01538500

01546500
015S15CO

01553500

01536500

01540500
01567000
01570500

01576000
01573310

02042500

01574000
01674500
EASTERN
01491000
01437000
01485000
MAMS
James River ac Cartersville, 7a.
Rappanannock River near
Fredericksburg, Va-
Potomac River at Hancock, Md.
So. Fork Shenandoah River at
Front Royal, Va.
ohenandoah River at Millville,
WV
Potomac River at Shepnerdstown , -,W
Potomac River at Point of
Rocks, MD
Potomac River ax. Washington, O.C.
Wast Branch Suscuehanna River at
Williamsport , Pa .
West Branch Susquehanna River at
Lawisburg, PA
Susquenanna River at Wilkes-
3arre, PA
Susquenanna River at Danville, PA
Juniata River at Newport, Pa.
Susquehanna River at Harrisburg,
PA
Susquenanna River at Marietta, PA
Susquenanna River at Conowmgo,
MD
Chickahominy River at Providence
Forge, VA
Mattaponi River at Bowling Green, VA
Mattaponi River at Beulanviiia, VA
SHORE GAGto {SUM OF FLOWS)
Chop tank River or. Greensboro, MD
Manticoke River or. 3ridgavilla, OE
Pocomoka River or. Hillards, MD

16,206
4,134

10,550
4,253

•7,374

15,374
24,996

29,942
14,716

17,734

25,300

29,060
3,537
52,400

57,313
70,193

642

666
1,357
544
(233)
(194)
(157)
HYORCLCGY
X
X

:<


X




X
X



X


X


X
X

X

X
X
X



WATER QUALITY
X
X

X
X

X

X
X

X


X



X
X
X


X









Gage locations  are  shown in Figure 12.
                                 -40-

-------
I
I
l'
I
I
I
I
I
I
I
I
I
I
              50
          ,   i  I
             Miles
           LEGEND

         — — BASIN BOUNDARY

         O    CHAMNEINODE

              RESERVOIR
    LEGEND


• Calibration Gage

O Channel Node
                                        SUSQUEHANNA
                                            RIVER
                                            BASIN
                                      JAMES
                                      RIVER
                                      BASIN
                          Figure 12.  Map Showing Basin Model Channel Network

                                      and Calibration/Verification Gages
             NOTE:  See Table 2 for Map Key
                                               -41-

-------
hydrology calibration focused on developing parameter  estimation  methods
that could be applied to ungaged watersheds.  This  approach  has previously
been used to calibrate hydrologic models of several watersheds in the
Northern Virginia portion of the Chesapeake Bay Basin  (_27,_28) .

    A summary of the major model parameters which were related to river
basin physical features is shown in Table 3.  LZSN  (lower  zone soil  moisture
storage capacity) and INPIL  (infiltration rate) are the most important
parameters for simulations of annual streamflow volumes.   An increase  in
LZSN will inprease the storage of water in the idealized lower zone  of  the
soil, thereby lowering seasonal and annual streamflows by  increasing the
depletion of soil moisture through evapotranspiration  and  by reducing  the
frequency of saturated soil conditions.  An increase in INFIL will likewise
lower annual streamflows by  reducing direct runoff  due to  higher
infiltration and increasing  soil moisture depletion through
evapotranspiration.  INFIL is most often used to modify seasonal  streamflows
after LZSN and a reasonable  range of INFIL values which achieve acceptable
annual streamflows have been identified.  As has been  the  case in previous
studies (.l,_9f_2Z'.2!3'.22'.32'_3_P  , LZSN in the Chesapeake Bay Basin was found to
generally be directly related to average total water holding capacity  and
depth of soil above the restrictive layer.  Calibrated LZSN  values ranged
from approximately 2.0 to 4.0 in. in the Appalachian Plateau and  mountainous
areas of the Appalachian Ridge and Valley to approximately 6.0 to 8.0  in.  in
the Coastal Plain, lower Piedmont, and portions of  the Appalachian Ridge and
Valley.  INFIL was found to  be related to indicators of surface runoff
potential such as hydrologic soil group, soil permeability,  and soil
texture.  Calibrated INFIL values ranged from approximately  0.01  to  0.015
in./hr in the Appalachian Plateau and sections of the  Coastal Plain  and
lower Piedmont with "D" hydrologic soil groups to 0.06 to  0.07 in./hr  in
sections of the Coastal Plain with "B" hydrologic soil groups.  K3 is  an
evapotranspiration index that is generally set at reasonable levels  to
reflect vegetative cover, and then held constant while LZSN  and INFIL  are
calibrated.  K3  values on the order of 0.3-0.45 were used  throughout the
Chesapeake Bay Basin, with the higher values typically associated with areas
characterized by high forest cover to account for the  effect of vegetative
cover on evapotranspiration.  Although it is not as sensitive a parameter  as
either LZSN or INFIL, UZSN (i.e., soil moisture storage near the  soil
surface most closely related to depression storage) can have some effect on
seasonal and annual streamflow volumes because of its  impact on individual
storm events.  UZSN is most  often related to the calibrated  LZSN  value  and
in the Chesapeake Bay Basin  was typically found to  be  5%-15% of LZSN.   After
acceptable agreement between simulated and observed streamflows is achieved
on an annual and seasonal basis, the interflow coefficient INTER  is  adjusted
to redistribute  streamflows  between surface runoff  and subsurface flows in
order to match the shapes of the recession limbs of observed hydrographs.
Relatively low INTER values  assign more of the streamflows to surface
runoff, thereby  resulting in higher peak flows and  steep recession limbs.
Like previous studies (_1,_9,^7,^*3) , Chesapeake Bay Basin calibration  results
indicated that INTER could be related to average land  slopes, with values
ranging from 1.0 in relatively flat Coastal Plain areas to 1.5 in

-------
I
I
I
I
I
I
I
I
I
                        n   3
                        a
    CO

  ,  *
-3  3
a  £-1
Q  <
O  a
S  Cu

X  J
o  <


O  CO
os  :*
Q  —
>«  O,
                             2  M
                             a  co
                             a  <
                             g  a

                             a  a
                             to  3
                                 co
                             co
                             O.  Q
                             CO
                             z  co
                             O  ca
                             M  a
                             tr-  r-i

                             3  2
                             a  a
                             cs  «i
                                         73
                                                     S-l
                                                     0)
'-1

CU

4-1
fl






CO
CU
M
H
i
o
M
S-i
S
3
f£



















Z
O
M
E*
(— i
M
fa
a
Q





3

rH
fl
4J
P
E-i
0
4J
TJ
0)
4J
fl
rH
0)
U

en
•H
z
co
N
J
4J
C
,
j-i
•H
O
fl
a
fl
o









>1
•H
O
fl
I?
CJ

Oi
C
•H
-o
1— t
0



, o
CT1 4J VI
0 -H ^ (U
f-t — 1 OJ >
0 -H > -r-,
i-l ^2 fl 4J
t3 fl HH fl
> CD 4-1
= £ s-i a)
l-i (U Oi
o a) a m
•u a, a >
3
T3 -H 0
O -H 14-1 4J
•I-1 0 0
fl CO 13
r-* CU tU
CU • S-l JJ
s-i a 3 fl
3 4J _|
w o x cu
•H SH 0) 1^
CJ c^
J U)
HI *H ^ .H
fa -H -H
2 o O rn
w CO CO ^





C
0
•H
4J
fl
tj
•H
u a
>u 5>
4J C
g fl
S V4
fl 4J
l-i 0
fl a
a 03
CU 0)
JJ
fl ~1
U fl
3
C 4J
0 0
•H fl
4J
fl 0
Lj ! 1






Z
CO
N
t-3
0
4-1
T3
,
^
a
fl
ij
w
o
a
o
4J

0
4J

•o
CU
JJ
fl
r^
CU
s^
tn
.^

cs
a
E-t
z
M

















4J
C
CU
•fl
0
'^i
(44
M-l
(JJ
0
o

5
0
fl O -U fl CL u_i



C r
•f-V
S
iu
-1
M
3
^-1 X
•H CO
«-i T>
C C
M M
= 3
•H
S !-l
0 0
2 ^
^

-------
mountainous areas.  After INTER had been set, the baseflow recession
coefficient (KK24) was fine tuned to improve the agreement between  simulated
and observed dry weather flows.

    Results.  Calibration/verification runs were terminated for each
streamgage when it was determined that the differences between simulated and
observed strearaflows could not be improved with further parameter set
adjustment and a sufficient number of model runs had been completed to
develop reasonable regional parameter sets.  The calibrated parameter values
for the sub-basins located above the fall line are summarized in Table 4.
The hydrology parameter sets produced by these calibration/verification runs
are quite reasonable based upon some previous continuous simulation modeling
studies in the Chesapeake Basin (.1,.9,if'iZ'2Z'2§.)  and other literature
values (_5,_3JJ,3i2) .

    Comparisons of simulated and observed streamflow data based upon
streamflow volumes (annual and seasonal)/ daily streamflow time series, and
daily flow-duration plots (period of record and seasonal) generally indicate
very good calibration and verification results.  Table 5 summarizes
comparisons of simulated and observed annual streamflow volumes for 14
streamgages.  As may be seen, differences between simulated and observed
annual volumes are typically within the range of observation errors (e.g.,
±20%)  associated with meteorologic and streamflow data collection activities
(33).   In general, the greatest differences between simulated and observed
streamflow were associated with winter periods and drought periods.  Winter
periods, which were typically somewhat undersimulated by the Model, tend to
be characterized by frozen ground conditions which are not simulated by the
River Basin Model as well as the highest raingage errors due to freezing
conditions.  Since winter periods are not expected to be used for water
quality assessments due to the relatively high assimilative capacities of
most receiving waters, the higher errors in winter flow simulations are not
of serious concern.  Streamflow errors during drought periods, which were
typically oversimulated by the River Basin Model, can be attributed in large
part to the tendency of the Thiessen-weighting procedure to exaggerate the
areal distribution of localized storms which tend to be more significant
during droughts.

    Comparisons of simulated and observed daily flow—duration curves are
shown in Figures 13 through 21 for both calibration and verification periods
at nine major river streamgages.  As may be seen, the agreement between
simulated and observed curves is typically quite good, with goodness-of-fit
for the verification period typically almost as good as for the calibration
period.  In general, the calibration period exhibits better agreement for
low flow periods than does the verification period, which was characterized
by drought periods that presented difficulties with the development of a
representative mean segment rainfall record.

    As another indication of goodness-of-fit, Table 6 summarizes statistics
on the correlation between simulated and observed weekly streamflows at the
nine gages covered by Figures 13 through 21.  Correlation coefficients are
                                    -44-

-------
I
I
f
I
I
I
 I
 I
 I
*3*


0)
  I
  I
  I
  I
      8
      j
      §
      Q
                  ,Z
                            OS
                            U
                            EH
                            2
J
M

2
                  isi
                  D
                      < o
                          X 0
                         — (13
                            a
                            w
0
«3
CN
O
r~
rH
rH
o
o
o
•3-
in
ro
o
0
'a1
o
T
T
*3>
rH
in
rH
O
o
o
i— i
CN
O
O
0
in
(N
r-
(N
vO
U3
rH
O
rH
1
0
o
ro
o
o
00
o
o
o
o
•q-
in
ro
o
o
^*
o
o
r»
ro
rH
CN
iH
a
0
o
rH
CN
O
o
o
CN
CN
<£>
O
r—
rH
ro
O
CN
U
r£

o
in
U
a
^0
UD
(N
O
rH
VO
r-l
O
0
o
ro
m
ro
O
0
T
o
in
rH
m
VO

rH
CN
'T
a
r~
CN
in
rH
O
co
B
u
CS
o
ro
CN
O
0
CN
rH
o
O
0
ro
in
ro
o
0

O
ON
u
»
f
ro
CN
O
O-l
O
rH
O
0
O
ro
in
ro
o
o
•31
o
a*
CN
*3"
03
in
CN
O
O
r-
03
CN
O
"*
rH
03
ro
• VO
o
p~
in
rH
O
o
rH
B
8
in
CTi
CN
0
ro
\o
o
o
o
"O
m
m
ro
o
o
•^
0
CN
0
ro
rH
ro
O
0
r~
in
in
o
CN
03
in
in
VC
01
o
CN
rH
O
rH
rH-
B
U
cc
l>-
CN
ro
0
r~
T
o
o
o
0
T
in
ro
o
0
T
o
0
m
CN
rH
O1
•^
ro
O
O
in
^
«3
O
in
in
T
V£>
O
in
TT
rH
in
o
CN
rH
O
IX
O3
in
ro
o
rH
ro
o
0
O
O
^T
in
ro
0
O
i"
O
ro
O
CN
rH
CM
10
r-l
O
O
10
rH
r^
0
m
ia
rH
r^
T
03
m

rH
O
^
rH
O
a
0
in
CN
o
^O
CN
rH
O
O
O
ro
in
ro
O
0
<*
O
o
en
ro
rH
•q-
CN
0
0
0
0
in
0
o
o
o
in
^0
O-i
to
m
CO
o
'-O
iH
1
O
ro
CN
0
*£>
tr
r-l
0
O
0
ro
in
ro
o
o
•3"
O
O
UJ
T
rH
O
"31
O
O
O
O
in
o
o
0
0
in
CN
in
in
in
CA
o
r~
rH
U
X
ro

r-
ro
rH
ro
ro
CN
0
0
CN
0
O
rH
O
rH
O
!-•
in
rH
rH
^O
rH
0
CO
rH
O
»
                                                                       -45-

-------
13
0)
D
C
c
o
o
T

ai
     OS
     §
     Q
03
W
OS


C/3
W

>-3
Z
Z

n
££
OS
63
g

J
1— I
Ct4
M
2
W
N

z
w
N
. T
t—4

o "w
i— I QJ
W
X O
^ (0



z
M
W
c^
03
1
03
D

en
CO
CN
0
vo
CN
rH
O
O
O
CO
in
ro
o
CN
ro
o
CN
in
•w
iH
O
CO
o
0
o
r-
o
rH
O
CO
CN
r-
cn
O
•
in
o
r-l






o
en
rH

33
O
CN
CO
CN
0
VO
r-t
rH
O
0
o
in
ro
0
ro
(-0
O
o
ro
**"
ro
cn
CN
o
o

in
o
i— i
T
CN
CN
r-
• CN
in
r—
•
o
01






o
o
CN

33
O
o
CN
ro
0
O
in
0
o
o
0
CO
o
o
•*
o
o
VO
CN

en
ro
0
O
o
co
vo
o
o
o
ro
vo
co
CO
TJ1
I
CO
vo






o
rH
CN

33
O
O
CN
ro
O
O

O
o
o
o
in
ro
0
o
T
O
o
en
i-t

in
rH
O
O
O
vo
in
o
o
vo
in
m
CO
CN
ro
•
en
in






o
CN
CN

EC
O
o
I— 1
ro
0
O
vo
0
o
o
o
in
ro
0
in
T
o
o
0


VO
CN
0
0
o
CN
VO
0
O
in
iH
vo
•3"
r-t
•
CN
0
r-(






0
co
CN

33
U
OS
ro
O
co
0
O
in
o
o
in
CN
in
ro
0
in
<3-
o

in
CN
in
*3*
CN
O
O
I—
CN
<""
0
CN
VO
ro
vo
00
^
^*
•
VO
r-l






in

CN

33
o
OS
vo
oo
CN
0
en
CN
0
o
oo
in
ro
o
in
"
o
r-(
m
r-l
rH
ro
rH
CN
O
O
CO
in
en
o
r-
i-H
co
VO
vo
rH
o
•
CN
CN






O

CN

X
o
OS
0
rH
ro
O
O
vo
0
o
o
o
in
ro
o
in
-jr
O
O
o
ro
r-t
VO
CN
O
O
O
CN
VO
O
o
in
r-l
vo
CN
in
en
•
rH
CN






O
in
CN

33

•=r
a
o
in
rH
rH
VO
CN
O
O
O
ro
r-
o
o
in
CN
t-~
en
rH
O
•
VO
CO
0
CN
O
f-H
ro
^
o
o
ro

33
8
CO
vo
ro
O
vo
CN
O
0
VO
in
ro
0
o
•5T
o

VO
rH
rH
r-
rH
O
0
en
vo
r~
o
o
VO
r-
vo
en
vo
CN
•
CN
CN


O
T
ro
^
O
ro
ro

X
%
                                        -46-

-------
I
I
                        OS
f
I
I
I
I
I
I
•o
 OJ
 a
 c
                c
                0
                o
f

0)
                10
                EH
QS
a
EH
       <  CQ
       Pu
                      rj
                      O
                      i-J
                      O
                      QS
                      a
                        J
                        HH

                        Z
                                        z
                                        w
                                        N
                                        D
                                        z
                                        w
                                    rH  
ro
|

00



*c£
I—I
£H
C/3
o
o

CO
^O
ro
O
in
CN
o
0
CM
in
ro
o
o
T
o
in
in
rH
rH

rH
O
O
CO
CO
r~
o
o
in
U5
•
in
cn
o
in
•^
0
CO
rH
rH
rH

CN
O
O
T
CO
on
o
CM
in
o
^
CM
in
cn
in
rH

Z
M
SI
0

10
r*»
in
ro
CO






W
1
«3*
CO
ro
o
cn
rH
O
O
0
O
m
m
ro
o
0
"
0
ro
o
rH
rH
O>
TJ"
O
o
CO
o
f*-
o
o
cn
•w
tr>
ro
O
ro
^
cn

X
D
EH
CU

OS
o
CM
r-
CN
o
rH
rH
0
0
o
0
in
m
cn
0
o
•=r
0

rH
O
rH
ro
«q«
O
0
*»•
CO
co
o
in
o
•q-

o
ro
•
rH
rH




CJ

O
g
CK
                                                                                  -47-

-------
                        OS
       CO
 C
•H
4J

 O
 O
•V

 

EH
           CO
           c
       O.   W
       K   <
-------
I
I
f
<
>
3"

K U) O
u o; o
> y o
sss
< o
« a IN
w o
55 =
"3
K
g
a
v
8
<
a
2
KSBURG, VA.
u
a: -—
u o
a o
cu o
tt 3J
U. 
•J3
td W
z —
1
a! - —
u *; o
28§
ce o '">
3 -H
2 ™" O
§< =
Cu
rfi
>
w *
2 3 «
K 3 3"
M O
S > w»
< J J3
O J ^
5*3
u S =
w

a;
cu
>
M
oi

i
u
Q
2 —
>-. c
— O
W in
< 0
«2
< 0
z *
<
a*
£ <
a a •—
i a
< w o
i SS
SdS
W » -*
d*. °
w * —
w
tlKAHMA RIVEI
5
M
3
W
tf
01
i
<
CM
H
a.
fo"
o
*H ^
J ^,
J in
(-1 1/1
Z3
< —
«C
a.
K - — ,
MHO
> X O
« O 0
a a. r-
2 v£
ess
< o
2 "* —
=i
n
DC <
u a.
>
(X < ^»

-------
   z a
           10  «
*o-iJ  03a.v3ioNi ;
                 3HU JO !N33H3d
                                         -50-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
(O


 I


en
w
   •
                               _l
                                                                                           u.
                                                                                           o
                                                                                         i;
                                                               2


                                                              !*
                                                                                         O <
                                                                                         O o
                                                                                         P.
                  S   2   <*
                 0*0000003000
                 oi*mK«>io*5>«   —


                   MOId  03J.V01QNI ^  3NU.  JO  !N33U3d
    C
    0

    4-1 4J  -r-l
    rO  ITS  £
    U

    144  OJ  W

           O
    
 C -^  OJ   •
 0     4=  3
•H  C W
4-1  O
 (0 -rt
 (-1 4-1  ••
 3  <0  W  r-l
Q  !H 'O  -H
 I  J2  O  >

 0 3 H  3
t—I  tO  01  -i-l
ti U &  2
0)

3


PL,
                                                                                                           O
                                                                                                        C  0
                                                                                                        o  o
                                                                                                       •H  C
                                                                                                       4J  10

                                                                                                        o  ~
                                                                                                       •H  4-1
                                                                                                       14-1  !0
                                                                                                    SJ -r-l

                                                                                                    C!  a)  a

                                                                                                    01     -H
                                                                                                    oj TI  as
                                                                                                    >  c
                                                                                                    i-( tO  O
                                                                                                    a     « —
                                                                                                    u —  e •"'.
                                                                                                       ^  O S
                                                                                                    C -^  4-1
                                                                                                    o     o tr
                                                                                                    •H C  CU 10
                                                                                                    4-1 0
                                                                                                    10 -H
                                                                                  ro
                                                                           4J  ••  r-»
                                                                           TJ  en  o
                                                                           Vj xi   *
                                                                           43  0  T
                                                                                                    3
                                                                                                    Q
                                                                                                    I
                                                                                                    3
                                                                                                    0
                                                                                                    -4  (0  Q)  Q
                                                                                                    t.  U Oi  S
                                                                                                                                             (I)
                                                                                                                                             u
                                                                                                                                             3
  t
                                                                                 -51-

-------
                                                  o
                                                  o
                                                  Q.
                                                  o
                                                  
                                 _l
                                 4
                              O a
                              8
                              K>
01
cri
en
     O
     co
O
(O
  o
  «r
SHU.  jo
3
ooooooooo
a>ar»(£tnTineu   —
                                                                                                                                          LTI
                                                                                                                 c      -~-
                                                                                                                 o  -u  -H
                                                                                                                •H  <3  S
                                                                                                                4J

                                                                                                                 O  0)  M
                                                                                                                •i-l  >

                                                                                                             u  •-<  ci
                                                                                                             0   !-*
                                                                                                            M-l   Q)
                                                                                                                >  O  ^
                                                                                                             W      0  —'
                                                                                                                C  C  <
                                                                                                             ^   fQ  fQ  £>
                                                                                                                    j=
                                                                                                                    13    -
                                                                                                                    a, cn
                                                                                                             c  — a, w
                                                                                                             0      fl  3
                                                                                                            •H   C  OS  XI
                                                                                                            4J   o      w

                                                                                                             S-l  -U   ••  CJ
                                                                                                             3   (3  M  ••-(
                                                                                                            C   ^  13  VJ
                                                                                                             I   XI  0  4)

                                                                                                             O  rH  S  0)
                                                                                                            r-4   (3   (U  i-l
                                                                                                            fa  U  Oi  fa
                                                                                                                               3
                                                                                                                               O
                                                                                                                               0)
                                                                                                                               S-l
                                                                                                                               3
                                                                                                                               cn
                                                                                                                               0      —
                                                                                                                              •H      -r^
                                                                                                                              *j       e
                                                                                                                               "3   k
                                                                                                                               o   o   cr
                                                                                                                              •H   a)   cn

                                                                                                                           l-i -H      O
                                                                                                                           0  ^(   1-1  ^D
                                                                                                                          y-t  d)   (U  LD

                                                                                                                           Cn     ••-(  rH
                                                                                                                           CD T3  Oi  i—I
                                                                                                                           >  C      —
                                                                                                                           1-4  (3   O

                                                                                                                          CJ -»  S  O
                                                                                                                               c3   O    •
                                                                                                                                  4J  Q
                                                                                                                                   O
                                                                                                                               c
                                                                                                                               o
                                                                                                                               JJ   O       C
                                                                                                                               (3  *H       O
                                                                                                                               Ij   *J   ••  4J
                                                                                                                               3   (3   cn   U1
                                                                                                                               a   ^  -a   c
                                                                                                                                  XI   O  -i
                                                                                                                               I
                                                                                                                               3
                                                                                                                               0
                                                                                                                              rH   IT3
                                                                                                                              fa  O
                                                                                                                           a
                                                                                                                           cn
                                                                                                                              fa
                                                                                                                                       cn
                                                                                                                                       nj
                                                                                                                                      3
                                                           -52-

-------
 I
 I
 f
I
I
I
I
I
I
I
I
I
I
f
0)
m
a>
O
   a
                                                                                                        0  M
                                                                                                        IW  (1)
                                                                                                        0)
                                                                                                        0)

                                                                                                        !u
                                                                                                        3
                                                                                                        U
                                                                                                                                   i"0 ^"i
                                                                                                                                   c   -
                                                                                                                                   C 0^
                                                                                                                               C  
^.i
3
O

C
O
•H
4J
Id
^1
3
Q
1
3
O
rH
t-
erif ication
^>

T3
C
id

^^
id
•*--•

c
o

4J
id

_Q
• H
r-H
id
O
Susquehianna
&
o
c
id
^
ca

4-1
CO
CU
3


••
cn
T3
0
•1-1
±4
u
0,
cu
4J
1-1
o

en
s
id
• H
r^
,-H
•H
3

4J
id

VJ
cu

-r^
OS









^-^
•H
g

tr
en

(N
ca

«.
in
—
                                                                                                                                                    3
                                                                                                                                                    C7>
                                                                                      -53-

-------
!
/
s ~ n
7 « / -
P 11-
01 '/
//
:
23 // '
5" // -
ill / -
Ul -
-1
_ 4
§ 0
0
I> 0
                   n
                   u.
                   o
                                                                                              2
                   <
                   a
                                                                                                         C
                                                                                                         O
                                                                                                        •rt -P
                                                                                                        JJ  (0
                                                                                                         tO
                                                                                                         o  uj --^

                                                                                                        ^  >  s
                                                                                                           K  &
                                                                                                               w
                                                                                                           (0
                                                                                                           c o
                                                                                                           c   C
 S-!  10
 3
O —
    a
 c --'
 o
•H  C
•U  0
 10 -H

 3  fl
Q  ^
 I  £1

 O ^H
r-l  HJ
Cu U
                                                                                                     r-l
                                                                                                     (N
                                                                                                     0)
                                                                                                     M
                                                                                                     3
                                
^1
3
O

C
0

4J
(0
u
3
Q
1
3
0
r-l
&4
' — '
C
o
•H
4J
•0
0
•H
4-1
'£
0)
>

^
c
(0

x~.
(0
>— '

c
o
•H
Jj
tO

ffl
.^
—1
iO

4J
o

3
0)
2

4J
(0
^
4)

• H
C£

(0
4J
fO
•H
C
3
1-3


.•
tn

o
• H
!-l
m
cu
















^-^
•H
S

cr


^*
LO
n
*
ro
**^>

^
04
                                                                                                    OJ
                                                                                                    U
                                                                                                    3
MOId
ooo
v   in  cj

   jo
                                            -54-

-------
I
1
f
1
1

1


Table 6
Correlation Coefficients for Weekly Streamf lows at Major




River Gages

CORRELATION COEFFICIENT
I

1

1
a
i§
I
••
I

I
1
(WEEKLY
STREAMFLOWS)
CALIBRATION VERIFICATION
STREAMGAGE PERIOD
James River at Cartersville, VA ' 0.94
Potomac River at Hancock, MD 0.91
Shenandoah River at Millville, WV 0.92
Potomac River near Washington, D.C. 0.95
Rappahannock River near Fredericksburg, VA 0.85
West Branch Susguehanna River at Williamsport, PA 0.85
Susguehanna River at Wilkes-Barre, PA 0.88
Juniata River at Newport, PA . 0.86
Susguehanna River at Marietta, PA 0.92
-

PERIOD
0.84
0.78
0.74
0.80
0.81
0.71
0.74
0.81
0.80


-55-

-------
based upon weekly streamflows because weekly-to-monthly flows were of
greatest interest for the pollutant transport study.  As may be seen in
Table 6, correlation coefficients for the calibration period are somewhat
higher than for the verification period, although coefficients for both
periods are within acceptable ranges.  The gages at the mouths of the three
largest river basins are characterized by the highest correlation
coefficients.

    Comparisons of simulated and observed hydrographs for Hurricane Agnes
(late June 1972) and Tropical Storm Eloise (late September 1975) reveal that
the Model appears to handle these relatively infrequent events rather well
at some gages.  Sample comparisons are presented in Figure 22 for the gages
on the West Branch Susquehanna River at Williamsport, Pennsylvania and on
the Potomac River at Hancock, Maryland.  It is felt that the River Basin
Model handled infrequent storm events better than some earlier models (1,27)
of smaller watersheds in large part because the rainfall during these storms
tends to be rather uniformly distributed over sub-basins with areas on the
order of 2,000  sq mi.

    In summary, both calibration and verification results are very good
particularly, in light of the very large rainfall segments (e.g.,
approximately 2,100 sq mi on the average).  Most of the remaining error can
probably be attributed to factors such as frozen ground conditions which
were not explicitly represented by the Model and errors in the mean segment
rainfall record due to localized rainstorms, low raingage densities in some
areas, missing  rainfall records, and freezing conditions (34).

Calibration/Verification of Nonpoint Pollution Loading Submodel

    Introduction.  Nonpoint pollution loading factors were developed for
urban and rural-agricultural land use categories by calibrating the nonpoint
pollution loading submodel with monitoring data collected by intensive test
watershed studies.  Approximately 70 years of nonpoint pollution monitoring
data collected  at more than 50 test watersheds was analyzed to develop the
nonpoint pollution loading factors used  in the Basin Model.

    A total of  27 test watersheds suitable for nonpoint pollution loading
characterizations were monitored from late 1979 through mid-1981 under a
$2.5 million  study funded by the EPA Chesapeake Bay Program  (EPA/CBP).
These EPA/CBP testing sites were distributed among the Occoquan River Basin
(6 sites) in  northern Virginia, the Pequea Creek Watershed (3 sites) in
southern Pennsylvania, the Ware River Basin  (4 sites) in southeastern
Virginia, and the Patuxent (5 sites) and Chester River  (9 sites) basins  in
Maryland.  As may be seen in Figure 23,  the  test watersheds were located  in
Coastal Plain and Piedmont river basins  in the vicinity of the Bay's
estuarine system.  In the majority of cases, the test watershed sites
covered only  one land use category.  Nonpoint pollution loading data
collected at  the EPA/CBP test watershed  sites was used  to develop loading
factors for the following rural-agricultural land use categories:
forestland, pasture land, high tillage cropland, and  low tillage cropland.
                                    -56-

-------
 I
 I
 f
 I
 I
 I
 I
 I
 I
I
1
I
I
                                             -57-

-------
                                   SUSQUEHANNA
                                       RIVER
                        POTOMAC
                          RIVER
"CHESAPEAKE
      BAY
Figure 23.  Map Showing  Location  of River Basins Subjected to EPA/C3P Test
           Watershed Studies:  Pequea Creek  (A), Patuxent River  (B),
           Occoquan River (C), Ware River (D), and Chester River  (E)
                             -58-

-------
    Urban nonpoint pollution loading factors developed  by  a  previous
twelve-month  (1976-1977) test watershed monitoring  study in  northern
Virginia were the principal sources of data for characterizing  total  P and
total N loadings from different urban land use categories  (_7,.8,.9) •  The
urban loading factors developed from northern Virginia  testing  site data
were verified at three urban test watersheds covered  by the  EPA/CBP
monitoring study.  In addition, nonpoint pollution  loading data collected at
10 urban testing sites in the Metropolitan Washington region under the
1980-1981 Nationwide Urban Runoff Program (NURP) was  used  to refine the
distribution between organic and inorganic nutrient fractions for nitrogen
and phosphorus loadings in urban runoff.

    Each test watershed site was equipped with a flowmeter,  automatic
sampler, and was served by a continuous recording raingage located within or
nearby the site.  Either a natural (e.g., ephemeral stream)  or  artificial
(e.g., H-flume, Parshall flume, storm sewer) drainage control was typically
used to establish stage-discharge relationships at  the  outlet of each
watershed.  The sampling interval was generally automatically initiated by
the flowmeter at the start of a runoff event.  For  studies which relied upon
flow-composite sampling methods, the flowmeter activated the sampler  at
preselected increments of runoff volume.  For studies which  relied upon
sequential-discrete sampling methods (i.e., collection  of discrete samples
at numerous points along the runoff hydrograph), the  flowmeter  activated the
sampler at preselected stage increments.  At the larger sites which
exhibited dry weather flow, baseflow samples were periodically  collected.
Runoff and baseflow samples were analyzed for plant nutrients and total
suspended solids, with periodic analyses for organics.

    At the time the EPA/CBP test watershed monitoring studies were designed
and implemented, a work program for data management and model calibration
had not yet been developed, although a follow-up modeling study was under
consideration.  The absence of a modeling study work  program at the start of
the monitoring studies tended to significantly complicate data
reduction/management activities during the modeling effort and  to reduce the
amount of monitoring data that was suitable for model calibration studies.
A later section of this chapter discusses specific  problems  and
recommendations for coordinating future test watershed  monitoring and
modeling investigations.

    Data Reduction/Management Requirements for Model  Calibration.  This
section focusses on the data reduction/management requirements  for the
EPA/CBP testing sites.  Requirements for the earlier  studies of urban  test
watersheds in northern Virginia and the metropolitan Washington region  were
similar and are summarized elsewhere (REFS).

    The test watershed monitoring investigators reduced the  data required to
characterize the runoff pollution loadings from each  runoff  and baseflow
sample.  Investigators who relied upon flow-composite sampling  techniques
reported mean flow rate and mean concentration data for the  runoff or
baseflow sampling interval, while investigators who relied upon
                                    -59-

-------
sequential-discrete sampling methods reported instantaneous flow rate and
instantaneous concentration for each runoff or baseflow sample.  All other
meteorologic and hydrologic data required for model calibration typically
had to be reduced by the modeling investigator.

    Since the test watersheds were relatively small, a 15-minute time step
was required for the continuous rainfall record to ensure that the  rainfall
interval was not significantly greater than the watershed's time of
concentration.  Drum raingage stripcharts used at the Pequea Creek  and Ware
River basin sites were manually reduced, while other rainfall stripcharts
were reduced with a Numonics digitizer equipped with software to create
files with the appropriate time-step.  Approximately 1.0-1.5 yrs of
stripchart record was reduced for the modeling studies.  Since the
monitoring investigators were only required to report and analyze water
quality monitoring data, raingage maintenance appeared to receive the lowest
priority of all equipment checks and considerable gaps were found in the
onsite records.  During periods when the sampling station was shut  down due
to a breakdown of either the flowmeter or automatic sampler, the raingage
was sometimes shut down until water quality sampling was resumed.   Since a
continuous simulation model requires rainfall records covering the  periods
between monitored runoff events in order to calculate antecedent soil
moisture and sediment accumulations for each monitored storm, missing
rainfall records for periods when the onsite raingage was shut down had to
be constructed from a nearby raingage.  Separate software was developed to
create a continuous rainfall file for input to the NFS model.

    The other meteorologic input file required for the NFS model calibration
study is a time series of daily potential evapotranspiration for the test
watershed monitoring period.  A continuous potential ET record was  developed
for each river based upon meteorologic data at the nearest NWS station.

    Since it is advisable to use a long-term runoff record to calibrate a
continuous simulation model, the reported flow records for runoff events
with water quality samples were expanded to include all flow records
collected during the monitoring period.  Although all monitoring sites were
equipped with continuous recording flowmeters, the Pequea Creek test
watersheds were the only sites with a complete daily streamflow record.  The
flow records at some sites were restricted to those runoff events with water
quality samples.  Stripcharts with additional flow records were reduced with
the same digitizer software used to construct rainfall records.

    Since the quality control programs of most monitoring investigators
concentrated on laboratory analyses, the modeling investigator was  required
to perform most of the quality assurance checks on the hydrometeorologic
datasets.  The digitized rainfall and runoff records were integrated for
each storm event and rainfall and runoff volumes were compared to identify
potential water balance problems due to such factors as backwater,  an
incorrect stage-discharge relationship or an error in the reported  flowmeter
setting.  Flow stripcharts for monitored storms were also checked for
"flat-top" hydrographs which indicated that the maximum stage exceeded the
                                    -60-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
full-scale setting of the flowmeter.  In cases of spurious  rainfall/runoff
ratios or "flat-top" hydrographs, the runoff and water  quality data  were
deleted from the observed dataset for model calibration.

    Software was developed to calculate total storm loads from datasets with
mean or instantaneous flow rates and concentrations for each  runoff  sample.
For test watersheds with flow-composite samples, the product  of mean flow
rate and mean concentration was multiplied by the storm duration to
calculate total load.  For test watersheds with sequential  discrete  samples,
the time series of instantaneous loading rates (i.e., product of
instantaneous flow rate and concentration)  was numerically  integrated
between the first and last sample time to calculate total loads for  each
storm.  Mean storm concentrations were also calculated for  the sequential
discrete sampling datasets.  Dry weather flow concentration statistics were
calculated for test watersheds with significant amounts of  baseflow.

    Land use and drainage area data was typically based upon  maps, drawings,
and tables compiled by the monitoring investigator, which were checked
through site inspections and with available aerial photographs and
topographic maps.  At one test watershed where a check of rainfall/runoff
ratios revealed a serious water balance problem, the authors  performed a
plane-table survey which produced a significant increase in the drainage
area and more reasonable rainfall/runoff relationships.  For  urban test
watersheds, percent imperviousness was determined by planimetering aerial
photographs or site plans.

    Soils characteristics for each test watershed were derived from  county
soil series maps and surveys.  The predominant hydrologic soil group was
determined and average values of permeability, total water  holding capacity,
and erodibility were calculated for use in deriving hydrologic and nonpoint
pollution model parameters.

    Average overland flow slope was typically reported by the monitoring
investigator.  Reported values were checked with 1:24,000 scale maps of the
test watershed and surrounding areas.

    For cropland sites, data on monthly vegetative cover and  the timing and
extent of tillage activities are required to accurately model soil loss.
Based upon discussions with local SCS staff and information on the timing
and extent of harvest and tillage operations at the test watershed,  a time
series of monthly ground cover was derived for each cropland  site.   Since
the NFS model does not permit the user to alter ground cover  time series
from year-to-year, two different input datasets were sometimes required to
model cropland watersheds with more than one year of monitoring data due to
changes in harvest and or tillage dates from one year to the  next.  The
monthly ground cover time series and tillage parameter values were refined
during model calibration to help achieve acceptable agreement between
simulated and observed sediment loadings.

    Although the monitoring investigators made an effort to select test
watersheds which included only one land use, finding an acceptable catchment
                                                -61-

-------
with a single land use was not always possible.  For test watersheds with  a
mixed land use pattern, aerial photographs of  the test watershed  were
checked to determine the character and extent  of any secondary land uses.
In cases where nonpoint pollution loading rates from the secondary land  use
were significantly higher than the loading rates for the primary  land  use,
the complexity of the model calibration study  was significantly greater.   An
example is the forestland test watershed (Site #2) in the Pequea  Creek basin
which included a high-tillage cropping site  (approximately 13% of drainage
area)  that contributed the majority of the nonpoint pollution loads during
major storm events.  In such a case, loading factors for the secondary land
use had to be set based on model calibration for a similar test watershed
prior to model calibration at the mixed land use site.  Also, the monitoring
dataset had to be screened to identify those storm events (e.g.,  minor
runoff events in the case of Peguea #2) which  are least likely to be
characterized by major nonpoint pollution loading contributions from the
secondary land use.

    EPA/GBP Test Watershed Model Calibration:  Site Selection.  Not all  of
the test watersheds were characterized by sufficient land use homogeneity
and hydrometeorologic data to permit NFS model calibration.  Sufficient
hydrometeorologic and water quality data was available to calibrate the NFS
model to 11 of the 12 acceptable sites in the  Occoquan River, Ware River,
and Pequea Creek basins.  A summary of the sites with monitoring  records
suitable for model calibration is shown in Table 7.  Due to the late start
of the Maryland test watershed studies, insufficient hydrometeorologic data
was available for model calibration of the single land use sites  in the
Patuxent and Chester basin sites.  Therefore,  only statistical analyses
could be performed on the Maryland testing site database prior to the
initiation of Basin-Model calibration/verification runs in the Fall of
1981.  Following completion of Basin Model production runs in the Spring of
1982, nonpoint pollution loading factors calibrated for the Virginia and
Pennsylvania testing sites were successfully verified at two Maryland
testing sites:

    1.   Browntown Rd. Low Tillage Cropland  (Chester River Basin) :  331  ac

    2.   Chestertown B Single Family Residential/Commercial (Chester River
         Basin):  49 ac

    For most EPA/CBP test watersheds, the nonpoint pollution monitoring
records were not extensive enough to permit  subdividing the dataset into
separate calibration/verification periods.  Consequently, the entire test
watershed monitoring dataset was used for NPS  model calibration.  The
calibrated NPS loading factors were verified through applications to mixed
land use river basins in the Chesapeake Bay drainage area.  Even  in the
absence of verification in the Chesapeake Bay  river basins, it is felt that
the risk of producing biased calibration results from the test watershed
modeling studies are significantly reduced by  the use of continuous
simulation calibration techniques which involve long-term simulations  and
parameter adjustments that are not keyed to  individual storm events.
                                    -62-

-------
I
I
I
I
I
I
I
I
I
I
I
I























r-

<1J
r-l
J2
nJ
E-i




















^^
U
18

r-
^*
CN
II
(8
O
0
,_4


tn
-U
i-t
3
W
i
Cn
0
i-t
0
V4
TJ
>,
X
T3
C
(8
0)
o
•H
4J
cn
•H
U
0)
.U
0
18
i-i
ia
r*
U
Watershed

4J
U)
•» i ^^ ^^

c
CO
2
2
M
Cu
J3
Cd
CC
8<
CO
SH
M
HH
rt
a



O cn TJ-
N r- i ii ii p. i | co i
CK * | li ii * 1 i i
1-H • 1 II || * 1 1 •

o o o



W CO i^Q
§ *: i ! ! ! i ^ ! i * !
$ ° 00




<8 £l O
2 CN| || || CNII rri
* cn i it ii CNII r~- I
T CN "*




CN
CO
g
£
co
0 S
g, <
g l
M
Z
o
2

co cn co cn in " o CN in r» COCN
coon cn cn cn cn vo cn cn cn cn
s >. . . .. ... .,
oo oo oo ooo oo



d IOCN com i-i r* OrHtn voo
i< r»r> cno coo r^i-irH coco
^ .. «• .« ••• ..
3 OO Oi-l O i-l O^r-( OO
0




_
, ^ r»
» if>p« cor»» iriirt cor^cn CN i— i
"1 H






a
S3 I
« c.
rti (0 • — i (—1 r-i
~
^
1 CNCN ^OCO fOCO OCOI" CNCN
} . . . . . . ... . .
4 mvo ^oui i-ico couii~- r^to
i-l rH CNCN f!,-l (Ni*.^ ^j.
--I i-l — •




g
M
CO
U
co
2
Q
5




Q
2 a
3 2
a, J
O 04
CS O 0
O ft! CN ^-4 r-t (ft CTl
O 3» 3fc 3» 5fc 3«s
Cd rn CN rr
O 3O CdZZ ZZ 3C Z J 3C
3r^o<<: «;< «;co< m
<3* . . «* T»
*CM,> W^JCT! > ><
a) .-o .a
s s s
nj -a o
                                              -63-

-------
    EPA/CBP Test Watershed Model Calibration;  Hydrology Results.  Due to
the relatively small size of the test watersheds, the majority of the
subsurface flows were often not detectable at the monitoring stations and
therefore, baseflow and interflow components of runoff had to be suppressed
in hydrologic model calibrations for most sites.  Typically, baseflow and
interflow were only included in models of forested watersheds where the dry
weather flow component represented a significant fraction of monitored flows.

    For most test watersheds, hydrologic calibration focused on achieving
acceptable agreement between simulated and observed storm volumes.  Each
test watershed model was iteratively executed with a continuous rainfall
record which bracketed the 1-2 year monitoring period and agreement with
monitored flows checked for each model parameter set.  A 3-6 month
antecedent rainfall period was used for most test watersheds to minimize the
impacts of the assumed soil moisture conditions at the start of the
simulation period.  Scatterplots and simple linear regressions of simulated
and observed runoff volumes were generated for each calibration run to guide
parameter adjustments.  Sample comparisons of the final simulated and
observed  runoff volumes for monitored storm events are shown in Figure 24.
Based on  the goodness-of-fit regression statistics for monitored storms
presented in Table 7, it was concluded that acceptable hydrology calibration
had been  achieved at the 11 test watersheds.  For the three Pequea Creek
test watersheds, continuous daily flow records were also available for
calibration.  As shown in Table 7, acceptable regression statistics were
achieved  for all three Pequea Creek test watersheds, with the lower slope
term for  the forested watershed (Pequea #2) probably due in large part to
the underlying Conestoga Valley limestone formations which appeared to
contribute subsurface flows that originated outside the drainage area.

    EPA/CBP Test Watershed Model Calibration;  Nonpoint Pollution Results.
After achieving an acceptable hydrologic calibration, nonpoint pollution
loading factors were calibrated for each test watershed by iteratively
executing the NPS model with continuous meteorologic records for the entire
monitoring period and checking model projections for monitored storms.  The
calibration of nonpoint pollution loading factors for total P and total N
focused on the agreement of simulated loadings with monitored storms which
exhibited acceptable hydrologic simulations.  Goodness-of-fit evaluations
were based upon conventional and nonparametric statistical analyses.

    For urban land uses, daily pollutant accumulation rates developed by the
1976-1977 northern Virginia study (J3,^)  for pervious and impervious
fractions were tested with the Chesapeake Bay Program monitoring data.  The
calibration techniques used to derive separate loading factors for the
pervious  and impervious fractions are described below.  These urban loading
factors were held constant in the models for Pequea #4 and Ware #5 to see
how well  they represented loadings in different regions under different
meteorologic conditions.  As was the case in the previous urban modeling
study, the daily pollutant accumulation rates were held constant from month
to month.
                                    -64-

-------
 I
 I
 f
 I
 I
 I
 I
 I
 I
 fc
 I
 I
I
I
  a
  z
   o
 2 O
      no  awmoA
-J (0
a. (o
O CJ
a: —
UJ
-J Z)
-J a
                         08
                               no)  3«nioA
                                                      aaivinwis
 TJ
  0)
  ^
  O
 J-)
 •H
  C
  0


  S-i
  0
 H-l

 tn
 0)
 O
 M-l
 0
    n
    03
    3
                                                                                          0)  CU
 tn -H
 A =*=
 O
    C
 T3  ITS
 C  3
 (0  cr
    0
 T3  O
 
 3 CN
 S =*
 •H
 ca c
   H3
 <4J 3
 o cr
   o
 w u
 c o
 o o
 en  ..
 0)  tn
 ij jj
 iTi  C
 01  
   W
 0)
-i  e
 a,  ^
 e  o
 (0 JJ
W OQ
                                                                                         
-------
    For cropland test watersheds, monthly variations in sediment potency
factors were required to account for such factors as fertilizer/manure
applications and crop harvest, with the higher potencies generally
associated with the months characterized by highest percentages of
vegetative cover and vice versa.  The monthly distribution of potency
factors is established during model calibration by deriving the upper limit
for the summer months of high ground cover and the lower limit for the
winter months of low ground cover.  For forestland test watersheds, monthly
variations in sediment potency factors and daily pollutant accumulation
rates were required to account for the variations in ground cover and leaf
litter.  Table 8 illustrates the relationship between monthly potency factor
and monthly ground cover for cropland and forest land uses.  For pasture
test watersheds, monthly variations in sediment potency factors were
generally not required to achieve an acceptable calibration.

    The conventional goodness-of-fit evaluations included scatterplots and
linear regressions of simulated and observed storm loads and comparisons of
simulated and observed volume-weighted mean concentrations for the entire
monitoring period.   The simulated volume-weighted mean concentration was
calculated by summing the loads and runoff volumes for all storm events
within the simulation period, including storms which were not covered by the
monitoring study.  The protocol for NPS loading factor adjustment after each
calibration placed greater emphasis on the agreement of volume-weighted mean
concentrations, since long-term loading trends were felt to provide the best
indication of the need for and direction of further loading factor
adjustments.  In other words, whenever a parameter adjustment decision
involved choosing between improving volume-weighted mean concentration vs.
improving the storm load scatterplots and linear regression statistics, the
former usually governed.

    Comparisons of simulated and observed volume-weighted mean
concentrations for the calibrated NPS loading factors are shown in Table 9.
As may be seen, the ratios of simulated to observed mean concentrations
typically fell within the range 0.75-1.25 which is comparable to the typical
errors inherent in hydrometeorologic gaging and laboratory analyses
(JLl'.lli'.-lJi) •  Thus, Table 9 indicates that the calibrated NPS loading factors
provide a good representation of long-term nonpoint pollution loads per unit
volume during the monitoring period.

    Some of the better scatterplots and regression results for storm loads
are shown in Figure 25.  As may be seen, agreement between simulated and
observed storm loads was quite good for selected sites.  However, regression
statistics for several other test watersheds were insufficient to
demonstrate goodness-of-fit for storm loads.  It is felt that much of the
difficulty in achieving acceptable regression statistics for storm load
comparisons can be attributed to the lower power of conventional normal
statistics for evaluations of small sample sizes characterized by skewed
(i.e., non-normal) distributions.
                                    -66-

-------
«•  a
f
' \.
-v/-f

-• "v
t
*•"'-'
-
-—

-^
c(.


f-ft
< Oi
i -
< j~-
a
o
< .
1 ^
8 _ .
c,
X
<
a
1 ^
&3 o-
* r-^j
1
O
p —
-,jr
O —
a --,
j
o<
<
s^
2

< *
h
2 ._

O vj?
g
-1 c^.
Z
o
i t-

Q ,—.-7
Z
I —

1 "*",
t '0














y
o



>
o

f"
°

Oi
Cd



=
2




J
o




z
i



X
g



&
(3*
<





OS
1


a

hi


z
i?


K
a

S
<
04
\
M
«
3
a
z




••»
•
en


o
o

o
o

o
o


o
o




o
* o




o
o
*H


o
o




o
o
••H




r*
•
0*


1-4
en
en

•H

S




~
»rf

a;
§
8
a
z
S 1
w o
2 -H
rf*
VO
o
iH


vO
O
i-H
VO
O
H
00
CO
•-4

co
CO
^H



00
0

a>
o
ON
O
O
IT)
r-
*H

»n
r»
»-H



in
r-
<-H



in
r-
<~4


^
a>
o



m
01
o




a\
v\
o


a\
»

a

w
A
0
^
i
I

g
2

|
S
£
Q
Cd
CQ
m

^-j
o
o


,-4
O
0
--(
o
0
(N
O
O

0*
O
o



CN
o
o



(N
0
0


^
o
o



•H
0
o




•H
O
O


o

0

,H
O
0
a*
Z
i

X
o


2
g
i
a
w
^

vO
O
o


M3
0
0
VO
O
O
>H
rH
O

>H
~4
0



'F-*
*H
o



^H
f-(
O


vO
O
0



vO
O
0




VO
O
O


VO
O

a

VO
O
O
„.,
z
1
w
1
EH
Z

X
o
z

1
H
Z

Q
01
U1

U1
i-H
O


in
^H
O
in
^-4
0
VO
(N
O

VD

a«
1

X
o
z

2
EH
Z

g
w
«

^
r-|
O


^*
pH
O
^"
pH
O
V
(M
O

^
CN
O



^
CN
0



•»
CM
0


^
iH
0



^*
rH
0





i-H
O


"V
rH

O

^
i-H
0
dP
a-
i
U
z
1
o

X
u
z

i
g
i
Q
u
CO
r-

-4 ^**
O
O


'•H
O
O
iH
O
O
fN
0
O


en ni
m

o **
en f*i
m



in «r
CO f>
m



O fN
in m
CN


O fN .
CN m
CN
*


0 (N
m
CN




O CN
i*l
CN


O ^O
r^-

_H

O V0
r*
rH

2
1

CO
CN



cn
CD
CN



fN
O
CN


(N
0
(N



rN
o
CN




fN
O
fN


fN
in

*H

CN
in
^
«.
z
i
CJ
z
a:
O

>>
O
z

§
g
i
a
«
ro

1*1 r-( O
O 
O O O


m ^H O
0 fN ro
000
<*! — 1 Q
o fN m
o o c
VO Cn f*-
o f*i in
o o o

vo 01 n-
o <*> -n
o o o



vo en r~

000



•V vO O
o CN *r
o o o


V VO O
O CN ^
o o o



f VO 0
O CN T
o o o




•V \O O
O fN V
o o o


M " ^ 0
o CN m

O O O

f*> cH O
O CN <-n
o o o
OP o*»
z z —
1 1
5 | |
1 s i

XXX
0 0 O
Z z Z
y y w
2 i i
g g s
y tij cd
Q a a
W U Cd
WWW
^ in vo

^H crv
fN O
0 0


F-I cr>
fN O
0 0
~* en
CN O
0 0
o r»
•V -H
o' o

o r*
rr »H
0 0



0 f--
T f-t
o o



CO fN
CN <-<
0 0


CD fN
CN -- 1
Z O
C3 E?
0 0

X X
U CJ
z z

E-. £•>
2 2
g g
dJ C>3
a a
QJ CO
M 01
r- co

                           -67-

-------
r--\
^r    in   **
                     os    m
                     CM    CN


                     o    o












•3
a
u _c
a jj
< c
H 0
U































S <-l --
Q ^?" ^
X
a

uOr z
*; /-> -j ^
w '
u
s
o
B* »
o *  M
2^-- £
^~v
H4
2
S ~"~ 5
^






QC

H
g
s
1
«
a
z




os ^ tn
OS ^ vtf
n ^H



OS i-1 *
os (N m
m ^




rM ^H ^«
os fN m
fl rH



tn rs* IH
co m 04
 OS
\C O OS

 ai a,
3 8 fi fi
itf Q U Ci3
M S Q Q

o w to
u
c-t
a


.
m
o




2
o



o

o
CO
o
0

CO
o

o

CO

o
Of

z
1

M
z

<
t*
CJ
1
g
g
s
Q

CO
*
in
10
i-4



in
^




in
,H



»H
(N
•H
OS
OS
0

Os
OS

O

OS
OS
0
MB

Z
t
u
E-t
E-*

Z
,.
CJ
z
u
2
z

a
u
CO
m
OS
CM
O



fN
O




(N
O



iH
r«i
o
r*

o

p*
•H

O


^
O

^
«. -1

a*
i


H
.*
U
z
I
g
a
Q
oJ
CO
y3
CN
o



PJ
0




CN
0



P*

O

"-*
0

21

o


"-<
o
^

On

o

z
a:
O
..
g
z
i
g
i
a

U]
"
O
o



in
o
o




in
o
o



•V
o
o
m
0
o

m
a

o


0
o

of


a*
i
s
a.
o
..
S
§
2
g

a

U)
33
                                                                 g
                                                                 o

                                                                 a
                                                                 z
                                                                      X
                                                                      o
                                                                           o
                                                                           z
                                                                           u)
                                                              *H    fN



                                                              O    O
                                                                      in    **    **
                                                                      r*    -H    O
                                                          U





                                                          1
O
z
                                                    2   2
                                                                                                                                   I

                                                                                                                                   I
          o



          o
                                                         -H   -«    (N    ^H



                                                         O   O    O    O
                                                                                                                                    8

                                                                                                                                    I

                                                                                                                                    I

                                                                                                                                    I
                                                                                                     o


                                                                                                     o
dV

Z
1
J
<
£-*
Z
1
O
M
z
I
o
z
t
<
HH
2
|
<
Z
t
01
£H
S
2
e*
5
a»

a,
i
j
1
e->
a.
1
CJ

z
I
o
00

0.

O
p
a;
O
u
z
                                          -68-

-------
I
I
f
I
I
I
 I
 I
 I
 I
 I
 I
  I
a
s
      I
                o '
                o

                o
      Cu
      U
      CO
























?
00 CU
3
U C
<
a
as.
A

a
v
<
z
M


a

>i
s
b]

<
[I]
a.

CO
H
9
o
8

a
>-<
i
Q.

g
<


o> o
O» i
£

o\ o
Ol PJ
o
o


m
"H
o
0





en
rH
0

0


m
»H
0

0


rO
-H
O
o


Ol
^H
0
o

0




m

o
o

0

ffl
•H
O
0

0

o\
»H
o
0
o

>H
in
o
o

o




p-4
Ul
O
O

o

M
in
o
o

o

^H
l/l
o
o
d

                           o
                           o
o

o
o

o







j
g
CJ
a

en
3


^
3
Cu
3
CO



a
I
g
^,
5
i

0* <

« 5


u to
Q 8
z »-•
O a!
cj a
!— < (N
0
o






•0

o
tl
X.


— '
z
1
^
s

o
o

B
S
g
f_,
5 ">;
If
O (0


S 5

<
z
CO 1
g y
> §
as u
Cd 3S
0» O
n



S
S
g
EJ
5
D
O
U


5


w
8
>
a:
v
o
o

0*




>1
u
TJ

en
13


Z
1
<
Z
1




B

g
E"
3
3
O

<

U
a:

CO
3
>
^
U
a.
m
o
o

o




"£
u

X
w



z
1
u
&-

c*




«
I
g
H
5
13
O
<

3
s
<
Ul

^
a:
jj
Oi
ya
a
o

o





i
TJ
X
o
«i
X
in
a
•H
"*
?
^
g

0
o

o
w
1
g
E^
5 >;
3 "3
CJ rtj
< ^
to



04
g y
« z
> <;
a; u
S o
r^
O
o

d
M
S
§
^
<
i?
u -a
u x
< o
ra

2 ^
< -H
CO
811
Si
a: .-
u ce
o. o
CO
                                                        -69-

-------
    o
    id
    a
    >
    o
                     o

                     o
                         o

                         o
          o

          o
i
z
a, i  u

si  S
«s  <
*8  s
              o

              o
              o

              o
                                                     •I
o
I
1



_
c
o
u
"^
J
M
w
a
en
2


^
£
•f
Cu


CJ
O

H
2
g
C-i
3
£
O >^
cj ro

< CJ
W ^

in -i
8-

> i
a j
U 4
0. EH

>-" EH
O
O
cu
fH
2
g
H
«:
8s
t3 UJ
O TJ
•£ U

§ cl

en
3 f
M CJ

as -z.
M <
Q* CJ

S o
0
o
a
g
g .
g*
*
s >i
5 -o
«c o
rtj
cd ^
< -i

tn
3f


a z
w a
a. g
r 5

0
0
Cd
I
g

2
i >
CJ t3
o \
< o
ra
Cd to
< -H

en
8|

> 5-i
ca a!
a, EH

f-» Z
0

Cd
EH
S
g
H
5
g
§ —
m

•£ U


S ~

> 1
n. J
u <
a* H

l-( fH
o
0
'i

§
s

i >
D (0
< 0
a
2 CO
< -H

01



a z
Id 4
£• CJ
Z a
""* O
o
o
a
S
S
f^
3

8^
0 «>
< 13
-x
SO
m
< CO
J

if

> 1
a o
w =
a, e-.
S a;
-* Q
             -70-

-------
I
I
t
I
I
 I
 I
 I
 I
 b
  I
  I
  I
  1
  I
     u

     z
     o
     u
o> i  g
S-S
32
h8
j
a
a.
  i  i
     z
     o
k
c
z
O.
u
•Jt
<:
J
=
0.
as
$
a
s

3
U

2
<
a.
U
en
a
Q
z
5


cn
ffi
os
OS
o\
 rf
a! f;
Cd O
a y

i!
2
§
|

I

1
*
g
>
OS
U
o,
(N
O
O
O
CN
O
O
o
CM
O
O
o
fN
O
o
o
tN
O
O
fN
O
O
0
rn
fN
O
O
o
(N
e
o
o
fN
0
0
O
fN
O
0
o
fN
O
o
o
CN
O
o
0

rtJ
T3

^
tfl
5
z
i

g
X
0

s
MULATl

<

I
n
8
>
a
o
o
o
o
o
r*
o
o
o
o
o
o
0
o
r-
o
o
o
o
o
o
o
o
r-
o
o
o
o
r*
O
o
o
o
r-
o
0
0
0
0
0
o
o
r-
o
o
o
r—
o
o
o
o
r*
0
0

X
o
o
0
r-
o
o
o
o
0
r»
o
o
o
o
0
o
r-
o
o
o
r-4
O
o
0
r-
o
o
o*
0
0
o
o
0
Q
r-
•H
o
0
0
r*
i-H
O
O

1

X
en
J3
Z
"
2
H

2
g
1
3
<

I
<
ui
8

a
CJ
Q.
O
o
o
o
o
o
o
o
o
o
o
o
r*
0
o
o
r*-
0
0
0
o
o
0
r*
o
o
0
r-
o
0
o
o
o
0
r-
c?
o
o
r»
o
o
o

— .

o

X
CO
~H
0.
^
I

2
z
o
MULAT]
3
B

<
<
g
>
ft,
-H
O
o
o
— I
o
0
o
i-H
O
0
o
i-H
0
O
o
o
o
o
0
0
0*
o
o
0
l-t
o
o
o
rH
o
o
0
,-t
o
o
o
,-H
o
o
o
H
o
0

•u
\
u
 2
a. P
M 3S
0. O
                                                           -71-

-------
       ej
       m
       a
                           a

                           o
                                       o

                                       o
   a.

   i
   z
              o

              o
                  O

                  O
              0

              o

              Q
   (J

   8
—  Oi
-  »•
              o
              Ul
              o
   i
      o
      t*l
      o

      d
                 §
I!
u <
                     ul
                     E-
                     3 *I


                     < U




                     CE -Q
             i>
             0 'D

             < O
                         w
    8 >
    O «I
    <^
    <£ U
              r  a?  3?  s
                 04 O
                 Z X
                 n O
                 -72-
^
K <
u 2
T S
5 5
                                  i >
^^?
                                  « Z
                                  ss
    < 0!
    W -i

    §;
    « o
    Sg
    sg

-------
I
I
f
I
I
I
I
I
 I
 k
 I
 I
 I
 I
  I
                                  Table 9

Comparison of  Simulated and Observed Volume-Weighted Mean Concentrations
                            SEDIMENT
TOTAL P
TOTAL H
SITE (UNO USE)
PEQUEA 3 (H.T. CROP)
WARE 7 (H.T. CROP)
OCC. 2 (L.T. CROP)
OCC. 10 (L.T. CROP)
OCC. 9 (FOREST)
PEQUEA 2 (FOREST)
WARE 8 (FOREST)
OCC. 1 (PASTURE)
OCC. 5 (PASTURE)
PEQUEA 4 (RES ID.)
WARE 5 (RES ID.)
OBS.
N
17
10
16
13
15
21
34
27
11
52
30
SIN. OBS. SIM. OBS. | SIM. OBS.
(HG/g_ (MG/g RATIO (Ms/g (Mc/g RATIO ' (MG/U (M_c/g RATIO
783 829 0.94 4.53 4.70 0.96 18.7 19.2 0.98
272 222 1.23 0.70 0.62 1.13 ,' 1.5 1.30 1.20
370 361 1.02 1.96 1.67 1.17 6.8 5.5 1.03
138 121 1.14 0.47. 0.40 1.18 j 4,6 3.8 1.22
83 70 1.19 0.13 0.13 1.0 1.1 0.9 1.22
99 159 0.58 - 0.1 0.13 0.74 3,8 3.5 1.05
64 71 0.90 0.05 0.06 0.83 0.37 0.4 0.93
561 670 0.84 0.94 -1.12 0.84 , 5.3 6.2 0.86
166 145 1.14 0.43 0.38 1.13 ; 2.5 2.2 1.15
115 194 0.56 0.24 0.30 0.80 ' 1.8 2.4 0.75
50 38 1.32 0.12 0.10 - 1.23 i 0.95 0.70 1.36
  f
                                                -73-

-------
                                                          —   o
                                                    aaivmwis
                                                                      in

                                                                      ±4
                                                                      o
                                                                      4J
                                                                      w
                                                                      0
                                                                      4J
                                                                      c
                                                                      o
                                                                      in
                                                                      01
                                                                      c
                                                                      <0

                                                                      2

                                                                      TS
                                                                      ID
                 0)
                 in
                                                                      c
                                                                      10
                                                                      •a
                                                                      0)
O


< 2
S
s
•- 3
  O
IS

II
                 e =*=
                •H
                W  (0

                in  3
                 0  0<
                   0)
                 ffl 3<
                 C
                 O TJ
                •H  C
                 in  10
                 en
                 a; o

                 IT* =st=

                QH  C

                 a)  3


                Is
                 .0  O
                w O


                m
                (N

                 0)

                 3

                •H
                                  -5   o  10  o  «  e
                                  Si   V  «  «r>  C\J  c

                                  p     (*qi)  avon
                                  LU
5  2
                               -74-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
    To provide a visual check of agreement between simulated  and observed
frequency distributions, box and whisker plots were developed for  the
simulated and observed storm load datasets for each test watershed.  As
illustrated in Figure 26, the box and whisker plot displays the 25th,  50th,
and 75th percentile values in the frequency distribution as well as  the
upper and lower extremes.  The departure of the 25th and 75th percentile
lines from the median line provides an indication of skewness of the
distribution.  The mean value is sometimes plotted on  the box and  whisker
diagram to highlight departure from the median value and the  non-normality
of the distribution.  Box and whisker plots for the calibrated NFS loading
factors are shown in Figures 27, 28, and 29 for cropland, forest/pasture,
and urban land uses, respectively.  To provide an indication  of
non-normality, the location of the mean in each diagram is indicated by an
"o".  Since it is the driving force for simulated cropland loadings,
sediment data is presented for the cropland test watersheds in Figure  27.

    As was the case with the simulated volume-weighted mean concentrations
reported in Table 9, the simulated box and whisker plots are  based on  all
storms which occurred during the test watershed monitoring period, including
those which were not monitored.  The inclusion of all  storms  tended  to
automatically skew the simulated distribution in the direction of  minor
storms which typically were not monitored in the field due to the  very small
runoff volumes.  This skewness can be attributed to the fact  that  whereas
the mathematical model will calculate runoff volumes and loads from minor
storms, the test watershed monitoring studies relied upon runoff volume
thresholds associated with more significant rainfall events.   Since  similar
runoff volume distributions are required to ensure meaningful statistical
comparisons of storm loading datasets, minor storms were generally deleted
from the simulated dataset prior to the development of the box and whisker
plots shown in Figures 27-29.  The establishment of the cutoff for minor
storms was based upon iterative analyses of box and whisker plots  and
nonparametric statistics for the simulated and observed runoff volume
datasets.  The storm runoff volume thresholds which resulted  in box  and
whisker plots and nonparametric statistics that indicated acceptable
agreement between the simulated and observed runoff volume distributions and
median values are summarized in the "runoff interval" column  of Table  10.  A
check of the total runoff volume and nonpoint pollution load  produced  by
storm events which fall within the specified runoff interval  indicates that
typically 90% or more of the total volume or load produced during  the
monitoring period was generated by these storms.  This check  suggests  that
the minor storms deleted from the simulated and observed datasets  were
relatively insignificant in terms of seasonal or annual loads.

    The runoff intervals shown in Table 10 were used as the basis  for  the
box and whisker plots of simulated and observed datasets in Figures  27-29.
Inspection of these box and whisker plots confirms the highly skewed
distributions of the runoff and nonpoint pollution loading datasets.    The
similarities between the simulated and observed plots, in terms of
distribution and median value, also serves as graphical evidence of model
goodness-of-fit.
                                                -75-

-------
 T UPPER  EXTREME
    UPPER  QUART1LE
     MEDIAN
     LOWER QUARTILE
 -1 LOWER EXTREME
Figure 26.  Configuration of a
        Box and Whisker Plot
    -76-

-------
I
I
f
I
I
I
I
I
 I
 I
 I
 I
 I
 I
3.JO
StTS
X
~ Z.2O
Ul
§,.«
o
1.10
u.
u.
1 0.39
c









'4!



















r









i








f
1










a°03 OS OS OS
PIOUE* 3 WMK 7 OCC. 2 OCC K>
(HTCftOP) (M.T. WOP) IU. CHOP) ICC CROP*
                                   §
                                   3
                                   X
1.3
OJ
a*
OJ
ao


-



4 U



1 <£



1;


r
? 'i


l
j



1 c
r*i rn


•
gj o
OS OS OS OS OS
occ. 3 pea 2 WARE a oec i occ 9
(FOREST) (FOREST) (FOREST) (PASTURE) (PASTURe)
ISO.
G iso.
S iso.
x 9a
Ul
5 60.
Ul
to
nn



. r



1 *
u;









F



i F



i



I
i




i
•

.

^H
KOUCA 3  WMC 7  OCC. 2   OCC K>
      (MX enow (U. awn (u: CHOP*
U.JW
•JO-**
V
 O2O
0
_J
"" O.IS
z
i a 10
o
•-aw









'1



!








I
*l







r

R m
9 T

,





r



<
i
H f-
0
T






'










f
ri
y f









1;
OS OS OS OS OS
oec t pea 2 WARE s occ i occ t
(FOIWST) (FORCST) (FORCST) (PMTUM9 (PASTUHC)
2.4
i
« •••
a)
" I.Z
2
K
O
^™ _ .
on






h






T






r






r

T r

>





I
k






\






hi






J
                                                                (O
                                                                a
    3
 PEOUCA 3  WAKC 7   OCC. 2  OCC O
 (H.I CROP)  (RT. CROP) (LT. CROP) (U CROP)
u









r I




r




F


3 r


1




C

,



1




g (

]
03 OS 03 03
                                   Figure 28. Comparison of
                                               Simulated (S) and
                                               Observed  (O) Box and
                                               Whisker Plots for
                                               Forest and Pasture
                                               Watersheds
        mu* 7  occ z  occ. 10
 (RTCHOP) (H.T. OtOPI (U. OWPt (ti CITOPJ
               Figure 27, Comparison of Simulated  (S)  and
                          Observed  (0)  Box and Whisker Plots
                          for  Cropland  Watersheds
                                                     -77-

-------
au
I
O
tu
§,.9
-1
O
u.
u.
109
=>
tr
nn











c





1 C





1


















r
p

-
.

•



*
1*1
" U
T 9
U.JO
£ 0.30
W 024
0
""" am
z
< 0.12
t-
o
•" aoe
no








4

t








» H







T
1 •!
r 9






4



1
-




» r*

• «

n


.

•j .

»
-
r
    03    03
  PEQUEA4   WARES
(RCSIOENT1AU  (RESIDENTIAL)
    03    03
  PCQUEA4  WARE 9
(ME3IOCNT1AO (BEHIOC?m*U
v/.uo
— 0^)5
O
V,
m
~ O03
a.
< 0.02
O
K OJJI
no





i •
t





i.





1
i?





4
1





H
•
J i
•

• -

    03    03
  PCQUEA4  WARE 9
dieSIOCNTUU (RHSIOCNTUU
Figiire 29.  Comparisons  of Simulated  (S)  and Observed (0) Box and
            Whisker Plots for Single-Family Residential Watersheds
                                   -78-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
                                      Table 10
      Nonparametric Goodness-of-Fit  Statistics  for Test Watershed Model
        Calibration:   Runoff Volumes  (R.O.) and  Total Phosphorus  (TP),
             Total  Nitrogen  (TN), and Sediment  (SED) Loadings  (Ibs)
SITE (LAND USE)

PEQUCA 3 (H.T. CROP)
HARE 7 (H.I. CROP)
OCC. 2 (L.T. CROP)
QCC. 10 (L.I. CROP)
OCC. 9 (FOREST)
PEQUEA 2 (FOREST)
HARE 3 (FOREST)
UCC. 1 (PASTURE)
OCC. 5 (PASTURE)
PEQUEA 4 (RESID.)
HARE 5 (RESID.)
RUNOFF
INTERVAL
(IN.)

> 0.025
(0.01-0.125
* HURR.)
>0.015
>0.025
>0.03
<0.09
(0.075-0.9)
(0.025-0.5)
>0.01
> 0.025
>0.09
sm.
STORfS
28
26
29
19
30
91
20
15
a
36
43
OBS.
STORfS
13
9
6
3
9
16
24
9
8
25
21
TWO-SIDED
K-S TEST WILCOXON RANK SUM
LEVEL OF SIGNIFICANCE
R.Q.
0.20
0.10
>0.20*
XJ.1T
>0.20*
0.42
XJ.20*
X).10*
XK20*
0.94
0.53
TP _
>0.20*
>0.20*
XUO*
>0.11*
?0.20'
0.34
0.10
>0.10*
>0.20*
0.39
0.22
_TN
>0.20*
>0.2(T
0.10
>0.il*
xuo-
0.26
0.10
>O.W
0.20
0.14
0.29
JED_
>0.20*
>0.20*
>0.20*
>O.U*
N/A
ft/A
N/A
N/A
N/A
N/A
N/A
TEST
LEVEL OF SIGNIFICANCE
_R_J._
0.58
0.17
0.71
0.65
0.56
0.30
0.86
0.40
0.37
0.76
0.31
TP_
0.96
0.23
0.91
0.97
0.45
O.U
0.31
0.95
0.54
0.70
0.64
JA
0.65
0.42
0.20
0.94
0.80
0.18
0.13
0.31
0.25
0.48
0.34
SED
0.83
0.34
0.39
0.98
N/A
N/A
H/A
N/A
N/A
N/A
N/A
•EXCEEDS MAXIMUM PROBABILITY VALUE CURRENTLY REPORTED W STATISTICAL TABLES FOR SAffltE SIZES
  f
                                                  -79-

-------
    Nonparametric statistical tests (3^7) were performed for a quantitative
assessment of the goodness-of-fit for the datasets plotted in Figures 6-8.
Nonparametric statistics assume no shape for the population distribution  and
therefore are valid for both normal and skewed distributions.  Consequently,
nonparametric statistical techniques have a much higher power than normal
statistical techniques for analyses of datasets, such as the monitored  storm
load dataset, which are characterized by small sample sizes and skewed
distributions.  The results of two-sided Kolmogorov-Smirnov  (K-S) and
Wilcoxon Rank Sum tests are summarized in Table 10.  The K-S analysis is  a
test for any significant deviation of the simulated distribution from the
observed distribution.  The analysis involves checking the maximum
difference between simulated and observed distributions to determine if it
exceeds a critical value.  Since it is a broad alternative test, the K-S
test has lower power for any specific alternative, such as a difference in
median values.  The Wilcoxon Rank Sum analysis compensates for this
deficiency since it is designed to test for differences in median values,
under the assumption that the simulated and observed distributions may
differ only with respect to this value.  The Wilcoxon Rank Sum test assigns
ranks to the combined dataset of simulated and observed values and
calculates the sum of the data ranks for each dataset.  If the simulated
median value differs significantly from the observed median value, the  sum
of the simulated data ranks will be higher or lower than the sum of the
observed data ranks.

    Based on a 0.05 probability cutoff for the 95% confidence interval, the
level of significance statistics in Table 10 indicate that the simulated
runoff volumes and nutrient loads do not vary significantly from those
monitored in each test watershed.  The high significance levels of the
nonparametric tests summarized in Table 10 meet the primary objective of  a
goodness-of-fit evaluation by indicating a low probability of accepting a
false model as true (Type II error).

    Urban Test Watershed Model Calibration:  Northern Virginia and
Metropolitan Washington NURP Studies.  Sixteen small watersheds  (26 ac
average) characterized by a homeogeneous urban land use pattern were
monitored from June 1976 through May 1977 to develop the required database
for loading factor calibration.  The distribution of monitoring sites among
urban land use classifications was as follows:  2 large-lot single family
(0.5-2.0 DU/AC) residential watersheds; 3 medium density single family
(2.0-8.0 DU/Ac) residential watersheds, 4 townhouse-garden apartment  (8-22
DU/Ac) watersheds, 3 high-rise residential (greater than 22 DU/Ac)
watersheds, 3 shopping center watersheds, and 1 central business district
watershed.  Separate daily pollutant accumulation rates (Ibs/ac/day) were
calibrated for the pervious and impervious fractions of each test watershed,
with the same loading factors found to provide an acceptable representation
of all test watersheds within a particular land use category (_9) .  The
calibrated urban loading factors were found to provide acceptable loading
simulations without altering the factors from month to month.  Five-minute
rainfall records collected within each  test watershed were used for NFS
model calibration.
                                     -80-

-------
I
I
I
I
I
I
I
I
I
 I
 I
 I
 I
 1
     In  order  to  derive  separate  pollutant accumulation rates for pervious
 and  impervious fractions  of  each urban testing site,  observed nonpoint
 pollution loadings  were separated into pervious and impervious area
 components.   Although the monitoring  data collected in each test watershed
 did  not provide  such  a  breakdown of  land surface loadings,  the contribution
 of each fraction can  be inferred from the projections of pervious and
 impervious area  runoff  volumes developed with a watershed model which has
 previously undergone  hydrology calibration.   The first step in the
 calibration procedure was the  execution of a separate model of the test
 watershed's pervious  fraction  to identify those monitored storm events which
 produced runoff  from  pervious  areas.   Those  storms which do not produce
 pervious area runoff  were designated  for the calibration of a separate model
 of the  test watershed's impervious fraction.   The impervious fraction model
 was  then executed with  the rainfall  record for the entire monitoring period
 and  pollutant accumulation rates were iteratively adjusted  until there was
 acceptable agreement  between simulated and observed loadings for those
 storms  which  did not  generate  any runoff from the test watershed's pervious
 fraction.   Differences  between observed pollutant loads and simulated
 impervious fraction loadings were used as calibration data  for iterative
 executions of the pervious fraction model.   Following the determination of
 pervious fraction accumulation rates  with the pervious fraction model,
•simulated pervious  and  impervious area loadings for each storm event were
 summed  and compared with  composite loading observations for goodness-of-fit
 assessments which are reported elsewhere (_9fJ10) .

     During calibration  of daily  pollutant accumulation rates,  care was taken
 to.develop an approximate equilibrium or balance between the accumulation of
 loads on the  one hand and pollutant washoff  and transport on the other.
 Extended dry  periods  produced  noticeable increases in surface pollutants and
 extended wet  periods  produced  decreases.   However,  accumulation rates and
 other model parameters  were  adjusted  to assure that the overall trend was
 relatively stable so  that accumulated pollutant loads were  not continually
 increasing or decreasing  during  the calibration period.   This equilibrium
 was  developed for both  the pervious and impervious fractions in each urban
 watershed.  It was  found  that  daily pollutant decay factors,  which
 approximate reductions  in pollutant loading  potential within the pervious
 (REFER)  and impervious  (REIMP) fractions resulting from wind and biochemical
 processes,  should be  set  at  5%/day to 8%/day to prevent pollutant
 accumulations from  becoming  unstable  during  extended  dry periods.   The
 results of test  watershed model  calibration  study also indicate that the
 necessary accumulation-washoff balance during periods of average rainfall
 can  be  maintained with  transport parameters  set at about 0.4 for the
 pervious fraction (KSER)  and 0.5 in./hr for  the impervious  fraction (RIMPX).

     Verification of Calibrated Nonpoint Pollution Loading Factors.   As
 previously indicated, the risk of producing  biased calibration results are
 significantly reduced by  the use of continuous simulation calibration
 techniques involving  long-term simulations and parameter adjustments that
 are  not keyed to individual  storm events.  Nonetheless,  four different
 modeling studies have been performed  to verify the NFS loading factors
 developed  from the  aforementioned test watershed  model'calibration studies.
                                                -81-

-------
    One of the verification studies involved testing  the  urban  nonpoint
pollution loading factors from the 1976-1977 northern Virginia  study  (_9)  at
two of the EPA/CBP test watersheds covered by the aforementioned  calibration
study:  Pequea #4 and Ware #5, both of which are characterized  by single
family residential land use patterns.  As indicated  in Figure 29  and  Table
10 / the total P and total N loading factors developed by  the earlier  study
achieved a satisfactory representation of monitored  loadings at these two
EPA/CBP residential testing sites.

    A second verification study involved applying calibrated nonpoint
pollution loading factors to two of the EPA/CBP testing sites located in
Maryland:  Chestertown B (single family residential)  and  Browntown Road  (low
tillage cropland).  Due to the timing of the Maryland test  watershed
monitoring studies, the schedule and budget for the  modeling study could
accommodate model verification work at only a few sites.  Based upon  a
review of available monitoring data for the Maryland  testing sites with a
relatively homogeneous land use pattern, Chestertown B and  Browntown  Road
were selected for the verification study.*  The NFS  loading factors
developed by the  1976-1977 northern Virginia test watershed study (9)  for
single family residential land uses produced a satisfactory representation
of monitored loadings at Chestertown B.  As may be  seen in  Table  11 which
presents nonparametric goodness-of-fit statistics,  there  is no  significant
difference between the simulated and recorded datasets at the 95% confidence
interval.  The Chestertown B results represent a third level of verification
of the total P and total N loading factors used for  single  family
residential land  uses.  For the Browntown Road site,  the  monthly  total N
loading factors which produced an acceptable agreement between  simulated  and
recorded (see Table 11) values were only 15% higher  than  the Basin Model
factors shown in Table 8 for low tillage cropland land uses.  This
represents an acceptable verification of total N loading  factors  for
*The following Patuxent River testing  sites were  not modeled  for  this study
because of mixed agricultural land use patterns:  Deale A  (131  ac));  Deale B
(606 ac); Zepp Farm  (55 ac); and Grey  Farm  (34  ac).  Even  if  these sites
were characterized by homogeneous land use patterns, the unavailability of
reduced rainfall and streamflow records  for this  study would  have prevented
modeling analyses as was  the case with the Patuxent Park forest site  (144
ac) which also lacked a land use breakdown.  Although reduced rainfall and
streamflow records were available for  all Chester River  testing sites, the
following were not modeled  for this  study because of mixed land use patterns
and/or  excessive drainage areas for  a  testing  site modeling study: Sutton
Farm mixed agricultural site  (804 ac); Millington A forest/pasture site
(1,295  ac); and Chestertown A residential/institutional  site  (46  ac).  Of
the 5 Chester River  sites with relatively homogeneous land use  patterns and
acceptable drainage  areas,  Still Pond  Road minimum tillage cropland site (29
ac) and Harris Farm  no till cropland site  (14  ac) were not modeled due to
the limited number of storm loading  observations  (i.e.,  1  to  3  storms),
while the Millington B forest site  (271  ac) was not modeled due to
difficulties encountered  in characterizing  the  hydrologic  responses (e.g.,
flat, swampy watershed with high water table conditions) at this  site.
                                    -82-

-------
I
I
I
I
I
I
I
I
I
u
I
I





































IH
I-H:
y
"O U
(U -U
.c -rt
03 2
Ll
4) fH
JJ (5
IT3 JJ
SO-.
EH M
-U -Q
to - ,H
O ^-K *"•*
EH CU
EH W
T3 — cn
C C
nj M -H
^ 3 T!
>i u rtj
U Q o
(0 0= J
S Q.
w _
>H 0 Q
O .C H
iw CU 01
VI r* ~~~
O (0 4J
•H -U C
4J 0 4>
W EH S
•H -H
-U -o -u
<3 C Q>
4J to Ol
01
^^ Tl
4J T C
•H O "3
fa •
1 OS
l*H ^-
O
1 
o
CJ M-l
U-l
0 0
•H C
M 3
-u as
41
fO
L_]
(0
a,
o
2





EH
01 H
W U
EH 2
ff
S 0
D HH
cn fa
HH
« 2
2 U
e£ IH
oS cn

2 fa
o o
X
O J
U Ed
J > <
IH Cd
S J C
o
2 CN
EH
O





a\
i ^r
H
o





• VO
D ui

« o'


Cd
o ;
EH 2 E
01 <
Cd CJ
EH HH
fa
01 IH
1 2
X U
IH C
Q 01 E
Ed
Q fa
M 0
01
1 i-3
O H
S > C
H a
J 0
in
Z f>

o




in
u n
H •
O

• 1C
> CO
S* <=>

u
• 2
cn or
ca c
O E-
O1
|

1 fTi
> -H
H
1


c/^
2 cc
M C
01 g-
o;

CN
in



,j
fa <
fa > —
O OS 2
2 W M
D EH ~
OS 2
IH

4"
c
o
2



2
01
D

Q
<
J


U
EH
IH
w

CQ --»
U

2 M
O EH
as cd
Cd Q
EH M
01 Ol
Cd Cd
E 2
o —
rH
CN

O







•)C






CO
CO

o


o
CN
•
O


^



*


o
CN
O
A



cn





in
CM




v\ &, •*
•H S -H
O fj .
O 2 O
• HH VI



Q
<
O
« -»
&
IS
O CJ

2 •
2 EH
0 •
X J
CQ -^






























0)
JJ
>PH
CQ

'O
H3
O
OS
vntown
o
u
CQ

O
IfH

0)
^ ]
J^
(0
rH
flj
fO

£
O

4J
(0
^i
dj
in
jj.
0
CU
f°H
(0
4J
o
4J

4)
o

>,
/H
C
o
*
                                               -83-

-------
low tillage cropland.  Total N was the only constituent which could be
verified for Browntown Road, due to the availability of only one  total P
observation and exaggerated sediment loads caused by plowing of the stream
channel in the vicinity of the monitoring station.

    A third verification study involved testing the NPS loading factors used
in the Basin Model (see Table 8)  in two mixed land use watersheds within
northern Virginia's Occoquan River Basin:  Occoquan Creek Watershed (343  sq
mi) and Bull Run Watershed (185 sq mi).  The wet-weather monitoring stations
used for this verification study were  located at the outlets of these two
watersheds.  These two watersheds, which are located in the Piedmont
physiographic province, are tributary  to the upper Potomac Estuary.  As may
be seen in Table 12, which summarizes  the land use patterns in the two
watersheds, pasture and cropland land  uses are more predominant (i.e., 43%
of total area) in Occoquan Creek Watershed while urban land uses  are more
predominant (i.e., 15% of total area)   in Bull Run Watershed.  The model used
in this verification study is the same one referred to in this report as  the
Basin Model.  A two-year period was used for NPS loading factor testing:
January 1, 1978 - December 31, 1979.  The year 1978 was characterized by
average to slightly below average streamflows while 1979 was characterized
by relatively high streamflows.  These two years were selected because they
bracket the start-up of an advanced'wastewater treatment plant in the Bull
Run Watershed which essentially eliminated point source loadings  of total
P.  Due to the AWT operations, the pollution loadings monitored at the
outlets of Occoquan Creek and Bull Run watersheds can be attributed
primarily to nonpoint sources, particularly the former which includes no
major wastewater discharges.  After reverifying previous hydrology
calibration/verification results  (38), simulated and recorded pollution
loadings delivered to the outlets of  these two watersheds were compared on
an annual and storm event basis.  Table 13 summarizes comparisons between
simulated and recorded annual loads for total P, total N, nitrate-N and
TKN.  After accounting for errors in  streamflow simulations, errors in the
annual load simulations for total P and total N are approximately ±20% or
less for both 1978 and 1979.  Since observation errors of 10%-20% are not
unreasonable for N and P measurements  (.34,_35) , these comparisons  indicate
satisfactory agreement on an annual basis.  To provide a visual check of
agreement between simulated and recorded storm load datasets, box and
whisker plots of runoff and loading frequency distributions are presented in
Figure 30 for the Occoquan Creek and Bull Run monitoring stations.  A total
of 38 monitored storm events were available for the Occoquan Creek watershed
for 1978-1979, while 24 monitored storm events were available for the Bull
Run watershed.  Visual inspection of  the box-and-whisker plots in Figure  30
indicates that there is good agreement between the simulated and  recorded
distributions for storm flow volumes  and N and P loads.  Table 14 summarizes
the results of the two nonparametric  statistical tests used to provide a
quantitative assessment of the goodness-of-fit for the NPS loading datasets
plotted in Figure 30.  Based on a 0.05 probability cutoff for the 95%
confidence  interval, the high level of significance statistics in Table 14
indicate that the simulated storm flow volumes and pollutant loads do not
vary significantly from those monitored in each watershed.  Thus, the
                                    -84-

-------
I
I
f
I
I
I
I
I
 I
 I
 I
 I
 I
  I
                                Table 12


                      SUMMARY OF EXISTING LAND USE

                IN OCCOQUAN CREEK  AND BULL RUN WATERSHEDS
          LAND USE
Urban Residential
Urban Commercial & Industrial
Other Urban
               Forest  &  Idle Land
                Pasture Land
High Tillage Cropland
Low Tillage Cropland
TOTAL:
% OF TOTAL WATERSHED AREA
                                       OCCOQUAN CREEK
  5.3%
  0.7%
  0.7%
                                           50.0%
                                           31.3%
  2.0%
 10.0%
  100%
                     BULL RUN
10.3%
 2.7%
 1.9%
                     .  58.1%
                       15.9%
 1.6%
 9.5%
 100%
                                                -85-

-------
                                     Table 13

                   Comparison  of Annual Total Flows and Loads
                for Calendar Years 1978 and 1979:  Occoquan Creek
                             and Bull Run Watersheds
                      (1.0  in.  =» 2.54 cm;  1.0 Ib  -  0.453 kg)
 A. OCCOQUAN CHEEK

    1978 Simulated

   . 1978 Recorded-

    1978 Error


    1979 Simulated

    1979 Recorded

    1979 Error


B. BULL RON*

   1978 Simulated

   1978 Recorded

   1978 Error
                       PLOW
                       Jin.)
         TOTAL P
          (Ibs.)
TOTAL N
 (Ibs.)
NITRATE-N
  (Ibs.)
  16.0     97,130    1.31  x  106     529,750

  14.2     71,628    1.35  x  106     680,020

12.7%     35.6%         -3.0%     -22.1%
 TKN
(Ibs.)
                           779,950

                           669,980

                             16.4%
 28.1   201,100   2.48 x 106    914,000    1.56 x 106

 29.3   180,000   2.12 x 106 .   870,740    1.24 x 106

-4.1%     LL.7%          17%       5.0%         25.5%




 17.6    89,540   1.26 x 106    48-9,400       775,000

 16.2   115,060   1.32 x 106    525,000       793,980

 8.6%    -22.2%        -4.2%      -6.8%         -2.4%
-Insufficient monitoring data for calculations of 1979 annual loads
                                   -86-

-------
I
I
f
I
I
 I
 I
 I
 I
  I
  I
  I
  I
  1.08

  9.90

a
J 0.7?


°. 0.54
8
— o.as


2 0.18


  o.o
                                               OCCOOUAN
                                               RESERVOIR
                                     AWT PLANT
                                     RESERVOIR
                                     MONITORING STATION

1.73
0.87
0.43
0.0
8UU HI*


l||

i i


i
occooww
'CREEK


£


1 I


•
1
3.1.0
_ 27-5
* 11.0
I 5.5
•
:
i
rou. am

, *
OCCOOUM
tfi£EJt

E
I •
lJ-
IXC  SIM

§ 1-04
§ 0.78
*o.s»
| 0.28
a. a
BUlt stm
'



1
3 M



OCCOqWH

1
t
.
•
J •
              SIM
                           DEC  SIN
                                     DEC  SIM
 BULL RUfl
                                        Jfl
                            KCC  SIM
                                     ttC  SIM
iU. IU
8 i«.n
g».«
C I0-os
* «.70
| 3.35
a.n
8UU. RUH
•
•


" (


1
)g
oca




i

fiL



] f

•
-
•
-
1 '
                                                       DEC  SIM
                                                                 R£C  SIM
O.M
3 0.55
§ 0.44
8
C 0.33
£ a.n
3
«
n rt
Mm 1^,

.





T
1 0






c
KCOQUM
CRtEH -




]
-

"

]l
                                                        KC  SIM
                                                                 »C  5IM
                                                                                   DCC  SIM
                                                                                            KC  SIM
                   Figure  30.  Comparison of  Simulated  (SIM) and Recorded  (REC) Box and

                                 Whisker  Plots  for Storm Volumes  and Loads at Outlets of
                                 Bull Run (185  sq mi)  and Occoquan Creek (343 sq mi)
                                 Watersheds:  1978 -  1979
                                                         -87-

-------
                                  Table 14

                   LEVEL OF  SIGNIFICANCE  FOR NONPARAMETRIC
             GOODNESS-OF-FIT TESTS:   OCCOQUAN CREEK  AND BULL RUN
              WATERSHED SIMULATIONS OF STORM LOADS FOR  1978-1979
  CONSTITUENT
TWO-SIDED K-S
Flow (cu. ft)
Total P (Ibs)
Total N (Ibs)
Ammonia-N ( Ibs)
TKN (Ibs)
Nitrate-N (Ibs)
DISTRIBUTION
OCCOQUAN CREEK
0.15
0.54
0.44
0.92
0.54
0.70
TEST
BULL RUN
>0.20*
>0.20*
>0.20*
>0.20*
>0.20*
>0.20*
    WILCOXON RANK SUM
	MEDIAN TEST	
OCCOQUAN CREEK    BULL RON
                                                        0.13
                                                        0.40
                                                        0.25
                                                        0.75
                                                        •0.19
                                                        0.79
                                            0.39
                                            0.41
                                            0.34
                                            0.15
                                           0.28
                                            0.49
*Exceeds maximum probability value currently reported in statistical tables
for sample sizes.

NOTE:  N = 38 storm events for Occoquan Creek Watershed
       N » 24 storm events for Bull Run Watershed
                                     -88-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
calibrated NFS loading factors used  in the Basin Model provided  a  very good
representation of nonpoint pollution  loadings  in two  mixed  land  use
watersheds with drainage areas on the order of  a few  hundred  sq  mi.

    A fourth verification study  relied upon streamflow-loading relationships
developed by the U.S.G.S. fall line monitoring  study  (11) funded by
EPA/CBP.  The U.S.G.S. monitoring study covered wet-weather loadings  at the
Susquehanna, Potomac, and James  river fall lines from January 1979 through
April 1981.  Unfortunately, since sufficient NWS rainfall records  were not
available for the period corresponding to the U.S.G.S. monitoring  period,
the Basin Model could not be operated for direct loads.  Instead,  regression
equations (.11/.32) relating daily streamflow to database, were used to
synthesize a daily loading database for a period with available  NWS rainfall
records:  January 1974 - December 1978.  The Basin Model was  then  operated
for the 1974-1978 period with the NPS loading  factors derived from the test
watershed statistics  (e.g., Table 8)  and simulated daily loading records at
the Susquehanna, Potomac, and James  river fall  lines  were compared with the
daily records synthesized from the fall line regression equations.  The
period 1974-1978 was  selected for model verification  because  its streamflow
statistics (i.e., flow-duration curves) at the three  fall line gages  are
reasonably similar to the streamflow  statistics during the  period  covering
the U.S.G.S. monitoring study.  The results of these  fall line comparisons,
which are summarized  in a later  section on receiving  water  submodel
calibration/verification, provide further evidence of the reliability of the
NPS loading factors developed from the testing site studies.

    Determination of Representative NPS Loading Factors.  The purpose of the
test watershed studies was the development of  nonpoint pollution loading
factors for application throughout the 64,000  sq mi drainage  area  of
Chesapeake Bay.  Of the 30 modeled test watersheds, only the  urban sites
relied upon a single  set of loading factors for each  land use category.
Since the urban loading factors developed from test watershed studies
provided a good representation of urban loadings in four different sections
of the study area (i.e., metropolitan Washington, D.C., southeastern
Pennsylvania, eastern Maryland, and southeastern Virginia), it was decided
that these loading factors would be used for all residential  and commercial
land uses in the Chesapeake Bay Basin.  The transferability of urban
nonpoint pollution loading factors is not surprising  because  impervious
cover is such an important contributor to urban nonpoint pollution loadings
(_8,_9)  and an urban land use tends to  exhibit similar  impervious  cover
patterns regardless of location.

    Differences in calibrated loading factors at the  test watersheds  in each
rural-agricultural land use category  can be attributed to variations  in
management practices  and in the significance of sediment loadings.  For each
land use category, volume-weighted mean concentrations for  modeled and
unmodeled (i.e., Patuxent and Chester rivers)  test watersheds were compared
to ascertain long-term loading differences among the  testing  sites.
Volume-weighted mean concentrations for forest, pasture,  and  cropping  land
uses are shown in Table 15.  For the  forest land use  category, a review of
                                                -89-

-------
                                Table 15
      Volum*  Weighted M«an Concentrations:  Chtsaocak* Bay  T»st Wat»rsh»ds
SITS
MLLINGTON A
MILLINGTCN 3
OCCC9 (FOREST)
PATUXENT PARK
PEC2 (FOREST)
WARS 3 (FORES
SITE
PEC3 (CONTILL
WAR; 2 (CROP)
WARf 7 (CROP)
SITE
3RCWNTOV.N ROAD
OCC510(MINTILL5
STILL POND ROAO
SITE
GcALS A
CEALE 3
GREY FARM
SUTTON FARM
USGS GAGE
ZEPP FARM
SITE
HARRIS FARM
OCCC2 (NO-TILL)
SITS
OCCC1 (POOR-PAST
OCCC5(GaOO-PAST
SITE
CHSSTERTOWN A
CHESTERTOHN Z
=>5C4 (RESIO)
WARg 5 (RESIC
3ASIN
CHESTER
CHESTER
CCCOCUAN
PATUXENT
PEQUEA
WARS
3ASIN
PEQUEA
HARE
feARe
3ASIN
CHESTER
OCCOQUAN
CHESTER
3ASIN
PATUXENT
PATUXENT
PATUXENT
CHESTER
CHESTER
PATUXENT
3ASIN
CHESTER
OCCQQUAN
3ASIN
OCCCQUAN
OCCOCUAN
3ASIN
CHESTER
CHESTER
PECUSA
MARS
• — LANOUSE»FOREST —
ACRES MNINCHRO
1173.0 0.08
271.0 0.33
75.8 0.13
144.0 0.11
123.0 0.04
17.4 0.51
ACRES MNINCHRO
115.20 0.22
43.20 0.52
16. T6 0.30
ACRES MNINCHRO
352.0 0.11
25.3 0.34
29.0 0.10
ACRES MNINCHRO
300 0.09
1300 0.09
. 40 0.16
304 0.14 •
8704 0.11
50 0.05
ACRES MNINCHRO
14.1 0.19
26.6 0.53
ACRES MNINCHRO
31.3 0.16
18.3 0.12
• LANOUSEaRESIOENTIAL
ACRES MNINCHRO
46.40 0.11
50.00 0.12
147.20 0.43
6.18 1.68
NTN
24
29
15
25
19
27
NTN
13
f
NTN
12
13
NTN
if
?!
17
NTN
15
NTN
25
10
NTN
26
21
57
22
MNTN
O.S6
1 .C5
0.91
1.11
3.52
0.43
MNTN
22.22
1:11
MNTN
22.53
3.91
3. 01
MNTN
till
3.94
6.31
5.50
3.90
MNTN
2.72
9.00
MNTN
5.97
2.12
MNTN
2.61
3.23
2.40
0.65
NTP
24
il
?!
34
NTP
13
10
NTP
11
NTP
H
34
30
17
NTP
15
NTP
2i
NTP
57
30
MNTP
0.07
C.10
0.13
C.22
0.19
C.06
MNTP
3.?6
t:l\
HNTP
0.57
0.40
1 .65
MNTP
1.41
1.02
2.13
0.22
1.74
1.91
MNTP
4.50
1.23
MNTP
1 .06
0.35
NNTP
1.03
0.91
0.30
0.10
KEY:  ACRES  =  Test Watershed Area  (ac)
      MNINCHRO = Mean Runoff Volume  (in)
      NTN =  #  Events Sampled For TN
      MNTN = Volume Weighted Mean TN  (rag/L)
      NTP =  #  Events Sampled for TP
      MNTP = Volume Weighted Mean TP  (mg/L)
                               -90-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 i
long-term loading statistics for 3 modeled and 3 unmodeled  sites  indicates
that Occoquan #9 is characterized by mean concentrations which  are  similar
to the mean concentrations at most other forest sites.  Therefore,  the
calibrated total P and total N loading factors for Occoguan #9  were selected
as the most representative forest loading factors for application throughout
the Chesapeake Bay drainage area.

    The selection of representative pastureland loading factors was
influenced by the limitations of the land use database for  the Chesapeake
Bay Basin which is based upon interpretations of LANDSAT satellite  images
from the period 1977-1979.  The LANDSAT data interpretations  tend to
emphasize reasonably well-managed pasture (e.g., Occoquan 15) rather than
poorly-managed pastureland (e.g., Occoquan #1)  since the latter is  difficult
to distinguish from low-tillage cropland.  Therefore, the calibrated total P
and total N loading factors for Occoquan #5 were felt to be most  appropriate
for application to the Chesapeake Bay Basin.

    For the cropland land use categories, variations in management  practices
such as manure applications produced different monthly and  average  annual
potency factors for each watershed.  Because sediment is modeled  as the
driving force for nonpoint pollution loadings, comparisons  of test
watersheds to identify representative loading factors were  based  upon
average annual sediment potency factors.  Monitored total P and total N
loads were regressed with monitored sediment loads for each site, and the
slope of the regression line was designated as an average annual  sediment
potency factor (i.e., pollutant mass/sediment mass) which could be  used  to
compare site loading factors with factors for the high tillage or low
tillage cropland datasets.  The monitored storm load datasets were  then
pooled by land use category, and separate pollutant load vs.  sediment load
regressions were performed for the high-tillage cropland and  low-tillage
cropland datasets.  In this manner, total P and total N loading factors  for
test watersheds which were not suited to model calibration  could  be compared
with factors for modeled watersheds.  Likewise, average annual  sediment
potency factors for calibrated watersheds could be compared with  average
annual values for the entire high tillage cropland or low tillage cropland
datasets.  For the high tillage cropland category, the average annual
sediment potency factors for total P and total N resulting  from the pooled
regressions fell approximately midway between the relatively  high values for
Pequea #3 and the relatively low values for Ware #7.  For the low tillage
cropland category, the average annual loading factors resulting from the
pooled regressions are felt to represent an approximate midpoint  between no
tillage and minimum tillage values.  Having identified average annual total
P and total N loading factors for basinwide application, monthly  factors
were based upon the relationship between the average annual values  for the
cropland category and a selected calibration site.  The "land usersite"
ratios of the regressed sediment potency factors were multiplied  by the
calibrated average annual sediment potency factors for Pequea #3  and
Occoquan #10 to develop average annual sediment potency factors for
high-tillage cropland and low-tillage cropland, respectively.  The  average
annual potency factor was then distributed to monthly values  based  upon  the
distributions calibrated for Pequea #3 and Occoquan #10.
                                                -91-

-------
    The resulting NPS loading factors for each land use are shown  in Table
8.  The loading factors for the organic and inorganic fractions of
phosphorus and nitrogen are based upon mean fractions (%)  calculated from
monitoring data for each land use category.  For each land use, the
percentage of P and N in each organic or inorganic form was multiplied by
the composite total P and total N loading factors to obtain the monthly
values for organic-N, ammonia-N, nitrate-N, organic P, and ortho-P shown in
Table 8.  In other words, this approach relied upon the test watershed model
calibration studies to derive NPS loading factors for total P and  total N,
which are then distributed between organic and inorganic fractions based
upon statistical analyses.  It ensures that the sum of each land use's NPS
loading factors for P and N fractions is equivalent to the composite total P
and total N loading factors derived for each land use category.
Consequently, this procedure was felt to be preferable to calibrating
separate loading factors for P and N fractions monitored at each testing
site and then trying to establish composite loading factors for each land
use.

    Comparisons of simulated and recorded loading data at the river basin
gages, as described in a later section, permitted some verification of the P
and N fractions assigned to each land use category.

    The P and N fractions for rural-agricultural land use categories are
based upon monitoring data from the EPA/CBP test watersheds, while the
fractions for urban land use categories were based upon monitoring data
collected at the NURP testing sites in the metropolitan Washington, D.C.
region (_3j3) .  As may be seen in Table 8, the organic fraction typically
predominates in runoff from most land use categories.

    Based on test watershed model calibration results, monthly ground cover
(COWEC) for urban and pasture land uses was set at 100% so that pervious
area loadings are governed entirely by the calibrated pollutant accumulation
rate rather than soil loss.  For the other land use categories, ground cover
was based upon the calibrated values for the test watershed used to derive
the representative loading factors:  Occoquan #9 for forestland, Pequea #3
for high-tillage cropland, and Occoquan -#10 for low tillage cropland.  The
forestland and cropland ground cover values shown in Table 8 were  used to
model the river basins in the southern half of the Chesapeake Bay  drainage
area (e.g., Potomac and James river basins).  For the Susquehanna  River
Basin, which occupies the northern half of the Bay's drainage area, the
ground cover and corresponding sediment potency factors for forestland and
cropland were shifted one month to represent the shorter growing season and
earlier crop harvest.

    Based upon discussions with State SCS staff and some initial Basin Model
test runs, cropland tillage dates were assigned to each major river basin.
On the specified tillage date, the accumulated sediment load on the soil
surface is reset to a user-specified value to account for the tillage.
operation.  Based upon the specified daily decay rate (REPERV), the
accumulated sediment load becomes less available for direct washoff with
                                    -92-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 i
each day that passes following the tillage operation.  The  accumulated
sediment value assigned to each tillage practice is  based upon  the  test
watershed model calibration studies described above.  Based  upon  the  testing
site studies, accumulated sediment is  reset to 1.0 ton/acre  on  each tillage
date for high tillage cropland and to  0.25 ton/acre  on the  single tillage
date for low tillage cropland.  For the river basins in  the  southern  half  of
the Chesapeake Bay drainage area, the  Basin Model assumed that  high tillage
cropland areas are tilled on April 20th and September 23rd  of each  year
(September 24 for 1976).  For the Susquehanna River  Basin,  the  Basin  Model
assumes high tillage cropland areas are tilled on May 10th  and  October 15th
(May 11 and October 16 for 1976), while low tillage  cropland areas  are
tilled on October 15th (October 16 for 1976).  Obviously, every farm  field
is not tilled on the same date in each river basin.  However, it  is felt
that assumed tillage dates and accumulated sediment  reset values provide a
reasonable representation of average changes in soil disturbance during the
2-3 week period when tillage practices typically occur in each  river  basin.
While the selected tillage dates are felt to be adequate for general
planning studies of the 64,000 sq mi drainage area of Chesapeake Bay, these
values can be changed for any follow-up modeling studies by  modifying the
TIMTIL parameter in the nonpoint pollution loading submodel.

    In order to compare annual nonpoint pollution loadings  from the various
land use categories, the NFS model was set up on a hypothetical single-land
use watershed with silt loam soils typical of the Piedmont  Province and a  2%
overland flow slope.  The NPS model was executed with hourly rainfall
records from the Virginia suburbs of Washington, D.C. to simulate annual
loadings for each land use based upon  the NPS loading factors in Table' 8 and
in Hartigan et al. (9) ,  Annual loadings were developed for  a year  of
average wetness (1967)  characterized by 40.6 in of rainfall, and a
relatively wet year (1975), characterized by 54.1 in of rainfall.   The
simulated annual unit area loadings of total N and total P  are  summarized  in
Table 16.  As may be seen, high tillage cropland produces the highest unit
area loads of total N and total P while forestland produces  the lowest unit
area loads.  Also of note are the higher loadings for wet year  conditions.
For example, Table 16 shows significantly higher cropland loadings  for wet
year conditions which can be attributed to percentage increases in  soil loss
that are much greater than runoff increases, while urban land uses  exhibit
increases proportional to runoff increases.

    Dry Weather Flow Concentrations.  As previously  indicated,  the
relatively small size of the test watersheds tended  to reduce the detection
of subsurface flow contributions at the testing site monitoring stations.
Based upon hydrology calibrations, it  appears that the majority of  the
baseflow/interflow from the small test watersheds is delivered  to a stream
channel downstream of the testing site outlet where  the monitoring  station
was located.  Consequently, baseflow/interflow contributions either were not
considered or were suppressed in the model calibration studies  for  most test
watersheds.  As discussed in a later section of this chapter, concentrations
of baseflow/interflow for the Basin Model were calibrated to large
sub-basins without regard for land use breakdown.  However, concentration
                                                -93-

-------
                                    Table  16

  Simulated Annual Surface Washoff of Total N (as (N) and Total P (as P) for
   Average  and Wet Years:  Silt Loam Soils Typical  of  Piedmont Province in
               Northern Virginia (l.'O Ib/ac/yr - 0.89 kg/ha/yr)
                                    TOTAL N LOAD - .           TOTAL P LOAD
          LAND USE	        (Ibs-N/acre/yr)          (Ibs-P/acre/yr)
FOREST.
PASTURE
SINGLE FAMILY RESIDENTIAL
(18% impervious)
COMMERCIAL (90% impervious)
LOW TILLAGE CROPLAND
HIGH TILLAGE CROPLAND
AVG. YH.
0.6
2.6
6.0
10.7
5.0
12.0
WET YR.
0.8
3.0
6.8
12.2
9.0
51.2
AVG. YR.
0.08
0.45
0.86
1.29
0.42
2.05
WET YR.
0.12
0.52
0.97
1.46
0.74
8.72
NOTE:  Based on rainfall for Northern Virginia gages

            o AVG. YR. (1967) - 40.6 in (104.1 on)

            o WET YR.  (1975) - 54.1 in (138.7 cm)
                                    -94-

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
i
statistics on the dry weather flow which was monitored at  some of  the
EPA/CBP testing sites should be of some assistance  in interpreting the
average sub-basin concentrations reported later in  this chapter.   As may  be
seen in Table 17, the highest P and N concentrations in dry weather flow  are
typically associated with cropland sites, the lowest values with forest
sites, and the midrange values with pasture and urban sites.

    Also of interest in Table 17 are the much higher concentrations at the
Pequea Creek sites, particularly in the case of nitrate-N.  For example,  the
mean nitrate-N concentration for the Pequea Creek forest site (Pequea #2) is
37 and 329 times greater than the dry weather flow  means for the Occoquan
and Ware forest sites, respectively.  Likewise, the mean nitrate-N
concentration for the Pequea Creek residential site (Pequea #4)  is on the
order of 30 times greater than the mean concentration for  the other
residential sites.  It is felt that the much higher dry weather flow
concentrations associated with the Pequea sites, particularly for  a readily
reached constituent like nitrate-N, can be attributed in large part to the
underlying limestone formations in the Pequea Creek basin.  In addition to
increasing monitored streamflow volumes at the outlets of  most Pequea Creek
testing sites, it is conceivable that the limestone formations can
accelerate the delivery of leachate to streams at certain  times of the year
and also cause highly concentrated leachate from other land uses to be
delivered to the outlet of a particular test watershed.

    Scale-Up to River Basin Models.  A five-step procedure was followed to
scale-up from the test watershed models to the Chesapeake  Bay river basin
models.

    First, the river basin models were subjected to the independent
hydrology calibration/verification study which was described earlier in this
chapter.  Hydrologic parameter sets developed from  the test watershed model
calibration could not be applied directly to the river basin models because
the subsurface flow component was often not detectable at  the testing
sites.  Further, the independent hydrology calibration for the river basin
model permits an accurate simulation of overland flow transport to the
stream channel system.  An accurate representation of overland flow
transport eliminates the need for application of a "sediment delivery ratio"
to simulated sediment and sediment-related loads.  Based upon test watershed
model sensitivity studies, the river basin models relied upon a transport
coefficient (KSER) equal to 0.4 which provides a stable representation of
runoff transport of detached pollutants.

    Second, an erodibility factor (KRER)  based upon the "K" factors
presented in the State SCS-5 reports was assigned to each  sub-basin based
upon average soils characteristics.  Since the NFS  loading factors for each
land use category are the same in each river basin, the erodibility factor
is one of the most important parameters to represent locational differences
in cropland nonpoint pollution loadings.   For example,  high tillage cropland
in a sub-basin with highly credible soils will produce higher NFS  loads of
total N and total P than the same land use in a sub-basin with less erodible
                                                -95-

-------
                                          Table 17
   I
•>
Statistics

LAND USE
FOREST


PASTURE







LOW-TILL
CROPLAND
HIGH-TILL
CROPLAND

SINGLE
FAMILY
RESIDENTIAL


MIXED
MONITORING
SITE
OCCOQUAN 9
PEQUEA 2
WARE 8
OCCOQUAN 1
OCCOQUAN 3
OCCOQUAN 4
• OCCOQUAN 5
PEQUEA 5
PEQUEA 6


OCCOQUAN 2

PEQUEA 3

WARE 2
OCCOQUAN 7

OCCOQUAN 8
PEQUEA 4
WARE 5
PEQUEA 1
on Dry
Weather Flow Concentrations
TOTAL P
N
52
26
10
72
4
38
8
22
22


49

21

11
68

65
22
10
22
MEAN
0.05
0.03
0.01
0.05
0.10
0.16
0.05
0.48
0.30


0.11

0.16

0.07
0.08

0.05
0.04
0.02
0.10
DEVIATION
0.03
0.03
0.02
0.03
0.09
0.10
0.02
0.34
0.24


0.20

0.05

0.08
0.22

0.03
0.04
0.04
0.05
N
52
25
10
72
5
38
9
22
22


49

•21

11
70

67
22
10
22
(mg/L)
ORTHO-P
MEAN
0.01
0.01
0.0
0.01
0.22
0.02
0.01
0.33
0.22


0.01

0.12

0.04
0.01

0.01
0.01
0.01
0.06
DEVIATION
0.
0.
0
0.
0.
0.
0.
0.
0.


0.

0.

0.
0.

0.
0.
0.
0.
01
01
.0
01
01
02
01
25
20


01

06

07
02

01
01
01
03
N
52
26
9
71
4
38
8
21
22


48

21

11
69

65
22
9
21
1
TOTAL N |
MEAN
0.39
3.57
0.12
3.34
2.18
3.32
1.12
7.75
7.27


5.10

21.2

0.44
0.69

0.64
3.74
0.27
6.53
DEVIATION •
0.16 1
1.11
0.06 1
1.39 1
0.17
1.17 1
0.61
1.76 J
1.37
m
^
2.19
|
I
2.41
1
0.25 1
0.82

0.40
1.16
0.15
1.03
NOTE:  No baseflow recorded at Ware  7  (High-Till Crop)  & Occoquan 10 (Low-Till Crop)
                                          -96-

-------
1
1
f
1
LAND USE
I FOREST

1
• PASTURE

1



L
W
LOW-TILL
• CROPLAND
1 HIGH-TILL
CROPLAND


SINGLE
FAMILY
• RESIDENTIAL

i
m MIXED
"Jl
Statistics
MONITORING
SITE
OCCOQUAN 9
PEQUEA 2
WARE 8
OCCOQUAN I
OCCOQUAN 3
OCCOQUAN 4
OCCOQUAN 5

PEQUEA 5
PEQUEA 6

OCCOQUAN 2

PEQUEA 3
WARE 2

OCCOQUAN 7

OCCOQUAN 8
PEQUEA 4
WARE 5
PEQUEA 1
Table 17 (continued)
on Dry Weather Flow Concentrations (mg/L)
TKN
N
52
26
9
72
4
38
8

21
22

49

21
11

69

65
22
9
22
MEAN
0.30
0.28
0.10
0.48
0.75
1.95
0.81

1.34
0.62

0.74

0.28
0.32

0.46

0.51
0.42
0.19
0.41
DEVIATION
0.15
0.20
0.06
0.26
0.73
1.43
0.47

1.14
0.47

0.74

0.30
0.13

0.66

0.30
0.20
0.12
0.16
N
51
26
10
72
5
38
9

22
22

49

21
11

70

67
22
10
22
AMMONIA-N
MEAN
0.05
0.03
0.0
0.10
0.18
0.36
0.17

0.16
0.09

0.11

0.06
0.0

0.09

0.17
0.05
0.0
0.04
DEVIATION
0.02
0.03
0.0
0.14
0.22
0.40
0.21

0.13
0.08

0.14

0.06
0.0

0.07

0.15
0.04
0.0
0.03
N
53
26
10
71
5
38
9

21
22

48

21
11

70

67
22
10
21
NITRATE-N
MEAN
0.09
3.29
0.01
2.86
1.56
1.37
0.35

6.43
6.65

4.35

21.1
0.12

0.23

0.13
3.31
0.10
6.13
DEVIATION
0.05
1.12
0.02
1.40
0.81
0.94
0.30

1.18
1.27

1.95

2.51
0.14

0.26

0.16
1.15
0.12
1.02
NOTE:  No baseflow recorded at Ware 7 (High-Till Crop)  & Occoquan 10 (Low-Till Crop)
                                            -97-

-------
soils, even though the sediment potency factors are the same in both
sub- basins.

    Third, the NFS loading factors shown in Table 8 were assigned to the
land use categories in each sub-basin.

    Fourth, a receiving water model with point source discharge files was
iteratively executed in a quasi- steady state mode for a typical low flow
condition  (i.e., 25th percentile flow).  Simulated baseflow concentrations
were compared to low flow monitoring data to establish an initial estimate
of basef low/interflow concentrations.

    Fifth, the sub-basin/receiving water models of each river basin were
calibrated for a two-year period (January 1974-December 1975) and verified
for a three-year period (January 1976-December 1978) .  The models were
iteratively executed for the two-year calibration period with the NFS
loading factors assigned in Step 3 to set instream process parameters and to
derive a final set of basef low/interflow concentrations for each sub-basin.
The NFS loading factors were not adjusted during the calibration/
verification of the river basin models.  The results of the receiving water
model calibration/verification are presented in a later section of this
chapter.

    Precipitation Loadings on Bay Surface Area.  The wetfall dataset
developed  by EPA/CBP. ( 39)  was analyzed to derive the following mean annual
concentrations of precipitation falling on the surface area of Chesapeake
Bay:
Organic N:
Ammonia-N :
Nitrate-N:
Organic P:
Inorganic P:
BOD2o:
0.67 mg/L
0.35 mg/L
0.57 mg/L
0.048 mg/L
0.016 mg/L
13.0 mg/L
These mean concentrations were then combined with the hourly rainfall
records for segments 5 and 6 (upper Bay) and segments 9 and 10  (lower Bay)
to develop daily "per acre" loading records for precipitation.  Daily
precipitation loading records were developed for 1966, 1974, and 1975  (i.e.,
production run periods)  in a format suitable for input to each  node in the
Main Bay water quality model.

    Recommendations for Future Test Watershed Studies.  Outlined below are
recommendations for improved coordination between the monitoring and
modeling efforts to ensure maximum usefulness of the test watershed database
(6) .  The recommendations address problems with site selection  and certain
elements in the monitoring work program which were encountered  during  the
modeling study described herein.
                                    -98-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
    A. Site Selection;  One problem which  reduced  the  applications  of
monitoring data from certain test watersheds was the  selection of mixed land
use catchments with significantly different loading factors.   Secondary land
uses which represent a relatively small percentage of  the  total catchment
area can distort monitoring characteri2ations of the primary  land use if
runoff concentrations for the secondary land use are  significantly  higher.
While it may not always be possible to identify single  land use watersheds
for monitoring studies, mixed land use catchments  with  a  secondary  land use
that is characterized by much higher NFS loading factors  than the primary
land use should not be designated as test  watersheds.

    Two of the test watersheds in the Pequea Creek basin were located over
limestone formations that affected the quantity and probably  the quality of
monitored baseflow during dry weather and  storm periods.   Since the  purpose
of test watershed monitoring studies is to collect nonpoint pollution
loading data that are representative of larger basins,  care should  be taken
during site selection to ensure that underlying geology as well as  land use
and upper soils characteristics are representative of  the  river basins in
which the data is to be applied.

    B. Monitoring Work Program;  As previously indicated,  the test  watershed
monitoring studies were designed and initiated in  the absence of a  specific
watershed modeling work program.  The earlier start-up  of  the monitoring
study was intended to ensure sufficient time for modeling  studies of the
monitoring data.  However, if the monitoring and modeling  work program had
been developed and implemented concurrently, it is likely  that:   some
different site selection decisions would have been made; collection  of the
continuous rainfall and runoff records required for model  calibration would
have received a higher priority in order to increase the amount of data
available for model calibration; data reduction requirements  could  have been
reduced considerably; and the results of model calibration studies would be
improved due to the expanded database.  After the  monitoring  studies have
started, periodic interactions between the modeling and monitoring
investigators can facilitate any mid-course corrections necessary to enhance
the applications of the monitoring database.  While it  is  often necessary to
initiate test watershed monitoring studies at the  earliest possible  date to
ensure the maximum amount of monitoring data and/or a sufficient amount of
time for data analysis, the advantages of  better coordination between the
monitoring and modeling efforts from start to finish merits consideration.

    The majority of the hydrometeorologic  data reduction required for model
calibration was performed by the modeling  investigator.  The  monitoring
investigators were not required to reduce  rainfall stripcharts and
reductions of flow stripcharts were generally restricted to the  storms which
produced water quality samples.  Consequently, the modeling investigator was
required to perform most of the quality assurance  checks on
hydrometeorologic data for the majority of the test watersheds.  These
checks included assessments of rainfall-runoff relationships  and comparisons
of runoff volumes recorded at the test watersheds  in each  river  basin.   Due
to the later start-up of the modeling study and delayed transmittal  of
                                                -99-

-------
monitoring data to the modeling investigator, initial quality assurance
checks on the hydrometeorologic dataset were not completed until most  test
watershed monitoring studies had ended.  As a result, onsite experiments to
resolve hydrometeorologic data problems could not be performed, and
mid-course corrections involving additional instrumentation, further
instrument calibration, or the selection of substitute testing  sites could
not be considered.  Further, an earlier quality assurance effort focusing on
model calibration needs probably would have flagged the significant gaps in
the hydrometeorologic records required for model calibration in time to
produce an expanded database.  Therefore, it is recommended that extensive
quality assurance checks be performed on the hydrometeorologic data very
early in the test watershed monitoring study so that problems and anomalies
can be identified in time for mid-course corrections.

    For certain test watersheds, relatively long sampling periods (e.g.,
24-72 hrs)  resulted in the inclusion of excessive baseflow volumes in  the
flow-composite samples for monitored storm events.  As a result, the
separation of baseflow volumes and loadings from -the reported storm volumes
and loadings was very difficult for these test watersheds, and model
calibration studies were significantly complicated.  To ensure the
development of a reliable nonpoint pollution loading dataset by test
watershed monitoring studies, it is recommended that the sample collection/
retrieval schedule be designed to minimize baseflow contributions during
monitored storms.

    For test watershed monitoring studies in four of the five river basins,
separate raingage and flowmeter recorders were used for the majority of the
monitoring period.  The use of separate recorders generally resulted in
unsynchronized rainfall and flow records due to inevitable differences in
chart speeds.  Consequently, one of the more time-consuming data reduction
tasks involved scanning the individual stripcharts to match rainfall and
flow records for monitored storms.  The use of a dual-pen recorder for the
raingage and the flowmeter would not only reduce data reduction requirements
for continuous simulation studies but would also facilitate the quality
assurance checks of hydrometeorologic data recommended above.

Receiving Water Submodel Calibration/Verification

    Methodology.  A total of 13 major U.S.G.S. water quality gages with
extensive monitoring records were selected for the receiving water model
calibration/verification study.  These gages are listed in Table 2 and shown
in Figure 12.  Comparisons of simulated and recorded data were performed at
10 of these gages.  Due to a limited number of observations for most
constituents of interest, water quality data for the following  five gages
was used only to develop subsurface flow concentrations, not for
goodness-of-fit assessments:  Rappahannock River near Fredericksburg,  VA
(#01668000); South Fork Shenandoah River at Front Royal, VA (#0163100);
Shenandoah River at Millville, WV (#01636500); Potomac River at Hancock, MD
(#01613000); and Potomac River at Shepherdstown, WV  (#01618000).
                                    -100-

-------
    Calibration of the receiving water submodel involved:   setting  the  rate
coefficients for instream water quality processes;  setting
baseflow/interflow concentrations which account for pollutant  loadings  from
subsurface flows as well as the relatively small wastewater discharges
(i.e., less than 0.5 MGD) that are not included in  the wastewater treatment
plant dataset and for net flux from stream channel  bottom sediments;  and
quality control checks on wastewater discharge files based  upon  instream
monitoring data.  To minimize calibration costs, the receiving water
submodel was initially operated in a quasi-steady state mode for a  typical
low flow condition to develop initial estimates for baseflow/interflow
concentrations and chlorophyll-_a growth parameters.  It was felt that a low
flow period would be most appropriate for setting initial receiving water
model parameters, since it would be very difficult  to distinguish runoff
loadings from subsurface and wastewater loads under high flow  conditions. •
Because the test watershed data could not be used to establish subsurface
flow concentrations for each land use category (i.e., since
baseflow/interflow contributions were not detectable at the test watershed
monitoring station) , average baseflow/interflow concentrations were
calibrated for each sub-basin without regard for land use breakdown.  While
it would have been preferable to establish subsurface flow  concentrations
for each land use category, the selected approach does provide a reasonably
accurate representation of the magnitude of subsurface flow contributions
upstream from each calibration gage as well as locational differences within
the Chesapeake Bay drainage area.

    The 25th percentile daily streamflow (i.e., streamflow  which is exceeded
75% of the time) at each gage for the period 1966-1978 was  selected for use
in the steady-state model applications.  Starting at the upstream gages and
working downstream,  the receiving water model was operated  'with the 25th
percentile daily streamflow, reported wastewater discharges, and typical
daily meteorologic data for May through October to  set instream process
parameters and tributary baseflow/interflow concentrations.  It should  be
emphasized that the baseflow/interflow concentrations are intended  to
approximate small wastewater discharges and errors  in the wastewater
discharge input files (e.g., treatment plant bypasses)  as well as subsurface
flow concentrations.  Simulated concentrations for  the quasi-steady state
model executions were compared with the observed median concentrations  for
low flow periods at each major gaging station.  Following the  refinement of
initial parameter estimates and baseflow/interflow concentrations with  the
steady-state model,  the Basin Model was calibrated  in a continuous
simulation mode for the period January 1974-December 1975.

    The receiving water model was executed for the  two-year calibration
period and the following parameters were adjusted in conjunction with the
baseflow/interflow concentrations for each river basin:  chlorophyll-_a
growth parameters; BOD sinking rate; chlorophyll-_a  sinking  rate; and  sinking
rates for organic N and P.  Rate coefficients for BOD decay and
nitrification were derived from stream channel geometry relationships
developed by Wright and McDonnell (41)  and Bansal (42).  With the exception
of nitrification rates in a few reaches, no further adjustment to these
                                    -101-

-------
coefficients was made during calibration.  Likewise, nonpoint pollution
loading factors developed from the test watershed modeling study were  not
adjusted during receiving water model calibration to provide a verification
of the previous loading factor calibration work.

    Parameter adjustments were based upon comparisons between simulated and
observed water quality data at two different levels:

    1.   Inspection of time series plots of simulated and observed
         concentrations at USGS monitoring stations and comparison of
         statistics for the simulated and observed concentration datasets.

    2.   Comparison of simulated loading records delivered to the mouths of
         the Susquehanna, Potomac, and James river basins with synthesized
         pollutant loading records developed from regression equations for
         the USGS fall line monitoring dataset.

In other words, calibration focussed on balancing the goodness-of-fit  for
upstream reaches with the match of pollutant loading relationships for the
three major river basins.  Comparisons of simulated loadings with
"flow-loading" relationships from the USGS fall line monitoring study  were
made on a daily, monthly, and annual basis.  Daily loading comparisons
provided an excessively rigorous test of the nonpoint pollution loading
factors derived from the test watershed modeling studies, while the monthly
and annual loading comparisons were used to guide adjustments to receiving
water model parameters and baseflow/interflow calibrations.

    Following calibration, the receiving-water model was verified by
operating it for the period January 1976-December 1978 with constant
parameter sets and baseflow/interflow concentrations.  Since nonpoint
pollution loading factors were not adjusted during the two-year calibration
period, the verification of the test watershed model calibration results
actually totalled five years.

    The 1974-1978 period was selected for receiving water submodel
calibration/verification for five reasons.  First, in order to perform a
meaningful model verification, the wastewater flows and effluent levels used
in the model must be reasonably close to the actual wastewater discharges
which produced the monitored receiving water quality.  Since wastewater
treatment levels changed significantly throughout the Basin during the early
1970's (i.e., prior to 1974) and sufficient data on actual discharges  is not
available for model input, it was advisable to use a verification period
(1974-1978) with available wastewater discharge records and relatively
constant wastewater flows and effluent levels.  Second, the period selected
for model verification should also have a defined land use pattern that is
consistent with the available water quality monitoring records in order to
ensure a meaningful verification of nonpoint pollution loading factors.
Since the existing land use data for the Basin Model was based upon LANDSAT
data  from the period 1977-1979, it was appropriate to select a verification
period that does not predate the LANDSAT scenes by a significant amount,
                                    -102-

-------
I
l'
I
I
I
I
I
I
I
I
I
I
particularly in light of the significant switchover  from  high  tillage  to  low
tillage cropland practices that occurred in mid-  to  late-1970's.  Third,
only a limited amount of plant nutrient data is available for  U.S.G.S.
monitoring stations prior to 1974.  Thus, even if an earlier verification
period was acceptable from the standpoint of wastewater discharges  and  the
Basin land use pattern, an insufficient amount of water quality data would
be available for comparison with Basin Model simulations.  Fourth,  the
1974-1978 period includes a good mix of normal and above-normal streamflow
periods which should provide a good test for parameter sets and
baseflow/interflow concentrations based upon low  flow periods.  Finally,  as
previously indicated, the 1974-1978 period could  be  used  to 'verify  nonpoint
pollution loading factors because its streamflow  statistics are reasonably
similar to the streamflow statistics (e.g., flow-duration curves and
annual/seasonal streamflow volumes)  during the 1979-1981  U.S.G.S. fall  line
monitoring study.

    Fall Line Loading Results.  Monthly, daily, and  annual fall line loads
were compared to measure agreement between the Basin Model simulations  and
the USGS flow-loading relationships.  The simulated  values represent the
loads the total point and nonpoint pollution loads delivered to the fall
line from the tributary river basin.  In all cases,  USGS  loads were based
upon daily loads calculated from simulated daily  streamflows at the mouth of
each river basin, which were then aggregated to produce either monthly  or
annual totals.

    Scatterplots and regressi-on statistics for the calibration and
verification periods that indicate goodness-of-fit for monthly and daily
loads are presented in Figures 31 through 48 for  the Susquehanna River  fall
line, Figures 49 through 66 for the Potomac River fall line, and Figures  67
through 84 for the James River fall line.  For each  set of river basin
figures, the first six cover the two-year calibration period (January 1,
1974 - December 31, 1975), the second six cover the  three-year verification
period (January 1, 1976 - December 31, 1978), and the final six cover both
the calibration and verification periods (January 1, 1974 - December 31,
1978).  Each group of six scatterplots covers monthly loads of total P,
inorganic P, total N, and nitrate-N and daily loads  of total P and total  N.
The "line of equal values" in each scatterplot is indicated by a "*".
Therefore, a visual inspection of the scatterplots to assess departure  from
the line of equal values can serve as one measure of goodness-of-fit.   The
month associated with each point in the scatterplot  can be determined from
the following key:

                             1 = January
                             2 = February
                             3 = March
                             4 = April
                             5 = May
                             6 = June
                             7 = July
                             8 = August
t
                                                -103-

-------
                             9 = September
                             0 = October
                             N = November
                             D = December

Therefore, it is easy to determine whether any months consistently exhibit
significant errors from a visual inspection of each scatterplot.

    In order to quantify goodness-of-fit, a least-square  regression  line was
fit to each scatterplot, and the R2 and slope terms associated with  the
regression are reported in each figure.  The coefficient  of determination
(R )  indicates the percent of total variance which is accounted for  by the
regression line or how well the simulated and USGS-based  loads are fit by a
least-squares regression line with the reported slope.  Acceptable
goodness-of-fit is demonstrated by regressions with a relatively high R
and a slope that is reasonably close to 1.0.  Calculations of t-statistics
on the significance of the slope and intercept terms in the linear
regression equations covered by Figures 31 through 84 revealed that  the
slope term was always significantly different from zero while the intercept
term was generally not.  Consequently, for regressions exhibiting relatively
high R  values, a reasonable indication of simulation error is provided by
the following relationship:

              Average % Error = (Slope - 1.0) * (100%)

    Based upon both visual inspection of Figures 31 through 84 and the
reported regression statistics, all three river basins typically exhibit
satisfactory goodness-of-fit for all three periods.  Most R2 values  are in
excess of 0.85 and many are 0.90 or higher.  Many regression line slope
terms are between 0.9 and 1.1 while most slopes are between 0.75 and 1.25,
which represents an acceptable range of error in light of errors on  the
order of  20% that can be encountered in water quality monitoring studies
(REF).  In general, inorganic P tends to exhibit the least satisfactory
goodness-of-fit in all three river basins, probably because of factors such
as:  (a) sewage treatment plant upgrading during the calibration/
verification period that is not reflected in the constant wastewater
discharge values used in the model; and (b) the approximations required to
model transport through the large river basins.  However, because total P
loads are accurately simulated and organic P is the predominant phosphorus
form delivered to the fall line, the poorer fits associated with inorganic P
are not of serious concern.  The potential role of wastewater discharge
errors in the James River Basin simulations is discussed  in a later
section.  Some of the other general conclusions about goodness-of-fit that
can be based upon the fall line loading comparisons are the following:

    o    While all three river basins exhibit goodness-of-fit for critical
         constituents, the Susquehanna River Basin appears to exhibit the
         best agreement between simulated and USGS-based  loads, followed by
         the Potomac River Basin and the James River Basin in that order.
                                    -104-

-------
 I
 I
 f
 I
 I
 I
 I
 I
 I
 V
I
I
I
I
                                                                    I C
                                                                    I C
  fflj-
  5-0
  •gf-
 rH ~T~

 SK
  •o
  (0
  i
  2
  O
CO
D
   C\J
 0)
 0]
CQ

01
 q
 (0
 IB
 3-J
 r1
 o
ice
•5W
lu «^

§E
01 rf
•r^ ^*"
     o
      X
      LU
oj o
      CO
                                                   N Is-
                                                   00 O>
                                                   o o
                                                   H  II
                                                      o
                                                      j
                                                      CO
           — c
                                            I
                                            c
                   (S81)  QVOl  13QOW

                                -105-
                                             ?
                                              Nisva
                                                                      03
                                                                      -J
+ c "'
!- 2
 -O
                                                                   IS
                                                                    £ Lul
                                                                     CO
                                                                     5
                                                                    § <
                                                                    IS
                                                                     co
                                                                     e

                                                                  ^<
                                                                   m
                                                                  c
                                                                * ^Q
                                                                I  C £f
                                                                        c
                                                                        c
                                                                        r
                                                                       <\

-------
 .3-
   c
   0
  •H
  JJ
   10
  C
  -H
     £L

     O
    0:
    O
 10 iT
 Q O

 !*
w
CJ)
en
3

o
    ro
 T3
 0)
 w
 (0
 aa

 TO
 Tl
 <0
TJ
C
(0

w
13
    OJ
 13 2
 •w LL.
 ra
 3

 •H
 w

 M-l
 O
  a:
  LU
  >
  Q:
  <
  2
  UJ
0) O
^CO
.?z>
fc CO
                                                         to  oo

                                                         d  —
                                                         M  II
                                                            m
                                                            Q.
                                                            O

                                                            CO

58
2. C
C
0-

c
c

X
X
c
^».

c
c
c
<:

C
1
X

X
!
TT

C
£
C
z

X

C
C
c
r
                     (SSI)   QV01  13QOW  N1SV9
                                                                           I

                                                                           I X
                                                                          •f o
                                                                           I C
                                                                          I ~
                                                                          ! s
                                                                          i ~
                                                                          i
                                                                          " CO

                                                                         •I"
                                                                          r>
                                                                              CO
                                                                            s >
                                                                          + c CC.
                                                                           § UJ
                                                                           " CO
                                                                            ffl
                                                                         \  c O
                                                                         +•  c

                                                                         \5
-------
 I
 I
f
I
I
I
I
I
I
I
I
I
I
I
I
t
  I
  a s
  3 ~
  5 *-*
  fo IO
  W N»
  O 1
c fO
0 x.
cu V,
3 Al
 0)
 01
 <0
 CQ

 01
 T)
 (0
      T)

      10
 § UJ
 J2
 >. —
 4J
 C
 0
 1C
  LU
IE
ui ^f
•*j ^^
a
o
o
00 GO
O cr
M  II


a^
   O
   _J
   CO
                                      <*
                                      p^;




                                      g




                                      r,




                                      C
                                                                             CO
                                                                             00
                                                                             _J
                                     ! £ O
                                     •+ c
                                     I C
                                     i c
                                     I (T
                                     !^
                                      r
                                      5
                                    "
                                                                             cc
                                                                             LU
                                                                             CO
                                                                             CD
                                                                             O
                                    i|§
                                    + ^ o
                                    I r- >«J
                                                                              !?
                                                                                 O
                                       co
                                                                        i
                                                                        + c
                                                                         S co'
                                                                         C  .
                                                                         - o


                                                                         s -^
                                                                         *.
                                                                         P 2
                                                                         * o


                                                                         r
  CO
                                   i  c
                                   !  S
                                   I  Cw
                                                                             + c
                                                                             ! S
                                                                        * c
                         Q
                         UJ
                         CO
                         <
                         00
                           C
c
c
IT

                          (SSI)  QVCTI  13QOW   NISV8
                                         -107-

-------
    LU
    H

    DC.
 0
 en
 U
 CO
 c
 o
 a


 13
 0)
 Ul
 (0
 oa
 O
 i-l
 C
 it]
ro
x.
c\l
 §
CO
'4-4

O
  IT
  LJ
                                                     in  ro
                                                     f>;  N

                                                     O  O
O
-I
                                                                         r
                                                                     + c
                                                                      r
                                                                      c
                                                                      r\
                                                                      n.



                                                                      I
                                                                          CO

                                                                          CD
                                                                       CO
                                                                     I C
                                                                     I C

                                                                     \-o
                                                                     I C
                                                                     I C
                                                                     I
                                                                     I
                                                                          <
                                                                    + c

                                                                      c
                                                                      T-



                                                                      c
                                                                      c
                                                                      c
                                                                    •f C
                                                                      c
                                                                      c
                                                                     l\
- LU
  CO

IS
                                                                    .1!
                                                                    " I u-
                                                                     r
                                                                     - CO
                                                                     lo
                                                                     c LU
                                                                     " CO
                                                                     IS
                                                                        cO
                                                                         s
                                                                         o
 CO
 D
 CO


C >= i
1 ! I


c
c
f
c


C c
§ §
C c
x £
rr •»•


^
g ' f
I 1
                    (sai)
                                13QOW  NISV8

                                -108-

-------
f
      a<
      c O
      Cn
      o If)
     W
     o
       ro
     CO
     (0
     m
     TJ
     C
     
      O
      CO
5= 8
r
c
js


c
c



c
£-

(SSI)

S
X
c
c
avoi

r
c
c
IT
c
c
i
"13QOIM
-109-

I I
t t
"• <\
Nisve

c
r



-------
  .§<

  *I
  2*"
  rtj
  H V.
  0) >-
  I-*' *^
   «

 ft!

 3
 c "•"
 IT3

 ^y
 it! —"
 Q —I
 •a ^
 aj (7^
 4J U-
 CO
u-i
0
o
o
LU
>
o:

<
  x
 . LU
S =)
a, O
^ CO
01 ID
£ co   —
                   (S8H)  QV01  13QOIAI  NISV9
                                -no-

-------
I
f
   <
   o
 •H
 .C
 CO

 !>
 *3
 rH -T-


 ft
       CO
I ^
c ro
o V.
a ^»
3 C\J
-o *"•

                                                            CO
                                                            CD
                                                           ? O
                                                              „ O
                                                              C -I

                                                               CO
                                                              IS

?
5
•c

g
^r

C
£
r\

C
f
C

1
c:

c
c
X
<:

c
1
c
r-

?
C

-------
   CL

   O
  li
  •H """*
  4J
 -H
 •o
 o -""^
 i-H jMt
 b 00
 I
 
.0
iiCO
O GO
O> CO
O O
H  a
Q.
O

CO
                                                                            CO
                                                                            CD
                ? cn
                                                                          ^ CO
                                                                           CD

                                                                          c^

                                                                          J CD
                                                                          I?
                                                                           CO
                                                                            •
                                                                           CD

                                                                           CO
                                                                         c UJ
                                                                         S CO
                                                                         - <
                                                                           ffi

                                                                         ?Q
                                                                         S <
— c c
5 5 f
=_ (T C
IT ~
"• a-


c
c
ir
r-
(S81

|

T
)

C
s
c
IT
avoi

s
>c

^
? 5 - S
f - e *
^ ? 5 ?
Ov r\ 	
C
s
c
IT
13QOW NISVQ
-112-



-------
I
I
 I

I
I
I
I
I
I
I
I
I
I
I
        Cn _

        ? o
       CQ
       O
       w
                                                       (Q  CD

                                                       CM
                                                       or
  0
  a
 u

UJ
CL
O

en
    CVJ
  01  I
  
-------
    r—  I
   z a:
   HI-
  in
  c
  •O
  § A
   GO
co r*»

I ^

= 10
o ^
a ^
3 CM
T3 "~*
<" i
Ul I
(0 '


:g
TJ r-
 T3
 C
 (0
  LU
2-J
-1-J

Id
•a <
^ u:
ta

IS
co «^
C
0
  cr
 . Ld
O —x
^r —^

V O

5 CO

?^5

- CO
                                   IT.
                                                 (VJ N-
                                                 0) 00
                                                  •  »
                                                 O O

                                                 H  u

                                                 CM  iii

                                                 « 0-

                                                    o


                                                    CO
                                                                  IS
                                                                  
-------
I
I
I
I
I
I
I
1
I
u
I
I
I
I
1
  12
  iP
 I w
 b N.
 S 10
 a
 TJ
 o;
 w
 TJ
 CQ

 0)
 •u
  LU
 S -1
 Q -J
 01
 •l-i
 v<
 r«

*s
IS
.? o
"• w

is
C
LT
r.

C
C
c
^,

1 ?'i <>Ar\

c c
5 F
f £
<\ _

c
r

i
1
IPS/TOO +
1
I

C
£

c
s
r-

X

C
(\
                (Sai) QVOl  13QOIM NISV8
                          -115-

-------
  z


  -I


, h-


•£ I—
W

o >"•
•H   I
-P •—J
s a
cr>
c  ••
•H ^-^

I?

?i
  ro
W
o
U3
O

C
0
a
3

Tf
01
U)
(0
oa

in
T3


3

•o
  CD
  UJ
"3

3
,— I

•H
4J

(0
M-l
O


0
tfl
-H
^1
(Q
ft


s
  (T
  UJ


  E

  <
  LU
  Z)
  o
  CO
'- CO
                       (sai)   avoi  13QOIN  NISVQ
                                     -116-

-------
    CL

    _J
  °fO
 •gl-

 I>-
 §•
 TJ
   00
 01  I
 rn  i
 03
W
T3
T3
  LU
1

1
1

1

1
•ta»
. >, —
3 -J
4J
1 ^
•a ^
4J l{_
rH
i ^
S w
U-l *^
°o=
0 ..
01 ^J
'^ 2
if
s<
CP
 LU
 z>
 o
 CO
 Z)
 CO
                                                 c\j ro
                                                 0> O
                                                   ,   •
                                                 O —
                                                 H  II


                                                its
                                                    o

                                                    CO
              r

              <%
                            X

                            c
                                                                        i r
                                                                        i c

                                                                         c
                                                                         c
                                                                           CO

                                                                           CD
                                                                           CO
                                                                       i
                                                                       i
                                                                       t C
                                                                       I C
                                                                       i r
                                                                       + c
                                                                       i c
                                                                       i
                                                                       i
                                                                       i
                                                                       i c
                                                                       i f
                                                                       i c

                                                                       I C
                                                                       i r
                                                                       r _
                                                                       i
                                                                       i
                                                                       i
                                                                      i
                                                                      i
                                                                      i
                                                                      i
                                                                      i  r
                                                                      I  c
                                                                      +  C
                                                                      i  c
                                                                      i  r
                                                                      I  v-
                                                                      i
                                                                      i
                                                                      i
                                                                      i
                                                                      i  c
                                                                      i  r
                                                                      +  c

                                                                      !  4
                                                                      I TT
                                                                      I
                                                                      I
                                                                        <•  LU
                                                                           CO
                                                                           CD
                                                                        §  O
                                                                           o
                                                                          o
                                                                          CO

                                                                          C5

                                                                          CO
                                                                          Q
                                                                          UJ
                                                                          CO


                                                                          CD
                                                                          <
                                                                          O
                     (S8~l)  QV01  "I3QOW  NISV8
                                   -117-

-------
    a.
    o
  |0

  o Z
  -H
  4J
  (0 S.
  -1 *l
  .
X
c
<1~
c
c
c
c
c
c
c
c
r
                    (sa~n  avoi  i3Qow  NISVS
                                                                      i r
                                                                      i c
                                                                      * c
                                                                      i c
                                                                      I T
                                                                      i
                     c
                   .. c
                     c



                     c

                   .. c

                    *^
                                                                        cn
                                                                        CD
                                                                        -J
                                                                        C/5
                                                                      r
                                                                     + r
                                                                     i c
                                                                     i *•
                                                                     i <•
                    cc

                    r
                    T
                    If


                    C

                   . r
                    _c

                    T
                                                                        cc
                                                                        UJ
                                                                        CO
                                                                        03
                                                                        O
                                                                      r

                                                                    *s
                                                                     fN.
                                                                     T

                                                                      C
                                                                    •  r
                                                                      c
                                                                     -c
                                                                       CO
                                                                       CO
                                                                       Q
                                                                       UJ
                                                                       CO
                                                                       
-------
 I
 I
 I'
 I
 I
 I
 I
 I
 I
I
I
I
I
I
  •H
  .C
  tfl

  O
  •H
  4-1
  13
    1-
    o
    I-
  K
  73
  m
 I

 0
iH
fa

M
O
w
D

C
0
a
 TJ
 01
 (0
 CQ
 W
 T3
 fl
H
z
o
2
 M

00
h-
v^

fO
>,
(—1
£
4->
C
O


TJ

-------
CO
CD  C\J
00  00
o  6
H  II

is
   o
   CO

g
"
r
c
c

£
c
c
c
•j-
r—

C
C
c
c
L"

X
J^
>
r.

C

A
X

c ,
c
\ i
u- t


r *
«• x

                                                           5 CO

                                                           
-------
I
I
  I
I
I
I
I
I
I
I
u
I
I
I
I
   w ^
   § I-
  •H
  4->

   " ^
  a =J
    CD
   10

 a
 I
 •H 	I
 (0 —•'
 Q -J
 'S ^
 ™ LL.
  >
  o:
I <
s
  x
  LU
  ID
  O
  CO

  CO  --
                    (S81)  QVOl  13QOW  NISV8
                                 -121-

-------
  ig
  •H
  4J
  (0 ,.
  -< ^"*
  0)
  *s
  £ Q
  03
  3^

    GO
   V.
 I

 o
r-)
C^

cn
 § ro
 c ^
 13
 03
 •a
 
-------
I
I
I
I
I
I
I
I
I
I
1
1
I
        § p
        2 1-
        Cn
        C

       O
       w
       a
       a
       3
       g
       W
       73
       (0
       to
       "S
         IO
      3  3
      >1
      I—I
      5  UJ
      §  z
      0)
      4J
      <0
      U-l
      O

      c
      0
      w
      ft
      E

 2
 rr
 UJ
 >
 tr
 o
 <
 2
 o
i-
o
Q.
                                                      W (\J
                                                      CD cn
                                                      6 6
                                                      . H  II


                                                      "iS
                                                         O
                                                         _J
                                                         CO
                                                                     c  cr>
                                                                              i

                                                                              • • c
                                                                               r
                                                                               c
                                                                      ! §
                                                                      4- r
                                                                       r
                                                                       T


                                                                       C
                                                                       r
                                                                      * c
                                                                      I C
 I
 I
 I C

 il=i
 T f-
 I
 I
 i ? 2
 t £ O
                                                                         CO
                                                                         in
                                                                                 CO
                                                                             I V
  C

•  C

  c
  V"

  c

+  r
i  c
i  -r
H
<

(T
LJ
CO
CD
O

O
                                                                      c' Q
                                                                      5 §
                                                                      I I
                                                                        CO

                                                                        9
                                                                        CO
                                                                        3
                                                                            i ~
                                                                            T 8
   Q
   UJ
   CO
   <
   CD

r c
'"1
0-

c
c
cr

C
1
S
p*.

s
X
c
c
c

J
r
c
j-

S
X
c
TT

C
X
c

g
s
7
C
8
                                                                            -  Q
                                                                            ,  <
                                                                            IS
                          (sail  avoi  -i3aow  NISVS
                                       -123-

-------
   Q_

   O
   2
^  <
«  CD
S  K
3  O
  a
  •H
  Q
  i—i
  i
  2
  O
  r-H
  fa

  cn
  o
  en

  c


  I
    c\j
 03

 3

 >,
 M
 JS
 4J

 0
i
D,

3
o
in
0)
Vl
cn
 LU
 2
 -J

 _J
 _J
 <
 Li.

 (T
   O
   <
  O
  a.
^r
C
C
(™*
C
c
c
c
X
r»»
f\
i
i
^'loooo +
I
1
C
S
rv
C
C
o:
c
LT
C
C
f\
X
5-
C
C
•c
c
r
__ \
c
                     (S81)  QV01  13QOIAJ  NISV9
                                   -124-

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
I
f
       •g  <
       §  t
       -  P
       «  1»
       0)
       * T
       S fE
       n •
       g 0
       co  »
         IO
       I
       •o
       
-------

r*-

SS'IOOOO

r
c

c
1
c
IT
•^

'lOOIXlOO

c
I
1?

S
1
ri

5 §
5 £
c r
IT C
(\ r\

8
X
C
u~

C
\

I
C
IP
(sai)  avcn  13001^  NISVS
            -126-

-------
I
I
I
I
I
I
I
I
I
  01   I
  c —I
  .2 <
  S I-
  3 O
  « f-
  en
  -5 x
  •s 2
    0
 a
    10
    ro
 TI
 T3
 C
 (13

 tn
 T3
   LU
 2  ±

 V  -J
 0)

 5  -J
§
u 0
 . <

c

c
n,
(\

C
r\

C
iT
P»

X
u"

C
r.

C
c
r

0
••»

|
§
LT

r
LT
                     (sal)  avoi
                                   -127-

-------
   03
   a
   •H
   J=
   W
   C
    h-
    o
-p
(0
H
0
OS

CT1
c
•H
T3
ITJ
  Da ~
  C
  0
 'O
 0)
 w
 (0
 03

 0]
 T!
 J-


 ^ UJ
 T)
 (1)
 3
 S
 ^
 I  w
o (j

 . <
^ 55
in ^

0) O

^ h-

•?0
fe CL

§ 5
£

c S
I f
5 ^

8
|
rs

C
I
£

c
1
?

S
r

g
c
!*
C
8
                      (S81)  QVOl   13QOW  NISVQ
                                    -128-

-------
I
I
I
I
I
I
I
I
I

I
I
I
I
    o
  r-l

  0)
  § 2
   00
 "S ro
 w V.
 fl ^
 m C\J
      (0
      w
      •n
      nj

      3
 C
 o
TJ
0)
   UJ
I
    in
    in
    .21  O
    C,  Q_
           -  s
                        (sen)  avoi  IHQOW  NISVG
                                     -129-

-------
    a.

    o
  •g o
 •H

 ~
 fa
 en
 en
 3
 C
 O
 01

 (0

 03


 01

 T3
 IT)
   GO
    I

   (O
   LjJ
 •n

 ,
r-l
.c
-US
c
o
s

T)
0)
I  <
en  li

U-<

0  £E

§  UJ
I Q1

S O
  O
  n
2
C
C
•T
C
C
C
C
X
C
r
r
ir.
c
C
£
C
C
                      (sai)   avcri  naaow
                                     -130-

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
   to
  3  -J
   «  5
   c  j—
   n  r^
  8 0
  to
 •u
  co
        o
        a


       T3
        01
        (/)
        (0
       CQ
 •s
 "S
    GO
   CVJ
 5
 §
 0)
 iJ
 in
 *—i
 3
      U-i
      O

      C
      o
      0)
         UJ
  UJ

-------
   £LU
  •H
  4J
  n3
  H
  0)
  (X
c
•rH
T)
nj
0
i-l
I

0
  co —
  CO
  ii
       a:
       h-
    GO
  c
  o
  a
  3
 s ^
 m (\J
 01 """
 TD

 T3
    I
   (0
 4J
 C
 0
 0)
 j-j
 ITJ
 (—^
 3
 E
 •H
 CO
 0 
 ^ _
 "3 f^
 a Q.
 o » .
u O
03
in
  O
  h-

1

c
c
•x

\
I
C

X
PC
5

c
*•
c
X
e

E
f
^

X
X
•c

r
c
1
if-

£
?
c
T

8 -
1 !

c
c
c
c
                     (SSI)  QV01  13QOW
                                    -132-

-------
I
I
I
I
I
I
I
I
I
I*
I
I
I
I
  01 n
  a QL
  •H
  JS

  S -J
  v ;-*
  « I-
  Cn
 •5 >-
 "S -I

 ?S
 go
   00
 3 fO
 £ <\J
 
-------
   a
  •.-I
  -W
  (0
  Q)
  K
  C
  •H
  CO
  o
  CO
    <
    o
  C
  o
  OH
 m

 at
 T3
 10

 a

 "0
 C
 nj
 (0
 0
GO
f^
V.

ro
v»
CVJ

 I

CD
   UJ
 -O
 OJ
 CO
   <
   ll
 0  o:
 o  UJ
 w  ^
3 o
S
  O
  I—
  o
  Q_
                                  c

                                  I
                                  c
c
£
c
c
fr
c
c
?
c
rC
C
?
?
r

                      (S8~l)  QVOl  13QOIM  N1SV8
                                     -134-

-------
I
I
I
I
I
I
I
I
I
I
I
I
I
I
 •H
 TS
 8 ~
 w <—»
 Doo
 
-------
    CL


    O
  0)

  K
  CO
  O  •«
  w «— »


  °2

  Ix
  w
  .
r-.-

* c
r

-,


F
• • r
r\



f
|

^1


C
c
c



c
r
CO

CD
_J
«••'

CO
z
o

l-
<


a:
LU
CO
CD
o


o
4^L
^ ^
0
<
0
	 1
^^
1
^
5
o
	 1
v^
u.


CO
C5

CO
3'

•M^
z

^

§

c

c
X

C
r

\

C
X
r
- ^ - w ^
C
                   (S81)  OVOl  13QOW   NISV8
                               -136-

-------
I
I
I
I
I
1
       a
      fcj ^

      8  -
      w «—•»
      D OD
      § K
      g-v.

      "S fO
     CO —
     "O
     T3  I
     Q  I
     s
     o
     £

     TJ
     0)
LU

13
_j
_j
<
    IU

    °£T

    o LU
    W %w
    H3 -^^
    a a:
      o
       _
    1 o
                        (ssi)   avoi  i3aoi/v
                                       -137-

-------
  f
  01 -J "•*
  .s x o:
  "g K K

  522
  4 0Z
  0 «5 J_
  3 -
  §
   GO
 I/I
 T3
 TJ
 73
 c
 nj
   ro
   \
   CVJ
 x:
 4J
 c
 o
 T)
 0)
 3
 s
LU
U)
•H
  rr
  UJ
!*
8 o
  O
  I-
i r
i r
1 C
1 C
1 C
i _
i

c
c
c
c
c
c
^.





c
c
c
• c
c




















^JH,
(
CO
CD
•
-J
*^^
                    (S81)  QV01  13QOIM  NISVQ
                                 -138-

-------
I
I
l(
I
I
I
I
I
I
I
I
I
I
I
       a, Q_
       S -I
       •2 <
         O
         h-
       TJ  I
       «J _J
       CO _
       o CD
       CO ~
       a h-
       0)
       CO f\J
       to i.2


       « I
       t3 '
       (0 —t.

       tJ ^

       -a ^

       m —

       w ^>
        >, UJ

        a ^~
        ^
        a
        r-l
        3
        g
        •r-l
        w
-J
-J
<
u.
         UJ
       § o
       o w

       t,
         H
         O
         Q.
              I I
                f
                (\
                      c
                      c
                      !\


                             (sen)  avoi  i3Qow  NISVS
                                           -139-

-------

c
F
<•

c
F
c
*r

x
X.
C
r\

C
F

F
c
tr

r
1
ft n noon +
1
1

C
F
rr

C
F
c
r\
(S81)
 13QOW  NISVQ
-140-

-------
I
I
l'
I
I
I
I
I
I
fc
I
I
I
  CO
  S
  §1-
   o
   10
lO  O
05  N
o"  d
n   it
                                                               O
                                                               J
                                                               CO
 "8 ro
 W
 (0
 03

 W
 •o
 T3

 W
t3
 IT)

5
j;
4J
CVJ
  -I
  -I
  <
  u.

  o:
  Ld
     3
                           O
                           m
                                              C

                                              c

                                             oa
                                                   C


                                                   in
         c

         c
                                                                          < §

                                                                           c

                                                                           >c



                                                                           c
                                                                          • c
                                                                           c
                                                                           in
                                                                           IT



                                                                           F

                                                                           c
                                                                           in
                                                                                CO
                                                                           c
                                                                         I  C
                                                                         +  c
                                                                           c
                                                                           in
                                                                           ^
                                                                         + c
                                                                         I C
                                                                         I C
                                                                         I V
                                                                         I

                                                                         ! g
                                                                         + c
                                                                         I C
                                                                         i in
                                                                         i <-i
                                                                         i
                                                                         i
                                                                         i
                                                                         i
                                                                         i c
                                                                         I C
                   CE
                   UJ
                   CO
                   CD
                   O
                                                                        + t $
                                                                        '•$ o
                                                                           + c  co
                                                                           i c  "^
                                                                         ir,
                                                                         <\'


                                                                         c


                                                                        " f  2
                                                                         ^  O
                                                                              cd
                                                                              D"
                                                                        + c
                                                                        I C ,_
                                                                        I in CO
                                                                           o
                                                                           LU
                                                                           i
                                                                           i
                                                                           i
                                                                           ! £
                                                                                I C
                                                                                in






                                                                               •»• C
                                                                               I
                                                                                   CD
                           (sen)  a von  "i3aow  NISVS
                                          -141-

-------
   o
  in ^r

  '£<
  m CD
 a
 en
   H-
   2
0)
01
113
CQ

tn
•o
 c
 <0
 a
 > •
   PO
   x.
   <\J
CO
  u.

  o:
3
cn
          sf
          Q. C
                    (S9T)   QV01  13QOW  NISVS
                                 -142-

-------
I

1
I
I
I
I
I
I
I*
I
I
I
I
I
  2,-J
    I-
    o
to
c
o
•H
4J

-------
    I
    O
         UJ
      -J
   S  -
   w
   n3
   ffl

   tn
   •o
   10
  •s
  >.
  I—I
  £
  0
  s
  0)
  4J
  OJ
  <— I
  3

  e
 C

 8 o:
 2 UJ
 <0 ^.
o Q:


g co
<, iu
S 5
Cn —r
ro
x
(\j

I
                s
                c
                (SSI)
                                    13QQW   NJSVg
                                  -144-

-------
I
V
I
1
I
I
i
      CO


      0
      •rH

      4J
      (0
        o
        h-
      CO
      u
      CQ
      o
     •a
     o
     w
     OJ
     CQ
     en
     •a
     03
        in
        N-
        \
        fo
        V.
        OJ

        I
     T3
     0)
     4J
     03
     r—I
     3
     g

     co
     o
     c
       CO
    a,  UJ
    £  <
    a,  -
                                   avcn  13QOW  Nisva
                                         -145-

-------
w
C
o
•H
4-1
(0
i-(
0)
a
  <


  O
(0
3
  <
  o
  10
  N-
  v*
  Jo
  \
  C\J

   I
CO
o
co
C
0
04
3
en
1
iH
H yj

° ?
"S 	i
D

•H
en

M-l
0
  £T
  UJ
a
e
  CO
  LU
                      (S81)
                                       13QOW   NISVQ

                                     -146-

-------
 I
 I
 f
 I
 I
 I
 I
 I
 I
I
I
I
I
  |Q.
  I -J
  ^h-
  "o
  0) >^
  « I—
  Cn
  CO
  O
  C
  O
  a
 03
   ro
 4J
 C
  UJ

c.
s o:
  LJ
Sir
•; co

-------
    Q.


   M O
  C
  o
  •H
  K
  M
  I®
  => r-

 T3 V.
  O _
 to
 TI
 
-------
I
I
I
I
I
I
I
I
       •H
       •S-J
       §<
        o
        I—
        '
       *Z
       o ^
       CO

       3*
      C
      0
      ft
      3

      T3
      m
      03
        CM
     •a
     0)
     3

     • H

     C/}
     c
     o
     tn
       <
       u_

       o:
       UJ
       o:

       CO
       LU
     a
     cri
     •H
     t,
2
H
s
1
in
in
c
1
c
c
in
8

-------
  tn
  04
  2 LU

  Ib
  •s o:
  •U I
  tt)

  K
  01
  TJ
  §
 § ~




 IP
 3 Is-
      UJ
      a:
 
a:

CO
LU
s
!
(M
I
I
1 75oooo *
I
J
g
c
o
o
in
I
I
2500OO +
I
1
8 i
f <
? |£
1 !
in
g
in
                     (S81)  QVOl  13QOIN   NISV9
                                    -150-

-------
 I

 f
 I
 I
 I
 I
 I
 I
I
I
I
I
I
  CO
  a

  I o-

  I -I
  % <
  •3 I-
•H
•s >.
3 2

8 <
£ Q
 03
 03
 D
T3

03 3 a CJ 4J 03 § I 8 r-~ r~ Cn > tr CO UJ CO CD § cc UJ CO CD O CD O -I O CO o CO r> 2 O Q UJ CO < GO Q < O (SSI) QV01 13QOIM -151- NISV8


-------
  w
  a
  •H
  ^r*t
  OT
  C
  0
  •H
  g1
  •H

  co
    GO
  Is
  "8 ro
  cn
  03
  ffl
  CO
  TJ
 CM

  I
 CO
 c
 "3
 a
 >,
 a
 TJ
 0)
   <
   LL
 O
 Wl
I-E
co
=
a
C75
UJ
                   (S81)  QVOl  -13QOW  NISV9

                                -152-

-------
I


}
I
I
I
I
I
I
 a. Q.
•r^
* I"
*u .^^

0 ^
« h-
- >:
tJ «_J
s^
w
o
-S c\J
•
^•^^
I^^^H

1


1

1

1

1
>, ~^
r— 1 . 	 ..
x: ^^
4J
1 w
z
(1) 1
•p -J
10
i -1
"<
o U.
§ o:
"S ^
QJ ^?^
E „ „
O /»>.
o Q«
<*' CO
^ UJ
OJ v
o  o

H  II




   o

   CO
                                                                   c
                                                                  . c
                                                                   c
                                                                   c
                                                                   If
                                                                   r

                                                                   c
                                                                   c
                                                                   IT
                                                                  + C
                                                                  I
                  f     I
                  5     r
                                   r
                                   c
                                                           c
                                                           c
                                                           rv
                                                                     CO

                                                                     CO
                                                                     CO
                                                                     5
                                                                     o:
                                                                     UJ
                                                                     CO
                                                                      I


                                                                     o
                                                                     CO

                                                                     O
                                                                     co"
                                                                     O
                                                                     UJ
                                                                     CO
                                                                     <
                                                                     CD


                                                                     O
                                                                     <
                                                                     o
                    (ssi)  avoi  i3Qow

                                 -153-

-------
 W f %
 ft ^

 if

 IS
 4J
 TJ
   IT
   O
 10

 S
 I


 O
 rH
 fc.
 Eg
 3 h-
 "8 ^
 I P
0
J
 (0

 cn
 T3
 03
 -U
 0)
 3
vw . ~
o U.
C
o
in
  (T
M UJ
a >
IE
O
03
  CO
  UJ
                                                               CO
                                                               CD
                                                              CO
                                                              Z
                                                              o
                                                               CC
                                                               UJ
                                                               CO
                                                               CO
                                                               o
 Q
 <
 O
 _J
 i

 O
 _J
 U.

 CO
 O
 CO
 3

 2
 O

 Q
 UJ
 CO
 <
ffi

Q
<
O

i]
— r-
n

C
r
r

C
S
r-»

C
C
Lf

C
f
IT
r\

C
C
C

C
r».

C
C
ir

c
                  (SSI)  QV01 13QOIM
                                      NISVQ
                             -154-

-------
 I
 I
f
I
I
I
I
I
I
I
I
I
I
 en
 ft
 •H
 JS
 01
 c
 o
 •H
 4J
 (0
 r-<
 0)
 05

 tn
    o
    H-
0 00
 TJ ^
   o^
                                                        d  d
                                                          o
                                                          en
I r
1 C
+ c
1 C
1 C
i r
1 V
i
i
1 C
1 C
1 C
•f C
1 C
1 J~
r IT
i
i
i
i r
1 C
1 C
* r
i r
1 C
i
i
i
1 C
i r
1 C
+ r
i r
1 L"
\ —
[
(
1
1 C
1 C
1 C
^ c
1 C
1 C
1 r-
1
r
i
I C
1 C
1 C
+ c
1 c
I -
1 ^*
1
>






















^^
CO*
CD
_J

*'•'*


CO
2J
O

h-
<


cr
UJ
CO
CD
o
                                                                         CD
                         :ssi)  avoi  laaoi/v  IMISVS
                                      -155-

-------
 U)
 ft
 •H
 JZ
 w

 o
 •H
 -p
 (0
 en
   LU
   H

   IT
   2UJ
 3 x
  CO
  LU
•S1 <
                                                                      CO
                                                                    _  (D
                                                                    £  _J
                                                                   -  o
              =  o
              £  UJ
              2  co
              '  <
                 CD


              is
              £  o
                                                                  * r
                                                                  i
                   (ssi)  avoi  13QOW  Nisva
                                -156-

-------
I
I

I
I
I
I
I
I
t
  CU
  •H

  5 0.


  Jj
 
f -J


^
fa Q
w
O M
 tJ

 (0
   QD
   fO
   X.

   CVJ
 >,
 ^H
 •H
        UJ
     •u
     (0
 g
•H
w
     ^ •
     o U_
     c
     S a:

     M LU
S o:
"' CO
  LU
00


0)
     !?
                                                        (0  c\j
                                                        GO  05
                                                        O  O
                                                       evj
                                                           m
                                                           Q.
                                                           O

                                                           CO
                                                                       I
                                                                       !
                                                                       i r
                                                                       i c
                                                                       + c
                                                                        c




                                                                        c
                                                                        c
                                                                       * c
                                                                       I r
I (
i r
+ c
i :
i —
i
i
i
i c
i c
+ r
I (\
I r-
I
I
I
I C
I r
                                                                       i c
                                                                       i c
                                                                      + c
                                                                       i r
                                                                       i '
                                                                       i
                                                                       i
                                                                       i
                                                                       i r
                                                                       i r
                                                                      + c
                                                                       i r-
                                                                       i •_-
                                                                       i
                                                                       i
                                                                       i
                                                                        r
                                                                      i  r
                                                                      +  c
                                                                      i "
                                                                      I T
                                                                      I
                                                                      I C
                                                                      I C
                                                                      + r
                                                                       r
                                                                       r
                                                                      + r




                                                                       c
                                                                       c
                                                                      + c
                                                                      i c
                                                                      i --
                                                                            CO
                                                                            CD
                                                                           CO
     5
     tr
     LU
     CO
     CD
     o

     CD
     ^
     O
     <
     o
     _)
     I

     o
                                                                                CO

                                                                                o
                                                                                CO
                                                                               a
                                                                               UJ
                                                                               CO
                                                                               <
                                                                               CQ
                     (S81)
                                         13QOW  NISVQ

                                         157-

-------
  M
  ft

  "
    o
  s i-
  CO
  O
  w
  D

  C
  O
 >
 _J
 <
 o
  M

 CD
 N
 x.
 fo
 x.
 C\J
 (0
 5
 a
 01
 J-.
 rtj
 tH
 3

 -H
 cn
 o
 
-------
    o    As expected, monthly comparisons typically exhibit  better
         goodness-of-fit than daily comparisons, due  to the  greater
         variability in daily loadings.  However, given the  typical  response
         times of the Bay and the tributary estuaries to pollutant loading
         inputs, it is felt that monthly loads are more appropriate  for
         goodness-of-fit assessments.

    o    In general, goodness-of-fit for the verification period  is  as good
         as the agreement between simulated and USGS-based loads.  Since  the
         calibration period was characterized by higher streamflows  than  the
         verification period, this indicates that the Basin-Model provides a
         reasonably good representation of a range of streamflow  conditions.

    o    In general, goodness-of-fit is somewhat better for  total P  and '
         total N than for inorganic P and nitrate-N,  respectively, although
         the results for the P and N fractions are typically quite
         acceptable.

    o    The acceptable goodness-of-fit results for the three basins
         indicates that the Basin Model typically provides as good a
         representation of fall line loadings as the  USGS flow-loading
         relationships*  The major difference between the two techniques  is
         that only the Basin Model can be used to determine  the sources of
         loads delivered to the fall line and to project reductions  in fall
         line loads for upstream water quality controls.

    Comparisons of annual fall line loads of nitrogen and phosphorus for
1974-1978 are shown in Tables 18, 19, and 20 for the  Susquehanna  River
Basin, Potomac River Basin, and James River Basin, respectively.  In
general, the agreement between the simulated annual load and the
U.S.G.S.-based annual load is quite good, with most errors in the range ±20%
and many in the range ±10%.  The agreement for annual loads  of total N and
total P is typically quite good.  The organic and inorganic  fractions of
total N tend to be simulated somewhat better than the fractions of total  P,
probably because the majority of the N loads tends to be contributed by
subsurface flow which is easier to model than surface runoff loads which
accounted for the majority of the P loads.  It should be noted that  some  of
the error in Tables 18 through 20 can probably be attributed to the  fact
that the USGS flow-loading relationships does not explicitly account for
seasonal differences (e.g., differences in cropland loadings that reflect
changes in ground cover)  in nonpoint pollution loadings whereas the  Basin
Model does.  While'the USGS-based loads do account for some  seasonal
differences, such as those associated with relatively high flows  in  the
spring and relatively low flows in the winter, the potential for  higher
surface runoff loadings in early summer than in late  summer  (i.e., due to
different ground cover conditions)  as a result of identical  daily
streamflows can only be represented by the Basin Model.  The fact that the
majority of the simulated P loadings are contributed  by surface runoff may
account in part for the greater differences between simulated and USGS-based
distributions of total P between organic and inorganic forms, since  the
                                   -159-

-------
S to
2§
cc.

a
§ sJ
I§3
CO
8s
z
j
j
=i
z










o
cn
o





















tn

tn























to

























s
to
3














0

XJ
ffi


O1
C

T3
0

1
Q
b



OJ
V
5
£

C

(Q
m
0
*j
a:


?

•o
ro
0
j

X
0
Gu


c


as
O
J
1

O
1*
Cu


0)
•8
z
c

ft
a
jj


a


s

•a
o

i
3
O
Du

*-*
CU


Z

c
W
iTJ



CO
CO
o



0
(N
^
O

10




cn
r*


in

CO
r*
fN

CO
o


_<
CO
CO

m

CO

to


^
cn
to
o

r*
tn



en

0


cn

CO

fN
O



fN



tn
CO
en
^
CO
0
fN
fN


CO

0

o
cn


0






0
fN
tn
m

CO



n
o
CO
o


cn
CN
CO

cn

fN

CO


^
CO
CO
p»
(N
to
txT


in
cn

0


\o
o
CO

fN

fN

cn
fN



in
2
00
en
CO
p»
fN


CO

O

m
o

r-


£


»H

CO


00
cn
m
-
CO

O


ra



il

T3
CO «
0 0
to J
O !

0




11
13
6
S
C

fl
a
0

0
*
j;
T3
CO fl
U 0

3
0

Eu
tH


O
Z
c
-H
fl
CO
o

iQ

i

T3
f« «
S 0

2 i
2
0
6k




0)
•8
Z

C

TI
0
o

•*J
cc

_
c

TJ

CO J
3 J
rH
b

^
0)
"8
z

C
tfl
R]
CO




cn

o


r-



m








o
tn
•^

r-
p-
to
^


m
**

3-
T

fN
O
fN
fN



O
cn
fN
0
m
,— i
m



CN

p-

co
V

en

n





o



fN
CO
^T


fN
CO
0

in
*f
10
*
o
cn
o
CO


a
CO
in
in

fN
to


'


CO
o




m
fN
fN





fN




„
CO
r-

3
to
fN


CO
fN

r-


fN

fN
T



O

CO
fN

T
m


0
fN

*H
cn
p«.
CO

p*

\o






s


0
rH
CO


cn
CO
o

fN
o

»

CO

-------
i!

ai
« S
g S
M U
Sg
Q Si

H S §

|§3
s f j

S 3 <
0

.u
IU

01


•o
tn ra
o o
W J
3 1

0
fe


01
"5
0
"^
c

T}
ffi
0
•vH
^J
a

g

T)
U 0
U) J
=1 1
0
Ex


OJ
•8
*
C
7n
Q
ffl


0



r-

in



r*

CO



(N
CO
m
{N.
m
f>
cn


r-
CN


en

r-
o
in
CO




~*
*o
CO
cn
r*
o
,_j


m CN
O O




CO **



m
m

VO O



^,

r-
p* 
C« CM
r— co

O CO
c- cn



CM CN
~H cn
-H 0

\G CM

r* CN
CO O
O HI
-H CN



GO r-
*O (M
cn o"
r^ in
-H ,H
^H fN




•


O
GO
cn

cn
CO
VJ3





— i

00
0
o
cn




cn
o

m
cn
*-*
cn
*
^



cn
0
cn
<*!

in
^


r-
in




in




m





m
cn
o
en
in
CD




O
^

CO
CO

?
CM
^



VO
in
cn
m
cn
CM
—H


cn








f- 1

cn





^

CO
CO
o
-H
-7


o
CO
0

tn
us
cn
in
CO
f
(N



VO
cn
v

cn
.—4

   0*


  3g
3

Z

Z
  a


  Cfl
5S §
M «

05 <






I
o

•u
5

S

*u
CO "3
LiS
3 1
O
(H
a.

to
?


^
CO
(0
03


O
i-H

CO

r*

o
cn
CO
r*



r-|
CN


m
CM

CO


>H
O
-i

nn

cn

o
cn
v
"H


^
CN

m
in
O

tn
—*

CO
CP
0

CO
o


CO
cn
0



o

-------
                                                                  •o
                                                                co  a
                                                                O  0
                                                                in j
                                                             CL

                                                             8
                      £

                      C '
                                                             2
                                                             •S
                                                             O
£

C
j
u
a
D  «

u  ^
                   S3

                   =  i
                       0
                      ^
                      b

                                                                   "J
                                                                   K
                                                                   O
                                                                  *H

                                                                  b
                                                                                   o    r^    ^
M,  b^  K

S5g
   *T  Hf
J
•£
b



rf
D
Z
CO  g

53
0

4J
a

i1

•o
II

0
Eh


a>
*5
0
z

c

CO

CQ


^
fH

CO
m
cn
m
CN
o
•*



1*1



CO

r-





en
o

^
cn
0
OS
^
r»



m

m


CO
r»




fN
••H
i-*

OS

m
m
^
m



en
CO
o

cn
CO
o




en
o
^

CO

^
CM
m
v



cn
0


CO
CO





in
cn
o

fN
cn
>£>
cn
m
CN
CO



in
o




CO

r-

                                                                co  ra
                                                                o  o
                                                                CO  J
                                                                                   •V    o    T
                      £

                      c

                      M
                                 y\    **    r-*
                                 -I    0>
                                 v    at

                                                                CO  "3
                                                                U  0
                                                                CO  J



                                                                   o
                                                                   ^-4

                                                                   CM

                      a!



                      X
                                          -162-

-------
seasonal differences in instream P transformations  (i.e.,  conversion  from
inorganic to organic forms due to chlorophyll-^ uptake) during  wet-weather
flow periods are handled by the Basin Model, but  not  by the USGS
flow-loading relationship.

    Tables 18 through 20 are also of interest because they indicate the
predominant form of N and P loads delivered to the  outlets of the  three
river basins.  In the Susquehanna and Potomac basins,  the  majority of the
annual N loads delivered to the fall line are in  the  form  of nitrate-N,
whereas organic-N loads predominate in the James  basin.  By comparison,  the
majority of the annual P loads delivered to the fall  line  are in the  form of
organic P for all three river basins, with the highest inorganic P fractions
(%) found at the Susquehanna River fall line.  This examination of organic
and inorganic nutrient fractions further illustrates  the different sources
of N and P loadings delivered to the fall lines.  As  indicated  in  Table  8,
the organic fractions of N and P typically exhibit  much higher  NPS loading
factors than the inorganic fractions.  Consequently,  the majority  of  N and P
loadings delivered to streams by stormwater runoff  is in the organic  form.
While instream transformations during transport to  the fall line can  result
in higher inorganic N and P fractions than were delivered  to streams  by
stormwater runoff, the organic fraction should still  constitute a  higher
percentage of the stormwater loadings of N and P  that reach the fall  line.
The fact that the majority of the annual P loads  are  in the organic form
whereas the majority of the annual N loads are in the inorganic form
suggests that stormwater runoff is probably not the most significant
contributor of N loadings.

    Tables 21 through 24 summarize calibrated baseflow and interflow
concentrations of nitrate-N, ammonia-N, inorganic P,  and BOD, respectively.
As may be seen, subsurface flow concentrations of nitrate-N are relatively
high and typically highest (i.e., on the order of 2.0 mg/L) in  the
Susquehanna River Basin, which has a higher percentage of  agricultural land
than most other major river basins.  By comparison, inorganic P
concentrations in interflow/baseflow are relatively low (i.e.,  0.01 - 0.02
mg/L)  in all river basins.  Thus, Tables 21 through 24 are further
indication of the greater significance of subsurface  flow  contributions  of
nitrogen.  Model production runs that are not described in detail  herein
confirm the nonpoint pollution loading trends evident in Tables 21-24 and
previous tables:  subsurface flow is the principal  source  of nitrogen loads
(i.e., primarily nitrate-N)  delivered to the fall line, while surface runoff
is the principal source of phosphorus loads.

    Results Based upon Upstream and Fall Line Gages.   Simulated and recorded
daily loads were compared at seven USGS gaging stations for the period
1974-1978:  4 stations in the Susquehanna River Basin; 2 stations  in  the
Potomac River Basin; and 1 station in the James River Basin.  These daily
loading comparisons fulfilled two major objectives:   (1)  Since  the USGS
flow-loading relationships used for fall line loading comparisons  were based
upon monitoring data collected from 1979 through  1981, goodness-of-fit
assessments based upon actual monitoring data for 1979-1981 provide an
                                   -163-

-------
i  1
!
en  «
2  3
u
Cd
Q
>
O

£
u
§
X
J
n
W
§
j
K
cu
OS
?
a
H
5
«
2 U
o Q.
-4 >«
CU £•>
z
w
s
1
31
n
yi
o o
fN S
CN fN
0 0
 m
o o

0 0
m m
m m
o o
m in
f\ ci
0 0
CN CN
0 0
CN CN
hH 01
in
r^
o
r-
^-i
0
\o
o o
CN CN
0 0
CN CN
0 0
0 0
m in
0 0
0 0
m m
0 0
m m
•v *
0 O
*»• -«r
0 0
in m
m f*>
0 0
in m
m m
o o
fN tN
0 0
fN CN
•M 03
O •
CN
fN
O
O
fN
a
u»
0
CO
>-t o
o o
CN fi
0 0
^-t o
0 0
CN m
o o
r-4 O
0 0
O O
r^ O
o o
CN m
0 0
CN 0
0 0
0 0
CN 0
o a
T ^
a o
CN 0
a a
*? ^
0 0
^ o
o o
CN m
O O
i-H O
0 0
CN m
0 0
*-4 O
0 0
fN m
o o
r-t O
o o
CN m
o o
•-t o
o o
fN f*>
0 0
— 03
O 0
vO o
CN Cl
o o
un cr>
CN ,-N
O O 0
•^ CO fN
CN CN i^l
in o o
PH r- — <
fN fN r-l
0 0
o o
^ f-t
0 0
0 0
cH <-t
o o
o o
!— t  CJ»
0 0
0 0
en ^\
0 0
0 0
en en
a o
a o
en en
0 0
M 03
a
en
O
("1
m
0
rn
fN
                              -164-

-------
II
II
II*
I
I
I
I
I

                   I
                   I
                   Ik,
m m
o o
o o
o o
o a
m ro
o o
0 0
m m
0 0
0 0
CO CO
0 0
0 0
CO CO
O 0
o o
CD CO
O 0
0 O
in in
o o
0 Q
m m
o o
o a
in m
0 O
0 O
in in
o o
o o
m m
o o
o o
« .


o
o
0
fN
o
in in
0 0
0 0
o o
o o
0 0
0 0
in in
o a
O 0
CO CO
0 0
O 0
CO 00
0 0
0 0
CO CO
0 0
0 0
in in
0 0
0 0
irt in
0 0
0 0
irt in
0 0
0 0
irt in
0 0
O 0
m in
0 O
0 0
IH a


0
o
10
o
in
in in
0 0
0 0
o o
0 0
0 0
0 0
in m
0 0
0 0
CO CO
0 0
0 0
CO 00
0 0
0 0
CO 0
0 0
0 0
in in
0 0
0 0
in in
0 0
O 0
in m
0 0
0 O
in in
o o
0 0
vn m
a o
0 0
n a

a
o
o
i
m in
0 0
0 0
0 0
0 0
m 1*1
o o
0 0
in in
o o
o > o
00 00
0 0
0 0
o o
0 0
CO CO
0 0
0 0
in in
0 0
0 0
in in
0 O
0 O
in in
0 0
o a
m in
0 0
0 0
in in
0 0
0 0
M aj


o
o
o
in in
0 0
0 O
0 0
o o
0 0
o o
in in
0 0
o o
0 0
o o
in in
o o
o o
in in
0 0
0 0
in "in
0 0
o o
0 0
a o
in in
0 0
a o
in m
a o
0 0
in in
o o
o o
» m


in
r^
o
o
•^
in in
o o
o o
in in
0 0
0 0
in in
o o
o o
in in
o o
o e>
in i/>
O 0
0 O
in irt
0 0
0 0
in in
0 0
0 0
in in
0 0
0 0
in in
0 0
O 0
in in
0 0
0 0
in in
o o
0 0
in irt
o o
0 0
M 03

0
fN
CN
0
0
fN
O
o
GO
rH
in irt
0 0
0 O
irt in
0 O
0 0
in m
0 0
0 0
in irt
0 0
0 0
irt in
0 0
0 0
Irt Irt
0 0
C3 O
a o
a o
in in
0 0
0 O
in in
0 O
0 0
in irt
0 0
0 0
irt in
0 0
0 0
in in
0 0
0 0
1-1 O

0 0
0 0
irt eft
fN (N
O O O

-------
    tn

    i
    a.
    tn
    o
    o
    M

    z
.a
CO

E*
    u
    U
    tn
    <
    m

&J

0

s

&
CO
a

J
^
z
^
^
J
IE
OS
s
w

2
«
3 M
O ft*
i?
Z
a
<
CO
1
cn
3

o a
0 0
0 0
0 0
0 O
0 O
0 0
0 0
0 0
0 O
o o
0 0
0 0
0 0
0 0
a a
0 0
0 O
0 0
0 0
o a
0 0
0 0
i- a
o
*r
O

a
o
o
a\

0 0
o a
0 0
0 0
0 0
0 0
a o
0 0
0 0
o a
0 0
CT 0
0 0
o a
o o
0 0
o a
0 O
a o
o o
a a
o o
0 0
n 03
in
r~
«4
a
r~
o
^
,-<

o a
0 0
a o
o a
o o
0 0
o o
o a
0 0
0 0
0 0
0 0
0 0
a o
0 O
o a
0 O
0 O
o o
o a
o o
a o
a a
H a
o
nj

CN tN
O O O
•* 03 CN
CN CN n
in a o
r^ f*. ^H
CN CN m
ul in
0 0
o a
in m
O 0
o o
in in
o o
0 0
in in
o a
0 0
in in
0 0
0 0
m in
0 0
0 0
in in
O 0
o o
m m
o o
« 0
in in
o a
o o
in m
0 0
in in
o a
o o
in in
0 0
ALL COASTAL AREAS I 0
B 0

0 0
0 0
0 0
0 0
o o
0 0
0 0
0 0
0 0
0 0
0 0
0 0
o a
a o
0 0
0 0
o a
0 O
o a
0 0
O 0
0 0
o o
H> ta
a
«r
m
o
m
fl
o
m
CN
                                                                                        3

                                                                                        O
                                          -166-

-------
I
1
1
I
I
1
a
O
a
 I
 I
 I
 I
3 O
n m
N 04
3 O
n in
N fN
3 O
n m
N 04
3 0
n in
N 04
3 0
n m
N 04
3 0
n m
N 04
3 a
N OJ
D 0
n m
N 04
n m
•4 04
3 O
n in
M OJ
3 0
n m
^4 04
3 O
n in
M O4
-i 03


D
9
3
O
3
3
O O
m in
r-4 *H
0 0
m m
— ! p— 1
o o
m m
•H i-4
O O
•H i-4
0 0
in in
*"* '""'
0 0
in m
- -
0 0
m m
»-(
o o
in m

in m
r-t r4
O Q
in in
~" "*
O C3
fH cH
O O
«f — *
- a,


o
o
0
m
0 0
in m
(N 04
0 0
in in
04 CN
O 0
in m
CN 04
o o
m m
O4 CN
0 0
in in
04 O4
O O
m in
04 04
0 0
in m
OJ O4
O O
m m
OJ OJ
m m
O4 04
o o
m m
O4 CN
o o
04 CN
o o
m m
04 04
w 03

O
O
(N
0
O
CD
0 0
m in
04 CN
o o
in m
04 OJ
0 0
m tn
CN O4
O O
m m
04 04
0 0
m in
04 04
O O
in in
CN 04
0 0
in m
04 04
O 0
in m
04 OJ
m m
04 04
O O
in m
04 04
0 0
04 OJ
a o
m in
04 04
M 03


O
Q
O
as
0 O
m in
04 04
o o
in m
04 04
0 0
in m
CN CN
o o
m in
01 04
o o
in m
04 04
0 0
m m
04 04
0 0
m m
CN O4
0 0
m in
04 04
in m
04  «5
O O
>O ^
0 0
o ^»
o o
tO tf
0 0
« ^
o o

0 0
en us
0 O
a\ \o
o
o
^ CO
o
0
,H CO
1-4 O)

0 0
04 m
o o
in ^
04 04
o o a
Tf CO OJ
01 oj m
in o o
f"» r* i— t
CN oj m
0 0
m m
01 OJ
0 0
LI in
CN O4
0 0
m m
CN OJ
o o
in in
OJ 04
0 0
m in
04 04
o o
in in
O4 04
0 0
m m
04 04
O O
m m
04 04
in m
CN CN
O 0
in m
CN CN
O o
m m
04 04
o o
in m
04 04
>-> 01
03
1
COASTAL
»j
J
<
0 0
m in
04 04
0 O
m m
04 rM
o o
in m
CN CN
Q O
04 CN
0 0
m in
04 OJ
0 0
in in
O4 O4
o o
in m
CN O4
O Q
in m
04 04
m m
04 O4
O 0
in in
04 OJ
O O
in »n
CN CN
o Q
m m
CN O4
w OS


O
o
o
O4
                                                                                      u u
                                                                                      ^ tn
                                                                                      z <
                                                                                      K4 CO
                                                   -167-

-------
< o
^.°
a
<
M a:
O 3
?s
S3-;
u z
z z
  o

  o
     W td
     Cd U

     HS
     r u
     z o
     M ^
     2 U
m •»  S

fl tn  o>
o m tn
en <*> CM
^« r- (N

000
u o u
o o o
1 1 1
o o u
z z z



in 03 03
> t 1
£ £ 5



5 ?! £
o o o'

z z z
«<
z z z
3S3
o o o
Z Z Z
in m o
r» o co
o o 5
1) Of 0*
1) 1) D
^ ." 2
see
Cd Ul Cd



in 03 oj
1 1 1
2 £ 5 £
<""""
e2

S en vo
o o a

z z a
<^
z z z
in m in
000
Z Z Z
,n,n a; 
-------
   s

— a1
S3
>• en
J u
H< en
•S 3
O «

en <
en z

z <

O -J

O Ol
        H
        cn u
        Cil O
       s y

       in b
          hH
       ^ Z
       Z U
       < M
       2 cn


       S o
       x
       o J
cn

z
       S-i

       Q
               I
o «r co
>H (*1 O
o o o
o o
< CM CN
zoo
A A
in
< o
z o
VO
O
0
N/A
Lognorraal '
Log normal
•* o *r
pi in co
fi in co
QQ 00 0]
a a a
0)  VO

en eo (a
a a a
CI 01 01
a a a
sag
(0 «J (TJ
en cn cn
ci ci ei
2X2
? 0 IS
0) 01 0)
a. a. a,
01 01  a> tu
-,-.-.
c c c


in ao ca
i t l
§
it

o o o

z z z

z z z
cn in i^
o o o
N/A
N/A
Log normal
n in oa

In" "t? *n
to as as
a a a
41 U 41
III
en cn cn
ex a a
3 £5.
WWW
01 4) 0)
13 jQ J
O O 0
4J 4J JJ
O O O
0 CJ CJ
S E S


ui co co
l i l
cn cn CTN
3
                                                    -169-

-------
  cr
  en
5-
a °"
5 5
W CO
rt U
ya
ii
EU 2;
,' z
  CJ

  I
  CO
     tn a
     u
     s o

     tn &«

     X Z
     z o
     O J

     3£
     h* CJ
     S J
 u
 CJ
 z
 <:
 cj

 b

 z
 o
       b
       O
        a
        ca
        a
r

tn

CN ^ ^
O O O
z z z
III
O o vo en
o o o
z z z
z z z"
-< en en
o' o o

2 Z Z

— ( ^H





333
Ot Of 0)
0*0*04
c c c
fiJ U U
z
u
i
z *n 03 ca
* i i i
a
t-
z

^J- CO CO
O 0 0
a z z~
z z z
O r4 *4
O O O
A

z z z
>n r- CN

>i >. >
a Q Q
0) 0) 01
a,|.|
*o m to
CO CO CO
O" Ot 0
232
333
S*fi
C C C
Cd Ci] Cd

in 03 03
1 1 1
Z
£-

o\ in r-
a o o
z Z z
z Z z
•H *nO O
O O O
A

Z Z Z
fj CN ao
eN in ^o
r* co
en
a
a
0)
e
fl
CO
or
2'
'S u ^?
a, o a,
O
0) 1 U
C "3 C
M X W

m a> oo
Z 1 1 1
|

eN -v ^H
O O O
III
z z z
CO O (N
0 0 0
A A

Z Z Z
m o m
in o in

to en ea
n fl &
a a a
41 Ot 0)
ttt
fl tf 1O
tn en to
01 O o
3X3
•o T; -a
800
01 01 0)
01 01 01
c c c;

m aj co
1 i I
b

m -4 -1
O O O
z z* z
z z z
** r» as
o o o
< < 4
•z z^z

r» o co





333
or 01 ot
c c c
u di a


i i i
3
                                 -170-

-------
II
II
I
I
I
 I
 1
 I
>. 3

S cr
X M

  *»
J in
  I
  I
  I
  I
Cd
O
Z
«c
o
HH
Cu
§0 r- o
^ U"> CN
to o o o
i,
O
J
a
1
a
S
.EVEL OP SIGNIFICANCE
DED SIMULATED OBSER\
)4 N/A N//
73 N/A N/;
19 N/A N/;
— DISTRIBUTION I
TWO-S
3 N/A 0 .
9 N/A 0.
9 N/A 0.

S (N r> co
01
0] 01 OT
IU «J *J
o u y
000
o o o
moo
.H OJ fN
c c c
5m a
j= j=
w ^ **
4J OJ OJ
n3 "3 rtj
4) n co ca
i i i
* R S S
<
S
«r o co
\C M3 CO
O 0 0
^!
^^^
z z z
CO GO O\
0 0* 0*
< < <
Z" 2 21
•H m oi

m rn co
 en CB
<
^ «» CT\
r- H
03 05
U-! U
O O
o o
O 0
O 0
04 OJ
C C
55

 ai
NITRATE-N
5 Ent
78 Ent
8 Ent]
•*• i i t
t- c* en a*
a:
f
H- *
Z
9 ^O CO
^H ro O
000
XXX
z z z
^<^
z z z
OS -V i-^
0 0* O
< rf^ <
z z z
o tr> -H

PJ m ^H
In
<->4
(0 CO CO
«-l  CO CO
( f 1
ON o\ cr»
Z
^
fr-
ui rn en
in ro \o
0 0 O
\\\
z z z
<^<
z z z
000
xxx
z z z
^ CN ^O

r-i OS OJ
~4 — (







333
4» H
o
0 0
\ * •
zoo
o
x. -:
z o
A
in
L4 N/A 0.
11 Lognormal *
16 Log normal *

m os oj
rH .H







TJ "S T5
000
0) 01 0)
OJ 1> 0)
c c c
W Cd Cd
m oj CD
1 I 1
as o* os
§
oj rn co m o
ao m i£ •-* v
o o O o o
X X X X X
z z z z z
< < < < <
X X X X X
z z z z z
co -H —* ^* en
o o o o o
^<<^^
z z z z z

CN m (*S 00 i-4
(-4 i— I
CO M Ul W M
o y y o o
o o o o o
o o o o o
m o in o ui
*-( 04 OJ C* (N
c c c c c
5 5 5 J: 5

a  a)  a> 
-------
s-
cn
u
EH
s
3
to
x
z
2
z
0
X

U
»-i
3







H
— . en
— . a
"S 6"
e

_ cr o
X M «
< ^
Q -H 3
X, tfl 03
3*G l— l
- &
J » EH
•— W

en •• o
§0
0 CO
M Tl t
£« <0 K
33
as
M *•
CO
V
a 01
< < o
hj en
M D
g- <
«i 2
o
tt. -
en
co a:
CJ L> t
M O
fri 0! Z
en
So
•C
EI e-i
to z
5 2
h
1 6"
&• -C
o
1 CB
to a
en >
w M en
z a e-
§^H
O I
O £ iJ

§ S
at en
<
H
«
a











g
CJ
M
CC
u
a.
H
•z
01
g
r"
f*
tn
3
8

H
CJ
z
o

b
Z

t— t
en
i.
o

J
g
2



S
z
g
b
i— (
Z
O
K*
en

b
O


1
J






g
M
1
»_<


u
M
a



n
o
3




E
H
n
























































n en en
*-* <3\ I3X
0 0 0







S rf < <
a x x \
u z z z
en
o
O
a
u



3 Z Z Z


en


g
9 ° S ^1
en ...
1 O O O


f*


x. \\
z z z









^O (J» Ol
i-H -H ^




^O CT f*J
f— t ^ ^*












OS W UJ
>" J1 >
a a a


^ ^ i— f
»*&
fl 113 fQ
cn w 05

o o a

3! 3 3


o cu o
•-^ .a -w
u O b
(U -u (U
a. o a,
O
01 1 01
-^ o —
C 1J C
W £ U






m as as
r- r*- r-
i i i

0%  >o r--
iTl ^H CN
o o o







< 0
"x • *
z o




o o
«€ <-* CN
"v, • •









o o o





xlxl \
z z z









»H CN •*




O «> (N
rn CN *r












u u]
^ 51
a a

 >t >-
a a a

(U 
1»
U
«
-172-

-------
o
2
     _ a
     x _
     3  £
     2U
     « en
     < o
     a en
     H
55f
000
f* H fr"
.. „ _
(0 (O rt3
a a a
V 01 
jj JJ JJ
U U U
Z Z Z
	
a a a
(0  a d
322
333
tU 4» 4»
a* a. o-
01 Ol 41
c c c
Ed U Cd


tn co co
t f t
£ S £
2
                                                     -173-

-------
r
                       a

                       §
                          a

                          05
                          M O




                          5 1
a
<; ai

3 «T

*°


d|
a 3



Is


U H
                          as
                          t> 05
                          •t X
                          ft.
                          b <
                          U


                          §
                               H  t
                               05 Ml


                               gz1'


                               SS
                               z u


                               S U]


                               2 b
                               O O
                               X

                               O J
                                6"


                                E"
^
^H
« * *
0
in
fS O
• * •
o o
« « «
0 Lognormal N/A
6 Lognormal N/A
6 Lognormal N/A
o *o us
In" "w" "a?
>i >. >
rtj TJ rtj
a a a
at a> d>
a- a-*
flj fl rtj
Ul CO W
O O Oi
— 3 3
0) D   > >1
fl (0 fl
a a o
a) a» cu
« 05 05
a a a
3 Z 2
131
a v a
a. a, a.
4> V 0)
c c c
Id U U
U°l 00 CO
? i 7
2 r^ r- r-
cTi OS <7\
J-.-V-,
r*« ax co
o r~ *4
odd
< < <
Z 2 Z
<^<
z z z
p^ ao ^
o in •-»
000
< < -c
\ x^ \
z z z
m r* o
»-4
m r- o
^•i
"00" *«" '«'
>*>•>.
itj fl tg
Q a a
0) V V
a-a-t
fl "j m
05 en 05
01 a a
s s a
m
o; a/  1 1 1
M r* r- r-
E-i o\ ijN <^
ffi
^
Z
-H r^   >. >t
HJ fl (TI
Q Q a
01 0) D
aa.a
us cj >d W U
oi at o>
a* cw cu
01 4) 0>
U W U
c c c;
ca en w
in 03 co
i t I
r- r*- r-
 tn ON
S
b
kO CO  >t >i
ttj fO iTj
Q 0 0
0) 01 01
a §4 a
Cfl U3 tfl
01 cu a
5 £ 2
D J3
m cr> ^
O O O
<^
z z z
<^^
z z z
d d d
<^
z z z
(N VJ3 CO
O l/> t/1
*r r^ -^



CJ 41 J
j ^ ji
o o o
u o o
o o o
iii
o o u
flj (a ij
Z S S
g-t CO CO
1 1 1
r* r* r^
en c« a\
U-
                                                             -174-

-------
opportunity to test the reliability of the fall  line  loading  comparisons  in
Figures 31 through 84 and Tables 18 through 20;  and  (2) For the Susquehanna
and Potomac river basins, the USGS gaging data for 1974-1978  permits
assessments of goodness-of-fit at upstream locations  in the river  basin.

    Recorded daily loads were based upon the product  of the mean daily  flow
and instantaneous concentration (i.e., grab sample)  reported  for each USGS
gage.  Simulated daily loads were based upon the sum  of hourly loads
calculated from the product of simulated hourly  streamflows and hourly
concentrations.  Therefore, in addition to error associated with gaging
measurements, some of the difference between simulated and recorded daily
loads can be attributed to the use of an instantaneous concentration for the
recorded dataset and to errors in streamflow' simulations.

    Comparisons of simulated and recorded daily  loads were based upon the
following simulated datasets:

    fl:  Entire Period (January through October of each year)
    #2:  March-October
    #3:  Entire Period (Water Quality Sampling Days Only)
    #4:  March-October (Water Quality Sampling Days Only)

Comparisons based upon simulated datasets fl and #2 are felt  to be the  most
rigorous, since they involve comparing simulated loadings on  each day of the
period (e.g., 1974-1975;  1976-1978; 1974-1978)  with recorded  loadings on
sampling days,  in other words, due to the relatively low water quality
sampling frequency during 1974-1978, evaluations based on simulated datasets
#1 and #2 involved comparing a relatively small number of recorded values
with a relatively large number of simulated values.  The assumption made for
goodness-of-fit assessments based on datasets fl and  #2 is the same one used
to support most monitoring programs—statistics associated with recorded
data collected on a weekly or biweekly basis is assumed to be representative
of long-term statistics.   Goodness-of-fit comparisons based upon simulated
datasets #3 and |4 involved matching recorded data with simulated values on
each sampling day.  In general, goodness-of-fit statistics based upon
simulated datasets #3 and #4 were higher than statistics based upon
simulated datasets #1 and #2.  Likewise, because they represented a little
more than two seasons rather than four, goodness-of-fit statistics based
upon the March-October period (datasets #2 and #4)  were typically higher
than statistics based upon the entire period (datasets fl and #3).  In  the
goodness-of-fit tables that follow, results for only one simulated dataset
are reported.  The selection of the simulated dataset which is reported in
the goodness-of fit tables is based upon the hierarchy shown  above, with the
most rigorous evaluations based upon dataset fl and the least rigorous  (but
still acceptable)  evaluations based upon dataset #2.  That is, dataset  fl is
considered first,  followed by f2, f3, and f4 in that order until a
satisfactory goodness-of-fit statistic is obtained.  Since the first
simulated dataset producing satisfactory statistics is reported, different
simulated datasets are sometimes reported for different constituents-   It
should be noted that, since the Basin Model is intended for simulations of
                                    -175-

-------
the period March through October, the comparisons based upon  simulated
datasets for the "Entire Period" (January-December) are probably excessively
rigorous.  Likewise, since the calibration and verification periods were
treated separately, goodness-of-fit statistics for the 1974-1978 period are
not as important as statistics for the other two periods.

    Box-and-whisker plots were developed for the simulated and  recorded
datasets for graphical assessments of goodness-of-fit.  Figures 85 through
91 present sample box-and-whisker plots for streamflow and P  and N loads  at
the Susquehanna River gage at Harrisburg.  As may be  seen, separate plots
are presented for the calibration (1974-1975)  and verification  (1976-1978)
periods and for the two periods combined (1974-1978).  The simulated
datasets in these figures are based upon the Entire Period (dataset fl),
while the recorded dataset is based upon the same period.  Based upon an-
inspection of these figures, there appears to be good agreement between
simulated and recorded datasets for all three periods for all constituents.

    Nonparametric statistical tests were used to quantify goodness-of-fit
for the daily loading comparisons at each gage.  Tables 25 through 31
present the goodness-of-fit statistics for the.seven  gages.   The entries  in
the "Dateset Limits" column indicates the simulated  (i.e., fl,  |2, 13, or
#4) and recorded datasets used for goodness-of-fit assessments.  In other
words, if simulated dataset #2 was used to derive the statistics reported in
the table, "March-October" appears in the Dataset Limits column and both
simulated and recorded data were restricted to March  through  October of the
indicated years.  In cases where a satisfactory level of significance was
obtained from the two-sided K-S test, the value is reported in  the column
labeled "Two-Sided."  In cases where the two-sided K-S test produced a level
of significance less than 0.05, one-sided K-S tests were performed on both
the simulated and recorded datasets to determine whether they both exhibited
a  lognormal distribution, with the resulting level of significance values
reported in the "Simulated" and "Observed" columns.  The one-sided K-S
statistic is reported in cases where graphical comparisons of simulated and
recorded distributions indicated similar distributions, but outliers in the
datasets resulted in a relatively low level of significance for the
two-sided K-S test.  In cases where both the simulated and recorded datasets
are shown to be characterized by lognormal distributions, the Wilcoxon Rank
Sum test may be used to evaluate whether there is a  significant difference
between the two datasets assuming that their variances are approximately
similar.

    Based on a 0.05 probability cutoff for the 95% confidence interval, the
level of significance statistics in Tables 25 through 31 indicate that there
is typically no significant difference between simulated and  recorded daily
values for all three periods.  Only total P at the James River  gage exhibits
an unacceptable agreement between simulated and recorded daily  values.
Potential sources of error in the simulation of total P concentrations on
sampling days at the James River at Cartersville gage are presented below.
In summary, the acceptable level of significance values in Tables 25-31 are
further evidence of the reliability of the earlier comparisons  based upon
                                    -176-

-------
I
I
I




CO
O>
«— «
•— 1


CO
1-4
I
1— «




en
1
*r
en



D
VJ
O
3

1 1 1 1 I


JTI
tLr









Ii
I



JTli
ilf





i r t i 1
M
0
w
o
• i~H
s ft
00 s
w
•H
3 2
O
C >*-!
, (0 E
._ 03
i § 2
O CQ 4J
CO r-t
O -H
-^ m
Q
0)
_.' i-l (1)
S D en
*r w («
CO XI O
o
T? di
c
tn
OBS.
Lmulated (SIM)
: at Harrisbun
CO 0)
i2 ° *
1/1 to 03
C C
0 C
tn m
M 0)
(C 3
i1 of
O 3
O w
00
ca
o •
CO
0)
en
•H
ro r*^ o <^ f*^
00 *^«f *~^ ^^ ^^
^O *^ ^* ^J ^^ CD
o o o* o o o
(SJO NOmiW) M01JWV3HJLS A1IVQ
                                                   -177-

-------
CO
(*•»
CTV
00
r*.
                                                                       o
                                                                       4-1


                                                                       M
                                                                       -P
                                                                       0
     W

     •H T3
     T3
     C 0<
     ,
     m r-<
     o -^
     •^ a3
       Q
     T3
     0)

 *   MO)
2.   0) tji
^   t/} o3

-------





00
en
1— 1
I
*r
01
r— 1







00
r-.

^•4
|

VO
cn
***









if)
i —
cn
•"•4
i

]*
cn
r— 4






O
£>
1 1 1 1 1








i faTl
i i"















.1 11

L_Lr




' I a 1 1 1
i°i r







. .^" "Tli

Im^MM




1 1 1 1 1

o
OT
*J
0
SI (
p^
•— t
00 ^
01
M
•r-( 73
*3
C S
(• iH
OO X ra
CD O 4->
o ca o
EH
03 >,
CQ r-l
O •H
~- ro
Q
Q)
to^ rt) ^|
to W ro
JQ C?
O
T3 ft
C
rB -
Cn
5 3
oo' co w
CO *~^ "^
c M
0) _jrj
(0 ""
rH 4J
3 rrj
CO 0)
4-4 -S
2-' 0 OS
*~? cn r8
1/7 C C
O C
W ra
•H ,-C
I-J 01
rd 3
E to
O 3
U CO
00
03
O r-
CO
0)
i-l
cn

O O O O O 0
in *r <"i evi •— > o
f
                                           Nomiw)  avoi  N  ivioi
                                          -179-

-------


t— t
I""""" 1.
1 i 	 I1

i ETIi
. tU
i

m
CTi
1
cn
_
1 a T 1
Ll, '

o to o ui o to c
m CM CM <— ' — • O c
M.
isker Plots foj
itrate-H Load

<
• «•
) Box and (
Nitrite +
,
0 0 -H
Q
T)
-( 0)
00 (03
a cr
6 tn
0 3
o w
CO
CO
•

-------
N
n
 i
  i
  i
  i
  i
1 1 1 1 1



1 HI
CO

CTi

1

^*>
f****
O^ f^
— « i nfi
Lu







CO
en
**"*

i
en pp
ll










in
^^•1.
cr>
»• *

i

^^
r^«»
Cft f**T


IMHIB







o
w
-J-J
o
t»4 ^
t^l 0)
^ TJ
W 03
53
3:
•z,
fO 1
C ^
OJ -H
1' • G
to x o
m og
o « J
en >,
ffl ^H
O -H
~— a)
Q
a)
S- (-1  a
H 4J
3 rt3
g
•H ^
W (1)
,£ 'S (5
J"? CO (tl

o c
en <3

M 0)
IT) 3
ft 01
e w
O 3
0 en
•
CO
ca
O en
CO
0)
Vl
3
tr>
fa
O  *t f^> CM «— ' O
• • • • • • •
o o o o o o o
                                    (AVQ/'S91  NOmiW)   (Mil   N-VINOWWV  A1IVQ
                                                 -181-

-------
                                                                   Sn
                                                                   0
00
co
UD
                                                       -Q
IT)

en
CTi
                                                        ^F
>ffl
                                                                   en
                                                                   4J
                                                                   0
X 73
 M (TJ
                                                                  3
                                                                     2
                                                                  'u I
                                                                  C O
                                                                  IT) -H

                                                                  X rfl
                                                               vi  o DI
                                                               CO  CQ S-i
                                                               o     o
                                                                  O -H
                                                                  •— (0
                                                                     Q
                                                                   M 0)
                                                                   0) CTi
                                                                   w -i

4J ^
ITS

3 nj

•H ^
W 01

u-i -H
O «

en n3
c c
o c
en ID

>-( 0)
(T3 3
a w
E tn
0 3
U CO
                                                              on

                                                              0
                                                              °
                o
                           en
                                                Lf>
      CM        CM         —i         •—'         O

 (Ava/'sai   Kiomiw)   avoi   N   :iNvsao   AIIVQ


                -182-
                                                           o
                                                           o"
°
01
                                                                  3
                                                                  cn

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 i
00
r-.
•o
                                         00
                                                      •0
                                                              oo
                                                             CO
                                                             en
                                                             o
                                               CM
                                                         CM         i—i         i-t         O

                                                            Nomiw)   avoi   NXI  AIIVQ
                                                                                                    o
                                                                                                    o
                                                                   S-l
                                                                   o
                                                                   w
                                                                   JJ
                                                                   o
                                                                   IS
                                                                  w
                                                                  •H
     C  nj
     ifl  O
     x  ^

     o  ^
     a  2
                                                                  w  >,
                                                                  CQ r-i
                                                                  O -H
                                                                  ^  (0
                                                                    o
                                                                                                              O)
                                                                  TJ
                                                                  0)

                                                                  M
                                                                  0)
                                                                  W  rd
                                                                  ga
                                                                                                            c
                                                                                                            n)
                                                                  E 3
                                                                  M X2
                                                                  C/3 Ul
                                                                  'O i-)
                                                                  
-------
fall line loading relationships and of the Basin Model's ability to
represent important pollutant loading/transport processes in the major river
basins.

    It is felt that the poor fit of simulated and recorded total P
concentrations for the James River at Cartersville gage can be attributed  in
large part to errors in the wastewater treatment plant discharge data  (i.e.,
municipal and industrial)  used for the 1974-1978 simulations.  During the
calibration study of the James River Basin, the wastewater discharge files
were changed three times by EPA/CBP staff in response to inquiries by NVPDC
staff based upon poor agreement between simulated and recorded concentration
data.  Sensitivity study runs with the calibrated James River parameter set
revealed that the deletion of all nonpoint pollution loadings (i.e., surface
runoff plus baseflow/interflow)  resulted in only 10%-20% reductions in the
simulated mean concentration of total P and inorganic P for the calibration
and verification periods.   This indicates that the assumed wastewater
discharges are the most significant contributors to simulated concentrations
for the majority of the calibration/verification period, and therefore that
errors in the discharge records can have a very significant impact on
comparisons of simulated and recorded concentrations.  As indicated below, a
comparison of recorded concentrations for 1974-78 (calibration/verification
period) and 1979-80 (USGS fall line monitoring study) indicates that the
latter period exhibited higher mean total P and inorganic P concentrations
even though the mean streamflows were on the order of 30%-90% higher than
those encountered during the earlier period.

                              RECORDED TOTAL P       RECORDED INORGANIC P
          PERIOD             N            MEAN        N            MEAN
    March-Oct., 1974-78      69        0.086 mg/L     19        0.022 mg/L
                1979-80      47         0.12 mg/L     39         0.05 mg/L

The much higher inorganic P concentration for 1979-80 is particularly
suggestive of different wastewater discharge contributions, since the
majority of nonpoint pollution loadings of phosphorus are  in the organic
form.  The simulated mean total P concentration for 1974-1978 is in good
agreement with the recorded mean for 1979-80, while the simulated inorganic
P value exhibits much better agreement with the 1979-80 mean than with the
1974-78 mean.  Given the significant differences between recorded mean
concentrations for 1974-78 and 1979-80 and the relative insignificance of
nonpoint pollution loadings on simulated means, more extensive checks of the
wastewater discharge records assumed for 1974-78 are required before any
conclusions can be reached regarding goodness-of-fit of total P
concentrations for the calibration/verification period.  In the absence of
further checks of the wastewater discharge files, it can be stated that
goodness-of-fit assessments based upon the USGS fall line  loading
relationships indicate that total P loads from the James River Basin are
satisfactorily simulated on an annual, monthly, and daily  basis.
                                   -184-

-------
I
I
f
I
I
I
I
I
I
1*
I
I
I
I
I
    Since the main stem reservoirs near the mouth of the Susquehanna River
were found to have a significant impact on the delivery of point  source
loadings to Chesapeake Bay, agreement between simulated and  recorded data
was checked for Conowingo Reservoir which is the largest of  the three  major
reservoirs and the one with the most extensive database.  Table 32
summarizes simulated and recorded means and medians for Conowingo Reservoir
for the period 1974 through 1978.  The simulated means are based  upon  each
day of the five-year period, as indicated by the relatively  high  N's
reported in the table.  As may be seen, simulated and recorded values  are
typically within approximately ±20% for phosphorus, nitrogen, and
chlorophyll-^.  The satisfactory agreement between simulated and  recorded
chlorophyll-_a is particularly encouraging, in light of subsequent model
production runs that indicate that chlorophyll-a levels in the lower
Susquehanna Reservoirs are one of the controlling factors for upstream point
sources.  For example, production runs that examined the impacts  of a
reduction in wastewater discharges of phosphorus indicated that the
resulting reduction in chlorophyll-_a within the lower Susquehanna Reservoirs
would pass higher loadings of dissolved nitrogen through the reservoirs  into
the upper Chesapeake Bay.  While this response seemed nonintuitive at  first,
further study revealed that the higher nitrogen loadings could be attributed
to the reduction in chlorophyll-a concentrations which reduced nitrogen
uptake conversion to suspended organic forms which could be  removed within
the reservoirs.  Similar trends have been noted in the upper Potomac Estuary
where reductions in wastewater discharges of phosphorus have reduced
chlorophyll-ja levels, thereby resulting in the delivery of higher dissolved
nitrogen levels to downstream reaches.  The relatively good  agreement
between simulated and recorded chlorophyll-^ concentrations  for the
calibration period, particularly the fact that the model typically
undersimulates rather than oversimulates chlorophyll-a in Conowingo
Reservoir, indicates that the Basin Model should provide a relatively
accurate characterization of secondary impacts of management strategies  that
affect algal growth in the lower Susquehanna Reservoir system.  Nonetheless,
since chlorophyll-a levels in these reservoirs appear to have a significant
impact on nutrient delivery to upper Chesapeake Bay, field studies of
existing eutrophication levels (e.g., follow-up studies related to _£3) and
relationships to long-term pollutant trap efficiency might be considered for
future funding.

    EPA/CBP did not supply any recorded chlorophyll-a^ data for comparison
with simulated data for the Potomac River and James River fall lines.
However, a general assessment of chlorophyll-^ simulations for the Potomac
River fall line can be based on the boundary conditions used for  previous
modeling studies of the upper Potomac Estuary by the Metropolitan Washington
Council of Governments (44).  Based upon analyses of recorded data and
Potomac Estuary Model calibration studies for periods of relatively low
streamflow, a mean chlorophyll-ji concentration of 25 ug/L is used as the
boundary condition for the Potomac River fall line.  By comparison, the
simulated mean concentration of chlorophyll-^ at the Potomac River fall  line
for the calibration/verification period (1974-78)  is approximately 16  ug/L.
By raising the chlorophyll-_a/algal phosphorus ratio by a relatively small
                                                -185-

-------
                      Table 32

COMPARISON OF SIMULATED AND OBSERVED CONCENTRATIONS
              FOR CONOWINGO RESERVOIR:
                 1/1/74 -  12/31/78
Parameter
TP
TN
BOD2Q
DO
NH3
NO 3
ORG. N
ORG. P
P04
Chla
Parameter
TP
TN
BOD
DO
NH3
NO 3
ORG. N
ORG. P
P04
Chla
Sim
Mean
0.058
1.250
3.341
11.156
0.063
0.939
0.247
0.035
0.023
11.841
Sim
Median
0.045
1.25
2.50
11.00
0.060
0.95
0.20
0.03
0.020
10.00
Sim
S.D.
0.030
0.361
2.733
2.577
0.043
0.268
0.165
0.023
0.016
4.6
Sim
MAX
0.342
4.203
29.786
15.888
0.386
2.067
1.774
0.253
0.183
46.906
S im OBS OBS
N Mean S.D.
43,824 0.075 0.066
43,824
43,824
43,824 9.4 3.3
43,824
43,824 0.775 0.469
43,824 — — •
43,824
43,824 0.026 0.039
43,824 14.82 12.64
Sim OBS OBS
Min Median MAX
0.029 0.059 0.400
0.522
0.354
4.225 9.6 15.2
0.005
0.286 0.70 2.482
0.032
0.005
0.010 0.010 0.230
2.198 10.80 52.87
OBS Sim Mean
N OBS Mean
67 0.77

__
68 1.18
—
67 1.21
—
—
46 0.88
70 0.80
OBS Sim Median
Min OBS Median
0 0.76
—
—
1.0 1.14
—
0 1.36
—
—
0 2.00
0.80 0.93
                      -186-

-------
I
I
f
I
I
I
I
I
I
I
I
I
I
I
 f
percentage, the simulated mean chlorophyll-_a levels for  the Potomac  River
fall line could be increased to more closely agree with  the uptake of
nitrogen and phosphorus in the Potomac River.  Since the  simulated mean
chlorophyll-_a concentraton is in the same ballpark and less than  the upper
Potomac Estuary boundary condition, it appears that the Basin Model  should
provide a reasonably good representation of management strategies which  will
affect chlorophyll-^ levels in the prototype.
                                                -187-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 1
                                  CHAPTER VI

                       FRAMEWORK FOR MANAGEMENT STUDIES
Introduction

    Following the calibration/verification of  the Basin Model,  it was  used
for production runs to assess the impact of alternate management strategies
on pollutant loadings delivered to the bay system.  This chapter covers  the
periods selected for the management studies, the interpretation of model
output, and a discussion of typical production runs for which the model  can
be used.

    A detailed summary of actual production run results is beyond the  scope
of this report.  Several management study runs were performed to generate
loading data for input to the stand-alone models of the Bay's estuarine
system.  Time series output files for these management studies  have
previously been transmitted to the Virginia Institute of Marine Science
(VIMS)  for input to the receiving water models of the Bay system.  In
addition, tabulations of simulated fall line data on seasonal loadings and
concentration-frequency relationships have been transmitted to EPA/CBP for
analyses and inclusion in upcoming EPA/CBP reports.  However, since the
budget for this modeling study did not cover a detailed report on production
run results (i.e.,  scope of study was restricted to transmittal of time
series output to VIMS), only the framework for model production runs is
presented herein.

Periods Selected for Management Studies

    The selection of the periods for which the Bay model package would be
operated was an important decision that was based on the type of management
strategies to be studied.  To provide an indication of long-term impacts of
point and nonpoint sources, the model package  should be set up to represent
a particular point/nonpoint management strategy and then operated for  the
period of record for available rainfall data.  The frequency of water
quality criteria violations simulated by the Main Bay model would represent
the long-term impacts of each management option.  In the case of low flow,
high flow, and long-term assessments, model operations should concentrate on
receiving water quality impacts during the spring, summer, and fall which
are the seasons when water temperatures are high enough to result in the
most critical water quality problems within the Bay's estuarine system.
Summarized below is the methodology for identifying the periods used for
management strategy studies.

    Analyses of Long-Term Point/Nonpoint Source Impacts.  It is technically
possible to execute the Bay model package with many years of rainfall and
streamflow records in order to assess long-term water quality impacts of a
particular management strategy.  However, since the costs of long-term
simulations with state-of-the-art models can be very high, it is becoming
                                                 -188-

-------
common practice to identify a short period (e.g., a full year; several
seasons)  which can serve as a less expensive surrogate for the multi-year
period of interest (If45_,46) •  This shorter "typical" period should be
characterized by streamflow and water quality statistics that are  reasonably
close to the period of record for the meteorologic data which would be  used
for a long-term simulation.  Production runs of the model package  with
meteorologic records for the typical period are then assumed to produce
streamflow and water quality statistics which approximate the statistics
that would result from a model production run covering the entire  period.
In terms of streamflow statistics, this typical period can also be referred
to as a period of "average wetness" since its flow-duation curve will most
closely approximate the long-term curve; however, the term typical period  is
felt to more accurately describe the applications for continuous simulation
model production runs.  Since the use of a "typical" period permits
simulations of long-term water quality impacts at a reasonable cost, this
approach was selected for management studies with the Basin Model.

    Since the spring, summer, and fall seasons are most critical from a
eutrophication management standpoint, it was decided that the assessments  of
long-term water quality impacts should focus on the seven-month period
extending from April 1st through October 31st.  Because the available
meteorologic record for Basin Model studies covered the period 1966-1978,
the selection of the typical period was based on analyses of the daily
streamflow statistics associated with April through October of each year in
this thirteen-year period.  The year with April-October streamflow
statistics which come closest to the statistics for the full thirteen-year
period can be designated as a typical year for assessments of long-term
impacts with the Bay model package.

    In order to ensure that the selected typical period was characterized  by
similar frequencies of occurrence for a range of daily flow levels,
simulated daily flow-duration curves served as the basis for comparison of
individual years with the full thirteen-year period.  The fall-line
monitoring study by the U.S. Geological Survey (11) had previously
demonstrated a positive relationship between daily streamflow and  pollutant
loadings at the mouths of the Susquehanna, Potomac, and James rivers.   The
USGS study produced  regression equations relating streamflow and pollutant
loadings which can be used  to produce loading-frequency relationships from a
daily flow-duration curve.  Therefore, a year characterized by a daily
flow-duration curve  that approximates the streamgage's long-term
flow-duration curve  is also characterized by loading-frequency relationships
which approximate long-term loading statistics.

    Simulated flow-duration curves for individual years were plotted with
the simulated flow-duration curve for 1966-1978 and the assessment of
similarity in distribution  was based upon visual inspection.  Comparisons
were based upon simulated flow-duration curves for streamgages located  at
the mouths of the Susquehanna River  (25,990 sq mi), Potomac River  (11,560  sq
mi), and James River  (6,257 sq mi) and a composite flow-duration curve  based
upon the sum of simulated daily streamflows for these three river  basins
                                    -189-

-------
(43,807 sq mi).  Based on a preliminary screening/ the following three years
were given detailed consideration for designation as the typical year:
1971, 1974, and 1976.  As may be seen in Figures 92 through 95, each gage's
flow-duration curve for the period 1966-1978 is most closely approximated by
the curves for 1974 and 1976.  Since 1974 exhibits closer agreement with the
long-term plots for the Susquehanna, Potomac, and "summation" gages, it was
selected for use as a typical year.  In other words, Basin Model executions
with meteorologic conditions for the period April-October 1974  (preceded by
executions for a sufficient antecedent period)  should produce strearaflow and
loading statistics for Bay Model input which closely approximate
April-October streamflow and loading statistics that would result from
executing the Basin Model for a full 13-year period (1966-1978).

    Analyses of "Worst Case" Point Source Impacts.  As suggested above,
model operations for an extended low streamflow period can be expected to
provide the greatest insights into the Baywide impacts of wastewater
treatment strategies for the Chesapeake Bay Basin.  Selection of a "dry
year" with a spring, summer, and fall characterized by an extended low flow
period was based upon comparisons of April-October flow-duration curves for
each year in the period 1966-1978.  Based upon a preliminary screening, the
following 3 years were selected for detailed analysis:  1966, 1968, and
1977.  In the case of the Susquehanna, Potomac, and "summation" gages, .
extreme low flows occur more often in 1966 than in 1968 and 1977.  In the
case of the James, 1977 appears to be characterized by the most frequent
occurrences of extreme low flows.  In terms of total streamflow volumes
during the April-October period, inspection of the areas beneath the
flow-duration curves indicates that 1966 exhibits the lowest total inflows
to the Bay for the Susquehanna and "summation"  gages,  while 1977 exhibits
the lowest total inflows for the Potomac and James gages.  Because it is
characterized by the highest frequency of extreme low flows in the two
largest river basins and the lowest overall volumes for the sum of
Susquehanna, Potomac, and James basin streamflows, April-October 1966 was
designated for dry year production runs of the Basin Model.

    Analyses of "Worst Case" Nonpoint Source Impacts.   Model operations for
a "wet year" characterized by relatively high streamflows provide the
greatest insights into "worst case" impacts of nonpoint sources of
pollution.  Based upon a preliminary screening of April-October
flow-duration curves for each year in the period 1966-1978, the following
three years were selected for detailed analyses:  1972, 1973, and 1975.  The
simulated flow-duration curves for the Susquehanna River Basin are shown in
Figure 96.  A review of simulated flow-duration curves for the Susquehanna,
Potomac, and James basins and the summation dataset indicates that 1972 is
the wettest year.for all four gages, with 1975 the second wettest.   It
should be noted that the period April-October 1972 includes Hurricane Agnes,
a storm event with a very low probability of occurrence.  Since a period
which includes an event as rare as Hurricane Agnes may not be an appropriate
"design condition" for basinwide assessments of nonpoint pollution controls,
the second wettest year, 1975, was selected for wet year production runs.
                                    -190-

-------
»*•
•0*


•am


*
•a »»
V
•o »


< • —


•o • —



•0 • —



o « —


•oin ^ —



•c • —

•c « •»• —
.


•c sn ^ —



CO *O V5 ^ —
r-

,^
"H T kO U3 W »
Q r- t— r- vo
2  
Q
Cil
£-•
<
U
M
Q
i— i
8

•j
<'
ex
u
a:
0

z
<
?
OS
a
E*
<
rj
o:
u

en
tH
1
CL,
^
H
OS
S^
cn
a
s
M
r j
O
'R ACTION
b









Vj
1)
>
.,_4
ai
i a)
J 2
•H O
<13 4J
Q 5
113 -
 1-1 '^
sj a M
3 < >i
0 ~ -H
ra
r < c
C 04 <
o
w-Durati
arietta,
cal Year
0 S -H
—( Q<
fc< 4J >,
r <0 c->
CN
cn
4)
S-l
3
01
.H
U4



 MOT.IWY3KXS
-191-

-------
 I
 I
 f
 I
 I
 I
 I
 I
 I
I
I
I
I
i
                                                      03
                                                      [*»
                                                      01
                                                      >H
                                                      I
                                                 ^O   vO
                                                 P>   «D
                                                 CTV   O1
                                                 i-l   .-I

                                                 II    II

*
»
VI


. «»


«

Ul
^


^


^



























.


S
* o>
c
c
o.
•a
 ,
1-1 x:
•H O1
fl 3
Q 0
u
T3 £
CJ 4-1
T3
W 4J
0 OT
O r-t
(11
W
OS rH
•H
(, ti
o a
U-l <
10
0) • CO
> O -H
^ • w
3 Q >,
O -H
" 2
"c o <
0 4J
•H CTi W
4J C nj
n3 -ft ID
14 ^ ><
3 31
C oj — i
1 S a}
3 0
o u •*
>— < .TJ Q<
t, 0) >,
: C EH
n
(31
0)
tj
3
S
A
3
                                                           2

                                                           fi
                                                           •o
                                                   O
                                                                        -192-

-------
•
V U»
«*• 10
'

» «3


*«0



•»«


^O


*• *o



, *




* v:




i




« •* vo




^ **



'
CO
p*1
0^ »•> ^ u»
(
Q r- r*» r- ^o
2 ^ (^ 0^ o^ ••• -o v
H P^ ^H ,—j ^_(
0
3 " " " "
^ ^T 0 CO - -CWT



•J1 -C ^


X X M X M £ £ X ~
xxSx*X>x
2?£>c£c?

5
^j
**


_
c
c





c
c



1 ?.
e




c
, «f

c






o

c




c
rv.




~
C


I
0



H
3

Q
Cd
^
0

Q
Z
M
0
J
f£
cS
a

o

z
=
^
1
(S^
§
o
en
H

§
^
s
-d
as
8"
en

a
£
H
E->

6

o

CH
O
'~












..
5
U -H
0) !-l
•H Qj
Q^
W CO

£ ttl
fl O
I-J -H
i-i C
0 ^
^
en 0

0 -I-1
r-( 01
IU — (
e <•"
cu ^
'-4 CU

CO 0
u
^rS
•H
n: ^
Q ™
3
T3 0
CU v-l
-0 -C
0
O -U
CU 01
QS r^
O -H
U-l ij
CO <
cu ~^ tn
> — (
u < w
3 > >i
O -H

: CU C
C rH <
O -1
•H -^ U
JJ > (0
(13 Ul O
U l-l >•
3 CU
Q 4J ^
1 k-i 13
O O -H
ft. -P Vi
r ,-05-"

•
cn
CO
3
01
•H





(SJD)
      -193-

-------
Q

a
o
                co
                r-
 i
iO
                                                   ! e  H
J
<
>

_ Q
- Cd
5 £-«
• <
c 0
M
Q
2
t— *
I 8
, •
e <
CX
u

05
™" f~\
c °
*. z
s <

15

05
CJ
--,
C <
IT a
fi 05
C ,„
w
CO
w

3
? 3
' &.
C S
<
M
a
E-
CO
— w
* s
• I— |
c £-

CL4
o

2
— O
M
r ,
C O
<
S
Cb






S-l
0 CJ
LU ' 4J
Q W
W r-(
3 - m
O C
^ O S-l
tu 4J (U
E cr> J
03 C O
OJ -H 4J
U Si 0
•u en o
CO 03
>•* 01
>-l S-l 3
•H 03 O
Si) V.
c a
4J
T3 SJ
1) CD 4-1
T3 ^ t/1
i_i --j - -j
M "^ ^™ T
-O as oi
O -I -H
4) O -H 01
25 (13 Sj >,
§— , r—
< (0
0 J-> — c
0 <
c o< •
0 < ^
•^ - > 03
-W < 
i  -H
S-i CD W CU
0 -H k >
"4-1 Sj 0) ft
(Tt J >
fy *rf
ai S U
0) 03 . .
> -u O -^
>J 03 73
3 JJ 0
O S-l 03 -H
0) jj
• > W 0)
C "i 1) £.
0 05 >
•H -M 13
-W 03 05 cy
03 C iJ
>-i c w os
3 13 
-------

CD
to


•
a

Jl 9)


Ifi
99

01 «


(f. 8»
n

-i«

win

IN rfi«


at <•!
Q
- £d
S- §
= 0
M
a
M

— O
c S
^ 1
c<
• y

ce
c °
?. z
c <
S
s
I S
' tf^ U4
* 0
w
hH

2
s 3
•» t.
c S
ss
IX
c_,
cr*
cn
— » rvj
o S

e f-.

0. ?O|
HACTION OF
ft.


JJ










^
OJ
>
•M
a:
nj -^
iS
rtj -H
x: u
a a)
3 a.
a*
Cfl T3
3 0)
cn 4J
03
*-i O
0 -I
U^ T^
01 rH
3
0 u-i
r-l Q
14-J
£ •"
rtj CO
0) rH
^ f*1
4J
fO (j
>^
rH 0
•~t 4J
it! O
Q 0
T3 ^
0) C7i
TJ 3
U 0
0 U
0 J=
0) 4-i
a:
^j
M tn
0 i-l
u-t
rH
W -H
01 U
> 2.
•4 <
3 — -
O *"
-Duration"
rietta, PA
sar Analysi
1 M s»
3 03 >
0 S
— i 4-1
fc, JJ 0)
: ft] S
iX)
^j

-------
I
I
f
I
I
I
I
I
I
    Farther Evaluations of Selected Periods.  The aforementioned
designations of a typical year, dry year, and wet year were  based  upon the
period 1966-1978 since this is the period covered by  the NWS  hourly  rainfall
records acquired for the EPA Chesapeake Bay Program.  The decision to
restrict the rainfall records to the period 1966-1978 was one based
primarily on costs in an effort to keep the budget for the acquisition and
analysis of NWS rainfall tapes from becoming prohibitive.  Since NWS hourly-
rainfall records for the Chesapeake Bay Basin recording gages are  currently
available for 17 additional years (i.e., 1949-1965),  it was  necessary  to
screen observed April-October flow-duration curves for each  year in  the
period 1949-1978 to demonstrate that the selected years were  the most
appropriate for model production runs.

    For the typical year comparisons, the flow-duration curves for the
period 1966-1978 were found to adequately approximate the curves for the
full 30-year period (1949-1978)  covered -by NWS hourly rainfall records,
indicating that selections of a typical year based on comparisons  with the
shorter period were certainly appropriate.  Further,  1974 exhibited  as good
an agreement with the 1949-1978 period as did the only two years (1950 and
1961) in the earlier period which merited consideration as a  typical year.

    For the dry year comparisons, three earlier years (1954,  1957, and 1963)
merited consideration as a design dry year.  However, although the total
April-October streamflow volumes for at least one (1963)  of  these  earlier
years are less than April-October 1966 volumes, the similar  and even higher
frequencies of extreme low flows during 1966 reinforces its  selection  for
the dry year production runs.

    For the wet year comparisons, it was determined that none of the earlier
years produced April-October flow-duration curves which approached 1972 and
only one year (1960)  approached 1975.  Inspection of the flow-duration plots
indicated that 1975 is typically very similar to 1960 in terms of  the
frequency of high flows and superior to 1960 in terms of the  frequency of
low-to-moderate flows.  Thus, the comparison of flow-duration curves for the
period 1949-1978 supports the selection of 1975 for the wet year production
runs.

Interpretation of Model Output

    For each production run, the Basin Model can be operated  in its entirety
and loading output for the April-October period (plus a one-month  antecedent
period)  will be input to the Bay system models to project receiving water
quality impacts.  In addition, loadings delivered to the Bay  system can be
tabulated on a seasonal, monthly, and daily basis for general assessments of
management strategies.

    By subjecting the output time series to frequency analyses,
concentration-frequency relationships at the fall line can be derived  for
each management strategy simulation.   Concentration-frequency relationships
for typical year production runs may be assumed to represent  the average
                                                -196-

-------
impacts over a thirty-year period.  In other words, the number of days  (or
hours)  with water quality criteria violations projected for the  typical year
simulation may be assumed to represent the average number of violations per
year if the management strategy were applied over a thirty-year  period
characterized by a good mix of wet years, dry years, and average years.
Likewise, concentration-frequency relationships for dry year and wet year
production runs define the duration of water quality criteria violations for
two periods which can be expected to occur relatively infrequently over a
thirty-year period.  Since the selected wet year and dry year represent
extreme conditions, one would expect the frequency of any water  quality
criteria violations during these two simulation periods to be greater than
under typical year simulations which are assumed to represent average algal
growing season conditions in the receiving waters.

    By relating the simulated concentration-frequency relationships to
beneficial use criteria, water quality damages may be inferred from fall
line concentration simulations.  Reductions in water quality damages
achieved by a particular management strategy can then be expressed in terms
of reductions in the frequency of undesirable concentrations.  However, it
should be cautioned that since the models are idealized representations of
the prototype river systems, water quality output is best viewed in a
relative sense to provide insights into differences among management
strategies.

Framework for Model Production Runs

    The first set of production runs to be carried out with the  Basin Model
package were daily loading projections for the existing land use pattern and
the land use pattern projected for the Year 2000.  As previously indicated
the existing land use pattern is based upon LANDSAT data interpretations.
The 2000 land use projection was derived by increasing the existing urban
land use in each sub-basin by the ratio of the 2000 population projection to
the 1980 population and reducing forest land to represent the consumption of
undeveloped land by urban development.  Wastewater discharges and water
supply diversions for the Year 2000 were generally derived by multiplying
the 1980 values by the  ratio of population values.  The assumption that only
forest land will be consumed by urban development occurring between 1980 and
2000 is a conservative one designed to ensure a worst case projection of
Year 2000 water quality impacts.  This assumption preserves the  agricultural
land use distribution at 1980 levels while increasing urban development to
reflect population increases, thereby resulting in what is probably a
conservatively high estimate of nonpoint pollution loadings in the Year
2000.  The comparison of existing and future land use impacts on the Bay's
estuarine system should provide considerable insight into the impacts of
urban development on Baywide water quality.  These two model runs will  also
establish the baseline  receiving water quality conditions (i.e.,
concentration-frequency and loading-frequency relationships) for evaluations
of alternate management strategies.
                                    -197-

-------
I
I
f
I
I
I
I
I
I
V
I
I
I
I
1
    Included among the management strategy evaluations which were the
subject of model production runs are the following:

    A.    General Comparisons of Point Source and Nonpoint Source Loadings;
         The contributions of point and nonpoint sources to fall line
         loadings were determined by operating the Basin Model without any
         point source loadings (i.e., with only the flow, temperature, and
         dissolved oxygen associated with each wastewater discharge).  The
         difference between simulated fall line loadings "with" and
         "without" point source inputs represents the fraction of the fall
         line load that can be attributed to point sources.'  Since these
         production runs were intended to evaluate both point and nonpoint
         sources, they were based on all three hydrologic conditions (i.e.,
         typical year, dry year,  and wet year).

    B.    General Comparisons of Water Quality Benefits Promised by Cropland
         Best Management Practices (BMP's);   Initial production runs
         focussed on the relative loading contributions of cropland areas.
         The Basin Model was operated without surface runoff loadings from
         forest, pasture, and urban land uses,  and no change in subsurface
         flow loadings.   The difference between fall line loadings for this
         run and previous runs represented the contribution of non-cropland
         land uses.  In general,  the non-cropland contributions were not
         more than about one-third of the total fall line load of N and P
         for average and wet years, indicating the significance of cropland
         contributions.   The principal cropland BMP which was studied
         involved converting high tillage cropland areas to low tillage
         cropland.  The potential benefits of this cropland BMP were
         determined by comparing  Basin Model projections "with" and
         "without" the low tillage BMP.

    C.    Assessments of Locational Differences in Nonpoint Pollution Loading
         Delivecy to the Bay's Sstuarine System;  As previously indicated,
         it would be incorrect to assume that all nonpoint pollution
         loadings released into the Basin's  receiving waters are transported
         to the Bay's estuaries.   The Basin  Model was used to develop
         "pollutant delivery ratios" for different sections of the three
         major river basins.  By  removing the nonpoint pollution loadings
         produced by sub-basin clusters one  cluster at a time (i.e.,
         including only flow,  temperature,  and dissolved oxygen time series
         for clusters deleted from the simulations) ,  the Basin Model was
         used to estimate the fraction of sub-basin nonpoint pollution
         loadings which is delivered to the  Chesapeake Bay's estuarine
         system.  The difference  between simulated annual loadings at the
         mouth of each river basin for "with" and "without" sub-basin
         cluster conditions may be assumed to be the  total annual nonpoint
         pollution loadings delivered to the Bay's estuarine system by each
         sub-basin cluster.   Pollutant delivery ratio estimates for
         different sections of the Chesapeake Bay Basin can be used by
         management agencies to identify those areas  where nonpoint
                                                -198-

-------
     pollution controls promise the greatest benefit.  For example, maps
     showing pollutant delivery ratio contours could be developed for
     use in comparing the Chesapeake Bay impacts of nonpoint pollution
     loads from the various physiographic provinces in the Basin and to
     determine whether nonpoint pollution controls should be
     concentrated in downstream sections (e.g., Coastal Plain and
     Piedmont provinces)  of the Basin or extended into upland sections
     (e.g., Appalachian Ridge & Valley and Appalachian Plateau
     provinces)  as well.   Since these production runs are restricted to
     nonpoint source impacts, they were based on typical year and wet
     year meteorologic conditions only.

D.   Assessments of Locational Differences in Point Source Pollution
     Delivery to the Bay's Estuarine System;  Using the same approach
     outlined under C, loadings from point source clusters were removed
     from the Basin Model one cluster at a time to develop estimates of
     each cluster's contribution to the total fall line load.  Estimates
     of locational differences in the delivery of point source loadings
     to the fall line were developed for different sections of the three
     major river basins.   Point source contributions were also defined
     for each state traversed by the river basin.  These production runs
     were based upon all three hydrologic conditions (i.e., typical
     year, dry year, and wet year).

E.   Assessments of Point Source Management Strategies:  The impacts of •
     higher wastewater treatment levels on fall line loadings were
     assessed with the Basin Model.  The majority of the wastewater
     management strategies involved more stringent effluent levels for
     phosphorus.  At least one management strategy involved more
     stringent treatment levels for both nitrogen and phosphorus.  These
     production runs were typically restricted to dry year and average
     year conditions.
                                -199-

-------
I
I
f
I
I
I
I
I
I
b
I
I
I
1
i
                                  REFERENCES

1.   Hydrocomp,  Inc.,  "The Occoquan Basin Computer Model:   Calibration,
     Verification,  and User's Manual,"  prepared for Northern Virginia
     Planning District Commission, Falls Church,  VA,  May 1978.

2.   Johanson, R.C.,  Imhoff,  J.C., and  Davis,  H.H.,  "Users  Manual  for
     Hydrological Simulation  Program -  FORTRAN (HSPF),"  BPA-600/9-80-015,
     U.S.  Environmental Protection Agency,  Environmental Research
     Laboratory,  Athens,  GA,  April 1980.

3.   Linsley, R.K.  and Crawford,  N.H.,  "Continuous Simulation Models in
     Urban Hydrology," Geophysical Research Letters,  Vol. 1,  No. 1,
     May 1974, pp.  59-62.

4.   Linsley, R.K.  and Crawford,  N.H.,  "Computation of  a Synthetic
     Streamflow  Record on a Digital Computer,"  Journal  of International
     Association of Scientific Hydrology,  No.  51,  1960,  pp.  526-538.

5.   Donigian, A.S. and Crawford, N.H., "Modeling  Nonpoint  Pollution from
     the Land Surface," EPA-600/3-76-083,  U.S.  Environmental Protection
     Agency,  Environmental Research Laboratory, Athens,  GA,  July 1976.

6.   Hartigan, J.P., Quasebarth,  T.F.,  and Southerland,  E.,  "Use of
     Continuous  Simulation Model  Calibration Techniques  to  Develop Nonpoint
     Pollution Loading Factors,"  Proceedings of Stormwater  and Water Quality
     Management  Modeling Users Group Meeting;   March 25-26,  1982,
     EPA-600/9-82-015, U.S. Environmental  Protection Agency,  Environmental
     Research Laboratory, Athens, GA, 1982, pp. 101-127.

7.   Northern Virginia Planning District Commission and  Virginia Polytechnic
     Institute and  State University,  "Occoquan/Four Mile Run Nonpoint Source
     Correlation Study,"  Final Report prepared  for Metropolitan Washington
     Council  of  Governments,  Washington, D.C.,  July 1978.

8.   Northern Virginia Planning District Commission,  "Guidebook for
     Screening Urban Nonpoint Pollution Management Strategies," prepared for
     Metropolitan Washington  Council of Governments,  Washington, D.C.,
     November 1979.

9.   Hartigan, J.P., et al.,  "Calibration  of Urban Nonpoint  Pollution
     Loading  Models,"  Proceedings of ASCE  Hydraulics  Division Specialty
     Conference  on  Verification of Mathematical and Physical  Models in
     Hydraulic Engineering, American Society of Civil Engineers, New York,
     NY, August  1978,  pp. 363-372.

10.   Grizzard, T.J., Hartigan,  J.P.,  and Randall,  C.W.,  "The  Development of
     Water Quality  Management Plans Using  Data  from Automatic Stormwater
     Monitoring  Networks,"  Proceedings  of  IAWPR International Workshop on
     Water Quality  Monitoring,  held at  Munich, Germany,  June  22-25, 1981.
 f
                                                -200-

-------
11.   U.S.  Geological Survey,  "Water  Quality  of  the Three Major Tributaries
     to the Chesapeake Bay,  January  1979  - April  1981:  Estimated Loads  and
     Examinations of Selected Water  Quality  Constituents,"  prepared  for
     USEPA Chesapeake Bay Program, November  1981.

12.   Hartigan,  J.P., et al. ,  "Areawide and Local  Frameworks for Urban
     Nonpoint Pollution Management  in Northern  Virginia," Stormwater
     Management Alternatives, Tourbier, J.T.  and  Westmacott, R. , eds. ,
     University of Delaware,  DE,  April 1980,  pp.  211-245.

13.   Biggers, D.J., Hartigan, J.P.,  and Bonuccelli,  H.A., "Urban Best
     Management Practices (BMP's) :   Transition  from  Single-Purpose to
     Multipurpose Stormwater Management," Proceedings of Seventh
     International Symposium on Urban Storm  Runoff,  Report  No. UKYBU121,
     College of Engineering,  University of Kentucky, Lexington, Kentucky,
     July 1980, pp.. 249-274.

14.   Northern Virginia Planning District  Commission, "Follow-up Assessments
     of Alternate AWT Operating Rules with Occoquan  Basin Computer Model,"
     prepared for Camp, Dresser,  McKee, Inc., consultants to Virginia State
     Water Control Board, Richmond,  VA, February  1980.

15.   Northern Virginia Planning District  Commission, "Recommended Operating
     Procedure for Nitrogen Treatment Facilities  at  UOSA AWT Plant,"
     prepared for Upper Occoquan  Sewage Authority, Fairfax  County Water
     Authority, and Occoquan Watershed Monitoring Laboratory, May 1981.

16.   Northern Virginia Planning District  Commission, "Modeling Study of
     Nonpoint Pollution Loadings  from Potomac Embayment Watersheds," Final
     Report prepared for Virginia State Water Control Board, Richmond, VA,
     March 1981.
17.  CH2M~HiH' "Stormwater Management:   A Comprehensive  Study  of  the
     Muddy Branch and Seneca Creek Watersheds,"  prepared  for Montgomery
     County (MD)  Planning Board,  Silver  Spring,  MD,  April 1975.

18.  Montgomery County Planning Board,  "Conservation and  Management  Report
     on the Rock Creek Watershed, Montgomery County,"  Silver Spring, MD,
     1978.

19.  Metropolitan Washington Council of  Governments,  "The Seneca Creek
     Watershed Model:  Documentation of  the Hydrologic Calibration and
     Verification," Washington, D.C., April 1982.

20.  Northern Virginia Planning District Commission,  "Technical Plan for  a
     Modeling Study of Nonpoint Pollution Loadings and Transport  in  the
     Chesapeake Bay Basin," Interim Report prepared for USEPA Chesapeake  Bay
     Program, Annapolis, MD, February 1981.
                                   -201-

-------
I
I
f
I
I
I
I
I
I
 I
 I
 I
 I
 I
21.   Roffe,  K.A.,  "Computerized Mapping  for Assessment  of  Environmental
     Impacts," presented at Annual Conference of  American  Institute  of
     Planners, held at Kansas City,  MO,  October 8-12,  1977.

22.   Arbruster, J.T.,  "Flow Routing in the Susquehanna  River  Basin:   Part  I
     - Effects of  Raystown Lake on the Low-Flow Frequency  Characteristics  of
     the Juniata and Lower Susquehanna Rivers,  Pennsylvania,"  U.S.
     Geological Survey Water Resources Investigations  77-12,  USGS,
     Harrisburg, PA, April 1977.

23.   Wischmeier, W.H.  and Smith, D.D., "Predicting  Rainfall Erosion  Losses
     from Cropland East of the Rocky Mountains,"  Agricultural Handbook
     No. 282,  Soil Conservation Service,  U.S.  Department of Agriculture,
     Washington, D.C., May 1965.

24.   Northern Virginia Planning District Commission, "Assessments of Water
     Quality Management Strategies with  the Occoquan Basin Computer  Model,"
     prepared for  Occoquan Technical Review Committee,  January 1979.

25.   Hydrocomp, Inc.,  "Hydrocomp Simulation Programming:   Water Quality
     Simulation Operations Manual,"  Palo Alto,  Cal., 1977.

26.   Zison,  S.W.,  et al., "Rates,  Constants, and  Kinetics  Formulations in
     Surface Water Quality Modeling,"  EPA-600/3-78-105, Environmental
     Research Laboratory, U.S.  Environmental Protection Agency, Athens, GA,
     December 1978.

27.   Northern Virginia Planning District Commission, "Water Quality'Modeling
     Study of Goose Creek, Broad Run,  and Sugarland Run Watersheds,"  Falls
     Church, VA, June 1980.

28.   Northern Virginia Planning District Commission, "Occoquan Basin
     Computer Model:  Summary of Calibration Results,"  Falls  Church,  VA,
     January 1979.

29.   Lumb, A.M. and James, L.D., "Runoff Files  for  Flood Hydrograph
     Simulation,"  Journal of the Hydraulics Division, ASCE, Vol. 102,
     No. HY10, October 1976, pp. 1515-1531.

30.   Lumb, A.M., "UROS04:  Urban Flood Simulation Model, Part  1:
     Documentation and Users Manual,"  School of Civil Engineering, Georgia
     Institute of  Technology, Atlanta, GA,  March  1975.

31.   Ross, G.A., "The Stanford Watershed Model:  The Correlation of
     Parameter Values Selected by a Computerized  Procedure with Measureable
     Physical Characteristics of the Watershed,"  Research  Report No.  35,
     Water Resources Institute, University of Kentucky, Lexington, Ky., 1970.

32.   Hydrocomp, Inc.,  Hydrocomp Simulation Programming:  Hydrology
     Simulation Operations Manual,  Palo  Alto, CA, January  1976.
                                                -202-

-------
33.   Winter,  T.C.,  "Uncertainties in Estimating  the  Water  Balances  of
     Lakes,"  Water  Resources Bulletin,  Vol.  17,  No.  1,  February  1981,
     pp.  82-115.

34.   Cavacas, A.,  et al.,  "Hydrologic Modeling  for Studies of  Pollutant
     Loadings and Transport in Large River Basins,"  Proceedings  of
     Stormwater and Water  Quality Management Modeling Users Group Meeting:
     March 25-26,  1982,  EPA-600/9-82-015,  U.S. Environmental Protection
     Agency,  Environmental Research Laboratory,  Athens, GA, 1982, pp.  69-89.

35.   American Public Health Association, Standard Methods  for  the
     Examination of Water  and Wastewater,  14th Edition, APHA,  Washington,
     D.C., 1976.

36.   U.S.  Environmental  Protection Agency,  "Methods  for Chemical Analysis of
     Water and Wastes,"  EPA-625/6-74-003,  National Environmental Research
     Center,  Washington, D.C., 1974.

37.   Hollander, M.  and Wolfe, D.A., Nonparametric Statistical  Methods, John
     Wiley and Sons, New York, NY, 1973.

38.   Northern Virginia Planning District Commission, "Reverification of
     Occoquan Computer Model for Post-AWT  Conditions,"  prepared  for NVPDC
     Occoquan Basin Nonpoint Pollution  Management Program,  Annandale,  VA,
     November 1982.

39.   Smullen, J.T.  and Taft, J.L., "Nutrient and Sediment  Sources to the
     Water Column  of the Chesapeake Bay,"  USEPA  Chesapeake Bay Program
     Synthesis Paper, U.S. Environmental Protection  Agency, Annapolis, MD,
     February 1982.

40.   Northern Virginia Planning District Commission, "NURP Study Quarterly
     Report:   October -  December 1981," prepared for Metropolitan Washington
     Council  of Governments, Washington, D.C., February 1982.

41.   Wright,  R.M.  and McDonnell, A.J.,  "In-Stream Deoxygenation  Rate
     Prediction,"  Journal  of Environmental Engineering  Division, ASCE,
     Vol.  105, No.  EE2,  April 1979, pp. 323-335.

42.   Bansal,  M.K.,  "Nitrification in Natural Streams,"  Journal of Water
     Pollution Control Federation, Vol. 48,  No.  10,  October 1976,
     pp.  2380-2393.

43.   Roy F. Weston, Inc.,  "Susguehanna  River Water Quality Study Pertaining
     to the Safe Harbor  and Holtwood Hydroelectric Projects:   Interim
     Report," prepared for Penna. Power and Light Co. and  Safe Harbor  Water
     Power Corporation,  February 1982.

44.   Personal communication from Stuart Freudberg, Metropolitan  Washington
     Council  of Governments, June 22, 1982.
                                   -203-

-------
I
I
l'
I
I
I
I
45.   Huber,  W.,  "Discussion Remarks,"  Proceedings of Seminar on Design Storm
     Concept,  Ecole  Polytechnique  de Montreal, Montreal,  Quebec,  1979.


46.   Southerland,  E.,  "A  Continuous Simulation Modeling Approach  to Nonpoint
     Pollution Management," Dissertation presented  to Virginia Polytechnic
     Institute and State  University, Blacksburg, VA, in December  1981, in
     partial fulfillment  of the  requirements  for the degree Doctor of
     Philosophy  in Environmental Sciences and Engineering.
                                                -204-

-------

-------
I
I
f
I
I
I
I
I
I
          APPENDIX A
        SUBBASIN MAPS


            WITH


PRECIPITATION GAGE LOCATIONS

-------
                  Figure Al
186980
  O
                             073570


o
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
1
MAP
NOT TO SCALE
                        A-l

-------
Figure A2
  180470


0
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
)
MAP
NOT TO SCALE
   A-2

-------
 I
 I
                   Figure A3
f
I
I
I
I
I
I
                                   LEGEND
O  HOURLY RAINGAGE

    DAILY RAINGAGE
      SEGMENT NUMBER

 xxxxxx GAGE NUMBER



A-3
                                                                 MAP
                                                            NOT TO SCALE

-------
Figure A4


o
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
I
MAP
NOT TO SCALE
    A-4

-------
                         Figure AS
                                                                 079595
                                                                  o
               LEGEND
O  HOURLY  RAINGAGE
    DAILY RAINGAGE
  )( SEGMENT NUMBER
xxxxxx  GAGE NUMBER
     f
     MAP
NOT TO SCALE
                         A-5

-------
                          Figure A6
                     363699
                      Q.
364166
 O
12
361(778
  o
                                                                      361589
                                                                          O
      189030
       o


o

A
LEGEND
HOURLY RAINGAGE

DAILY RAINGAGE
X SEGMENT NUMBER

xxxxxx GBGE NUMBER
1

MAP
NOT TO SCALE
                             A-6

-------
          Figure A7
464763
  O


O
A
LEGEND
HOURLY -RAINGAGE
DAILY RAINGAGE
X SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
1
MAP
NOT TO SCALE
             A-7

-------
Figure AS


O
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
1
MAP
NOT TO SCALE
    A-8

-------
I
I
f
I
I
I
I
I
I
I
I
I
I
I
Figure 9
           441322
                            446533
                              A


o

LEGEND
HOURLY RA1NGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
\
MAP
NOT TO SCALE
                                             A-9

-------
             Figure 10
18^030
 O


O
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
)
MAP
NOT TO SCALE
                 A-10

-------
I
I
f
I
I
I
I
I
I
                               Figure 11
 I
 I
 I
 I
442044

O
                                         448046
                              465739
                               O
                                             445120
                                              O


O
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
1
MAP
NOT TO SCALE
                                              A-11

-------
Figure 12
                                                  O


O
A
.LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
I
MAP
NOT TO SCALE
    A-I;

-------
I
I
f
I
I
I
I
I
I
Figure 13
 I
 I
 I
 I
 I
  f


o
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
f
MAP
NOT TO SCALE
                                               A-13

-------
                     Figure 14
188065
 O
             3*3298
              O
                                  362721
                                    O
                           20
                 467730
                   O
                                           184030
                                            O
                                                        367965
                                                          O


0
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
1
MAP
NOT TO SCALE
                        A-14

-------
I
I
f
I
I
I
I
 I
 I
                    Figure 15
 I
 I
 1
  I
   I
362298
 O

O
A
l- — —
LEG
HOURLY RAINGAGE
DAILY RAINGAGE
END
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
MAP
NOT TO SCALE
                                              A-15

-------
                               Figure 16
366916


o
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
X SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
I
MAP
NOT TO SCALE
                                   A-16

-------
 I
 I
 f
 I
 I
 I
 I
I
I
Figure 17
                365817
                  o
                                   LEGEND
                        HOURLY RAINGAGE

                    A  DAILY RAINGAGE
   X SEGMENT NUMBER

 xxxxxx GAGE NUMBER
      t
     MAP
NOT TO SCALE
                                        A-17

-------
Figure 18
 A-18

-------
Figure 19
                       3622^5
                        O
      36^001
       O


O
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
»
1
1
MAP
NOT TO SCALE
    A-19

-------
Figure 20
                                 369408
                                  O


O
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
t
1
MAP
NOT TO SCALE
   A-20

-------
Figure 21
                                      368905


o
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
1
MAP
NOT TO SCALE
    A-21

-------
Figure 22


o
A
LEGEND
HOURLY RAINGAGE
DAILY RAINGAGE
)( SEGMENT NUMBER
xxxxxx GAGE NUMBER
1
I
MAP
NOT TO SCALE
    A-22

-------
1
                                       Figure 23
1
  m
                                   LEGEND
                         HOURLY RAINGAGE

                     A  DAILY RAINGAGE
     SEGMENT NUMBER

xxxxxx  GAGE NUMBER
     f
     MAP
NOT TO  SCALE
                                           A-23

-------
                                  APPENDIX B

                LANDSAT LAND USE DATA ACQUISITION AND ANALYSIS

                               LANDSAT OVERVIEW

    Landsat is a satellite data acquisition system dedicated  to earth

resource investigation.  The three Landsat satellites are  in  circular,  near

polar orbits at an altitude of 570 miles  (920 km).  The orbit brings an

individual satellite over the same point on the earth, at  the same time of

day, every 18 days.  The satellites pass over the equator  from north to

south at approximately 9:30 a.m.

    The primary sensor on Landsat is the Multispectral Scanner (MSS).   The

MSS collects data for each Landsat scene over an area 185  x 185 km (13,000

square miles).  The data is collected in the following spectral bands:

                   Band 4              0.5-0.6 micrometers

                   Band 5              0.6-0.7 micrometers

                   Band 6           '0.7-0.8 micrometers

                   Band 7              0.8-1.1 micrometers

which correspond to the colors blue-green, green-yellow, red, and

near-infrared, respectively.  The instantaneous field of view (INFOV)  or

resolution of the MSS is 79 meters x 56 meters.  The data  is  collected  in

each of the four bands in west to east strips called scan  lines in a Landsat

scene.  Each scan line is divided into picture elements (pixels)  that are 56

meters in length.  Hereafter, when referring to the location  of individual

pixels, a scan line will be called a line, and the elements in a scan line

will be called samples.  Therefore, each Landsat scene is  comprised of  over

7,500,000 pixels (3440 scan lines    x 3300 PIXELS   )
                 (     LANDSAT SCENE        SCAN LINE)
                                     B-l

-------
Each pixel has an area of 1.1 acres (79 meters x 56 meters); and there are 4




spectral reflectance values for each pixel.




    Each land cover type (soil, vegetation, water,  etc.)  has a distinct




spectral reflectance pattern called its spectral signature.  Using the




spectral signatures for each land cover, one can assign or "classify" each




pixel to a land cover type from the pixel's spectral reflectance.   For more




information on Landsat, consult "Landsat Data Users Handbook," U.S.




Geological Survey, 1200 South Eads Street, Arlington, VA or "Manual of




Remote Sensing" by the American Society of Phtogrammetry, 105 N. Virginia




Avenue, Falls Church, VA.
                                     B-2

-------
I
I
I
                                  OBJECTIVES

    The objectives of this project were to:

    o    Produce a Level I land cover classification from Landsat data of

         the Chesapeake Bay watershed.

    o    Within the agricultural land cover, determine tillage practices.

    o    Tabulate the land cover statistics by river subbasin.

                              GENERAL PROCEDURES

    The land cover analysis was performed on the Eastern Regional Remote

Sensing Application Center's (ERRSAC) Hewlett-Packard 3000 computer.  The

land cover dataset was developed using the Interactive Digital Image

Manipulation System (IDIMS) and Geographic Entry System (GES) software

packages.

    An IDIMS file containing a whole Landsat scene was created for each of

the twelve scenes covering the Chesapeake Bay drainage basin.  For the eight

Landsat scenes with dates prior to 1979, the known Landsat distortions were

removed using IDIMS automated preprocessing programs (keskewing and rotating

to true north).  The three 1979 scenes were ordered in a geometrically

correct format.

    Each Landsat scene was divided into subscenes with similar physiographic

and land cover characteristics.  Land cover signatures were developed for

each subscene from subsampled areas by an unsupervised classification

program.  The signatures from the unsupervised classification were used to

produce a land cover map of the subscene.   Figure Bl shows the subscene

location and Table Bl the names of the images and Land Cover Spectral

Signatures statfile for each of the Landsat scenes.
                                                 B-3

-------
    High altitude color infrared, high altitude black and white, and




orthophotquads of ground truth sites were used to verify the land cover




dataset.  Thematic line printer maps (line printer maps with grouped




spectral classes) were produced for each of the ground truth sites, and the




verification work was done at the Northern Virginia Planning District




Commission (NVPDC) office."  If the land cover maps were not"adequate, a




supervised approach was used to augment the land cover signatures.




    The digitized subbasin boundaries were registered and superimposed on




the final land cover map.   Pixel counts for each land cover type were




tabulated by subbasin.




    To establish the validity of the land cover statistics, the land cover




statistics were compared with data from:  NVPDC, Maryland State Planning,




Susquehanna River Basin Commission, Soil Conservation Service Rockingham




County, and Piedmont Planning District.  In addition, eight sites of




approximately four square miles each were randomly selected and their land




cover statistics compared with land cover statistics interpreted from




1:222,000 scale color infrared photography.
                                     B-4

-------
    Figure Bl

LANDSAT SUBSCENE LOCATION
CHESAPEAKE BAY
       BASIN
     B-5

-------
w
kJ.
M
fa
1
H
H
fa
CO
<
3
CJ
X
a
*
Q
Z
3
EH
05
0
CJ



X
J
Q
Z
<
3
EH
§
O


X
05
E-i
05
O
CJ
s
•-3



>i
fa
co
<;
3
u


05

Q
1
EH
05
O
CJ



X
05
Q
Z
«S
J
EH
05
O
CJ























































































































1 i
i
X
05


	 i



' S
, 1



Q
Z
5
EH
05
O
CJ



































o
ro

1
vO
M^
i — <

X
3
<
05
O
CO
s
>n




>H
fa
CO
H
fa
co
<
J
u
X

J
*
z
1
<
05
O
CO




3
•
z
0
EH
Z
g
CJ
CO






















































































































i

X X
3 ' J





i

! . i



z

z
s
a
co





H
fa
CO
<;
j
o
j
3
EH
05
O
cu
CO
<
H

3
M
3
u3
3

EH
05
O
a.
CO
2
<
i— t
J
M
3

05
3
J
J
H
3
S
1^3
>H
fa
' CO
<
3
u
05
3
EH
05
o
Cu
CO
<
M
, J
h4
M
3
05
3

EH
05
O
Oi
CO
s
<
H
iJ
J
H
2

J
J
J
a
M
2
S
1-3
M
>H
fa
CO
rt
J
O
j
kH
EH
05
O
AJ
CO
<:
H
J
J
f— I
3
J
J

EH
05
O
PJ
CO
s
<
I-H
J
J
M
3

05
J
J
i-J
I-H
3
3
t-3
>H
fa
CO
<
l-d
O
05
J
EH
05
0
a.
CO
hH
J
•4
M
2
K
J
>
EH
05
O
a,
CO
s
<
M
J
J
H
2








































j |

05 . 05
: 3 J
i i

i



p-3 J
! 3 1 iJ
1 '
' . i

CH
05
o
cu
CO

M
J
J
>— <
3






a\
r~
X
r-
in












-— j




















n

I
r^
r— i

a
3
05
05
H
fa
CO
<
J
CJ

J
3
•
U
05
3
ca
CO
I-H
05
05
<



J
3
o
05
3
m
CO
M
05
05
<
X

J
3
05
05
<:
X
s
1— J


>H
fa
CO
$
CJ

05

O
05
3
03
CO
1— 1
05
05
<
X


05
3
CJ)
05
3
CO
CO
M
05
05
<


J
J
05
05
H
fa
CO
rij
o

J
J
t
o
05
3
CQ
CO
M
K
05
H
fa
CO
<
I-H
u

05
J

CJ)
cs
o
ca
CO
M
05
O5
<



05
J
O
05
3
(23
CO
IH
05
O5
<
X



05 05
3 J


I

|
l _ .,
| J iJ
^ ^ i
t !



u
05
3
CD
CO
M
05
K
S
i-L,






CTi
r-
X
[^
in













-------
•
1 &
1 3
fa
rH
 < M CC
ft < z co a
L O c^
" i a a a
» 3 2 a 2
n> < a X
" 2 U M 2
»j CO Q< M
i.
13
It
b
^^
18 3
in §
1 i
13 S

JMALTO



Jx
fa
CO
5
CJ
J
D
LTOONA .




i-J
o
2:
|
<
OS
'
JMALTO



>H
fa
CO
3
CJ
os
D
LTOONA .




OS
£3
Z
0
g
J
e£

JMALTO



>H
fa
CO
3
CJ
J
J
LTOONA .




^
1-3
i
|
<
OS
JMALTO



^~*
t-l
CO
5
CJ
QS
>_3
LTOONA .




QS
J
2
O
<

























as os
O hJ




D i-I i




^
Cu
cn

»-j
CJ
a:
'-j
^H
ALISBUR
CO


X
1-3
SH
QS
D
CQ
CO
M
CO





































































X X
133 H^




I
'


as
£3
CQ
CO
M
H
s
EH
Z
o
OS
fa
QS
JMFRONI
fa
CO

! )
CJ

OS
t
<
^4
O
OS
0
OS
fa
QS

,
J
<
>H
0
QS
H
2
O
OS
fa

JMFRON]
fa
CO

! 1
CJ

^

<
^H
0
QS
B
O
QS
fa

^
.
<
^H
O
OS
E-i
Z
O
OS
fa
OS
J
JMFRON]
fa
CO

l_]
CJ

as
.
<
Jx
0
QS
2
O
OS
fa
OS

<
J
rt
>H
O
OS
E-<
§
OS
fa

j
t
os os
IO i i-4



— _ - 	 	
i
J ' iJ
3 J


<
g
OS

2
O
QS
fa





en
r-
X
in





















ro
ro
1
rH

-------

H
fa
1
ff
E-i
cn
Q fa
H 2
M H
fa CJ M
H cn s
cn i <
cn CQ 2
rt| O
h4 CO
O


TABLE Bl (contd
RAW DATA
SUB-SCENE
NAME
2
2 O
0 H
M EH
£-• 3
CJ 03
a M
cn os
1 6-1
CQ cn
•— ^ |_^
co a

y a cn
< 2 N 05

< 2 CO 6-
Q Cd
i a J s
3 z a
< a x 2
o5 CJ i i i i
'/5 Oj

Q
05
4,
H
o,

O
fa
05
O
2
S

LASFY
CJ

J
O
•
x;
J
o

o
2
NORFOLK. UL
O5
pi
fa
O5
0
2


LASFY
O
•
OS
o
*
^
(_]
0

0
2
NORFOLK. UR
, ^J
j
fa
05
O
2


LASFY2
O

X

(
**s
ij
o

o
2
NORFOLK . LH J








































05 X
3 J
.
1

- 	 	



3




y|
J
O
fa
05
0
2




P^
LO
X
r~
Lfl





























(^

1
U1
1—1


Z3
CJ
H
05
2.

CLASFY ,
1
(_^
3

Q
2
0
2
X
CJ
H
G5
"
RICHMOND. UL
PS
D
X
O
M
05


CLASFY
•
05

*
Q
2
O
&
U
H
05
RICHMOND. UR

^
X
CJ
H
QS
5^

CLASFY
•
f^
i— 3
.
Q
2
0

X
CJ
H
K
RICHMOND. LL
OS
1 ~\
CJ
M
as


CLASFYl
•
05

•
Q
2
O
2
CJ
H
OS
!
RICHMOND. LR





05 05
D J



—

1

3 Jj



Q
2
g

*T*
CJ
M




cn

X
f-v.
in



























'X

f>1

1
kO

B-8




























X

CJ
2
| -i
•^

.CLASFYl
w
J
.
O
OS

CQ
ffi

g

.
LYNCHBURG. LH

Q^
O
2
( 1
s

.CLASFY j
X
05

CJJ
OS
3
CQ
X

£H

j
LYNCHBURG. RH



























































on



	 - -



5


O
05
3
CQ
X
CJ
2





r~
LTJ
X
r-
in



























J
•
m

I
r~



a
1
05
2

LASFY |
CJ

X
os

w
^
0
2
O

ROANOKE.RH



























































X
OS
1











M
^
O
!2
<
0




CTi
p~
X
r*»
u">


















OS
•3-
m

I
as
i — i



-------
I
I
   1
I
I
I
 I
 I
 I
 I
  I*
  I
  I
  I
  I
                    CRITERIA FOR  SELECTING  LANDSAT  SCENES

    The criteria for Landsat scene selection were determined by the

requirements of the Chesapeake Bay fluvial model.  The land cover dataset

needed to be a USGS Level I, present use, land cover classification.  Within

the agricultural class, minimum and conventional tillage as well as pasture

land covers needed to be identified.  A study was undertaken to determine

the feasibility of developing two land cover datasets for the Chesapeake'Bay

Basin:  a more detailed land cover dataset developed from a multitemporal

analysis for areas with an estimated river transport time to the Bay of less

than 5 days; and a less detailed dataset developed from a single date for

all other areas.  The multitemporal was to be done at ERRSAC by NVPDC

personnel and the single data work by Kennedy Space Center (KSC).  After

meeting with Kennedy Space Center personnel, it was mutually agreed that KSC

did not have the resources to participate in the project.  Therefore, it was

planned that a 'single land cover database for the entire Chesapeake Bay

Basin would be developed at ERRSAC by NVPDC and that it would not use the

multitemporal approach.  After discussions with District Conservationists

and Agriculture Extension Agents from Fauguier, Fairfax, Loudoun, and Prince

William Counties in Virginia and Agricultural Engineers from the University

of Maryland, it was decided that only Landsat scenes from 1977 to the

present would adequately reflect present agricultural practices.   To

differentiate between minimum and conventional tillage practices, the

Landsat scenes' date had to be after spring planting but before the crop

cover reached 20%.  To minimize the amount of undefined land cover, only

Landsat scenes with a cloud cover of 10% or less were used.
                                                   B-9

-------
    Table B2 lists the Landsat scenes in possession of the Chesapeake Bay




Program.  "Used in study" denotes the scenes used in the development of the




land cover database.  "Quality" denotes the image quality on a scale of 1 to




8 for bands 4, 5, 6, and 7, respectively.




    There are six areas—three in Virginia/ one in Pennsylvania and two in




New York—where there is not total coverage of Chesapeake Bay Basin by the




Landsat scenes.  In Virginia's George Washington National Forest near Back




Creek Mountain, there is an area of approximately 250 square miles that is




not covered.  This area .is over 95% forested, and it is not covered because




the Front Royal and Roanoke Landsat scenes do "riot overlap.  There are two




small portions of the drainage basin- south of Norfolk, Virginia that are not




covered by the Norfolk Landsat scene.  They are about 15 square miles each.




One is west of the Dismal Swamp and south of Route 58; the other is east of




Dismal Swamp and south of Great Bridge, Virginia.  The missing area west of




Dismal Swamp is about 80% forest, and the area east of the Dismal Swamp is




predominately marsh.  On the West Branch Susquehanna River south of




Curwensville, Pa., there is an area of approximately 300 square miles which




is 85% forested that is not covered by the Altoona Landsat scene.  In New




York State there'are two areas which fall outside of the Williamsport




Landsat scene.  The first is an area of approximately 150 square miles that




is west of Kenka Lake and north of Avoca; the second is 25 square miles in




area and  is east of Kenka Lake and north of Weston.  Both areas are




predominately agriculture.  For each of the missing areas it was assumed




that the  ratio of land cover types within the missing areas was the same as




that for  the entire subbasin.
                                     B-10

-------
                                             TABLE B2
                                  EPA/GBP AVAILABLE LANDSAT CCT'S
I
TAPE
NUMBERS
21
22
23
24
25
26
27
28
29 '
JO
31
32
33
34
35
36
37
PATH ROW
15-33
15-34
16-30
16-31
16-32

16-33
16-34

17-31
17-32

17-33


17-34
18-34
SCENE I.D.
30422-15025
30422-15031
28321-44815
28141-44935
21192-14425
28681-44715
21192-14432
28501-44855
21192-14434
21193-14482
30082-15125
28151-45545
28151-45605
21229-14505
30082-15131
21589-15062
30101-15193
DATE
5-1-79
5-1-79
5-3-77
4-15-77
4-28-78
6-8-77
4-28-78
5-21-77
4-28-78
4-29-78
5-26-78
4-16-77
4-16-77
6-4-78
5-26-78
5-30-7-9
6-14-78
CLOUD COVER
10%
0%
10%
10% '
10%
10%
10%
10%
0%
10%
0%
0%
0%
10%
10%
10%
10%
QUALITY
8888
8888
8888
8888
5888
8888
8888
8888
8888
8858
8888
8888
5888
8888
5888
8888
8588
USED
IN STUDi
*
*
*
*
*

*
*

*


*


*
*
                                                B-ll
 a

-------
                       LAND COVER SIGNATURE DEVELOPMENT




    An unsupervised classifier was the first step in developing land cover




signatures for each subscene.  Only spectral data from every other line  and




sample pixel (within a subscene)  were used as input to the  unsupervised




classifier.  The unsupervised classifier, using an iterative clustering




algorithm, grouped the spectral data into 50 unique spectral classes.  The




statistics from the 50 classes (mean, variance, and covariance in the  4




spectral bands) were used as the land cover signatures and  stored in a




statfile.  The entire subscene (every line and sample) was  then classified




by a "maximum likelihood" classifier using the signatures generated by the




unsupervised classifier.




    Thematic line printer maps were output for the various  ground truth




sites within each subscene.  The line printer map in conjunction with  the




ground truth were used to assign each of the 50 spectral classes to one  of




the following land cover classes:  undefined (clouds and shadows), forest,




winter agriculture, conventional tillage, minimum tillage,  pasture, water,




marsh, urban, strip mine, and idle land.




    If all the land cover classes were not adequately represented by the 50




spectral classes developed by the unsupervised classifier,  a supervised




approach was used to develop the missing land cover signature.  The




supervised signature development was done at ERRSAC and the typical




procedure was:




    1.   The Landsat data corresponding  to the ground truth site was




         displayed on a color cathode ray tube (CRT).




    2.   Training sites (locatable areas on a Landsat scene with distinct




         spectral features)  for the missing land cover classes were selected
                                     B-12

-------
I
I
                      and their statistics calculated  and  appended  to the existing
                      statfile.
•               3.   The subscene was reclassified  using  the  updated statfile.
                 4.   The resulting classified  subscene  was  examined  on the  CRT  and
                      checked with ground truth.
                 5.   If the new classification was  not  adequate, steps 2-4  were repeated.
                 When a satisfactory land cover classification was  developed for all 36
•           subscenes, the data was overlaid with the geographic data and tabulated.
I
I
 I
 I
 I
 I
 I
 I
 I
 f
                                                  B-13

-------
                          GEOGRAPHIC ENTRY OVERVIEW

    In order to facilitate the registration of ground data to Landsat

imagery, the Geographic Entry System (GES)  was used.   GES is an interactive

system capable of entering,  editing, and storing geographic information

taken directly from a map surface.   GES has no analytical capabilities.  Its

only purpose is to allow geographic information to be put into a data

structure that is accessible to other programs such as IDIMS.

    The GES program uses three storable entities as it acquires geographical

information.  They are points, lines, and polygons.  Discrete points consist

of a coordinate and a point name.  Such points may be used as ground control

points for use in generating registration transformations between ground

data and digital imagery.  Discrete lines consist of a series of

coordinates.  Discrete lines are useful for digitizing roads, pipelines,

trails, and to a limited extent, contour lines.  Polygons consist of lines

that intersect at nodes called junction points.  Polygons may be used to

digitize any geographic boundary.

    The GES program can also construct a latitude/longitude or UTM base

grid.  The grid facility is very useful for cell by cell stratification of

data.

    All digitized information in GES is done with reference to a

"geoblock."  A geoblock is a rectangular subsection of the earth's surface

whose edges are oriented N-S and E-W.   The user defines the position and

dimensions of the geoblock, and then the GES program creates an internal

representation of the geoblock and superimposes a grid over it.  The grid's
inData Stratification and the Geographic Entry System:  A User's Manual,"
 ESL Inc. Technical Memo. ESL-TM 991, 1978.
                                    B-14

-------
I
I
JB         individual cells are called internal units.  As a result, ground  resolution


             varies with the geoblocks's size.  When the size of the geoblock  increases,


•           the ground resolution, as defined by the internal units, decreases.  GES
I




•

•




•




I
 I

 I

 I

 I

 I
             also recognizes three additional coordinate systems:  latitude/longitude,


             UTM, and digitizer coordinates.


                 The data digitized in GES is stored in six different Multiprogramming


             Executive (MPE) files.  They are:  the geoblock directory, the overlay, the


             polygon, the line, the junction point, and the discrete point files.  These


             six files exist for each geoblock, and the files are unique for that


             geoblock.  Table B3 lists the geoblock name, MPE files and the information


             stored in them for the Chesapeake Bay Program.
                                                  B-15

-------
                   TABLE  B3




              GES DATA  STRUCTURE
GEOBLOCK NAME:  WHITE








MPE FILES





ARCPBSBB





CPCPBSBB





JPCPBSBB





LSCPBSBB





OVCPBSBB





TXCPBSBB





WHITE
INFORMATION STORED





POLYGONS





DISCRETE POINTS





JUNCTION POINTS





LINES





OVERLAY
GEOBLOCK DIRECTORY
                         B-16

-------
I
I
•                                          GEOGRAPHIC DATA


                 All geographic data was digitized from nineteen 1:250,000 scale, USGS



•           topographic maps.   The  following  information was digitized from these base



             maps:   river subbasin boundaries,  state boundaries, populated area



•           boundaries, contiguous  boundaries  of the Landsat images, and the control



M           points used to transform the geographic data to the Landsat images.  Table



             B4  lists the GES  overlay number,  its name,  and the geographic data it



I           contains.   The following information on the GES overlays:  the overlay



             number,  the overlay name,  the GES  class, and the description of that class,



•           is  contained in a subsequent section entitled "Tape Catalog."



                 The river subbasins were divided into two overlays.   Overlay 1



 I         "  containing fifty-six subbasins covers the basins above the fall line, and



             overlay 2  containing twenty-nine  subfaasins  covers the basins below the fall



             lines.



 9               The first step in digitizing  the subbasin boundaries on these two



             overlays was locating the USGS gaging stations on the maps.  This was done



 9           using  information the USGS provided on the  gage's latitude and longitude



             coordinates and the description of the gage's location.   From this point,


             the boundaries of the subbasins were hand drawn on the 1:250,000 scale


 _          maps.   The subbasins were then digitized as polygons which are stored in GES



 •          overlays 1, "SUBBASIN," and 3, "COASTAL."



 fl              State  boundaries in the basin  for Delaware,  Maryland, New York,



             Pennsylvania,  Virginia  and West Virginia were digitized  from the base maps.



 §|          The boundaries were digitized as polygons on overlay 5,  "STATE."  This



             information was not needed for the hydrologic tabulations, but was digitized


 1
 H          for possible future use.
                                                 B-17

-------
    Populated areas below forty degrees north latitude which have  relatively




short travel times to the Bay, as highlighted on the base maps, plus




populated areas on the Harrisonburg base map were digitized.  The  USGS




criteria for defining populated areas on the USGS 1:250,000 scale  maps  is




that any area large enough on this cartographic product with detectable




housing or construction was a populated area.  The purpose for digitizing




these populated areas was to obtain data on the distribution of urban,




suburban, and rural areas in the subbasins.  The populated areas were




digitized as polygons on GES overlay 6, "URBAN."




    Due to the fact that adjacent Landsat  images overlap between 20 and  30




percent, there is a problem of double counting the areas within the




overlap.  Double counting the area within  the overlap can be prevented  by




merging the images together or dividing the images along a common  boundary




where the images overlap.  Rather than merging the twelve Landsat  images




together, a costly and time-consuming task which in this case required  too




much data storage, it was decided that the most efficient method for




correcting the problem of double counting  the area of overlap would be  by




using contiguous boundaries.  To do this,  an accurate approximation of  the




area covered by each image was needed.  Based on 9 x 9 inch hard copy prints




of the  images, the boundaries of each image were drawn on the base maps.




Noting  the overlap between the areas covered by the images, contiguous




boundaries for the images were then drawn  in the areas of overlap.  Also, at




the edge of the basin, the image boundaries were extended past the basin




boundary.  This insured the maximum coverage of the basin by the Landsat




images.  The outer boundaries drawn on the map contained the total basin.
                                      B-18

-------
I

f
I
I
I
I
I
I
fc
 I
 I
 I
 I
The contiguous boundaries and the outer boundaries were then digitized  as


polygons on GES Overlay 2, "Images."


    In order to generate a GES-to-Landsat transformation/ landmarks  that


were identifiable on both the IDIMS digital display and the GES base maps


were selected as control points.  The control points were marked and labeled


on the base maps and their line and sample values from the Landsat image


were noted and saved in a file.  A total of 570 points, more or less


uniformly distributed throughout the basin, were chosen.  The control points


on the base maps were then digitized on the point overlay, overlay 4,


"CONTROL."
                                                  B-19

-------
                              TABLE B4
                     GEOGRAPHIC ENTRY SYSTEM
                             OVERLAYS

OVERLAY HO.                 NAME            DESCRIPTION

    1                     Subbasin          Non-Tidal River Subbasins

    2                     Images            Contiguous Boundaries of the
                                              Landsat Images

    3                     Coastal           Tidal River Subbasins

    4                     Control           Control Points

    5                     State             State Boundary Lines

    6                     Urban             Populated Areas Highlighted
                                              on the USGS, 1:250,000
                                              Scale Maps
                               B-20

-------
I
I
I
I
I
I
I
 I
 I
 I
 I
 1
                        REGISTRATION  OF GEOGRAPHIC DATA


    The general procedure for registering the geographic data  to  the Landsat


scenes was as follows:


    1.  The Landsat scenes were displayed on a CRT and significant  landmarks


(ground control points) were located.  The line and sample coordinates of


the landmarks along with their identifying labels were stored  in  a


mensuration file (a unique file for each Landsat  scene).  The  location of


the landmarks along with their identifying labels were marked  on  the


1:250/000 scale USGS topographic maps.


    2.  The ground control points were digitized off the 1:250,000  scale


maps using GES software.  The identifying labels and their coordinates (in


internal units)  were saved in overlay 4.


    3.  Using IDIMS software, a third order polynomial equation was


developed (a unique equation for each Landsat scene)  which converts GES


internal unit coordinates to IDMS line and sample coordinates.  The


residuals from the transformation equations were all within ±3 pixels.


    4.  Using the transformation equation, the geographic data was


transferred into an IDIMS image format such that they had the  same  number of


lines and samples as their corresponding Landsat scenes.  For  each Landsat


scene, an image was created containing the Landsat scene boundaries (overlay


2).  For Landsat scenes where the data was digitized, an image was created


containing the urban boundaries (overlay 6).


[NOTE:  The Landsat data was not transformed to a ground coordinate system


(Latitude/Longitude, Universal Transverse Mercator, State Plane)  because of


increased computer time and storage limitations.  The GES files will be
                                                 B-21

-------
available for future users to do the transformations.  See TECHNICAL




MEMORANDUM "A Procedure for merging Landsat data with other geographic data




using the Geographic Entry System (G.E.S.)  and the Interactive Digital Image




Manipulation Systems (IDIMS)" by Wayne A. Hallada, Computer Science




Corporation, April 22, 1981)].
                                     B-22

-------
I
I
l'
I
 I
 I
 I
 I
 I
                         TABULATION OF LANDCOVER DATA


    The landcover data was tabulated by river subbasin for each of  the


thirty-six subscenes.  The general procedure for tabulating the land cover


data was:
I
                 1.   The classified subscene was overlayed by  the  image containing  the


•             ,       Landsat scene boundaries.  Only the portion of  the subscene  inside


                      the Landsat scene boundary was used in  the tabulation.


|               2.   The classified subscene was then overlayed by the image  containing


                      urban boundaries.  The subscene areas inside  the urban boundaries


•                    were changed to either high, medium, or low density  urban  depending


•                    on their spectral characteristics.


                 3.   The subscene was then overlayed by the  image containing  the  river


                      subbasin boundaries and the land cover  types tabulated by  subbasin.
                                                  B-23

-------
                           LANDSAT ACCURACY  ANALYSIS




    The verification of the Landsat land cover data set followed two




different procedures.  The first procedure was to randomly select areas




within the Chesapeake Bay drainage basin and compare the Landsat land cover




statistics with statistics from photointerpreting color infrared




photography.  The second was to compare land cover statistics from Landsat




with land cover statistics from various agencies in the Chesapeake Bay basin.




    The general procedure for comparing the Landsat land cover statistics




with the photointerpreted land cover statistics was:




    1.   The classified Landsat image was displayed on a CRT.  An IDIMS




         function was run which randomly selects blocks whose area was




         selected by the operator from within the Landsat scene.  The land




         cover statistics of the block were tabulated.




    2.   The positions of the blocks were located on the color infrared




         photography, a mylar overlay was placed on the photography and land




         cover types drawn and labeled.  The blocks were labeled for




         geographic features near their location.




    3.   The mylar was then transferred to  a digital planimeter and the area




         of each land cover calculated.  These were compared with the values




         tabulated from the classified Landsat data.




    The only requirement on the location of the randomly selected areas was




that it could be identified on the color infrared photographs (which was the




best ground truth available) on loan from Maryland State Planning.  There




was at least one randomly  selected site in  each of the Chesapeake Bay basin




physiographic provinces.  The Appalachian Plateau, Appalachian Ridge and
                                     B-24

-------
I
I
•           Valley, and the Blue Ridge each have one site.  The Piedmont  has  two  sites,


             and the Coastal Plain has three randomly selected areas.  The randomly


•           selected areas fell across three different Landsat scenes (Front  Royal,


             Washington, Salisbury).  Each of the sites was 50 x 50 pixels.  For the


g           Front Royal and Washington scenes, that corresponds to 4.3 square miles.


_           For the Salisbury scene the area is 3.1 square miles.


•               Because the randomly selected areas were located on three different


•           Landsat scenes, on different dates, and represent all 5 geographic


             provinces, it was decided that, when aggregated, their statistics would


•           adequately represent the accuracy of the land cover over the  entire basin.


                 There was some evidence to suggest that the randomly selected site near


I           Loch Raven Reservoir should not be included in the accuracy analysis.  The


             site,  10 miles north of the Baltimore beltway, was in a state of  transition


             between agriculture land use and residential land use.  There were


             •indications on the aerial photography that many of the agricultural fields


             had been left idle.  In one field a road network was in place for a


 fl           subdivision development.  If the fields had been left idle,  their spectral


             characteristics would not be representative of the land cover type and


 H           therefore should not be included in the analysis.   The photointerpreted land


             covers for Loch Raven Reservoir in winter agriculture and pasture were 2.2%


 9           and 41.44%, respectively.   There was no other site that had such  large


 M          discrepancies between these two land cover types.   There were no  indications


             that the classification errors were the result of physiographic influences.


  PI          Indeed, another site, Liberty Reservoir which is only 15 miles from Loch


             Raven Reservoir, has a land cover classification that correlated well with
  m
  t>q
  fei          the photointerpreted land cover.
                                                 B-25

-------
    Table B5 contains the Landsat land cover accuracy results.  Lines A-H




are the 8 randomly selected sites and their tabulation of land cover types.




Line I is the sum of all 8 sites and Line J is the sum of 7 sites with  line




E (Loch Raven Reservoir) excluded.  Line K lists the percent error  for  the




lines I and J.  Columns 1-5, 9, and 10 are the land cover types tabulated




from the randomly selected sites.  The value in column 1  (FOREST) includes




Idle land.  Columns 6-8 are different aggregations of the agriculture land




covers.  From inspection of line K, it can be seen that only columns 2,  5,




and 9 (winter agriculture, pasture, and urban)  are significantly effected by




the deletion of Loch Raven Reservoir from the accuracy analysis.  The change




in percent error in winter agriculture from 50.3% to 3.43% and in pasture




from -11.45% to +15.91% are reasonable.  After visual inspection of the




classified data and comparison of the classified data to other land cover




data sets, it was determined that the percent error tabulated without Loch




Raven Reservoir was more representative.  The percent error in the  urban




classification is probably not representative of the error in urban




classification basinwide.  Most of the error in the urban classification is




caused by Catoctin Mountain and Cumberland sites.  In both cases, the urban




areas were located in the shadows of mountains, which is not representative




of urban areas basinwide.  Visual inspection of the classified data and




comparison of the classified data with other land use data sets also




indicates that the urban area classification errors are less than the




accuracy analysis indicates.




    In addition to the  randomly selected sites, the Landsat land cover  data




set was compared with land use data sets from various state and federal




agencies.  The following river basins were used in this comparison:
                                     B-26

-------
I
I
•           Potomac, James, York, Rappahannock, Patuxent, Gunpowder, Occoquan,


             Appomattox, and Susquehanna.  The Landsat land cover data set was compared


•           with the Environmental Protection Agency (EPA)  Chesapeake Bay Program


             Historic Land Use data base.  The EPA data base was compiled from


•           Agricultural Census data (1978)  and Timber Survey data (1975).   Tables B6 -


•           B9 show the comparisons for the Potomac, James, York, and Rappahannock river


             basins.  For the Patuxent and Gunpowder river basins, the Landsat data set


•           was compared with data from the Maryland Department of State Planning's


             Maryland Automated Geographic Information (MAGI)  System.  The MAGI's 1978


B           land use data set was developed from high-altitude color infrared


             photography with a minimum mapping area of 10 acres.   Table BIO shows the


•           comparison of the Landsat land cover data with MAGI land use for the


g           Patuxent and Gunpowder river basins.  For the Occoquan River Basin,  the


^^r         Landsat land cover data was compared with the land cover statistics


•           planimetered from a 1979 1:48,000 scale land use map  developed by the


             Northern Virginia Planning District Commission.  Table Bll shows the


•           comparison for the Occoquan River Basin.  In the Appomattox River Basin,  the


             Landsat land cover statistics were compared with land cover statistics from


 B           the Pidemont Planning District Commission and Forest  Statistics for the


 «           South Piedmont of Virginia.  This comparison is found in Table B12.   Table


 ™           B13 shows a comparison of the percent forest developed by the Landsat data


             Bbase with percent forest figures developed by the Susquehanna River Basin


             Commission.
                                                 B-27

-------

0
0 g
rH «Q
3

z
<
a.
cn a
~


rj in
< •*•
*•/
CO J +
< cn
^ ^
gd
j
J
M
f« -.
T
p- j j.
< cn
E* +
p fN

B
2 en
* 5
W J ~
5 VO gd
> O M
j SH
2
<
U
LD XI 2
03 0 p
3 Ul P
« 1 s
i-i o 2
03 y
,-r* ^
<; **•
EI r-,
"* > -J
co ^ £ 3
a o M
!

s.i
fl Kt k-
S H

«
£o
CN 2 <
M
2t



CH
W
- §
o
u.



i ^* CN
1 CN .H
: ° d

' Sg
J d




*-< T
VO •"*
r- r-


rH 10
*r rn
TT ry




-H ^i
vo o
T ^"



O CO
IN IN
CO CD
CN CN


O CN
CO -<
•H o

in T
CJ» O
O rH
en en

10 O
vO O
m in
»H rH



en 10
cn en
en in
CN CN

Liberty % Photo
Reservoir % LANDSAT
Area=2782 Acres

i i
i i
i i

en vo
rH rH
r* o




IO Tf
VO IN
rH O
rH CN


O IN
vo in
^ p*




O CO
vo •«•
^T p*



ID CN
O P-
r- IN
rH


1 T
1 O
d

T ID
en in
•H en

10 CN
IN cn
en en



cn o
rH VO
rt cn
co p>

Catoctin % Photo
Mountain % LANDSAT
Area=2782 Acres

*T
1 O
1 <
1 O

O CN
en en
•H



CO CN
10 M
in en
•H CN


p* ^"
•H 10
^ •*?




CN
i in
i
i ^



.H CO
in T
^ CO
•H >H


p- IN
*H *H
-4 d

VO
1 p-
1
1 CN

IO
1 P-
1 •
1 »H



IN CN
O in
d in
p* P*

Cumberland % Photo
\ LANDSAT
Area=2782 Acres

CD O
CN 0
-H d
*H -H

P- T
•H CN
tH iH



CN TT
in 10
en vo


P- O
in o
O T




0
1 0
i .
1 en



in ^ .
a\ 10
IN IN


r- o
in IN
d d

0
1 CO
1 •
1 CN

0
1 O
1
1 •-*



en IN
O r*
^r IN
CO CO

Deep Creek % Photo
Lake % lANDSAT
Area=2782 Acres

i i
i i
i i

in *»
en co
cn o
t-i



CN *T
in Ti-
en d
in in


O IN
o in
CN c-H
FH m




en o
•
in TT

Swan Creek * Photo
% LANDSAT
Area=2782 Acres

cn CN
in rH
en en
^ *r

in vo
cn n
d d



Tf VO
*-» iH
in cn
m IN


VO VO
*H p»
cn vd
CN IN




CO O
.H O
VO TT
CN CN



CO O
cn TT
in IN


co vo
cn p*
IN CN

"4 IN
r-1 iH
ui en
CN IN

r- co
O CO
-i d



CO VO
en en
d >
IN CN

Tolcljester % Photo
Beech % LANDSAT
Area=2007 Acres
i
rH
CN 1
. 1
0 1

CO
1 0
1 t
r o



a* IN
P» e>
•^ rH
en cn


.H CN
p- CO
co eo




0 O
CO CN
-H IN
^ T



cn IN
vo in
p» en


o o
o o
cn vo
en ^

IN T
IN O
cn in
IN CN

CO vO
CO -H
CO P^
«"H rH


O CN
in eo

Lang ford % Photo
% LAUDSAT
Area=2007 Acres
i
0 T
-J CO
vo m

p* in
x cn
^r ^



in co
cn co
cn .4
en T


in IN
vO IN
in cn
IN IN




in CN
CO T
c en
CN CN



O VO
en vo
T IN
rH rH

0 0
eo co
T in

O CO
cn cn
in m
rH rH

in T
Cn T
TT p*



CO en
0 en
cn o
TT in

TOTAL % Photo !
% LANDSAT
Area=20, 706 Acres

m m
o r-
p- vO

0 p.
c m
T O



rH p.
cc in
r^ o
en 
+ +


IN .n
cn -H
en *r
rH
"*" *



p» en
cn vo
d d
rH
•t- +


in rH
*r cn
rH '^1
rH rH
1 *-

en IN
co in
o co
IN ^H
+ +
O cn
in IN
d d
+ I
en
en *r
d en
in

•f -t-
in o
in -H
CN CN
•f -t-

TOTAL* % Error
LANDSAT % Error**
Area=20,706 Acres














1>
01
Z)
•s
3
t:
• J) CD
01 JJ
=> C1J
u
•O 2^
IP
SM
w o
Q JJ
Z O
<£ s:
3 a
n n
-a -a
t) 01
> -w
t. u
HI cj
01 a
J X
O a

0)
W Ul
a •-*
i §
Vj
u
O wl
O 0)
rH CC,
X C
— D
TJ >
HI eg
JJ U
0 J<
CD O
a-o Q
(Observed-Ex
* %ERROR= Expecte
** Tabulated without L
 Q UJ l_t_ ten —• - — 	 , ^>
B-28

-------
I
I
f
I


I


I


I


I


I
LAND COVER


PASTURE


CROP


FOREST


OTHER
              TABLE B6


LAND COVER COMPARISON POTOMAC RIVER BASIN



      POTOMAC ABOVE FALL-LINE    POTOMAC BELOW FALL-LINE
LANDSAT


 18.42%


 16.35%


 61.00%


  4.23%
                       EPA/CPB*  LANDSAT


                        18.32%     15.90%


                        16.53%     15.23%


                        57.00%     54.47%


                         8.15%     14.40%
                                                                                   EPA/CPB*


                                                                                      7.64%


                                                                                     14.59%


                                                                                     51.70%


                                                                                     26.07%
          * Environmental Protection Agency Chesapeake Bay Program Historic Land Use Database
I


I


I

I
                                                B-29

-------
LAND COVER





PASTURE





CROP





FOREST





OTHER
                 TABLE 37





LAND COVER COMPARISON JAMES RIVER BASIN








         JAMES ABOVE FALL-LINE





         LANDSAT      EPA/CPB*





           14.62%      14.49%





           12.19%       7.16%





           65.23%      74.21%





            7.96%       4.14%
JAMES BELOW FALL-LINE





LANDSAT      EPA/CPB*





  10.81%        2.75%





  14.64%       12.78%





  62.68%       61.52%





  11.87%       22.95%
* Environmental Protection Agency Chesapeake Bay Program Historic Land Use Database
                                       B-30

-------
                                  TABLE  E3





                 LAND COVER COMPARISON YORK RIVER  BASIN








                            YORK ABOVE FALL-LINE





LAND COVER                  LANDSAT    EPA/CPB*





PASTURE                       12.88 %     9.11 %





CROP                          13.90%    11.00%





FOREST                        72.80%    69.70%





OTHER                          0.42%    10.19%
YORK BELOW FALL-LINE





LANDSAT    EPA/CPB*





  11.84%      3.25%





  19.26%     12.33%





  68.82%     71.59%





   0.05%     12.83%
* Environmental Protection Agency Chesapeake Bay Program  Historic  Land  Use Database
                                       B-31

-------
                                    TABLE B9





               LAND COVER COMPARISON  RAPPAHANNOCK RIVER BASIN
LAND COVER





PASTURE





CROP





FOREST





OTHER
                       RAPPAHANNOCK ABOVE  FALL-LINE
LANDSAT




  26.21%




  12.08%




  61.42%




   0.29% •
EPA/CPB*




 23.59%




 15.32%




 55.23%




  5.86%
                                 RAPPAHANNOCK BELOW FALL-LINE
LANDSAT




  15.1%




  21.64%




  62.20%




   1.06%
EPA/CPB"




  3.22%




 19.94%




 65.96%




 10.88%
* Environmental Protection Agency Chesapeake  Bay  Program Historic  Land Use Database
                                      B-32

-------
I
•
 _






 I



 I



 I
                                          TABLE BIO
                       LAND COVER COMPARISON PATUXENT AND GUNPOWDER RIVER BASINS
Forest



Agriculture



Urban
 PATUXENT



LANDSAT    MAGI*
  53%



  41%



   6%
                                                          48%



                                                          38%



                                                           9%
 GUNPOWDER



LANDSAT




   47%




   44%
                                                                                    MAGI*



                                                                                      41%




                                                                                      48%
I

_
•



I



I
                _

                *From Maryland Automated Geographic Information  (MAGI) , 1978 Dept. of
                 State Planning
                                                 B-33

-------
                                   TABLE Bll




                  LAND COVER COMPARISON OCCOQUAN RIVER BASIN







                                    LANDSAT                      NVPDC*




Forest                                 62%                         59%




Tilled Agriculture                     14%                         11%




Pasture                          '      20%                         22%




Urban                                   4 %                          8 %
*Planimetered from a 1:48,000 scale Land Use Map NVPDC 1979
                                       B-34

-------
I
I
l'
I
I
I
I
I
I
 I
 I
 I
 I
                                  TABLE B12



              LAND COVER COMPARISON APPOMATOX RIVER BASIN,  VIRGINIA




                                                PIEDMONT  PLANNING        FOREST

                          LANDSAT*              DISTRICT  COMMISSION**    PUBLICATION**_*


Forest                      69%                         65%                  72%


Pasture                     18%                         20%                  	


Agriculture              .   12%                         13%                  	
                From  Subbasin  in  Appomatix  River  Basin
                Average  of  Amelia and  Prince  Edward Counties
                Average  of  Amelia and  Prince  Edward Counties from "Forest Statistics
                for the  South  Piedmont of Virginia" 1976
                                                 B-35

-------
                               TABLE B13

           LAND COVER'COMPARISON SUSQUEHANNA RIVER BASIN
ZONE

 1

 2

 3

 4

 5

 6
LANDSAT
% FOREST
57
53
60
77
72
44
SRBC*
% FOREST
57
54
55
80
66
40
AREA SQUARE
MILES
4771
2596
3700
6900
3404
5288
 TOTAL
61
60
26600
ZONE

 1
 2
 3
 4
 5
 6
                 NAME

       Susquehanna River, Upstream From Athens, Pa.
       Chemung River
       Susquehann River, Sayre, Pa. to Sunbury, Pa.
       West Branch Susquehanna River, Source to Mouth
       Juniata River
       Susquehann, Sunbury, Pa. to Mouth  (excluding
        Juniata River)
 *  From "Assessment of the Water Quality of Streams in the Susquehanna River
    Basin," Susquehanna River Basin Comm., January 1976
                                  B-36

-------
I
I
l'
I
I
I
I
I
I
 I
 I
 I
 I
 I
  t
  I
                           LANDSAT IMAGE INFORMATION


                             CHESAPEAKE BAY BASIN


    The following sections contain information on the location of  the


Landsat imagery, the location of  the subscene on the imagery, the  land  cover


classification scheme, and the tape number where each Landsat image  is


stored.  Table B14 provides the information needed  to locate and access the


classified images for each Landsat scene.  It lists the Landsat image's


path/row location, the name of the file containing  the classified


information by the subscene name,  the location of the subscene on  the images


by the starting line, starting sample, number of lines and number  of


samples, the number of classes in each file, and the tape number where  the


information is stored.  Table B15 lists the identification numbers for  he


land cover types.  Tables B16 through B27, which use the land cover


identification numbers,  relate the land cover types to the subscene class


numbers.  These tables provide all the information  needed to access the  land


cover data for the Chesepeake Bay Basin.
                                     B-37

-------
TABLE B14
  LOCATION ON ORIGINAL IMAGE
PATH ROW
16-30

16-31

17-31







16-32



17-32



15-33

16-33



17-33



15-34


16-34



17-34

18-34

SUBSCENE NAME
CORTLAND . LH . CLASFY
CORTLAND . RH . CLASFY
SCRANTON.UH. CLASFY .
S CRANTON . LH . CLASFY
WILLIAMSPORT . UL .
1 CLASFY1
WILLIAMSPORT . UR.
CLASFY
WILLIAMSPORT . LL .
CLASFY1
WILLIAMSPORT . LR.
CLASFY
HARRISBURG - UL . CLASFY2
HARRISBURG . UR. CLASFY
HARRISBURG . LL . CLASFY1
HARRISBURG . LR. CLASFY
ALTOONA . UL . CLASFY
ALTOONA . UR . CLASFY
.ALTOONA . LL . CLASFY
ALTOONA . LR. CLASFY
SALISBURY . UH . CLASFY
SALISBURY . LH . CLASFY
WASH. UL. CLASFY
WASH. UR. CLASFY
WASH. LL. CLASFY
WASH. LR. CLASFY
FRONT . ROYAL . UL . CLASFY
FRONT . ROYAL . UR. CLASFY
FRONT . ROYAL . LL . CLASFYl
FRONT . ROYAL . LR . CLASFY
NO RFOLK . UL . CLAS F Y
NORFOLK. UR. CLASFY
NORFOLK . LH . CLAS FY 2
RICHMOND . UL . CLASFY
RICHMOND . UR. CLASFY
RICHMOND . LL . CLASFY
RI CHMOND . LR . CLAS F Y 1
LYNCHBURG . LH . CLASFYl
LYNCHBURG . RH . CLASFY
ROANOKE . RH . CLASFY

STARTING
LINE
1100
1100
1
1171
1

1

1171

1171

1
1
1171
1171
1
1
1171
1171
1
1492
1
1
1171
1171
1
1
1171
1171
1
1
2101
1
1
1171
1171
1
1
1
B-38
STARTING
SAMPLE
1
1501
1
1
1

1717

1

1717^'

1
1719
1
1719
1
1683
1
1683
441
441
1
1721
1
1721
1
1721
1
1721
1
1433
1
1
1723
1
1723
1
1799
1000

NUMBER
OF
LINES
1241
1241
1170
1170
1170

1170

1170

1170

1170
1170
1170
1170
1170
1170
1170
1170
1491
1492
1170
1170
1170
1170
1170
1170
1170
1170
2100
2100
883
1170
1170
1170
1170
2983
2983
1350

NUMBER
OF
SAMPLES
1500
1500
3343
3343
1716

1717

1716

1717

1718
1719
1718
1719
1682
1683
1682
1683
1936
1936
1720
1721
1720
1721
1720
1721
1720
1721
1432
1000
2432
1722
1722
1722
1722
1798
1798
2373

NUMBER
OF
CLASSES
48
50 '
50
50
49

50

50

50

51
51
52
50
50
50
50
52
55
50
50
50
46
50
50
50
51
50
50
50
51
48
45
50
46
50
50
50

TAPE
NO.
1
1
2
2
3

3

3

3

4
4
4
4
.
5
5
5
6
6
/
7
7
7
8
8
8
8
9
9
9
10
10
10
10 ,
^
11
11
12


-------
1
1
^^
f

1

1
1
1

1

1
•



NUMBER
1
2
3
4
5
6
7
8
9
10
11
I
I
I
I
I
                                             TABLE B15



                              IDENTIFICATION NUMBERS  FOR LAND COVER TYPES



                                                          LAND COVER TYPE


                                                          Unclassified


                                                          Forest


                                                          Winter  Agriculture


                                                          Conventional Tillage


                                                          Minimum Tillage


                                                          Pasture


                                                          Water


                                                          Marsh


                                                          Urban


                                                          Strip Mine


                                                          Idle
                                                 B-39

-------
                                                ?A7H-?.CW 16-30
              TABLE B16
LAND COVER TYPES 3Y SU3SCZNS CLASSES
— ^-s-.*-. r - -v

£.
3
4
5
o
™
3
?
10
— —
12
13
14
15
15
* -*
i ~
13
20
21
^ ^
23
24
25
25
27
23
29
30
31
32
33
34
J ^
35
37
33
3?
1 n
41
42
43
t-T
T C
45
LAND COVER
CORTLAND . LH . CLASFY
1
6
2
5
3
4
2
D
3
6
2
2
6
6
6
7
6
5
2
9
2
5
2
2
6
o
2
2
6
6
2
1
6
2
5
2
3
2
2
3
6
2
6
6
2
1
TYPE ^^
CORTLAND . RH . CLASFY
1
o
2
5
6
4
6
6
2
6
6
2
6
6
2
2
3
6
9
5
2
5
6 4ft
2 ^
D
6
2
2
2
2
2
7
. 6
6
2
5
2
4
2
2
o
6
2
2
6
5
4" 62
4o
49
2


.
5 j 5
Z —
52 B-40

-------
I
                                                 TABLE  B17




                                   ^ND COVER  TY?ES  BY  SUSSCZMZ CLASSES
                                                                                    ?ATH-=.CW 16-31
,
,
2
•3
4
5
0
7
3
Ci
10
11
	 J_
13
14
15
15
LAND COVER
SCRANTON . UH . CLASFY
A
2
3
2
5
2
^
2
3
2
2
2
6
2
2
7
TYPE
SCRANTON . LH . CLASFY
4
2
6
2
4
2
5
2
3
2
6
2
6
2
2
2
17 63
13
19
20
21
22
23
24
25
25
27
23
29
30
"« 1
32
33
34
3:
3c
37
33
39
.in
• T
_i T
43
44
4:
45
4~
4-=
49
-'w
2
6
2
5
2
D
2
3
2
2
2
5
9
5
7
4
2
6
2
5
2
5
2
3
2
2
2
6
7
o
2
D
2
6
6
2
5
2
5
2
3
2
2
2
5
9
2
7
4
2
3
2
5
5
2
3
2
9
2
6
2
2
3
2
o
2
; _
=2 B-41

-------
              TABLE B18
LAND COVER TYPES BY SU3SCENS CLASSES
CLASS NO.
1
^
3
•?
0
5
/
-*
3
~2
' Q
• 1
1 I
±2
14
15
15
— ;
15
^ ^
20
21
22
23
24
23
25
27
23
29
30
2 i
3 2
33
34
3-
•^ •-
JC
2 7
33
39
40
41
42
43
44
i £
•s —
46
47
43
i (-
-» ^
0 w
WILLIAiMSPORT.
UL.CLASFvi
4
1
2
1
3
1
2
1
5
•
2
1
5
1
2
6
1
7
1
6
1
2
1
5
1
2
1
2
1
7
1
4
1
2
1
6
1
2
1
5
1
2
1
5
1
2
1
0
1
-t

WILLIAiMSPORT.
UR.CLASFY
1
o
5
2
4
2
2
2
3
2
6
2
6
2
2
2
6
2
5
2
6
2
2
2
6
2
2
2
6
2
9
7
6
2
5
9
5 '
2
2
3
6
•?
6
2
2
6
5
5
->
6
WILLIAiMSPORT.
LL.CLASFY1
4
2
6
2
6
2
2
3
2
2
2
2
6
2
2
2
6
2
6
2
5
2
2
2
6
2
2
2
2
2
2
7
4
2
6
5
2
2
3
2
2
6
5
5
s
2
5
2
O
2
•
WILLIAiMSPORT .
LR.CLASFY
4
2
5
2
3
2
5
2
4
2
2
2
5
2
7
2
6
2
6
2
6
2
5 4|
2 ™
5
2
2
2
6
5
2
7
4
2
5
2
3
5
2
5 ,
5
9
2
6
5
5
2
2 ^
3 V
2
fl B-42
52

-------
I
                                                                                             16-32
                                                TABLE B19
                                   U\D COVER TYPES BY 3U3SCENE CLASS!
CLASS :ic.
<
^
3
4
3
C
7
3
o
10
' i
12
13
14
15
15
; "7
15
19
20
21
22
23
24
25
25
27
23
29
30
31
32
33
34
35
3 c
37
•« f^
j a
39
TU
41
42
"1 J
44
45
45
• —
4o
-» ^
50
~ •
52
HARRISBURG.
UL.CLASFY2
4
2
5
2
3
2
5
2
5
2
D
2
b
2
2
7
5
2
0
2
3
9
6
2
6
2
5
2
6
2
5
7
4
2
5
3
5
6
2
6
2
5
3
2
5
7
5
2
D
2
10

HARRISBURG .
UR . CLASFY
4
2
5
2
3
2
5
2
5
2
5
2
6
2
2
7
3
2
5
2
3
9
6
2
6
2
5
2
o
2
5
7
4
2
^
3
5
6
2
6
2
5
3
2
5
7
5
2
5
2
10

HARRISBURG.
LL.CLA3FY1
5
3
2
4
2
5
2
5
5
5
2
5
2
6
2
1
0
3
2
4
2
6
2
6
5
3
2
5
2
2
7
4
2
3
2
5
9
6
2
6
5
6
9
6
2
5
2
4
2
1
2
9
HARRISBURG.
LR. CLASFY
4
o
4
2
3
5
6
7
6
2
5
2
3
2
5
7
6
2
c;
9
3
2
2
7
6
2
5
8
6
2
2
7
4
2
4
2
3
5
5
7
6
2
5
7
6
2
5
7
5
9


53 B-43

-------
                                                = Y---?.C-v 17-32
         TABLE B20


iND COVER TYPES 3Y  Su3SCENE CLASSES
--^cc -;c_
7
2
3
4
0
5
^
3
~t
10
^. _
]_2
13
i i
15
15
i /
15
19
20
21
22
23
24
25
25
27
23
29
30
2 i
32
33
34
35
35
37
33
39
40
4 1
42
43
f 1
**-*
45
-* c
T .
43
49
— ^
ALTOONA .
UL.CLASFY
3
4
4
•5
6
2
2
2
5
2
2
2
6
2
2
9
6
4
6
5
2
2
2
2
2
2
2
2
2
2
2
7
3
5
6
7
9
2
2
9
6
2
2
2
2
2
7
3
5
7
ALTOONA .
UR.CLASFY
3
4
2
5
3
6
2
9
6
6
2
2
6
6
2
4
3
4
2
5
2
5
2
7
2
5
2
2
2
2
2
7
3
5
2
2
2
6
2
4
6
6
•j
2
6
6
7
-,
5
2
ALTOONA .
LL.CLASFY
3
4
5
4
3
6
5
2
f.
2
2
2
6
2
2
2
3
6
2
5
2
2
6
2
3
2
2
2
2
2
2
7
3
6
6
5
2
5
6
2
2
2
2
2
2
3
6
2
9
2
ALTOCNA . ^^
LR.CLASFY
3
4
6
2
6
6
6
2
V
4
2
2
5
5
6
9
3
5
2
2
6
5
2
2 A
3
5
2
2
2
5
2
7
3
4
6
2
5
6
2
5
6
5
2
2
2
5
2
2
2 J|
5 ™
51 9
32

B-44

3
z 3

-------
I
                                            TABLE B21



                                  LAND COVER TYPES 3Y SUBSCENE CLASSES
                                                                                             15-3
_, . _,. ,.n
L
2
3
SALISBURY.
UH.CLASFY
4
5
3
SALISBURY.
LH.CLASFY
4
9
3
4 77
5
5
7
3
9
10
— . J.
12
13
14
15
15
17
13
•19
20
21
22
23
24
25
25
27
23
29
30
2 i
32
33
34
35
35
J /
33
39
40
4 1
42
43
TT
45
45
-,"
43
49
3
2
2
7
5
2
2
7
5
7
2
7
d
9
3
7
5
2
5
7
5
2
2
7
5
2
2
7
4
2
3
7
6
2
2
5
2
2
5
7
2
. 4
2
2
6
6
2
2
7
5
2
2
7
4
2
2
"7
4
7
3
7
4
7
5
5
2
2
7
5
7
2
4
9
3
7
5
7
2
5
2
2
4
7
2
4
2
2
7
6
7
50 2 2
— -» 2
52
5

52 2 B-45

-------
                                                 ?Ar-:-?.cw  16-33
          TABLE B22

LAND COVER TY?ES  SY  SU3SCSNE CLASSES

1
-1
3
4
o
6
1
3
n
10
• T
12
13
14
15
15
: —
i ^
19
20
21
? 7
23
24
25
25
27
28
29
30
3 i_
32
• 33
J -I
3 5
35
37
33
39
40
f T
•^ —
42
43
« 1
Tt
-9 3
45
47
4c
49
-^
WASH . UL .
CLASFY
4
b
3
2
0
o
6
o
-
q
o
9
5
9
6
7
4
2
' 3
2
6
5
3
2
5
2
3
2
6
2
6
7
4
5
1
2
6
6
9
5
9
3
2
5
7
• 6
7
5
2
3
WASH . UR .
CLASFY
4
2
S
7
3
y
2
7
6
5
5
7
6
2
2
4
2
6
7
3
7
2
5
2
5
7
6
7
2
4
2
5
7
3
9
2
5
5
3
3
2
2
4
2
6
T
7
7
5
2
WASH . LL .
. CLASFY
3
2
4
2
4
2
2
7
6
2
2
7
s
2
2
6
2
4
7
6
5
2
6
2
5
7
5
2
2
3
2
5
2
6
2
2
7
3
2
2
7
fi
3
2
5
3




•
WASH . LR .
CLASFV
4
5
5
7
3
7
2
7
3
2
6
6
7
2
6
2
5
5
7
2
6
2
2 A
2
7
2
4
5
5
3
7
2
6
2
2
3
7
2
5
2
s
1
7
6
7
S
7
7 ^1
J ^
5
= 1
s:
5: B-46

-------
I
I
l'
I
I
I
I
I
I
            TABLE B23


LAND COVER TYPES  BY SUBSCZN'Z
                                                 PATH-ROW  17-33
CLASSES
-LA =3 .1C
i
2
3
t
-»
5
5
"7
3
£)
10
•; i
L2
13
14
15
15
;_7
13
^_a
20
21
1 "*
4* «
23
24
25
26
27
23
29
30
2 T
32
33
34
35
35
37
33
39
i n
t w
• 1
"t _
t -)
"t —
t 1
-Ij
• t
•9-1
"! Zi
4c
T /
-»O
45
c ^
FRONT. ROYAL.
UL.CLASFY
6
2
o
2
4
2
2
2
5
2
2
2
6
2
2
2
3
2
5
2
6
2
2
2
3
2
2
2
2
9
7
3
2
2
2
5
2
2
2
6
0
6
9
1
2
2
5
2
^
6
FRONT. ROYAL.
UR . CLASFY
6
2
4
2
6
2
6
2
6
5
5
2
3
2
3
7
3
2
4
2
6
?
9
2
. 3
5
5
6
2
2
2
7
3
2
4
2
6
2
6
2
6
5
5
2
6
2
5
->
3
4
FRONT. ROYAL.
LL.CLASFY2
3
2
4
2
4
2
D
2
6
2
6
2
6
2
2
2
6
2
5
2
6
2
5
2
6
2
2
2
6
2
5
2
3
2
5
2
5
2
2
3
6
3
2
3
5
2
6
2
3
2
FRONT. ROYAL.
LR. CLASFY
3
2
4
2
4
2
6
2
^i
2
o
2
6
2
6
7
3
2
5
2
6
2
2
2
3
2
5
2
o
2
2
7
3
2
3
6
5
b
2
o
-
6
o
3
2
-i
->
3
t
3
51 7
52 B-47

-------
                                           PATH-ROW  15-34
     TABLE 324

COVER TYPES 3Y SUBSCZNZ CLASSES
CLASS ::c.
-
7
3
4
5
c
~
3
n
NORFOLK .
UL.CLASFY
4
2
3
7
6
7
2
7
5
NORFOLK.
UR.CLASFY
4
7
5
7
3
7
2
7
5
NORFOLK . ^^
LH.CLASFY2
4
2
3
7
o
7
2
7
D
10 878
1 '
1 ^
13
T t
15
.L C
• •"
13
19
20
^ j.
22
25
24
25
26
27
23
29
30
31
32
33
34
35
3c
37
33
39
40
_il
42
43
-t-5
,! ^
45
4"
43
2
7
5
7
6
7
4
2
2
7
5
7
2
5
7
2
6
7
2
7
4
8
3
6
7
2
5
7
2
5
7
2
4
2
2
6
2
5
2
7
6
"7
8
4
7
2
3
2
5
7
2
2
8
4
5
3
2
b
7
2
2
7
5
7
7
2
8
5
7
2
2
8
4
5
3
2
2
7
5
7
6
7
4
2
2
7
6
8
2 A
5 ™
8
2
7
5
7
2
4
2
2
7
6
7
2
5
2
2
7
6
2
4
2
2
6
7 ^
49 7 5 o ^
^ 0
2
7
5
51 9
52 B-48

-------
         TABLE B25
P.ND COVER TYPES  BY  SU35CZNE CLASSES
                                                PAO:-?.CW 16-34
CLASS :;c.
T_
2
T
!
-S
C
5
7
3
3
10
•L —
12
13
14
15
16
17
_c
1 9
20
21
22
23
24
25
26
27
23
7Q
30
3 1
32
33
34
35
36
37
33
33
40
1 1
-i *.
42
43
44
45
4c
4"
4o
49
RICHMOND.
UL.CLASFY
4
6
3
2
6
3
9
5
6
3
2
3
2
2
7
4
2
3
9
6
2
5
2
5
2
2
2
6
2
2
7
4
2
3
2
6
5
2
2
5
2
2
5
2
7
5
2
3

RICHMOND.
UR.CLASFY
4
6
<•
2
5
5
2
4
2
3
7
3
2
2
4
6
2
7
6
5
2
7
5
2
2
7
5
2
2
4
6
2
5
6
2
4
3
7
5
2
2
4
2
2
7




RICHMOND .
LL.CLASFY
1
5
3
2
6
5
2
2
4
6
6
2
5
2
2
9
1
o
2
2
6
9
2
2
4
2
2
2
6
2
2
7
1
5
3
2
5
5
2
6
6
2
2
5
2
2
2
4
5
RICHMOND.
LR.CLASFY1
4
6
2
2
3
3
2
4
2
3
7
5
2
2
4
6
2
7
6
9
2
7
5
2
2
7
5
2
2
4
6
2
5
6
2
4
3
7
5
2
2
4
2
2
7
9



5: 2
31
;- B-ay
: j

-------
TABLE B26

CL--.53 *IC .

i
^
3
-I
5
0
T
3
3
10
• i
j. ^
13
14
15
la
^ /
15
19
20
21
22
23
24
25
25
27
23
29
30
J ^
32
33
34
35
35
37
33
39
4C
41
-1 T
43
-t-T
-» ^
45
-T I
4o
-» ^
5C
^ ^
c _
^At-iL) LUViK TYPES 3Y 5UBSCZME CLASSE
LYNCHBURG.
RH CLASFY

1
2
3
5
5
5
2
5
4
2
2
2
6
2
2
5
2
3
6
6
6
2
2
4
6
2
2
6
2
6
7
2
3
5
5
5
2
2
4
2
2
2
6
2
2
2
1
2
2
2

B-50
3
LYNCHBURG. ^^
LH.v_LASFY2
3
2
2
6
4
4
5
5
9
2
2
2
2
2
2
5
2
6
2
5
6
5
6 A
2 W
2
2
2
2
2
2
2
7
-5
2
2
6
6
5
6
2
2
2
2
2
2
2
2
2 -
6 4
2 ^



-------
                                           ?ATH-=CW 13-34
     TABLE B27
COVER TYPES BY SU3SCZME  CLASSES
CL.




















4































1.53 NO.
•
t.
3
I
T
5
c
1
3
a
uO
, ^
_2
i3
.4
.5
.5
. f
.5
.9
20
/I
Z2
23
24
25
25
27
2S
29
50
; 1
if — •
j2
)3
,4
! ^
55
7
;3
9
0
_
2
3
4
5
8
~
o
y
C
—
-
ROANOKE . RH . CLASFY
3
6
2
2
2
2
4
2
2
4
2
5
6
6
2
9
2
6
2
2
2
2
5
2
2
2
2
2
5
2
7
3
6
2
2
2
2
5
2
2
6
2
2
6
6
2
2
6
2
2

B-51

-------
                                 TAPE CATALOG




    All the Landsat land cover data and ancillary information is stored on




twenty 9-track tapes.  Tapes 1-12 contain the classified data, the Landsat




edge boundary, the subbasin boundaries, and where available, the urban




boundaries.  Tapes 13-18 contain the Landsat scenes preprocessed by IDIMS




software.  Tapes 19 and 20 contain the mensuration files, statfiles, and the




GES files.




    The classified data (Tapes 1-12)  is stored by subscene without the




spectral classes being aggregated into land cover types.  If the classified




data is to be overlayed with the geographic data'(river subbasins, urban




areas, Landsat edge), the following procedure should be followed:




    1.   Each of the classified scenes should be loaded off the tape.




    2.   The spectral classes should be aggregated into land cover types




         using Tables B15 - B27.




    3.   The subscenes should be reunited or mosaiced together.  They will




         now overlay the geographic, pixel for pixel.




    Tables B28 - B30 list the data contained on each tape.  Each file on the




tape contains all the data for a subscene.  Within each file are a given




number of  records.  Each record is one scan-line or line of data in a




subscene.  Each record is a given number of bytes in length, and each byte




is one pixel.  If there is an odd number of pixels in a line, an extra pixel




of intensity zero is added to make the number of pixels even.




    The GES files for the Chesapeake Bay program which are stored on Tape 19




are found  in Table B31.  The GES files can only be used with GES and IDIMS




software.
                                     B-52

-------
    Tape 20 contains the files for mensuration and transformation.  The




mensuration files for each Landsat image are listed  in Table B32.   (The




mensuration file for the Salisbury Image was lost because of tape errors.)




The transformation file contains the third degree polynomial transformation




equations for each Landsat scene and are prefixed by a six letter code




listed in Table B33.  The six letter code is used to identify and access  the




transformation file for each Landsat image.  The mensuration and




transformation files can only be used with GES and IDIMS software.




    Tape 20 also contains the land cover spectral signature files




(statfiles).  A list of the statfiles are found'in Table B34.  The  format of




the statfiles is as follows:




    1.   There is 1 record for each spectral signature of class in  the




         statfile.  See Table B14 for the number of classes in each statfile.




    2.   Each record contains 168 words, each word is 16 bits.  Figure B2




         represents the makeup of the record in a statfile.




    The Chesapeake Bay Basin Landsat data has been transferred to the




University of Maryland Remote Sensing Systems Library (RSSL)  for permanent




storage.  The RSSL will make copies of the tapes available at cost  to




interested parties.  Contact the Department of Civil Engineering, University




of Maryland, College Park, Maryland, for information.
                                    B-53

-------
TABLE B28
TAPE NUMBER/
DENSITY
1
(6250JBPI


2
(6250)BPI


3
(6250)BPI




4
(6250)BPI





5
(6250)BPI





6
(6250)BPI



7
(6250)BPI





SUBSCENE NAME
CORTLAND . LH . CLASFY
CORTLAND . LH . CLASFY
CORTLAND. EDGE
CORTLAND . SUBBASIN
SCRANTON . UH . CLASFY
SCRANTON . LH . CLASFY
SCRANTON. EDGE
SCRANTON . SUBBAS IN
WILLIAMSPORT . UL . CLASFY1
WILLIAMSPORT. UR. CLASFY
WILLIAMSPORT. LL . CLASFY1
WILLIAMSPORT . LR. CLASFY
WILLIAMSPORT . EDGE
WILLIAMSPORT . SUBBAS IN
HARRISBURG . UL . CLASFY2
HARRISBURG . UR. CLASFY
HARRISBURG . LL . CLASFY1
HARRISBURG . LR. CLASFY
HARRISBURG. EDGE
HARRISBURG . SUBBAS IN
HARRISBURG . URBAN
ALTOONA . UL . CLASFY
ALTOONA . UR . CLAS F Y
ALTOONA . LL . CLASFY
ALTOONA . LR . CLASFY
ALTOONA. EDGE
ALTOONA . SUBBAS IN
ALTOONA. URBAN
SALISBURY . UH . CLASFY
SALISBURY . LH . CLASFY
SALISBURY. EDGE
SALISBURY . SUBBASIN
SALISBURY . URBAN
WASH. UL. CLASFY
WASH. UR. CLASFY
WASH . LL . CLASFY
WASH. LR. CLASFY
WASH. EDGE
WASH. SUBBAS IN
WASH. URBAN
FILE
NO.
1
2
3
4
1
2
3
4
1
2
3
4
' 5
6
1
2
3
4
5
6
7
1
2
3
4
5
6
7
1
2
3
4
5
1
2
3
4
5
6
7
NO. OF
RECORDS
1241
1241
1241
1241
1170
1170
2340
2340
1170
1170
1170
1170
2340
2340
1170
1170
1170
1170
2340
2340
2340
1170
1170
1170
1170
2340
2340
2340
1491
1492
2983
2983
2983
1170
1170
1170
1170
2340
2340
2340
LENGTH OF
RECORD (BYTES)
1500
1500
3000
3000
3344
3344
3344
3344
1716
1718
1716
1718
3434
3434
1718
1720
1718
1720
3438
3438
3438 •
1682
1684
1682
1684
3366
3366
3366
1936
1936
2376
2376
2376
1720
1722
1720
1722
3442
3442
3442
BYTES PER
PIXEL
1
1
1
1
1
1
1
1 •
1.
1
1
1
1
1
1
1
1
1
1
i m
i •
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
i
      B-54

-------
                                           TABLE B29
rf
i
i
i
i
F TAPE NUMBER/
DENSITY SUBSCSNE NAME 	
8 FRONT . ROYAL . UL . CLASFY
(6250)BPI FRONT. ROYAL. UR. CLASFY
FRONT . ROYAL . LL . CLASFY1
FRONT . ROYAL . LR . CLASFY
FRONT . ROYAL . EDGE
FRONT . ROYAL . SUBBAS IN
FRONT . ROYAL ..URBAN
9 NORFOLK. UL. CLASFY
(6250) BPI NORFOLK. UR. CLASFY
NORFOLK . LH . CLASFY
NORFOLK. EDGE
NORFOLK. SUBBAS IN
NORFOLK. URBAN
10 RICHMOND . UL . CLASFY
(6250) BPI RI CHMOND. UR. CLASFY
RI CHMOND . LL . CLASFY
RICHMOND . LR. CLASFY2
RICHMOND. EDGE
RI CHMOND . S UBB AS IN
RI CHMOND . URBAN
| 11 LYNCHBURG. LH.CLASFY1
™ (6250) BPI LYHCHBURG . RH . CLASFY
LYNCHBURG. EDGE
LYNCHBURG . SUBBASIN
LYNCHBURG . URBAN
1 2 ROANOKE . RH . CLAS F Y
(6250) BPI ROANOKE. EDGE
ROANOKE . SUBBAS IN
ROANOKE . URBAN
13 ROANOKE (RAW DATA BAND 4)
(1600)BPI ROANOKE(RAW DATA BAND 5)
ROANOKE (RAW DATA BAND 6)
ROANOKE (RAW DATA BAND 7)
14 CORTLAND(RAW DATA BAND 4)
(1600) BPI CORTLAND(RAW DATA BAND 5)
CORTLAND(RAW DATA BAND 6)
CORTLAND(RAW DATA BAND 7)
FILE
NO.
1
2
3
4
5
6
7
1
2
3
4
5
6
1
2
3
4
5
6
7
1
2
3
4
5
1
2
3
4
1
2
3
4
1
2
3
4
NO. OF
RECORDS
1170
1170
1170
1170
2340
2340
2340
2100
2100
883
2983
2983
2983
1170
1170
1170
1170
2340
2340
2340
2983
2983
2983
2983
2983
1350
1350
1350
1350
2340
2340
2340
2340
2340
2340
2340
2340
LENGTH OF
RECORD (BYTES)
1720
1722
1720
1722
3442
3442
3442
1432
1000
2432
2432
2432
2432
1722
1722
1722
1722
3444
3444
3444
1798
1798
3596
3596
3596
2373
2373
2373
2373
3372
3372
3372
3372
3358
3358
3358
3358
BYTES PER
PIXEL
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
                                                  B-55

-------
TABLE B30
TAPE NUMBER/ FILE
DENSITY SUBSCENE NAME NO.
15
(6250)BPI










16
(625COBPI










17
(6250)BPI










18
(6250JBPI


19
(6250)
20
(6250)

FRONT. ROYAL (RAW DATA BAND 4)
FRONT. ROYAL (RAW DATA BAND 5)
FRONT. ROYAL (RAW DATA BAND 6)
FRONT. ROYAL (RAW DATA BAND 7)
HARRISBURG(RAW DATA BAND 4)
HARRISBURG(RAW DATA BAND 5)
HARRISBURG(RAW DATA BAND 6)
HARRISBURG(RAW DATA BAND 7)
CORTLAND.SB(RAW DATA BAND 4)
CORTLAND.SB(RAW DATA BAND 5)
CORTLAND.SB (RAW DATA BAND 6)
CORTLAND.SB (RAW DATA BAND 7)
RICHMOND (RAW DATA BAND 4)
RICHMOND (RAW DATA BAND 5)
RICHMOND (RAW DATA BAND 6)
RICHMOND (RAW DATA BAND 7)
NORFOLK. SB (RAW DATA BAND 4)
NORFOLK. SB (RAW DATA BAND 5)
NORFOLK. SB (RAW DATA BAND 6)
NORFOLK. SB (RAW DATA BAND 7)
ROANOKE.RH(RAW DATA BAND 4)
ROANOKE.RH(RAW DATA BAND 5)
ROANOKE.RH(RAW DATA BAND 6)
ROANOKE.RH(RAW DATA BAND 7)
SCRANTON (RAW DATA BAND 4)
SCRANTON (RAW DATA BAND 5)
SCRANTON (RAW DATA BAND 6)
SCRANTON (RAW DATA BAND 7)
WILLIAMSPORT (RAW DATA BAND4)
WILLIAMSPORT(RAW DATA BANDS)
WILLIAMSPORT (RAW DATA BAND6)
WILLIAMSPORT (RAW DATA BAND7)
WASH (FILE!) (RAW DATA BAND 4)
WASH (FILE!) (RAW DATA BAND 5)
WASH (FILED (RAW DATA BAND 6)
WASH (FILED (RAW DATA BAND 7)
ALTOONA (RAW DATA BAND 4)
ALTOONA (RAW DATA BAND 5)
ALTOONA (RAW DATA BAND 6)
ALTOONA (RAW DATA BAND 7)
GIS FILES

MENSURATION
ANO
STAT FILES
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4
5
6
7
8
9
10
11
12
1
2
3
4





NO. OF
RECORDS
2340
2340
2340
2340
2340
2340
2340
2340
1241
1241
1241
1241
2340
234.0
2340
2340
2983
2983
2983-
2983
1350
1350
1350
1350
2340
2340
2340
2340
2340
2340
2340
2340
2340
2340
2340
2340
2340
2340
2340
2340





LENGTH OF
RECORD (BYTES)
3442
3442
3442
3442
3438
3438
3438
'3438
3000
3000
3000
3000
3444
3444
3444
3444
2432
2432
2432
2432
2373
2373
2373
2373
3434
3434
3434
3434
3434
3434
3434
3434
3442
3442
3442
3442
3366
3366
3366
3366





BYTES PER
PIXEL
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1 ^
1 t
1 ^
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1



M
1
      B-56

-------
                   TABLE B-31
                 GES MPE FILES
FILE NO.
FILE NAME
                               DATA DESCRIPTION
  1           ARCPBSBB. WHITE. CIS
  4       CPCPBSBB. WHITE. CIS
  6       JPCPBSBB. WHITE. CIS
  9       LSCPBSBB. WHITE. CIS
 11       OVCPBSBB. WHITE. CIS
 13       TXCPBSBB. WHITE. CIS
 14       WHITE. WHITE. CIS
                               Polygons
                               Control Points
                               Junction points
                               Line segments
                               Overlay
                               Text
                               Geoblock
                      B-57

-------
    TABLE B-32



MENSURATION FILES
FILE NO.
3
4
5
6
7
8
9
10
11
13
14
FILE NAME
JMWMALTO . USERFILE . IDIMS
JMWMCORT . USERFILE . IDIMS
JMKSEROEL USERFILE . IDIMS
JMWMHARR . USERFILE . IDIMS
JMWMLYNC . USERFILE . IDIMS
JMWMNORF .USERFILE . IDIMS
JMWMRICH . USERFILE . IDIMS
JMWMROAN .USERFILE . IDIMS
JMWMSCRA . USERFILE . IDIMS
JMWMWASH . USERFILE . IDIMS
JMWMWILL. USERFILE. IDIMS
PATH/ROW
17-32
16-30
17-33
16-32
17-34
15-34
16-34
18-34
16-31
16-33
17-31
SCENE NAME
ALTOONA
CORTLAND
FRONT ROYAL
HARRISBURG
LYNCHBURG
NORFOLK
RICHMOND
ROANOKE
SCRANTON
WASHINGTON
WILLIAMSPORT
          B-58

-------
                                TABLE B-33
                           TRANSFORMATION FILES
FILE NO.
                      FILE NAME
   12
              JMWMTRNF.USERFILE.IDIMS
   CODE
 ALTOON
 NEWCOR
 FRONTR
 HARRIS
 LYNCHB
 NORFOL
 RICHMO
 NEWROA
 SALISB
 SCRANT "
WASHIN
WILLIA
_PATH-ROW
  17-32
  16-30
  17-33
  16-32
  17-34
  15-34
  16-34
  18-34
 15-33
 16-31
 16-33
 17-31
 SCENE NAME
 ALTOON
 CORTLAND
 FRONT ROYAL
 HARRISBURG
 LYNCHBURG
 NORFOLK
 RICHMOND
 ROANOKE
 SALISBURY
 SCRANTON
WASHINGTON
WILLIAMSPORT
                                    B-59

-------
FILE NO.
16
17
18
19
20
21
22
23
24
' 25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
TABLE B-34
STATS FILES
FILE NAME
JMALTOLL, STATS . ID IMS
JMALTOLR .STATS . ID IMS
JMALTOUL . STATS . IDIMS
JMALTOUR. STATS . IDIMS
JMCORTLH .STATS . IDIMS
JMCORTRH. STATS . IDIMS
JMFRONLL .STATS . IDIMS
JMFRONLR . STATS . IDIMS
JMFRONUL . STATS . IDIMS
JMFRONUR . STATS . IDIMS
JMHARRLL . STATS - IDIMS
JMHARRLR. STATS . IDIMS
JMHARRUL . STATS . IDIMS
JMHARRUR . STATS . IDIMS
JMLYNCLH . STATS . IDIMS
JMLYNCRH . S TATS . ID IMS
JMNORFLH . STATS . IDIMS
JMNORFUL. STATS. ID IMS
JMNORFUR. STATS . IDIMS
JMRICHLL. STATS . IDIMS
JMRICHLR . STATS . IDIMS
JMRICHUL. STATS . IDIMS
JMRICHUR .STATS . IDIMS
JMROANRH .STATS . IDIMS
JMSALILH . STATS . IDIMS
JMSALIUH . STATS . IDIMS
JMSCRAUH . STATS . IDIMS
JMSCRAUH . STATS . IDIMS
JMWASHLL. STATS . IDIMS
JMWASHLR. STATS . IDIMS
JMWASHUL . STATS . IDIMS
JM WAS HUR. STATS . IDIMS
JMWILLLL. STATS . IDIMS
JMWILLLR .STATS . IDIMS
JMWILLUL .STATS . IDIMS
JMWILLUR . STATS . IDIMS
PATH/ROW
17-32
17-32
17-32
17-32
16-30
16-30
17-33
17-33
17-33
17-33
16-32
16-32
16-32
16-32
17-34
17-34
15-34
15-34
15-34
16-34
16-34
16-34
16-34
18-34
15-33
15-33
16-31
16-31
16-33
16-33
16-33
16-33
17-31
17-31
17-31
17-31
SCENE NAME
ALTOONA. LL
ALTCONA.LR
ALTOONA. UL
ALTOONA . UR
CORTLAND . LH
CORTLAND.RH
FRONT . ROYAL . LL
FRONT. ROYAL. LR
FRONT. ROYAL. UL
FRONT.ROYAL.UR
HARRIS BURG. LL
HARRIS BURG. LR
HARRISBURG.UL
HARRISBURG.UR
LYNCHBURG . LH
LYNCHBURG . RH
NORFOLK. LH
NORFOLK. UL
NORFOLK. UR
RICHMOND . LL
RICHMOND. LR
RICHMOND. UL
RICHMOND. UR
ROANOKE.RH
SALISBURY. LH |
SALISBURY. UH
SCRANTON.LH
SCRANTON.UH
WASHINGTON.LL
WASHINGTON . LR
WASHING TON. UL
WASHINGTON. UR
WILLIAMSPORT . LL
WILLIAMSPORT . LR
WILLIAMSPORT . UL
WILLIAMSPORT .UR
B-60

-------
i
i
i
i
                                      o
                                      o
           to
           uo
           IT)
           LO
o
H
(XL
O
                  o
                  o
                  LU
                  CJ
 i
 i
 i
 t
           co
H


CQ

CN

II
                           O
           cr>
           co
                                     co
                                                 o
                                                 2:
                                                    cu
                       cu  a;
                       O  Q-r
                      <*- i— _a
                       3
                      CO
                      4-> T3
                      O O)
                                                       6uLuie.li  #
                                                       •AOO'^C
                                            spueg
                                                d OJ M
                                                 -a  c
                                                                 to  JZ  (O
                                                                 rO  ra  
                                                                                                     10

                                                                                                     CU
                                                                                                     U
                                                                                                     a
                                                                                                     to
                                                                                       (tJ
                                                                                       >
                                                                                       o
                                                                                                (/I

                                                                                                It}
                                                                                                01
                                                                                                                 •o
                                                                                                                 s_
                                                                                                                 o
                                                                                                         CO
 in     +     +
-o
 S-     VO   CM
 O          •—»

 S     "S    -a
 c     o     c
f-     3     10
                                                                                                      o
                                                                                                      o
                                                                                                      a>
                                                                                                                        C
                                                                                                                        cu
                                                                                                                  +

                                                                                                                  o
                                                                 •*->
                                                                 C
                                                                 
                                                                                              O
                                                                          B-61

-------
                           GEOGRAPHIC ENTRY  SYSTEM




    This section contains a list of the base maps and the GES overlay




information needed to access the digitized geographical information.  Table




B35 lists the names of the 1:250,000 scale, USGS maps used for the GES base




maps.   Tables B36 through B41 list the GES classes and the geographic




information contained in the classes for each overlay.
                                     B-62

-------
           TABLE B35
GEOGRAPHIC ENTRY SYSTEM BASE MAPS
       Scale:  1:250,000
MAP NAME
Wilmington
Washington
Baltimore
Cumberland
Salisbury
Newark
Binghamton
Elmira
Harrisburg
Pittsburgh
Scranton
Warren
Williamsport
Charlottesville
Eastville
Norfolk
. Richmond
Roanoke
Bluefield
Clarkesburg
STATE
Del.
D. C.
Md.
Md.
Md.
N. J.
N. Y.
N. Y.
Pa.
Pa.
Pa.
Pa.
Pa.
Va.
Va.
Va.
Va.
Va.
W. Va.
W. Va.
             B-63

-------
                            TABLE B36



                     GES OVERLAY NO. 1  -SUBBASIN
CLASS
88
89
87
81,
74,
76,
79,
60,
55
72,
61,
66,
32,
30
54,
13,
33,
58,
38,
35,
23
12,
1, 2
7, 8
11
10



83
75
85
80
78, 84

73, 77
82
67
51, 52, 53

56, 57
29, 31, 34
39
59
40
36, 37

22
, 3, 5, 6
, 9


SUBBASIN 10
20
30
40
50
60
70
80
'90
100
110
120, 130, 140, 150
160
170
180
190
200
210
220
230
240
250, 260
270
280
290
300, 310, 320
                                B-64

-------
                      TABLE B37



               GES OVERLAY NO.  2   IMAGE







CLASS	DESCRIPTION OF CLASS





  1                             ROANOKE LANDSAT IMAGE





  2                             LYNCHBURG LANDSAT IMAGE





  3                             RICHMOND LANDSAT IMAGE





  4                             NORFOLK LANDSAT IMAGE





  5                             FRONT ROYAL LANDSAT IMAGE





  6      ^                       WASHINGTON LANDSAT IMAGE





  7                             SALISBURY LANDSAT IMAGE





  8                             ALTOONA LANDSAT IMAGE





  9                       .      HARRISBURG LANDSAT IMAGE





 10                             WILLIAMSPORT LANDSAT IMAGE




 11                             SCRANTON LANDSAT IMAGE





 12                             CORTLAND LANDSAT IMAGE
                            B-65

-------
                           TABLE B38
                   GES OVERLAY  NO.  3 COASTAL

CLASS	DESCRIPTION OF CLASS
  69                           LAND SEGMENT CLASS  1

  42                           LAND SEGMENT CLASS  5

  65                           LAND SEGMENT CLASS  6

  19                           LAND SEGMENT CLASS  8

  17                           LAND SEGMENT CLASS  9

  68                           LAND SEGMENT CLASS  11

  47                           ANACOSTIA

  10                           APPOMATTOX

  64                           BALTIMORE HARBOR

  71                           BOHEMIA

  48    '   "                    CHESTER

  14                           CHICKAHOMINY

  44                           CHOPTANK

  25                           ELIZABETH

  20                           GREAT WICOMICO

  62                           GUNPOWDER

  15                           JAMES

  24                           NANSEMOND

  46                           NANTICOKE

  91                           OCCOQUAN

  63                           PATAPSCO

  41                           PATUXENT

  28                           POCOMOKE

  21                           POTOMAC

  18                           RAPPAHANNOCK
  90                           SEVERN

  49                           WICOMICO

  43                           WYE

  16                           YORK
                                  B-66

-------
I
                                           TABLE B39
                                    GES OVERLAY NO. 4 CONTROL
                 CLASS
        DESCRIPTION OF CLASS
                   1
                   2
                   3
                   4
                   5
                   6
                   7
                   8
                   9
                  10

                  11
                  12

                  13
                  14
                  15

                  16
                  17
                  18
BLUEFIELD MAP: PT 7, 8, 144, 145
ROANOKE MAP: PT 1-6, 10-49
RICHMOND MAP: PT 50-119, 131, 132
NORFOLK MAP: PT 120-130
EASTVILLE MAP: PT 133-143
CHARLOTTESVILLE MAP: PT150,_151, 326, 330, 332-351
WASHINGTON MAP: PT 152-169, 180-241, 250, 251
SALISBURY MAP: PT 278-281, 290-297
WILMINGTON MAP: PT 276, 277, 288, 289, 431, 440
BALTIMORE MAP: PT 242-249, 252-275, 283-287,  .
               299-303, 398, 400, 402, 405-408, 433-439
CUMBERLAND MAP: PT 304-325, 327, 329, 331, 395, 404
NEWARK MAP:  PT 419-421, 430
HARRISBURG MAP: PT 358-368, 378-388, 396, 397, 399,
                401, 413-415, 417, 418, 422-429
PITTSBURGH MAP: PT 354, 357, 370-374, 377, 389-394, 403
SCRANTON MAP: PT 491-496, 507-514, 528-531
WILLIAMSPORT MAP:  PT 410-412, 416, 482, 484, 489,
                  497-506, 515-527, 532, 53.3,  548, 549,
                  551,  555-561,  565, 566, 568-570
WARREN MAP:  PT 353,  355, 356, 552-554, 562-564, 567
BINGHAMTON MAP: PT 445-459, 465-473, 477, 478, 490
ELMIRA MAP:  PT 441-444, 460-464, 474, 476, 479-481,
            483,  485-488, 540-547, 550
                                                B-67

-------
                     TABLE B40
              GES OVERLAY NO. 5 STATE

CLASS                            DESCRIPTION  OF  CLASS
  1                              VIRGINIA
  2                              MARYLAND
  3                              DELAWARE
  4                              WEST VIRGINIA
  5                              ARLINGTON  COUNTY,  VIRGINIA
  6                              ALEXANDRIA CITY, VIRGINIA
  7                              FAIRFAX COUNTY,  VIRGINIA
  8   '                           PRINCE WILLIAM COUNTY,  VIRGINIA
  9                              LOUDOUN COUNTY,  VIRGINIA
 10                              PENNSYLVANIA
 11                              NEW YORK
                               B-68

-------
                                             TABLE B41

                                      GES OVERLAY NO. 6 URBAN



                    CLASS	DESCRIPTION OF CLASS



                      1                                  SALISBURY MAP

                                                         EASTVILLE MAP


                                                         NORFOLK MAP


                      2                                  RICHMOND MAP


                      3                                  ROANOKE MAP


                      5                                  BLUEFIELD MAP


                      6  "                                WASHINGTON MAP


                      8                                  CHARLOTTESVILLE MAP


                      9                                  WILMINGTON MAP


                                                         BALTIMORE MAP


                     10                                  CUMBERLAND MAP


                     11                                  HARRISBURG MAP
i
I
I
I
                                                  B-69

-------
                                 GROUND TRUTH


    This section contains lists of aerial photography (Table B42) and USGS


orthophoto quads (Table B43)  used as ground truth for the Landsat study of


the Chesapeake Bay basin.
I
I
I
                                     B-70

-------
H
it
i
t
t
 I
 I
 I
 I
 I






































CM
CQ

id
CO

S '











Z
\ — 1
CO
*3*
CQ

>•
CQ
g

U Cu
/3
Q3
~ 8
H ^
•^ o
23 ~
=H a,

Cl. ^
2 ' <
Z) M
o cs
Ct3 Cd
O <
Cd
^
<£
O*
&J
J
O
03
0
o 6
X 2
ft,
< Q
CO •
< H
Z
^

CJ
Cd
O
03 Cd
a, Q
O
* CJ
CO
CO


Cd
04
£4





EH
V
§
O
t4




J
f^
0
CO


w
£-<
 m r^ [^» r^ fr"] fT1 fr1 O O O O O
ininini£>voinu>>>>>>>



03Q3030303(Xa:Q3
x.x.x,\x.x,x.\
rj^n^iHi-irin
oooooooo
"o O OO 0 O O OX,x.x,x.x.x,x.\
CJCJCJCJOaCJCJCQCacQCQCQCQcaCQ

U U4. r)
(l)CUQJCl)(l)(DCi^Cl](l)ClJCl)OJCl)DrDCU
n, -lJ-l 3 S-lrH^tt)
0)0 OnODiJfdrH
•-HJ-I SiJDUXiCDiJrH
—1 CJ cnwOJfJCW'OXl -H
H i-i SE^ eUSO£tOt/)>
>fT3?fl) ^^tSUlS-i^rflO^^^iC
(A O 0 ~> -H -H -H O -r-l -H 5 jH C UH
£ _C3 (0 C tl rH rH i-J f^ 4_) rH r*H ^4 ^ Vl H™*
(^ -i-l OJfi30HH[T3(L)!l2 'H O O 0 fO 'H
^rHZI — ^SSCJSCdSSSEHES

in
CM
, — 1

1>














•g
"3
Q
^







X.
CQ

^

n.
frj
a,
cn
X


o
o
o
CO
p.*

rH
CM
r-l
o
r-i
vjQ
r"



CO
o
CO
3


it!
cu








S1
0
^Q
CO
•r-4
Jj
^J
f^J
32

xf
0*i
(•**_
in







xT
01
f-%.
LO
H
p^
LO
rH

(*!-)
f-x.
LO



a:


o
0
o

I]
M
Q)
p
fG
O4
0^
X


0
o
0
CO
CM
rH
rH

o

rH
ro
""-

5!
(fj*
2
t/}
,
ch
4J
Jj
Q
Q
CO

CN
CNJ
i—4
^r







CN
CN
r-l
xT
rH
(T>
o
CN

LO
r^
LO



03
X
M
0
0
o


Oj
d*
(T3
CM
cn
X


o
o
o
cn

r— 1
rH

o

o

p-

s

^
CO







^
c
•H
"c
(Tf
r— 1

i J
r^
3
Q
s

in
OO
rH
T







m
ro
rH
"^
rH
en
o
CM

m
[^
in



03
X.

o
Q
o"

LJ
M
OJ
p
(^
CM
cn
X


0
o
o
cn

•H
rH

o
in
o

^

y
^
<

f-C>
z


(0
*^»






n
rH
.2
"-W
J^
Q)
jj
cn
c^
r^
o

!_T) ^J*
CO Lf
1— 1
^r







LO
CO
r-l
<^p
rH
a^
o
CN
O^
*
m cc
r^ [jj
in |>



03
X.
^
0
0 X.
U CQ

t, (.
M >H
03 Q)
d, O4
crj fO
&4 CU
cn cn
X X


O 0
o o
0 0
r^ 0

rH
rH rH
com
0 rH

0 0

r- r-

s

<: co
CO O
*£ CO
Z D


n3 ro
?* w>








o

O i-l
d) CQ
J^
CU Cfl
M £
<^s o
3 E-1
.
co *vj* co CN O ^
i_n i — i cr* 0*1 LO co
rH


rH













O a. ffl w 03 03
ffi rJ 2J 2 2H 2
H O W W tJ W
^> ^ [> I> i> £>







\ \ \ X. X. \.
CD CQ CQ CQ CQ CQ

(, tl It t, (j (,

Pi Q_i Clj PH ^i PH
T3 (T3 fd  ^> ,> £> ^> ^>



OJ
cn

rH
•H JH
> 3
_Q C
o; wo
rH ^3 E J-*
ft3 *H fH C7^
p^-l CO O "H i
GJ QJ i-4 C rH -rH
C JyJ Q eg --H .>
•H n3 > o 'r~' ^
^ J M E-t 2 M

r-
i—

















a.
J
Q
^







\
CQ

(,

Q-i
ft3
a.
cn
X


o
o
o
o
CO

r-H

CM
CO
0

r-



CO
O
CO
D


ftj
£>








CU
c
Q
4J
C/}
^
nj
i-H
U

0"^
CN



rH













Oi
rJ
Q
£>







-x^
CQ

I.

Qw
tO
&4
CTl
X
cn

o
o
o
0


rH

CN
PO
O

r-



co
o
CO
D


n3
[>








rH
rH
n3
U
QJ
j^
Q
4-J
CO

CO
< — 1

















•^fi
^
o
£>







x^
CQ
M
d)

CU

CO
rH
CO


o
o
o
CO
[-^

rH

CN
ro
O

^



CO

CO
D


03
^>









C
n3
C
(^
^c
(J
-J
CQ

cc
CN

















•K
O
Q
^







X.
CQ
!H
0)

Qj

CO
i- H
X
00


o
o
o
CO


rH

CM
C*")
O

^



CO

CO
ID


T3
£>






>i
rH
rH
>

rrj
rH
Q
C
M
5

cc
O*1

















J.
-•j
Q
^







\
CQ
^
0)
n
WH
n3
p_t

00
rH
CO


o
o
o
CO
ps.

rH

CN
OO
O

r~



CO
CU
CO
o


n3
£>








0)
rH
( — j
•H
^
UJ
,_^
,-rj
Q

Q
O
rH
















X
^
Q
^







x^
CQ
!H
CU

Q,

CO
rH
CO


o
o
o
co


rH

CN
on
O

^



CO

CO
.r>


03
^>










cu
•y*
0
c
fQ
Q
cc

CO
CN
f — 1
















•ic
itD
Q
^







\_
CQ
rH
a
n
*-!-(
Qj

CO
r- H
CO


o
o
o
CO


rH

CM
OO
0

"^



CO

co
o


tQ
[>








(T.
0

CTi
£
>p^
5^
ty
t-i

1 1
1 1
1 |

1 1
1 1













I |
! |
| |
1 1



O3 ££
X. X,

0 0
rH i— I
O 0
0 CJ

•
CO CO
z z

03 Q3
H EH
en cn
X X
cn cn

o o
o o
0 0
CM CN
CM CM
rH r— |
rH rH
CO CN
CM rH

o o

r-- r-

2 ^
<1 f£
S! -<3
CO CO
^ ^
2 Z


'O 13
2 ^*





1 (0
W -U

2 EH
                                           B-71

-------
                                    TABLE B43

                       GROUND TRUTH:  CHESAPEAKE BAY BASIN
                                 ORTHOPHOTO QUADS
                                 SCALE = 1:24,000
MAP NAME
                          STATE
MAP NAME
                                                                              STATE
Barton
Binghamton West
Cortland
Endicott
Guilford
Holmesville
New Berlin North
New Berlin South
Norwich
Oxford
Pitcher
• Seeley Creek
Sherburne
Spafford
Waver ly
Altoona
Beech Creek
Berwick
Black Moshannon
Crooked Creek
Harveys Lake
Howard
Julian
Kingston
Knoxville
Lopez
Mifflinetown
Millersburg
Phillipsburg
Ransom
Schellsburg
State College
Tamaqua
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
NY.
Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
• Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
Pa.
Beach
Bentonville
Bon Air
Bowers Hill
Charlottesville, East
Charlottesville, West
Harrisonburg
Hopewell
Lake Drumond -
Luray
McKenney
Midlothian
New Kent
Newport News North
Petersburg
Prince George
Strasburg
Walkers
Warfield
Waynesboro East
Waynesboro West












Va.
Va.
Va.
Va.
Va.
Va.
Va.
Va.
Va.
•Va.
Va.
Va.
Va.
Va.
Va.
Va.
Va.
Va.
Va.
Va.
Va.












                                        B-72

-------
*r.
               ••>?-\'»;f
                -., t _,;.ijVA

-------
1       •   i-'sft* ''i
                                                              m   3 o
                                                              <-   o =•
                                                              O   Q. ?f

                                                              W   ai CD
                                                              cn  ~o ID
                                                              -^   o ^<

-------