Chesapeake Bay Program
Watershed Model Application to
Calculate Bay Nutrient Loadings
Final Findings and Recommendations
Prepared by the Modeling Subcommittee
of the Chesapeake Bay Program
July 1994
CBP/TRS 157/96
EPA 903-R-94-042
-------
EPA
July 1994
CHESAPEAKE BAY PROGRAM
WATERSHED MODEL APPLICATION TO
CALCULATE BAY NUTRIENT LOADINGS
Final Findings and Recommendations
Anthony S. Donigian, Jr.
Brian R. Bicknell
Avinash S. Patwardhan
AQUA TERRA Consultants
Mountain View, California 94043-1004
Lewis C. Linker
Chesapeake Bay Program Office
Annapolis, Maryland 21403
Chao-Hsi Chang
Robert'Reynolds
Computer Sciences Corporation
Annapolis, Maryland 21403 >
Project Officer
Mr. Robert Carsel
Assessment Branch
Environmental Research Laboratory
Athens, Georgia 30613
CHESAPEAKE BAY PROGRAM OFFICE
U.S. ENVIRONMENTAL PROTECTION AGENCY
REGION III
ANNAPOLIS,, MD 21403
-------
FOREWORD
The EPA Chesapeake Bay Program (CBP) has sponsored a series of projects to
develop and implement a 'Watershed Model' of the Chesapeake Bay drainage area
that could be used effectively to estimate nutrient loadings to the Bay and to
evaluate the impacts of agricultural Best Management Practices (BMPs). A phased
approach was planned to develop the Watershed Model and implement its calibration
and refinement. The U.S. EPA Hydrological Simulation Program - FORTRAN (HSPF) was
selected as the framework for the CBP Watershed Model.
Phase I of the effort was designed to improve the nonpolnt loading
representation, refine the data input to the Model, and perform a preliminary
calibration to available water quality data for the 1984-85 period.. This period
was used as the basis for the 40% nutrient loading reduction goal defined by the
1987 Chesapeake Bay Agreement.
Phase II was designed to focus on a better representation of the effects of
agricultural BMPs, including application of the detailed Agrichemical Modules
(AGCHEM) of HSPF to allow a more deterministic, process-oriented approach to BMP
analysis and evaluation. AGCHEM provides a nutrient balance approach to modeling
cropland areas so that sensitivity to nutrient inputs can be represented. In
addition, HSPF code enhancements were performed to allow specific consideration
of sediment-nutrient and bed interactions within the stream channel. The Phase
II Watershed Model was then applied for the 1984-87 period to the Bay drainage
area to include sediment erosion, -sediment transport, and associated nutrients,
in addition to the current modeled water quality constituents, to provide input
to the Chesapeake Bay 3-D Water Quality Model.
This report presents the results of the combined Phase I/Phase II effort. Model
calibration results are presented for hydrology, nonpoint loadings, -and instream
water quality for numerous sites throughout the Bay drainage. In addition, model
estimates of Fall Line loads for various nutrient forms are compared with
regression estimates used in the calibration of the Bay Model: The results
indicate that the Watershed Model is a valid representation of nutrient loadings
to the Bay, and provides a framework for assessing the impacts of land use
changes and alternative nutrient and agricultural management practices.
Since the completion of Phase II, the Watershed Model has been used to: 1)
determine the distribution of point and nonpoint source loads and the
controllable and uncontrollable portions of the loads; 2)determine the quantity
of loads reduced under different management actions; 3) determine the nutrient
loads to the Bay under different Clean Air Act scenarios; and 4) quantify the
loads under future (year 2000 conditions). These loads were used as input
conditions for the Chesapeake Bay Water Quality Model. Following this Foreword
is a brief technical summary of the status and use of the Watershed Model for
management and planning in the Chesapeake Bay Region. Work is continuing in a
current Phase III effort to extend the simulation period, investigate and improve
the plant nutrient and reservoir simulations, and refine ..the calibration for
recent water quality conditions.
-------
THE CHESAPEAKE BAY WATERSHED MODEL
L.C. Linker*, G.E. Stigall*. C.H. Chang**, and A.S. Donigian***
* U.S. Environmental Protection Agency, Chesapeake Bay Program
, 410 Severn Ave. Annapolis, MD 21403
" Computer Sciences Corporation
410 Severn Ave. Annapolis, MD 21403
" Aqua Terra Consultants
2672 Bayshore Pkwy. Mountain View, CA 94043
ABSTRACT
The Chesapeake Bay Watershed Model is designed to
simulate nutrient loads delivered to the estuary under different
management scenarios. The nutrient loads are differentiated
into anthropogenic loads amenable to management, and
nonanthropogenic loads which are considered to be
uncontrollable. The model divides the Bay basin into sixty-
three model segments with an average area of 260300
hectares. Hydrology, sediment and nonpoint source loads
Deluding atmospheric deposition) are simulated on nine
.jid uses. In the river reaches, sediment, nonpoint source
loads, point source inputs, and water supply diversions are
simulated on an hourly time-step. Paniculate and dissolved
nutrients are transported sequentially through each segment,
and ultimately to the tidal Chesapeake Bay.. The input data
used for development of the Watershed Model water is
reviewed. Model scenarios of existing loads, Chesapeake
Bay Agreement loads, year 2000 loads, and loads under limit
of technology controls are described. The model will be used
to assist the Chesapeake Bay Program in the restoration of
Chesapeake Bay water quality through nutrient load
reductions.
INTRODUCTION
"Decision makers who assess which environmental
control strategies to implement are particularly wary of two
possibilities:
1. Reducing waste inputs to a water body and
observing little or no improvement in water quality
(the environmental engineering equivalent ofbuilding
half a bridge) or,
2. Mandating control actions that are subsequently
shown to be [excessive]...(the environmental
engineering equivalent of building a bridge to
nowhere)."
Robert V. Thomann
The Chesapeake watershed drains the waters (and nutrient
loads) of seven mid-Atlantic States from New York to
Virginia. Six major basins of the watershed are the
Susquehanna, Potomac, Rappahannock/York, James, West
Shore Chesapeake, and East Shore Chesapeake (Figure 1).
Taken together, the three largest basins of the Susquehanna,
Potomac, and James, comprise 80% of the total basin area.
Land use in the watershed is predominately forest (60%),
cropland (20%), pasture (9%), and urban land (1.0%) (Figure
2). By the year 2000, urban land in the watershed will
increase to 13% due to the conversion of forest (2% decrease)
and cropland (1% decrease).
Western Shore
Eastern Short
Rappahannock/York Rivers
Susquehanna River
Potomac River
James River
Figure 1. Chesapeake Bay watershed: major basins.
ii
-------
THE CHESAPEAKE BAY WATERSHED MODEL.
,_ o.
2
o
ฃ 0.
c .
o
&0.
I
0
a
6-
4-
2
(n-8)
(n-6)
(n-5)
3 4
Segment
PO4 --A--ORGP
Figure 5. Watershed average phosphorus transport
-------
_THE CHESAPEAKE BAY WATERSHED MODEL
Susquehanna Basin
segment 6a->,
segment 6b->
>segment 5a->segment 4a->
>segment &->segment 2>segment 1>discharge to tidal Bay
segment 6c->segment 5b->segment 4b->
/
segment 5c->segment 4c->
Segments 6a, 6b, 5a, and 4a represent the East Branch Susquehanna. Segments 6c, Sb, and 4b represent the West Branch Susquehanna.
Segments 5c and 4c represent the Juniata Branch and Segments 3,2, and 1 represent the Susquehanna main stem.
Not surprisingly, phosphate P is rapidly attenuated
through incorporation into riverine sediment by
phytoplankton uptake or adsorption to particulates. This
removal mechanism is dramatically reversed during periods
* high flow and scour of sediments. During a high flow
-riod in November, 1985 (peak discharge 33 times the
average discharge of the 1991 record) in the Upper Potomac
segments, the phosphorus discharged was 1.8 times that
amount input by annual edge-of-streara loads.
As can be seen from Figure 5, attenuation of nutrient
loads is dependent on the nutrient form and the time or length
of transport Otherfactors of attenuation arenutrient limitation
and the presence of reservoirs in the system which in some
cases greatly increase the basin residence time. Examination
of the nutrient concentrations reveals these systems to be
consistently under phosphorus limitation. The slope of the
phosphate P attenuation is -0.125. The slope of the organic
P attenuation is -0.030.
Average transport losses of ammonia, nitrate, and organic
nitrogen are diagramed in Figure 6. Organic nitrogen, as in
the case of organic phosphorus, includes biomass, labile and
refractory organic and paniculate inorganic nitrogen.
Ammonia is removed quickly (slope = -0.105) in riverine
transport by incorporation into organic nitrogen through
biomass uptake and adsorption to particulates. Ammonia is
"so removed through nitrification. Nitrate (slope = -.014)
-moval mechanisms include uptake by biomass and
denitrification. The transport time in the basins is correlated
with increasing organic nitrogen primarily through conversion
of inorganic nitrogen.
MANAGEMENT SCENARIOS
Several key scenarios werecompleted in order to develop
die basic inventory of loads under specific management
conditions.
BASE CASE SCENARIO
This scenario is the base case yean 1985 loads to the
Chesapeake Bay.
BAY AGREEMENT SCENARIO
This scenario represents the nutrient loads to be reduced
by the year 2000 under the Bay Agreement. The reduction
is a 40% reduction of controllable nutrient loads. Controllable
loads were defined as the base case loads minus the loads
delivered to the Bay under the conditions of an all forested
watershed. In other words, controllable loads are defined as
everything over and above the total phosphorus or total
nitrogen loads that would have come from an entirely forested
watershed. In this definition, point source loads are considered
entirely controllable.
LIMIT OF TECHNOLOGY SCENARIO
The limit of technology scenario is defined as all cropland
in conservation tillage; the Conservation Reserve Program
fully implemented; nutrient management, animal waste
controls, and pasture stabilization systems implemented
where needed; a 20% reduction in urban loads; and point
source effluent controlled to a level of 0.075 mg/1 total
phosphorus and 3.0 mg/1 total nitrogen. An important aspect
of the limit of technology scenario is that it determined the
feasibility of the 40% nutrient reduction.
viii
-------
THE CHESAPEAKE BAY WATERSHED MODEL.
YEAR 2000 SCENARIO
This scenario represents the growth in population and
the projected changes in land use by the year 2000. The
controls in place in the base case scenario are applied to the
year 2000 point source flows and land use representing the
loading conditions without the nutrient reductions of the Bay
Agreement
1991 PROGRESS SCENARIO
This scenario applies the actual reductions made in
nitrogen and phosphorus by the year 1991.
Figures 7 and 8 show the base case total phosphorus am
total nitrogen loads relative to the Bay Agreement, the"'
of technology, the year 2000, and the 199 1 Progress
Chesapeake TP loads by scenario.
4
In all basins, the limit of technology phosphorus loads
are less than the Bay Agreement loads. For nitrogen loads
only the point source dominated basins of the Potomac
James, and the West Shore have limit of technology loads
appreciably less than Bay Agreement loads. For nonpoint
source dominated basins of the Susquehanna, Rappahannock
York, and the East Shore, the limit of technology
loads are essentially equivalent to the Bay
Agreement loads.
Smquduuma PCXOOMC June* lUplTork Weft Shore EUL Shore
Base S1991 Progress BBay Agreement BLinit of Tech. QYeซr2000
The Year 2000 scenario has increased loads
of phosphorus and nitrogen for all .basins. Th.
point source dominated basins of the Potomac,
James, and West Shore, which experience the
greatest urbanization in the Year 2000 scenario
have the greatest increase in nutrient loads.
The 1991 Progress Scenario indicates the
phosphorus reductions are on track to reach th
year 2000 goal The nitrogen reductions posted
by the 1991 progress scenario are m'
Overall progress in nitrogen reduction^
need to increase by a factor of 3.7 in order to
reach the year 2000 goal
Figure 7. Chesapeake total phosphorus
loads by scenario.
Figure 8. Chesapeake total
nitrogen loads by scenario.
Chesapeake TN loads by scenario.
70
Swquetanm Potomac Jtmei Rip/York Wett Shore E** Shore
Base ฃ31991 Progress ฎBซy Agreement 0 Limit of Tech. OYeir 2000
ix
CSC.MOU:
-------
.THE CHESAPEAKE BAY WATERSHED MODEL
Another aspect of the Progress Scenario is that the Bay
Agreement reductions become nutrient load caps after the
year 2000. The Watershed Model is envisioned as the tool
which will track the changes in watershed loads to ensure the
caps are not exceeded in any basin. .
CONCLUSION
The Watershed Model is a successful water quality
simulation which quantifies the nonpoint source and point
source nutrient loads from all basin sources. The model was
essential in establishing a consistent method of accounting
for the nutrient loads, among all sources, and among the
basin jurisdictions of the Bay Program. The model is used to
examine the level of controls achievable from different
management practices. The model has examined management
practices, which when combined into strategies of nutrient
control, will meet the Chesapeake Bay Program water quality
' objective of a 40% reduction in controllable nutrients in the
Chesapeake watershed and do so in a cost-effective and
equitable manner.
ACKNOWLEDGMENTS
The authors wish to acknowledge the state, regional, and
federal members of the Chesapeake Bay Program Modeling
Subcommittee for their essential guidance and direction
throughout the development of the Chesapeake Bay
Watershed Model, and in particular Dr. Robert Thomann for
his expertise and experience gained through 23 years of
water quality modeling work on the Chesapeake and its
tributaries. .
REFERENCES
Donigian. Jr.. A.S., and H:H. Davis (1978). Users Manual for
Agricultural Runoff Management (ARM) Model. U.S. EPA
Environmental Research Laboratory. Athens. GA.
Donigian, Jr.. A.S., B.R. Bicknell, L.C. Linker, J. Hannawald,
C.H. Chang, and R. Reynolds (1990). Watershed Model
Application to Calculate Bay Nutrient Loads: Phase I
Findings and Recommendations. U.S. EPA Chesapeake
Bay Program; Annapolis, Mb.
Donigian, Jr., A.S., B.R. Bicknell, A.S. Patwardhan, L.C.
Linker, C.H. Chang, and R. Reynolds (1991). Watershed
Model Application to Calculate Bay, Nutrient Loads:
Phase II Findings and and Recommendations, U.S. EPA
Chesapeake Bay Program, DRAFT. Annapolis, MD.
Conservation Technology Information Center (1985) 198S Crop
Tillage Data Base. West Lafayette, IN.
Hartigan. J.P. (1983). Chesapeake Bay Basin Model Final
Report prepared by the Northern Virginia Planning District
Commission for the U.S. EPA Chesapeake Bay Program.
Annapolis, MD.
Bicknell, B.R., J.C. Imhoff, J.L. Kittle. Jr.. A.S. Donigian
Jr., and R.C. Johanson. 1993. Hydrological Simulation
Program - Fortran: User's Manual for Release 10. U.S.
EPA, ERL, Athens. GA. EPA/60O-R-93/174.
National Atmospheric Deposition Program (1982); (1983);
(1984); (1985); (1986); (1997). (IR7)/National Trends
Network. NADP/NTN Coordination Office, Natural
Resource Ecology Laboratory, Colorado State University.
Fort Collins. CO.
U.S. Bureau of the Census (1984). 1982 Census of Agriculture.
Volume 1, Geographic Area Series. U.S. Department of
Commerce. Washington, DC.
U.S. Department of Agriculture (1984). Soil Interpretations
Records (SCS-SOI-5 data file) and the National Resources
Inventory (NRI), Soil Conservation Service, Washington,
DC.
U.S. Geological Service (1974) Geographic Information
Retrieval and Analysis System (GIRAS) 1:125,000 land
use/landcover (Anderson Level II classification.
Washington, D.C.
CSC.MOU.7/93
-------
ABSTRACT
Since 1985 the EPA Chesapeake Bay Program (CBP) has sponsored a series of
projects to convert the proprietary HSP-NPS model to the public domain HSPF
model, make it operational on the CBP VAX computer system, and implement a number
of model refinements so that the resulting 'Watershed Model' could be used
effectively to estimate nutrient loadings to the Bay and to evaluate the impacts
of agricultural Best" Management Practices (BMPs). A comprehensive workplan
developed in September 1987 proposed a two-phased approach to address Watershed
Model deficiencies and produce an improved and re-calibrated model. Phase I was
designed to improve the nonpoint loading representation, refine/re-evaluate the
data input to the Model, and perform a preliminary re-calibration to available
water quality data for the 1984-85 period. This period was used as the basis
for the 401 nutrient loading reduction goal defined by the 1987 Chesapeake Bay
Agreement.
Phase II was designed to focus on a better representation of the effects of
agricultural BMPs, including application of the detailed Agrichemical Modules
(AGCHEM) of HSPF to allow a more deterministic, process-oriented approach to BMP
analysis and evaluation. AGCHEM provides a nutrient balance approach to modeling
cropland areas so that sensitivity to nutrient inputs can be represented. In
addition, HSPF code enhancements were performed to allow specific consideration
of sediment-nutrient and bed interactions within the stream channel. The Phase
II Watershed Model was then applied to the Bay drainage area to include sediment
erosion, sediment transport, and associated nutrients, in addition to the current
modeled water quality constituents, to provide input to the Chesapeake Bay 3-D
Water Quality Model.
This report presents the results of the combined Phase I/Phase II effort. Model
calibration results are presented for hydrology, nonpoint loadings, and instream
water quality for numerous sites throughout the Bay drainage. In addition, model
estimates of Fall Line loads for various nutrient forms are compared with
regression estimates used in the calibration of the Bay Model. The results
indicate that the Watershed Model is a valid representation of nutrient loadings
to the Bay, and provides a framework for assessing the, impacts of land use
changes and alternative nutrient and agricultural management practices.
Refinements and recommendations are provided to improve model simulations and its
use as a tool to assist in the 1991 Re-evaluation of the 402 nutrient reduction
goal of the Bay Agreement.
This report is submitted in fulfillment of Work Assignment No. 7(B) by AQUA TERRA
Consultants under EPA Contract No. 68-CO-0019 with the EPA Athens Environmental
Research Laboratory. Earlier work for Phase I of the CBP Watershed Model re-
calibration was performed under EPA Contact No. 68-03-3513.
xi
-------
CONTENTS
Page
Foreword i
The Chesapeake Bay Watershed Model . .. il
Abstract. < xi
Tables xv
Figures. , xviii
Acknowledgments >..... xxii
List of Acronyms , ; xxiy
1.0 INTRODUCTION . . . .......... % ... 1
1.1 Background ' . . 1
1.2 Overview of the Phase I/Phase II Efforts .................... 2
1.3 Phase I Preliminary Findings and Recommendations ..... 9
1.3.1 Hydrology Conclusions and Recommendations ......:....... ' 9
1.3.2 Phase I Nonpoint Loading Conclusions and
Recommendations 10
1.3.3 Phase I Water Quality Simulation Conclusions, and
Recommendations ..... 11
1.4 Phase II Findings and Recommendations 12
1.4.1 Phase II Hydrology Conclusions and Recommendations.... 12
1.4.2 Phase II Nonpoint Loading Conclusions and
Recommendations. 13
1.4.3 Phase II Water Quality Simulation Conclusions and
Recommendations. 16
1. 5 General Watershed Model Recommendations 20
1.6 Format of this Report 21
2.0 KETEOROLOGIC AND PRECIPITATION DATA BASE DEVELOPMENT PROCEDURES.... 22
2.1 Introduction 22
2.2 Regional Meteorologic Data 22
2.2.1 Cloud Cover and Dewpoint Temperature 25
2.2.2 Air Temperature and Wind Speed 25
2.2.3 Solar Radiation ..'...:.,. ... 25
2.2.4 Potential Evaporation ^ 27
2.3 Hourly Precipitation Data '.'.- 'ป 28
2.3.1 Development of Observed Rainfall Data Base ....... 28
2.3.2 Basin Segmentation and Development of
Weighting Factors 28
2.3.3 Generation of Weighted Hourly Rainfall Records ....... 29
Xii
-------
3.0 LAND USE DATA AND WASTE LOAD INPUTS . 31
3.1 Land Use Data 31
3.1.1 Land Use for 1978 32
3.1.2 Land Use for 1985 33
3.1.3 Cropland Tillage . . 33
3.1.4 Water Acres 33
3.1.5 Manure Acres 34
3.1.6 Urban Land Subcategories 35
3.1.7 Land Use by County .... 37
3.1.8 Land Use by Model Segment 37
3.1.9 Land Surface Cover and Erodlbillty Parameters -
COVER, SLSUR, KRER ..... 37
3.1.10 Crop Distributions for Conventional (CNT) and
: Conservation (CST) Tillage. 38
3.1.11 Comparison of Forest Land. 39
3.2 Point Sources ; '._.. 39
- 3.3 Surface Water Diversions 44
3.4 Atmospheric Sources. ... 45
4.0 Nonpoint Loading Simulation Modules and Simulation Results........ 47
4.1 Overview of Cropland and Non-Cropland Simulation.... 47
4.1.1 Hydrologic, Sediment Loading, DO, and Water '
Temperature Simulation 54
4.2 Simulation of Non-Cropland Areas 60
4.2.1 Forest, Pasture, Urban Area Procedures.. . 60
4.2.2 Animal Waste Procedures 63
4.3 Simulation., of Cropland Areas . 64
4.3.1 AGCHEM Overview 64
4.3.2 Soil Nutrient Processes Modeled by AGCHEM.. 68
4.3.3 Composite Crop Representation............ป............ 72
4;3.4 Development of Model Segment Nutrient Application
Rates...; ,... 79
4.3.5 AGCHEM Application Procedures for Model Segments 103
4.3.6 AGCHEM Simulation Results.... 121
4.3.7 AGCHEM Conclusions and Recommendations 135
4.4 Comparison of Expected and Simulated Nonpoint Loading Rates.. 136
5.0 RIVER TRANSPORT AND TRANSFORMATION SUBMODEL. 151
5.1 Overview of the Instrearn Model .' 151
5.2 Model Structure and Processes , 154
5.2.1 Hydraulics 154
5.2.2 Water Temperature 155
5.2.3 Sediment Transport.. 157
5.2.4 Water Quality Processes. 161
5.3 Idealized Channel Network 174
6.0 HYDROLOGY CALIBRATION RESULTS.... 178
6.1 Overview and Steps in the Hydrologic Calibration Process..... 178
6.2 Snow Simulation 179
6.2.1 Identification of Areas Needing Snow Simulation 179
6.2.2 Snow Calibration and Simulation Results 182
6.3 Adjustment of Hydrologlc Parameters for Land Use 187
xiii
-------
6.4 Phase I Hydrologlc Re-Calibration for 1974-78 189
6.5 Phase .II Hydrology Calibration for 1984-87 193
7.0 WATER QUALITY CALIBRATION RESULTS '. 206
7.1 Overview and Steps in the Water Quality Calibration Process.. 206
7.2 Sediment Loading and Instream Transport Calibration... v 207
7.2.1 Potency Factor Adjustments 215
7.3 Water Quality Calibration Results.......; 215
7.3.1 Instream Water Quality Concentrations 216
7.3.2 Fall Line Load Simulation and Regression Comparison... 220
7.4 Analysis of Segment and Basin Loads 221
8.0 REFERENCES 274
APPENDICES .(Available from CBPO, Annapolis, MD)
A. Hydrology Calibration Results
B. Water Quality Calibration Results
C. Nonpoint, Loading Simulation Results
and Nutrient Application Rates
Addendum to Appendix C
D. Meteorologic and Precipitation Data Development
E. Land Use and Selected Parameter Values
F. Point Sources, Diversions, and Atmospheric Deposition
-------
TABLES
Page
1.1 Subbasin, Segment Name, and Segment Numbers of the Chesapeake
Bay Watershed Model. . 6
2.1 Correspondence "Between 1974-78 Regional Data and
Published Data 23
2.2 Stations Used to Develop 1984-87 Regional Meteorologic Data . 23
3.1 Estimated Quantities of Voided Manure from Livestock and
Poultry. Normalized to 1,000 Pounds Body Weight.. 34
3.2 Crop Distributions for Conventional and Conservation Tillage
for Each Watershed Model Segment 40
3.3 Municipal Nitrogen Default Percentages of Ammonia, Nitrate
and Organic Nitrogen, and the Default Concentrations of Total
Nitrogen. Defaults Based on Levels of Secondary Treatment 42
3.4 Municipal Phosphorus Default Percentages of Phosphate and
Organic Phosphorus, and the Default Concentrations of Total
Phosphorus. Defaults Based on Levels of Secondary Treatment..... 42
3.5 Basin-Wide Average Annual Atmospheric Deposition Loads 45
4.1 Land Use Categories in the Watershed Model. 48
4.2 Water Quality Constituents Simulated in the Watershed Model 49
4.3 HSPF Modules used to Simulate Nonpoint Loadings from Each
Land Use in Phase II 53
4.4 Crop Cover Values for all .Regions and Practices 75
4.5 Fertilizer Application Rates Supplied by the NRTF ... 82
4.6 Manure Application Rates Supplied by the NRTF. 84
4.7 Percentage of Cropland Receiving Fertilizer and/or Manure
as Supplied by the NRTF. 86
4.8 Estimated Percentages of Multi-State Segments in Various States.. 90
4.9 Annual Nitrogen Atmospheric Deposition Loads 93
4.10 Model Procedures for Timing and Placement of Fertilizer/
Manure Applications and Atmospheric Deposition. 100
4.11 Example Calculations of Final Nutrient Application Rates 101
4.12 Example Calculations of Timing and Placement of Final
Nutrient Application Rates 102
4.13 Final Calculated Model Segment Nutrient Application Rates for .
Conventional Tillage 104
4.14 Final Calculated Model Segment Nutrient Application Rates for
Conventional Tillage 106
4.15 Final Calculated Model Segment Nutrient Application Rates
for Hay 108
.xv
-------
4.16 Typical Nitrogen Balance for Major Crops When Nitrogen is
Applied at Agronomic Rates. 112
4.17 Typical Phosphorus Balance for Major Crops When Phosphorus
is Applied at Agronomic Rates 113
4.18 Crop Nutrient Uptake Targets 114
4.19 Nutrient Transformation and Sorption Parameters used in the
.AGCHEM Simulations 119
4.20 Plant Nutrient Uptake Rates used in AGCHEM Simulations 120
4.21 AGCHEM Results for Juniata Segment 100 (HT Composite Crop) 122
4.22 AGCHEM Results for Juniata Segment 100 (LT Composite Crop) 123
4.23 AGCHEM Results for Juniata Segment 100 (Hay) 124
4.24 AGCHEM Simulated Nutrient Runoff Losses for Selected Model
Segments 128
4.25 Summary of Observed Concentrations from Selected Field
Monitoring Stuflies .... 131
4.26 Summary of Rappahannock (Segment 230) Simulated Concentrations
by AGCHEM .. 132
4.27 Summary of Patuxent (Segment 340) Simulated Concentrations
by AGCHEM. 133
.4.28 Summary of Monocacy (Segment 210) Simulated Concentrations
by AGCHEM... ...:... '. 134
4.29 Expected Ranges of Land Use Loadings Developed in Phase I..-7 138
4.30 Mean and Range of Loading Rates for Each Land Use from
Various Sources for the Chesapeake Bay Region..... 140
4.31 Nutrient Expert Coefficients from Beaulae et al. (1980)...... 141
4.32 Watershed Model Segment 60 Annual Loading Rates 142
4.33 Watershed Model Segment 180 Annual Loading Rates. 143
4.34 Watershed Model Segment 280 Annual Loading Rates 144
4.35 Summary of Average Annual Loading Rates for Each Land Use in
Selected Model Segments; .... 145
4.36 Overall Simulated Mean and Range of Loading Rates from
Selected Segments Listed in Table 4.35 149
t .
5.1 Water Quality Constituents and Processes Simulated in
the Watershed Model 162
5.2 Instream Water Quality Parameters 175
5.3. Channel Reach Information. - 177
6.1 Elevation, Temperature and Snow Depths for Watershed
Model Land Segments 180
6.2 Land Use Cover Values Used in the Watershed Model for
Non-Cropland Areas ; 188
6.3 Watershed Model Hydrology and Water Quality Calibration
Stations ,.... 190
6.4 Summary of Chesapeake Bay Watershed Hydrologic Calibration
^ vs. Simulated Flows , 191
6.5 Summary of Watershed Model Hydrologic Calibration Seasonal
R2 Values, 1984-1987.. 205
7.1 Total Annual Suspended Sediment Loads at Selected Locations 209
7.2 Susquehanna River Fall Line Loads: Comparison of Regression
and Model Estimates 222
-------
7.3 Potomac River Fall Line Loads: Comparison of Regression
and Fall Line Estimates 223
7.4 James River Fall Line Loads: Comparison of Regression
and Model Estimates . 224
7.5 Patuxent River Fall Line Loads: Comparison of Regression
and Model Estimates 225
xvii
-------
FIGURES
Page
1.1 Chesapeake Bay watershed and major above fall line
subbasins 3
1.2 Above fall line watershed model segments. 4
1.3 Below fall line watershed model segments 5
2.1 Meteorologic regions and principal stations in the
Chesapeake Bay Watershed Model 24
2.2 Diurnal variation of air temperature 26
2.3 Diurnal variation of wind speed 26
2.4 AFL Daily and Hourly Precipitation Stations ...-... 28,1
2.5 BFL Daily and Hourly Precipitation Stations.-.. 28.2
4.1 PERLND structure chart. ..... 50
4.2 I^PLND structure chart impervious land-segment application .
module ;.......... 51
4.3 Hydrologic processes simulated by the PWATER module of HSPF 55
4.4 Flow diagram for SEDMNT section of PERLND application module..... 57
4.5 Flow diagram for PQUAL section of PERLND application module...... 61
4.6 Nutrient storages and transport modeled by the AGCHEM Module..... 65
4.7 Soil profile representation by the AGCHEM Model.. 67
4.8 Nutrient transformations simulated by. the AGCHEM Module 69
4.9 Crop COVER region boundaries 74
4.10 Composite crop COVER values for Segment 100 - conventional
tillage...... i.. 77
4.11 Composite crop COVER values for Segment 100 - conservation
tillage . T...... 78
4.12 Flow chart for development of model input nutrient ' ' '
application rates 80
4.13 Composition of nutrient inputs to AGCHEM Model ....... 95
' 4.14 Plant nutrient inputs - Juniata Segment 100 125
4.15 Comparison of nutrient losses Juniata 100. 126
4.16 Nutrient runoff losses - Juniata 100 . .................... 127
5.1 RCHRES structure chart 152
5.2 Structure chart of the RQUAL section of the RCHRES module. 153
5.3 Flow diagram for HTRCH section of the RCHRES module 156
5.4 Flow diagram of inorganic sediment fractions in the SEDTRN
section of the RCHRES module. 158
5.5 Simulated bed shear stress in Reach 290. 160
5.6 Instream nitrogen dynamics in the Watershed Model. 164
5.7 Instream phosphorus dynamics in the Watershed Model...... 165
; xviii
-------
5.8 Flow diagram of dissolved oxygen in the OXRX section of the
RCHRES module , 166
5.9 Flow diagram of BOD in the OXRX section of the RCHRES module 167
5.10 AFL model segments and river channel reaches 176
6.1 Watershed model segments for snow simulation 181
6.2 Simulated and observed snow depth for Model Segment 20, 1984-87.. 184
6.3 Simulated and observed snow depth fpr Model Segment 50. 1984-87.. 185
6.4 Simulated and observed snow depth for Model Segment 160, 1984-87. 186
6.5 Simulated and observed flow frequency for Harrisburg, PA
Segment 80, 1984-87. 195
6.6 Simulated and observed daily flow for Harrisburg, PA,
Segment 80, 1984-87 196
6.7 Simulated and observed flow frequency for Conowingo
Reservoir, MD, ^Segment 140, 1984-87.... 197
6.8 Simulated and observed daily flow for Conowingo Reservoir
MD, Segment 140, 1984-87 198
6.9 Simulated and observed flow frequency for Potomac River,
Washington, DC, Segment 220, 1984-87 199
6.10 Simulated and observed daily flow for Potomac River,
Washington, DC, Segment 220, 1984-87. ... 200
6.11 . Simulated and- observed flow frequency for James River,
Cartersville; VA, Segment 280, 1984-87..... 201
6.12 Simulated and observed daily flow for James River,
Cartersville, VA, Segment 280, 1984-87. 202
7.1 Suspended sediment concentration Susquehanna River at
Conowingo * ......ป.........;..............;. 210
7.2 Suspended sediment concentration - Potomac River at Chain
Bridge.. . . .. 211
7.3 Suspended sediment concentration - Juniata River at Newport 212
7.4 Suspended sediment concentration - Patuxent River at Bowie 213
7.5 Suspended sediment load - Juniata River at Newport....; 214
7.6 Simulated and observed water temperature (ฐC), Susquehanna
River at Harrisburg, PA, 1984-87. 226
7.7 Simulated and observed dissolved oxygen (mg/1), Susquehanna
River at Harrisburg, PA, 1984-87. ..r..... .227
7.8 Simulated and observed suspended sediment (mg/1), Susquehanna
River at Harrisburg, PA, 1984-87. 228
7.9 Simulated and observed nitrate (mg/1), Susquehanna River
at Harrisburg, PA, 1984-87.... ...... 229
7.10 Simulated and observed ammonia (mg/1),, Susquehanna River
at Harrisburg, PA, 1984-87. 230
7.11 Simulated and observed particulate nitrogen (mg/1), Susquehanna
River at Harrisburg, PA, 1984-87 231
7.12 Simulated and observed total nitrogen (mg/1), Susquehanna ,
River at Harrisburg, PA, 1984-87. ,. 232
7.13 Simulated and observed phosphate (mg/1), Susquehanna River
at Harrisburg, PA, 1984-87 ., 233
7.14 Simulated and observed particulate phosphorus (mg/1).
Susquehanna River at Harrisburg, PA, 1984-87 234
xix
-------
7.15 Simulated and observed total phosphorus (mg/1), Susquehanna
River at Harrisburg, PA, 1984-87 235
7.16 Simulated and observed total organic carbon (mg/1), Susquehanna
River at Harrisburg, PA, 1984-87 236
7.17 Simulated and observed chlorophyll A (ug/1) , Susquehanna
River at Harrisburg, PA, 1984-87 . 237
7.18 Simulated and observed water temperature (ฐC), Susquehanna
River at Conowingo, MD, 1984-87 238
7.19 Simulated and observed dissolved oxygen (mg/1), Susquehanna
River at Conowingo, MD, 1984-87................. 239
7.20 Simulated and observed nitrate (mg/1), Susquehanna River
at Conowingo, MD, 1984-87 240
7.21 Simulated and observed ammonia (mg/1), Susquehanna River
at Conowingo, MD, 1984-87 241
7.22 Simulated and flbserved particulate nitrogen (mg/1),
Susquehanna River at Conowingo, MD, 1984-87 242
7.23 Simulated and observed total nitrogen (mg/1), Susquehanna
River at Conowingo, MD, 1984-87 243
7.24 Simulated and observed phosphate (mg/1), Susquehanna River
at Conowingo, MD, 1984-87 ...; . 244
7.25 Simulated and observed particulate phosphorus (mg/1),
Susquehanna River at Conowingo, MD, 1984-87 245
7.26 Simulated and observed total phosphorus (mg/1), Susquehanna
River at Conowingo, MD, 1984-87... .7 246
7.27 Simulated and observed total organic carbon (mg/1), ."',''..
Susquehanna River at Conowingo, MD, 1984-87.... ... 247
7.28 Simulated and observed chlorophyll A (ug/1), Susquehanna
River at Conowingo, MD, 1984-87..'..' -.. 248
7.29 Simulated and observed water temperature (ฐC), Potomac River
at Washington, DC, 1984-87 ; 249
7.30 Simulated and observed dissolved oxygen (mg/1), Potomac River
at Washington, DC, 1984-87 4. '-,. 250
7.31 Simulated and observed nitrate (mg/1), Potomac River at
Washington, DC, 1984-87.. , 251
7.32 Simulated and observed ammonia (mg/1), Potomac River at
Washington, DC,. 1984-87.......... 252
7.33 Simulated and observed particulate nitrogen (mg/1), Potomac
River at Washington, DC, 1984-87. . ....-.- 253
7.34 Simulated and observed total nitrogen (mg/1), Potomac River
at Washington, DC, 1984-87. . 254
7.35 Simulated and observed phosphate (mg/1), Potomac River at
Washington, DC, 1984-87 , 1.' 255
7.36 Simulated and observed particulate phosphorus (mg/1),
Potomac River at Washington, DC, 1984-87 256
7.37 Simulated and observed total phosphorus (mg/1), Potomac
River at Washington, DC, 1984-87 257
7.38 Simulated and observed total organic carbon (mg/1), Potomac
River at Washington, DC, 1984-87. 258
7.39 Simulated and observed chlorophyll A (ug/1), Potomac River
at Washington, DC, 1984-87... 259
7.40 Simulated and observed water temperature (ฐC), James River
at Cartersville, VA, 1984-87...... 260
OCX
-------
7.41 Simulated and observed dissolved oxygen (mg/1). James River
,at Cartersviile, VA, 1984-87..... 261
7.42 Simulated and observed suspended sediment (mg/1), James River
at Cartersviile, VA, 1984-87 262
7.43 Simulated and observed nitrate (mg/1), James River at
Cartersviile, VA, 1984-87 '. 263
7.44 Simulated and observed ammonia (mg/1), James River at
Cartersviile, VA, 1984-87 264
7.45 Simulated and observed particulate nitrogen (mg/1), James
River at Cartersviile, VA, 1984-87 .... 265
7.46 Simulated and observed total nitrogen (mg/1), James River
at Cartersviile, VA, 1984-87 266
7.47 Simulated and observed phosphate (mg/1), James River at
Cartersviile, VA, 1984-87 267
7.48 Simulated and observed particulate phosphorus (mg/1), James
River at Cartersviile, VA, 1984-87 268
7.49 Simulated and observed total phosphorus (mg/1), James River
at Cartersviile, VA, 1984-87.. 269
7.50 Simulated and observed total organic carbon (mg/1), James
River at Cartersviile, VA, 1984*87.. 270
7.51 Simulated and observed chlorophyll A (ug/1), James River
at Cartersviile, VA, 1984-87...... . 271
7.52 Simulated and observed-total nitrogen (mg/1), Patuxent River
at Bowie. MD, 1984-87 ..... 272
7.53 Simulated and observed total phosphorus (mg/1), Patuxent River
at Bowie, MD, 1984-87, :....,..... 273
xxi
-------
ACKNOWLEDGMENTS
A large number of individuals and organizations contributed to the successful
completion.of this Watershed Model re-calibration and re-application effort. The
financial support was provided by the Chesapeake Bay Program with administrative
and contractual assistance through the EPA Athens Environmental Research
Laboratory (ERL) in Athens, GA. Mr. Robert Carsel was the Project Officer for
the Athens ERL, and Mr. Charles App was the Watershed Modeling Coordinator for
the first phase of the study; Mr. Lewis Linker subsequently replaced Mr. App and
assumed both technical and administrative responsibilities on the project. All
these individuals contributed to the project by their assistance in
administrative and contractual matters; this assistance is gratefully
acknowledged. .
The project was guided by a number of technical committees that met on a regular
basis to provide thorough technical review of all aspects of the study. These
committees included the Modeling Subcommittee, the Nonpoint Source Committee, the
Nutrient Reduction Task Force (NRTF), and the Science and Technical Advisory
Committee. It would be impossible to accurately acknowledge the specific
contributions of all the individuals on these committees, but their
participation, review, and input was a major factor in the conduct of this work.
Mr. James Collier, as Chair of the Modeling Subcommittee, played a key
.coordination and oversight role among the various agencies and committees.
The NRTF deserves special mention because its various state members were
instrumental in providing the information needed to develop appropriate input for
the AGCHEM modeling of the cropland areas, including the nutrient application
rates, crop cover values, tillage practices, etc. Mr. Jin Cox, as Chair of the
NRTF, was a key force in this effort, in addition to the various state
representatives including Dr. Robert Summers and Mr. John McCoy of MDE, Mr. Mike
Flagg of VA DSWC, Dr. Elizabeth Gasman of ICPRB, Ms. Veronica Kasi of PA BUSC and
Mr. Fred Samadani of MDA. the AGCHEM model application would not have been
possible without the valuable assistance from these individuals. The assistance
of Joel Myers, SCS is acknowledged for.his contributions to the Pennsylvania
agricultural data. In addition, Mr. James Hannawald, SCS. played a major role
in the land use evaluation, soil parameter assessment, and crop cover
representation and provided invaluable knowledge of agricultural conditions and
practices within the Bay drainage.
The CBP Watershed Model Evaluation Group was convened at various times throughout
the study to provide expert technical review of the modeling procedures and
results. Mr. Thomas Bamwell of Athens ERL, Dr. Alan Lumb of the USGS, and Dr.
Robert Thomann of Manhattan College provided helpful and insightful review
comments as members of this group. Their participation is sincerely appreciated.
xxii
41
-------
In addition to the authors, various individuals at AQUA TERRA Consultants and the
CBPO provided key technical and staff support roles throughout the study. At
AQUA TERRA, Mr. John Imhoff was a key participant in the development of the
sediment-nutrient interaction capabilities and assisted in the water quality
parameter review and calibration; Mr. Jack Kittle assisted in the Conowingo
Reservoir operation program, the rainfall generation program, and the ANNIE/WDM
application; and Ms. Dorothy Inahara and Ms. Mary Shandonay provided
administrative, word processing, and graphics support.
At CBPO, in addition to the authors, several individuals provided guidance and
assistance. Mr. Ed Stigall, Technical Programs Director, and Mr. Lynn Shuyler,
Nonpoint Source Coordinator, assisted in bridging the gap between the technical
aspects of the model and its application to water quality problems in the
Chesapeake Bay. Their assistance is gratefully acknowledged. Many of the staff
of the Computer Scierices Corporation were essential for the successful execution
of a number of tasks throughout this project, with special appreciation to Scott
McDiarmid and Lacy Williams for satisfying the variety of requisites, large and
small, made of them by HSPF or it's modelers. Ms. Diane Allegre, the Modeling
Intern, is thanked for her many contributions to the appendices and the report.
xxiii
-------
TABLE OF ACRONYMS AND ABBREVIATIONS
AFFIX
AFL
AGCHEH
ARC-INC
ARM
ATEMP
BFL
BMP
BOD
BtU
BWSC
C
CBP
CBPO
COVER
CNT
CSC
An HSPF parameter .representing the incorporation of detached
sediment into the soil matrix where it would be unavailable for
transport. . .
X ..
Above Fall Line.
Agricultural/Chemical section of HSPF. An HSPF module which
simulates agricultural nutrient loads in detail in PERLND.
A type of CIS software.
Agricultural Runoff Model. . . ' .
An HSPF module used to adjust air temperature for adiabatic
differences between the measuring station and the average.
altitude of the model segment; .
Below Fall Line.
Best Management Practice. . .
Biochemical Oxygen Demand.
British Thermal Units heat metric.
Bureau of Water and Soil Conservation (PA).
Centigrade temperature.
Chesapeake Bay Program.
Chesapeake Bay Program Office. ,
An HSPF parameter used to specify the degree to which the soils
are protected for rain splash soil detachment by vegetation,
organic debris, or vegetative litter.
Conventional Tillage land use.
Computer Science Corporation.
xxiv
-------
CTIC
CST
3-D Water
Quality
Model
DETS
DSUC
DO
EPA
ERL
F
FHA
FTABLE
CIS
HSPF
HSP-NPS
HTRCH
HYDDAY
HYDHR
HYDR
HYDSDY
IMPLND .
ICPRB
IQUAL
Conservation Technology Information Center.
Conservation Tillage land use.
The Chesapeake Bay Three-Dimensional Time-Variable Model, a
coupled hydrodynamic and water .quality model of the Chesapeake
Bay. The Watershed Model simulates the nutrient loads used as
input in the 3-D Water Quality Model.
An HSPF parameter used to represent detached sediment storage
in SEDMNT.
Department of Soil and Water Conservation (VA).
Dissolved Oxygen concentration.
Environmental Protection Agency.
Environmental Research Laboratory (US EPA).
Fahrenheit temperature.
Federal Highway Administration.
An HSPF input file table which relates water volume to depth,
surface area, and flow rate for a reach or reservoir.
Geographic Information System. '
Hydrologic Simulation Program - Fortran. Model code used as,
a basis for the Watershed Model.
Hydrologic Simulation Program - Nonpoint Source - A proprietary
model code. .
An HSPF module used to simulate heat exchange and water
temperature in RCHRES.
App D.
App D.
An HSPF module used to simulate hydraulic behavior in RCHRES.
App D.
Impervious Land simulation section of HSPF. An HSPF module
used to simulate the pollutant loads from impervious land uses.
Interstate Commission of the Potomac River Basin.
An HSPF module which simulates water quality in IMPLND.
XXV
-------
IWTGAS
K
KPL
KRER
KSER
LU/LC
MAGI
MDA
MDE
MGD
MS1AY
N
NADP
NASA
NHAPP
NITR
NOAA
NPDES
NPS
NRI
NRTF
NUTRX
NVPDC
An HSPF module which simulates water temperature and dissolved
gas concentration in IMPLND.
A USLE parameter used to represent soil credibility.
An HSPF parameter representing plant uptake rates of nutrients
in AGCHEM. .
An HSPF parameter representing the erodibility of soil.
An HSPF parameter used to adjust the amount of soil fines
detached by raindrop impact.
Land Use/Land Cover.
^
Maryland Geographic Information System.
Maryland Department of Agriculture.
Maryland Department of Environment
Millions of Gallons per Day.
An HSPF module which simulates solute transport in AGCHEM.
Nitrogen nutrient. ,
National Atmospheric Deposition Program.
National Aeronautics and Space Administration.
National High-Altitude Photography Program.
An HSPF module which simulates nitrogen in AGCHEM.
National Oceanographic and Atmospheric Administration.
National Pollution Discharge Elimination System.
Nonpoint Source, a term describing pollutants from diffuse
sources. . .
National Resource Inventory; .
Nutrient Reduction Task Force.
An HSPF module which simulates primary inorganic nitrogen and
phosphorus reactions in EQUAL.
Northern Virginia Planning District Commission.
-------
NVSI
ORGN
ORGP
OXRX
An HSPF parameter used to represent net external additions or
removals of sediment caused by atmospheric deposition or wind.
Organic Nitrogen.
Organic Phosphorus.
An HSPF module used to simulate primary DO and BOD reactions
in RQUAL.
Phosphorus nutrient.
PANEVAP
PERLND
PLANK
PRECIP
PQUAL
PHOS
PSTEMP
PWATER
PWTGAS
RADCLC
RCHRES
RQUAL
SCS
A FORTRAN subroutine used to develop pan evaporation rate time
series for the Watershed Model.
Pervious Land simulation section of HSPF. An HSPF module used
to simulate the pollutant loads from pervious land uses.
An HSPF module which simulates plankton in RQUAL.
A FORTRAN subroutine used to develop an hourly time series of
rainfall data.
Vater quality section of HSPF. An HSPF module which simulates
water quality in PERLND.
Detailed phosphorus simulation section of HSPF. An HSPF module
which simulates phosphorus in AGCHEM.
Soil temperature section of HSPF. An HSPF module Which
simulates soil temperature in PERLND.
Water budget section of HSPF. An HSPF module used to simulate
the water budget in PERLND.
Water temperature section of HSPF. An HSPF module which
simulates water temperature and gas concentrations in PERLND.
A FORTRAN subroutine, used to develop solar radiation time
series for the Watershed Model.
Reach and Reservoir section of HSPF. An HSPF module used to
simulate transport and transformation in the river reaches and
reservoirs.
An HSPF.module used to simulate water quality constituents in
RCHRES. .
Soil Conservation Service.
xxvii
-------
SEDMNT
An-HSPF module which simulates sediment to the edge of stream
in PERLND.
SEDTRN
SLSUR
SLSED
SNOU
SOLIDS
TAUCD
TAUCS
TO
TOC
TP
UAN
USDA
USGS
USLE
VIRGIS
WDM
An HSPF module which simulates sediment transport in RCHRES.
An HSPF parameter that represents the average land slope of a
model segment.
HSPF parameter that represents external lateral input of
sediment from an upland segment.
Snow simulation section of HSPF. An HSPF module used to
simulate snow accumulation and melt.
^ '
An HSPF module which simulates the accumulation and removal of
solids in IMPLND.
An HSPF parameter which specifies the critical shear stress for
sediment deposition in SEDTRN.
An HSPF parameter which specifies the critical shear stress for
sediment scour in SEDTRN. .
Total Nitrogen. -o '
Total Organic Carbon.
Total Phosphorus.
Urea, Ammonium, Nitrate liquid fertilizer.
United States Department of Agriculture.
United States Geological Survey.
Universal Soil Loss Equation model.
Virginia Geographic Information System. .
Watershed Data Management system. System data files by HSPF.
xxviii
-------
SECTION 1.0
INTRODUCTION
1.1 BACKGROUND
Since 1985 Che EPA Chesapeake Bay Program (CBP) has sponsored a series of
projects to convert the proprietary HSP-NPS model to the public domain HSPF model
(Jchanson et al, 1984), make it operational on the CBP VAX computer system, and
implement a number of model refinements so that the resulting 'Watershed Model'
could be used effectively to estimate nutrient loadings to the Bay and to
evaluate the impacts of agricultural Best Management Practices (BMPs). A
comprehensive workplan developed in September 1987 proposed a two-phased approach
to address Watershed Model deficiencies. Phase I was designed to improve the
nonpoint loading representation, refine/re-evaluate the data input to the Model,
and perform a preliminary re-calibration to available water quality data for the
1984-85 period. This period was used as the basis for the 40Z nutrient loading
reduction goal defined by the Chesapeake Bay Agreement.
Phase II was designed to focus on a better representation of the effects of
agricultural BMPs through application of the detailed Agrichemical Modules
(AGCHEM) of HSPF which allow a more deterministic, process-oriented approach to.
BMP analysis and evaluation. In addition, HSPF'code enhancements were performed
to allow specif ic consideration of sediment-nutrient and bed interactions within
the stream channel so that runoff, and subsequent:'delivery to. the Bay, of
dissolved and sorted nutrient forms can be more accurately modeled. The Phase
II Watershed Model was then applied to the Bay drainage area to include sediment
erosion, sediment transport, and associated nutrients, in addition to the current
modeled water quality constituents, to provide input to the Bay model.
This report presents' the results of the combined Phase I/Phase II effort. The
Phase II effort, tn addition to the agricultural and sediment-nutrient emphasis
noted above, also addressed the preliminary conclusions and recommendations
resulting from the Phase I work as discussed in the Phase I report (Donigian et
al, 1990). In summary, the .Phase I effort Identified the need for a longer
simulation period (i.e. more observed data), more in-depth analysis of animal
waste and point source contributions, .and simulation problems with selected
nutrient forms and phytoplankton as issues that needed to be addressed in Phase
II. Thus, the results presented herein represent the integrated efforts of both
phases of the Watershed Model refinement and re-calibration work. However, the
detailed results of Phase I are not repeated in this report since they have been
superseded by Phase II and were presented in the Phase I report.
The Watershed Model represents the entire drainage area to the Chesapeake Bay as
a series of land segments each with relatively uniform climatic and soils
-------
conditions. Within each model segment a variety of different land use categories
are each modeled with its own parameter values, and each land use provides
surface and subsurface nonpoint loadings to the stream draining that model
segment. Each model segment also corresponds to a single channel reach which are
linked sequentially with other channel reaches in other segments to represent the
major and minor river systems that comprise the Bay drainage. Figure 1.1 shows
the primary subbasins for the areas above the Fall Line, which include the three
major subbasins - Susquehanna, Potomac, James - and a number of smaller river
systems. Figures 1.2 and 1.3 show the model segments for the Above Fall Line
(AFL) and Below Fall Line (BFL) areas, respectively; Figure 1.2 also shows the
primary model calibration stations. A complete list of the model segments is
included in Table 1.1. This brief overview is provided to establish a common
basis for the following sections on the Phase I/II efforts and associated
conclusions and recommendations. The model segmentation is discussed further in
Sections 4.0 and 5.0V
It should be noted that the entire Phase I/II project has been a joint effort
between AQUA TERRA Consultants and the Chesapeake Bay Program Office (CBPO),
including the support of Computer Sciences Corporation, in Annapolis, Maryland.
AQUA TERRA was responsible for overall guidance and review of the modeling
effort, with the lead on specific tasks such as meteorologic data update
procedures, snowmelt/hydrology calibration, nonpoint parameter evaluation, AGCHEH
application, water quality Calibration guidance and review, and filial model input
sequence development. CBPO has taken the lead on database update for
meteorologic data, diversions, point sources, and observed water quality data;
land use re-evaluation; and urban nonpoint source loading assessment. The final
water quality calibration was a joint effort with AQUA TERRA preparing final
input sequences for all basins, developing a preliminary calibration for the
Susquehanna basin, and providing essentially final calibrations for the remaining
basins and below Fall' Line areas. CBPO staff continued calibration -of the
Susquehanna, performed selected refinements on the other basins, executed the
final runs, and prepared'the statistical analyses and appendices of the final
results. Through this joint effort CBPO staff have become thoroughly trained in
the application and many of the intricacies of the Watershed Model.
1.2 OVERVIEW OF THE PHASE I/PHASE II EFFORTS
The primary tasks in the Phase I project involved selected model refinements,
updating model input, and preliminary hydrology and water quality re-calibration;
these primary tasks and major sub tasks are listed below:
WATERSHED MODEL REFINEMENTS
Incorporate snowmelt simulation / '
Allow variable hydrology for each land use category
Include animal waste contributions
-------
Chesapeake Bay Watershed
and. Ma j or A'F L S ub -b-a s i n s
Figure 1.1 Chesapeake Bay watershed and major above fall line
-------
A F L Mo'de-1' Segments and
C a 1 i b r a ti on Static ns
Figure 1.2 Above fall line watershed model segments.
4
-------
Baltimore
Harbor
Severn
Bohemia
Anacostla
Chickahominy
Nansemohd
Elizabeth
Figure 1.3 Below fall line watershed model segments.
-------
TABLE 1.1 SUBBASIN, SEGMENT NAME, AND SEGMENT NUMBERS OF THE CHESAPEAKE BAY
WATERSHED MODEL
SUBBASIN
ABOVE FALL LINE
SUSQUEHANNA
POTOMAC
RAPPAHANNOCK
YORK
JAMES
APPOMATTOX
PATUXENT
SEGMENT NAME / GROUP
East Branch Susquehanna (EBSUS)
Vest Branch Susquehanna (VBSUS)
Juniata (JUNIATA)
Lower Susquehanna (LOSUS)
Conowingo (CONOW)
Upper Potomac (UPPOT)
Shenandoah (SHENA)
Lower Potomac (LOPOT)
Rappahannock (RAPPA)
Mattaponi (MATTA)
Pamunkey (PAMUN)
James (JAMES)
Appomattox (APPOM)
Patuxent (PATUX)
SEGMENT NUMBERS
10, 20, 30, 40
50, 60, 70
90, 100
80, 110
120, 140
160, 170, 175
190, 200
180, 210, 220
230
235. 240
250, 260
265, 270, 280, 290
300, 310
330, 340
UPPER EASTERN SHORE
BFL_1A
1
BFL_1B
LOVER EASTERN SHORE
BFL_2
VEST CHESAPEAKE
BFL_3A
BFL_3B
^
Coast_l
Bohemia
Chester
Wye
Chop tank
Nanticoke
Wicomico
Pocomoke
Coast_4
Coast_ll
Coast_6
Gunpowder
Baltimore
Patapsco
360
370
380
390
400
410
420
430
440
450
460
470
480
490
-------
TABLE.1.1 (Continued)
SUBBASIN
BELOW FALL LINE
SEGMENT NAME / GROUP
SEGMENT NUMBERS
PATUXENT/MID CHESAPEAKE
BFL_4 Patuxent
Severn
Coast 5
POTOMAC
BFL 5
RAPPAHANNOCK
BFL 6 ,
JAMES
BFL_7A
BFL 7B
Potomac
Anacostia
Occoquan
Rappahannock
Coast_8
Great Uicimico
York
James
Chickahomlny
Nansemond
Elizabeth
Coast 9
500
510
520
530
540
550
560
570
'580
590
600
610
620
630
640
-------
UPDATE MODEL INPUT
Update meteorology and precipitation to 1984-85
Update diversions, point sources, and observed
streamflow/water quality data to 1984-85
t Re-evaluate land use for 1974-78 and update to 1984-85
WATERSHED MODEL RE-CALIBRATION
Re-evaluate nonpoint parameters and loadings
Re-calibrate Watershed Model to 1984-85 with model
refinements and updated input
The Watershed Model refinements in Phase I focused on areas that had been noted
as specific deficiencies in the earlier applications. Snow accumulation and melt
had not been included in the first Watershed Model formulation, and thus was not
included in the HSPF version resulting from the HSP-NPS model conversion study
(Donigian et al, 1986). Since snow accumulation is significant in much of the
northern portions of the watershed, simulating snow processes was considered an
important element in year-round simulations produced by the Watershed Model.
Also, the converted Watershed Model used the same hydrologic parameters for all
land uses within a specific hydrologic segment; this was a limitation of the NFS
model which was eliminated when the Watershed Model was converted to HSPF.
Thus, allowing a different set of hydrologic parameters for each land use was
considered important to representing the impacts of land use changes.
Furthermore, animal waste had not been specifically considered in earlier efforts
and contributions from this source needed to be evaluated.
A major portion'of the Phase I effort involved evaluating, refining, and updating
the required input data to the Watershed Model. This effort was severely
hampered by the lack of detailed documentation and loss of back-up information
for the original Watershed Model (NVPDC, 1983), especially for the diversions,
point sources, and meteorologic data. Also, the accuracy and validity of the
land use information used in the original model was questioned, leading to the
need to re-evaluate the land use for both the 1974-78 period and the 1984-85
period. Although not originally envisioned as part of Phase I, direct
atmospheric deposition of nitrogen was an added input to the model (for water
surfaces only) to allow for this source. To avoid future problems, such as those
experienced in this study in updating the model, special efforts were expended
to develop detailed documentation and back-up for all procedures and analyses
performed in this model development effort.
The Watershed Model re-calibration effort involved both hydrology and water
quality. The hydrology re-calibration was needed because of the inclusion of
snow simulation and the parameter adjustments and refinement for better land use
representation; addition of an animal waste land segment and changes to the land
use also required a re-assessment of the hydrology. The water quality re-
calibration involved re-evaluation of the nonpoint parameters (i.e. potency
factors, subsurface concentrations, accumulation/washoff rates) and the expected
loadings from each land use represented in the Model, and then adjustment of
these parameters and selected instream water quality parameters was performed as
' 8 .
-------
part of the calibration process. Although the model had been previously
calibrated to the 1974-78 period, differences in the point loads and the
additional effects of animal waste and atmospheric contributions required
considerable re-calibration for the 1984-85 period in Phase I. Problems
identified during the re-calibration were discussed in the Phase I report, and,
as noted above, helped to focus the Phase II effort.
The emphasis in the Phase II work was directed to more detailed modeling of
agricultural BMPs, addition of sediment erosion and instream transport modeling,
and incorporation of instream sediment-nutrient interactions in order to better
represent nutrient delivery to the Bay. This effort involved replacing the
empirical .potency factor/subsurface concentration approach with the process-
oriented Agrichemical modules in HSPF for the cropland areas of the watershed.
In addition, the instream module (RCHRES) of HSPF was enhanced to include
sediment-nutrient interactions, and the resulting enhanced version was re-
applied/re- calibrated to the drainage area. The original workplan for Phase II
identified the following major tasks and subtasks:
DETAILED EVALUATION OF AGRICULTURAL PRACTICES
Identify/inventory current practices/BHPs
Identify/evaluate field site data
Apply detailed agricultural models .
Adapt Watershed Model to reflect field site results
SEDIMENT/NUTRIENT INTERACTIONS IN HSPF RQUAL
Enhance HSPF RQUAL to include sediment/nutrient interactions
Test model enhancements on selected subwatersheds
APPLICAtlON/RECALIBRATION OF ENHANCED WATERSHED MODEL
Calibrate erosion and instream sediment transport
. Apply/calibrate enhanced RQUAL
Integrate Task A results into Watershed Model .
Apply/Calibrate Enhanced Watershed Model to 1984-87 data
1.3 PHASE I PRELIMINARY FINDINGS AND RECOMMENDATIONS
In this section we briefly summarize the primary conclusions and recommendations
resulting from the Phase I work, in terms of hydrology, nonpoint loadings, and
water quality simulation, in order to set the stage for describing the Phase II
efforts and final results. These findings and recommendations were discussed in
greater detail in the Phase I report.
1.3.1 Hydrology Conclusions and Recommendations
a. Overall,, the hydrology calibration results for 1974-78 were very good
to excellent. The agreement in mean annual flow volumes and flow
frequency results was excellent. Although some differences remained
in the daily timeseries flow comparison, the basic conclusion was
-------
that the model was a good representation of the observed data for
that time period.
b. The hydrology results for 1984-85 were not nearly as -close to
observed data values as in the earlier 1974-78 period. Both the daily
and monthly R2 values were considerably lower at most stations, and
the frequency curves generally showed greater deviations with lower
peak flows and higher base flows than observed.
c. The agreement could have been considerably improved for 1984r85 if a
complete re-calibration had been performed, instead of limiting our
. adjustments to the input evaporation. The Phase I recommendations
for additional efforts in Phase II included the following:
1. Extend the data base and simulation .through 1988 or 1989 to
provide a better basis for evaluating the model performance for
this latter period.
(
2. Snow depth data should be developed for the 1984-89 period to
check and evaluate the snow simulation, and allow for any
additional calibration if needed.
3. The Conowingo Reservoir simulation needs to be investigated and
re-evaluated because the rule curve needed to be adjusted during
the 1974-78 calibration.
4. The storm of November 1985 in the Upper Potomac was not well
simulated, in terms of peak flow, by the Watershed Model,
although the monthly volume was accurate. Further efforts should
be made to investigate the problems with'simulating this event
once the simulation period has been extended.
1.3.2 Phase I Nonpoint Loading Conclusions and Recommendations
a. The simulated annual nonpoint loading rates were within the range of
expected values, with some deviations such as the rates for P04 and
NH4 from forest, and PO* rates from pasture.
b. Comparing Conventional and Conservation tillage segments,
Conventional produced higher loading rates for all pollutants except
N03, .where Conservation is the higher rate.
c. The highest rates shown were for the manure segment, followed by
urban and/or conventional tillage, conservation tillage, pasture, and
forest. The order changed slightly for individual pollutants. The
manure segment loading rates need to be further evaluated to ensure
they provide a reasonable means of representing animal waste inputs.
d. Analysis of Total Segment Loads indicate that agricultural cropland
(i.e. both conventional and conservation) tended to be the highest
loader of both Total N and Total P; however, the percent
contributions varied .from segment to segment, and other sources
-
10
-------
dominated in selected segments. Both urban and animal waste
contributions were significant, generally varying between 10X and 30Z
of TN and/or TP.
e. Point sources are still a major source of TP, often in the range of
202 to 302 or more, and a moderate source of TN. The assumptions in
developing the point source loads should be re-assessed to insure
their representation is accurate.
f. The atmospheric contribution tends to be only a few percent or less
for most areas, since only direct, deposition to water surfaces is
. considered. Atmospheric deposition to land surfaces is assumed to be
implicitly included in the nonpoint calibration.
g. It should^be noted that these results from Phase I were based on only
two years of simulation. Loading rates from nonpoint sources are
. notoriously variable from year to year. Thus, the recommendation to
extend the simulation in Phase II through 1988 or 1989 (as discussed
above for hydrology) is also supported by the need to better define
the simulated range and mean of the nonpoint loading rates.
1.3.3 Phase I Water Quality Simulation Conclusions and Recommendations
a. Water temperature and DO simulation was generally very good to
excellent throughout the Bay drainage.
b. The majority of the nutrient concentrations were simulated quite
well, reflecting the general range of the observed data values,
especially for Total N and Total P at the major Fall Line stations.
'...".'. ' . ' ' '
c. Larger deviations were seen'in the individual nutrient forms, such as
PO* in the Upper Susquehanna and the James rivers. The focus of
Phase II, on sediment simulation and sediment-nutrient interactions,
should help to improve the phosphate simulation and should be further
investigated.
d. The Potomac simulation results were significantly better than the
results for the Susquehanna and James, especially for the phosphorus.
components and Chi a. The Potomac data is more extensive than in the
other basins, allowing a better representation of the expected
variability in the observations. .
e. The river systems show considerable variability within the Bay
drainage. Differences in P0t, Chi a, and other parameter between
regions and individual subbasins need to be further investigated as
part of Phase II. .
f. The storm of November 1985 in the Upper Potomac was also under-
simulated in terms of particulate associated nutrients. In Phase II,
the water quality simulation should be further' investigated
especially in terms of the nonpoint loadings contributions.
11
-------
g. Phase II should also focus on some of the smaller subbasins (e.g.
Patuxent, Appomattox) where further investigation and calibration is
needed to present a consistent level of simulation throughout the Bay
drainage.
h. For Total N and Total P loads at the Fall Line, the model and
regression annual loads generally agreed within about 20X; problems
with Organic P in 1985 (partly due to the November 1985 storm) lead
to greater differences for that year.
i. Nitrate showed the closest agreement between the two load estimates,
at least for the three major Fall Line stations. Interestingly,
nitrate dominates the Total N and all other nutrient loads in terns
of total mass, especially for the Susquehanna, Rappahannock, and the
Patuxent,^and to a lesser extent, the Potomac. In the James, Organic
N dominates the Total N load and is significantly greater than the
nitrate load.
j. The differences between the simulated and regression loads were
generally much greater for the minor tributaries for most nutrient
forms, than for the major Above Fall Line basins.
k. Monthly load timeseries showing the simulated and regression loads
(plus Standard Error Bounds) indicate that the model estimates,
especially for TO and TP, generally fell within the envelope defined
by the two error bounds,
1. Simulation of Total Organic Carbon (TOO) should be considered in
Phase II, either to replace BOD or supplement it.
1.4 PHASE II FINDINGS AND RECOMMENDATIONS
The general conclusions of the Phase II -effort are discussed below in terms of
hydrology, nonpoint loadings and AGCHEM simulation, and water quality simulation:.
1.4.1 Phase II Hydrology Conclusions and Recommendations
The Phase II hydrology calibration results indicate the following:
a. The agreement between mean annual flow volumes is excellent;,
simulated values are within 62 of observed values. The year-to-year
variations shown in the appendix are greater but they still indicate
good agreement. ,
b. The daily R2 values for simulated and observed streamflow are
generally 0.70 or greater, and usually in the range of 0.75 to 0.84
for the three major basins. The monthly R2 values are consistently
in the range of 0.77 to 0.90 for all subbasins.
c. The daily R2 values for 1984-87 are consistently higher than the
corresponding values for 1974-78, possibly due to the greater number
. ' of rain gages used in Phase II.
.12
-------
d. Both the timeseries and flow frequency comparisons for all
calibration sites show good to very good agreement. Although
differences exist, the daily simulated timeseries and frequency
curves track the observed timeseries quite well for most all sites.
The largest deviations are consistently at the smaller subbasins
where more detailed spatial representation of precipitation may be
needed. Also, these differences are greater when reservoirs are
present in the smaller subbasins (e.g. Pamunkey, Patuxent);' further
study of the reservoir operating rules and finer spatial segmentation
may be needed.
i
e. Very little improvement was possible in Phase II for simulation of
the November 1985 storm in the Potomac. The storm was well simulated
in the Upper Potomac and Shenandoah model segments, but the recorded
rainfall amounts, intensities, and spatial distribution in.the Lower
Potomac could not sustain the extreme observed storm flows.
f. In Phase II, the Conowingo Reservoir hydraulic simulation was
investigated and adjusted to include a minimum summer release rate of
: 5000 cfs and minor adjustments to the weekly outflow coefficients
Both the daily timeseries and the frequency results show very good
agreement between simulated and observed values, with some deviation
in the low flow simulation since these flows are controlled entirely
by the operating procedures of the reservoir.
g. Seasonal R2 values were calculated by CBPO based on the following
seasons:
Season 1 January - February
Season 2 March - May
Season 3 June - September
Season 4 October - December
'"''
The results show that the winter (Season 1) is consistently simulated
worst than the other three seasons, and that the summer and fall
(Seasons 3 and 4) are usually the best, at least for the three major
basins. For most of the smaller basins, the spring (Season 2) is
often the best simulated.
Overall, the Phase II hydrology calibration results are a very good
representation of the Chesapeake Bay basin hydrology and provide a sound basis
for nonpoint load generation and delivery of pollutant loads to the Bay. Ongoing
updating of the meteorologic, diversion., and land use information is recommended
to allow for complete verification of the hydrology using the 1988-91 period.
1.4.2 Phase II Nonpoint Loading Conclusions and Recommendations
a. Generally, the simulated annual loading rates are within the range of
expected values, with some deviations. Annual rates for P04 from
forest and pasture, and NH4 from forest occasionally tend to be
toward the lower end of the defined range. Annual P04 rates from the
cropland areas are somewhat higher than the defined range.
13
-------
b. For non-cropland categories, the Total N and Total P simulated values
compare favorably with both the expected means and ranges.
c. For the cropland categories of Conventional Tillage, Conservation
Tillage, and Hay, the Total N and Total P simulated values are
generally close to the mean lumped cropland data; the Hay values are
generally less than the mean, while the tillage categories are
usually greater than the mean but well within the observed range.
d. Comparing Conventional and Conservation tillage segments,
Conventional produces higher loading rates for most model segments
for. all pollutants except N03, where Conservation is sometimes the
higher rate.
e. The highest rates for Total N and Total P are for the manure segment,
followed by conventional tillage, conservation tillage, urban, hay
land, pasture, and forest. The order changes slightly for
individual pollutants.
f. The manure segment loading rates are the most uncertain since there
is very little information on which to assess their validity.
g. For NH3 and P04 from cropland, the simulated ranges are generally 0.5
to 4.0 Ib/ac and 0.2 to 2.0 Ib/ac, respectively; these ranges are
generally higher than the limited observed data for these forms, but
they are not unrealistic based on the general literature.
h. Urban pervious - and impervious areas provide loadings which are
comparable to the Hay and Pasture categories, and for some nutrient
species (etg., NH3) the loadings are similar to .the tillage
categories. Thus, urban land can be a significant; source of total
nonpoint loadings in urbanized model segments.
1.4.2.1 AGCHEM Conclusions and Recommendations
1 . - - --
The AGCHEM application to the cropland areas of the Chesapeake Bay drainage has
satisfied the objectives outlined at the beginning of Phase II, namely to (1)
represent nutrient balances for the major cropland categories in each model
segment, (2) allow investigation of sensitivity to climate variations and
agricultural practices, and (3) provide a mechanism to project impacts of
nutrient management alternatives .for agricultural cropland. The results are
consistent with expected nutrient balances, observed ranges of runoff
concentrations, and expected ranges of nutrient loadings from cropland areas.
The AGCHEM approach to modeling the nutrient balances on agricultural cropland
is exactly the type of approach recommended by the Chesapeake Bay Nonpoint Source
Evaluation Panel in their report to the U.S. EPA and the CBP Executive Council.
In March 1990, the EPA Administrator convened the Panel to "assess the
effectiveness of current efforts to reduce nonpoint source loadings of nutrients
entering the Bay system". They reviewed and evaluated the effectiveness of a
wide range of nonpoint source programs, including program design, implementation,
budgets, modeling and assessment methods, and research efforts. One of the key
14
-------
recommendations of the Panel related to new approaches to nutrient control was
as follows:
"The Panel recommends that the Bay jurisdictions and the federal
agencies develop a mass balance accounting system, where nutrient
loadings are balanced by the nutrients removed from the system
plus those which are introduced and stored" (Chesapeake Bay
Nonpoint Source Evaluation Panel, 1990).
The AGCHEM application provides an initial assessment of this 'mass balance' for
cropland areas,- a tool for assisting in the 1991 Re-evaluation of the 40X
nutrient reduction goal of the Bay Agreement, and a framework for future efforts
on nutrient management based on the mass balance approach. Recommendations to
improve the use and application of AGCHEM by CBP include the following:
*
a. A detailed review and refinement of the current fertilizer and manure
application rates, and changes expected under nutrient management
alternatives should be initiated since these rates are the starting
point for nutrient reduction-from cropland.
b. Detailed site applications of AGCHEM, such as in the Patuxent,
Nomini, and Owl run watersheds, should be diligently pursued by
either CBLO and/or state agencies to further refine and improve the
simulated nutrient balances and regional differences.
c. Variations in soil nutrient storages, throughout the Bay drainage and
for different land uses, should be included in future model
/improvements. .Use of CIS capabilities and soils databases should be
used to improve spatial definition .and regional variations in model
parameters. ,
d. In line with the CBP NPS Evaluation Panel recommendations, future
Watershed Model refinements should include application of mass
balance and soil process modeling procedures for all land uses to
represent nutrient balances in all areas.
e. Recommended improvements to the AGCHEM module to facilitate Watershed
Model application and operation, and refinement of selected nutrient
processes include:
1. Capability to simulate NH3 volatilization from soils and
manure applications.
.
2. Modifications to simplify specification of nutrient
application rates and agricultural operations; more than one-
half of the current model input sequences are devoted to
defining this input for cropland areas.
3. Capability to specify wet and dry atmospheric deposition
rates for all nutrient forms.
15
-------
U. Refinement of soil phosphorus cycle and processes to allow a
better representation of sorption, fixation, and plant uptake
mechanisms. Additional research and algorithm development
may be needed to accurately model the soil phosphorus cycle.
1.4.3 Phase II Water Quality Simulation Conclusions and Recommendations
Review of the water quality calibration results indicates the following general
conclusions and recommendations, some of which were also presented in the Phase
I report:
a. Water temperature and DO simulation is generally very good to
excellent throughout the Bay drainage. In Phase I, water temperature
was calibrated at three instream sites within the basin, where
continuous water temperature data was available, and then
extrapolated to the other model segments. The simulation was
extended and refined in Phase II.
The DO simulation is generally very good for most calibration sites
but with some deviations at selected sites for low summer DO
observations. At most sites, the simulation is quite good except for
one or two extreme points whose observed values may be questionable.
Some of the low summer values may be due to benthal demands which are
not currently calculated in the model.- ,
b. The majority of the nutrient constituents are simulated quite well at
most sites, reflecting the general range of the observed data values,
especially for Total N and Total P at the major Fall Line stations,
with the .notable exception of the Susquehanna Basin (discussed
below). Peak simulated concentrations are often somewhat higher than
observed values, as would be expected when comparing a continuous
simulated timeseries with discrete, individual observations.
c. The Phase II results are a significant improvement over the Phase I
results due to the water quality parameter re-evaluation, the AGCHEM
nutrient loading simulations, and the additional calibration effort.
Although some problems remain, the results show improved simulation
of seasonal variations in nitrate and Chi a, closer representation of
individual nutrient species, and better agreement with most observed
concentrations and loads.
d. Large deviations are still seen in the individual nutrient forms at
some sites, such as P0t in portions of the Susquehanna and inorganic
nutrients in the minor tributaries. Chi a observations were limited
at many calibration sites, thus restricting the calibration process.
e. The Potomac simulation results are the best of all the major basins
and are generally a good to very good representation of nutrient
concentrations throughout the Bay basin. This could be partially due
to the fact that the Potomac data is more extensive than in any of
the other basins, allowing a better representation of the expected
variability in the observations and a better calibration. The total
16
-------
nutrient concentrations, and the individual species, are well
simulated throughout the entire period. Concentration peaks do
exceed individual observations, especially for NH3 and P04, but they
are generally within the range of the observed data.
f. The Susquehanna simulation also shows good agreement between
simulated and observed concentrations, but this was accomplished only
after intensive calibration effort of the entire basin, with
particular focus on the reservoir simulation. Generally, Total N and
component concentrations at Conowingo are somewhat under-simulated.
P04 and Total P concentrations are somewhat over-simulated throughout
the Susquehanna, possibly due to interactions with sediment and metal
oxides, and increased sorption under low pH conditions.
USGS data* for pH show extremely low values for two main stem stations
in the East and West Branch subbasins, with mean monthly values in
the 3.0 to 4.0 range. Detenbeck and Brezonik (1991) show greatly
increased P binding by lake sediments for pH in the range of 4.5 to
6.0, with greatly decreased diffusive P fluxes under these
conditions. Consequently, increased sorption and retention, and
possibly precipitation of P04 from the overlying water may be a
reason for most of the observed concentrations being close to
detection limits.
Clearly, further investigation is warranted and should focus on
instream processes in the Upper and Lower basins for both nitrogen
and phosphorus delivered to the Conowingo Reservoir, and processes
within the reservoir. This should be performed in conjunction with
: a thorough loadings analysis to confirm the relative loadings
contributions from all sources. Further calibration of the reservoir
water quality is heeded to improve the sediment and Chi a
simulations, which may also assist in improving the phosphorus
simulation. '
g. The, simulations for the Janes River are a significant improvement
over .the Phase I results. Although the concentration agreement is
not as good as in the Potomac, the Total H and Total P are
consistently within the range of the observations, and the component
nutrient species are reasonably well simulated. The data base for
the James was much more limited compared to the other sites, thus
limiting the calibration effort.
h. The river systems show considerable variability within the Bay
drainage, as noted above concerning the P04 simulation problems; in
the Susquehanna. In general, the smaller tributaries and those with
reservoirs are not simulated as well as the larger basins with free-
flowing reaches. When both these conditions occur, such as in the
Pamunkey and Appomattox basins, the differences between simulated and
observed values usually increase. In these cases, finer segmentation
and more investigation into reservoir operations is needed. For the
short-term, simulation differences in the smaller basins should have
17
-------
minimal impact on the Bay simulation since they are a small fraction
of the delivered nutrient load.
i. Although the Patuxe'nt River falls in the category of a small basin
with reservoirs, the simulation results are exceptionally good due, to
the wealth of observed data for calibration and the point source
monitoring that accurately defined the point source loads inx the
Watershed Model. The Total P simulation reflects the impact of the
phosphorus ban in December 1985, with the 1986-87 concentrations
being significantly lower than in 1984-85. The few instances where
simulated values are high are due primarily to under-simulation of
extreme low flows during late summer.
j. The storm of November 1985 in the Upper Potomac was under-simulated
in terms ,of peak flow, and many of the particulate associated
nutrients were also under-simulated for this event. In Phase II, the
HSPF code was modified to allow user-specified particulate NH4 and P04
concentrations for scouring of the bed sediments; however, much of
the scoured material is organic. Consequently, the particulate N and
P concentrations are under-simulated for this event in both the
Potomac and James basins, and since these components dominate during
a major scour event, Total N and P concentrations are also under-
simulated. . Future efforts should consider adding the capability to
allow user-specified bed concentrations for organic components.
k. Simulation of Total Organic Carbon (TOC) was included in Phase II
because TOC is a required input for the Bay model and observed TOC
values were more readily available at more sites than BOD. BOD is
still simulated by the Watershed Model because it is a required
parameter. .BOD is used in the current version as an indicator of
TOC, i.e. BOD nonpoint and point source Ipadings are converted to TOC
by a conversion factor which was used as a calibration factor to
adjust the TOC simulation. The simulation results for TOC are
generally quite good for most sites, and are similar in appearance to
the organic N and organic P simulations.
1. The sediment simulation results are generally adequate but could be
improved with further calibration. The model tends to over-simulate
large storms, and under-simulate small storms and concentrations
.during low flow periods. Additional sediment calibration should
concentrate on evaluation of the simulated sediment loadings and the
target values derived from the HRI data. Also, reservoir sediment
simulations should be further investigated with additional
calibration. Finer model segmentation .and spatial detail of both
land segments and stream reaches is probably needed for improved
sediment simulation.
The following general conclusions are derived from comparison of the simulated
Fall Line loads with regression estimates:
a. For Total N and Total P, the model and regression annual .loads
generally agree within about 251 to 302. Problems with Organic P
18
-------
and, to a lesser extent Organic N, in th Potomac in 1985 (partly due
to the November 1985 storm) lead to greater differences for that
year.
b. In the Patuxent and the Potomac, the closeness of the concentration
comparison is also reflected in the Fall Line load ratios. For both
these rivers, extensive data was available for developing the
regression and performing the model calibration.
c. Nitrate generally shows the closest agreement between the two
estimates, at least for the three major Fall Line stations.
Interestingly, nitrate dominates the Total N and all other nutrient
loads in terms of total mass, especially for the Susquehanna.
Rappahannock, and the Patuxent.
^
d. The differences between the two estimates are generally much greater
for the minor tributaries for most nutrient forms (with the exception
of the Patuxent), than for the major Above Fall Line basins. For the
minor tributaries and the James, significantly less data was
available for developing the regression. Consequently, the
uncertainty of the regression is greater for these basins.
1.5 GENERAL WATERSHED MODEL RECOMMENDATIONS
The Watershed Model provides a framework for quantifying and evaluating nutrient
loadings to the Bay, and a mechanism for assessing the impacts of land use
changes and alternate nutrient and agricultural management practices. However,
the watershed' modeling component of the Chesapeake Bay Program should be an
ongoing, evolving process with continuing data collection, model development and
testing, and calibration/validation efforts. The current Watershed Model,
although providing 'a sound basis for current decision-making, needs continuing
updating and refinement in a number of areas. As all planning tools, the
Watershed Model should not be static. Continuing efforts should be devoted to
the following areas:
a. Further calibration and validation with new data (e.g. 1988-90) with
specific focus on problems areas, such as animal -waste, reservoir
modeling, and P04 in the Susquehanna.
b. Further investigation and refinement of reservoir hydrology and water
quality simulation for model segments with major reservoirs.
c. Refinement and improvement in urban -and animal waste simulation
procedures and management alternatives.
d. 'Refinement of AGCHEM nutrient balances based on field site
applications, regional conditions, and seasonal variations.
e. Finer spatial segmentation and representation of land areas and
stream reaches to provide more detailed input to regional water
quality issues and problems.
H
19
-------
f. Direct linkage between the CBLO capabilities with geographic
information systems (CIS) and model parameter/input development.
This is a major focus of current development efforts in watershed
management and nonpoint source modeling. .
g. Extend nutrient balance concepts of AGCHEM to simulation of all land
uses to provide a consistent mass balance approach for all portions
of the Watershed Model.
h. Develop improved capabilities in the Watershed Model .for user
Interface, output analysis, and specification of nutrient
applications and agricultural management practices.
.20
-------
1.6 FORMAT OF THIS REPORT
Section 2 describes the procedures used to update the meteorologic and
precipitation data for the 1984-87 period, while Section 3 briefly describes the
land use and waste load development techniques. Section 4 describes the nonpoint
loading simulation procedures used for both cropland and non-cropland land
categories, including an overview of the AGCHEM module and the application
procedures used for cropland areas; nonpoint simulation results are also
presented in Section 4 for all land uses. Section 5 provides an overview of the
RCHRES module sections used in the Phase 11 effort, along with discussion of the
sediment-nutrient capabilities incorporated into the HSPF code. Sections 6 and
7, respectively discuss the hydrology and water quality results, with the primary
focus on the Phase 11 results. As noted above, due to the vast amount of data
and information, developed for and generated by the Phase 11 watershed modeling
work, the majority of the results and input related information are provided in
the Appendices.
21
-------
SECTION 2.0
METEOROLOGIC AND PRECIPITATION DATA
BASE DEVELOPMENT PROCEDURES
2.1 INTRODUCTION
This section documents the procedures that were used to update the regional
meteorologic data base and hourly precipitation data base for 1984-1985 and
subsequently, 1986-1987. The regional meteorologic data base consists of wind
speed, air temperature, dewpoint temperature, evaporation, cloud cover, and solar
radiation. The goal of this task was to produce a data base that was both valid
and, insofar as possible, consistent with the 1974-78 data with regard to the
observation stations, computation and disaggregation methods, and level of
detail. Because documentation of the existing (1974-78) regional data base was
hot available, this data base was exhaustively analyzed and compared to observed
and generated data in an attempt to determine data sources and computational
methods. Based on this analysis, observation stations were selected and
procedures were developed to update the data base for the 1984-85. Development
of the hourly precipitation data was accomplished using modified versions of the
Thiessen procedure, aggregation software, and observed data base used to develop
the 1974-78 data base. The same procedures were also followed in completing the
data base for the 1986-1987- period; however,: small changes in the network of
observed stations were necessitated by the discontinuance of several stations.
The final data base is stored in Watershed Data Management (WDM) System files.
WDM files are unformatted, direct access files that may contain several types of
data and many individual data sets of each type. The data sets are identified
by unique data set numbers and other attributes such as a name, descriptive
information, and time interval. WDM file format is the primary.data storage'
format for the HSPF program. In addition, several intermediate data formats were
utilized in developing the data base. These formats are ASCII, HSPF-readable,
hourly and daily formats that are identified in this document as HYDDAY, HYDSDY,
and HYDHR, respectively. The formats are described in Appendix D.I
2.2 REGIONAL METEOROLOGIC DATA
The 1974-78 regional met data was compared with published data for major stations
in the basin. As a result of these comparisons, the cloud cover and dewpoint
temperature data sets were found to exactly match the published data as shown in
Table 2.1 .
22
-------
TABLE 2.1 CORRESPONDENCE BETWEEN 1974-78 REGIONAL DATA
AND PUBLISHED DATA
REGION
1
2
3
4
5
6
7
CLOUD COVER
STATION
Binghamton, NY
Williamsport, PA
Harrisburg, PA
Roanoke, VA (1)
National Airport
Roanoke , VA
Richmond, VA
DEWPOINT TEMPERATURE
STATION
Binghamton, NY
Binghamton, NY (2)
Harrisburg, PA
Roanoke, VA
National Airport
Roanoke, VA
Richmond, VA
(1) Region 4 cloud cover data are computed from Roanoke
data and monthly factors.
(2) Region 2 dewpoint temperatures for March-November
are the Binghamton data multiplied by 1.1.
The other regional met data sets (wind speed, air temperature, potential
evaporation, and solar radiation) did not match these stations, but were found
to exhibit other relationships. Generally, these relationships involved constant
monthly factors between the regions. Since the methodology used to develop these
factors is unknown, it was decided to base the regional data entirely upon
observed data from representative stations in or near each of the regions. The
stations selected were those that were used in the 1974-78 data base for dewpoint
and cloud cover, with two exceptions. The weather stations at Elkins, WV and
Dulles Airport were chosen as more representative of Regions 4 and 5,
respectively, than Roanoke, VA and National Airport. Table 2.2 contains the
stations used to develop the 1984-87 regional data base, and Figure 2.1 shows the
seven meteorologic regions and the locations of their corresponding stations.
TABLE 2.2 STATIONS USED TO DEVELOP 1984-87 REGIONAL
METEOROLOGIC DATA
REGION
STATION
STATION NUMBER
1
2
3
4
5
6
7
Binghamton, NY
Williamsport, PA
Harrisburg, PA
Elkins, WV
Dulles Airport
Roanoke, VA
Richmond , VA
300687
369728
363699
462718
448903
447285
447201
23
-------
Principal Meteorologlc Station
Figure 2.1 Meteorologic regions and principal stations in
the Chesapeake Bay watershed model.
24
-------
2.2.1 Cloud Cover and Dewpoint Temperature
The Watershed Model utilizes daily values of cloud cover and dewpoint
temperature. The daily records for the stations listed in Table 2.2 were down-
loaded from the NOAA data tape, reformatted, and transferred into the WDM files
without any modification of the data.
2.2.2 Air Temperature and Wind Speed
The Basin model utilizes hourly values of wind speed (miles/hour) and air
temperature (deg F). The daily records for the stations listed in Table 2.2 were
1) downloaded from NOAA data tapes, 2) reformatted, 3) distributed to hourly
values using customized software, and 4) transferred into the WDM files.
The daily max-min air temperatures were converted to hourly values based on a
smooth variation with the daily minimum at 6 AM and daily maximum at 4 PM.. This
variation is illustrated in Figure 2.2. The program TMPDIS, which is included
in Appendix D.2, was used to perform this distribution. TMPDIS reads a daily
data file containing maximum and minimum temperatures in HYDSDY format and the
daily observation hour (1 to 24). . The output hourly air temperature data are
written to the output file in the HYDHR format,
The daily wind movement (miles/day) data were distributed to hourly Values using
a slight diurnal variation having a minimum value from midnight to 7 AM and a
maximum at 3 PM. Figure 2.3 illustrates this variation. The program WNDDIS
(Appendix D.3) was used to perform the distribution. This program reads the
total daily wind movement (miles per day) from an input file, using the format
HYDDAY, and outputs the hourly wind speed (miles per hour) to the output file in
the HYDHR format.
2.2.3 Solar Radiation
Analysis of the solar radiation data and the corresponding cloud 'cover time
series indicates that the 1974-78 radiation used by the Watershed Model was
computed using a model based on latitude and cloud cover. Such radiation models
generally use curves of clear sky radiation as a function of latitude and day of
the year, along with an empirically derived correction factor based on cloud
cover or percent sunshine. Since measured solar radiation data are no longer
readily available, it was decided to generate the 1984-87 radiation data using
a model.
Two simple solar radiation models were compared to the 1974-78 clear sky data in
an attempt to match the 1974-78 regional data. The model that more closely
matched the 1974-78 data was the RADIATION generation utility contained in the
HSP QUALITY program (Hydrocomp, 1977). Using approximate area-weighted latitudes
for each of the regions, this program generated curves that were nearly identical
to the 1974-78 clear sky data. Small differences (1 to 41) occurred in the
spring and .autumn months depending on the latitude used in the computation.
While neither the clear sky radiation nor the cloud cover correction factor were
reproduced exactly, the differences in generated radiation are negligible.
25
-------
60
60
40
Temperature (F)
0 ซ 18 24
Time of Day (hours)
Figure 2.2 Diurnal variation of air temperature.
16
Wind Speed (miles/hour)
14
10
06 12 W 24
Time of Day (houra)
Figure 2.3 Diurnal variation of wind speed.
'26
-------
A FORTRAN subroutine (RADCLC) was developed to implement this model. The
algorithm is based on empirical curves developed by Hamon et al., (1954), and
documented in the HSP II Manual (Hydrocomp, 1978)). Also, since the daily
radiation must be distributed to hourly values for input to HSPF, the -subroutine
RADDST was adapted from the HSP QUALITY program to perform this distribution.
This routine generates hourly radiation values using the daily total, the month,
and the day. These two routines were implemented in a program that reads the
daily cloud cover data and the latitude for a region, and generates the daily and
hourly solar radiation for each day of the year. Listings of the subroutines are
included in Appendix D.4 along with a table of inputs and outputs of the
-routines.
The area-weighted latitudes for the regions are based on the following model
segment groupings:
^
Region Segments Latitude
1 10 20 30 41.75
2 40 50 60 70 90 100 40.9
3 80 110 120 140 40.8
4 160 170 175 190 200 39.25
5 180210220230330340 39.75
6 265 270 39.1
7 .235 240 250 260 280 290 300 310 37.8
2.2.4 Potential Evaporation
Daily and monthly totals of potential evaporation from the regional data sets
were compared to observed data from the pan evaporation stations located in the
Chesapeake Bay watershed. No definitive relationships were observed between the
regional and recorded data in these comparisons. However, comparisons between
the individual regional data sets showed that the .daily values are related to
each other by monthly factors. Since the evaporation data were not derived from
observed pan data, but are apparently related .to each other, it was hypothesized
that the data for one region were computed using an evaporation model; and then
the other regions were obtained from the computed time series and long term
relationships between the regions. This hypothesis was reinforced by comparison
of. the Region 5 data to daily pan evaporation data that was estimated using the
Penman equation. Dewpoint, air temperature, wind speed, and solar radiation data
from Region 5 were used as parameters in the Penman,model. When these estimated
values were compared with the Region 5 data set, it was found that the daily
values were related by. constant monthly factors for the years 1974-76.
The Penman (1948) method is a simple, commonly-used algorithm for estimation of
pan evaporation. The method, as described by Kohler et al. (1955), uses daily
inputs of solar radiation (Langleys), wind speed (miles/day), dewpoint (F), and
average air temperature (F). The average air temperature is defined here as the
mean of the maximum and minimum daily values. The output from the model is daily
pan evaporation in inches. This algorithm was implemented in a FORTRAN
subroutine entitled PANEVP that reads the appropriate input data files, and
outputs the daily data in the HYDDAY format. The PANEVP subroutine is listed in
Appendix D.5. .
27
-------
Development of Che regional potential evapotranspiration data sets was
accomplished using the following two-step procedure:
1. Regional daily pan evaporation totals were computed using the Penman
method as implemented in the subroutine PANEVP.
2. Factors were developed to account for the pan coefficient and apparent
overprediction by the Penman method during winter months. These
factors were applied to the Penman data to obtain the final potential
evapotranspiration. .
2.3 HOURLY PRECIPITATION DATA
Significant effort was devoted to development of the hourly rainfall data for the
watershed model because this data is the principal driving force for the model
and because of the quantity of data involved. This task Included three principal
subtasks. First, a large amount of observed hourly ..and daily data was
reformatted from NOAA tapes to WDM format using the ANNIE program and associated
utility software. Second, an iterative procedure, using CIS software (ARC-INFO),
coordinate locations of the observed stations, and digitized model segments was
used to develop the precipitation segmentation and weighting factors needed to
combine the data. Third, the observed data were combined and distributed to
hourly values utilizing software adapted from the procedures used to generate the
1974-78 data base. The resulting data base consists of 32 data sets of generated
hourly rainfall corresponding to individual model segments or, in some cases,
several aggregated model segments.
2.3.1 Development of'Observed Rainfall Data Base
This task involved reformatting the 1984-85 and 1986-1987 daily and hourly data
for all rainfall stations located within or near the basin to WDM format (Figure
2.4 and Figure 2.5). This data was obtained from standard NOAA tape files for
each of the states in the basin, i.e., NY, PA, W, DE, MD, and VA. The ANNIE
program and several additional stand-alone utility programs were employed in this
effort, which involved reading and obtaining an inventory of the NOAA tape files,
selecting stations to extract based on coordinate locations, breaking the large
data files into smaller files, and transferring the data to a WDM file. Details
of the procedures and software are documented in Appendix D.6.
2.3.2 Basin Segmentation and Development of Weighting Factors
A Thiessen analysis was performed to determine the weighting factors to use in
the rainfall generation. The stations and Thiessen weights from the original
1974-78 rainfall generation were reviewed, and it was decided to redo the
analysis since several of the original stations were no longer available, and the
model areas and segments associated with several of the generated rainfall
records were not documented. Also, it was likely that the availability of
additional observed stations would improve the analysis.
The final 1984-85 Thiessen analysis was based on the digitized model segment
boundaries and the coordinate locations (latitude and longitude) of a selected
28 .
-------
A.F L Daily and Hourly
Precipitation Stations
HOURLY PRECIPITATION STATIONS
DAILY PRECIPITATION STATIONS
Figure 2.4 AFL daily and hourly precipitation stations..
28.1
-------
BFl Dai ly and Hourly
Precipitation Stations
HOURLY PRECIPITATION STATIONS
DAILY PRECIPITATION STATIONS
Figure 2.5 BFL daily and hourly precipitation stations.
'28.2
-------
subset of the observed stations. It was performed using the ARC-INFO CIS
software. The outputs from the procedure were the observed stations and the
corresponding area-based weights for each model segment. The procedure required
a sequence of steps as described in Appendix D.7. The resulting Thiessen network
consisted of 32 separate rainfall segments and utilized a total of 173 individual
stations.
When the 1986-87 data were being generated, it was discovered that several
stations included in the 1984-85 Thiessen network had been discontinued or had
significant missing data. Therefore, these stations were removed from the
network and their Thiessen weightings were added to the nearest station in the
set of stations for that segment. The differences between the 1984-85 and 1986-
87 Thiessen networks are documented in Appendix D.7. Generally, these
differences are very small.
^ -
2.3.3 Generation of Weighted Hourly Rainfall Records
The final step in creating the hourly precipitation was to run the PRECIP program
to generate the data using the Thiessen network and the data base of observed
hourly and daily stations. The PRECIP program is a FORTRAN program designed to
generate an hourly time series of rainfall that is a weighted average of several
observed time series. For each execution of the program, which generates one
hourly time series, the following inputs are required:
Between one and ten observed time series of hourly rainfall
Between zero and ten observed time series of daily rainfall
Weighting factor for each of the above time series (sum of all factors
must be 1.0)
The program then proceeds daily through the following steps to produce an hourly
time series: . - .
1. Computes the daily totals for all input (observed) time series (daily
and hourly). (Data are obtained from the WDM file.)
2. Computes the weighted daily total rainfall using the weighting factors
and the daily totals computed in Step 1. (Note that any station
having missing data during the day is eliminated from the computations
on that day, and the weighting factor for that station is distributed
among the remaining stations in proportion to their original weights.)
3. Compares the weighted daily total with each of the daily totals of the
observed hourly time series. Selects the observed hourly station
having the closest daily total to the weighted daily total.
4. Computes the generated hourly time series by distributing the weighted
daily total using the hourly distribution of the station selected in
Step 3. .
5. Stores the resulting hourly time series in the WDM file.
.29
-------
6. Writes a summary of the data generation for each day that data were
generated. If the program experiences difficulty because of too much
missing data, it also writes an explanatory message to the output
file, and skips the current day. .'
During generation of several of the rainfall records, the PRECIP program skipped
a number of days on which precipitation occurred because none of the hourly
stations had data on that day. When the total quantity of skipped precipitation
was judged to be non-negligible, the skipped rainfall was distributed manually
over the day, and the hourly values were inserted into the WDM file using ANNIE.
Additional documentation of the PRECIP program, including operational details,
input formats and a program listing, may be found in Appendix D.8.
30
-------
SECTION 3.0
LAND USE DATA AND WASTE LOAD INPUTS
3.1 LAND USE DATA
This section documents the methods used to provide a 1985 base year land use data
set for use in theปWatershed Model. The 1985 base year was chosen to be
consistent with the 1987 Bay Agreement, and because it was a recent year that had
sufficient land use information coverage from the different sources used. The
Phase II Watershed Model land uses are forest, conventional till cropland,
conservation till cropland, cropland in hay, pasture land, animal waste areas,
and water areas. The hay cropland was defined as a separate category in Phase
II and extracted from the total cropland acreages used in Phase I.
There are several methodologies, developed at the state level, for the assessment
of land use. These methodologies focus on .different planning and .assessment
issues and at differing levels of detail. As a consequence, their methods, land
use definitions, and base years vary from state to state. Some excellent
examples of state level land use assessments include Maryland's Geographic
Information System (MAGI) , Virginia's Geographic Information System (VIRGIS) , and
Pennsylvania's detailed County Level National Resources Inventory (NRI) data
base.
Another methodology with partial basin coverage is the USGS Land Use and Land
Cover System (USGS LU/LC).. The USGS LU/LC has consistent coverage for the basin
for the 1972-73 period, except for the basin area within Hew York. Other basin-
wide land use assessments include the U.S. Census Bureau which has surveyed
agricultural land. use since the 1800's. The U.S. Forest Service has been
conducting state forestry surveys since the 1940's. More recently, the U.S. Soil
Conservation Service has conducted the National Resources Inventory, a data base
of land use and geographic characteristics, to the extent possible, consistent
with the objective of obtaining a basin-wide methodology, all of the above land
use methodologies have been utilized in the development of the 1978 and 1985 data
sets. ' '
A consistent methodology of determining land use for the entire Bay basin was
developed which obtained particularly detailed information on agricultural
cropland. The principle sources of information,used provided data on land use
at a county level throughout the basin. Principle sources were the U.S. Census
Bureau, the U.S Forestry Service, and the U.S. Department of Agriculture. Also
used to advantage were the U.S. Geological Survey LU/LC data for areas of water
(rivers, lakes, and reservoirs) and urban land.
31
-------
3.1.1 Land Use for 1978
The first step in the creation of the 1985 data base was to review the 1978 data
set used by the early versions of the model. The refined 1978 land use data set
provided a base for the update to 1985 land use.
Refinement of the preliminary 1978 land use data was achieved through utilizing
several existing sources. These sources include the series, Census of
Agriculture for 1978 and Census of Agriculture for 1982. Volume 1, Geographic
Area Series (U.S. Census Bureau, 1980, 1984) published for each state the by the
U. S. Department of Commerce, .Bureau of the Census. The 1978 Census of
Agriculture provided consistent, reliable data on land areas of cropland, and
pasture. Combining the major cropland categories from the 1978 Census of
Agriculture produced^cropland acreage figures by county. Major categories of
pasture land were also aggregated to obtain the total pasture area. The 1982
Census of Agriculture was used to determine the total land area of each county
in the basin. .
Woodland area was derived from USDA Forest Service Timber Surveys (U.S. Forest
Service, 1974, 1978, 1978, 198Q, and 1982)- The latest available publications
of the Timber Surveys were used to obtain woodland area by county. Total "other
land" found in the Ag Census was aggregated with the woodland area to create a
land use defined as natural woodlots, planted woodlots, timber tracts, cutover
land, and land not harvested, grazed, or used for buildings or roads.
Neither the Census of Agriculture nor the Forestry Surveys quantify urban land
use. Urban land area for the preliminary 1978 land use data was determined by
subtracting the cropland, pasture, and woodland acreage from the total acreage
of each county. Any negative differences resulted in the assumption of zero
urban land area. Negative differences were then reconciled to the total county
land area by adjustment of the woodland acreage.
The resulting preliminary 1978 land use data set was then sent to the State
Offices of the USDA, Soil Conservation Service for revision and/or confirmation.
All states made revisions. Each state used its own methods to revise the county
data set. Among the methods used were work load analysis studies, National
Resources Inventory data, state planning agencies, and local county field offices
of the Soil Conservation Service. The 1978 data set after the State revisions,
is presented in Appendix E.I. This data set has four categories of land use on
a county basis, cropland, pasture land, woodland and urban land.
j -
Assisting with review and corrections to the refined 1978 land use data base
were:
David Benner, Asst. State Conservationist, Delaware
Jeffery Loser, State Resource Conservationist, Maryland
Phillip Nelson, Asst. State Conservationist, New. York .
Robert Heidecker, State Resource Conservationist, Pennsylvania
Willis Miller, Asst. State Conservationist, Virginia '
Bruce Julian, State Resource Conservationist, Virginia
Ken Carter, Soil Conservationist, Virginia
Dixie Shreve, State Resource Conservationist, West Virginia
Ernest L. Moody, District Conservationist, District of Columbia
32
-------
3.1.2 Land Use for 1985
To obtain the 1985 land use data, the 1978 data were further processed. The 1978
data were sent to the State Offices of the Soil Conservation Service for updating
to 1985. The state contacts reviewed and updated the land use to 1985 as shown
in Appendix E.2. Pennsylvania land use data were updated by the Pennsylvania
Department of Environmental Resources.
Assisting with review and corrections to the 1985 land use data base were:
David Benner, Asst. State Conservationist, Delaware
Jeffery Loser, State Resource Conservationist, Maryland
Phillip Nelson, Asst. State Conservationist, New York
Robert Heidecker, State Resource Conservationist, Pennsylvania
Timothy Murphy, Hydraulic Engineer,. Pennsylvania
Bruce Julian, State Resource Conservationist, Virginia
Dixie Shreve, State Resource Conservationist, West Virginia
Ernest L. Moody, District Conservationist, District of Columbia
The 1985 land use data were further refined as described below.
3.1.3 Cropland Tillage
The Watershed Model has three categories of cropland; conventional tillage,
conservation tillage, and hay land. Conventional tillage represents fall plowed
and/or spring plowed conventional tilled cropland. Conservation tillage
represents those tillage practices that result in a residue cover of at least 30Z
at the time of planting. .
Tillage information on a county level for the 1985 data base was obtained from
the Conservation Technology Information Center (CTIC), West Lafayette, Indiana.
The CTIC is a clearinghouse for information on soil conservation and, in
particular, cropland tillage practices. The CTIC conducts an annual survey by
county of acres of crops grown under different tillage systems. CTIC data for
1985 was the basis for the conventional and conservation cropland distribution
in the 1985 land use data set (CTIC, 1986).
The CTIC tillage data were processed to eliminate area overestimation due to
double cropping. Percent of cropland under conservation tillage was calculated
for each county (Appendix E.3).
Hay acres were compiled from the 1982 Census of Agriculture from the category of
harvested, "hay, alfalfa, and other tame, small grain, wild, grass silage, or
green chop*. The hay acres were transformed to a percentage of the Census of
Agriculture total harvested crop acres, and the area of hay acres was determined
by this proportion applied to total model cropland area.
3.1.4 Water Acres
Water acres are defined as the area in rivers, creeks, streams, canals, lakes,
and reservoirs. Only nontidal waters of the basin are considered to be in this
land use category. Tidal waters are included in the simulation of the 3-D Water
Quality Model.
33
-------
The water land use area was obtained from the CBPO Geographical Information
System (CIS), USGS Land Use/Land Cover (LU/LC) compiled from NASA high-altitude
aerial photographs and National High-Altitude Photography Program (NHAPP)
photographs. The CIS was used to directly determine the acres of water in each
segment, with the exception of the area of New York state which was not
available. Data from the National Resources Inventory (NRI) were used to
determine acres of water for New York.
3.1.5 Manure Acres
Manure acres is a derived land use which represents the production of nutrients
from manure produced in a segment. These acres do not represent acres of
concentrated animals, nor do they represent manure piles or manure stacking
facilities, rather the manure acres are used to represent the aggregate of all
of these activities.'
Animal units in each model segment were estimated from the livestock numbers in
the 1982 Census of Agriculture. An animal unit is defined as 1000 pounds of
animal weight. For the purpose of this analysis, this corresponds to 0.71 dairy
cows, 1 beef cow, 5 swine, 250 poultry layers, 500 poultry broilers, or 100
turkeys. Table 3.1, modified from Animal Waste utilization on Cropland and
Pastureland (U.S. EPA, 1979) tabulates the voided manure of different animal
types. The tabulated wet weight of manure does not include the wet weight of
bedding material, spilled food, or soil.
i ... ' -
TABLE 3.1 ESTIMATED QUANTITIES OF VOIDED MANURE FROM LIVESTOCK AND POULTRY.
NORMALIZED TO l.OOQ POUNDS BODY WEIGHT. FROM, ANIMAL WASTE
UTILIZATION ON CROPLAND AND PASTURELAND (U.S. EPA. 1979)
Animal
Type
Dairy
Beef
Swine
Poultry
Poultry
Turkeys
Animals/
Animal Unit
0.71
1.0
5.0
layers 250
broilers 500
100
Tons of
. Wet Manure
Voided
. 14.9
6.7
11.7
9.7
13.1
10.2
P (Ibs)
21
18
37
100
110
84
N (Ibs)
123
61
160
235
390
304
COD (Ibs)
3,340
1,510
2,080
4,353
5,915
4,599
Animal units of poultry, swine, beef, and dairy were adjusted to account for the
predominant manure handling practices. Poultry are usually raised in houses.
The houses are cleaned out completely once a year and the manure is spread on
fields. As the spread manure is already accounted for in the land uses of
-------
conventional cropland, conservation cropland, and hayland, manure from poultry
are not counted in the manure land use.
Likewise, swine are housed during a. significant portion of the year in finishing
sheds. Manure is washed daily from the finishing pens to storage ponds and
ultimately applied to cropland as liquid manure. As for poultry, spread manure
is already accounted for in the (mostly) county level estimates of manure
applications to conventional tilled, conservation tilled, andhaylands. Only the
period that swine are allowed on ground for breeding and gestation (115 days) are
accounted for as manure acres.
Following Gasman (1990), a ratio of indoor weight to outdoor weight can be
estimated. Assumptions are 10 sows per boar, an adult pig weight is 350 Ibs, two
litters of 10 pigs per sow per year, and that piglets increase weight from 30 to
250 Ibs in 6 months with an estimated annualized weight of 140 Ibs per finished
pig. Therefore: '
indoor weight - (140 IbWIO finished Digs)+(3SO lb)*(250/365 davsl - 13.6
outdoor weight (350 lbs)*(l.l adults)*(115/365 days)
Therefore, swine animal units were adjusted down by the factor of 0.0735,
Animal units of cattle were segregated into two categories. The first category
is beef cattle, defined here as beef cattle, heifers, heifer calves, steers,
steer calves, bulls and bull calves. Animal units of beef cattle are decreased
by a factor of 0.7 to account for the large proportion of cattle that are for
most of the year, unconfined to small pastures or enclosures. The second
category is dairy cows. No adjustments are made in animal units of dairy cows.
The total adjusted animal units were divided by a "compromise animal density" of
145 animal units per acre, yielding the number of manure acres. Manure acres
are listed by model segment in Appendix E.4.
Manure acres were originally taken from the pasture land use acres. . As manure
is stored and treated through the implementation of a control program', a portion
of these acres will revert back to pasture acres for simulation purposes. As
control actions of guttering, diversions, and manure containment are not totally
effective, some proportion of each manure acre will revert to pasture for the
full implementation of controls and some smaller proportion will remain in the
manure land use to account for control efficiencies less than 100X.
The overall simulation of manure land use allows the model to track agricultural
waste storage practices, and show any increases due to increases in animal
populations or decreases due to control actions.
3.1.6 Urban-Land Subcategories
The CIS system was used to differentiate the urban land into five sub-categories!
The urban land uses are listed below as described by Anderson et al. (1976).
Residential - ranging in density from very high in urban cores to low density
with units on more then one acre.
35
-------
Commercial - including urban central business districts, shopping centers,
commercial strip developments, warehouses, etc.
Industrial - including light arid heavy manufacturing as well as mining
operations, stockpiles, and spoil areas.
Transportation - roads, railroads, airports, seaports, and facilities associated
with the transportation of water, gas, oil, electricity, and communications.
Institutional - urban parks, cemeteries, open land, playgrounds, golf courses,
zoos, and undeveloped urban land in an urban setting. .
Urban land imperviousness was determined for each model segment based on the five
subcategories of urban land. The model simulates one urban land use based on the
area-weighted parameter of imperviousness of the above five subcategories. The
following lists the five classes of urban land with the reported range of
imperviousness and the single impervious value chosen for use in the model. The
range of imperviousness, is from the EPA report, National Urban Runoff Program
(U.S. EPA, 1982) except for the transportation subcategory, which was obtained
from the Federal Highway Administration report, Retention. Detention., and
Overland Flow for Pollutant Removal From Highway Stormwater Runoff (FHA, 1988).
Reported Range Impervious Value .
Land Use of Imperviousness Used in Model
residential
commercial
industrial
transportation
institutional
5-60X
65-851
75-851
5-201
27-1001
30X
75X
BOX
10X
50X
Using the proportion of the total urban area in the different subcategories and
the value of imperviousness described above, a single area-weighted
imperviousness was determined for the single aggregate urban land use modeled.
Appendix E.5 lists the percentage of urban land in each of the subcategories and
the calculated imperviousness for the area-weighted total.
The area weighted imperviousness was used to determine the expected urban load
in each model segment. The degree of imperviousness of the urban area determined
the expected annual urban load using the method described by Schueler (1987):
L - l(P)(PJ)(0.05 + 0.9)(I)/12] (C)(2.72)
where: ' /
L - expected annual load
P - rainfall depth over an annual period .
Pj - fraction of rainfall events that produce runoff
I - imperviousness
36
-------
C - mean concentration of pollutant based on national NURP average
(C values were obtained for phosphate, total phosphorus, ammonia,
nitrate, total nitrogen and BOD)
Following Schueler, the annual loads were adjusted to include baseflow loads and
a curve of expected annual urban loads was generated for the range of
imperviousness encountered. The urban land uses in each model segment were
calibrated to the expected urban load.
3.1.7 Land Use bv County
The county land use data were converted to a model segment basis. Consistent with
the level of spatial detail of the model, it is assumed that all land uses are
evenly distributed within a county. Counties within the Bay basin are shown in
Figures 1-6 (Appendix E.6).
3.1.8 Land Use bv Model Segment
Land uses by county are proportioned by percent of the county in each model
segment (Appendix E.7). The percent of county area in each segment was
determined by CIS. With few exceptions, model segments are larger than counties
in area. Usually several counties were aggregated up to compile the model
segment, data. Figures 7 through 26 in Appendix E.8 show the relationship of
model segments to county boundaries throughout the Bay basin. Appendices E.9 and
E.10 list the land uses by model segment for 1978 and 1985, respectively.
3.1.9 Land Surface Cover and Erodibilitv Parameters - COVER. SLSUR. KRER
COVER represents the fraction of the land surface that is covered by canopy, crop
residue, leaf litter, etc. and is subsequently protected from raindrop erosion.
Cover is a one of the primary determinates of the generation of sediment fines
that can be transported by runoff as part of the erosion process. The twelve
monthly values used in the model represent the land cover on the FIRST day of the
month. Cover is interpolated daily in the model between the user-defined monthly
cover values. COVER values for conventional cropland and conservation cropland
are based on the crop types grown in the segment. Major crop types aggregated
from harvested acres in the 1987 Ag Census were used to obtain unique cover
Values for conventional and conservation cropland in each model segment.
Aggregated crop categories obtained from the 1987 Ag Census include corn for
grain, sorghum for grain, wheat for grain, barley for grain, buckwheat, oats for
grain, rye for grain, sunflower seed, tobacco, soybeans, potatoes, sweet
potatoes, corn for silage, sorghum for silage, and vegetables. Section 4.3.3
discusses the development of the COVER values used for the cropland categories
in each model segment. COVER values for pervious land areas are tabulated by
land use in Appendix E.ll. .
Appendix E.ll also includes the values of the SLSUR parameter for each land use.
SLSUR represents the -slope for overland flow. This parameter influences the
simulation of hydrology and sediment erosion. Land slope data were derived from
the National Resources Inventory (NRI) data base. The county-based NRI
distribution of slopes was combined with the proportion of different county land
uses to develop an average slope for cropland, woodland and pasture in each
37 .
-------
segment. Urban data were not available from this data set. The slope of the
urban land was set equal to that of cropland.
The parameter KRER, a coefficient in the model soil fines detachment equation is
also listed in Appendix E.ll. KRER, in conjunction with the COVER parameter
controls the amount of fine sediment detached by raindrop impact and is then
available to be transported by overland flow. It is usually estimated by
assuming it is equal to the erbdibility factor, K, in the Universal Soil Loss
Equation (USLE).
3.1.10 Crop distributions for Conventional (CNT) and Conservation (GST)
Tillage
Due to computational limitations for modeling at the scale of the Bay drainage,
the Watershed Model required the concept of a 'composite crop' representation for
the CNT and CST categories in order to evaluate land cover, nutrient application
rates, and expected plant uptake rates. The 'composite crop representation' is
introduced and discussed in Section 4.3.3 as part of the nonpoint load modeling
for the cropland categories. In order to develop a 'composite crop' for each
cropland model segment, the crop distribution in the CNT and CST categories is
needed. These distributions were developed as follows:
1. Total Cropland was determined as the sum of the Phase I CNT and CST
categories after they had been adjusted for removal of the hayland
\ acres (as discussed in Section 3.1.3 above).
\ . ' '
2. The 1987 Ag census information was used to develop the crop
distributions
-------
Corn Soybeans Grains
PA 50.9 X 52.6 X . 29.6 X
MD 74.0 X 79.8 X 54.1 X
VA 67.4 X 66.0 X 50.0 X
WV 54.0 X 73.6 X 29.1 X
NY 20.6 X 7.2 X . 6.6 X
5. The acres in each crop/tillage category were then adjusted to maintain
the same split of conventional and conservation acres as the original
Phase I cropland acres.
6.' The Ag Census information was then used to determine the breakdown of
corn-grain, and corn-silage acres (and percentages) so that separate
parameter values, especially COVER (see Section 4.3.3), could be used
for each corn type.
The final 'percentages for the crop distributions for the four major categories
listed above for Conventional and Conservation Tillage are shown in Table 3.2 for
each model segment, both above and below the Fall Line.
3.1.11 Comparison of Forest Land .
Forest acreage for 1950, 1978. and 1985 were compiled from the 1950 Timber Survey
Data and from the 1978 and 1985 County Land Use Data submitted by SCS State
Offices. Appendix E.12 contains a summary of this data by county and the totals
for each state.
The CBPO Cultural Data file for 1950 does not have data for all counties in
the basin. Forest land estimates include idle land and wetlands. This was done
because the model has a limited number of land uses available.
\ , '
These data give an indication of changes in forest land from 1950 to 1985. Since
the data for 1950 are limited, however, it is difficult to show
total changes in basin acreage between 1950 and 1978. .
From the data available, forest acreage increased overall in the basin by 8X
between 1950 and 1978. Forest acreage in the Bay basin decreased by 3X between
1978 and 1985. These estimates are consistent with previous reports (U.S. EPA,
1983) which ascribe an increase in forest land between 1950 and 1978 to a
decrease in cropland acreage due to changes in agricultural practices. The
decrease in forest land between 1978 and 1985 may be due, in part, to increases
in urban land acreage. . .
3.2 POINT SOURCES
Point source inputs to the Watershed Model were developed for the 1984-87 period.
These data represent all loadings from municipal wastewater and - industrial
facilities which discharge to channel reaches in the basin. Point sources
discharging below the Fall Line were considered to be a direct discharge to the
tidal Bay' and were not included as part of the Watershed Model input data.
39 .
-------
TABLE 3.2 CROP DISTRIBUTIONS FOR CONVENTIONAL AND CONSERVATION TILLAGE FOR
EACH WATERSHED MODEL SEGMENT
SEGMENT
10
20
30
40
50
60
70
80
90
100
110
120
UO
160
170
175
180
190
200
210
220
230
23S
240
250
260
265
270
280
290
300
310
330
340
ANACOSTIA
BAL.T HAK
BOHEMIA
CHESTER
CHICKAHOM
CKOPTAMK
COASTAL1
COAST ALII
COASTAL4
COASTALS
COAST AL6
COASTALS
COASTAL9
ELIZABETH
GREAT UIC
GUNPOWDER
JAMES
NANSEMOND
N ANT 1 COKE
OCCOOUAN
PATAPSCO
PATUXENT
POCOMOKE
POTOMAC
, RAPPAKANN
SEVERN
UICOMICO
WYE
YORK
CORN
GRAIN
33. n
31. 5X
37. U
52.9X
46.2X
51.3X
49. 7X
42.6X
41. 6X
46.3X
46.1X
49.2X
47.9X
36.3X
35. 5X
40. 1X
37.7X
19. 7X
25. 6X
34. 3X
47.9X
34 .9X
24. 2X
23.8X
30.8X
21. 3X
30.9X
30.SX
24. 5X
9.6X
17.2X
34. 8X
47. 4X
54. U
55. 5X
51. 4X
54. OX
49.2X
25. 6X
31. 2X
37.2X
49.8X
25. 2X
59.4X
52. 7X
24.7X
20. OX
0.2X
21 .4X
49.4X
31. 6X
56. 2X
40.2X
35. 2X
43. 8X
52. 4X
33.3X
31. 5X
22.3X
40. 7X
28.4X
41. 6X
24. 2X
CONVENTIONAL TILLAGE
CORN
SILAGE SOYBEANS GRAINS
34. 3X
54. 2X
45. 7X
5.9X
12.7X.
,19. 6X
11. 2X
18. U
33. 2X
25. 3X
15.5X
24\9X
14.7X
31. 5X
56.5X
35. 4X
28.3X
57.2X
43. 8X
21. 5X
17. U
39.3X
13.5X
2.0X
32.3X
10.2X
67.4X
54. IX
20. IX
23. 1X
24. 1X
10.SX
7.2X
8. OX
S.8X
6.0X
7.U
3.3X
5. OX
2.4X
2.1X
16.0X
O.OX
3.3X
9.4X
0.4X
0.7X
O.OX
O.OX
7.3X
3.U
1.0X
2.2X
27.6X
9.7X
2.2X
0.9X
1.7X
0.6X
3.3X
0.6X
2.5X
2.4X
0.6X
O.OX
0.3X
11. IX
1.SX
4. IX
11 .ex
6.9X
1.2X
2.8X
9.8X
7.5X
7.5X
0.6X
0.8X
1.8X
4. IX
0.8X
3.8X
9.4X
17.5X
8.0X
40.8X
48.2X
16.9X
40.2X
O.OX
O.OX
33. 4X
44. 3X
29.3X
22.3X
1S.2X
13.6X
19.2X
11.5X
15. 6X
21. U
34.6X
27.n
30.SX
13.U
45. ZX
24. 2X
11. n
47.0X
60.2X
S5.0X
41.0X
11. 5X
46.8X
27.3X
17.9X
10.2X
12.2X
25.0X
42.3X
34.9X
43. IX
15.7X
49. OX
26. 4X
44. 8X
31. 4X
. U.3X
16.8X
30. IX
39.7X
24. 9X
27.2X
32.4X
24. OX
25. 6X
28.6X
18.SX
29.9X
31. 6X
7.2X
22.7X
29.9X
22.4X
26.9X
34. 9X
17.5X
17.7X
21. 5X
26.1X
20. IX
28.3X
1.7X
15.4X
22.0X
23.0X
29.5X
32.5X
30. 2X
24. 2X
19.5X
31. IX
23. 3X
26.3X
34.8X
38.8X
30.2X
21 .OX
29.6X
13.2X.
26.2X
27.9X
19. IX
- 44.8X
37.6X
31. 7X
18.4X
15.5X
39.7X
27.0X
34. 4X
20. 5X
23.SX
31 .9X
34.0X
40.3X
22. OX
29. 5X
28.6X
TOTAL
100. OX
100. OX
100.0X
100.0X
100.0X
100. OX
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100. OX
100. OX
100.0X
100.0X
100. OX
100.0X
100.0X
100.0X
100.0X
100. OX
100.0X
100.0X
100.0X
100.0X
100.0X
100. OX
100.0X
100. OX
100.01
100. OX
100. OX
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100.0X
100. OX
100. OX
100.0X
100. OX
100. OX
100. OX
100.0X
100.0X
100.0X
100. OX
100.0X
100. OX
100. OX
100.0X
100.0X
100. OX
100. OX
100. OX
CORK
GRAIN
42. OX
34. 9X
41 .3X
63. 9X
60.4X
60. OX
58. 7X
52.5X
48.4X
54. 4X
55. U
55.0X
56. U
44.6X
36.8X
45. 8X
44.8X
22 .3X
29.8X
41. 5X
52.4X
38.7X
28.0X
28.4X
34.7X
25.7X
31.2X
33. 2X
28.3X
11. 2X.
20. n
42.5X
53.7X
59. 4X
57.8X
59.6X
58.4X
53. U
27.5X
35. 4X
39.5X
54. 9X
30.7X
S8.4X
59.0X
29.8X
23. IX
0.2X
27.4X
57.5X
36.1X
62.2X
48. 1X
41.2X
51. 7X
53.6X
32.9X
36.3X
27.9X
49.3X
26.8X
44. 7X
29.3X
CONSERVATION TILLAGE
CORN
SILAGE SOYBEANS GRAINS
42.7X
60. U
50.8X
7.1X
16.6X
23.0X
13.3X
22.3X
38.7X
29.8X
18.6X
27.8X
17.2X
38.7X
58.6X
40.4X
33. n
64.7X
51. OX
26.0X
18.7X
43.SX
15.6X
2.4X
36.4X
12.3X ^
68.0X
58.8X
23.2X
27. 1X
29.0X
12.8X
8.2X
8.8X
6.0X
6.9X
7.7X
3.5X
5.3X
2.7X
2.2X
17.7X
O.OX
3.2X
10.6X
0.4X
0.8X
O.OX
O.OX .
8.5X
3.6X
1.1X
2.6X
32.3X
11. 5X
2.3X
0.9X
2.0X
0.7X
4. OX
0.6X
2.6X
2.9X
0.5X
O.OX
0.4X
14.4X
2.0X
5.2X
1S.OX
9.0X
1.5X
3.5X
12.5X
8.9X
12.2X
1.2X
1.9X
3.9X
6.4X
0.8X
4.4X
14.4X
20.0X
8.3X
44.4X
S4.1X
17.9X
45. 5X
O.OX
O.OX
36.2X
48.7X
33.2X
2S.6X
23.9X
20.7X
27.n
18.SX
23.4X
31 .6X
S1.6X
43.6X
45. OX
17.4X
51.8X
33.0X
18.2X
53.4X
65. 4X
70.3X
49.3X
18.6X
S0.2X
28.4X
29.6X
11.2X
19.9X
35. 4X
56.5X
45. 3X
50.8X
26.4X
64.0X
39.5X
51 .OX
14. 7X
5.0X
7.6X
u. n
21.0X
11. 8X
13.0X
16.2X
11. 3X
12.2X
13.8X
8.4X
14.5X
15.6X
2.n
9.9X
15. IX
12.2X
14.9X
18. IX
8.8X
9.5X
12.0X
15.1X
11. OX
16.5X
0.8X
8. IX
12.3X
13.0X
17.2X
19.2X
14.2X
11. OX
8.4X
15.0X
10.5X
11.8X
15.5X
18.3X
13.3X
10.0X
17.5X
5.4X
12.2X
16.3X
10.7X
29.5X
23.3X
15. 4X
10.2X
8.3X
19.7X
15. 3X
16.9X
8.7X
9.7X
16.SX
20.6X
20.3X
8.6X
13.2X
16.8X
TOTAL
100. OX
100. OX
100. OX
100.0X
100. OX
100. OX
100. OX
100.0X
100. OX
100.0X
100. OX
100. OX.
100.0X
100.0X
100.0X
100. OX
100. OX
100.0X
100. OX
100. OX
100. OX
100.0X
100. OX
100. OX
100.0X
100.0X
100.0X
100. OX
100.0X
100. OX
100. OX
100.0X
100. OX
100.0X
100. OX
100.0X
100.0X
100.0X
100.0X
100. OX
100.0X
100.0X
100.0X
100.0X
100.0X
100. OX
100.0X
100.0X
100.0X
100. OX
100.0X
100. OX
100. OX
100. OX
100.0X
100.0X
100.0X
100.0X
100.0X
100. OX
100. OX
100. OX
100. OX
40
-------
The Point Source Atlas was used as the universe of possible point sources.
Municipal dischargers were selected based on a flow of 0.4 million gallons per
day (MGD) or greater. This criteria captured more than 96X of the municipal
point source flow. The remaining (approximately) 4X of point source flow was
from numerous small discharges..
Industrial dischargers have highly variable concentrations of discharged
nutrients, ranging from zero to many times the concentrations associated with
municipal discharges. Accordingly, Industrial dischargers were included in the
point source data set if the load from the industrial source was equivalent to
the TN, TP, or BOD load of a 0.4 mgd municipal point source with secondary
treatment. Industrial dischargers were selected if any of the following were
true: total phosphorus load greater than or equal to 12 Ib per day; or, total
nitrogen load greate^r than or equal to 40 pounds per day; biological oxygen
demand greater than or equal to 100 pounds per day.
Data for all facilities that discharge to streams within a model segment were
aggregated to obtain a single set of point source loads for the corresponding
model reach. The constituents and units are listed below:
CONSTITUENT UNIT
flow ac-ft
BOD Ibs
DO Ibs
ammonia (NH3) Ibs
nitrate (N03) ,lbs
qrganic-N Ibs
phosphate (P04) Ibs
. " organic-P Ibs
heat BTU
The data sets consist of monthly total loads for each reach for the 1984-87
period. It was determined that monthly values represented the most appropriate
resolution for the available data.
Data for these .parameters were derived in the following manner. If State
(National Pollution Discharge Elimination System, NPDES) data were available,
they were used preferentially. When no State NPDES data were available, data
from the 1985 Point Source Atlas were used. As a last resort, defaults were
calculated for missing data, as described below. Permit Compliance System data
were found to contain many errors and were not used to compile the initial data
set. . . ' . /
Defaults for municipal dischargers were based on default concentrations applied
to the municipal flows. The record of flows for municipal dischargers was good,
and only measured, values of flow were used. The default value for dissolved
oxygen assumed a dissolved oxygen concentration of 6.25 milligrams per liter for
municipal dischargers. A flow-weighted load was calculated from the default.'
Tables 3.3 and 3.4 summarize nitrogen and phosphorus default concentrations and
chemical-species percentages used when nutrient data were missing from municipal
flows. '
41
-------
TABLE 3.3 MUNICIPAL NITROGEN DEFAULT PERCENTAGES OF AMMONIA, NITRATE AND ORGANIC
NITROGEN, AND THE DEFAULT CONCENTRATIONS OF TOTAL NITROGEN. DEFAULTS
BASED ON LEVELS OF SECONDARY TREATMENT
Maryland Virginia **
Year
1984
1985
1986
1987
1988
Ammonia/
Nitrate/
Org. N,
percent
50/35/15
50/35/15
50/35/15
50/35/15
50/35/15
Default
Total
Nitrogen
mg/1
18.0
18.0
18.6
18.0
18.0
Ammonia/
Nitrate/
Org. N,
percent
73/11/16
73/11/16
73/11/16
73/11/16
75/09/16
Default
Total
Nitrogen
mg/1
18.7
18.7
18.7
18.7
18.7
Pennsylvania
Ammonia/
Nitrate/
Org.N,
percent
50/35/15
50/35/15
50/35/15
50/35/15
50/35/15
Default
Total
Nitrogen
mg/1
18.5
18.5
18.5
18.5
18.5
**
Virginia default concentrations and percentages were based on 7
1 Hampton Roads Sanitation District processes.
Note: Tabulated point source data for 1988 are npt currently included in the
Watershed Model.
TABLE 3.4 MUNICIPAL PHOSPHORUS DEFAULT PERCENTAGES OF PHOSPHATE AND ORGANIC
PHOSPHORUS, AND THE DEFAULT CONCENTRATIONS OF TOTAL PHOSPHORUS.
DEFAULTS BASED ON LEVELS OF SECONDARY TREATMENT **
Maryland
Year
1984
1985
1986
1987
1988
Default
Ortho-P/
Org. P,
percent
85/15
85/15
85/15
85/15
85/15
Total
Phos.,
mg/1
7.0
7.0
3.0
3.0
3.0
Virginia
Default
Ortho-P/
Org. P,
percent
85/15
85/15
85/15
85/15
81/19
Total
Phos . ,
mg/1
6.4
6.4
6.4
6.4
2.5
Pennsylvania
Default
Ortho P/
Org. P,
percent
85/15
85/15
85/15
85/15
85/15
Total
Phos.
mg/1
8.0
8.0
8.0
8.0
8.0
** Primary or tertiary treatment level ratio of ortho-phosphate to organic
phosphate is 45/55.
Note: The tabulated point source data for 1988 data are not currently included
in the Watershed Model.
42
-------
The. shift in default values of TP for Maryland in 1985 and for Virginia in 1988
reflects the phosphate ban on detergents that went into effect for Maryland in
December 1985 and for Virginia in January, 1988. .
Missing municipal biochemical oxygen demand (BOD) values were replaced with the
average of values for the months for which data was available. No defaults were
applied if all BOD data was missing.
A different procedure was developed for industrial load defaults. Industrial
flows are highly variable and, when nutrients are discharged, are characterized
by low volume and high concentrations. In addition, some industrial flows are
very large t such as water used for condenser cooling or hydroelectric generation,
and have no nutrient loads whatsoever. As a consequence, industrial flows are
assumed to have zero flow (and are, of course, not counted as surface water
diversions) . Nutrients from industrial sources were input to the channel as a
load based on the dischargers' reported values.
Since industrial flow is not quantified, dissolved oxygen loads for industrial
dischargers are not required. Industrial loads are primarily BOD, P04, organic
P, or NH4. Industrial loads of nitrate and organic N are negligible and not
included. The total industrial nitrogen load is assumed to be entirely ammonia.
The industrial phosphorus load is assumed to be 851 ortho-phosphorus and 1SZ
organic phosphorus.
All of the data sets generated as described above were then reviewed by the
States of Virginia, Maryland, and Pennsylvania for accuracy. The reviewed data
sets were then further processed for use as model input.
State contacts who provided input to the collection and review of the point
source data are:
Maryland: Narendra Pandy
Maryland Department of the Environment
2500 Broening Highway
Baltimore, Md. 21224
Pennsylvania: Ken Bartel / Cedric Karper / George Fctchko
Pennsylvania Bureau of Water Quality Management
Fulton Building, 12th Floor
, Harrisburg, Pa. 17120
Virginia: Alan Pollock / John Kennedy / Arthur Butt
Virginia State Water Control Board
2111 North Hamilton Street
Richmond, Va. 23230
The point source dischargers were assigned to appropriate model segments by the
NPDES record latitude and longitude of each discharger. Point source loads by
model segment were then reformatted and stored in the WDM files for the
appropriate river basins. During model simulations, the monthly values were
divided evenly over all simulation intervals (hourly) in the month. The
following tables are included in Appendix F:
43
-------
Appendix F.I details the total basin-wide point source loads by year.
Appendix F.2 displays loads by model segment for each year.
Appendix F.3 displays loads by basin for each year.
Appendix F.4 lists the total municipal loads for each parameter by State.
Appendix F.5 lists the total industrial loads for each parameter by State.
Appendix F.6 lists the point source discharges for which data was compiled.
3.3 SURFACE WATER DIVERSIONS . ,
Ma j or' surf ace water diversions from channel reaches are represented in the
Watershed Model for the 1984-1987 simulation period. Data were obtained from
U.S. Geological Survey offices in Pennsylvania, Maryland, and Virginia (USGS,
-1988) and from the U.S. Army Corps of Engineers report Water Needs Assessment of
the Susquehanna River Basin (COE, 1988).
^
The data were reported in monthly or annual diversions for each water user. If
only annual diversions were available, then equal monthly divisions of the annual
total were assumed.
Categories of water use included as surface water diversions were:
domestic water use - water for household purposes.
industrial water use - water used for .industrial fabrication, processing, washing
and cooling. - . .
irrigation water use -water used.to assist in growing crops or to maintain
vegetative growth.
public supply - 'water' withdrawn by public and private water suppliers and
delivered to groups of users. , .
Several categories of water uses were assumed to involve a "once through"
nonconsumptive use and were not included in the model data of surface water
diversions. These categories are hydroelectric, thermoelectric, and industrial
cooling. Mining is assumed to be a water use which provides its own source of
wash or process water and does not involve consumptive use. Ground water sources
and water that is continually recycled in industrial use or in mineral processing
were also not included in the surface water diversion data.
Commercial water, use is a subcategory of the public supply water use and is not
included to avoid double-counting. National figures (USGS Circular 1004,
Estimated Use of Water in the United States in 1985) show that 93X of commercial
water is from public supply (82Z) or ground water sources (112).
The surface water diversions are assigned to model segments by the reported
latitude and longitude and. summed for each segment. The surface water diversion
data is reformatted and stored in WDM files in time steps of one day and in units
of cfs. -The WDM files are the total surface water diversions in each model
segment. Appendix F.7 lists the average annual surface water diversions in cfs
in each model segment.
44
-------
3.4 ATMOSPHERIC SOURCES
The Watershed Model accounts for the atmospheric deposition of nitrogen and
phosphorus directly onto water surfaces for the 1984-87 period of. simulation.
Deposition to water surfaces is explicitly modeled.
Deposition of inorganic nitrogen to land surfaces is explicitly included in the
AGCHEM land uses (conventional cropland, conservation cropland, and hayland)
through the inclusion of atmospheric loads with nutrient applications of
fertilizer and manure (see Section 4.3.4). Atmospheric deposition to PQUAL land
uses (forest, urban, pasture) is implicitly included by calibration to the annual
loads observed in field measurements.
The wetfall ammonia and nitrate loads for each model segment were determined by
the use of annual isopleths produced by the National Atmospheric Deposition
Program (NADP, 1982 - 1987). The wetfall nitrate and ammonia loads vary
spatially by model.segment with the highest deposition generally in the northwest
areas of the bay basin, (Vest Branch Susquehanna and Juniata), and the lowest
deposition in the southeast area of the basin (lower James, lower Eastern Shore).
The year-to-year variation of the wetfall nitrate assigned to model Segments for
the 1982-87 period is about 16X. The year-to-year variation in wetfall ammonia
is about 44Z for the same period.
Following Tyler (1988), the dryfall nitrate loads are assumed Co be equal to the
wetfall nitrate. Accordingly, the dryfall nitrate was set to the long term
(1982-1987) average of the wetfall nitrate. The dryfall nitrate is spatially
distributed by model segment, but has no year-to-year variation. Dryfall ammonia
is assumed to be negligible.
Orthophosphate, organic nitrogen and organic phosphorus are not typically
monitored by NADP. Annual loads of these constituents were derived from the.
report'Chesapeake Bay Program Technical Studies: A Synthesis (U.S. EPA, 1982).
The Synthesis Report gives no indication of year-to-year or spatial variation of
these parameters, so constant loads are used for all years. Table 3.5 lists the
basin-wide average annual atmospheric deposition loads.
TABLE 3.5 BASIN-VIDE AVERAGE ANNUAL ATMOSPHERIC DEPOSITION LOADS
(LB/AC/YR)
ORGANIC TOTAL ORTHO TOTAL
YEAR NH3 N03 NITROGEN NITROGEN PHOSPHORUS PHOSPHORUS
1984
1985
1986
1987
1.95
1.50
1.68
1.70
6.94
6.70
6.86
6.41
6.07
6.07
6.07
6.07
15.0
14.3
14.6
14.2
.143
.143
.143
.143
.566
.566
.566
.566
The average total nitrogen load compares favorably with other estimates of
atmospheric deposition in the basin (Jaworski and Linker, 1991):
45
-------
The data was reformatted as a monthly load to the total model segment water area
and added to the WDM files. The data is input to the model in the same manner
as the point source inputs, i.e., the monthly totals are divided evenly over each
hour of the month. Atmospheric loads by year and model segment are included in
Appendix F.8.
46
-------
SECTION 4.0
NONPOINT LOADING SIMULATION MODULES
AND SIMULATION RESULTS
This section discusses the HSPF modules used to simulate the nonpoint loadings
from each of the land uses considered in the Phase II Watershed Model. Table 4.1
shows the land use categories in the model, while Table 4.2 lists the water
quality and nonpoint "loading constituents simulated in Phase I. Note that TOC
(Total Organic Carbon) and Chlorophyll A are simulated only as instream
constituents.
The focus in Phase II was to provide a more detailed simulation of agricultural
cropland areas so that a cause-effect relationship could be represented between
the nutrient applications (fertilizer and manure) on these areas and the
subsequent runoff loadings from these land uses. Consequently, the empirical
approach to nonpoint simulation used in Phase I was replaced with the more
detailed mechanistic approach provided by the Agrichemical Module sections of
HSPF. Also, a significant element of the Phase II work was the discovery that.
hay cropland was a major portion of the total agricultural cropland as. defined
in Phase I. Hay cropland represents SOX of the total cropland above the Fall
Line, and 10Z of the cropland below the Fall Line. Since'the nutrient loadings
from hay areas are expected to be significantly lower than from cultivated
croplands, the acreage of hay cropland was removed from the total' cropland and
simulated as a separate land use category.
The procedures.used in simulating the nonpoint loadings from both cropland and
non-cropland areas are discussed in the following subsections, along with the
estimation of selected input parameters needed for model application at the scale
of the Chesapeake Bay drainage area. The loading simulation results for all land
uses are discussed in the final subsection, Section,4.4. Further details on the
model equations and algorithms are contained in the HSPF Users Manual (Johanson
et al. , 1984) and the HSPF Application Guide (Donigian et al., 1984).
4.1 OVERVIEW OF CROPLAND AND NON-CROPLAND SIMULATION
Within the HSPF framework, the PERLND and IMPLND module sections are used to
simulate hydrologic, sediment, and water quality processes that produce nonpoint
loadings to a stream from both surface and subsurface pathways. 'The structure
charts for the PERLND and IMPLND module sections are shown in Figures 4.1 and
4.2, respectively. Each module section is comprised of subroutines, or groups
of subroutines, that perform the simulation of individual processes. For
example, the SNOU module in Figures 4.1 and 4.2 performs the snow accumulation
and melt simulation for both pervious and impervious land areas, while the PWATER
module in Figure 4.1 performs the hydrologic simulation for all pervious land
47
-------
TABLE 4.1 LAND USE CATEGORIES
IN THE WATERSHED MODEL
Simulated Land Uses:
Pervious -
Forest
High Tillage Cropland
Low Tillage Cropland
Hay Cropland
Pasture
Urban
Impervious
Animal Waste
Urban
Urban Categories:
Residential
Commercial and Services
Industrial
Transportation
Institutional
48
-------
TABLE 4.2 WATER QUALITY CONSTITUENTS
SIMULATED IN THE WATERSHED MODEL
Dissolved Oxygen (DO)
Water Temperature
Sediment
Biochemical Oxygen Demand (BOD)
Nitrate-Nitrogen (NO3-N)
Ammonia-Nitrogen (NH3-N)
Organic Nitrogen
Orthophosphorus (PO4-P)
Organic Phosphorus
Total Organic Carbon (TOO
Chlorophyll A (CHL A)
Total N
Total P
49
-------
PERLND
Simulate
a pervious
land
segment
ATEMP 1 SNOW 1
Correct air S
tempera- s
ture ic
PSTEMPI
Estimate
soil
tempe'a-
ture(s)
imulate
now and
:e
PWATER 1 SEDMNT!
Simulate Simulate
water , sediment
budget
PWTGAsI
Estimate
water
tempera-
ture
gas
and
con-
centrations
PQUAL 1
Simulate
general
quality
constit-
uents
MSTLAY
PEST
NITR
PHOS
TRACER
Estimate
solute
transport
Simulate
pesticides
Simulate
nitrogen
Simulate
phosphorus
Simulate
a tracer
(conserva-
tive)
'
AGRI-CHEMICAL SECTIONS
Figure 4.1 PERIOD structure chart
50
-------
IMPLND
'
Perform
computa-
tions on a
segment of
impervious
land
ATEMP 1 SNOW I WATER I
(see
module
PERLND)
(see
module
PERLND)
-
Simulate
water
budget
for
Impervious
land
segment
-
SOLIDS 1 IWTGAS | IQUAL
Accumulate
and remove
solids
Simulate
water
tempera-
tures and
dissolved
gas cones.
f
Simulate
quality
constit-
uents
using
: simple
relation-
ships with
solids
and/or
water
yield
Figure 4.2 Structure chart - impervious land-segment application module,
-------
areas. The specific processes simulated by each module are identified in the
figures.
Since HSPF has the ability to use ซither simple or detailed (i.e., AGCHEM)
approaches for nonpoint loading simulation, different modules were used for
different land use categories to combine the simplified simulation approach from
Phase I for non-cropland areas with the detailed AGCHEM approach for the cropland
categories. Table 4.3 lists the specific HSPF modules used to simulate the
individual nonpoint loading constituents from each land use category.
Typically, nonpoint loadings are calculated.in nonpoint models by using one or
more of three basic approaches, which include the following:
a. Potency factors and subsurface concentrations
b. First-order washoff and subsurface concentrations
c. Detailed modeling of land surface and soil processes
In the 'potency factor' approach, nonpoint loadings from the land surface are
calculated as a function of Che sediment loading rate, which in turn may be
calculated from the Universal Soil Loss Equation (Uischmeier and Smith, 1978),
or one of its modifications, or by detailed modeling of storm runoff and soil
erosion. The potency factor approach has been used in ~a number of nonpoinC
loading estimation techniques and models, including Che NFS model (Donigian and
Crawford, 1976b) which was. Che basis for Che original NVPDC and Phase I CBP
watershed modeling efforts. .
1 v
The 'first-order washoff approach includes a daily calculation of pollutant
accumulation/deposition on Che land surface, .and a subsequent washoff of
pollutants for storm events as a first-order function of Che scorn runoff race.
This approach was originally developed for urban areas where impervious surfaces
provide the major fraction of surface nonpoinC loadings; Che veil-known EPA Scorn
Water Management Model (SWMM) (Huber and Dickinson; 1988), the Modified NFS Model
used in the NVPDC Study, HSPF, and numerous other models use Che first-order
washoff approach. - " .
Both the potency factor and first-order washoff approaches are used exclusively
for surface loadings; they rely on user-specified subsurface concentrations to
define the subsurface pollutant contributions.
The 'detailed modeling' approach involves Che representation of soil chemical and
biochemical processes that, in conjunction with hydrologic and erosion modeling,
calculate both the surface and subsurface nonpoinC loading contributions: For
modeling nonpoint nutrient loadings, this often involves Che calculation of
mineralization, nitrification/denitrification, immobilization,
sorption/desorption, plant uptake, and other soil nutrient processes as impacted
by nutrient applications (fertilizer and manure) , agronomic practices, hydrologic
conditions,' and soil moisture and temperature.
Although this approach requires more data and input than the other empirical
methods, detailed modeling allows a direct linkage between nutrient application
rates and agricultural practices, and resulting runoff loadings. The HSPF AGCHEM
modules, CREAMS (Knisel eC al., 1980), CREAMS-NT (Deizman and Mostaghimi, 1991),
' . ' . ; ' 52
-------
TABLE 4.3 HSPF NODULES USED TO SIMULATE NONPOINT LOADINGS FROM EACH LAND USE IN PHASE II
en
OJ
Hydrology
Dissolved Oxygen
Water Temperature
Sediment Loading
Organic* Loading
NOj-N
NHj-N
Organic N
P04-P
Organic P
Conventional
Tillage
PWATER/SNOU
PWTGAS
STEMP
SEDNNT
POUAL
A6CHEN
AGCHEN
AGCHEN
AGCHEN
AGCHEN
Conservation
Tillage
PUATER/SNOW
PWTGAS.
STEMP
SEDNNT
POUAL
AGCHEN
AGCHEN
AGCHEN
AGCHEM
AGCHEM
Hay
PUATER/SNOW
PWTGAS
STEMP
SEDNNT
POUAL.
AGCHEN
AGCHEM
AGCHEM
AGCHEN
AGCHF.N
Forest
Pasture
Urban
Pervious
Urban
Impervious
Manure
Acres
PWATER/SNOW PWATER/SNOU PWATER/SNOW IWATER/SNOW IWATER/SNOW
PWTGAS PWTGAS PWTGAS IWTGAS
STEMP STEMP STENP IWTGAS
SEDNNT SEDMNT SEDMNT - -
POUAL POUAL POUAL IOUAL
POUAL POUAL POUAL IOUAL
POUAL POUAL POUAL IOUAL
POUAL
POUAL
POUAL
IOUAL
-------
and other agricultural models use this general approach to varying degrees of
process representation.
The PQUAL and IQUAL modules of HSPF allow both the potency factor and first-order
washoff approaches to nonpoint loading simulation. As shown in Table 4.3, these
modules were used in Phase 11 to simulate the nonpoint loadings for all
constituents from forest, pasture, and urban (pervious and impervious) land uses,
and for the BOD/organics loading from the cropland categories. The AGCHEM
modules were used for nutrient loadings from the cropland areas. The nutrient
loadings from the animal waste 'manure acres' were derived from the runoff and
defined waste concentrations as described below in Section 4.2. That section
also discusses additional details of the non-cropland loading simulation.
4.1.1 Hydrologic. Sediment Loading. DO. and Water Temperature Simulation
* ;
As shown in Table 4.3, the hydrology, sediment, DO (Dissolved Oxygen), and runoff
water temperature simulations are based on the same procedures, utilizing the
same HSPF modules, for all pervious . land categories. The hydrology for
impervious land categories (i.e., urban impervious and manure acres) uses a
different module, although the snow simulation is identical. Since these modules
are common to all the land uses, a brief overview of the simulation approaches
is provided below with further details available in the HSPF User Manual.
The hydrologic submodel, PWATER, in HSPF was derived originally from the Stanford
Watershed Model (Crawford and Linsley, 1966), and subsequent refinements through
the HSP (Hydrocomp, Inc., 1976) , ARM (Donigian and Crawford, 1976a; Donigian et
al., 1977), and NPS models. Figure 4.3 is a flowchart of the .hydrologie
processes simulated in PWATER; the snowmelt processes are simulated in the
separate SNOW modules of HSPF. PWATER calculates a complete water balance for
each land use category within the watershed, or model segment, by converting
input rainfall and evaporation data into the resulting surface runoff, changes
in soil moisture storages for various portions of the soil profile, infiltration
of water, actual evapotranspiration, and subsequent discharge of subsurface flow
(both Interflow and baseflow) to the stream channel. During storm events,
rainfall is distributed among surface runoff and soil moisture storage
compartments based on nominal storage capacities and adjusted infiltration rates.
Between storm events, water storage in the soil profile is depleted by
'evapotranspiration and subsurface recharge, thereby freeing up soil moisture
capacity for rainfall inputs from the next storm (NVPDC, 1983)i
The SNOW module requires input of additional meteorologic timeseries, that is in
addition to precipitation and evaporation, including solar radiation, air
temperature, wind, and dewpoint. Energy balance calculations, in terms of heat
fluxes, are performed to evaluate melt components associated with radiation
(solar and terrestrial), heat exchange due to condensation and convection,
rainfall temperatures, and subsurface heat. In conjunction with the melt
calculations, air temperatures are used to determine whether precipitation is
falling as rain or snow, and snowpack characteristics are simulated including
snow density and compaction, areal coverage, albedo, evaporation, heat loss, and
liquid water storage. The end result is a continuous simulation of snow depth,
water equivalent, liquid water storage, and subsequent release of melt water to
the land surface. This melt water release is then received by the PWATER modules
54
-------
Actual '
ET /
Potential ET
Precipitation
Temperature
Radiation
Wind, Dew point
-S/-
Snowmelt
Interception
Storage
Interception
Subroutine)
i r^- '
( Input )
("Output j
' I Sforage~~~\
0Decision
ET-Evapo-
transpiration
Lower Zone
Storage
Surface
Runoff
Interflow
Overland
Flow
Upper
Zone
Storage
Interflow
Deep or Inactive'
Groundwater
Groundwater
Storage
/To
\ Stream
Figure 4.3 Hydrologic processes simulated by the PWATER module of HSPF.
-------
as input for the soil hydrologic processes shown in Figure 4.3. The equations
used in SNOW are based on work by the U.S. Army Corps of Engineers (1956),
Anderson and Crawford (1964), and Anderson (1968); they are identical to the
equations used in the HSP, ARM, and NFS codels.
For impervious land surfaces, the IWATER oodule of HSPF performs the hydrologic
simulation which includes only the processes of detention or retention of
incident rainfall, evaporation from retention storage, and overland flow routing
of the rainfall excess. Retention accounts for surface depressions and ponding
on various impervious surfaces (e.g., roof tops, parking lots, road side
depressions) from which the retained water can then evaporate. Thus, impervious
runoff simulation would be analogous to including only the 'interception' and
'overland flow' portions of the hydrologic processes shown in Figure 4.3 (for
pervious surfaces).
The SEDMNT module of HSPF performs the sediment loading simulation for all
pervious land use categories, i.e., all land uses except urban-impervious and
manure acres. Figure 4.4 shows the sediment processes and fluxes simulated by
SEDMNT, including detachment of sediment particles by raindrop impact, net
vertical sediment input (or export), attachment or aggregation of fine sediment
particles, washoff of detached sediment, and scour of the soil matrix by overland
flow.. All these processes are simulated in the Watershed Model except for scour
of the soil matrix which requires information at finer special scale than was
available for the model segments.
The equations used to produce and remove sediment are based on the ARM and NFS
Models (Donigian and Crawford 1976 a,b). The algorithms representing land
surface erosion in these models were derived from a sediaent model developed by
Mcshe Negev (Negev 1967) and influenced by Meyer and Wischmeier (196?) and Onstad
and Foster (1975). The equation which., represents the scouring of the matrix
soil, which is not included in ARM or NFS, was derived from Negev's method for
simulating gully erosion.
Two of the sediment fluxes shown in Figure 4.4, SLSED and HSVI, are . added
directly to the detached sediment storage variable DETS. SLSED represents
external lateral input from an upslope land segment. It is a time series which
the user may optionally specify, this option was not implemented in the Watershed
Model. NVSI is a parameter that represents any net external additions or
removals of sediment caused by human activities or Wind.
Impacts of tillage operations on sediment availability are affected in SEDMNT by
re-setting the Value of DETS to represent an .increased -amount of sediment fines
available for erosion, at the time of the tillage operation.
The washoff process involves two parts: the detachment/attachment of sediment
from/to the soil matrix and the transport of this sediment. Detachment (DET)
occurs by rainfall. Attachment occurs only on days without rainfall; the rate
of attachment is specified by parameter AFFIX. Transport of detached sediment
is by overland flow. .
56
-------
SLSED
lateral
input of
sediment
to sur-
face
NVSI
net ver tical
sediment
input
WSSD
washoff of
detached
sediment
by water
DETS
detached
sediment-
storage
AFT ix
sediment
attachment
SOSED
total
removal
of .soil &
sediment
from sur-
face by
water
DET
dolachmenl
of soil by
rainfall
soil matrix
(assumed to
have '
unlimited
storage) .
scour of
matrix
soil by
water
Figure 4.4 Flow diagram for SEDMNT section of PERLND application module
-------
Kinetic energy from rain falling on the soil detaches particles which are then
available to be transported by overland flow. The equation that simulates
detachment is:
DET - DELT60*(1.0 - CR)*SMPF*KRER*(RAIN/DELT60)**JRER (4.1)
i ' .
where: .
DET - sediment detached from the soil matrix by rainfall in
tons/acre per interval
CR - fraction of the land covered by snow and other cover
SMPF - supporting management practice factor
KRER - detachment coefficient dependent on soil properties
RAIN - rainfall in in./interval
JRER - detachment exponent dependent on soil properties
^
The variable CR is the sum of the fraction of the area covered by the snowpack
(SNOCOV), if any, and the fraction that is covered by anything else but snow
(COVER). SNOCOV is computed by section SNOW. COVER is a parameter which for
pervious areas will typically be the fraction1 of the area covered by ve.getation
and mulch. It can be input on a monthly basis.
When simulating the washoff of detached sediment, .the transport capacity of the
overland flow is estimated and compared to the amount of detached sediment
available. The transport capacity is calculated by the equation:
STCAP - DELT60*KSER*((SURS + SURO)/DELT60)**JSER (4.2)
where:
STCAP - capacity for removing detached sediment in .
tons/acre per interval
DELT60 - hr/iriterval .
KSER - coefficient for transport of detached sediment
SURS - surface water storage in inches .
SURO - surface outflow of water in in./interval
JSER -exponent for transport of detached sediment
When STCAP is greater than the amount of detached sediment in storage, washoff
is calculated by:
WSSD - DETS*SURO/(SURS + SURO) (4.3)
If the storage is sufficient to fulfill the transport capacity, then the
following relationship is used:
WSSD - STCAP*SURO/(SURS + SURO) (4.4)
where:
. WSSD - washoff of detached sediment in tons/acre per interval
DETS - detached sediment storage in tons/acre
WSSD is then subtracted from DETS.
58
-------
SEDMNT also simulates the re-attachment of detached sediment (DETS) on the
surface due to soil compaction, particle aggregation, etc. Attachment to the
soil matrix is simulated by merely reducing DETS by a daily rate input by the
user. Since the soil matrix is considered to be unlimited, no addition to the
soil matrix is necessary when this occurs. DETS is diminished at the start of
each day that follows a day with no precipitation by multiplying it by (1.0 -
AFFIX), where AFFIX is the user defined parameter. This represents a first order
rate of reduction of the detached soil storage.
Water temperature of runoff and the DO concentration are calculated by the PSTEMP
and PWTGAS modules, respectively, for the pervious land categories. Actually,
the PSTEMP module calculates soil temperatures for each of the defined soil
layers surface, upper zone, lower zone, and groundwater zone - and the flow
component originating from that zone is assumed to be at the calculated soil
temperature, except that the water tempera'ture cannot be less than freezing.
Thus, surface runoff occurs at the calculated surface soil temperature, interflow
occurs at the calculated upper zone temperature, and groundwater occurs at the
calculated groundwater zone temperature. In addition, these calculated
temperatures are used to perform adjustments to the input soil nutrient reaction
rates in the AGCHEM module (see Section 4.3 below)
For the Watershed Model, the surface and upper zone soil temperatures are based
on a linear regression with air temperature in each time interval. The lower
zone and groundwater temperatures are assumed to be the same and are input
monthly values defined by the user; these temperatures correspond to the lower
root zone and shallow groundwater that contributes baseflow to the stream.
The surface soil temperature and surface runoff temperature is then used in the
. PWTGAS module to calculate the DO concentration of the overland flow, which is
assumed to be at saturation. PWTGAS uses the following empirical nonlinear
equation to relate dissolved oxygen at saturation to water temperature (Committee
on Sanitary Engineering Research, 1960):
SODOX - (14.652 + SOTMP*(-0.41022 +
SOTMP*(0.007991 - 0.000077774*SOTMP)))*ELEVGC (4.5)
where:
SODOX - concentration of dissolved oxygen in surface outflow, in mg/1
SOTMP - surface outflow temperature in degrees C
ELEVGC - correction factor for elevation above sea level
The concentration of dissolved oxygen in the interflow and the active groundwater
flow cannot be assumed to be at saturation. Values for these concentrations are
provided by the user. He may specify a constant value or 12 monthly values for
the concentration of each of the gases in interflow and groundwater. If monthly
values are provided, daily variation in values will automatically be obtained by
linear interpolation between the monthly values.
For impervious surfaces, the procedures for runoff water temperature and DO are
identical to those used for pervious surfaces; the IWTGAS module uses a linear
regression to calculate impervious overland flow temperature, which is then used
with Equation 4.5 (above) to calculate the DO concentration.
59
-------
4.2 SIMULATION OF NON-CROPLAND AREAS
Based on the hydrology and sediment loading simulations, the PQUAL module
simulates the nonpoint pollutant loadings, from the non-cropland areas of forest,
pasture, and urban-pervious, while the IQUAL module simulates these loadings for
the urban-impervious land use category. Animal waste contributions from manure
acres are derived from IWATER/SNOW modules and user-defined concentrations, as
discussed in Section 4.2.2 (below).
4.2.1 Forest. Pasture. Urban Area Procedures
Both PQUAL and IQUAL allow use of either the potency factor or first-order
washoff approach for any of the nonpoint constituents that are simulated. In.the
Watershed Model, the potency factor .approach was selected for all constituents
from pervious land areas, while the first-order washoff approach was selected for
all constituents from impervious land areas within the non-cropland categories.
This was the approach used in Phase I and is consistent with current nonpoint
modeling procedures. . .
Figure 4.5 is the flow chart for the PQUAL module. The corresponding diagram for
the IQUAL module is identical to the top half of Figure 4.5, since pollutant
contributions from scour, interflow and groundwater are not allowed from
impervious land areas.
The potency .factor approach simply assumes that the contaminant loading is a
direct function of the sediment load, as simulated by the SEDMNT module in each
simulation timestep. The relationship is shown as follows:
WASHQS - VSSD*POTFW
where:
VASHQS - flux of quality constituent associated with
detached sediment washoff in quantity/acre per interval
USSD - washoff of detached sediment in tons/acre per interval
POTFtf - washoff potency factor in quantity/ton .
In Figure 4.5, the pollutant contribution calculated by the above equation is
shown as 'associated with detached sediment'. Potency factors indicate the
pollutant strength relative to the sediment removed from the land surface.
Certain constituents that are strongly adsorbed to sediment particles, such as
metals, phosphates., organics,etc., actually move with the eroded sediment;
whereas, for other less-highly attached contaminants, (e.g., ammonia, nitrate,
BOD) this approach simply means that sediment loadings are a reasonable indicator
of the expected contaminant load in surface runoff.
Potency factors are specified separately for each land use category, and can be.
input on a monthly basis for pollutants that demonstrate a definite seasonal
variability. In conjunction with the potency factors, subsurface (i.e.,
interflow and baseflow) concentrations are needed to define the subsurface
contributions; these concentrations can also be defined on a monthly basis for
seasonal variation.
60
-------
. -
/
I
> <
storage of
OUAL on
surface, (or
direct wash*
oil by
overland
Mow
,1
t ' ' 1 ซ '
storage ol
/QUAL '\
affsoclateq
with soil1 ..
( /Dal r 1 x v'
/
5S5ES9BP>^^**(
storage ol j
QUAL I
associated 1
with 1
Interflow I
aL^Ii^^^j^ai
storage ol
QUAL
associated
with active
groundwater
IwSSM^ftMMi^taJ
s" '
accumulation
direct
washofl
by
overland
Lflowl 30QO
washoll
JM- '" ' ''l| associated
H storage ol II with
II OUAL M detached
N associated with D ^v sed- jr
g II
T " '"" ^ ' - (
' <^r ^^^. ftCOAA AAA4
. scour
of QUAL
associated
1 with soil .
matrix
1
total
. Dull IOW
ol OUAL
trom
surface
^
^^. 100
^^' ^^^ I**W
outflow I -. .!_...
- 1 r>n AI 1 OU{ HOW
Ol OUAL I _ i /-MI AI
ซociii.d 1 ฐ'w?,UhA
seSTen, ^erMo,
^^^ ^K
out now
Of QUAL
wi th
active
ground
l^vvater
ii A
:)
S
1
I
POQUAL
^
Total
i out Mow
of OUAL
^^^
Figure 4.5 Flow diagram'for PgUAL section of PERLND application module.
-------
In the first-order washoff approach, simulation of surface runoff pollutant loads
involves two components: daily accumulation and depletion (non-runoff related)
of contaminants on the land surface, and washoff by overland flow during storm
events. The constituent can be accumulated and removed by processes which are
independent of storm events such as street cleaning, decay, and wind erosion and
deposition, or it can be washed off by overland flow. The accumulation and-
removal rates can have monthly values to account for seasonal fluctuations.
When there is surface outflow and some quality constituent is in storage, washoff
is simulated using the commonly used relationship:
SOQO - SQO*(1.0 - fXP(-SURO*WSFAC)) (4.6)
where: ^ '
SOQO - wa,shoff of the quality constituent from the land
surface in quantity/acre per interval
SQO - storage of available quality constituent oh the surface
in quant ity/acr^e
SURO - surface outflow of water in in./interval
WSFAC - susceptibility of the quality constituent to washoff
in units of I/in. . . '
EXP - Fortran exponential function .
The storage is updated once a day to account for accumulation and removal which
occurs independent of runoff by the equation:
SQO - ACQOP + SQOS*(1.0 - REMQOP) (4.7)
where: ; .
ACQOP - accumulation rate of the constituent,
quantity/acre per day
SQOS - SQO at the start of the interval
REMQOP - unit removal rate of the stored constituent, per day
The module computes REMQOP and USFAC for this subroutine according to:
REMQOP - ACQOP/SQOLIM , (4.8)
where:
SQOLIM - asymptotic limit for SQO as time approaches infinity
(quantity/acre), if no washoff occurs
and .
WSFAC - 2.30/WSQOP ' . (4.9)
where: '
WSQOP -rate of surface runoff which results in a 90 percent
washoff in one hour, iri./hr
62
-------
Since the unit removal rate of the stored constituent (REMQOP) is computed from
two other parameters, it does not have to be supplied by the user.
As with the potency factor approach, subsurface concentrations which can be input
on a monthly basis are used to define the subsurface contributions.
4.2.2 Animal Waste Procedures
To represent animal waste contributions within the Watershed Model, the CBPO
developed estimates of 'manure acres' for each model segment (see Section 3.0)
which are used to calculate runoff and associated pollutant loads from animal
waste. These areas are included in the Watershed Model as 'impervious' land
surfaces in order to conceptually model a rain-driven source of pollutants with
constant concentrations. The available literature was reviewed to estimate
reasonable hydrologic parameters and concentrations to use in calculating the
waste contributions. The following hydrologic parameters are currently used in
the model:
Retention parameter - 0.35 in.
Roughness factor 0.15
Slope , - 2.0 X Above Fall Line
1.0 X Below Fall Line
Slope Length - 150 ft.
The most critical of these parameters is the retention parameter which specifies
the surface storage to be satisfied prior to runoff. Since evaporation occurs
from this storage, it also determines the total fraction of rainfall that becomes
runoff. The value of 0.35 inches is consistent with the limited available
literature from Edwards et al. (1985, 1983) and Phillips and Overcash (1976) for
paved animal feedlots; this value results in about 70Z runoff of precipitation.
The remaining parameters are consistent with small feedlot size areas, and they
have no significant impact on runoff volumes.
The calculated runoff volumes are multiplied by assumed constant concentrations
to calculate the pollutant loadings to the stream channel from the 'manure
acres'. Initial values for concentrations were derived from a review of the
available literature on animal waste runoff; these initial values were then
adjusted, primarily reduced, during the calibration process based on instream
concentrations and relative contributions from all sources. The .final
concentrations used in both Phase I and Phase II are listed below, along with
annual per acre loadings corresponding to 25 inches of runoff: .
Concentration Annual Loading
(og/D (Ib/ac/yr)
Organic N .300 1698
Ammonia N 40 226
Nitrate N 10 57
Total N 350 1981
Organic P 60 340
Phosphate P 10 57
- 63
-------
Total P 70 397
BOD (ultimate) 700 3962
Actual segment loadings are based on the actual runoff volumes for each segment;
the above loadings based on 25 inches of runoff are shown to .provide some
perspective on the general magnitude of loadings associated with animal waste in
the Watershed Model. These concentrations are considerably lower than those
expected from feedlots, primarily because the manure acres concept does not
represent feedlot areas, manure piles, or manure stocking facilities. . It is
simply a means of accounting for animal waste nutrient contributions, and
allowing for this contribution to be reduced as treatment and storage programs
are implemented (See Section 3.1.5).
4.3 SIMULATION OF CROPLAND AREAS
The primary objectives in applying the AGCHEM modules for simulating nonpoint
nutrient loadings from the cropland areas were as follows:
a. Represent nutrient (nitrogen and phosphorus) balances, including
applications (fertilizer and manure), crop uptake, soil
; transformations, and runoff loadings for the major cropland categories
in each model segment.
b. Provide a mechanism for investigating.and representing sensitivity to
climate variations and agricultural practices.
c. Provide a means of evaluating the water quality impacts of current
nutrient application rates and practices, and projecting impacts of
changes in rates and practices as potential nutrient management
alternatives.
This section provides an overview of the processes simulated by the AGCHEM
modules of HSPF, and discusses the general application procedures used in
applying the modules to the model segments within the Watershed Model. Details
of the equations used are provided in the HSPF User Manual (Johanson et al.,
1984). Following the overview of AGCHEM processes, we discuss some of the key
issues involved in applying the AGCHEM module at the scale of the Watershed Model
segments, including the development of a 'composite crop' representing a weighted
combination of the major crops grown within the Basin, and specification of
current fertilizer and manure application rates, chemical forms, and application
procedures for defining model input. The final subsection, Section 4.3.6,
summarizes the AGCHEM simulation results.
4.3.1 AGCHEM Overview
The AGCHEM modules of HSPF attempt to represent the major nitrogen and phosphorus
processes occurring within the soil profile that determine and control the fluxes
of soil nutrients within the soil/plant/terrestrial environment. Figure 4.6
shows the soil nutrient storages, nutrient application modes, and subsequent
movement of soil nutrients modeled by the AGCHEM module. Runoff, soil moisture,
and sediment values calculated by the corresponding sections of HSPF are used
64
-------
.Nutrient
application
Crop uptake and
denitrification
Application
mode
soil incorporated
surface
applied
Surface
storage and
transformation
infiltration
i
Upper zone
storage and
transformation
percolation
1
Lower zone
storage and
transformation
Nutrients with sedment
(adsorbed organic)
Nutrients in overland flow
(soluble)
Nutrients in interflow
losses to groundwater
1
Groundwater
storage and
transformation
Nutrients in
Groundwater discharge
To
stream
Figure 4/6 .Nutrient storages and transport modeled by the AGOHEM
.module. :
65
-------
by the AGCHEM module to provide the moisture storage and transport values needed
for the nutrient simulation. Fertilizers, animal wastes, plant residues, and
other nutrient inputs (e.g., atmospheric deposition, sludge application) are
applied in their chemical form (i.e., NH<,, N03, P04, Organic N, Organic P) either
as a surface application or incorporated into the soil.
Figure 4.7 shows the AGCHEM conceptualization of the soil profile. Nutrients are
stored in four depth layers: surface zone, upper zone, lower zone, and
groundwater zone. . The depths of each zone are defined by input parameters
estimated by the model user. The shallow surface layer is a continuous mixing
zone which may be defined functionally as the zone of.interaction of surface
applied chemicals, from which surface runoff, sediment erosion, and associated
chemicals are transported to a waterbody. The surface zone depth is typically
estimated at about one centimeter, although values ranging from 0.5 to 2.0 cm
have been used in specific applications with AGCHEM and other agricultural runoff
models.
The upper zone extends from the bottom of the surface zone to a depth often
ranging from 10 to 20 cm. This usually corresponds to the depth of major tillage
operations and/or incorporation of applied nutrients. It also defines the depth
used to calculate the mass of soil and nutrients through which interflow and
percolating water pass to reach the lower zone regardless of the method of
chemical application.
The lower zone is the primary source of plant evaporation, and it regulates the
amount of soluble, chemicals that can be introduced into groundwater since the
chemicals must pass through this layer. For agricultural applications, the lower
zone depth is typically set to the maximum depth of the crop root zone, usually
ranging from 50 to 150 cm.
The groundwater layer represents the depth of shallow groundwater that actively
contributes baseflow -to the stream channel. It can also be visualized as a
shallow mixing depth within the surface aquifer that controls the chemical
transformations and associated contributions to baseflow concentrations. In
reality, this depth is highly variable and difficult to define; for watershed
modeling purposes, depths of 1. to 3 m are typically used.
In Phase II, the soil depths for these layers were set to the following values
for all model segments:
Bottom Depth
Soil Layer. . Thickness from Soil Surface
Surface 1 cm .1 cm
/ Upper Zone 14 cm 15 cm
Lower Zone 105 cm 120 cm
Groundwater Zone 152 cm 272 cm
Although variations in the depth of the groundwater zone for each model segment
may be appropriate, and should be considered in any future studies, the use of
a constant value was consistent with the both the objectives and resources
available for the Phase II effort.
66
-------
Outflow to
Stream with:
Layer:
(Ti
Sediment,
Surface runoff
Interflow
Ground
water
Lower
?^???T5?JC??!?A5l!3S?^5^5iteซ5?!iFT6'???rS?
Ground
water
Figure 4.7 Soil profile representation by the AGCHEM Module.
-------
4.3.2 Soil Nutrient Processes Modeled by AGCHEM
Within the AGCHEM module, the NITR and PHOS sections perform the reactions and
transformations of nitrogen and phosphorus compounds, respectfully,, i'n the soil
profile as a basis for predicting the nutrient content of agricultural runoff.
The transformations, nutrient storages, and reaction rates considered by NITR and
PHOS are shown in Figure 4.8. The nutrient simulation primarily assumes first-
order reaction rates and is derived from work by Mehran and Tanji (1974), and
Hagih and Amberger (1974). The AGCHEM module is a translation and enhancement
of the soil nutrient capabilities of the Agricultural Runoff Management (ARM)
Model (Donigian and Crawford, 1976b; Donigian et al., 1977; Donigian and Davis,.
1978). The processes simulated include immobilization, mineralization,
nitrification/denitrification, plant uptake, and adsorption/ desprption (for
which an equilibrium isotherm option is available).
* '
The nutrient module simulate both nutrient transformations and movement in the
watershed by water or sediment. Transformations of nutrients determine the
nutrient forms in each soil layer and their resulting susceptibility to movement.
Nutrient transport processes are simulated only to the extent that is needed to
predict r'unoff quality and quantity. The model does not consider the generally
secondary movement of soil chemicals by concentration and thermal gradients.
However, it does model the lateral and downward transport of chemicals by water
from the soil zones. Lateral transport of nutrients towards the stream can occur
from the surface and upper zone storages. Inflow of nutrients to the groundwater
zone and nutrient transformations in the groundwater are modeled as a basis for
predicting the outflow of dissolved constituents with groundwater flow.
The diagrams in Figure 4.8 show the transformation pathways, storages, and names
of the reaction rate input parameters. Reaction rates are input on a per day
basis for each soil layer. Nitrite (NO2) transforms so quickly in most
agricultural soils that it is not considered separately. The adsorbed phase
represents the nutrients in a complex form along with those adsorbed on the soil:
The plant uptake rates (KPL) are input monthly as a function of the stage of crop
growth. These monthly uptake parameters are adjusted to represent the crop
uptake of N and P from the soil storages and to distribute it. throughout the
growing, season. .
Soil temperature is presently the only environmental factor that is modeled as
affecting the reaction rates, with the exception that the transformations are
stopped entirely at very low moisture levels as often occurs in the surface
layer. Soil temperatures for the surface and upper zones are determined by
regression equations based on air temperature, while the daily soil temperature
for the lower zone and groundwater is interpolated from average monthly input
values. The input first-order rate parameters are defined as the maximum, or
optimum, rates at 35 degrees C. .Using the calculated soil temperature for each
soil layer in each model time-step, the input rate parameter is corrected for
soil temperatures below 35 degrees C by the generalized equation:
68
-------
N2
PLNT-N
KDNI
KPLN
NO3
(+N02)
KNI
NH4 - A
KADAM
KDSA
NH4 - S
Key
KAM
N - Nitrogen
P - Phosphorus
A - Adsorbed
S - Solution
K - Reaction Rate
Parameters
* - Linear Equilibrium
Sorptlon (optional)
KIM AM
ORG -N
KIMNI
Nitrogen transformations in NITR module
P04 - A
KDSP*
tfARP*
PLNT - P
4
PO4
KPLP
1/IK4D '
- S
ซ n, nnp P
kTRXP
Phosphorus transformations in PHOS module
Figure 4.8 Nutrient transformations simulated by the AGCHEM Module.
,69
-------
KK - K*TH**(TMP-35.0) (A.10)
where: .
KK - temperature corrected first order transformation rate in
units-of per simulation interval
K - optimum first order reaction rate parameter
TH - temperature coefficient for reaction rate correction
(typically about 1.06)
IMP - soil layer temperature in degrees C
When temperatures are greater than 35 degrees C, the rate is considered optimum,
that is KK is set equal to K. When the temperature of the soil layer is below
4 degrees C or the layer is dry, no biochemical transformations occur.
The corrected reaction rate parameters are determined every biochemical reaction
interval and multiplied by the respective storages as shown in Figure 4.8 to
obtain the reaction fluxes. Plant uptake can vary monthly and can be distributed
between nitrate and ammonium by a user-specified input parameter which designates
the fraction of plant uptake from each species of N.
Other factors deemed as having a constant influence during the simulation period,
such as soil pH, are represented by adjustments to the input reaction rates.
Also, all input parameters are specified separately for each soil layer; thus,
variations in characteristics of the soil prof ile. can be accommodated by using
different reaction rates for the different soil layers.
Although HSPF has been applied to hundreds of watersheds across the United States
and abroad, the detailed agricultural runoff simulation provided by'the ACCHEM
modules has not had the same degree of application experience. However, initial
testing and application was performed on. field sites in Georgia and Michigan as
part of the ARM model development work (Donigian and Crawford, 1976; Donigian et
al., 1977), and subsequent applications including both field and watershed-scale
sites have been conducted in Iowa (Donigian et al., 1983; Imhoff et al., 1983),
Nebraska (Gilbert et al., 1982), Florida (Nichols and Timpe. 1985), and Tennessee
(Moore et al., 1988).
( '
4.3.2.1 General Limitations of AGCHEM
As with any model, AGCHEM has a number of inherent limitations in representing
soil nutrient processes that need to be discussed in order to fully appreciate
the model results and their use in the Phase II effort. Some limitations are due
to current capabilities of the HSPF code, others are due to the scale of the
Chesapeake Bay application, while others are a reflection of the current state-
of-the-art of modeling soil nutrient processes. Each limitation and its
potential impact is discussed below:
a. Ammonia volatilization is not currently simulated by the AGCHEM model..
Volatilization can be a significant loss of nitrogen when applied as
manure or as urea-based fertilizers. In order to account for this
loss, ammonia losses from manure applications were estimated
separately .and deducted from the total manure N application (see
Section 4.3.4.4). Subsequent ammonia losses, i.e., following the
70
-------
applications, from the soil were represented by increased
immobilization/fixation. Although this did not eliminate the
volatilized N from the soil, it did reduce soil ammonia values and
subsequent runoff, and the resulting increase in soil organic N was
minimal.
b. Urea is not modeled as a separate form of nitrogen in AGCHEM.
Consequently, the urea component of .fertilizer was assumed to be
readily hydrolyzed to ammonia so that all nitrogen fertilizer in the
urea form was input to the model as ammonia (see Section 4.3.4). This
is expected to have very little impact on the simulation results, and
would only affect storms occurring within a few days of a fertilizer
application.
c. ' Plant uptake is represented as a first-order process in AGCHEM, based
on the soil nutrient-storages in the various soil zones. As described
above, monthly rates are input by the user for each soil zone to mimic
the time-variation in rpot development and associated plant nutrient
uptake from the various soil zones. The input*rates are corrected for
soil temperature and uptake is stopped when moisture levels are zero;
this primarily occurs in the surface layer. Thus, uptake is a direct
function of soil nutrient storages arid application rates; plant
growth, root development, stresses associated with environmental
conditions, etc. are not represented. Consequently, uptake rates are
primarily determined through calibration to expected annual crop
uptake (see Section 4.3.S). The rates and resulting annual uptake
need to be re-evaluated whenever application rates are changed in
order to ensure that'crop needs are satisfied.
d.. The soil phosphorus cycle and processes represented in AGCHEM are
considerably simplified compared to the complex reality of how
phosphorus behaves in the soil environment. Chemical fixation and
precipitation reactions that produce iron and aluminum complexes with
phosphate are not; represented, and the distinction between sorbed
phosphate and organic. P is difficult to model, and somewhat arbitrary,
since these forms'are not usually reported in field studies .due to the
difficult analytical problems they present. In effect, the sorbed P04
in the model represents an 'exchangeable P04' which controls the
amount in solution, while the organic P effectively represents both
organic P and tightly-bound P0t not available for desorption. Section
4.3.5.3 discusses the estimation of model parameters based on this
view of the phosphorus state variables. Very few agricultural models
currently available attempt to represent the complex phosphorus cycle
in soils; further model development and testing work is needed.
e. Only sediment-bound organic N and P are simulated by AGCHEM; no
dissolved organics are modeled. This is generally considered to be a
reasonable approximation for surface runoff contributions, where
dissolved organics are usually small components of the total nutrient
load. In Phase II, a BOD component from the cropland areas is
included in the model simulations providing a source of dissolved
organic material for the instream processes.
, 71
-------
f. Croo rotations are not explicitly represented by AGCHEM in the
Watershed Model. AGCHEM is designed to represent a constant annual
land cover/land use pattern for each model segment throughout the
simulation period. Consequently, crop rotations and associated
practices at the field scale are not directly represented. The
implicit assumption is that, at the scale of a model segment ranging
from 100 sq. mi: to 1000 sq. mi. or more, rotations, within *. segment
will offset each other so that the total area in each land use and
crop will remain approximately constant throughout the period. That
is, although crops and practices on an individual field may change
from year to year, the net change in crops and practices within a
model segment will be minimal.
Other limitations of AGCHEM, primarily related to the large scale of the
Watershed Model application, are discussed below in Sections 4.3.4 and 4.3.5.
Recommendations for. improvements to AGCHEM, based on the above limitations, and
refinements of the Watershed Model application are discussed in Section 4.3.7.
4.3.3 Composite Crop Representation . . '
The Phase II Watershed Model includes 63 individual model segments, comprised of
34, segments above the Fall Line (AFL) and 29 segments below the Fall Line (BFL).
The AFL segments include simulation of nonpoint loadings from all the land .use
categories plus the instream flow and water quality constituents, while the BFL
segments provide the nonpoint loadings directly to the estuaries and main Bay
'segments modeled by the 3-D Water Quality Model. In each segment .of the
Watershed Model, each of the eight land use categories (pervious and impervious)
are modeled for flow, sediment, and all nonpoint loadings; over 500 combinations
of model segments and land use categories (i.e., 63 segments times 8 land use
categories equals 504 combinations) are simulated. This is a significant
computational burden even with the modern increases in computer processing speed.
Moreover, the computations increased in Phase II due to the addition of the 'hay'
land use category and the use of the more detailed computations of the AGCHEM
module for all the cropland categories. , .
Due to this computational burden and limitations inherent in the. spatial
representation of land use within a model -segment at the scale of the Bay
drainage, it was necessary to develop the concept of a 'composite crop' in order
to evaluate segmentwide parameter values for the Conventional Tillage and
Conservation Tillage categories. The 'composite crop' concept is simply a means
of calculating selected parameter values for a model segment by weighting .crop-
specific values by the relevant percentages in each crop category. The crop*
percentages, or crop distributions, for both tillage land use categories were
developed as described in Section 3,1.10; the percentages shown in Table 3.2 were
used to calculate the segment parameters.
The key model parameters affected by the cropping patterns include monthly land
surface cover (i.e., the COVER parameter), nutrient application rates, and
expected composite crop nutrient uptake values. This section discusses the
calculation of COVER values for the model segments. Section 4.3.4 (below)
discusses the development of the nutrient application rates and the associated
72
-------
influence of che composite crop approach, while Section 4.3.5 describes the
expected crop uptake values and how they were used in the application of the
AGCHEM simulations.
The COVER parameter represents the fraction of the land surface protected from
direct raindrop impacts by either crop canopy or residue on the surface. Its
primary influence in the AGCHEM model is on the calculated sediment loadings for
each land use category within each model segment. Monthly COVER values are input
to the sediment model representing the land surface cover on the first day of
each month; values for days during the month are linearly interpolated from the
neighboring monthly values. Figure 4.9 shows the division of the Bay drainage
area into three regions for definition and variation of the monthly COVER values;
the COVER values for each major crop for both Conventional and Conservation
Tillage, and for the hay cropland are shown in Table 4.4. Figures 4.10 and 4.11
show the weighting procedure used to calculate the composite crop COVER values
for model segment number 100 (within the Juniata Basin) for the Conventional
(CNT) and Conservation (CST) Tillage categories, respectively.
The monthly COVER values for CNT, CST, and Hay for all model segments are
provided in Appendix E.ll, and were estimated as follows:
a. Monthly COVER values for each of the four major crop categories and
hay were estimated for Region 2 in conjunction with the Soil
Conservation Service (J. Hannawald, 1990, Personal Communication).
b. Values for Region 1 and Region 3 were estimated by adjusting the
expected planting dates two weeks earlier for Region 3 and two weeks
later for Region 1, based on data from U.S.D.A. Handbook No. 628
(USDA.1984) and from State representatives on the CBP Nutrient
Reduction Task Force (NRTF). The resulting monthly values were
reviewed by. the NRTF members.
c. The composite crop COVER values for CNT and CST were then calculated
using the corresponding crop distributions for each tillage practice,
as shown in Figures 4.10 and 4.11 for CNT and CST in model segment
100. The hay COVER values are used directly by the model since it is
represented as a separate land use category. The final numbers were
reviewed by the NRTF.
In reviewing the COVER values, please note the following:
1. As noted above, all COVER values are for the first day of each month,
and values between months are linearly interpolated.
2. There are only slight differences between the CNT and CST values for
Corn-Silage, primarily in March and April, since the stover is removed
in both practices.
3. The values for Grains in June-September are 'dashed' because they are
not included in the weighting, under the assumption that they will be
followed by one of the other crops.
73
-------
Crop Cover
Region B ounda r i e s
REGION 1
REGION 2
REGION 3
Figure 4.9 Crop COVER region boundaries.
74
-------
TABLE 4.4 CROP COVER VALUES FOR ALL REGIONS AND PRACTICES
COVER VALUES FOR CONVENTIONAL TILLAGE
CORN GRAIN
REGION 1
REGION 2
REGION 3
JAN.
50
50
50
FEB.
45
45
45
MAR.
0
0
0
APR.
0
0
5
MAY
5
10
45
JUN.
45
50
70
JULY
70
75
88
AUG.
88
93
93
SEP.
93
93
93
OCT.
80
85 ,
90
NOV.
55
70
85
DEC.
55
55
55
CORN SILAGE .
REGION 1
REGION 2
REGION 3
SOYBEAN
REGION 1
REGION 2
REGION 3
GRAINS
REGION 1
REGION 2
REGION 3
JAN.
5
5
5
JAN.
10
10
10
JAN.
40
45
50
FEB,.
4*
4
4
FEB.
5
5
5
FEB.
40
50
65
MAR.
0
0
0
MAR.
4
3
6
MAR.
60
65
80
APR.
0
0
5
APR.
5
4
0
APR.
80
75
90
MAY
5
10
45
MAY
0
0
5
MAY
90
90
90
JUN.
45
50
70
JUN.
3
5
60
JUN.
.. .
--
JULY
70
75
88
JULY
60
65
80
JULY
...
, -- .
-I-
AUG.
88
93
93
AUG.
82
87
87
AUG.
--
-'.
-
SEP.
93
93
93
SEP.
87
87
87
SEP.
--
--
OCT.
10
10
20
OCT.
65
75
80
OCT.
35
40
45
NOV.
8
8
15
NOV.
15
15
15
NOV.
40
45
50
DEC.
6
6
10
DEC.
15
15
12
DEC.
40
45
50
COVER VALUES FOR CONSERVATION TILLAGE
CORN GRAIN -
JAN. FEB. MAR. APR. MAY
REGION 1 65 60 55 50 50
REGION 2 65 60 55 50 55
REGION 3 65 60 55 50 60
JUN. JULY AUG.
60 70 88
65 75 93
70 88 93
SEP. OCT. NOV. DEC.
93 85 7.5 70
93 90 80 70
93 93 85 70
CORN SILAGE
JAN. FEB. MAR. APR. MAY
REGION 1 53 335
REGION 2 5 44 4 10
REGION 3 5 5 5 5 45
JUN. JULY AUG. SEP. OCT. NOV. DEC.
45 70 88 93 10 8 6
50 75 93 93 10 8 6
70 88 93 93 20 15 10
75
-------
TABLE 4.4 CROP COVER VALUES FOR ALL REGIONS AND PRACTICES (continued)
SOYBEAN
REGION 1
REGION 2
REGION 3
JAN.
10
15
20
FEB.
9
12
15
MAR.
8
10
10
APR. MAY JUN.
8 8 15
10 10 20
10 15 60
JULY
60
65
87
AUG.
87
87
87
SEP.
87
87
87 .
OCT.
60
75
80
NOV.
20
25
30
DEC
15
20
25
GRAIN
REGION 1
REGION 2
REGION 3
JAN. FEB. MAR. APR. MAY
50 55^ 60 75 90
55 60 70 80 90
60 65. 80 90 90
JUN. JULY AUG.
SEP. OCT. NOV. DEC.
- - 45 50 50
50 55 55
55 60 60
COVER VALUES FOR HAY
HAY
REGION 1
REGION 2
REGION 3
JAN. FEB. MAR. APR. MAY
85 85 88 65 65
85 85 88 93 93
85 .85 88 93 93
JUN. JULY AUG.
93 93 93
93 65 65
93 65 65
SEP. OCT. NOV. DEC.
88 85 85 85
73 85 85 85
73 85 85 85
76
-------
Percent Coverage - Conventional Tillage
Juniata Segment 100
Percent Coverage
Corn-gr.
Corn-sll.
Soybeans
Grains
Composite
% of o
Cropland
Corn-gr.
Corn-sil.
Soybeans
Grains
Composite
Time (Months)
Figure 4.10 Composite crop COVER values for Segment 100 - Conventional.Tillage.
-------
-o
oo
Percent Coverage - Conservation Tillage
Juniata Segment 100
Percent Coverage
Corn-gr.
Corn-sil.
Soybeans
Grains
Composite
% of
Cropland
AugSep
MayJun
Corn-gr.
Corn-sil.
Soybeans
Composite
Time (Months)
Figure 4.11 Composite crop COVER valu *or Segment 100 - Conservation Tillage
'/'
-------
4. We've reduced the maximum values (e.g., 96 and 100) which are commonly
used for some crops to account for borders, non-cropped areas, spatial
variability (e.g., dry spots), etc.
5. Values on the scale of the size of the model segments will likely be
somewhat different than those measured in a single field since we need
to use average values that are adjusted for variations in practices,
e.g., 20% fall plowing in CNT, or 75Z of CST areas leave residues.
6. Differences of a few percentage points (i.e., 3 to 5) will have little
or no significant impact on the simulation; the COVER values primarily
affect the simulated sediment loading rates. Differences of 10 to 20
points for specific crops in specific months could have a significant
impact if the changes also result in similar differences in the
calculated*Composite Crop COVER values.
4.3.4 Development of Model Segment Nutrient Application Rates
ซ
The 'key' element in the application of the AGCHEM model to the cropland areas
of the model segments was the development of the nutrient (fertilizer and manure)
application rates used as input to the model. This information was developed
over a period of 6 months with continuous interaction and discussions among Aqua
Terra staff, CBPO, and the NRTF members. The outcome of this effort was the
specific values of fertilizer and manure nutrients, including composition and
methods of application, that were used to'model the' expected nutrient balances
for each cropland category, Moreover, this effort also demonstrated the
relatively high degree of uncertainty in the information available in each state
to categorize these rates. This is a critical issue since any efforts to reduce
nutrient loadings from agricultural cropland must consider current application
rates and procedures as a starting point. In other words, application rates are
the means of implementing nutrient reduction from nonpoint source loadings from
cropland. Thus, accurate,information on current fertilizer and manure nutrient
applications is essential to any management approach attempting to reduce
nutrient loadings from/agriculture.
Development of the model input application rates involved aggregating input from
each of the major CBP states (i.e., PA, MD, VA), developing assumptions
appropriate to the scale of the Bay drainage, and calculating.rates corresponding
to the 'composite crop' discussed in Section 4.3.3 (above). Figure 4.12 is a
flowchart that demonstrates the information used and steps performed in
calculating the model segment nutrient application rates. The items in 'boxes'
are the input data and assumptions used, while the 'ovals' represent the
calculations .performed in developing the input, rates. The major data needs and
issues involved in these calculations included the following:
a. Fertilizer and manure application rates, procedures, and timing for
each major crop, plus atmospheric deposition estimates.
b. Crop distributions for Conventional and Conservation Tillage.
c. Composition (i.e., organic and inorganic fractions) of fertilizer and
manure nutrients.
79
-------
Development of Model
Input Nutrient Applications
Application
and Volatil-
ization Loss
Assumptions
Crop
Distribution
Percent of
Cropland
Receiving
Fertilizer/
Manure
Fertilizer
and Manure
Composition
Nutrient
Application
Timing and
Methods
Application Rates
for Each Crop
CBLO
Atmospheric
Deposition
Fertilizer
Adjust N
for Volatilization
Calculate Composite Rates and
Crop Components
Calculate Rates Weighted
by Percent of Cropland and
Associated Rates
Calculate Composition of
Final Fertilizer and Manure Rates
Calculate Timing and
Placement of Applications
Model
Segment
Input
Values
Figure 4.12 Flowchart for development of model input nufrient application rates.
80 .
-------
d. Application/volatilization losses of manure nitrogen.
e. Model representation of application procedures and timing.
The following sections discuss each of these issues, except the crop
distributions which were discussed in Section 3.1.10, followed by a description
of the Lotus spreadsheet used to calculate the application rates and input values
for each model segment. Appendix C, provides copies of the spreadsheets for all
model segments; the spreadsheets are also available on diskette from CBPO.
In reviewing the following sections, the reader should note that many of the
problems and issues of applying AGCHEM (or any other detailed agricultural runoff
model) at the scale of the model segments would not exist if the model was being
applied at the field-scale. At the field level, farmers would be able to specify
(or data could be obtained to define) fertilizer and manure compositions, timing
and methods of applications, application rates for individual crops, etc. For
model segments with croplands approaching hundreds of square miles, field and
farm level practices must be adapted to define what is 'typical' or
. 'representative' of practices throughout the entire model segment. Much of the
effort in developing the application rates and procedures involved the definition
of representative conditions and practices, and resolution of conflicting
information from individual agencies in order to prepare consistent and
compatible model input for all portions -of the Bay drainage.
4.3.4.1 State 'Supplied application rates
Table 4.5 lists the fertilizer application rates for each major crop category for
each model segment which were developed by the state representatives on the NRTF.
The corresponding manure application rates are shown in Table 4.6. while Table
4.7 shows the estimated percentages for each crop category receiving applications
of fertilizer only, manure only, and both fertilizer and manure. -
As noted above, this information was initially supplied by the states and
subsequently refined in NRTF meetings and telephone/fax communications in order
to clarify how the rates (and percentages) would be used and interpreted within
the framework of the AGCHEM modeling calculations. In reviewing this
information, please note the following:
a. All nitrogen (N) and phosphorus (P) rates are in terms of the
elemental nutrients, i.e., N and P. P is calculated as 0.44 x
b. Application rates are usually the same for corn-grain and corn-silage,
with a few exceptions.
c. The rates for corn, soybeans, and grains were used to calculate the
weighted rate for the composite crop, while the hay rate was used
unchanged (except as discussed below for manure N volatilization) as
input to the 'hay' cropland category.
81
-------
TABLE 4.5 FERTILIZER APPLICATION RATES SUPPLIED BY THE NRTF (Ib/ac)
MODEL
SEGMENT
10
20
30
40
50
60
70
80
90
100
110
120
140
160
170
175
180
190
CORN
N
120
120
120
120
120
120
120
90
95
95 *
120
120
120
140/31a
150/105b
150/105b
142
150/105b
P
24
24
24
24
26
26
26
37
59
59
44
44
44
22/0*
26/22b
35/26d
* ซ_
26/22b
35/26d
30
26/22b
BEANS
N
34
34
34
34
34
34
34
34
34
34
20
20
20
20
20/18b
20/18b
25
20/18b
P
8
8
8
8
8
8
8
8
8
8
18
18
18
26 b
29/22b
29/22b
19
29/22b
GRAINS
N
34
34
34
34
34
34
34
34
34
34
34
34
34
100
100/80b
100/80b
76
100/80b
P
8
8
8
8
8
8
8
8
8
8
18
18
18
0
35/26b
35/26b
2
35/26b
HAY
N
60
60
60
60
60
60
60
15
0
0
20
20
20
0
36/20b
36/20b
7
36/20b
P
29
29
29
29
26
26
26
35
16
16
26
26
26
79
3Q/1&
SO/IS1*
63
30/16**
200
210
220
230
235
240
250
260
265
270
280
290
300
310
330
340
35/26
15Q/105b 26/22b 20/18b 29/22b 100/80b 35/26b 36/20b 30/161
35/26^
144
148
150
150/160ฐ
29
23
22
22/35ฐ
23
20
20
20ฐ
35
31
24
24/29ฐ
43
82
100
100ฐ
21
29
21
31/26ฐ
5
24
60
36/30
39
33
21
30/25ฐ
31/44ฐ
160/150ฐ 35/22ฐ 20ฐ 29/24ฐ
1
100e
26/31c
30
25
150
150/160ฐ
150
150/105b
150ฐ
150/160ฐ
150
150
160/49a
160/50a
22
31d
22/35ฐ
31/44d
26
35d
26/22b
35/26d
22/26ฐ
31/35d
22/35ฐ
31/44d
22
22
18/0a
18/0a
20
20ฐ
20
20/18b
20ฐ
20ฐ
20
20
io
20
24
24/29ฐ
29
29/22b
24/29ฐ
24/29ฐ
24
24
13
13
100
100ฐ
100
100/80b
100ฐ
100ฐ
100
100
50
50
31
31/26ฐ
35
35/26b
31/35ฐ
/
31/26ฐ
31
31
13
13
36
36/30ฐ
36
36/20b
30/36ฐ
30/36ฐ,
30
30
30
30 ,
30
30/25ฐ
30
30/16**
1
25/30ฐ
25/30ฐ
25
25
13
13
82
-------
TABLE 4.5 FERTILIZER APPLICATION RATES SUPPLIED BY THE NRTF (Ib/ac)
(continued)
MODEL
SEGMENT
Anacostia
Baltimore
Bohemia
Chester
Coast_l
Coast_4
Coast_5
Coast_6
Coast_8
Coast_9
Coast 11
Chicka.
Choptank
Elizabeth
Gr. Wicomico
Gunpowder
James
Nansemond
Nanticoke
Patapsco
Patuxeht
Pocomoke
Potomac
Occoquan
Rappahannock
Severn
Wicomico
Wye
York
CORN
N
143/25a
160
120
130
130
160
160
180
160
160
158
160/150C
130
160
160
160
160/150C
160
145
160/43*
135/21a
145
144
150
160
160
145
130
160/150a
P
25/Oa
18
18
16
16
35
18
42
35
44d
35j
44d
43
35/22c
44/31d
16
35
44d
3*.
44d
18
35/22c
44/31d
3*.
44d
30
18/0a
35/6a
30
34
22H
31d
35
44d
18
30
16
35/22a
BEANS
N
\ 20
20
20
20
20
20
20
0
20
20
7
20C
20
20
20
20
20C
20
17
20
7
17
13
20
20
20
17
20
20a
P
44
13
18
35
35
29
13
26
29
29
23
29/24c
35
29
29
13
29/24ฐ
29
13
13
11
13
18
24
29
13
13
35
29/24*
GRAINS
N
45
50
80
60
60
100
50
30
100
100
26
100C
60
100
100
50
100C
100
74
50
90
74
95
100
100
50
74
60
100a
P
26
13
18
18
18
26
13
26
26
26
23
26/31c
18
26
26
13
26/31c
26
20
13
20
20
25
31
- _ '
26
13
20
18 -
26/31a
HAY
N
0
30
20
30
30
30
30
0
36
30
7
30/26C
30
30
36.
30
30
30
30
30
30,
30
30
36
. 36
30
30
30
30/36a
P
40
13
26
26
26
25
13
31
25
25
27
25/30=
26
25
25
13
25
25
26
13
13
26
19
.30
25
13
26
26
25/30*
44/31ฐ
a - Two fertilizer rates used, manure applied .with both fertilizer rates;
b - Two fertilizer rates used, manure applied with second fertilizer rate only;
c - Two fertilizer rates used, in cases where the second rate separated by a
slash (/) is not shown means it is same as the first rate;
d -.Phosphorus fertilizer application for corn-silage (Nitrogen application is
same as Corn Grain)
.83
-------
TABLE 4.6 MANURE NUTRIENT APPLICATION RATES SUPPLIED BY THE NRTF (Ib/ac)
MODEL
SEGMENT
10
20
30
40
50
60
70
80
90
100
110
120
140
160
170
175
180
190
200
210
220
230
235
240
250
260
265
270
280
290
300
310
330
340
CORN
GRAINS
HAY
N
N
N
100
100
100
100
180
180
180
180
150 *
150
180
180
180
109
265
265
107
265
265
104
37
0
0
o
0
0
0
265
0
0
0
0
111
110
18
18
18
18
32
32
32
32
26
26
32
32
32
27
60
60
23
60
60
22,
9
0
0
0
0
0
0
60
0
0
0
0
26
25
140
140
140
140
140
140
140
140
140
140
180
180
180
0
265
265 .
121
265
265
113
37
0
0
0
0
0
0
265
0
0
0
0
111
110
25
25
25
25
25
25
25
25
25
25
32
32
32
0
60
60
25
60
60
24
9
0
0
0
0
0
0
60
0
0
0
0
26
25
140
140
140
140
140
140
140
140
140
140
180
180
180
109
265
265
121
265
265 .
113
37
0
0
0
0
0
0
265
0
0
0
0
111
110
25
25
25
25
25
25
25 .
25
25
25
32
32
32
27
60
60
25
60
60
24
9
0
0
0
0
0
0.
60
0
0
0
0
26
25
0
0
0
0
0
0
0
0
150
150
180
180
180
109
50
50
71
50
50
81
37
0
0
0
0
, o
0
50
0
0
0
0
111
110
0
0
0
0
0
0
0
0
26
26
32
32
32
27
16
16
16
16
16
18
9
0
0
0
0
0
0
16
0
0
0
0
26
25
84
-------
TABLE 4.6 MANURE NUTRIENT APPLICATION RATES SUPPLIED BY THE NRTF (Ib/ac)
(continued)
MODEL
SEGMENT
Coast_l
Bohemia
Chester
Wye
Chop tank
N ant i coke
Wicomico
Pocomoke
Coast_4
Coast_ll
Coast_6
Gunpowder
Baltimore
Patapsco
Patuxent
Severn
Coast_5
Potomac
Anacostia
Occoquan
Rappahannock
Coast_8
Gr. Wicomico
York
James
Chtcka.
Nansemond
Elizabeth
Coast 9
CORN
N
173
109
155
170
197 ''
245
241
230 *
0
135
121
109
124
117
114
126
110
67 .
118
.0
0
0
0
0
0
0
0
0
0
P
36
23
38
37
40
51
51
49
0
27
26
25
27
28
29
23
27
19
31
0
0
0
0
0
0
0
0
0
0
BEAN'S
N
173
109
155
170
197
236
246
231
0
135
121
109
124
117
114
126
110
67
118
0
0
0
0
0
0
6
0
0
0
P
36
23
38
37
40
50
52
50
0
27
26
25
27
28
29
23
27
19
31
0
0
0
0
0
0
6
0
0
0
GRAINS
N
173
109
155
170
197
244
238
227
0
135
121
109
124.
117
114
110
110
67
118
0
0
0
0
0
0
0
0
0
0
P
36
23
38
37
40
52
51
49
0
27
26
25
27
28
29
25
27
19
31 ,'
0
0
0
0
0
0
0
0
0
0
HAY
N
30
20
, 155
170
205
252
253
238
0
135
121
109
124
117
114
126
110
67
118
0
0
0
0
0
0
.0
0
. 0
0
P
26
26
38
37
42
53
54
51
0
22
26
25
27
28
29
23
27
19
31
0
0
0
0
0
0
0
0
0
0
85
-------
TABLE 4.7 PERCENTAGE OF CROPLAND RECEIVING FERTILIZER AND/OR MANURE AS
SUPPLIED BY THE NRTF
CROP PERCENT (%)
MODEL
SEGMENT
10
20
30
40
50
60
70
80
90
100
110
120
140
160
170
175
180
190
200
210
220^
230
235
240
250
260
265
270
280
290
300
310
330
340
CORN
F M
70 10
70 10
70 10
70 10
70 10
70 10
70 10*
70 15
70 15
65 15
45 25
45 25
4.5 25
77
85
85
70
49
51
68
93
100
79
21b
97
3b
100
93
7b
100
91
6b
98
2b
94
6b
100
100
65
80
F+M
20
20
20
20
20
20
20
20
20
20
30
30
30
12
15
15
30
51*
49a
32
7
3a
30
5a
16
4a
BEANS
F M
70 10
70 10
70 10
70 10
70 10
70 10
70 10
60.15
65 15
65 15
45 25
45 25
45 25
100
85
85
73
49
51
76
94
100
79^
21b
97
3b
100
93
7b
100
91
6b
98
2b
94
6b
100
100
70
84
F+M
20
20
20
20
20
20
20
25
20
20
30
30
30
15
15
27
51a
49a
24
6
3a
30
16
GRAINS
F M-
80
80
80
80
80
80
80
80
80
80
80
80
80
64 24
85
85
70
49
51
64
91
100 ""
79
21b
97
3b
100
93
' 7b
100
91
6b
98
2b
94
6b
100
100
60
75
F+M
20
20 v
20
20
20
20
20
20
20
20
20
20
20
12
15
15
30
51a
49a
36
9
3a
30
10a
16
9a
HAY
F
50
50
.50
50
50
50
50
50
50
50
50
50
50
30
89
89
30
58
58
30
30
100
79
100
100
91
9b
100
91
8b
44
56b
97
3b
100
100
30
30
M F+M
30
30
30
30
30
15
lla
lla
22
42a
42a
24
6
21b
la
<
30
16
86
-------
TABLE 4.7 PERCENTAGE OF CROPLAND RECEIVING FERTILIZER AND/OR MANURE AS
SUPPLIED BY THE NRTF (continued)
MODEL
SEGMENT
Anacostia
Baltimore
Bohemia
Chester
Chickahominy
Chop tank
Coast_l
Coast_4
Coast_5
Coast_6
Cpast_8
Coast_9
Coast_ll
Elizabeth
Gr. Wico.
Gunpowder
James
Nanticoke
Nansemond
Occoquan
Patapsco .
Patuxent
Pocomoke
Potomac
Rappahannock
Severn
Wicomico
Wye
York
CORN
F M
77
83
96
91 2
73
b*
ซ,
84
92 1
100
87
97
100
100
26 9
100
100
71 4
73
27b
31
100
100
82
91
19
94
90w
10b
65
26
85 2
85
15b
F+M
20
3a
17
4
7
16
7
13
3
65
25
69
16
2a
8
1*
81
6
35
74
13
CROP PERCENT (%)
BEANS
F M F+M
69 11 20
2 17
1 4
5 7
81
95
88
73h
27b
80 20
92 1 7
100
87 13
96 13
100
100
'26 9 65
100
100
56 19 25
22
100.
100
77
92
21
97
90
10b
65
31
86
85
15b
78
7 16
8
79
3
35
69
13
GRAINS
F M F+M
69 11 20
2
1
81
95
89
27b
84
92 1
100
87
96 1
100
100
39
100
100
68 7
73h
27b
30
100
100
82 16
90 2
15
94
90
10b
9
2
75
23
77 10
85
15b
17
4
16
7
13
3
61
25
70
4
8
85
6
16
75
13
HAY
F M F+M
30 20
30 17
30 4
30 7
27b
30 14
30 4
100
30 13
30 3
J.QO
100
38 55
100
100
30 25
100
30 34
100
100
30 16
30 16
3
4
30
63
9V
10b
30 35
30 61
30 14
85h
15b
87
-------
TABLE 4.7 PERCENTAGE OF CROPLAND RECEIVING FERTILIZER AND/OR MANURE AS
SUPPLIED BY THE NRTF (continued)
CROP PERCENT (%)
MODEL CORN BEANS GRAINS
SEGMENT F M F+M F M F+M F M F+M
SEGMENTS WHERE PERCENTAGES WERE DIFFERENT FOR CONSERVATION TILLAGE
160 78 12 100 40 18 12
175
180
210
220
330
340
Baltimore
Chester
Chop tank
Gunpowder
Nanticoke
Patapsco
Patuxent
Pocomoke
Severn
Wicimico
83
66
70
93
65
81
83
91
84
72
32
83
92
19
65v
5
10a
17
, 34
30
7
30
5*
16
3a
17
2 7
16
3 ,25
68
16
la
8
81
35
95
83
73
76
94
70
84
80
90
84
63
28
80
92
23
65
32
17
27
24
6
30
16
3 17
3 7
16
12 25
72
4 16
8
77
35
68
83
60
52
89
50
64
82
85
82
61
23
79
88
100
65
2
17
40
48
11
30
20*
16
20a
1 17
15
18
14 25
77
5 16
4 8
20 16
98
Wye 86 1 13- 86 1 13 80 7 13
a - Receives manure with the second fertilizer rate
b - Receives second fertilizer rate
F - Crop receives fertilizer only
M - Crop receives manure only
F+M - Crop receives both fertilizer and manure
88
-------
d> The fertilizer application rates for corn are relatively consistent
among the various states and model segments, whereas manure
application rates and fertilizer rates for the other crops are much
more variable, -even between neighboring model segments and within
model segments that cross state boundaries (see discussion below).
e. In Maryland and Virginia, different fertilizer rates were provided by
the states when fertilizer was used alone, and when it was applied in
addition to manure. The second (usually smaller) number shown in
Table 4.5 is the rate used when manure was also applied.
The basic information available to the state representatives on which to base
these application rates varied in detail from state to state. In general, the
rates for fertilizers, especially for corn, are probably more reliable than the
manure rates. Much of the information was extracted from county-level SCS or
Extension Service data, supplemented by 'best estimates' when information was not
available. In many cases, estimates were used when a particular crop represented
a minor fraction of the cropland in a specific model segment, e.g., races for
soybeans in the Upper Susquehanna were often estimated due tp lack of data, but
they often represented less than 5X of the cropland.
The rates were then adapted to model segments prior to being .transmitted to Aqua
Terra and CBPO; the only exception was for Virginia which provided rates for
subregions within the state which were then transformed into the model segment
values shown in. the tables. '.
For model segments that crossed state boundaries, a weighing procedure was used
to estimate application rates when the appropriate information was available from
the adjoining states and the areas were significant (e.g., more than 10X of the
model segment). Table 4.8 shows .the estimated percentages of the multi-state
model segments in the various states. The weighting was performed for each crop
application rate prior to calculating the model segment composite rate for the
composite crop. The percentages of croplands receiving fertilizer, manure, and
both were usually consistent and values for one state were adopted.
Although Delaware, New York, and West Virginia include portions of the Chesapeake
Bay drainage area, they are not active members of the Chesapeake Bay Program.
Consequently, information was not available for nutrient applications in these
states and the rates had to be estimated from the rates in the other CBP member
states.
The specific approach and relevant issues in individual model segments were as
follows:
Segments 10 and 20: PA rates were used since application rate information
for NY was not available. Although 67X and 91X of the area of these two
segments respectively reside in NY, the cropland (CNT, CST, Hay) comprises
20X and 18X, respectively, of the model segments.
89
-------
TABLE 4.8 ESTIMATED PERCENTAGES OF MULTI-STATE SEGMENTS
IN VARIOUS STATES - .
MODEL
SEGMENTS NY PA DE MD VA WV
10 67.0% 33.0%
20 ' 91.0% 9.0%
160 19.0% 36.0% 45.0%
170 7.0% 93.0%
175 ' * 13.0% 19.0% 1.0% 67.0%
180 36.0% 23.0% 19.0% 22.0%
200 92.0% 8.0%
210 23.0% 77.0%
220 33.0% 67.0%
Bohemia 20.0% 80.0%
Chester 10.0% 90.0%
Choptank 17.0% 83.0%
Coastal 1 2.0% 98.0%
Coastal 11 37.0% 63.0%
Gunpowder 2.0% 98.0%
Nantlcoke 63.0% 37.0%
Pocomoke 6.0% / 89.0% 5.0%
Pocomoke 95.0% 5.0%
Potomac . 53.0% 47.0%
Wicomlco 1.0% 99.0%
90
-------
Segment 160: MD rates were used for the entire segment because no
information was available for UV, MD comprised the next largest portion,
and MD provided detailed information on Z of cropland receiving
fertilizer, manure, and both. The overall impact should be small since
only 3.3Z of the segment is in Cropland and 6.7Z is in Hay.
Segments .170 and 175: Rates from the neighboring segments in VA,
Shenandoah Segments 190 and 200, were used with some adjustment of the 'X
receiving manure', because it was felt that the nature of agriculture in
these segments would be more similar to that in VA than in the other
states. Based on animal populations reported by Lugbill (1990). the
percentages of cropland receiving manure were reduced to 30Z of the values
in Segment 200. The overall impact of these assumptions should be small
since only l.OZ of Segment 170 and 3.0Z of Segment 175 are in Cropland,
while 4.0Z andr5.OX, respectively, are in Hay. ' '
Segment 200: VA rates were used for the entire segment since only 8Z Is in
WV where no data was available.
Segments 180. 210. and 220: The rates for these segments were weighted
between two states: PA and MD for Segments 180 and 210, VA and MD for
Segment 220. The segment rates were calculated from the percentages for
each state shown in Table 4.8. MD information on 'Z receiving fertilizer
only, and both fertilizer and manure7, was used for all three segments
since values were provided for every model segment, and the values were
reasonably consistent with the information from PA and VA.
Coastal 11 and Potomac (BFL): Weighting of PA and MD rates for Coastal 11,
. and VA and MD rates for the Potomac BFL segment, was performed. MD
information on percentages receiving fertilizer and/or manure was used.
Remaining BFL Segments; For the remaining multi-state BFL model segments
shovn in Table 4.8, MD data was used since the majority of the model
segment area was contained in MD and no information was available for the
portions in DE.
Although the assumptions described above are reasonable, and were necessary to
complete the AGCHEM modeling within the project schedule, any future efforts
should include an attempt to verify the application rates and percentages used
especially for areas in the non-member states of NY, WV, and DE.
4.3.4.2 Atmospheric deposition
At the request of the NRTF, atmospheric deposition of nitrogen was included in
the nutrient balance of the cropland areas simulated by AGCHEM by adding in
deposition estimates as part of the nutrient input to the land surface. The
Phase I Watershed Model included atmospheric deposition directly on water
surfaces as input to the stream channel simulation in each model segment. In
Phase Tl, these deposition values for nitrogen were also added to the nutrient
input to the cropland areas. Although the HSPF code does not currently contain
a capability for daily vet or dry deposition values, the estimated annual
nitrogen deposition for each model segment 'was included as an addition to the
91
-------
fertilizer/manure applications so that it could be considered in the annual
nutrient balance. Table 4.9 shows the model segment values for 1982-87 along
with the annual average; the annual average value for each segment (rounded to
whole numbers) was included in calculating the total nutrient applications to the
cropland areas (see Section 4.3.4.6 below). The development of the atmospheric
deposition values was described earlier in Section 3.4; the same values are used
for the cropland areas and the deposition onto water surfaces in each model
segment.
4.3.4.3 Composition of nutrient inputs
All nutrients that are input to the AGCHEM model must be in one of the specific
nutrient forms simulated. As discussed in Section 4.3.1, the nutrient forms
simulated by AGCHEH include N03, NH< (adsorbed and dissolved), Organic N, PO^
(adsorbed and dissolved), and Organic P. Thus, all fertilizer, manure, and
atmospheric deposition inputs are defined as one or more of these five nutrient
forms. Figure 4.13 shows the assumed forms of each of these three nutrient
sources.
For fertilizers, information provided by the NRTF indicated that the primary
forms of nitrogen fertilizers, are applied as urea, ammonium nitrate, and so-
called 'nitrogen solutions' which are typically referred to as UAN (urea ammonium
nitrate) solutions (Tisdale et al., 1985). For all three primary CBP states, UAN
solutions dominate the nitrogen fertilizer use. Similarly, phosphorus
fertilizers are almost exclusively applied in some .form of phosphate material,
either alone or in a combination with other nutrients such as ammonium or potash.
Based on the above information, all phosphorus fertilizer inputs were assumed to
be in the phosphate (POJ form. For nitrogen, we assumed a typical 301 nitrogen
UAN solution as defined by Tisdale et al. (1985), which contains a 50X/25X/25X
distribution of nitrogen as urea/ammonium/nitrate, respectively. Since 'urea'
is not a nutrient form simulated byCAGCHEM, all the urea nitrogen was input to
the model in the ammonium (NH4) form.. Consequently, all fertilizer was input as
752 NH4, and 25X N03. Tisdale et al. (1985) discuss the soil reactions (i.e.,,
hydrolysis and biomediated decomposition) that breakdown urea in the soil to
ammonia; they note that the reactions occur rapidly, continue .to occur at
significant rates even at temperatures of 2ฐ C and lower, and often completely
transform urea to ammonia within a few days in warm moist soils. Thus, inputting
all urea as ammonia appears to be a reasonable approximation.
For atmospheric deposition, the loading estimates developed by CBPO indicated
that about BOX of the nitrogen was in the nitrate form, with the remaining 20X
as ammonia. Due to the relatively rapid nitrification that occurs in surface
soils, all the deposition was input as nitrate in the model. Phosphorus loading?
in atmospheric deposition were assumed to be insignificant.
For manure nutrients, nutrient composition is much more variable, being a
function of animal type, diet, storage/handling, season, etc. . The primary issue
in terms of modeling is the distribution between the organic and inorganic
nutrient components because the inorganic fraction is partitioned between the
adsorbed and dissolved forms, while the organic fraction is assumed (in the
AGCHEM model) to be entirely bound to soil particles and sediment. Also, for
.92 .
-------
TABLE 4.9 ANNUAL NITROGEN ATMOSPHERIC DEPOSITION LOADS
Average
MODEL SEGMENT
10
20
30
40
50
60
70
80
90 *
100
110
120
140
160
170
175
180
190
200
210
220
230
235
240
250
260
265
270
280
290
300
310
330
340
ANACOSTIA
BALT HARBOR
BOHEMIA
CHESTER
CHICKAHOMINY
CHOTANK
COASTAL 1
COASTAL 11
COASTAL 4
COASTAL 5
1982
8
8
8
8
10
11
9
9
10
10
9
8
8
9
10
10
9
9
9
9
9
8
8
8
8
8
9
8
8
7
7
7
8
8
- 8
8
8
8
8
8
8
8
7
8
.95
.10
.51
.80
.92
.39
.21
.78
.27
.72
.57
.30
.61
.66
.80
.13
.97
.72
.19
.53
.33
.96
.38
.35
.59
.32
.90
.85
.03
.88
.45
.24
.91
.91
.58
.58
.58
.48
.31
.44
.48
,58
.97
.38
1983
8.75
8.77
8.50
8.81
9.89
9.82
8.87
8.67
9.52
9.55
8.34
8.10
7.87
9.45
8.55
8.98
8.61
8.28
8.38
8.11
8.11
7.94
7.50
7^47
7.50
7.43
7.71
7.34
7.13
7.07
6.63
6.56
8.17
8.17
7.83
7.83
7.70
7.66
7.57
7.56
7.60
7.77
7.02
7.63
1984
10.23
9.87
9.39
9.56
11.72
11.78
10.23
10.10
10.94
11.17
9.15
8.78
8.75
10 . 34
9.43
10.00
9.29
8.55
9.26
8.92
8.92
8.75
8.38
8.28
8.38
8.24
8.52
7.88
7.88
7.81
7.17
7.10
8.85
8.85
8.72
8.72
8.72
8.41
7.77
8.11
8.01
8.72
7.44
8.04
1985
9.15
8.91
8.77
9.01
10.15
10.22
9.35
8.87
9.92
9.75
8.40
7.96
7.93
9.52
8.94
9.51
8.87
8.21
8.91
8.44
8.44
8.14
7.63
7.60
7.70
7.56
8.04
7.33
7.20
7.13
6.62
6.62
8'. 23
8.23
7.83
7.83
7.83
7.73
7.63
7.62
. 7.66
7.83
6.88
7.36
1986
8
9
9
9
11
11
9
9
10
10
8
8
8
10
10
11
9
9
10
8
8
8
7
7
7
7
8
7
7,
7
6
6
8
8
7
7
7
7
7
7
7
7
6
7
.70
.65
.38
.22
.51
.65
.83
.15
.06
.70
.81
.10
.07
.87
.43
.67
.76
.02
.19
.78
.78
.21
.70
.47
.84
.43
.72
.47
.34
.20
.76
.62
.24
.24
.83
.77
.77
.46
.23
.36
.66
.77
.95
.29
1987
9
9
8
10
11
11
9
8
11
10
8
7
7
9
8
9
8
7
8
8
8
7
7
7
7
7
7
7
6
6
5
5
7
7
7
7
7
7
6
7
7
7
6
6
.38
.12
.58
.22
.79
.86
.69
.75
.07
.91
.21
.77
.74
.52
.21
.59
.68
.73
.58
.17
.17
.67
.50
.40
.50
.36
.43
.06
.59
.39
.95
.95
.77
.77
.56
.63
.63
.26
.35
.23
.53
.70
.42
.96
Total N
9
,9
8
9
11
11
9
9
10
10
8
8
8
9
9
9
9
8
9
8
8
8
7
7
7
7
8
7
7
7
6
6
8
8
8
8
8
7
7
7
7
8
7
7
.19
.07
.86
.27
.00
.12
.53
.22
.30
.47
.75
.17
.16
.89
.39
.98
.20
.58
.08
.66
.62
.28
.85
.76
.92
.73.
.39
.66
.36
.24
;76
.68
.36
.36
.06
.06
.04
.83
.48
.72
.82
.06
.12
.61
93
-------
TABLE 4.9 (continued)
Average
MODEL SEGMENT 1982 1983 1984 1985 1986 1987 Total N
7.91
7.52
6.63
6.79
7.40
7.98
6.96
6.79
7.58
7.99
8.00
7.60
7.34
7.67
7.45
7.91
7.44
7.60
7.12
COASTAL 6
COASTAL 8
COASTAL~9
ELIZABETH
GREAT WICOMICO
GUNPOWDER
JAMES
NANSEMOND
NANTICOKE
OCCOQUAN
PATAPSCO
PATUXENT
POCOMOKE
POTOMAC
RAPPAHANNOCK
SEVERN .
WICOMICO
WYE
YORK
8
8
7
7
8
8
7
7
8
8
8
8
8
8
8
B
8
8
7
.51
.34
.13
.60
.28
.54
.67
.60
.38
.54
.54
.38
.28
.41
.31
.51
.31
.38
94
7
7
6
6
7
7
7
6
7
7
7
7
7
7
7
7
7
7
7
.77
.40
.93
.99
.53
.80
.13
.99
.49
.80
.80
.63
.39
.67
.30
.77
.43
.36
.20
8
7
7
7
7
8
7
7
7
8
8
7
7
7
7
8
7
7
7
.38
.81
.13
.20
.81
.28
.27
.20
.84
.08
.28
.84
.74
.94
.77
.25
.77
.84
.34
7
7
6
6
7
7
6
6
7
7
7
7
6
7
7
7
7
7
6
.76
.12
.25
.31
.05
.79
.65
.31
.36
.93
.79
.36
.99
.39
.36
.69
.09
.56
.99
7
7
6
6
7
7
6
6
7
7
7
7
7
7
7
7
7
7
7
.70
.19
.38
.58
.12
.86
.92
.58
.43
.86
.86
.29
.12
.46
.16
.70
.23
.29
.06
7.37
7.26
5.98
6.05
6.59
7.60
6.12
6.05
6.96
7.74
7.74
7.09
6.52
7.13
6.82
7.57
6.82
7.16
6.18
94
-------
Composition of Nutrient Inputs
Nitrogen
Phosphorus
Fertilizer
25% NO3
75% NH3/Urea
100% PO4
Atmospheric
Deposition
100% NO3
V0
in
Manure
40% NH3
24-28%
16-12%
60% Org. N
18%
42%
Volatilization
Losses
Input as
Organic N
30-34%
Input as NH3
50% PO4
50% Org. P
Figure 4.13 Composition of nutrient inputs to AGCHEM model.
-------
nitrogen, volatile losses for the inorganic component (primarily NH3) can be a
significant fraction of the total manure nitrogen application.
A variety of articles and reports were reviewed in order to estimate a reasonable
value of percent inorganic nitrogen for manure applications to cropland. The
following table summarizes the range and mean of values found in the literature,
along with the literature source and the manure type:
Percent
Inoranic N
6 to 59 I
24 to 62 X
15 to 75 X
Mean: 26X
Mean: 33X
Mean: 38X
39 X
45 X
37 X
55 X
30 to 48 X Mean: 39 X
Manure
Type
Beef
Dairy
Swine
Dairy
Dairy
Dairy. liquid
Swine, liquid
Dairy, Liquid
Reference
ReddyeCal., 1979
"
Loehr, 1984
VA DSWC (Flagg, 1990
personal communication)
PA DER Manual
by R.E. Wright (1990)
Fox and Piekielek, 1990
Although, a number of the references are general in nature, the information from
the VA Department of Soil and Water .Conservation, the PA Department of
Environmental Resources, and the work of Fox and Piekielek (1990) are all
specific to the Chesapeake Bay region. Based on this review and with primary
emphasis on beef and dairy wastes, a value of 40X inorganic N was assumed for all
'nitrogen applied as manure. This assumption then provided the basis for
subsequent calculations of volatile ammonia losses from manure applications , as
discussed below in Section 4.3.4.4.
The remaining 60X organic N fraction in manure is comprised of both unstable and
stable organic compounds , with the unstable portion derived from urea* type
materials. Consequently, the unstable organic nitrogen will become available as
inorganic species (primarily NHซ) through mineralization processes-: much more
quickly than the stable fraction. Based on information from Banvel (1990), the
Maryland Department of Agriculture (F. Samadani, 1990, personal communication,
and the Virginia Department of Soil and Water Conservation (M. Flagg, 1990,
personal communication), about 30Z to 50X~6f the organic N can become available
to the crop as inorganic N during the first year; 30X to 40X can become available
relatively quickly within 3 to 4 weeks. The remaining portion of the organic N
is either non-mineralizable or slowly mineralizes over subsequent years; about
20X of the organic N is in this latter category based on the sources noted above.
Since AGCHEM includes only one mineralization rate for organic N. we assumed that
30X of the organic N is quickly converted to NH\ for input to the model with the
remaining fraction added to the soil organic N storage. In Figure 4.13, this 30X
is shown as the 18X of Total N (i.e. , 30X x 6. OX Organic N - 18X NHซ) added to the
96
-------
non-volatilized ammonia.
For the phosphorus composition in manure, very little useful information was
found due largely to the complex chemistry and associated analytical problems in
measuring the various organic and inorganic farms of P. Sommers and Sutton (1980)
provide a brief overview of some of these problems, indicating that most of the
P in animal wastes is part of relatively insoluble compounds and that much of the
organic P is not well defined and of unknown chemical structure. They cite work
by Townshend et al. (1969). where swine, beef cattle, and dairy cattle wastes were
found to have soluble P (expressed as a percentage of Total P) values of .43 X,
54 X, and 46 X, respectively. Based on this limited data, we assumed a 50/50
split between the inorganic and organic P fractions of manure P input to the
model, as shown in Figure 4.13. Clearly, this split is not quite as critical as
the corresponding one for N, since PQi does not undergo volatilization losses
during and following'application to the land.
4.3.4.4 Application/volatilization losses of manure nitrogen '
Since AGCHEM does not account for volatilization losses of nitrogen, these losses
associated with manure application were estimated and the application rates were
reduced accordingly to subtract the volatilized fraction. Ammonia volatilization
from manure applications is highly variable, and depends on animal waste type and
characteristics, meteorologic conditions (e.g., wind, air temperature).
application procedures, etc. Since the inorganic (primarily ammonia) fraction
is the volatile component, various literature sources were reviewed in order to
estimate the fraction of the inorganic component that would likely be lost by
volatilization. The following summarizes the results obtained from this review:
Percent NH3 Loss
4 to 96 X loss, dairy, surface applied
(mostly in 60 to SOX range)
39 X (incorporated within 24 hrs)
over 50 X (incorporated within 24 "hrs)
over 50 X (from surface within 24 hrs)
over 50 X (loss within 24-48 hrs, incorp.)
Reference
Reddy et al.,1979
Stevens and Logan, 1987
Thompson et al,, 1987
Lauer et al,, 1976
Fox and Piekielek, 1990
Gasman (1989) reviewed a number of literature sources on animal waste
volatilization losses associated with collection, storage and handling, and
application, and found that application losses were generally in the range of 10
to 60 X of Total N. In a spreadsheet for calculating animal waste application.
for croplands in Maryland, Banvel (1990) provides the user with options ranging
from 20X loss of NH, for incorporation within 1 day, to SOX loss for
incorporation within 5 days. He also allows an option for 100X loss (by both
volatilization .and leaching) if the manure is not incorporated at all, or later
than 5 days. . .
97
-------
Information on manure application procedures were provided by the state members
of the NRTF. In VA manure is either incorporated after 7 days or not at all. In
PA, the practice is to incorporate within 2 days about 40Z of the time, and after
2 days about 602 of the time. Practices in J1D typically fell between these two
extremes: incorporation within 2 to 5 days. In discussions with the NRTF, and
considering the literature information cited above, the following 'percent
ammonia losses' were assumed for each state:
Pennsylvania 60 Z loss
Maryland 65 Z loss
Virginia 70 Z loss
These percent losses were applied to the manure application rates specified by
each state as shown in Table 4.6; the state-supplied rates represented the
amounts applied to the land surface, accounting for any collection, handling, and
storage losses, so that application and volatilization losses were the only
additional mechanisms to be considered.
4.3.4.5 Nutrient application methods and timing
Once fertilizer and manure application amounts were estimated, application
practices and procedures were studied in order to define the appropriate timing
and placement of the nutrient applications for input to the AGCHEH model for each
model segment. 'Timing' refers to the period during the year when nutrients are
applied for each major crop, and 'placement' refers to how the nutrients are
applied, usually either surface application or soil'incorporated. The USDA
Handbook No. 628 (USDA. 1984) provided information on usual planting dates for
the major crops; information from .the NRTF, and subsequent discussions with state
representatives, helped to further refine typical application practices. . Table
4.10 summarizes the protocols used to distribute the estimated nutrient
application amounts, in both time and space, for both fertilizer and manure
applications to each major crop, along with assumptions on the atmospheric
deposition component. .
For fertilizer applications, both surface application and soil incorporation were
considered typical procedures depending on whether the application is at, or
prior to, planting or during the growing season as a side-dressing. For surface
applications, the application amount was added, in the model, to the 1 cm surface
zone. For soil incorporation, different assumptions were used for Conventional
Tillage (CNT) versus Conservation Tillage (CST). For CNT, we assumed that 10X
of the application would remain in the surface zone, and 90X would be added to
the 14 cm upper zone; this is based on the assumption that incorporation would
produce a uniform distribution of the chemical in the top 10 to 15 cm of the
soil. For CST, we assumed 15X would remain in the surface, and 85X would be in
the upper zone, derived from assuming a linear distribution of chemical from the
surface to zero at the depth of incorporation of 10 to 15 cm. These assumptions
were based on reviews of literature to evaluate the impacts of BMPs on model
parameters (Donigian, et al., 1983).
For manure applications, we assumed a 50/50 split between the surface and upper
zones for all nutrient components of the animal waste. Although manure
.98
-------
applications are typically surf ace-applied, some incorporation does occur. Also,
the small depth of the 1 cm surface zone relative to the large mass of manure
applied (typically 10-15 tons/ac) would lead to some natural mixing of manure
with native soil below the 1 cm depth, which the model, represents as the upper
zone. Since the AGCHEM model assumes all organic N and P is sediment bound, any
organics applied to the surface remain in the surface except for losses by
mineralization and attached to eroded sediment. For these reasons, the 50/50
split was needed to more closely represent manure nutrient applications within
Che model framework. .
For the Hay segment, we also assumed a 50/50 split for fertilizer applications
because re-seeding would likely involve, soil incorporation of applied fertilizer
but this would occur only once every two to three years. In the other years, the
fertilizer would be surface applied. Since some fraction of the farmers would
be re-seeding hay fields every year, we assumed the 50/50 split would be a
reasonable compromise for the complex combinations of practices that would occur
each year. .
Based on discussions with the NRTF, the frequency of manure applications can vary
from daily spreading in winter, to weekly or monthly applications for hay, to one
or two spring or fall applications for field crops. Unfortunately, the current
procedures for specifying nutrient applications in AGCHEM are too cumbersome to
allow daily or weekly applications at the scale of this modeling effort; each
application must be defined explicitly for each model segment and each year of
Che simulation, and each specific state variable impacted by the application.
Consequently, Che minimum frequency used in Che model is monthly applications for
Che hay segment and bi-monthly (i.e., every two months) for the other crops,
during Che typical application period appropriate for each crop.
.In general, the application protocols described in Table 4.10 produced a
reasonable representation of nutrient application procedures throughout Che Bay
drainage. Some compromises between practices in different states were needed in
order to maintain consistency in the model representation. Any future studies
with the Watershed Model by the CBPO and/or individual states should consider
refining the assumptions and practices as needed to better represent local and
regional conditions.
4.3.4.6 Calculation of model input values
Calculation of the final model input nutrient application amounts was performed
with a Lotus spreadsheet for each model segment, following the flowchart shown
in Figure. 4.12 and Che assumptions discussed in the above sections. Tables 4.11
and 4.12 show portions of the spreadsheet for Model Segment'100 (Juniaca model
segment) to demonstrate the format of the spreadsheet and the steps in the final
calculations.
Table 4.11 shows the calculations of the annual application amounts for the CNT
category within Model Segment 100; separate portions of Che spreadsheet are used
Co calculate the rates for CST and Hay. These portions are not shown, but they
follow che same steps as in Table 4.11 except that the calculations for Hay are
simpler since no compositing is involved.
99
-------
Table 4.10 MODEL PROCEDURES FOR TIMING AND PLACEMENT OF FERTILIZER/MANURE
APPLICATIONS AND ATMOSPHERIC DEPOSITION
CORN: ......
Fertilizer- SOX N and 100X P applied at planting and incorporated
50Z N sidedressed 5-6 weeks after planting
Manure - Bimonthly (i.e., every 2 months) applications, November
through April, with a 50/50 split between Surface and Upper
Zone -
SOYBEANS;
Fertilizer- 100X of N and P applied at planting, incorporated, about 3-4
weeks after corn planting dates
Manure- Bimonthly (i.e., every 2 months) applications November through
April, same as corn; if amounts are small, use double
applications in November and April (for Fall and Spring).
Split 50/50 between surface and upper zone.
NOTE: If total application amount is small, less than 5 pounds/ac, add
to the corn application, i.e., don't put in as separate applications.
GRAINS;
Fertilizer- 100X P and 20X N ^application in Fall (September),
incorporated, and 80 X N application top-dressed in Spring
(March)
Manure- Bimonthly summer applications, June through September, 50/50
split incorporation in Surface and Upper Zone
HAYc ' . ;'.' . .'" . .
Fertilizer- Use 40/30/30 split applications between August, mid-March, and
June to account for different seeding times (August and March,
incorporated 50/50 split between Surface and Upper zones) and
single top-dressing in June.
Manure- Monthly application April- through October, with 50/50 split
between surface and upper zone. .
ATMOSPHERIC DEPOSITION:
Apply all deposition amounts as N03 (since NH3 will be nitrified quickly)
and add to die fertilizer component, but only to the SURFACE zone.
100
-------
TA51E 4.11 EXAMPLE CALCULATION OF FINAL NUTRIENT APPLICATION RATES
SEGMENT 100
j-Modc
Crap Distribution,
State-Supplied
Application RAtes,
and AtBos. Dep.
Cropland Category
Initial Composite
Based on Crop
Distribution
State-Supplied
Fractions for
Application Rate
NH3 Volatilization
and Nutrient
Coapoaitian
Assuaptions ,
FINAL Composite
Rates and
Crop Coaponents
CALCULATION OF NUTRIENT APPLICATION RATES -
Date: 03/08/91
State-Supplied Fertilizer and Manure Appl. Rates and CBLO ATM. Depos.
State Nutrient Application Rates
I Segment Nunfcer
Corn
Corn-si I
Soybeans
Grains
Total's
CNT
CROP X
46.3X
2S.3X
2.8X
2S.6X
100.0X
CST
CROP X
54. 4X
29.9X
3.5X
12. ZX
100.0X
Fertilizer *
N
95
95
34
34
P
59
59
8
8
1 Manure
N
150
150
140
140
P
26
26
,25
25
ATM. DEP
. N
10
10
10
10
Fertilizer * 2
N
0
0
0
0
P
0
0
0
0
.-G
Conventional Tillage
Fertilizer
Manure ATM. DEP TN
(less NH3 vol. losses)
TP
Coapocition of
Final Segacnt
Application Rates
1 N P N P
Composite 77.7 44.5 111.8 25.7
Corn 68.0 42.2 81.6 18.6
Soybeans 1.0 0.2 3.0 0.7
Grains 8.7 2.0 27.2 6.4
Area Weighted Application Rates - CNT
N
10.0
7.2
0.3.--
2.6
Fraction of area receiving fertilizer * 1 and manure
Fraction of area receiving fertilizer * 1
Fraction of area receiving fertilizer * 2 and manure
Fraction of area receiving fertilizer * 2
Fraction of Manure available as NH3
Fraction of Manure available as ORGN
Fraction of Manure available as P04
Fraction of Manure available as ORGP
Fraction of fertilizer available as NH3
Fraction of fertilizer avajlable as N03
Fertilizer Manure
N P N P
Composite 67.3 38.1 35.1 8.0
Corn ^57.8 35.9 28.6 6.5
Soybeans 0.8 0.2 1.0 0.2
Grains 8.7 2.0 5.4 1.3
Fertilizer and Manure Conposit ion of TN and TP
Fertilizer * ATMOS. Dep.
ATM.
NH3 N03 N03 P04 NH3
Composite 50.5 16.8 10.0 38.1 15.5
Com 43.4 14.5 7.235.9 12.8
Soybeans 0.6 0.2 0.3 0.2 0.5
Grains 6.5 2.2 2.6 2.0 2.4
0.34
0.42
0.50
0.50
0,75
0.25
ATM. DEP
N
10.0
7.2
0.3
2.6
199.5 70.2
156.8 60.9
4.2 0.9
38.5 8.4
Corn Beans Grains
0.20 0.20 0.20
0.65 0.65 0.80
0.15 0.15 0.00
0.00 0.00 0.00
X
TN TP
112.4 46.2 ->
93.5 42,4 FINAL
2.1 0.4 SEGMENT
16.7 3.3 RATES
'
. Manure . TN TP
ORGN P04 ORGP
19.1 4.0
15.8 3.3
0.6 0.1
3.0 0.6
4.0 112.4 46.2
3.3 93.5 42.4
0.1 2.1 0.4
0.6 16.7 3.3
101
-------
TABLE 4.12 EXAMPLE CALCULATION OF TIMING AND PLACEMENT OF NUTRIENT
APPLICATION
Tiarfi* an
Fertilize
Applicati<
of
Application*
Conventional
TUlaซa
Contanwtian
Till***
ay land
Convantfonal
TIUaซa
Camera t Ian
TllUa*
FEITILIKI TIMING AND PIACEMEHT I SEGMENT 100K Model ScgBซU NuBtaar
CUT
ซ:HN3
II:N03
f:KX
CST
ซ:NH3
:N03
P:P04
KAT
Spr. Cr.
SZ
5.22
3.79
0.00
Spr. Cr.
SZ
2.49
.1.81.
0.00
Nardi
SZ UZ
ป:*a o.oo o.oo
:N03 3.00 0.00
' 1.20 1.20
Plant C.
SZ UZ
2.17 19.51
t.30 *.SO
3.5932.32
tant C.
SZ UZ
3.83 21.70
5.49 7.23
0.34 35.93
Juw
SZ
0.00
3.00
2.40
riant I.
SZ UZ
0.06 O.SS
0.30 0.18
0.02 0.17
tant 8.
SZ UZ
0.11 0.64
0.39 0.21
0.04 0.20
SZ UZ
0.00 0.00
4.00 6.00
1.60 1.60
S.Or. C
SZ
21.68
10.81
0.00
s.or. e
sz
25.53
12.72
6.00
Til T*
10.00 8.00
rail Cr.
SZ UZ
0.13 1.18
0.56 0.39
0.20 1.84
fall Cr.
SZ UZ
0.09 0.53
0.28 0.18
n rr
77.33 38.15
M Tป
83.23 43.49
0.150.83
Total
Fartllitar
Application.
[sttMEIIT 100|- I
MANURE TIMIttG MB HACEMEUT SECMEKT 100[- Modal tl|ซant Ma*ar
CUT
mo '
OtCM
M4/OK*
CST
no
OftGN
04AKG*
MY
COM/SOT (3 app(.)
Mov/Jon/Nar
SZ UZ
2.21 2.21
2.73 2.73
O.S4 O.S6
COtMTSOT C3 appt.)
ov/Jan/Nar
SZ UZ
2.61 2.61
3;22 J.22
0.66 0.64
OUINS (3 appl.)
Mar/All/Sap
SZ UZ
0.41 0.41
O.JO O.SO
8.11 0.11
CM!** (3 appl.)
May/M/Sap
SZ UZ
0.19 0.19
0.24 0.2*
O.OS 0.8S
KAT (6 appl.) Nar/AtVJul/Aui/Sap/Oct
SZ UZ
.1.281.28
1.58 1.S8
0.33 0.33
Total
Hanuro
Appdeatie
n Surfaea Soil Zona
UZ - Uppar Soil.Zona
102
-------
At the top of the spreadsheet, the input values for the cropping distributions,
the state-supplied rates for each crop, and the atmospheric deposition amounts
are shown. In the next section, the initial composite value and the individual
crop components are calculated based on the input cropping distribution. These
rates also allow for NH3 volatilization losses using the value for 'fraction of
manure available as NH3' , which is.shown in the middle portion of the table. For
example, the value shown in Table 4.11 is 0.34; this accounts for 60Z loss by
volatilization of the inorganic N, plus the fraction of organic N available as
NH3 (as shown in Figure 4.13 and discussed above).
The middle section of Table 4,11 shows the state-supplied fractions of each crop
(i.e., corn, beans, grains) which receives the various combinations of manure and
up to two different > fertilizer rates; this middle section also shows the
fractions used to calculate volatilization losses and nutrient composition. The
initial composited crop component rates are then weighted by the crop fractions
for the various combinations in order to calculate the final segment rates, and
the crop components, shown in the second to last block near the bottom of Table
4.11. The final block in Table 4.11 shows the composition of the final segment
rates, i.e., the distribution of the final rates into organic and inorganic
forms.
The components of the final segment rates, attributable to each crop is carried
through the calculations so that the protocols for timing and placement for each
crop can be represented. In Table 4.12,. the final segment rates from Table 4.11
are then distributed in time and space for specific applications on each cropland
category. The top half .of Table 4.11 shows the values for fertilizer
applications, while the bottom half shows the manure applications for CNT, GST,
and Hay. The values shown in this table- (which are the same as in the
spreadsheet) are the specific values input to the model and can be identified in
the model segment input sequence. .
For fertilizer applications, the values in Table 4.12 include the NH3, N03, and
P04 amounts epplied to.the surface zone (SZ in the table), and the upper, zone
(UZ) for applications associated with spring grains, planting of corn and
soybeans, side dressing of. corn, and fall grains. For the Hay segment, .the
fertilizer amounts associated with the three separate applications (discussed in
Table 4.10) are shown. .''--. ''.'.-
For manure applications, the bottom half of Table 4.12 shows the inorganic and
organic amounts, the.number and timing of the applications, and the placement in
the surface and upper zones. . .
The final application rates for all model segments for the CNT, CST, and Hay
cropland categories are .summarized in Tables 4.13, 4.14. and 4.15, respectively.
Appendix C includes copies of the individual spreadsheets for each model segment,
which are also available from the CBPO as Lotus files.
4.3.5 AGCHEM application procedures for Model Segments
As noted in the introduction to Section 4.3, the primary goal of the AGCHEM
application in Phase II was to represent nutrient balances on the cropland areas
so that the impacts of climate, agricultural practices, and nutrient management
103
-------
TABLE 4.13 FINAL CALCULATED MODEL SEGMENT NUTRIENT APPLICATION RATES FOR
CONVENTIONAL TILLAGE (Ib/ac)
MODEL FERTILIZER MANURE ATMOS. DEP. TOTAL
SEGMENT N P N P N N P
10
20
30
40
50
60
70
80
90
100
110
120
140
160
170
175
180
190
200
210
220
230
235
240
250
260
265
270
280
290
300
310
330
340
84.3
97.4
95.3
77.1
77.5
86.3
78.7
59.4
68.9
67.3
62.6
71.5
63.4
110.9
138.9
130.5
117.5
117.9
113.7
97.5
114.1
130.7
87.0
76.8
118.0
83.8
149.2
138.6
95.6
81.1
97.2
104.8
101.0
110.4
17.2
19.7
19.3
15.9
17.1
18.9
17.3
, 22.2
39.5
38.1
26.8
28.8
27.1
13.4
30.7
30.2
21.2
29.1
28.7
26.8
25.5
22.0
27.2
29.8
27.0
26.6
32.2
31.6
26.5
27.4
27.4
26.3
14.8
15.4
22.4
22.6
22.6
23.4
33.1
35.7
34.6
43.0
35.4
35.1
61,5
66.4
60.9
21.8
28.6
28.6
24.7
97.3
93.5
26.1
1.9
0.0
0.0
0.0
0.0
0.0
0.0
5.7
0.0
0.0
0.0
0.0
29.4
16.8
5.3
5.3
5.3
5.5
7.7
8.4
8.1
10.1
8.1
8.0
14.4
15.5
14.3
7.3
9.0
9.0
7.1
30.6
29.4
7.5
0.6
0.0
0.0
0.0
0.0
0.0
0.0
1.8
0.0
0.6
0.0
0.0
9.3
5.2
. 9
9
9
9
11
11
10
9
10
10
9
8
8
10
9
10
9
9
9
9
9
8
8
8
8
8
. .. 8
8
7
7
7
7
8
8
115.7
129.0
126.8
109 . 5
121.6
133.1
123.2
111.5
114 . 3
112.4
133.2
145.9
132.4
142.7
176.6
169.1
151.2
224.2
216.2
132.6
125.0
138.7
95.0
84.8
126.0
91.8
157.2
152.3
102.6
88.1
104.2
118.8
138.3
135.2
22.5
25.0
24.6
21.4
24.8
27.2
25.4
32.2
47.6
46.2
41.2
44.3
41.3
20.7
39.7
39.2
28.3
57.7
58.1
34.3
26.1
22.0
27.2
29.8
27.0
26.6
32.2
33.4
26.5
27.4
27.4
26.3
24.1
20.5
104
-------
TABLE 4.13 FINAL CALCULATED MODEL SEGMENT NUTRIENT APPLICATION RATES FOR
CONVENTIONAL TILLAGE (Ib/ac) (continued)
MODEL
SEGMENT
Anacostia
Baltimore
Bohemia
Chester
Chickahominy
Chop tank
Coast_l
Coast_4
Coast_5
Coast_6 '
Coast_8
Coast_9
Coast 11'
Elizabeth
Gr. Wico.
Gunpowder
James
Nansempnd
Nanticoke
Occoquan
Patapsco
Patuxent
Pocomoke
Potomac
FERTILIZER
N
96.7
109.3
94.9
85.4
89.9
72.4
74.6,
79.0
111.7
119.6
77 .5
64.3
100.9
56.1
80.0
103.7
82.4
112.5
93.9
123.5
101.4
92.8
74.2
82.7
Rappahannock 79.0
Severn
Wicomico
Wye
York
91.9
66.7
77.3
79;7
P
26.9
15.8
17.9
19.8
29.2
22.6
22.2
29.6
* 16.1
35.8
29.7 '
29.7
33.3
27.7
29.2
14.9
29.2
32.1
23.0
27.1
14.7
25.7
20.5
25.5
29.1
14.7
19.5
20.8
29.3
MANURE ATMOS. DEP. TOTAL ,
N
22.8
16.4
3 .5
11.7
0.0
24.4
10.2
0.0
10.6
3.0
0.0
0.0
72.1
0.0
0.0
25.6
0.0
0.0
127.6
0.0
16.7
7.6
137.8
2.4
0.0
27.7
130.4
21.5
0.0
P
8.1
4.8
1.0
3.9
0.0
6.7
2.9
0.0
3.5
0.9
0.0
0.0
19.2
0.0
0.0
7.9
0.0
, 0.0
36.4
0.0
5.4
2.6
40.1
0.9
6.0
7.3
37.4
6.3
0.0
N
8
8
8
. 8
7
8
8
7
8
8
8
7
8
7
7
8
7
7
8
8
8
8
7
8
7
8
7
8
7
N
127.5
133.7
106.4
105.0
96.9
104.9
92.8
86.0
130.3
130.6
85.5
71.3
181.1
63.1
87.0
137.3
89.4
119.5
229.5
131.2
126.1
108.4
219.0
93.1
86.0
127.6
204.1
106.8
86.7
P
35.0
20.6
18.9
23.7
29.2
28.8
25.1
29.6
19.6
36.7
29.7
29.7
52.6
27.7
29.2
22.8
29.2
32.1
59.3
27.1
20.2
28.3
60.5
26.5
29.1
22.1
56.9
27.2
29.3
105
-------
TABLE 4.14 FINAL CALCULATED MODEL SEGMENT NUTRIENT APPLICATION RATES FOR
CONSERVATION TILLAGE (Ib/ac)
MODEL
SEGMENT
10
20
30
40
50
60
70
80
90
100
110
120
140
160
170
175
180
190
200
210
220
230
235
240
250
260
265
270
280
290
300
310
330
340
FERTILIZER
N
96.7
104.3
102.1
86.1
90.9
95.2
86.8
65.3
74.6
73.2
71.0
77.5
70.7
115.8 .
139.7
133.0
124.5
121.6
117.6
108.3
116.5
134. 5;
87.2
75.1
121.2
82.9
149.6
142.1
96.8
80.4
98.3
107.2
106.1
115.4
P
19.5
20.9
20.5
17.5
19.8
20.7
19.0
25.4
- 44.7
43.5
28.5
30.0
28.4
16.8
30.5
29.5
25.1
28.9
28.4
28.4
25.1
22.1
26.9
30.4
26.6
25.9
32.2
31.4
26.0
27.0
26.8
25.4
15.2
15.8
MANURE
N
22.6
22.7
22.7
23.9
36.7
38.2
37.1
48.2
37.8
37.5
68.6
71.3
68.3
22.3
28.6
32,4
28.1
97.3
93.5
25.7
1.9
0.0
0.0
0.0
0.0
0.0
0.0
5.7
0.0
0.0
0.0
0.0
29.5
16.5
P
5.4
5.4
5.4
5.6
8.6
8.9
8.7
11.3
8.6
8.6
16.1
16.7
16.0
7.5
9.0
10.2
8.1
30.6
29.4
7.4
0.6
0.0
0.0
0.0
0.0
0.0
0.0
1.8
0.0
0.0
0.0
0.0
9.3
5.1
ATMOS. DEP. TOTAL
N
9
9
9
9
11
11
10
9
10
10
9
8
8
10
9
10
9
9
9
9
9
8
8
8
8
8
8
8
7
7
7
7
8
8
N
128.3
132.8
133.9
118.9
138.6
144.5
133.9
122.5
122.4
120.8
148.6
156.8
147.0
148.1
177.3
175.5
161.6
227.9
220.1
143.0
127.4
142.5
95.2
83.1
129.2
90.9
157.6
155.9
103.8
87.4
105.3
114.2
143.6
139.9
P
24.9
25.7
25.9
23.2
28.4
29.7
27.7
36.7
53.3
52.1
44.5
46.7
44.4
24.3
39.5
39.7
33.1
59.5
57.8
35.8
25.8
22.1
26.9
30.4
26.6
25.9,
32.2
33.2
26.0
27.0
26.8
25.4
24.5
20.8
106
-------
TABLE 4.14 FINAL CALCULATED MODEL SEGMENT NUTRIENT APPLICATION RATES FOR
CONSERVATION TILLAGE (Ib/ac) (continued)
MODEL
SEGMENT
Anacostia
Baltimore
Bohemia
Chester
Chicka.
Choptank . .
Coast_l
Coast_4
Coast_5
Coast_6
Coast_8
Coast_9
Coast 11
Elizabeth
Gr. Wico.
Gunpowder
James
Nansemond
Nanticoke
Occoquan
Patapsco
Patuxent
Pocomoke
Potomac
FERTILIZER
N
97.3
117.4
92.2
84.4
77.5
69.2
70.3
77.0
107.9
128.9
75.4
62.0
108.1
43.9
77.0
112.3
82.6
115.3
93.1
127.8
112.2
85.5
65.8
76.6
Rappahannock 76.2
Severn
Uicomico
Wye
York
98.7
57.0
76.1
78.0
P
28.3
16.2
17.9
21.6
29.3
24.7
24.5
30.3
- 16.1
37.1
30.4
30.2
34.4
28.1
29.9
15.4
29.5
32.6
23.0
26.5
15.8
25.1
19.4
25.3
29.6
15.1
18:3
23.4
29.8
MANURE ,
N
22.6
16.2
3.6
11.5
0.0
23.7
10.6
0.0
10.6
3.0
0.0
0.0
73.6
0.0
0.0
25.3
0.0
0.0
127.5
0.0
15.8
7.0
136.9
2.3
0.0
32.0
136.8
18.6,
0.0
P
8.0
4.8
1.0
3.8
0.0
6.5
3.0
0.0
3.5
0.9
0.0
0.0
19.6
0.0
0.0
7.8
0.0
0.0
36.4
0.0
5.1
2.4
39.6
0.9
0.0
8.2
39.6
5.5
0.0
ATMOS . DEP . TOTAL
N
8
8
8
8
: 7
8
8
7
8
8
8
7
8
7
7
8
7
7
8
8
8
8
7
8
7
8
7
8
7
N
127.9
141.7
103.8
103.9
84.5
100.9
88.9
84.0
126.4
139.9
83.4
69.0
189.7
50.9
84.0
145.6
89.6
122.3
228.7
135.8
136.0
ioo.s
, 209.7
86.9
83.2
138.6
200.8
102.7
85.0
. P
36.3
21.0
18.9
25.4
29.3
31.2
27.5
30^3
19.6
37.9
30.4
30.2
54.0
28.1
29.9
23.2
29.5
32.6
59.4
26.5
20.9
27.6
59.1
26.2
29.6
23.4
, 57.9
28.9
29.8
107
-------
TABLE 4.15 FINAL CALCULATED MODEL SEGMENT NUTRIENT APPLICATION RATES
FOR HAY (Ib/ac)
MODEL
SEGMENT
ATMOS. DEP. TOTAL
N N P
10
.20
30
40
50
60
70
80
90
100
110
120
140
160
170
175
180
190
200
210
220
230
235
240
250
260
265
270
280
290
300
310
330
340
30.0
30.0
30.0
30.0
30.0
30.0
30.0
7.5
0.0
0.0
10.0
10.0
10.0
0.0
34.2
34.2
2.1
29.3
29.3
1.5
7.2
60.0
34.7
30.0
36.0
35.5
36.0
34.6
33.4
30.2
30.0
30.0
9.0
9.0
14.5
14.5
14.5
14.5
13.0
13.0
13.0
17.5
* 8.0
8.0
13.0
13.0
13.0
23.7
28.5
28.5
18.9
24.1
24.1
11.7
9.9
21.0
29.0
25.0
30.0
29.6
30.0
28.7
27.8
25.2
25.0
25.0
3.9
3.*
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
34.2
34.2
41.0
41.0
41.0
12.1
4.0
4.0
11.6
15.1
15.1
14.4
" 1.6
0.0
0.0
0.0
0.0
0.0
0.0
0.4
0.0
0.0
0.0
0.0
24.6
13.0
0.0
0.0
0.0
0.0,
.0.0
0.0
0.0
0.0
7.8
7.8
9.6
9.6
9.6
4.1
i.a
1.8
3.5
6.7
6.7
4.3
0.5
0.0
0.0
0.0
0.0
0.0
. 0.0
0.2
0.0
0.0
0.0
0.0
7.8
4.0
9
9
9
9
11
11
10
9
10
10
9
8
8
10
9
10
9
9
9
9
9
8
8
8
8
8
8
8
7
7
7
7
8
8
39.0
39.0
39.0
39.0
41.0
41.0
40.0
16.5
44.2
44.2
60.0
59.0
59.0
22.1
47.2
48.2
22.7
53.4
53.4
24.9
17.8
68.0
42.7
38.0
44.0
43.5
44.0
42.9
40.4
37.2
37.0
37.0
41.6
30.0
14.5
14.5
14.5
14.5
13.0
13.0
13.0
17.5
15.8
15.8
22.6
22.6
22.6
27.8
30.2
30.2
22.4
30.8
30.8
16.0
10.4
21.0
29.0
25.0
30.0
29.6
30.0
28.9
27.8
25.2
25.0
25.0
11.7
7.9
108
-------
TABLE 4.15 FINAL CALCULATED MODEL SEGMENT NUTRIENT APPLICATION RATES
FOR HAY (Ib/ac) (continued)
MODEL
SEGMENT
Anacostia
Baltimore
Bohemia
Chester
Chicka.
Choptank
Coast_l
Coast_4
Coast_5
Coast_6
Coast_8
Coast_9
Coast 11
Elizabeth
Gr. Wico.
Gunpowder
James
Nansemond
Nanticoke
Patapsco
Patuxent
Pocomoke
Potomac
Occoquan
Rappa .
Severn
Wye
Wicomico
York
FERTILIZER
N
0.0
9.0
6.6
9.0
31.6
9.0
9.0
30.0
9.0
0.0
30.0
30.0
2.7
30.0
30.0
9.0
30.0
30.0
9.0
9.0
9.0
9.0
18.9
36.0
30.6
9.0
9.0
9.0
30.9
P
12.0
3.9
7.8
7.8
26.4
7.8
7.8
* 25.0
3.9
9.3
25.0
25.0
10.3
25.0s
25.0
3<.9
25.0
25.0
7.8.
3.9
3.9
7.8
12.0
30.0
25.5
3.9
7.8
7.8
25.8
MANURE
N
17 .,5
16.9
3.2
8.0
0.0
21.2
5.1
0.0
10.6
2.7
0.0
0.0
55.7
- 0.0
0.0
20.2
0.0
0.0
63.4
13.9
13.5
- 5.3
2.0
0.0
0.0
; 32.6
17.6
114.2
0.0
P
6.2
4.6
0.9
2.7
0.0
5.9
1.4
0.0
3.5
0.8
0.0
0.0
14.9
'0.0
0.0
6.3
0.0
0.0
18.0
4.5
4.6
1.5
0.8
0.0
0.0
8.0
5.2
32.9
0.0
ATMOS. DEP. TOTAL
N
8
8
8
8
7
8
8
7
8
8
8
7
8
7
7
8
7
7
8
8
8
7
8
8
7
8
8
7
7
N
25.5
33.9
17.2
25.0
38.6
38 . 2
22.1
37.0
27.6
10.7
37.0
37.0
66.3
37.0
37.0
37.2
37.0
37.0
80.4
30.9
30.5
21.3
28.9
44.0
37.6
49.6
34.6
130.2
37.9
P
18.2
8.5
8.7
10.5
26.4
13.7
9.2
25.0
7.4
10.1
25.0
25.0
25.1
25.0
25.0
10.2
25.0
25.0
25.8
8.4
8.5
9.3
12.7
30.0
25.5
12.0
13.0
40.7
25.8
109
-------
alternatives on NFS nutrient loadings and water quality could be evaluated. The
AGCHEM modules were originally designed for application to areas ranging from
field-size to small-to-moderate size watersheds on the order of the size of a
single model segment. For these size areas when the relevant soil, crop, and
runoff data are available, the AGCHEM calibration involves the establishment of
a reasonable simulation of soil nutrient storages through adjustment of plant
uptake rates, soil transformations, partition coefficients, and percolation
parameters, followed by evaluation of the nutrient runoff simulation (in
comparison with observed values) and parameter refinement as needed. The general
guidance provided for nutrient calibration as described in Che HSPF Application
Guide (Donigian, et al., 1983) is as follows:
1. Evaluate initial soil nutrient parameters from information available
in the literature, and include fertilizer manure and rainfall sources
of nutrients as input to the model.
2. Calibrate initial mineralization rates so that annual amounts of
plant-available nutrients correspond to expected values.
3. Adjust leaching factors based on any data available for a tracer such
as chloride.
4. Adjust plant uptake rates to develop the expected nutrient uptake
distribution during the growing season and the estimated total uptake
amount expected for the crop.
5. Adjust nutrient partition coefficients based on available soil core
and runoff data.
6. Refine the leaching, uptake, and partition parameters based on
observed runoff data and the expected sources of nutrient runoff,
i.e., surface, interflow, groundwater. '
Due to both data and resource limitations, the type -of detailed AGCHEM
calibration described by the above steps could not be performed in Phase II.
Even if the data were available for selected sites within the Bay drainage, the
calibrated model parameters would still need to be refined or adjusted to the
scale of the model segments and to represent the 'composite crop' within each
model segment. AGCHEM model calibrations were initially planned as part of the
Phase II effort for sites in the Patuxent, Rappahannock (Owl Run), and Potomac
(Nomini Creek) model segments; however, schedule and data constraints precluded
completion of these efforts. Plans are to complete these applications in future
studies either by the CBPO, the individual states, or through joint efforts.
Because of the above constraints and limitations, the focus of the AGCHEM
simulations was a 'regional' scale application designed to represent the expected
nutrient balances on each of the cropland categories. The literature was
reviewed to define typical nutrient balances for. each of the major crop
categories, and the AGCHEM simulations were calibrated to represent these
nutrient balances while maintaining the nonpoint nutrient loadings within the
range of expected values.
, 1
.110
-------
4.3.5.1 Expected crop nutrient balances
Tables 4.16 and 4.17, respectively show the typical nitrogen and phosphorus
balances expected for the major crops when the nutrients are applied at agronomic
rates meeting crop needs. Although there is a fairly large range in many of the
components of the nutrient balances, it is clear that nutrient applications and
crop uptake are the dominant portions of the balance. This is especially true
for corn which generally comprises more than half of the cropland (excluding hay)
for the model segments above the Fall Line; below the Fall Line, soybeans and
grains are more extensive but corn still comprises about 30X to 60Z of the
cropland in most segments.
Moreover, current total nutrient application rates including both fertilizer and
manure, as shown in Tables 4.5 and 4.6, are considerably larger than expected
crop nutrient needs. ป Table 4.18 shows the expected mean and range of crop uptake
rates along with the assumed crop yields on which the rates are based. The
uptake rates were initially derived from the general literature sources shown in
the table, and then adjusted through discussions with the NRTF based on
subsequent information provided by state representatives on actual crop yields
experienced in the individual states.
Because of the dominance of the application rates and crop uptake on the nutrient
balance, these two components were the primary focus of our efforts with AGCHEM
to simulate these balances. The extensive efforts to develop reasonable and
realistic nutrient application rates were described above in Section 4.3.3. The
critical importance of the application rates cannot be over-stated; these rates
defined the source and quantity of nutrients available for crop uptake and runoff
loadings. Based on the defined application rates, the steps in the AGCHEM
calibration were as follows:
1. Estimate soil nutrient storages and parameters from available
literature and state-supplied data. '
2. Calibrate uptake, mineralization, and immobilization (fixation) rates
to reflect expected nutrient balances.
3. Compare simulated runoff loadings and concentrations to available
literature data.
Initial estimates of both soil nutrients and parameters were derived from past
modeling studies using ACCHEM at a variety of field and watershed sites in
Georgia and Michigan as part of the ARM model development work (Donigian and
Crawford, 1976; Donigian et al., 1977), and in Iowa (Donigian et al., 1983;
Imhoff et al., 1983), Nebraska (Gilbert et al., 1982), Florida (Nichols and
Timpe, 1985), and Tennessee (Moore etal.. 1988). However, many of these studies
focussed on small field sites with little or no subsurface nutrient components,
and phosphorus modeling was either not attempted or de-emphasized due to limited
data, availability. In fact, the current application of AGCHEM is the most
comprehensive and extensive application to date, requiring more detailed
assessment of cropland nutrient processes than any previous study.
Ill
-------
TABLE 4.16 TYPICAL NITROGEN BALANCE FOR MAJOR CROPS WHEN
NITROGEN IS APPLIED AT AGRONOMIC RATES (Ib/ac)
Corn Soybeans Grains Hay
Inputs:
Fertilizer/Manure 100-160 25-35 50-100 30-60
Atmos. Deposition 7-10 7-10 7-10 7-10
Mineralization 25-40 25-40 25-40 25-40
Totals 132-210 57-85 82-150 62-110
' ' t.
Outputs:
Plant Uptake 120-150 25-40* 60-90 30-55
Surface Runoff 2-5 1-3 2-4 ,1-3
Leaching & Subsurface 10-25 10-15 5,-15 5-15
Runoff
Volatilization & 15-25 5-15 .10-20 10-20
Dinitrification
Totals 147-205 41-73 77-129 46-91
A Storage -15 to +5 +16 to +12 +5 to +21 +16 to +10
* - Represents uptake from the soil, approximately 25% of total
uptake, with 75% supplied by fixation (Tisdale et al.,'1985).
112
-------
TABLE 4.17 TYPICAL PHOSPHORUS BALANCE FOR MAJOR CROPS WHEN
PHOSPHORUS IS APPLIED AT AGRONOMIC RATES (Ib/ac)
Corn Soybeans Grains Hay
Inputs:
Fertilizer/Manure 20-40 10-30 10-30 10-30
Atmos. Deposition 0-1 0-1 0-1 0-1
Mineralization 2-5 2-5 2-5 2-5
Totals 22-46 12-36 12-36 12-36
Outputs:
Plant Uptake 20-30 12-20 12-22 12-25
Surface Runoff 1-2 0-1 0-1 0-1
Leaching & Subsurface 0-1 0-1 0-1 0-1
Runoff - -
Totals 21-33 12-22 12-24 12-27
A Storage +1 to -1-13 0 to +14 0 to +12 0 to +9
113
-------
4.3.5.2 Initial soil nutrient storages
Soil nutrient storages required by AGCHEM include initial values for all nutrient
forms (i.e.,, organic and inorganic) for each soil zone, including the surface,
upper, lower and groundwater zones as described in Section 4.3.1. Although a
variety of soil nutrient data was supplied by the state representatives on the
NRTF, the data was generally sporadic, incomplete for selected nutrient forms,
often based on different analytical techniques, and did not provide the vertical
spatial definition needed for the model soil zones. In addition,' limited
resources precluded efforts to establish different nutrient storage values for
each model segment. Consequently, the nutrient storages were developed from a
few selected detailed studies, past experience with AGCHEM, and general soils
information, and then compared with available state information to insure
consistency.
Parker et al. (1946) prepared national maps of Total N and P205 for the surface
foot of soil which showed percentages of 0.05 to 0.19 for these two forms,
respectively. Cunningham (R.L. Cunningham, 1991, personal communication) noted
that Total N was typically in the range of 0.13 to 0.16 X for Pennsylvania. Data
from Fox and Piekielek (1983) for a variety of cropland sites throughout
Pennsylvania were consistent with the earlier information. Bandel et al. (1975)
reports detailed information for N content at three research sites in Maryland
for different soil depths. Ibison (1990) reported Total N, Total P. and
Inorganic P soil concentrations for a variety of Maryland and Virginia sites in
coastal regions of the Bay as part of an effort to assess nutrient contributions
from eroding banks. '
The above data indicated extreme, variability in soil nutrient concentrations at
individual sites as a function of native soil characteristics, cropping
practices, historical fertilizer and manure applications, etc. Lacking the
resources to develop segment-specific values, we developed .initial nutrient
storages that were consistent with the general range of reported literature data
and state-supplied information. Brady (1990) reports that the vast majority of
nitrogen (i.e., 90-952 or more) is in the organic, form, while for.phosphorus the
inorganic form tends to dominate especially in subsurface soils. Moreover, for
our .cropland categories, nutrient applications and crop uptake primarily impact
the inorganic nutrient forms leading to significant variations in soil storages
during the annual simulations; this is consistent with the reported data for
-field sites. - '
Based on the- above information and data sources, we developed the following
organic (ON, OP) and inorganic (IN, IP) nutrient storages which were used, with
only slight adjustments, for all model segments: . .*
115
-------
The initial nutrient storages shown above provided a reasonable basis for
simulating the regional nutrient balances as part of the AGCHEM application.
Although the resources available for this effort were inadequate to analyze the
universe of available soil nutrient data and develop nutrient storages specific
to each model segment, any future efforts to refine or improve the AGCHEM
modeling should include a more detailed assessment of the variation in soil
nutrients throughout the region. .
4.3.5.3 Estimation of AGCHEM model parameters
In addition to the initial soil nutrient storages, the AGCHEM nutrient parameters
can be grouped into the transformation rates, sorption parameters, and plant
uptake rates. Initial values for most parameters were obtained from selected
literature,sources and past applications of AGCHEM in Georgia and Michigan as
part of the ARM model development work (Donigian and Crawford, 1976; Donigian et
al.,,-1977), and applications in Iowa (Donigian et al., 1983; Imhoff et al.,
1983). Nebraska (Gilbert et al., 1982), and Tennessee (Moore et al., 1988).
Stanford and Smith (1972) and Reddy et al. (1979) provided ranges of values for
nitrogen mineralization rates, along with temperature adjustment factors. Reddy
et al. (1979) and Reddy and Patrick (1984) also discussed expected values for
nitrification, denitrification, and volatilization rates for soils and sediments.
Recently, Deizman (1989) summarized literature values on rates for
mineralization, nitrification, and ammonia volatilization for soils amended with
organic wastes. Gilliam and Hoyt (1987) discuss the impacts of conservation
tillage on nitrogen soil processes and summarize much of the available
literature.
To represent the sorption process, a modified Freundlich isotherm is used in the
AGCHEM application in Phase II whereby a permanently bound concentration of the
chemical must be exceeded prior to partitioning by the Freundlich isotherm
parameters. Thus, for each soil layer three parameters are required to represent
NH4 end P04 sorption: the permanently bound concentration, XFIX (ppm), and the
Freundlich coefficient, Kl, and exponent, 1/N1 (Nl is the input parameter). The
literature on sorption was much more limited for NHซ than for P04, but a wide
range of values was reported based on soil conditions, sampling procedures, and
analytical procedures, especially for P0ซ. Work by Ardakani and McLaren (1977),
Simon (1989), Reddy (1989; 1990, personal communication) , and Reddy et al. (1988)
provided values from field observations for NH4. For the more highly sorptive
P04, a number of articles for a wide range of soils and conditions were reviewed,
including Berkheiser et al. (1984), Reddy et al. (1980), Sharpley et al. (1981),
Mansell et al. (1977), Fitter and Sutton (1971), Oloya and Logan (1980), Taylor
and Kunishi (1971), and McDowell et al. (1980).
In spite of the extensive Literature available on soil nutrient processes,
information on rate values that could be used directly in AGCHEM without
calibration was limited. In addition, the regional scale of the modeling and the
'composite crop' representation precluded the direct use of literature values
developed from specific field and plot-scale data collection efforts for specific
crops, soils, practices, etc. Consequently, the final AGCHEM parameter values
were determined primarily through calibration to represent the expected nutrient
1 *
117 .
-------
TABLE 4.19. NUTRIENT TRANSFORMATION AND .SORPTION PARAMETERS USED IN THE AGCHEM
SIMULATIONS
NITROGEN: Surface
Mineralization
Nitrification
Dinitrif ication
Mobilization/
Fixation
PHOSPHORUS:*
Mineralization 0:0007
Immobilization/ 10. Q
Fixation
FREUNDLICH SORPTION PARAMETERS:
NH^* Sorption Surface
Upper
Lower
G.W.
0.001
10.0
0.0
5,0/6.0
0.00015
5.0
0.0/0.005
2.0/3.0
0.00015
3.0
0.005
0.20
0.00
0.50
0.03
0.00
XFIX (ppm)
Kl
Ni
Sorption
XFIX (ppm)
Kl
Nl
5.0
1.0
1.5
25.0
5,0
1.5
0.00003
2.0
Upper
5.0
1.0
1.2
15.0
5.0
1.5
0.00005
0.10
Lower
0.7
0.5
1.2
10.0
5.0
1.5
0.00
0.00
G.W.
0.3
0:5
1.1
12.0
6.0
1.5
- Rates are in units of 'per day'
119
-------
d. The P parameters and representation of the P balance are the most
uncertain aspects of the AGCHEM simulations. Most previous AGCHEM
applications were limited to nitrogen simulations, largely because the
nitrogen cycle is much better documented and understood than the
phosphorus cycle in soils. The complexity of the. phosphorus cycle and
POt chemistry is well documented, but poorly understood quantitatively
(e.g., see Sample et.al. (1980), Olsen and Khasawneh (1980), and
sorption references cited above). Some of the limitations of AGCHEM
for phosphorus simulation were discussed -in Section 4.3.2. In
calibrating the phosphorus balance, it was necessary to use high plant
uptake rates, relatively low sorption parameters, and high
immobilization/fixation rates. The sorption parameters are consistent
with.the low end of the range of values in the literature, while rapid
fixation of applied P04 fertilizer (as bound or precipitated complexes
of iron, aluminum, calcium, etc.) is generally accepted.
4.3.6 AGCHEM Simulation Results
The AGCHEM model provides storages and fluxes of each nutrient form for each soil
zone and each simulated process for every timestep of the simulation.
Consequently, an extensive volume o.f information is produced for each model
segment. Tables 4.21, 4.22, and 4.23 summarize the simulation results for the
Conventional Tillage, Conservation Tillage, and Hay land categories, respectively
. for model segment 100 in the Juniata Basin. The tables provide annual totals and
the annual average for the. four-year simulation period, for the following
variables and fluxes:
Nutrient applications .
Runoff and components
Sediment loading. . '
N03 and Ammonia loadings and components
Sediment N loadings (ammonia and organic N) ,
P04 loadings and components
Sediment P loadings (PQ4 and organic P)
N and P plant uptake from each soil zone -
N and P mineralization '
Denitrification
Ammonia and PO* immobilization '
The results for model segment. 100 in the Juniata Basin are representative of the
AGCHEM results throughout the Bay drainage. Appendix C.2 includes AGCHEM summary
tables for the three cropland categories for selected model segments within each
subbasin.
Figures 4.14, 4.15, and 4.16 graphically show the inputs, losses, and runoff
losses of N and P for the Juniata model segment 100. Figure 4.14 shows the
composition of the nutrient inputs of N and P; Figure 4.15 shows the composition
and amounts of the N and P losses; and Figure 4.16 shows the composition of the
nutrient runoff losses. Finally, Table 4.24 summarizes the Total N and Total P,
and components, in runoff losses from selected model segments for relative
comparisons. All the information in these tables and figures are partial
extracts of model results included in the tables in Appendix C.2.
121
-------
TABLE 4.22 AGCHEM RESULTS FOR JUNIATA SEGMENT 10P (LT COMPOSITE CROP)
Pert. + Manure
Application
Runoff (in)
1 Surface
Interflow
GW
Total
Sed. Loaa (t/a)
Nutrient Loss, (Ib/a)
N03
Surface
Interflow
GW
Total
NH3
Surface
Interflow
GW
Total
NH3 Sed. (Ib/a)
ORG. N Sed. (Ib/a)
Total N (Ib/a)
P04
Surface
Interflow
GW
Total
P04 Sed. (Ib/a)
ORG. P Sed. (Ib/a)
Total P (Ib/a)
Plant Uptake (Ib/a)
Nitrogen
Surface
tipper
Lower
Total
Phosphorus
Surface
Upper
Lower
'Total
Net Mineralization (Ib/a)
Nitrogen
Phosphorus
Denitrification (Ib/a)
NH3 IMMOB. (Ib/a)
PO4 IMMOB. (Ib/a)
1984
1221b N/a
531b P/a
5.98
6.67
8.15
20.80
0.61
0.31
*' 13.23
4.87
18.41
1.12
1.25
0.05
2.42
6.53E-03
2.57
23.41
0.65
0.57
0.00
1.22
2.69E-02
0.68
1.93
0.01
59.53
32.57
92.11
0.04
20.01
9.82
29.87
/a)
33.03
2.44
1.48
19.73
29.50
1985
1221b N/a
531b P/a
4.73
4.73
6.36
15.82
Q.56
0.63
8.15
3.33
12.11
2.24
0.81
0.01
3.06
7.33E-03
2.36
17.54
0.86
0.48
0.00
1.34
2.71E-02
0.63
2.00
0.04
67.48
27.78
95.30
0.04
19.37
2.56
21.97
32.26
2.13
1.20
20.77
27.42
1986
1221b N/a
531b P/a
4.26
5.78
6.59
16.63
0.29
0.43
7.66
3.25
1-1.34
1.14
0.89
0.01
2.04
3.53E-03
1.21
14.59
0.91
0.73
0.00
1.64
1.33E-02
0.32
1.97
0.06
67.58
29.21
96.85
0.07
21.14
1.89
23.10
32.54
2.23
1.16
21.46
25.56
1987
1221b N/a
531b P/a
2.18
3.93
6.48
12.59
0.12
0.05
7.94
3.08
11.07
0.38
0.41
0.01
0.80
1.28E-03
0.50
12.37
0.34
0.29
0.00 .
0.63
5.56E-03
0.13
0.77
0.03
67.61
30.49
98.13
0.05
20.94
1.79
22.78
32.52
2.19
1.16
21.26
27.47
Mean
1221b N/a
531b P/a
4.29
5.28
6.90
16.46
0.40
0.36
9.25
3.63
13.23
1.22
0.84
0.02
2.08
4.67E-03
1.66
16.98
0.69
0.52
0.00
1.21
1.82E-02
0.44
1.67
0.03
65.55
30.01
95.60
0.05
20.37
4.02
24.43
32.59
2.25
1.25
20.81
27.49
123
-------
Plant Nutrient Inputs (Ib/ac)
Juniata Segment 100
Conventional Tillage
Fertilizer
114
Mineralization
30
Fertilizer
48
Mineralization
i
Nitrogen Phosphorus
Conservation Tillage
Fertilizer
122
Mineralization
33 .
Fertilizer
53 '
Mineralization
1
Nitrogen Phosphorus
Hay
Fertilizer
44
Mineralization
30
Fertilizer
16
Mineralization
2
Nitrogen Phosphorus
Figure 4.14 Plant nutrient inputs - Juniata Segment 100.
125
-------
Nutrient Runoff Losses (Ib/ac)
Conventional Tillage
Juniata Segment 100
NH4-N
2.3 11.7%
Organic N
2.4 12.2%,
Organic P
0.67 34.9%
NO3-N
14:9 76.0%
PO4-P (Ads.)
0.03 1.6%
Nitrogen
PO4-P (Dlss.)
t22 63.5%
Phosphorus
Conservation Tillage
Juniata Segment 100
NH4-N
2.112.4%
Organic N
17 10.0%,
N03-N
13.2 77.6%
Nitrogen
PO4-P (Diss.)
1272.3%
Phosphorus
\Hay
Juniata Segment 100
Organic N
0.619.0%
NO3-N
5.9 86.6%
Nitrogen
Organic P
0.16 20.8%
PO4-P (Ads.)
" 0.011.3%
PO4-P(DiSS.)
0.6 77.9%
Phosphorus
1984-1987 Annual Mean Values
Figure 4.16 Nutrient runoff'losses - Juniata Segment 100,
127
-------
Review of the AGCHEM model results presented in these tables and figures, and
included in Appendix C, indicate the following:
a. 'The nutrient applications of fertilizer and manure and the associated
plant uptake of N and P dominate the input and output; portions,
respectively, of the nutrient balance of the croplands.
b. Mineralization is a significant source of plant available N, but not
a significant source for plant available P in these simulations.
c. Plant uptake is a direct function of the nutrient application rates on
each segment and land use category. Annual uptake of N is typically
702 to 902 of .the application, while for P the uptake is typically
. about 502 of the application amount. However, there is considerable
variation from segment to segment due to variations in the application
rates.
d. Uptake amounts were generally less than calculated 'composite target'
amounts when the segment application rates were similar to or less
than the targets, and greater than the targets when the application
rates were greater than the targets by a significant amount, this
- simply indicated that on individual segments higher application rates
than expected (or required by the crop) produced higher yields and
uptake than was used in our calculation of the ''composite' target
uptake.
e. Plant available N is generally reduced by the processes of
denitrification, NH3 volatilization, fixation, immobilization, and
leaching; plant available P is reduced primarily by fixation and.
immobilization- Since NH3 volatilization and N and P fixation are not
. specifically modeled, .immobilization was used as the primary
mechanisms for losses of plant available nutrients. HH3
immobilization was typically in the range of 152 to 302 of application
amounts, while P04 immobilization was typically about 502 of the
. application. Denitrification was assumed to be small.
f. The calibration process focused on maintaining plant uptake amounts
close to target levels, representing reasonable losses of plant
available nutrients, and keeping N03, NH3, and P0^ runoff losses in the
range of expected values. T
g. Total N and Total P runoff losses are generally in the range of 10Z to
.302, and 52 to 152 of application amounts, respectively. N03 is the
largest component of Total N losses, followed by soluble NH3, Organic
N and sorbed NH3. For Total P, soluble P04 is the largest, followed
by Organic P and sorbed P0ซ. The loadings of sediment associated
organics, and sorbed NH3 and P04 are direct functions of the calibrated
sediment loading rate, which includes the impact of the assumed
delivery ratio of 152. Consequently, the sediment associated loadings
are considerably less than expected edge-of-field values due to
deposition processes affecting delivery at the scale of the model
segments. .
129
-------
Table 4.25 SUMMARY OF OBSERVED CONCENTRATIONS FROM SELECTED FIELD MONITORING STUDIES (Means and Ranges, mg/l)
NH4-N
Mean Range
N03-N
Mean
Range
Total
Mean
N
Range
Ortho-P
Mean Range
Total P TN/TP ratio
Mean Range
Owl Run Watershed (In Segment 230) 7.8 ซq. ml.
OOA
OOB
ooc
000
Nomlni
ONI
QN2
- 4.5
- 0.2
- 1.8
- 1.3
q.
q-
q.
q.
ml.
ml.
Ml.
nl.
Cr. Watershed
- 5.8
- 0.9
q.
aq.
mi.
ml.
2.44
0.61
0.32
0.43
(In Potomac
0.33
0.30
0.098-28.9
0.0-2.6
0.0-1.2
0.0-1.35
Baain, Below
0.01-1.48
0.0-1.58
2.90
1.61
, 1.17
2.60
0.0-9.78
0.07-6.5
0.0-3.9
0,0-7.75
Fall Line) - 6.7 aq.
0.63
1.15
0.11-1.86
0.12-2.09
12.60
7.92
6.29
8.40
mi.
5.75
7.85
1.7-89.5
2.4-33.1
0.0-15.2
0.0-28.9
1.80-15.1
2.46-39.7
0.68 0.02-4.9
0.55 0.05-1,38
0.04 0.0-.15
0.26 0.0-1.56
0.03 0.0-0.14
0.03 0.0-0.16
2.96 0.29-17.6 4.3
1.48 0.29-2.95 5.4
0.94 0.0-16.2 6.7
1.60 0.0-9.4 . 5.3
0.91 0.05-12.9 6.3
1.57 0.10-16.7 5.0
Jug Bridge, Monocacy River (in Segment 210) : 817 sq. ml. .
Base Flow data 0.37 0.05*0.98 2.100.05-3.57 3.23 0.65-5.37 0.23 0.08-0.52 0.44 0.14-0.88 7.3
Storm Event 0.25 0.11-0.65 1.77 0.05-3.10 4.32 0.95-9.1 0.16 0.01-.44 1.10 0.06-1.9 3.9
Patuxent Watershed - Field Sites (Segment 340) -
Station H01 - 5 ac. 0.09 0.0-0.26 0.88 0.07-5.3 3.87 0.77-18.8 0.63 0.01-1.50 1.21 0.30-5.2 3.2
Station CA1 - 8 ac. 0.07 0.02*0.24 0.33 0.02-0.70 7.49 1.60-45.0 0.13 0.03-0.40 1.54 0.40-3.6 4.9
-------
Ul
OJ
Table 4.27 SUMMARY OF PATUXENT (SEGMENT 340) SIMULATED CONCENTRATIONS by ACCHEM (Means and Ranges, mg/l)
NH4-N N03-N Total N P04-P - Total P TN/TP ratio
- '
Mean Range Mean 'Range Mean Range ' Mean Range Mean Range
CONVENTIONAL TILLAGE
Surface & Interflow 0.38 0.04-2.51 2.58 0.0-17.60 9.96 1.83-37.10 0.41 0.12-1.90 2.340.35-10.80 4.3
Total 0.02 0.01-2.33 1.41 0.05-14.60 1.60 0.50-36.10 0.009 0.00-1.54 0.0560.0-10.50 28.6
CONSERVATION TILLAGE
Surface t Interflow 0.45 0.06-4.74 2.30 0.01-16.40 10.30 1.68-35.10 0.41 0.15-2.61 2.42 0.35-6.31 4.3
Total 0.02 0.02-4.25 1.26 0.07-14.90 1.41 0.44-31.60 0.007 0.0-2.12 0.042 0.0-6.03 33.6
HAY
Surface & Interflow 0.13 0.01-1.53 0.35 0.0-1.16 5.38 0.31-14.20 0.26 0.04-1.20 1.57 0.11-3.96 3.4
^
Total 0.01 0.0-1.42 0.99 0.06-2.96 1.07 0.45-13.40 0.004 0.0-1.03 0.023 0.0-3.72 46.5
Notes: 1. Surface I Interflow concentrations are calculated as daily load/dally flow for these components
when the dally flow (surface and Interflow only) exceeds 0.1 in.
2. Total concentrations are based on both storm (I.e. surface and Interflow) and
baseflow values for the entire simulation period
-------
The extreme values of the ratio, in the range of 100 to 500, for total
concentrations in the Rappahannock (Segment 230) in Table 4.26 are due to
extremely low mean TP concentrations as reflected in the observed water quality
data. Since baseflow will have higher N and lower P than storm runoff, the TN/TP
ratio will increase as baseflow and subsurface contributions increase as
reflected in the AGCHEM results.
4.3.7 AGCHEM Conclusions and Recommendations
The AGCHEM application to the cropland areas of the Chesapeake Bay drainage has
satisfied the objectives outlined at the beginning of Section 4.3, namely to (1)
represent nutrient balances for the major cropland categories in each model
segment, (2) allow investigation of sensitivity to climate variations and
agricultural practices, and (3) provide a mechanism to project impacts of
nutrient management alternatives for agricultural cropland. The results are
consistent with expected nutrient balances, observed ranges of runoff
concentrations, and expected ranges of nutrient loadings from cropland areas
(discussed below in Section 4.4).
The AGCHEM approach to modeling the nutrient balances on agricultural cropland
Is exactly the type of approach recommended by the Chesapeake Bay Honpoint Source
Evaluation Panel in their report to the U.S. EPA and the GBP Executive Council.
In March 1990, the EPA Administrator convened the Panel to "assess the
effectiveness of current efforts to reduce nonpoint source loadings of nutrients
entering the Bay system". They reviewed and evaluated the effectiveness of a
wide range of nonpoint source programs, including program design, implementation,
budgets, modeling and assessment methods, and research efforts. One of the key
recommendations of the Panel related to new approaches to nutrient control was
as follows: ;
"The Panel recommends that the Bay' jurisdictions and the federal
agencies develop a mass balance accounting system, where nutrient
loadings are balanced by the nutrients removed from the.system plus'
those which are introduced and stored" (Chesapeake Bay Nonpoint
Source Evaluation Panel, 1990). ,
The AGCHEM application provides an initial assessment of this 'mass balance' for
cropland areas, a tool for assisting in the 1991 Re-evaluation of the 40X
nutrient reduction goal of the Bay Agreement, and a framework for future efforts
on nutrient management based on the mass balance approach. . Recommendations to
improve the use and application of AGCHEM by CBP include the following:
a. A detailed review and refinement of the current fertilizer and manure
application rates, and -changes expected under nutrient management
alternatives should be initiated since these rates are the starting
point for nutrient reduction from cropland. Plant uptake amounts under
reduced nutrient application rates will need to.be investigated to
ensure that crop needs are met; plant uptake rates may need to be
adjusted.
b. Detailed site applications of AGCHEM should be diligently pursued by
/
135
-------
In the Phase I, the .first steps in the nonpoint calibration effort involved a
review and evaluation of nonpoint loading rates associated with individual land
uses and nonpoint parameters used in the original Watershed Model. The goal was
to define the expected range of loading rates from the available literature, as
a basis for evaluating and calibrating the model predicted loading rates, and
determine if any changes or adjustments to the original nonpoint parameters could
be justified. The State representatives on the CBP Nonpoint Source Workgroup
provided data summaries of monitoring projects and studies conducted in their
respective regions to supplement the efforts of the CBPO and AQUA TERRA, on this
task. Table 4.29, developed in Phase I, provides a brief summary of the results
of this effort, with ranges of loading rates for individual nutrient forms for
the major land use categories in the Watershed Model.
Unfortunately, the available data was not sufficiently extensive or detailed to
allow delineation of'different loading rates for different portions of the Bay
drainage area. As shown in Table 4.29 and in the original literature, the rates
are quite variable with cropland showing the greatest variability and forest the
least variation. For urban pervious and impervious areas, the average annual
National Urban Runoff Program (NURP) loads, as complied by Schueler (1987) were
used to supplement the information in Table 4.29 and guide the calibration
adjustments. In summary, as shown in Table 4.29, the model predicted rates could
fall within a relatively large range and still be consistent with the literature
data. - . - ^ "' '
In fact, the second part of this effort involved an assessment of consistency
between the original model rates and the literature. In order to assess whether
refinements to the potency factors and washoff parameters were Justified, we
compared the model predicted loading rates (which are based on the input potency
factors and washoff parameters) with the loading- ranges shown in Table 4.29. Our
reasoning was that if the model loading rates were consistent with the literature
values, we did not have a basis for changing the parameters except through
calibration. The comparison showed that the model rates were consistent with the
literature values, partly due to the wide range in the literature.
Consequently, no changes in potency factors or washoff rates were made initially
until the calibration indicated that changes were justified.
However, the subsurface concentrations used in the original Watershed Model were
based on calibration at instrearn sites during low flow periods. Thus, the sane
concentrations were used for all land uses. In the Phase I work, we developed
different subsurface concentrations for each land .use for consistency with the
hydrologic/land use variability and to better represent effects of land use
changes. Initial subsurface concentrations were estimated for NH3, N03, P04,
BOD, and 00 from a variety of sources including past HSPF applications within the
Bay drainage, data from sites within the Patuxeat and Monocacy rivers, and
selected literature review articles. Generally, the highest pollutant values
were estimated for cropland areas, with conservation tillage having somewhat
higher values than conventional tillage. Urban subsurface concentrations were
the next highest, followed by pasture, with forest areas having the lowest.
concentrations. Subsurface 00 values were assumed to be the'same for all land
uses, and the same seasonal (i.e. monthly) variations as used in the original
model were imposed. These initial values were then adjusted as needed during
137
-------
the calibration process based on the observed data and model predictions for the
calibration sites throughout the drainage area.
In Phase II, Table 4.30 was developed by the CBPO from further analyses of field
studies specific to the .Chesapeake Bay drainage area; many of the studies were
provided by members of the NRTF. Note that a number of the entries in Table 4.30
indicate 'no data' due to insufficient information for selected land uses and
individual nutrient species. Table 4.31 from Beaulac (1980) and Reckhow et al
(1980) show the range and median of expected loading rates (or export
coefficients) only for Total N and Total P based on nation-wide data from field
studies. Typically, Total N and Total P are better defined than the individual
components of NH3, PO*. and organics; N03 loading rates are usually better
defined than the other species since it is the most commonly measured nutrient
form. ,
Tables 4.32, 4.33, and 4.34 are tabulations of the simulated loading rates from
the Phase II simulations for each nutrient form, plus BOD, for each year of the
simulation, along with the mean annual value, for model segments 60, 180, and
280, respectively. These three segments are the largest segments in each of the
three major basins Susquehanna, Potomac, and James. Similar tabulations are
provided for additional model segments in Appendix C. . Note that Total N and
Total P are calculated from the other nutrient forms and BOD as described in
Section 5.0 and the table footnote.
Table 4.35 summarizes the mean annual rates for 17 model segments'throughout the
Bay drainage, area, including both above and below Fall Line segments; the
detailed tabulations for these model segments are included in Appendix C.I.
Table 4.3.6 provides a capsule summary of the mean and ranges of loading rates for
the segments lisced in Table 4.35; although this summary is not based on all
model segments (segment tabulations are being currently compiled), ' it does
provide a good indication of the simulated mean and range for each land use and
selected nutrient forms. Comparing the mean annual rates in Table 4.36 with the
expected means and ranges from Tables 4.29 through 4.31 indicates the following
general conclusions:
a. Generally, the simulated annual loading rates are within the range of
expected values shown in Table 4.29 with some deviations. Annual
rates for P0ป from forest and pasture, and NH4 from forest occasionally
tend to be toward the lower end of the defined range. Annual P04
rates from the cropland areas are somewhat higher than the defined
range.
b. For non-cropland categories., the Total N and Total' P simulated values
compare favorably with both the expected means and ranges.
c. For the cropland categories of Conventional Tillage, Conservation
Tillage, and Hay, the Total N and Total P simulated values are
generally close to the mean lumped cropland data; the Hay values are
generally less than the mean, while the tillage categories are usually
greater than the mean but well within the observed range.
139
-------
TABLE 4.31 NUTRIENT EXPORT COEFFICIENTS FROM BEAULAC (1980) AND
RECKHOW ET AL. (1980)
Phosphorus
Land Use
Forest
Pasture
Cropland ,
Urban
Land Use
Forest
Pasture
Cropland
Urban
Minimum
0.017
0.12
0.09
0.17
Minimum
'.
1.23
1.32
1.96
1.32
Median
0.18
0.72
0.89
0.98
Nitrogen
Median
2.32
4.63
8.03
4.91
Maximum
0.74
4.37
1.96
5.56
Maximum
5.58
27.52
71.0
34.3
Source: Nutrient export coefficients from (Beaulac (1980) and
Reckhow et al. (1980).
141
-------
TABLE 4.33 WATERSHED MODEL SEGMENT 180 ANNUAL LOADING RATES (LB/AC/YR)
LAND USE
FOREST
CONV.TILL
CONSER. TILL
PASTURE
URBAN-PERV.
HAY
URBAN-IMPV.
MANURE SEC.
YR
84
85
66
87
AVERAGE
84
85
86
87
AVERAGE
84
85
86
87
AVERAGE
84
85
86
87
AVERAGE
84
85
86
87
AVERAGE
84
85
86
87
AVERAGE
84
85
86
87
AVERAGE
84
85
86
87
AVERAGE
NH4
0.072
0.053
0.046
0.059
0.057
2.73
3.05
1.38
2.79
2.49
3.82
2:08
1.42
2.98
2.58
0.10
0.07
0.06
0.08
0.08
0.31
0.20
0.16
0.22
0.22
0.42
0.13
0.15
0.30
0.25
1.42
1.34
1.36
1.39
1.34
243.0
198.2
170.7
202.0
203.5
K03
4.12
3.41
2.99
3.47
3.50
22.71
12.74
9.15
13.70
14.58
21.85
11.20
7.06
11.63
12.94
6.08
5.16
4.40 '
5.10
5.19
7.61
6.23
5.20
6.19
6.31
4.68
1.72
1,78
2.18
2.59
3.04
2.85
2.89
2.95
2.93
60.7
49.5
42.7
50.5
50.9
P04
0.012
0.005
0.004
0.009
0.008
1.46
1.29
0.92
0.97
1.16
2.06
1.36
1.00
1.16
1.40
0.06
0.04
0.03
0.04
0.04
0.25
0.14
0.11
0.17
0.17
2.14
0.69
0.88
1.49
1.30
0.66
0.62
0.63
0.64
0.63
60.7
49.5
42.7
50.5
50.9
1 BOD
7.91
2.81
2.93
5.09
4.69
400.00
86.00
84.90
174.00
186.23
181.00
44.80
31.10
80.20
84.28
48.00
29.80
25.00
35.30
34.53
54.90
29.50
22.40
36.60
35.85
203.00
42.70
27.40
89.10
86.25
86.50
81.00
82.30
84.00
83.45
4,251.6
3.468.5
2.987.8
3.535.7
3.560.9
ORC-N
5.86
1.20
1.47
2.44
2.74
4.66
0.89
0.67
1.78
2.00
1.97
0.36
0.27
0.79
0.85
1.822.1
1,486.5
1,280.5
1.515.3
1.526.1
ORG-P
1.61
0.33
0.41
0.67
0.76
1.23
0.24
0.18
0.47
0.53
1
0.53
C.W
0.07
0.21
0.23
364.4
297.3
256.1
303.1
305.2
TN
4.61
3.61
3.19
3.80
3.80
37.66
18.36
13.35
.21.70
22.77
33.21
14.88
9.64
17.67
18.85
8.72
6.81
5.79
7.05
7.09
10.83
7.99
6.55
8.35
8.43
10.30
2.89
2.63
4.69
5.13
.04
.48
.61
.79
.73
2,125.8
1.734.2
1,493.9
1,767.9
1.780.5
TP
0.072
0.026
0.026
0.048
0.043
3.98
1.82
1.52
2.04
2.34
3.70
1.70
1.25
1.81
2.12
0.42
0.27
0.22
0.31
0.31
0.67
0.37
0.28
0.45
0.44
3.13
0.88
1.01
1.90
1.73
1.31
1.23
1.25
1.28
1.27
425.2
346.8
298.8
353.6
3S6.1
* For all land uses, except CUT. CST. HAY and MANURE land uses, TN and TP are calculated by:
TN * NH3 * N03 + 0.053*800 TP ซ P04 * 0.0076*600
For CNT. CST, and HAY land uses, TN and TP are calculated by:
TN ซ NH3 ป N03 + ORGN * 0.3*0.053*800 TP P04 ป ORGP * 0.3*0.0076*800
For the MANURE land use, TN and TP are calculated by:
TN * NH3 + N03 + ORGN TP ซ P04 * ORGP
143
-------
TABLE 4.35 SUMMARY OF AVERAGE ANNUAL LOADING RATES FOR EACH LAND USE IN SELECTED
MODEL SEGMENTS (LB/AC/YR) '
SEGMENT
SEGMENT 10
E. Branch Susquehama
SEGMENT 30
E. Branch Susquehanna
' -
SEGMENT 40
E. Branch Susquehanna
.. /
SEGMENT 50
U. Branch Susquehanna
SEGMENT 60
U. Branch Susquehanna
1 . '
SEGMENT 110
Lower Susquehanna
SEGMENT 160 .
N. Branch Potomac
LANOUSE
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till'
Cons. Till
Hay
Pasture
Manure Seg
Forest?
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg .
forest
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest:
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
NM4
2.39
2.12
0.76
0.08
212
0.04
0.59
2.31
2.83
1.02
0.09
255
0.05
1.07
3.21
3.15
1.13
0.09
287
0.05
1.34
6.95
8.77
1.41
0.11
301
0.10
0.65
.4.93
5.14
1.33
0.10
261
0.09
0.72
2.08
2.39
0.49
0.21
221
0.09
0.95
4.64
3.34
0.31
0.11
240
0.06
0.66
N03
13.80
12.10
8.84
3.00
53
1.91
4.53
13.00
12.80
10.00
5.89
64
5.10
7.17
12.60
11.20
8.58
5.88
72
6.02
7.03
22.00
20.50
15.00
4.26
75
3.00
5.97
20.80
19.40
14.20
5.16
65
4.62
7.74
11.60
11.90
4.79
7.60
55
6.04
8.26
13.20
9.83
3.83
6.98
60
3.61 '
5.58
P04
0.95
0.93
1.10
0.02
53
0.01
0.31
0.96
1.07
1.17
0.02
64
0.01
0.38
1.12
0.96
1.65
0.02
72
0.01
0.66
1.38
1.38
1.30
0.02
75
0.01
0.26
1.19
1.22
1.06
0.02
65
0.01
0.33
1.31
1.37
0.75
0.01
55
0.01
0.36
0.96
1.09
2.29
0.08
60
0.01
0.38
ORG-N
3.91
2.48
1.20
1.06
1.593
0.27
3.28
3.49
2.37
1.08
1.39
1.915
0.34
3.56
5.79
3.35
1.29
1.88
2,155
0.43
4.13
4.25
3.43
1.29
1.06
2.257
0.35
3.88
4.62
3.98
1.36
1.62
1.962
. 0.33
3.64
18.02
9.71
5.92
3.39
1.657
0.35
5.99
6.76
4.03
1.93
2.26
1.799
0.24
3.72
ORG-P
0.76
0.49
0.23
> 0.15
319
0.04
0.42
0.69
0.46
0.21
0.20
383
0.05
0.50
1.14
0.70
0.25
0.27
431
0.06
0.58
0.79
0.51
0.23
0.15
452
0.05
0.56-
0.95
0.58
0.26
0.23
393
0.05
0.52
3.06
1.73
0.97
0.49
.332
0.04
0.85
1.16
0.70
0.33
0.32
360
0.04
0.54
TN
20.10
16.70
10.80
4.U
1,858
2.22
8.40
16.80
18.00
12.10
7.37
2.234
5.49
11.80
21.60
17.70X
11.00
7.85
2,514
6.50
12.50
33.20
32.70
17.70
5.43
2,633
3,45
10.50
30.55
28.52
16.89
6.88
2,287
5.04
12.10
31.70
24.00
11.20
11.20
1.933
6.48
15.20
24.60
17.20
6.07
9.35
2.099 ,
3.91
9.96
TP
1.71
1.42
1.33
0.17
372
0.05
0.73
1.65
1.53
1.38
0.22
447
0.06
0.88
2.26
1.66
1.90
0.29
503
0.07
1.24
2.17
1.89
1.53
0.17
527
0.06
0.82
2.14
1.80
1.32
0.25
458
0.06
0.85
4.37
3.10
1.72
0,50
387
0.05
1.21
2.12
1.79
2.62
0.40
420
0.05
0.92
145
-------
TABLE 4.35 (Continued)
SEGMENT
SEGMENT 310
Appomattox
SEGMENT 340
Patuxent
SEGMENT 380
Chester
SEGMENT 400
Choptank
SEGMENT 430
Pocomoke
SEGMENT 450
Lower Susquehanna
SEGMENT 500
Lower Patuxent
LANDUSE
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban"
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till
Cons. Till
Hay
Pasture
Manure Seg
Forest
Urban
Conv. Till
Cons. Till
Hay
Pasture .
Manure Seg
Forest
Urban
NH4
1.90
1.68
0.57
0.05
242
0.05
0.69
1.55
1.22
0.11
0.07
216
0.04
0.64
1.59
1.14
0.25
0.07
190
0.04
0.45
1.88
1.18
0.30
0.07
190
0.04
0.48
3.20
2.19
0.16
0.07
190
0.04
0.55
1.55
1.16
0.21
0.08
212
0.05
0.68 .
1.15
0.71
0.13
0.08
211
0.05
0.58
H03
12.50
11.50
4.33
1.51
61
0.68
2.35
10.20
8.69
2.47
2.64
54
1.33
4.72
11.20
9.92
3.59
4.73
48
2.20
4.69
11.90
10.20
3.80
4.69
48
2.28
4.83
17.80
16.60
2.94
4.54
.*
2.30
*ป
11.00
10.20
3.10
5.55
53
2.35
5.25
8.27
6.42
2.57
4.95
53
2.35
I
5.00
P04
0.93
0.89
1.01
0.03
61
0.01
0.30
0.63
0.46
0.15
0.03
, 54
0.01
0.28
0.42
0.35
0.18
0.04
48
0.01
0.22
0.53
0.44
0.22
0.04
48
0.003
0.23
1.34
1.22
0.21
0.04
48
0.004
0.26
0.89
0.63
0.35
0.03
:53
0.003
0.31
0.56
0.39
0.15
0.04
53
0.003
0.27
ORG-N
8.60
5.02
2.82
2.31
1,815
0.28
2.93
6.75
ซ.59
2.18
0.53
1.623
0.22
3.98
4.71
3.14
1.60
1.04
1.425
0.26 .
2.11
3.82
2.52
0.92
0.95
V*25
0.10
2.20
3.80
2.41
0.74
1.74
1,425
0.13
2.13
3.75
2.94
1.71
0.89
1,594
0.10
2.31
9.58
6.07
4.91
1.25
1,583
0.14
2.29
ORG-P
1.62
0.99
0,54
0.33
363
0.04
0.42
1.72
1.15
0.53
0.07
325
0.03
0.57
0.93
0.62
0.29
0.15
285
0.03
0.30
0.68
0.46
0.16
0.13
285
0.02
0.31
0.68
0.44
0.13
0.25
285
0.02
0.30
0.86
0.65
0.31
0.13
319
0.02
0.34
,2.33
1.44
0.89
0.18
316
0.02
0.33
TN
23.00
18.20
7.72
3.87
'2,118
1 .01
5.97
18.50
14.50
4.76
3.24
1,893
1.59
9.34
17.50
14.20
5.44
5.84
1,663
2.50
7.25
17.60
13.90
5.02
5.71
1,663
2.42
7.51
24.80
21.20
3.84
6.35
1.663
2.47
7.42
16.30
14.30
5.02
6.52
1.859
2.50
8.24
19.00
13.20
7.61
6.28
1,847
2.54
7.87
TP
2.55
1.88
1.55
0.36
424
0.05
0.72
2.35
1.61
0.68
0.10
379
0.04
0.85
1.35
0.97
0.47
0.19
333
0.04
0-52
1.21
0.90
0.38
0.17
333
,0.02
0.54
2.02
1.66
0.34
0.29
333
0.02
0.56
1.75
1.28
0.66
0.16
372
0.02
0.65
2.89
1.83
1.04
0.22
369
0.02
0.60
147
-------
TABLE 4.36 OVERALL SIMULATED MEAN AND RANGE of LOADING RATES FROM SELECTED
SEGMENTS LISTED IN TABLE 4.35 (LB/AC/YR)
NH3-N N03-N P04-P TN TP
FOREST
Mean 0.06 2.77 0.01 3.09 0.04
Range 0.04-0.10 0.55-6.04 0.003-0.03 0.81-6.50 0.02-0.09
CONVENTIONAL TILL
Mean 2.57 13.19 0.95 22.04 2.18
Range 0.90-6.95 8.27-22.00 0.42-2.17 , 15.20-33.20 1.21.4.37
^ " ' .
CONSERVATION TILL
Mean 2.24 11.69 0.91 17.90 1.68
Range 0.71-8.77 6.42-20.50 0.35-1.97 11.0-32.70 0.90-3.10
PASTURE
Mean
Range
URBAN*
Mean
Range
HAY
Mean
Range
MANURE
Mean.
Range
0.10
0.05-0.25
0.65 !.
0. 21-1. 34
0.51
0.11-1.41
230
190-301
4.70
1.01-8.53
5.49
1.27-9.73
6.13
2.47-15.00
57
-48-75
0.06
0.01-0.28
0.33
0.16-0.66
0.87
0.15-2.29
57
48-75
6,47
2.28-11.20
9.27
3.43-15.20
8.72
3.84-17.70
2010
1663-2633
0.29
0.10-0.91
0.78
0.42-1.24
1.26
0.34-2.62
402
333-527
* - Includes area weighted totals from both pervious and impervious segments..
149
-------
SECTION 5.0
RIVER TRANSPORT AND TRANSFORMATION SUBMODEL
This section discusses the HSPF modules used to simulate the instream transpo'rt
and water quality processes considered in the Phase II Watershed Model. Each
Above Fall Line (AFL) Watershed Model segment is associated with a corresponding
river channel/reservoir which receives nonpoint, point, and atmospheric
deposition loadings. These inputs of flow, heat, sediment, and inorganic and
organic nutrients then undergo the various instream processes, including
phytoplankton growth, and are transported downstream toward the Bay. Table 4.2
lists the water quality constituents simulated in Phase II.
A major focus of Phase II was to include a more detailed representation of
instream nutrient processes, particularly with respect to sediment-nutrient
interactions. With the exception of a simple organic N and P settling algorithm,
previous versions of the Watershed Model have not considered sediment transport
and the associated sediment-nutrient interactions that are critical to an
accurate representation of the water quality of the .rivers in the basin.
Consequently, the Phase II effort included significant modifications to the
nutrient simulation routines in HSPF to allow inorganic N (ammonium) and P
(orthophosphate) to adsorb to sediment and undergo settling, resuspension, and
transport. Also, the denitrification algorithm was modified to provide a more
flexible representation of nitrate losses via this process.
The methods used to simulate instream processes in run-of-the-river and reservoir
reaches are described in the following subsections. Additional details of the
model equations and algorithms are contained in the HSPF Users Manual (Johanson
et at, 1984).
5.1 OVERVIEW OFTHE INSTREAM MODEL
Within the HSPF program, the RCHRES module sections are used to simulate
hydraulics, sediment transport, water temperature, and water quality processes
that result in delivery of flow and loadings to the Bay. The structure chart of
the RCHRES module sections is shown in Figure 5.1. Each module section consists
of a group of subroutines that perform the individual process simulations. For
example the HYDR module performs the flow/hydraulic simulation that-drives the
sediment transport and pollutant advection. The water quality processes in the
Watershed Model are performed by the RQUAL module, which is futther subdivided
into submodules for performing simulation of different water quality processes
as shown in Figure 5.2. All processes in the Watershed Model are simulated using
a one hour time step.
151
-------
RQUAL
OXRX
Simulate
primary
dissolved
c-xygen and
BOD
reactions
Simulate
constituents
involved in
biochemical
transform-
ations
NUTRX
Simulate
primary
inorganic
nitrogon and
.phosphorus
reactions
PLANK
Simulate
behavior of
plankton and
associated
reactions
PHCARB
Simulate pH
and inorganic-1
carbon
reactions
Figure 5.2 Structure chart of the RQUAL section of the RCHRES module.
153
-------
Geological Survey (Arnibruster, 1977) and the Susquehanna River Basin Commission
(Jackson, 1979). ,
5.2.2 Water Temperature'
Water temperature is simulated in the Watershed Model because of its fundamental
effect upon water quality processes. Dissolved oxygen saturation level and the
rates of BOD decay, nitrification, and phytoplankton growth are all directly
affected by water temperature.
Instream water temperature processes are simulated by the HTRCH module, shown in
Figure 5.3. Changes in heat content of a reach result primarily from advective
transport, runoff and rainfall entering the reach, and heat transfer processes
across the air-water interface. The advective transport is simulated by treating
water temperature as*a thermal concentration. This allows use of the standard
method within the model for computing advective transport of dissolved
constituents.
The mechanisms affecting heat transfer across the air-water interface are
absorption of solar (shortwave) radiation, absorption/emission of infrared
(longwave) radiation, conduction-convection, and evaporation. Each mechanism is
evaluated separately, and the net transfer is the sum of the individual effects.
Solar radiation absorption is computed by the following equation:
. -, - *
QSR - 0.97 * CFSAEX * SOLRAD * 10. (5.1)
where:
QSR - radiation absorbed (kcal/mz/hr)
0.97 - fraction of incident radiation absorbed
CFSAEX - fraction .of: gage radiation that is incident to the surface
SOLRAD - solar radiation (langleys/hr)
10. - conversion from langleys to kcal/m2
The net infrared radiation transfer, considering both atmospheric radiation and
back radiation from the water surface, is dependent on temperature and cloud
cover. It is computed using the following equation: .
where:
QB - SIGMA * (TWKELV* - KATRAD * 10ซ * CLDFAC * TAKELV*) (5.2)
'
QB - net transport of infrared radiation (kcal/m2/hr)
SIGMA - Stefan-Boltzmann constant multiplied by 0.97 (kcal/m2/hr/K*)
TWKELV - water temperature ('K) . '
KATRAD - atmospheric longwave radiation coefficient
CLDFAC - cloud factor - 1.6 + (0.0017 * CLOUD2)
CLOUD - cloud cover (tenths)
TAKELV - air temperature corrected for elevation (*K)
Air temperature is corrected for elevation by application of a lapse rate to the
gage temperature. If it is raining, the lapse rate is 0.00194 C/ft; otherwise,
the lapse rate is selected from a table of 24 hourly values ranging from 0.0019
to 0.0028 C/ft. .
155
-------
Conductive-convective heat transfers are caused by the temperature difference
between the air and the water; the net transfer is computed as follows:
QH - CFPRES * KCOND * 1
-------
Since sediment migration characteristics and adsorptive properties of sediment
vary with particle size, SEDTRN considers three size fractions (sand, silt, and
clay), each with its own set of parameters. The silt and clay fractions are
categorized as cohesive sediments, and are simulated differently from the sand
or noncohesive fraction. The module maintains storages of each size fraction in
the bed and in suspension, and computes scour and deposition fluxes between these
storages. It also computes a bed depth based on sediment porosities and width
and length of the bed. SEDTRN assumes a single bed layer, so that all sediment
is available for scour; burial and armoring are not considered.
Scour and deposition of cohesive sediments depends on the bed shear stress
computed in'the HYDR module, and is based on equations developed by Krone (1962)
and Partheniades (1962). Whenever bed shear stress in the reach is less than the
user-defined critical shear stress for deposition, sediment settles to the bed.
Conversely, whenever the shear is greater than the critical shear stress for
scour, erosion of sediment from the bed occurs. The following two equations are
used for each size fraction to compute the deposition and scour fluxes,
respectively.
DEPCNC - CONG * (1.0 - EXP(-W/AVDEPM) * (1.0 - TAU/TAUCD)) (5.5)
where:
DEPCNC - concentration of suspended sediment lost to deposition (mg/L/hr)
CONC - concentration of suspended sediment at start of Interval (mg/L)
V - settling velocity of size fraction (m/hr)
AVDEPM - average depth of reach (m) -
TAU - bed shear stress (kg/m2)
TAUCD - critical shear stress for deposition (kg/m2)
SCRCNC - M/AVDEPM * 1000 * (TAU/TAUCS - 1.0) (5.6)
where: . .
SCRCNC - concentration of suspended sediment added- to .
suspension by scour (mg/L/hr)
M -credibility coefficient (kg/mVhr)
1000 - conversion from kg/m3 to mg/L
TAUCS -critical shear stress for scour (kg/m2)
The principal calibration parameters for cohesive sediments are the settling
velocity (W), credibility coefficient (M), and the critical shear stresses (TAUCD
and TAUCS). The critical stresses affect the timing of deposition and scour,
while V and M are useful in adjusting the magnitudes. In the Watershed Model,
initial values of TAUCD and .TAUCS were determined from time series plots of the
simulated bed shear stresses in each reach. These adjustments were made so that
occurrence of scour and deposition was restricted to extremely high and low
stress periods, respectively. As an example, based on the plot of shear stress
shown in Figure 5.5, the TAUCS and TAUCD values for REACH 290 in the James River
were set to the following values:
159
-------
Critical Shear Stresses for REACH 290
(lbs/ft2)
DEPOSITION SCOUR
SILT 0.030 0.137
CLAY 0.029 0.135
Values of M and KSER (the soil erosion parameter for pervious areas) were Chen
calibrated by comparing simulated and observed suspended sediment concentrations
during, scour events and periods of high loading.
Sand transport in SEDTRN is simulated by one of three optional methods. The
Toffaleti and Colby methods are detailed formulations primarily applicable to
wide rivers, while the user-specified power function (of velocity) is simpler and
more generally applicable. The power function method is used in the Watershed
Model because of the variety of channel types encountered in the basin, and
because of the greater degree of control over the sand simulation that this
option provides.,
Sand scour and deposition in SEDTRN depends on the amount of material the flow
is. capable of carrying. If the amount of sand being transported is less than the
flow can carry for the hydrodyriamic conditions of the reach, sand will be scoured.
from the bed. Conversely, deposition occurs if the sand in suspension exceeds
the flow's capacity. The potential sandload or sand carrying capacity for the
power function method is computed directly as: '
PSAND - KSAND * AVVELEEXI>sm> (5.7)
where: '
PSAND - sand carrying capacity of the flow (mg/L)
KSAND - coefficient in the sandload equation
EXPSND exponent in the sandload equation
AWELE - average velocity of the flow (ft/s).
5.2.4 Water Quality Processes
Water quality simulation in the Watershed Model is performed by the RQUAL module
which consists of four interacting submodules. Three of these suboodules are
used in the Watershed Model: 1) OXRX, which simulates oxygen and BOD dynamics;
2) NUTRX, which simulates inorganic nitrogen and phosphorus reactions; and 3)
PLANK, which handles phytoplankton and other organic materials. These submodules
are executed sequentially in the order listed, and each submodule requires the
execution of those above it. For example execution of NUTRX requires that OXRX
be executed, and PLANK requires execution of both OXRX and NUTRX. OXRX,execution
does not require any other submodule.
In this section, overviews of the nitrogen and phosphorus cycles are described
first, and then the individual processes simulated in each submodule are
described under the corresponding section below. Table 5.1 provides a summary
of all of the water quality constituents and processes simulated in the Watershed
Model along with the submodules that handle them.
161
-------
The complete schematic of instream nitrogen cycling in the Watershed Model is
shown in Figure 5.6. The distinct forms of nitrogen are nitrate (N03), dissolved
ammonia (NK3), particulate ammonia (ads-NH3), degradable organic N (BOD), living
organic N (phytoplankton), and dead refractory organic N (ORN). The principal
changes made as part of Phase II are the addition of the particulate ammonia and
a more flexible algorithm for denitrification of N03.
The parallel schematic for phosphorus is shown in Figure 5.7. As for nitrogen,
improvements made to the Phase II model are the addition of the particulate
orthophosphate constituent and its associated processes sorption/desorption and
settling/resuspension.
5.2.4.1 Dissolved oxygen and BOD dynamics
The primary reaction's of dissolved oxygen and BOD simulated in the Watershed
Model are performed by OXRX, while processes simulated in the NUTRX and PLANK
submodules also affect oxygen and BOD concentrations. Figures 5.8 and 5.9 are
the flow diagrams for oxygen and BOD; the processes simulated in the Watershed
Model are delineated by dashed lines.
Reaeration is computed as the product of a coefficient dependent on hydraulic
conditions, and a driving force which is the difference between the actual
dissolved oxygen concentration and the saturated value for the current
conditions. The general expression is:
. DELDOX - KOREA * (SATDO - DOXS) (5.8)
i ,
where:
DELDOX - change in oxygen concentration over the interval (mg/L) .
KOREA - reaeration coefficient
SATDO - oxygen saturation level at the current water temperature (mg/L)
DOXS - oxygen concentration at start of interval (mg/L)
The saturation value is based on temperature as follows:
SATDO - 14.652-+ TW*(-.4102 + TW*(.00799 - .7777E-4*TW)) * CFPRES (5.9)
where:
TW - water temperature (*C)
CFPRES - ratio of atmospheric pressure to pressure at sea level
Estimation of the reaeration coefficient for free flowing river channels uses a
modified version of Churchill's empirical relationship of depth and velocity
which has the following form: .
KOREA - 0.726 * AVVELE0-'*9 * 1.024(TW ' "'VAVDEPE1-673 (5.10)
where:
AWELE - average velocity (ft/s)
AVDEPE - Average depth (ft)
163
-------
WATER COLUMN
uptake
adsorption/
derorptlon
PO4-P
(adsorbed)
LIVING ORGANIC
P
(phytoplankton)
PO4-P
(dU(Olvod)
death/
reซp.
deposition/
oour
3-
BOO
dtoay
death/
respiration
OEORAOABLE
ORQANIC P
(BOO)
tiling
REFRACTORY
ORQANIC P
(ORP)
ttllng
tiling
1 PO4-P i
(adsorbed) |
I Infinite
1 ouroe/ซlnk i
Infinite
Ink
Infinite
Ink
I I
I I
1
Infinite
Ink
I _ __ J
: I
SEDIMENT
Figure 5.7 Insteam phosphorus dynamics .in the Watershed Model.
-------
IBOD
Inflow
to
RCHRES
O\
phy to-
pi an K ton
death
storage of BOD
settllnd
benthlc
release
outflow
'through
exit N
BOO
decay
Figure 5.9 Flow diagram of BOD in the OXRX section of the RCHRES module.
-------
TW - water temperature (*C)
TAM - total dissolved ammonia-N. concentration (mg/L)
This process is dependent upon a suitable supply of dissolved oxygen; i.e.,
nitrification is set to zero if the DO concentration is below 2 mg/L. In HSPF,
the amount of oxygen used during nitrification is 4.33 mg oxygen per mg NH3-N
oxidized completely to N03-N.
Denitrification is the reduction of nitrate by facultative anaerobic bacteria
such as Pseudomonas, Micrococcus, and Bacillus. These bacteria can use N03 for
respiration in the same manner that oxygen is used under aerobic conditions.
Facultative organisms use oxygen until the environment becomes nearly or totally
anaerobic, and then switch over to N03 as their oxygen source. The end product
of denitrification is assumed to be nitrogen gas.
Oenitrification does not occur unless the dissolved oxygen concentration is below
a user-specified threshold value (DENOXT) . If that situation occurs,
denitrification is assumed to be a first-order process based on the N03
concentration. The amount of denitrification is calculated by the following
equation:
DENN03 - KN0320 * (TCDEN**(TW-20)) * N03 (5.13)
where:
DENN03 - amount of N03 denitrified (mg N/L per hour)
KN0320 - N03 denitrification rate coefficient at 20*C (/hr)
TCDEN - temperature correction coefficient for denitrification
N03 - nitrate concentration (mg N/L)
NUTRX simulates the exchange of ammonium between the dissolved state and
adsorption on suspended sediment, i.e., adsorption/desorption is not simulated
In the bed sediment. The sorbents considered are sand, silt, and. clay, which are
simulated in section SEDTRN. The adsorption/desorption process for each sediment
fraction is represented with an equilibrium linear isotherm (K^) which is
described as follows :
ADSNH3 - DISNH3 * KDNH3 (5.14)
where : - . .
ADSNH3 - equilibrium concentration of adsorbed ammonia on sediment
fraction J (mg/Tcg)
DISNH3 - equilibrium concentration of dissolved ammonia (mg/L)
KDNH3 - adsorption parameter (or Kj) (L/kg)
The advective processes for ammonia adsorbed to a sediment size fraction are:
1. Inflow to the RCHRES of ammonia attached to suspended sediment.
4
2. Migration of ammonia from suspension to the bed as a result of
deposition of the sediment to which the ammonia is adsorbed.
,169
-------
TALGRH, no growth occurs. Between TALGRL and TALGRM, the correction factor
increases linearly from 0 to 1, and between TALGRM and TALGRH, the correction
factor is 1. In the Watershed Model, the high temperature threshold is 95*F, and
the optimum temperature (TALGRM) is 86 *F. The low temperature threshold was set
to 26.7*F for most of the basin, except for James and-other Virginia tributaries,
where lower values (> -10*F) are heeded to produce sufficient algal growth during
winter.
Algae depend on uptake of orthophosphorus to provide the phosphorus necessary for
growth. In situations where phosphorus is the limiting factor, the growth rate
depends on .both phosphate and nitrate as follows:
MALGRT * P04 * N03
CROP - : (5.15)
(P04 + CMMP) * (NO, -I- CMMNP)
where: .
CROP - unit growth rate based on P-limitation (/hr)
MALGRT - temperature-corrected maximum growth rate (/hr)
CMMP - phosphate Michaelis-Menten constant for P-limited growth (mg/L)
CMMNP - nitrate Michaelis-Menten constant for P-limited growth (mg/L)
Nitrogen, which is also essential to algal growth, is obtained from the total
pool of dissolved NH3 and N03 in the reach. The ratio of N03 uptake to total
Inorganic N (N03 + NH3) uptake is governed by a preference factor (ALNPR) which
ranges from 0.25 to 0.35 in the Watershed Model. The nitrogen-limited growth
rate is calculated by the following equation:
MALGRT * (NH3 + N03)
GRON - (5.16)
' CMMN + (NH3 > NOj)
where:
GRON - unit growth rate based on N-limitation (/hr)
CMMN - Michaelis-Menten constant for N-limited growth (mg/L)
During each time .step, the model computes the amount of radiation available to
support algal growth by applying light absorption rates to the incident light
available at the surface, and assuming that all phytoplankton are at the mid-
depth of the reach. The overall light extinction is the sum of the base
extinction coefficient, a chlorophyll A factor, and a sediment factor. Finally,
the light-limited growth rate is computed as: .
MALGRT * LIGHT
GROL - (5.17)
CMMLT + LIGHT . -
where:
GROL - unit growth rate based on light-limitation (/hr)
CMMLT. - Michaelis-Menten constant for light-limited growth (Ly/min)
LIGHT - light intensity available to algae (Ly/min)
171
-------
parameter used in the Watershed Model "to adjust algal losses from the system.
.5.2.4.4 Organic nitrogen, organic phosphorus, and total organic carbon
Dead refractory organic matter is simulated as three state .variables in the
model. ORN, ORP, and ORC are the nitrogen, phosphorus, and carbon species,
respectively. These constituents undergo advection, they are generated as a
result of phytoplankton death (see above), and they are lost from the system via
settling. They are.constituents in the respective total organic quantities
(TORN, TOR?, TORC) which are also made up of contributions from BOD and
phytoplankton. It is these simulated total organic state variables that are
compared directly with the "observed" organics concentrations (organic N, organic
P, and TOG). The following expressions, based on the Watershed Model
stoichiometry, define the total organic state variables:
*
TORN - ORN + 0.0863 * PHYTO + 0.0529 * BOD (5.19)
TORP . - ORP + 0.0123 * PHYTO + 0.00756 * BOD (5.20)
TORC - ORC + 0.49 * PHYTO + 0.301 * BOD . (5.21)
where:
PHYTO - phytoplankton concentration (mg/L as biomass)
BOD - biochemical oxygen demand (mg/L as oxygen)
In the Phase II Model, total organic carbon (TOC) replaces-BOD as the primary
indicator of carbon species. This change was implemented because of the greater
availability of observed TOC data for direct comparison with the simulation.
However, in order to simulate TOC, it was necessary to characterize the loadings
of carbon since the Phase I model did not account for the total carbon loadings.
Previously, loading of nonliving carbon species occurred entirely through the BOD
loading. In the Phase II model, refractory organic carbon (ORC) loadings from
point and nonpoint sources are explicitly included by assuming that ORC loads are
proportional to the existing BOD loads. Therefore, the BOD loads are multiplied
by a factor and input to the system as refractory organic carbon.
5.2.4.5 Review and adjustment of instream water quality parameters .
The chemical and biological parameters were.reviewed and adjusted in Phase II.
Some of the parameter values were atypical of the ranges used in surface water
modeling. This may have been due to the magnification or reduction of certain
parameters in order to compensate for phenomenon that are not simulated by the
model. For example, in the initial, versions of the model, temperature endpoints
of the smoothing functions were set at atypical levels to permit phytoplankton
growth in all four seasons. In addition, in previous versions of the model
-maximal algal growth rate may have been set artificially high to account for both
free-floating and benthic algae as well, as adsorption-sedimentation mechanisms
for reducing inorganic P and N during instream transport.
A series of steps, following the hierarchy of model sections were executed during
the process of adjusting the water quality parameters. For instance, hydraulics
simulation was finalized before temperature was adjusted and the dissolved oxygen
simulation was completed before proceeding to the adjustment of inorganic
nutrients.
173
-------
TABLE 5.2 INSTREAM WATER QUALITY PARAMETERS
Parameter Description
Value
Unit Ref.
TCBOD Temperature Coefficient
for BOD decay
TCGINV Temperature Coefficient
for gas invasion
TCNIT Temperature Coefficient
for nitrification
TALGRH Temperature above which
algal growth ceases
TALGRL Temperature below which
algal growth ceases
TALGRM Temperature below which
algal growth is retarded
CMMLT Half-saturation constant
for light
CMMN Half-saturation constant
for nitrate
CMHNP N03 Concentration for .
. Phosphorus limiting growth
CMMP P04 Concentration for
Phosphorus limiting growth
KBOD20 BOD Decay rate at 20 *C
KODSET BOD settling rate
KNH320 Nitrification rate at *C
MALGR Maximal algal growth rate
EXTB Base light extinction
coefficient
ALR20 Algae respiration rate at
20 *C
ALDH , High algal.death rate
PHYSET Phytoplankton settling rate
REFSET Dead refractory settling
rate , . '
RATCLP Chlorophyll-a to P ratio
in biomass
ALNPR Fraction of N03 as nitrogen
source for algal growth
1.047 N/A 1,2
1.024 N/A 1,2
1.045 N/A 1,2
95.0 *F 2
26.7 *F 2
86.0 *F 2
0.040 (ly/min) 1,2
0.025 (mg/L) 1,2
0.0001** (mg/L) 1,2
0.005 (mg/L) 1,2
0.004 (1/hr) 1,2
0.021*** (ft/hr) 1,2
0.004 (1/hr) 1,2
0.142 (1/hr)
0.575 (I/ft)
1,2
*
0.005 (1/Hr) 1,2
0.003 (1/hr) 1,2
0.027*** (ft/hr) 1
0.021*** (ft/hr) 1
0.68 N/A 1,2
0.35 N/A 1
1. Reference 1 - Bowie et al., 1985
2. Reference 2 Potomac Model, Thomann and Fitzpatrick, 1982
* calculated from tidal fresh monitoring data.
** Set at a value as to make the parameter inoperative.
*** Adjusted by calibration for each reach.
175
-------
TABLE 5.3 CHANNEL REACH INFORMATION
LENGTH
SUBBASIN REACH
REPRESENTED
SUSQUEHANNA
EAST BRANCH
20
30
40
WEST BRANCH
60
70
JUNIATA
100
LOWER
110
CONOWINGO
140 ,
PATDXENT
340
POTOMAC
UPPER
170
175
SHENANDOAH
200
LOWER
210
220
RAPPAHANNOCK
YORK
MATTAPONI
240
PAMUNKEY
260
JAMES
270
280
290
APPOMATTOX
310
10
184.4
,94.9
68.4
50
88.2
42.1
90
111.9
80
24.2
120
14.0
330
21.0
160
126.6
43.3
190
57.8 .
180
64.5
41.0
235
35.7
250
44.5
265
114.0
107.3
39.0
300
13.8
TRIBUTARY
(MI) AREA (MI2!
108.0
4991
2440
1492
91.0
4262
1330
27.0
2418
50.6
1922
19.6
295
12.0
215
,
92.3
1431
1257
139.0
1406
78.6
971
955
230
27.6
338
16.3
737
12.0
2917
2989
512
108 .4
158
2647
S
S .
S
1428
S
S
937
S
2286
S
743
L
133
S
1349
S
S
1623
S
2507
S
S
73.0
257
S
334
S
348
S
S
S
1194
L
) TYPE* PRIMARY TRIBUTARY
S EAST BRANCH
EAST BRANCH
EAST BRANCH
EAST BRANCH
S WEST BRANCH
WEST BRANCH
WEST BRANCH
L RAYSTOWN RESERVOIR
, JUNIATA RIVER
S SUSQUEHANNA RIVER
SUSQUEHANNA RIVER
L CONOWINGO RESERVOIR
CONOWINGO RESERVOIR
L PATUXENT RESERVOIRS
PATUXENT RIVER
S N. BRANCH POTOMAC
S. BRANCH POTOMAC
POTOMAC RIVER
S SHENANDOAH RIVER
SHENANDOAH RIVER
S POTOMAC RIVER
' MONOCACY RIVER
POTOMAC RIVER
1604 S RAPPAHANNOCK RIVER
S MATTAPONI RIVER
MATTAPONI RIVER
L LAKE ANNA
PAMUNKEY RIVER
L JAMES RIVER
JAMES RIVER
JAMES RIVER
JAMES RIVER
S APPOMATTOX RIVER
LAKE CHESDIN
* S - River reach
L - Lake/reservoir reach
177
-------
SECTION 6.0
HYDROLOGY CALIBRATION RESULTS
this section discusses the results of the hydrology calibration, including
discussion of tasks performed in both Phase I and Phase II. Much of the text
from the Phase I report is retained in order to provide a complete description
of the hydrology calibration effort; however, only the results from Phase II are
presented here since they supersede the Phase I results for the 1984-87
simulation period. Readers interested in the 1974-78 hydrology results are
referred to the Phase I report (Donigian et al, 1990).
6.1 OVERVIEW AND STEPS IN THE HYDROLOGIC CALIBRATION PROCESS
the purpose and justification for the Phase I hydrology re-calibration was to
include snow simulation., evaluate any changes resulting from the revised land use
data and addition of the animal waste segments, and incorporate any changes
necessitated by allowing different hydrologic parameters for each land use
category. In Phase I, the 1974-78 period was chosen for the primary re-
calibration effort because the existing data base covered this period and the
model had been previously verified for the same period. In effect, the goal of
the effort was to evaluate the cumulative impact of all the changes to the
hydrologic parameters, and then perform the parameter adjustments needed to 're-
calibrate' the model. The model was then run on the 1984-85 period and the
results were analyzed, evaluated, and provided the basis for the recommended
.focus of the hydrology calibration tasks in Phase II.
As discussed in Section 1.3, the Phase I recommendations included extending the
simulation and calibration through 1988 in order to provide a more realistic
evaluation of the Watershed Model' s ability to represent the hydrology during the
critical 1984-88 period. This period was used as a basis for both the Bay model
and the 40Z Chesapeake Bay Agreement. Additional Phase I recommendations
included evaluation of the snow simulation for 1984-88, investigation of the
Conowingo Reservoir operation and simulation, and further investigation of the
simulation of the November 1985 storm. Due to limitations on data availability
for 1988 at the time the meteorologic was needed, the simulction could only be
extended through 1987.
The steps performed in the combined Phase I/II efforts were as follows:
a. Identify watershed areas needing snow simulation.
b. Perform snow calibration for 1974-78 to observed snow depth data.
c. Evaluate hydrologic parameters for each land use category.
178 '
-------
TABLE 6.1 ELEVATION, TEMPERATURE AND SNOW DEPTHS FOR WATERSHED MODEL LAND SEGMENTS
Sub
RCH
10(1)
20(1)
30(1)
40(2)
50(1)
60(1)
70(2)
^80(2)
90(2)
100(2)
110(2)
120(2)
140(3)
160(1)
170(2)
175(2)
180(2)
190C3)
200(3)
210(2)
220(3)
230(4)
250(4)
260(4)
265(4)
270(4)
280(4)
290(4)
300(4)
basin
Elev
(feet)
3300-650
3300-650
2000-650
1640-650
3300-650
3300-650
3300-500
1640-330
3300-650
3300-300
1640-330
1000-500
1000-
2400-650
3300-650
1640-650
2000-400
3600-650
1640-400
2000-300
2000-
1000-
1000-
.1000-
3300-1640
3300-650
1000-
500 -
500 -
Average*
Elev
(feet)
1860
1512
1368
930
1980
1260
524
6UO
, 4
- 772
469
,510
40
1693
' 1868
1288
547
1880
460
710
910
500
: 220
320
2579
1060
916
164
.164
C
Normal S
Temp
27.3
25.9
27.2
27.3
26.4
28.9
30.7
32.3
30.9
31.5
33.3
32.2
34.4
. 31.4
34.7
33.3
33.4
35.4
34.5
33.3
35.8
34.5
37.9
38.7
33.9
40.0
38.5
39.9
39.4
ECEMBER
now on
Max
10.3
9.7
8.5
4.4
6.0
6.9
3.3
1.9
2.2
1.7
3.2
1.5
4.7
3.2
.9
.6
.2
.0
.8
.8
.2
.3
.6
2.3
T
T
T
T
Ground*
Min
T
T
T
T
0.0
0.0
0.0
0.0
T
0.0
0.0
0.0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
ป
Normal S
Temp
22.5
20.8
22.5
23:3
22.1
24.4
26.2
28.3
26.9
27.5
29.3
28.5
30.4
27.8
31.3
29.5
26.6
32.1
30.8
29.4
31.4
31.2
34.7
35.8
30.8
35.3
35.4
36.6
37.0
IANUARY
inow on
Max
13.5
14.8
12.8
5.5
11,9
10.5
8.1
6.7
6.2
5.7
8.1
4.0
11.3
.5
.1
.9
.3
.1
.3
.3
4.9
4.0
5.5
6.2
4.2
3.8
3.2
3.2
Ground*
Min
3.3
3.1
3.1
T
1.3
1.1
T
T
T
T
0.0
0.0
1.0
T
T
0.0
0.0
0.0
T
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
1
Normal !
Temp
23.5
22.0
, 23.8
25.0
23.2
25.7
28.2
30.3
28.5
29.2
31.2
30.5
32.2
29.7
33.3
31.2
31.8
34.1
32.8
31.6
33.6
33.0
36i4
38.1.
32.8
37.8
37.4
38.9
39.2
'EBRUARY
Inow on
Max
13.9
15.7
14.9
6.4
15.2
11.7
7.4
7.7
7.0
6.1
9.6
6.9
12.9
9.0
8.6
7.1
6.5
4.9
9.2
5.9
4.7
2.9
4.0
5.4
4.3
3.9
2.7
2.7 .
Ground*
Min
4.8
4.3
4.1
T
3.1
2.6
1.1
1.6
T
T
T
0.0
2.2
T
1.5
,T
T
T
T
T
Tr
0.0
0.0
T
0.0
0.0
o.o
0.0
* - Average of show measuring stations elevation (ft) in that Subbasin (see attached Notes).
* - Average maximum snow on ground and average Bininum snow on ground (see attached Notes).
T - Trace of snow. ' / .
Segment Categories:
(D-Temperature below 30 (F) and max snow on ground above 10 (in) for two or nore winter months.
(2)-Temperature below 32 (F) and max snow on ground below 10 (in) and above 6 (in) for two or
tore winter months.
(3)-Temperature below.32 (F) and max snow on ground below 10 (in) and above 6 (in) for one month.
(4)-Temperature above 32 (F) and/or max snow on ground below 6 (in) for two or more months.
NOTE* For subbasins 40, 110 and 200 the snow gages are located at the lower end of the elevation
range. So, an exception is made to include these subbasins under the next highest category.
480
-------
extreme, segments under Category //4 do not require snow simulation due to above
freezing temperatures and minimal observed snowfall.
With regard to the 'marginal' segments under Categories #2 and #3, the actual
snow depth records were reviewed (as discussed under step c. above) to assess
both the maximum depths achieved and the typical duration of snow cover during
the winter months. Based on this assessment, segments under Category #3 were
eliminated from contention because maximum depths rarely exceeded 10 to 12 inches
and usually lasted only a few days. Similarly, under Category |2, segments 110,
120, 140, 180, and 210 were eliminated due to depth and duration of snow cover;
depths rarely exceeded -12 inches and cover occasionally lasted up to two weeks
at the most. ' -
However, segments 40,^70, 80, 90, 100, 170. and 175, under Category 02 frequently
showed snow cover lasting a month or more, and'depths often exceeded 12 to J.8
inches. Consequently, we felt that these segments would have significant snow
depths, and subsequent storage and carry-over of moisture from one month to.
another; in our opinion, this justified the need for snow simulation.
* ' i
Based on these analyses, the dividing line for snow simulation (as shown in
Figure 6.1) occurs about midway through the Lower Susquehanna, at the northern
boundary of the Lower Potomac, and at the southern boundary of the Upper Potomac.
Since each segment is simulated with one set of input and parameters, the
dividing line follows segment boundaries. To add or delete areas for snow
simulation, an entire segment would need to be added or deleted. Additional
areas can be added for snow simulation in future efforts to those currently
simulated, if justified. However, the overall hydrologic calibration results
(discussed below) appear to confirm the snow segment selection.
6.2.2 Snow Calibration and Simulation'Results
In both Phase I and Phase II, snow calibration was performed on Segment 20 in the
East Branch of the Susquehanna, Segment 50 in the West Branch of the Susquehanna,
and Segment 160 in the Upper Potomac. these segments were chosen to specifically
span the geographic extent of the snow simulation region and include segments
with some snow depth measurements; see Figure 6.1 for the segment locations
within the snow region. Since snow depth is highly variable between locations
within a segment, we included as many observations as possible in the.
calibration.
The primary goals of the calibration were to capture the general time period of
snow cover, the magnitude cf the depth of the snow pack, and the general time
frame for disappearance or melt of the pack in the spring. Parameters relating
to the 'catch' of the precipitation gage (SNOVCF), the magnitude of the melt
components (CCFACT), the snow density/moisture content (RDCSN), and the ground-
associated melt (MGMELT) were adjusted in Phase I to obtain the best overall
agreement between the observed and simulated timeseries and frequency of snow
depth values.
The Phase I report included the calibrated frequency curves for snow depth for
Segments 20, 50, and 160, and the daily observed and simulated snow depths for
Segment 20 for each winter period from 1974-78 starting in January 1974. Both the
182
-------
00
ou
27
24
21
g 18
& 15
O
^ป
o
^ 12
9
6
V
i i i i i i i i i i i i
-
QFrMI
rnoii
t/Ur\ 1 L
RIKJPH
DIINon
-
- '
i
"1
j
j
i i i i i i i i i i i
r.NT20
AND
^MTON
?, '
i i i i
ii
|1
/('
i
*!
:
il
fป
; rl
, Y
'
-------
25.0
22.5
20.0
17.5
g 15.0
ฃ
& 12.5
00
- o
i/j
10.0
7.5
5.0
2.5
i i i I r
TI i i i i i i i n i i i iii i n n i r
0
SEGMENT 160
SAVAGE RIVER DAM
. Ill 1 I
n II I I I I
J FMA MJ J ASONDU F M A.M J J A S 0 N DIJ F M A M J J A S0 N Old FMAMJ J A SOND
1984 I 1985 I 1986 I 1987
SNOW DEPTH - SEGMENT 160
Figure 6.4 Simulated and observed snow depth for Model Segment 160, 1984-87.
-------
TABLE 6.2 LAND USE COVER VALUES USED IN THE WATERSHED MODEL
. FOR NON-CROPLAND AREAS
JAN FEE MAR APR MAY JUN JUL AUG SEP OCT NOV DEC
FOREST 0.97 0.97 0.97 0.97 0.97 0.97 0.97 0.97 0.97 0.97 0.97 0.97
PASTURE 0.90 0.90 0.90 0.91 0.91 0.91 0.93 0.93 0.93 0.91 0.90 0.90
N ' - .
URBAN 0.90 0.90 0.90 0.91 0.93 0.93 0.93 0.93 0.93 0.91 0.90 0.90
hydrologic simulation, the monthly rainfall interception values in the Model were
adjusted to be consistent with the seasonality and magnitude of the COVER value
for each land use.
Values for SLSUR and KRER were provided by CBPO based on soils information for
all segments in the watershed. Slope values for SLSUR were derived from
knowledge of the land use categories typically associated with different soil
classifications. Values for KRER were estimated by the 'credibility factor' in
the Universal Soil Loss Equation based on tabulations from the SCS.
LSUR values were estimated with. respect to corresponding slope values for each
land use; areas with high slopes were assigned .low values of LSUR and low slopes
were assigned higher values, based on experience with the model *
NSUR is the Manning's roughness factor for overland flow. The following values
assigned to each land category were initially developed in Phase I and
subsequently revised in Phase II based on information in KnLsel et al (I960),
Engman (1986), and Donigian et al (1983):
, . . i -'.''."'.'
Forest 0.35
Pasture 0.15
Urban Pervious 0.10
Conventional Tillage 0.05-0.06
Conservation Tillage 0.16 - 0.185
Urban Impervious 0.03 ' ,
INFILT values are primarily based on calibration. For each land segment, we
imposed a distribution of values based on experience that had the highest values
for forest areas, the lowest values for cropland, and intermediate values for
urban pervious and pasture. During the calibration effort, this ordering of
values was maintained with individual segment values adjusted as dictated by the
calibration process.
188
-------
TABLE 6.3 WATERSHED MODEL HYDROLOGY AND WATER QUALITY CALIBRATION STATIONS
Model Sta. # Station Name
Status Hydro.
SUSQUEHANNA RIVER BASIN
20 01515000 Sus. R, nr Waverly, NY
.40 01540500 Sus. R. nr Danville, PA
70 01553500 WB Sus. R. @ Lewisburg, PA
100 01567000 Juniata R. @ Newport, PA
80 01570500 Sus. R/@ Harrisburg.PA
110 01576000 Sus. R.
140 01578310 Sus. R.
POTOMAC RIVER BASIN
@ Marietta, PA
@ Conowingo, MD
JAMES RIVER BASIN '
280 02035000 James R. @ Cartersville, VA
290 02037500 James R. nr Richmond, VA
270 02025500 James R. (ง Holcombs Rock, VA
OTHER SITES
310 02041650
220 01646500
180 01638500
210 01643000
230 Q16680QO
240 01674500
260 01673000
340 01594440
330 01592500
Appomattox R. (ง Matoaca, VA
Pot. R. @ Lit. Falls Dam, DC
Pot. R. @ Point of Rocks, MD
Monocacy R. @ Jug Bridge, MD
Rappahannock R. nr Fred., VA
Mattaponi R. nr Beulah., VA
Pamunkey R. nr Hanover, VA
Patuxent R. nr Bowie, MD
Patux.. R. nr Laurel, MD
P
C
C
C
C
P
C
175 01613000 Potomac R. @ Hancock, MD C
220 06146580 Pot. R. @ Chain Bridge, DC C
200 01636500 Shenandoah R. @ Millville, W C
180 01618000 Potomac R. @ Shepardstown, WV P
C
P
C
C/P
P
C
P
C
C
C
C ,
P
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
"Water Cont. 'Co'nt.
Qual. Temp. Sed.
X
X
X
X
.X
X
X
X
X
X
X
XX
X X
X(Buchanan)
X
X
X
X
Proposed Stations for Hydrology and Water Quality Phase I Recalibration
C - calibration station; P - potential calibration stations.
190
-------
TABLE 6.* (continued)
Model USGS
Seg. Station
Number Nunfcer
Station
1974-78
Mean Ota. Mean Sim. ft* R*
Annual Annual Ave. Ave.
flow (In) Flo* (In) Dally Monthly
1984-87
Mean Ota. Mean Sin. R2 R*
Annual Annual Ave. Ave.
FloM (In) Flou (in) Dally Monthly
260 01673000
310 02041650
340 01594440
Pamunkey R. 8
Hanover, VA.
Appomattqx R. 8
Matoaca, VA.
Patux. R. nr.
Bowie, ND.
11.48
11.43
0.3658 0.9392
14.26 14.10 0.7170 0.8323
13.31 13.88 0.7328 0.8153
11.3*4 11.75 0.5928 0.7729
vO
-------
The end result of these differences and inconsistencies between the two periods
was to preclude the use of the 1984-85 as an effective period for verification
of the hydrology simulation in Phase I. Since the evaporation difference was the
major concern, the only calibration efforts performed for the 1984-85 simulation
were constant adjustments to the input evapotranspiration timeseries. That is,
we adjusted a single number, usually in the range of 0.8 to 0.95, multiplying the
input data to obtain improvement in annual flow volumes. These were the only
adjustments during the Phase I calibration for the 1984-85 period.
As reported in the Phase I report, .the results for 1984-85 were not nearly as
close to observed data values as in the earlier 1974-78 period. Although mean
annual flow volumes were calibrated with the evaporation adjustment to maintain.
close agreement, both the daily and monthly R2 values were considerably lower at
most stations. In addition, the frequency curves generally show much greater
deviations, such as the lower peak flows and the higher base flows than observed.
However, the daily flow comparison still showed good tracking between the
simulated and observed values, except for selected peak flow events and during
winter snowmelt periods.
Due to the need for better flow simulation for the 1984-85 period, our
recommendations for additional efforts in Phase II were presented in Section 1.3
and are briefly listed below:
a. Extend the data b*ae and simulation through 1988 or 1989, and
calibrate as needed to improve the flow simulation.
b. Obtain snow depth data for the 1984-89 period to check and evaluate
the snow simulation (AS discussed above).
c. Re-evaluate and investigate the Conovlngo Reservoir simulation for
1984-89.
d. Investigate the storm of November 1985 in the Upper Potomac since it
was not well simulated, in terms of peak flow, in the Phase I effort.
In Phase II, all these tasks were performed, except that the meteorologic data
base could only be extended through 1987 since the 1988 data was hot available
at that time. With the extended data base, the model required only minor
adjustments to the evaporation for most subbasins to produce the final results
shown in Table 6.4. This, in effect, was a pseudoverification of the Phase I
calibration, since the seme model parameters were used except for minor changes.
to the evaporation multiplier. For the smaller subbasins and portions of the
Susquehanna Basin, additional minor adjustments .were needed to the infiltration
(INFILT) and groundwater recession (AGURC) parameters to improve the seasonal
flows and frequency curves.
Figures 6.5 through 6.12 show the flow frequency and daily timeseries curves,
respectively, for the four major basin stations: Susquehanna River at Harrisburg,
Susquehanna River at Conowingo Reservoir, Potomac River at Washington, D.C., and
James River at Cartersville. Complete results for all simulation years and all
194
-------
LOWER SUSQUEHANNA RIVER AT SEG. 80
FLOW (CFS)
RED DASHED: SIM.. BLUE SOLID: DBS.
s
I
M
500000-
400000
300000 -
a
n .
d
0
B
S
200000 -
100000 -
0-
1 I I I r~1I I I I T
T"TT"T
00000000000 00 00 00000000 000 0000000 00000000000000 00
1111111111111111111111111111111111111111111111111
JFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJ
AEAPAUUUECOEAEAPAUUUECOEAE A P A U U U E C 0 E A E A P A U U U E C 0 E A
N B R B Y N L G P T V C N B R R Y N L G P T V C N B R R Y N LGPTVCNBRRYNLGPTVCN
8888888888888888888888888888888888888888888888888
4444444444445555555555556666666666667777777777778
..''' : : ' - 'DATE '-.'.'
FMyure 6.0 .Simulated and obacirved daily flow for llarrisbury, 1>A, Suymunt 80,
-------
S
I
M
500000-
400000 -
300000 -
0
B
S
200000 -
100000-
SUSQUEHANNA RIVER AT CONOWINGO
FLOW (CFS)
RED DASHED: SIM.. BLUE SOLID: OBS.
I i i I i i
I i
i i i i
i i i
ill i i
000000000000000000 000 00000000. 00000000000000000000
i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i
JFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJ
AEAPAUUUECOE A E A PAUUUECOEAEAPAUUUECOEAEAPAUUUECOEA
NBRRYNLGPTYCNBRRYNLGPTVCNBRRYNLGPTVCNBRRYNLGPTVCN
888888888888888888888888888 8888 888888888888888888
4444444 44 44455555 55 5555566666 666 6 6 667 7 7 77 77777778
- ' /' ' " ' ." DATE '' -.
. Pitjin(. 6.0 Simulated and obnorvod dally J.low for Cunowingo. Ruuu-rvul r, Ml), Mi-yini-nt 14O, 1UU1-UV.
-------
LOWER POTOMAC RIVER AT SEG. 220
FLOW (CFS)
RED DASHED: SIM.. BLUE SOLID:
OBS.
300000 -
S
I
M
200000-
ง
0
B
S
100000-
0-
I I III I
T~~I n i T~T i riiii n i i i i i i i i i n i i i i i i i TT~T~T n i i i
ooooooo ooooooo ooooooooo ooooooooooo oooo ooooooooooo
1111111111111111111111111111111111111111111111111
JFMAMJJASONDJFMAMJJA80NDJFMAMJJASONDJFMAMJJASONDJ
AEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEA
NBRRYNLOPTVCNBRRYN L 6 P T VCNBRRYNLGPTVCNBRRYNLGPTVCN
88888868888888888688888888688888 688 88888-888888888
44444444444455555 555 55556666666666667777777777778
Figure 6.10
DATE
Simulated and observed daily flow for Potomac River, Washington, DC, Segment 220,
1QP4-87.
-------
JAMES RIVER AT SEG. 280
FLOW (CFS)
RED DASHED; SIM.. BLUE SOLID: DBS.
300000-
ro
to
S
I
M
a
n
200000-
0
B
S
100000-
0-
iinFIiiir
oooooooooo
i
F
E
B R R
8
4 4
1111111
M A M J J A S
A P A U U U E
N L G P
Tiiiiiniiirnir
000000000000000
111111111111111
DJFMAMJJASONDJF
I
i i r
00000000000000000000000
1
N
C 0
E A E A P A U U U E
T V C N B R R Y N L G P TV
111111111111
M A M J J A S-0 N D J F
APAUUUECOEAE
RRYNLGPTV C N B
1
A
1 1
J A
A P A U U U
R R Y N L G
1 1
0 N
C 0
V
1
J
A
N
888888888888888888868888888888888888888888888888
44444444445555555555556666666666667777777777778
DATE :
Figure 6.-12 Simulated and observed' daily flow for James River, Cartersville, VA, Segment 280, 1984-87.
-------
f. In Phase II, the Conowingo Reservoir hydraulic simulation was
investigated and adjusted to include a minimum summer release rate of
5000 cfs and minor adjustments to the weekly outflow coefficients (see
Section 5.2.1.1). Both the daily timeseries (Figure 6.8) and the
frequency (Figure 6.7) results show .very good agreement between
simulated and observed values, with some deviation- in the low flow
simulation since these flows are controlled entirely by the operating
procedures of the reservoir.
g. Seasonal R2 values calculated by CBPO are shown in Table 6.5 and in
Appendix A. The seasons were defined as follows:
Season 1 > January - February
Season 2 March May
Season 3 June - September .
Season 4 October - December
The results in Table 6.5 show that the winter (Season 1) is
consistently simulated worst than the other three seasons, and that
the summer and fall (Seasons 3 and 4) are usually the best, at least
for the three major basins. For most of the smaller basins, the
spring (Season 2) Is often the best simulated.
Overall, the Phase II hydrology calibration results are a very good
representation of the Chesapeake Bay basin hydrology and provide a sound basis
for nonpoint load generation and delivery of pollutant loads to the Bay. Ongoing
updating of the meteorologic, diversion, and land use information is recommended
to allow for complete verification of the hydrology using the 1988-91 period.
204
-------
SECTION 7.0
WATER QUALITY CALIBRATION RESULTS
This section discusses the results of the water quality calibration, including
discussion of tasks performed in both Phase I and Phase II. Much of the text
from the Phase I report is retained in order to provide a complete description
of the water quality calibration effort; however, only the results from Phase II
are presented here since they supersede the Phase I results for the 1984-87
simulation period. Readers interested in the preliminary 1984-85 results are
referred to the Phase I report (Donigian et al., 1990).
7.1 OVERVIEW AND STEPS IN THE WATER QUALITY CALIBRATION PROCESS
In both Phase I and Phase II, water quality calibration was begun after
completion of the hydrology simulation." However, in Phase I the land use
revisions were also important for the water quality modeling, and in Phase II
additional input updates were required for point sources, atmospheric deposition,
and animal waste contributions. Although the water quality portions of the
Watershed Model had been calibrated in the original NVPDC effort, all these
revisions to the input for 1984-87 required re-assessment and re-calibration of
the water quality parameters. .
x " - " ' -. *
Whereas flow modeling deals with a single constituent - water quantity - and a
single primary source - .precipitation, water quality modeling must consider
numerous constituents, various forms or species, and multiple sources. Thus,
nutrient modeling must consider various forms of nitrogen and phosphorus, their
interactions and transformations, and other parameters or constituents (e.g., DO,
BOD, water temperature, sediment) that affect these interactions. In the current
version of the Watershed Model, the sources considered include point (municipal
and industrial) discharges, watershed land uses, atmospheric deposition, and
animal waste contributions. Nutrient contributions from all these sources are
estimated, hydrologic transport .processes are superimposed, and then water
quality modeling is performed to allow adjustments in parameters and sources as
part of the calibration process.
Water quality calibration is an iterative process; the model predictions are the
integrated result of all the assumptions used in developing the model input and
in representing the modeled processes. Differences in model predictions and
observations require the model user to re-evaluate, these assumptions, in terms
of both the estimated model input and model parameters, and consider the accuracy
and uncertainty in the observations. At the current time, water quality
calibration is more an art, than a science, especially for integrated simulations
of nonpoint, point, and instream water quality.
'206
-------
2. Calibrate sediment delivery from each land use by adjusting the KSER
parameter.
3.' Develop initial values of instrearn parameters (critical shear stresses
for scour and deposition) based on simulated shear stresses in each
channel reach.
4. Adjust KSER and instrearn parameters to calibrate sediment at selected
sites where observed data are available.
The sediment loading calibration was based on "expected sediment loading rates"
(in units of tons/acre/year) for pervious land areas. Expected loading rates for
pasture, cropland, and forest areas were developed for each model segment by
applying the USLE (using county soil data from the 1982 NRI survey), and assuming
a uniform delivery "ratio of 0.15 for all areas. The loading rates for
conservation tillage cropland were assumed to be 60X of the conventional tillage
rates, and values for hay areas were estimated by adding one third of the
difference between pasture and conservation tillage cropland to the pasture
rates. A constant value of 0.16 tons/acre/year* was assumed for all urban areas
in the watershed, based on NURF data. The resulting expected sediment loading
rates for all model segments are compiled in Appendix C.3.
The sediment erosion model, which is described in Section 4.1.1, involves
sediment detachment/attachment from the soil matrix and transport of the detached
sediment. Parameters for the sediment detachment (by rainfall impact) and
attachment were developed (in Phase I) from soil data, while calibration was
largely performed by adjusting the parameter, KSER, which is the coefficient in
the transport equation. The sediment loading for all segments was calibrated so
that the average of the simulated annual sediment loads for 1984-87 approached
the expected sediment loading rates. Appendix C.3 contains the final calibrated
average annual loading rates and the cbrresponding KSER values by model segment
and land use. ซ
The instream sediment simulation (described in Section 5.2.3) considers loading,
scour, deposition, and advective transport of three sediment size fractions,
.sand, silt, and clay. Since the loading model considers a single sediment size
fraction, the total load was apportioned into three size fractions using constant
factors estimated from soil composition data and different delivery ratios for
the three sizes. . . J
Initial values of the ins tr earn scour/deposition parameters for silt and clay were
developed by examination of daily time series plots of simulated velocity and
shear stress in each reach. The sand transport parameters were set so that sand
always settles to the bed, except at high flow velocities. The critical shear
stress for scour (TAUCS) was set so that channel scour would occur only during
extreme flow events, while the critical shear stress for deposition (TAUCD) was
set so that silt and clay deposition would occur-only at very low flaws. The net
effect is that the channels are not .eroded over time, and loadings of silt and
clay generally remain suspended, while sand deposits.
208
-------
2000
1800
1600
1400
J 1200
o
uu
1000
800
600
400
200
0
i i n I i i i i ii ii ii ii i n r
SIMULATED
OBSERVED
i i i i i i i i i i i i i i i i i i i i i i i
FMAMJ J ASONDU FMAMJ J ASOND
1984
r
J FMA.MJ.J ASONDU FMAMJ J.ASOND
1986
r
1987
1985
POTOMAC RIVER AT CHAIN BRIDGE
.SUSPENDED SEDIMENT (MG/L)
Figure 7.1 Suspended sediment concentration - Susquehanna TUver at Conowinqo.
-------
800
720
640
560
5 480
.tE 400
NJ I . I
ft 320
240
160
80
0
1IIIIIIFiIIIII
i i i n r i i i 11 ii- i i ii r
/LJ'I
SIMULATED
OBSERVED
1,
I
J FMA.'MJ J A SOND
1984
J F MA M J J A SO ND
1985
J F MA M J .J A S 0 N D
JFMAMJJASOND
1986 1987
JUNIATA RIVER AT NEWPORT
- SUSPENDED SEDIMENT (MG/L)
Figure 7.3 Suspended sediment concentration - Juniata River at Newport.
-------
UJ
I I I I III II II I II I I I I I III II I I I I I I I I III I II II I I II
SIMULATED
OBSERVED
i i i ii ii
J F MA MJ J A SON D
1984
JFMAMJJASOND
- 1985
J F MA M J J A.S 0 NDIJ FMA M J J A S 0 N D
1986
1987
JUNIATA RIVER AT NEWPORT
SUSPENDED SEDIMENT LOAD (TONS/DAY)
Figure 7.5 Suspended srdiinont load - Junint-.n Rlvo'r nซ Newport.
-------
The calibration involved numerous model runs and iterations at each selected
calibration site in order to maintain parameter consistency within and among the
basins. In the sections below we present and discuss the final Phase. II results
of the comparisons performed in steps d and e above for the concentration and
Fall Line loads. In Phase I, the greatest effort was devoted to the three Fall
Line stations due to the need to provide preliminary Basin loads to the Bay
model. In Phase II, calibration was performed on the Fall Line stations,
numerous subbasins, and the minor tributaries, comprising 13 separate calibration
sites. Due to the large number of plots produced by this effort, the model
results presented below will focus primarily on the Fall Line sites, but the
discussion will include issues relevant to all sites. Complete results for all
water quality calibration sites are provided in Appendix B.
7.3.1 Instream Water Quality Concentrations -
Once nonpoint and point source loading rates were deemed to be reasonable, the
ins tr earn water quality calibration focused on adjustments to selected ins t ream
parameters to improve agreement with observed concentrations. The primary
parameters of concern were the settling rates for phytoplankton and BOO,
nitrification rates, phytoplankton growth rates, algal nutrient uptake
parameters, and partition coefficients and bed concentrations for NH4 and P04.
Final model parameters were discussed in Section 5.2.4.
The figures provided at the end of this section include the water quality
simulation results for four major stations in the basin, including the three Fall
Line Stations on 'the SusqUehanna, Potomac, and James rivers. The. four stations
and corresponding figures are as follows: i
. . . Model .. .'.."'.'. . ' . . .'.'.''''.
Station - Fiures
.80 Susquehanna R. @ Harrisburg, PA 7.6 - 7.17
140 Susquehanna R. @ Conowingo, MD 7.18 - 7.28
220 Potomac R. @ Washington, D.C. 7.29-7.39
280 James R. @ Carters villa, VA 7.40-7.51
In these figures, daily simulated and observed values are provided for the
following water quality constituents for each station in the order shown below:
Water temperature (ฐC)
Dissolved oxygen (mg/1)
Suspended sediment (mg/1)
Nitrate plus nitrite (mg/1)
Ammonia, dissolved (mg/1)
Particulate nitrogen (mg/1)
Total nitrogen (mg/1)
Phosphate, dissolved (mg/1)
Particulate phosphorus (mg/1)
Total phosphorus (mg/1)
Total organic carbon (mg/1)
Chlorophyll a (ug/1)
216
-------
timeseries with discrete, individual observations.
c. The Phase II results are a significant improvement over the Phase I
results due to the water quality parameter re-evaluation, the AGCHEM
nutrient loading simulations, and the additional calibration effort.
Although some problems remain, the results 'show improved simulation of
seasonal variations in nitrate and Chi a, closer representation of
individual nutrient species, and better agreement with most observed
concentrations and total loads.
d. Large deviations are still seen in the individual nutrient forms at
some sites, such as P04 in portions of the Susquehanna and inorganic
nutrients in the minor tributaries. Chi a observations were limited
at many calibration sites, thus restricting the calibration process.
e. the Potomac simulation results are the best of all the major basins
and are generally a good to very good representation of nutrient
concentrations throughout the Bay basin. This could be partially due
to the fact that the Potomac data is more extensive than in any of the
other basins, allowing a better representation of the expected
variability in the observations and a better calibration. The total
nutrient concentrations, and the individual species, are well
simulated throughout the entire period. Concentration peaks do exceed
individual observations, especially for NH3 and P04, but they are
generally within the range of the observed data; ,
f. The Susquehanna simulation shows good agreement between simulated and
observed concentrations, but this was accomplished, only after
intensive calibration effort of the entire basin, with particular
focus on the reservoir simulation. Generally, Total N and component
concentrations at Conowingo are .somewhat undersimulated. PO^ and
Total P concentrations are somewhat over-simulated throughout the
Susquehanna, possibly due to interactions with sediment and metal
oxides, and increased sorption under low pH conditions.
USGS data for pH show extremely low values for two main stem stations
in the East and Vest Branch subbasins, with mean monthly values in the
3.0 to 4.0 range. Detenbeck and Brezonik (1991) show greatly
increased P binding by lake sediments for pH in the range of 4.5 to
6.0, with greatly decreased diffusive P fluxes under these conditions.
Consequently, increased sorption and retention, and possibly
precipitation of PO4 from the overlying water may be a reason for most
of the observed concentrations being close to detections limits.
Clearly, further investigation is warranted and should focus on
instream processes in the Upper and Lower basins for both nitrogen and
phosphorus delivered to the Conowingo Reservoir, and processes within
the reservoir. This should be performed in conjunction with a
thorough loadings analysis (see Section 7.4) to confirm the relative
loadings contributions from all sources. Further calibration of the
reservoir water quality is needed to improve the sediment and Chi a
simulations, which may also assist in improving the phosphorus
218
-------
the TOG simulation. the simulation results for TOC are generally
quite good .for most sites, and are similar in appearance, to the
organic N and organic P simulations.
7.3.2 Fall Line Load Simulation and Regression Comparison
The Phase I water quality calibration included comparison of simulated nutrient
loads at the Fall Line stations' with regression loads developed by the U.S.
Geological Survey (L. Zynjuki R. Summers, and T. Cohn, 1989,, personal
communication) and HydroQual, Inc. (1989), TheU.S.G.S. regression was used for
all stations except the Rappahannock, James, Mattaponi, and Pamunkey rivers,
where the HydroQual regression was used. In Phase II, regression loads for 1984-
86 were provided by the U.S. Army Corps of Engineers so that the same regression
loads would be used^by both the Watershed Model and the Bay model. In both
phases, the calibration effort considered both the available observed
concentrations (as discussed above) and the regression loads whenever parameter
adjustments were proposed. If parameter changes would improve one comparison at
the expense of the other, we placed greater emphasis and priority on the actual
.observed concentrations.
Tables 7.2 through 7.5 show a comparison of annual simulated and regression Fall
Line loads for the three major basins, and the Patuxent. along with their ratios
for each year, 1984 through 1986, and for the entire three-year period.
Review of the comparisons in the tables indicates the following general
conclusions and recommendations: ' '
'
: a. For Total N and Total P, the model and regression annual loads
generally agree within about 10-202; for the Potomac, the maximum
difference is closer to 25Z. Problems with Organic P and, to a lesser
. extent Organic N, in the Potomac and James in 1985 (partly due to the
November 1985 storm) lead to greater differences for that year.
b. In .the Potomac and to some extent the Patuxent, the closeness of the
concentration comparison is also reflected in the Fall Line load
ratios. For both these rivers, extensive data was available for
. developing the regression .and performing the model calibration.
c. Nitrate generally . shows the closest agreement between the two
estimates,- at least for the three major Fall Line stations.
Interestingly, nitrate dominates the Total N and all other nutrient
loads in terms of .total mass, especially for the Susquehanna,
Rappahannock, and the Patuxent, and to a lesser extent, the Potomac.
In the James, Organic N dominates the Total N load and is
significantly greater than the nitrate load.
d. The differences between the two estimates are generally much greater
for the minor tributaries for most nutrient forms (with the exception
of the Patuxent), than for the major Above Fall Line basins. For the
minor tributaries and the James, significantly less data was available.
for developing the regression. Consequently, the .uncertainty of the
regression is greater for these basins.
220
-------
TABLE 7.2 Susqueharma River Fall Line Loads: Comparison of Regression and
Model Estimates - (Millions of Pounds)
Constituent Yc
N03 1
1
I
1
84-M
NH3 1
1
1
1
4-U
Org H (
1
1
4-M
Tot N 1
1
1
1
4-tl
P04 1
1
1
1
4 -ft
Org P
'
**
Tot P
4-
TOC
4-
Regression . Model
ar Estimate Estimate
M 123 128
6 80 74
16 108 W
17 - - . 77
Ntan 104 . 99
* 11.0 14.6
5 7.3 7.4
I* 9.5 9;4
17 7.3
, Haan 9.3 10.5
14 60 52
B 33 27
ft 46 36
ff - . 29
. aปan 46 38
ft 194 194
B 120 108
ft 163 13*
ff 113
kftMn 159 ' U7
ft 1.6 1.0
B 0.9 0.8
ft 1.0 0.8
ff 0.
kซaซป 1.2 0.
7.9 8.
3.1 4.
5.5 5.
4.
MM 5.5 6.
9.5 9.
4.1 S.
6.5 6.
$.
MR 6.7 7.
417 634
209 235
299 336
263
Mn 308 402
Model
Regr.
1
0
0
0
1
1
0
1
0
0
0
0
1
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
Est. /
Est.
.04
.93
.87
.
.95
.33
.01
.99
.
.13
.86
.81
.78
.
.82
.00
.90'
.85
.
.93
.63
.89
.80
.
.74
.08'
.39
.05
.
.13
.00
.24
.02
-
.05
.52
.13
.12
' .
.30
To compare modปl aattwtaa with regression estimates, the following model outputs Mere used:
1. N03 WOt ซ ซS .
2. NH4 Oiuolvad ซH4 * Adsorbed MH4
3. Organic ซ6-K * (0.08631 PHYTO)* (0.05295 BOO)
4. Total Ultrotan 1 * 2 * 3
5. P04 ป Dissolved P04
6. Organic ORC-P + (0.01233 * PHYTO) ป (0.00756 * BOO) * Adsorbed P04
7. Total Photpnorous = 5*6
8. Total Organic Carbon = ORG-C ป (0.49 PHTTO) * (0.3006 BOO)
222
-------
TABLE 7.4 James River Fall Line Loads: Comparison of Regression and
Model Estimates - (Millions of Pounds)
Constituent
N03
NH3
Org N ,
Tot N
P04
.\
Org P
Tot P
. :
TOC
- .
Regression
Year Estimate
84
85
86
87
84-86 Mean
84
85
86
87
84-86 Mean
84
85
86
87
84-86 Mean
84
85
86
87
84-86 Mean
84
85
86
87
84-86 Mean
84
. 85
86
87
84-86 Mean
: 84 - " '
, -. 85 '
86
87
84-86 Mean
84
85
86
'87
84-86 Mean
6.2
. 4.4
2.3
.
4.3
0.73
0.64
0.36
.
0.58
7.9
5.6
3.2
.
5.6
15
12
6
-
11
1.2
1.2
1.0
. .
1.1
2.7
5.2
0.7
2.9
3.9
6.4
1.7
4.0
130
165
44
113
Model
Estimate
7.8
4.9
3.4
7.0
5.4
0.57
1.08
0.60
1.10
0.75
6.9
7.8
1.3
11.8
5.3
15.3
13.8
5.3
19.9
11.5
1.1
1.6
0.4
1.5
1.0
1.5
1.4
0.3
2.0
1.1
2.6
3.0
0.7
3.5
2.1
101
169
17
231
96
Model Est. /. ,
Regr. Est.
1.26
1.11
1.48
.
1.25
0.78
1.69
1.67
-
1.30
0.88
1.40
0.40
. .
0.96
1.02
1.15
0.88
.
1.04
0.92
1.33
0.40
.
0.91
0.56
-0.27
0.43
0.37
0.67
0.47
0.41
; . .
0.53
0.78
1.02
0.38
0.85
To compare model estimates with regression estimates, the following model outputs were used:
1. . N03 * N02 ซ N03
2. NH4 * Dissolved NH4 + Adsorbed NH4
3. Organic N ORC-M + (0.08631 * PHTTO) + (0.05295 * BOO)
4. Total Nitrogen * 1 + 2 * 3
5. P04 ซ Dissolved P04
6. Organic P ซ- ORG-P + (0.01233 * PHYTO) * (0.00756 BOO) * Adsorbed P04
7. Total Phosphorous ซ 5 * 6
8. Total Organic Carbon ซ ORG-C * (0.49 * PHYTO) * (0.3006 BOO)
224
-------
LOWER SUSQUEHANNA RIVER AT SEG, 80
WATER TEMPERATURE (C)
RED DASHED: SIM.. BLUE STARS: DBS.
40
s
I
M
en
a
n 20-
d
o
B :
s *
"
ซ
' "Si H
: 'f ',V
," 'ปซ*
Kl' i* M
? %' '
' -VW ' * '
i ^ i
J J . I
1 h '"
' il |i "
I j 2/
I I
T I I I I T1TTT
i~n
i i ri i i i i i r*ri" i i i i .r'TTT
0000000000000000000000000000000000000000000000000
i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i
JFMAMJJASO NO J F M A M J J A SONDJFMAMJJASONDJFMAMJJASONDJ
AEAPAUUUECOE A E APAUUUECOEAEAPAUUUECOEAEAPAUUUECOEA
NBRRYNLGPTVCNBRRYNLGPTVCNBRRYNLQPtVCNBRRYNLGPTVCN
8888888888888888688888888888888888888888888888886
4444444444445555555555556666666666 66 7777777777778
DATE
Figure 7.6 Simulated -:ind observed water temperature (ฐC)
Susquehanna River at Harrisburq, PA, 1984-87.
-------
LOWER SUSQUEHANNA RIVER AT SEG. 80
. TOTAL SUSPENDED SOLIDS (MG/L)
RED DASHED: SIM. TOTAL. BLUE BARS: DBS. TOTAL
700 -
600-
S
1 500-
M
*
400-
a
r*
.
d 300-
0
S 20ฐ"
'.
100-
0-
. '. ' :'.-.
1 ' ./ ':':
'it1 . ' ' ' .
!'|!
i,il . . .
!'
iซ< 'i
1 * l 'l
ijt ซ !'
i|i n 'f. .
iซi n - i| .
1 1 n !'."' 1
! nl ii i*
: ป5 'i ij
1 n* h n
* iih - i| l'
* ll'i *i M
n MM iiir n
^ i\l \ f\ซ !\fซ i'v'i Lii J.'ซ
\j\ ^ 'f/'v \M' ^__iniซ ^ jj* V^jfvP'^A^V ^
II 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 11 1
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
JFMAMJJASONDJF MA M J. J A S 0 N
AEAPAUUUECOEAEAPAUUUECO
NBRRYNLGPTVCNB R R Y N L Q .P T -V
88888888888888888888888
..'."!':
!!
n :
n
n
ii
n
n
X
II
II .
II
II
II
II
II
II
II
II
J II 1
ff !''
8 L' J i
S ' ! I
ii * ! 5i i
V./*!ปX"%!ป^ซ ^j
i i iii i i.i i i
0000000000
DJFMAMJJAS
EAEAPAUUUE
C N B R R Y N L G P
8888888888
5666666666
'' '". ' "
. . ' . .
.
i i
' ป "
! i! .
'
J j! ซ.; | .
j :! , J \.
i }t',( ^ฃ jl j!
LI* \4* ^'MS-Ai^f vซ*^*i i*vi_*'^
1 1 1 1 1 1 1 1 1 1 1 1 1
0000000000000
ONDJFMAMJJASO
COEAEAPAUUUEC
TVCNBRRYNLGPT
8 8 8 8 8 8 8888 8 88
6667777777777
.,
,
f '', iซ
i,1^*!^'1
1 1 1
00 0
14 4
1 1
N D J
0 E A
V C' N
,888
778
DATE
Figure 7..n Simulated and observed suspended sediment (mg/1), Susquehanna River at llarrisburg, 1'A, 1VM/1-H7.
-------
LOWER SUSQUEHANNA RIVER AT SEG. 80
NH3 CONCENTRATIONS (M6/L)
RED DASHED: SIM.. BLUE STARS: OBS.
EST
1.1-
1.0-
0.9-
0.8-
0.7-
0.6-
.
0.5-
0.4-
0.3-
0.2-
0.1-
0.0-
.
* . ' -
*l
j
1
j
i 11
S, t i i
1 i
\f J > . l^'-.s'Mi'
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
000000000000000
J F M AM JJASONDJFM
AEAPAUUUECOEAEA
- ' *
i- . . ' .
' ' . i -
. ' ' ' j
II 1
n
11 i
n S
n i
ii *
1 !l i I 1 ! I
| n j , i i T
MtJW*m TR ป" ซ* ^m ป" . *
\ \ i i i i 1 \ i \ i i i i i i i i ii i i i i i i i i i i i i
000000000 000 0000 00 0000 000000 00 00
AMJJASONDJFM AMJJASONDJFM A MJJASON
P A U U U EC 0 E A E A P AUUUECOEAEA PA U U U EC 0
_
K
1 l
o o
D J
E A
N B R R Y N L 6 PTVCNBRRYNL.6PTVCNBRRYNLGPT V C NBRRYNL6PTVCN
8888888888888688888888688888888888888888888888888
44444444444455555555555566666666666677777777777 7 8
DATE
Figure 7.10 Simulated and observed ammonia (mg/1), Susquehanna River at Harrisburg, PA, 1984-87.
-------
LOWER SUSQUEHANNA RIVER AT SEG. 80
TOTAL N CONCENTRATIONS (M6/L)
RED DASHED: SIM.. BLUE STARS: OBS.
EST
5:
4-
3-
o
S3 '.
2-
1-
0-
'."
t ' ~
I M '
il* it * ป
f*\ 1 1 p
rU Ml * M
TO! iv "V/H-^i
r VJ ^ 1 \ ' ' *f j^l M/1 ^ il
'i\* M ง !/ป ^J
, "*" \ป/N v *
'' ' .' : *
.---
I 1 1 1 1 , 1 1 1 1 1 1 1 1 1 1 1 1 1
000000000000000000
JFMAMJJASONDJFMA M J
AEAPAUUUECOEAEAPAU
N B R R Y N L 0 P T V C N B. R R Y N
888888888 868 6 8 88 88
" " :".'' '*' ' . : '
'.'"'' '..-,-
".'' .- . . ;'. ' _/-.
n
. - . . l|
. . *ป ll w
M ^ [| ^
f ;V* V ."*f''1 ^ 1* 1 ^W
^ \ *^C L if ' '\l * 1
NH, W ' f ' * ^m*!/^!'
^1 ' . .
' ". 'v " /. : "* '
i i i i ii i i i t i i i i ii ii
000000000000000000
JASONDJFMAMJJASOND
UUECOEAEAPAUUUECOE
L6PTVCNBRRYNL6PTVC
868888888888888886
555555666866666666
*
t -
*
' * '"..
** 8 * ซ i
m/4 K r[ i J )l
II ll \ IJ II .1
Jjft/jfi "iw Jl*l* d
ป" iffflWjvk j
j/ ~ 1 ?ซ P
? -M\j
VI
m
1 1 I 1 i i > i i
000000000
JFMAMJJAS
AEAPAUUUE
N B R R Y N L 6 P
888888886
I,
i^X4/ *V
JK
till
0000
0 N D J
C 0 E A
T V C N
8888
DATE
. . ...
Figure 7.12 Simulated and observed total nitrogen (mg/1), Susquehanna River at llarrisburg, I'A, I9H.4-H/.
-------
LOWER SUSQUEHANNA RIVER AT SEG. 80
PARTICULATE-P CONCENTRATIONS (MG/L)
RED DASHED: SIM.. BLUE STARS: DBS.
EST
0.36-
0.34-
0.32-
0.30-
0.28-
0.26-
0.24-;
0 . 22 -
0.20-
O.lB'-j
0.16-
0.14-
0.12-^
0.10-
0.08-
0.06-
0.04-
0.02
o.oo:
.
1
j
j
!
j
i
/
t
i
^
\ i
0 0
1 1
i't\
yilW
\
*
* m
i i i
OOOl
i i i ,
l'
II
1
3
1
ซ '
1 .
i
'-';-
' . 1 i
1 'j'i 8
MI ;v ' ซ
lU f' \ป i
Vl/ "! /
" W
m x
' '. * ' '
i i i i i i i i
00000 00 0
11111111
1
7
.''''
1 '
1 1
1 I
1 I
,,,Pii,,lj
>-fy \M
A**
X M * * J
M --..JK gf
W ' ' '' . m
1 J l.| 1 1 II 1 II 1 1
00 00 O 000 00000
11 11 i i i i i i 111
[
K
,v
;
.K
K
1
0
1
' ' ' W
m
>
1
1
1
1
H i i
ii* s h
ijwj/v rrsfci
JW^MW*
;-. ซซ V^
M v~y w^
* s J^
)({)){ )K "^
K
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
0000 000 000000 OC
1
1
I
I
1
2
III I
v3 1-
s',-v
v i'V
ป 5K K * *
*
1 1 1 1 1 1
(000000
L 1 1 1 1 1 1
.
JFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJ
AEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEA
NBRRYNL6PTVCNBRRYNLQPTVCNBRRYNL6PTVCNBRRYNL6P TV C N
88 BBSS 8688888888686888 6888888888 8 88888868888688 88
444444444444 55 555555555 56 6 6 666 6666667777777777778
' ' :.' ' 'DATE .;_'.-.
Figure 7.14 Simulated and observed particulate. phosphorus (mg/1), Susquehanna River at llarrisburg, I'A,
-------
M
Ul
I I I I I I I I I I I I I I I I I I I I I I I I I I I I
JFMAMJJASOND
1984
JFMAMJJAS ON 0
1985
JFMAMJJASOND
1986
|J F M A M J J A S 0 N P
1987
HARRISBURG TOC (MG/L)
Figure 7.16 . Simulated and observed total organic carbon, (mg/1), Susquehanna River at Harrisburg, PA,
1984-07.
-------
SUSQUEHANNA RIVER AT CONOWINGO
WATER TEMPERATURE (C)
RED DASHED: SIM,. BLUE SOLID: DBS.
40-
S
I
M
30-
CO
oo
a
n
d
0
B
S
20-
10-
\/
I I J
0 00
111
J F M
A E A
N B R
888
444
I I I I l I I l
00000 0 0 0
11111111
AMJJASON
I I I I. I I I I I II I I .1 II I I I I. I I II I FJ. TIT I II I I f I I
000000000 0*0 000000000000000000000000000
i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i i
DJFMAMJJASONDJFMAMJJASONDJFMAMJJASONDJ
PAUUUECOEAEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEA
RYNLBPtVC NBRRYNL6P TV CNBRRY.MLBPTV. CNBR'RY N L 6 P T V C N
8866888888686888668868888868888888888888888888
4444444445555555555556666666666667777777777778
DATE
Figure 7.18 Simulated and observed water temperature (ฐC), Susquehanna River at Conowlnpo, MD, 1984-87.
-------
SUSQUEHANNA RIVER AT CONOWINGO
N023 CONCENTRATIONS (MG/L)
RED DASHED: SIM.. BLUE STARS: DBS.
EST
B:
A
3
M
^^
+*
o '
2
1
0-
' -
* ' . . . .
/" ill * ft
f\ /! ! l\.
/L/ i. A f. * "IV"
Vvwป!! /wMv
/ if ? } ป* *;ซi/ 1 r U _
n * 1 ป * I ซ 1 ^fc *
'' * - ^IVN % A sr - v\
V'^v/ V V
' ; . . -"* . v
i i i i i i i i i i i i i i i i i i i i
0 00 0 0 00 0 000 0 00 00 000 0
14444444444444444444
1111111111111111111
JFMAMJJASONDJFMAMJJA
AEAPAUUUECO E A E A P A U U U
NBRRYNL6PTVCNBRRYNL6
88866886688888888888
' . - *
'*
t. JL ft ' *
m 'ป *% !l *
I '\ ' * I
ป^N ^ 1 \ . fl*\A w \ * *r*
Xw \ /T.i rj. \ 1 J* V V l-rซ W V # J*v
. f^ }i v'V% vS w /? ^i Aซ v * ^ *
rf i - Jt ^tr.^. r/
* / 1 A /~ * * VM */ *
-\x-v; V!
^ \/
1 1 i i 1 i i 1 i 1.1 i i i.i i i i i i 1 i i i ii ii i
0 0 000000000000000000000000000
S 0 N D J FMAMJJASONDJFMAMJJASONDJ
E C 0 E A E A P A U U U E C 0 E A E A P A U U U E C 0 E A
P'T'V C N B R R Y N L 6PTVCNBRRYNLGPTVCN
88888888888888686888868668868
4 A A A A 4 A A A A A A S S S S S S B ฃ S S S B 6 6 6 6 6 6 6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 7 B
' . ' ' . . ' .- - DATE ';-'
Figure 7.20 Simulated and observed nitrate (nig/I), SuH(|iielinnnu Klvcr at Conowliigo, Ml), I984-H7.
-------
SUSQUEHANNA RIVER AT CONOWINGO
PARTICULATE-N CONCENTRATIONS (MQ/L)
RED DASHED: SIM.. BLUE STARS: OBS.
EST
4H
a-
1-
x
x
ooo oooooooooo ooo oo oo oooooo od ooooooooooooooooooo
11111111111111111111111111 111111111111111111111
JFMAMJJA30NDJFMAMJJASONDJFMAMJJASONDJFMAMJJASON
A ฃ A PA UUUECOEAEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECO
N B R R Y N L 9 P T V C N B R R Y N L Q P T V C N B R R Y NLGPTVCNBRRYNLGPTV
88686886888888668888886886888688688888888888888
44444444444455555555555866666666666677777777777
. . ' . . ''.. DATE . ' '
Figure 7.22 Simulated and observed pnrtlculnte nitrogen (mg/l), Susquohnnnn Rfvcr ;it Conowln^n, Ml), lซ)H/i--H/,
0 0
1 1
D J
E A
C N
8 8
7 B
-------
SUSQUEHANNA RIVER AT CONOWINGO
P04 CONCENTRATIONS (M6/L)
RED DASHED: SIM., BLUE STARS: OBS.
EST
0.16-
0.15 -
0,14-
0.13-
0.12
0.11
0.10-j
0.09
;: 0.08
0.07-
0.06-
0.05-
0.04-
Q.03-
0.02-
0.01
0.00
.*'
* *
II
' 4
\ป *xW, x xx
-------
SUSQUEHANNA RIVER AT CONOWINGO
TOTAL P CONCENTRATIONS (MG/L)
RED DASHED: SIM.. BLUE STARS: DBS.
EST
1.5-
1.4-
1.3-
1.2-
1.1-
i.o-
0.9-
0.8-
0.7-
0.6-
0,5-
0.4-
0.3-
0.2-
0.1-
0.0-
X
-'!' A
ป' '""*
i i i i
0000
J F M A
A E A P
N B R R
8888
/ - . ' . " ' . ' - . *
.'''' '"'.'' ' '
*
'
i t .-.'....." .
II ' ' ' ' ' ' " ' ' ' '
~ *
1 'I 1 T * |
\jV\ j ' * \\ fป ^ i \ '\ ^.A iiซr * /v . M
ii i i i i i. i i i. i i i r i ! 1. i 1.1 i ii i i ii i i i i i i i i ii i ii i i i
00 0 0 0 00 000 0 00 0 000 00 0 0 0 0 0000 0000000000000000
MJJASONDJFMA M J J A SON 0 J FMAMJJASONDJFMAMJJASON
AUUUECOEA E A P A UUUECOEAEAP A U U UECOEAEAPAUUUECO
YNL6PTVCNBRRYNLOPTVCNBRRYNLGPTVCNBRRYNLGPTV
8888888888888 88 B 888 88888888888888888888 88 88
,1.
****
1 T
0 0
14
a
D J
E A
C N
8 8
DATE
Figure 7.26 Simulated and observed total phosphorus (tng/1), Susquehnnna River at Conowingo, MD, 1984-87.
-------
SUSQUEHANNA RIVER AT CONOWINGO
CHL_> CONCENTRATIONS (UQ/L)
RED DASHED: SIM.. BLUE STARS: DBS.
EST
110-
100 -
90 -
80-
'
70 -
60-
to -'
ซ 50- 1. J
K ,n
" A,i%
*:./ 4
20- , |f
10- j If
1
o- **
i i i i
0 0 0 0
J F M A
A E A P
N B R R
8888
. ' .. './V:'. ' -..':.
fv'l
P,
1
1
* 1
* 1 '
l( l\ 1 lซc
JWvji jc'^
/ '- ^rfij/ 1 riW| !
' n/ i! !
w U l| '
f h .
* * * IK ' MW JK '
. JK * *.-.. . ' .
i i i ii lit II i i.i i i i i i i
00000000000 000 000 0 0
MJJASONDJFMAMJ JASON
AUUUECOEAEAPAUUUECO
Y N L Q P T V C N BRRYNLOP TV
8888880888888888888
^
' ' ' '' . .'' '
1
' \ ''..
I
I M
| '. J\'> i
i i 11 ll L I* ^ 1 1 1 ^
jlซ ^ 1*1* VJ- ll*l *l/ Fib I* K ^w
* fl A ^V II r 1 MI 5l * *
A f 'lu'if. ^ j/*- ^ l/ป
V " 'f , J/V .:!.;
.' ' ' ' .- ' ' : ' "
** '.'.',. .
i i i i i i i i i i i i i i i i t i i i i i i i i i
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
DJFMAMJJASONDJFMAMJJASONDJ
E A E A P A U U U EC 0 E A E A P A U U U E C 0 E A
CNBRRYNLQPTVCNBRRYNLQPTVCN
868 88 8888888 888 888 88 8888 88
DATE
Figure 7.28 Simulated and observed chlorophyll A .(ug/1) , Susquehanna River at Conowingo, MD, 1984-87.
-------
LOWER POTOMAC RIVER AT SEG. 220
DISSOLVED OXYGEN CONCENTRATIONS (M6/L)
RED DASHED: SIM.. BLUE STARS: OBS.
N)
Ln
o
M
17-
15-
14
13-
12-
11-
10-
d 8
7
0 6
B e
2
1
0
M
M
I I
I I
I I I I
I I I I
I I
II I
I I
I I I I I III I I IT
00000000000 00 000 000 0 0000000 000 0000000 000 000000000
111 11111111111111 111 111111111 1111111 .111111111 1111
JFMAMJJASONDJFMAMJJASO NO JFMAMJJASONDJFMAMJJASONDJ
AEAPAUUUECOEAEAPAUUUECOEAEAPAUUUE C 0 E AEAPAUUUECOEA
NBRRYNL&PTVCNBRRYNL6PTVCNBRRYNL OPT V C N B R R Y NLGPTVCN
8 88 8 88 8 88 88 8 8 B 88 B 88 8888888888888888 88 888888888888
4444444444445555555555556666666666667777777777778
' ' '-.''. " DATE ' ' ' '
Figure 7.30 Simulated and observed dissolved oxygen (my/1), Potomac River at Wushinlun, U.C., J'J01-U7.
-------
LOWER POTOMAC RIVER AT SEG. 220
NH3 CONCENTRATIONS (MG/L)
RED DASHED: SIM.. BLUE STARS: DBS.
10
EST
0.8-
0.7-
0.6-
0.5-
0.4-
0.3-
0.2-
0.1-
0.0-
\ ซ(!
1*1
J*Uikli '
!Hfw1
MMrt
iiWflH."ir
III 1
0000
J F M A
A E A P
N B R R
L
MM*
i i
0 0
M J
A U
Y N
'.'.. .' ' "- ' ' -
-
'.' '
---'...' ''". /. ''.-' : '
<; .
. "~-
1 ''.'- . ' . ' ' " .
' J
*' I - | |'| '
i ir Vi '/'" -i/i v.j
1 1 I 1 -b" iS'l^ H fc. *l 1 lll'T it Ll i
kLdkAA^ Md^ปlfe*vuiii&
i i i i i i i i i i i i i i i i 1 1 1 i i i i i i i i i i i i i i i i i
000000 0 000 000 00000000000000000000000
JASONDJFMAMJJAS ON D J F M A M J J A SO NO J F M A M J
UUECOEAEAPAUUUECOEAEAPAUUUECOEAEAPAU
LGPTVCNBRRYNLGPTVCNBRRYNLGPTVCNBRRYN
.
II
x 1 H
I '1 * ,
4 n*./i
/ 'USilf
S(Jlf*5KTK THfffX^K
1 1 1 1 1 I 1
0000000
J A S 0 N D J
U U E C 0 E A
L 6 P T V C N
8888
8 88888888888888888 8 BBS 888 88 888888888888888868
4444444444445555555555556666666666667777777777778
- ; '. ';' ' '/' DATE ' . :.
I'igure 7.32 Simulated and ohscrved ammonia (mg/l), I'utomac Kiver at Washington, DC, 1984-87.
-------
LOWER POTOMAC RIVER AT SEG. 220
TOTAL N CONCENTRATIONS (MG/L)
RED DASHED: SIM.. BLUE STARS: DBS.
EST
9-
8-
7-
6-
5-
K
* 4-
3-
2-
1-
0-
' .
-''''ซ ':.
ป . . ''*_.
* - ' '
Sl JK
I ll w M !|l
4fiซ-j^i V* ( * N 'i' w*n
^i^j<^ fcj i ijf^" V \?
. v . y ' . .' ^ *^*jr*
"... . ' . .....
. - : . '
i i i i i .1 i i i i i i 1 i i i i i i i i .1 i i i i i i
0000000000000000000000000000
JFMAMJJASONDJFMAMJJASONDJFMA
AEAPAUUUECOEAEAPAUUUECOEAEAP
NBRRYNL6P TV C N B fl R Y N L 0 P T V C N B R fl
8888888888888888888888888888
'" - : . *
1 .1 i !
I ' n^^ ^' Ivnii i(W n
|l jj >< gw J"l L F "1?
\r n ซซw * || TJ^ Ir^yKl^^ M/ i^J V jiciK
i g__ . j ^y m j(f ^ f^tj/I ^* n
* - '
i ii i ii i i i i i ii i i i lit
00000000000000000 0.0
MJJASONDJFMAMJJASON
AUUUECOEAE A P A U U U E CO
Y N L 6 P T VCNBRRYNLGPTV
8888 88 88888 88888888
m
'*
1 r
0 0
1j
l
0 J
E A
C N'
8 8
7 8
. -. ..' -.DATE ' ' . - . -
Figure 7.3A Simulated and observed total nitrogen (mg/1), Potomac River at Washington, DC, 198/^-87.
-------
LOWER POTOMAC RIVER AT SEG. 220
PARTICULATE-P CONCENTRATIONS (MQ/L)
RED DASHED: SIM.. BLUE STARS: DBS.
EST
3
2-
Ln
K
000 0000 0000000 0000 00 00000 000 000 00000000 0000000000
ill 11111 111111 11 i Ji i i iiiiiiiiiiiiiiiiiiiiiiiii 11 i i
ONDJFMAMJJASONDJFMAMJJASOND
0 E A E A PA UUUECOEAEAPAUUUECOE
JFMAMJJASONOJ
AEAPAUUUECOEA
V C N B R R Y N L 0 P T V C N B R R Y N L Q P T V C N B R R Y N L 6 P T V C N
J F M A M J J A S
AEAPAUUUEC
N B R R Y N L G P T
6868688888886888888688888868688688888888888888888
4444444444445555555555556666666666667777777777778
' ' . ''.'' ':'' .DATE . ' ' . ..-';
Figure 7.36 Simulated and observed particulate phosphorus (mg/1), Potomac River at Washington, DC, I98/I-H7.
-------
M
in
03
60
54
48
42
36
3, 30
24
o
o
12
6
0
\ n i i i i i i i i i i n FI I rn n i i n i i i i i i i i i i r i i i i i i i i i i
SIMULATED
* *' OBSERVED
i i i i i i i i i i i i i i i i i i i i i- i ill ii i i i
J F MAMJ J A SON D
1984
JFMAMJJASONDIJFMAMJJASOND
1985 I 1986
LOWER POTOMAC TOC (MG/L)
JFMA'MJJASOND.
1987
Figure 7.38 Simulated and observed total organic Carbon (mg/1), Potomac River at WnshintjUm, I) c
1984-87.
-------
JAMES RIVER AT SEG. 280
- WATER TEMPERATURE (G)
RED DASHED: SIM.ป BLUE STARS: DBS.
.40 H
N)
a\
o
n
0
B
S
30 '-]
10-
0-
i i i i i i T~T r i i i r~i i i i i i i r~i i i r~r ' i i i i i i i r~i i i rn i i i i r'T'i r r
0 0 0 000 00 0 0 00 0 0 0 0.0 00 00 00 00 000.00000000 0000000 000000
1 1 1 1 1 1 1 1 1 1 l' 1 1 1 1 1 1 1 11 1 1.1 1 11 11 1 11111111111111111 1 1 1
: J F M A M J J A S 0 N D \J F M A M J J A S !0 N D J F M A M J J A S 0 N D J F M A M J J A S 0 N D J
AEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEA
NBRRYNLGPTVC N 8 R R Y N L G P T V C N B R R Y N L G P T V C N B R R Y N L G P T V C N
86686668868886888686886 666 66666668886866888886888
444444 44 4444555555555555 66 6.6 666666667777777777778
. ' . ': .. . ' DATE ; .'..--. . ...
Fiyure 7.40 "Simulated and observed water temperature (ฐC)ป James River at Cartersvilie, VA, 11JM4-H7.
-------
JAMES RIVER AT SEG. 280
TOJAL SUSPENDED SOLIDS (MG/L)
RED DASHED: SIM. TOTAL. BLUE BARS: DBS. TOTAL
1800 -
1700-
16.00-
1500 -
S .1400-
I 1300-
M 1200-
iiop-
a 1000^
n 500
d 800-
700-
0
Q 60ฐ-
B
g 500 -
. 400-
300 -
200 -
100-
0'
-
''.'.' . . . - ;
" * p
. ' ' _
. .. . i
> ' ' i
i
" " ' ' i
i
~ ' - i
.' . ' i
'''" 1
i ^
'. '-.' . i
' .1
'''!" - i
m ซ" ' l
i . 1
.ฃ''' , ..'''" '
<, fejjfc ^ jjl^AJL Jw.. ปw . -A* J
*', ** ' *..* ^
ii i i i i ii i i i ii i.i.i i i i iii i
0 0 000 0 0 0000000000000000
J F M A M J J A S 0 N D J F M A M J J A S 0 N
AEAPAUUUECOEAEAPAUUUECO
NBRRYNLGPTVCNBRRYNLGPTV
888888888888888888888 88
. ' ' - .. .
* - .
, *
".'". ' - ' ' . '
i - - . .
i . .
i
i -. . '
f
- ' ' ' .
' *
i . :' ' . - :.'..*
m . .
i . . . . . . ,
i i
i
i
5 ''. ' ' '
\t . .. j- \.\
-
iii i i i ii i ii i i i iiii i ii i
0000000000000000000000
DJFMAMJJASONDJFMAMJJAS
EAEAPAUUUECOEAEAPAUUUE
CNBRRYNLGPTVCNBRRYNLGP
6888888888888883888388
566666 6666 .6 6 6 777 777777
i
r.
i
i
i
i
Lx..
iiii
0000
0 N D J
C 0 E A
T V C N
8888
7778
Figure 7.42 Simulated and observed suspended sediment (mg/1), James River at Cartersville, VA, 1904-07.
-------
JAMES RIVER AT SEG. 280
' NH3 Concentrations (mg/1)
RED DASHED: SIM.. BLUE .STARS: DBS.
0.7-
0.6-
s
1 0.5-
M
*
0.4-
a
n
d O.Y
0
I P-a-
0.1-
0.0-
. t\
1*1
:
i
I
ซ'
i
I'll f i
*VI 9f ' *ซ 1
' ' ' ' -' - ' .-'..- ' '' ' ..-.- '-'''.''
'. ''''', '."
m ' '
'''.' ' ' " " '
'. . " 'I
". .-.-.... ..in
-.-..: - n .
. . 1 M
..- 1 '1
* II
..M
/i I ' V
M, /! , 111 1
' J 1
i ^'i . j\ v I I in] ''' "MI 1 I '! ' r i * i !
!! if V V1 A /! lii ง ! j}yrlLl/ N1! 1 1 ! jl,1 L / "Vi i n 1 1
'#r* V W4/ 'AA^y "Ir'w * * WAซ-d'V^ ; ^/Jitt i'1 "1
1 1 III II 1 III 1 1 1 1 1 1 1 1 1 1 II 1 1 1 1 1 1 1 1 II 1 1 1 1 II 1 III
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 6 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
J F M A M J J AS ONDJFMAMJJASON D J FMAMJ JASON DJFMAMJJ
AEAPAUUUECOEAEAPAUUUECOEAEAPAUUUECOEAEAPAUU
NBRRYNLQPTVCNBRRYN L 6 P T V C.N B R R Y N L 6 P T V C N B R R Y N L
1
i
- ji
;i
i> ' i*
,j' IK 'A.
"
i i i
000
A S 0
U E C
G P T
-
|,
Af.
4
' V
'(
ป
1 1 1
000
N D J
0 E A
V.C'N
8686888888888886888888888866808888888888888888888
4444444444445555555555556666666666667777777777778
' . . . .- ' 'DATE- '-' -..- . '.
Figure 7.A4- Simulated and observed ammonia (mg/1), James River at Cartersville , VA,
1984-87.
-------
JAMES RIVER AT SEG. 280
Total N Concentrations (mg/1)
RED DASHED: SIM.. BLUE STARS: DBS.
NJ
4-
s -
I 3'
M
a
n 2-
d
0
B
s i-
0-
: . . /
' ' "'.*
* . ' .
ป 'ซ' i i : '!
fm^^AJjl^^
* '*.' x *
*.-. '
1 1 II 1 1 1 -1 1 II 1 1 1 1 1 1 1 1 II 1 1 1 1 II 1 1 II 1 1
000000000000000000000 00 0 0 00 0 000 0 0
JFMAMJJASONDJFMAMJJASONDJFMAMJJAS
AEAPAUUUECOEAEAPAUUUECOEAE A P A U U U E
NBRRYNLGPTVCNBRRYNL6PTVCNBRRYNLGP
88 888888 88 8888 888 88 8888 88 888 8 8 88 8
444 4 444444 44535 5 55555 S 5 5666 6 66666
*
*
f
'
ii
!\
J W
*
1 1 1 1 1
00000
0 N D J F
C 0 E A E
T V C N B
88888
66677
- -
M "' i
' ''' S -' '!
Ait j
\j*
m
i i i i i i i
0000000
M A M J J A S
A P A U U U E
R H i Y N L 6 P
8888888
7777777
. .
1 .
1A
^ X
Till
0000
0 N 0 J
C 0 E A
T V C N
8888
7778
DATE
Figure 7.46 Simulated and observed total nitrogen (mg/1), James River at Cartersvllle, VA, 1984-87.
-------
JAMES RIVER AT SEG. 280
Particulate-P Concentrations (mg/1)
RED DASHED: SIM., BLUE STARS: DBS.
s
I
M
a
B
S
1.5-
1.4-
1.3-
1.1-
0.9-
0.8-
CT.7-
0.6-
0.4
0.3
0.2
0.1
0.0
I
U
\ i i i r i i i7} r i i i i i i i i i i i i i i i i i i i i i i i i i i TJi i m i i n i r
0000000000.000 000 00 000.0000000000000 000000000 0000 00
1111,11 111111111 1111111 1111 111 11111111111 111111 111
J F M A M J d A S 0 N O J F M A M J J A S 0 N D J F M A M J d ASONDdFMAMddASONOd
A E A P A U U U E C 0 E A E A P A U U U E CO E A E A P A U U U E C 0 E A E A P A U U U E CO E A
NBRRYNLGPTVCNBRRYNLGPTVCNBR R Y N LGPTVCNBRRYNLGPTVCN
88888888 8 888 888B 88888888 88 8 88888888 88 8 888 8 8888888
444444444 4 4 455555 5. 555 5 55666666 66666 6,7 777777777778
'. . : DATE.
Figure 7.48' Simulated and observed particulate phosphorus (mg/1), Jnmes River at CartersvilIc, VA,
-------
to
~J
o
O
O
80
72
64
56
48
40
32
24
16.
.8
0
i r
i i i i i i i i i i i i rn i i i ri i i i i i i i i i i i i i i i i \ TIT T T
SIMULATED
* x OBSERVED
F MA M J J A S O.N 0
1984
J FMA MJ J A SO NO
1985
J FMAMJJASONDIJ FMAMdJASOND
1986
r
1987
JAMES RIVER TOC (MG/L)
Figure- 7.SO Simulntud and observed total oryunic carbon (my/i) , .hi
1984-87.
-n R.ivt, VA,
-------
PATUXENT RIVER AT SEG. 340
Total N Concentrations (mg/1)
RED DASHED: SIM.. BLUE STARS: DBS.
14-
s
I
M
a
n
0
B
8
13-
12-
11-
.
lo-
g-
s'
.
7-
6-
5-
4-
3-
2-
1-
IT I ป I I
TT 1
0000000 00' 000 O 00000000 OOOO 00000
lllllllllliilillilillillllllll
JFMAMJJASONDJFHAMJJASONDJFMAMJ
AEAPAUUUECOEAEAPAUUUECOEAEAPAU
M B R R Y N L 6 P T V C N B R R Y N L G P T V C N B R R Y N
888888886888688888888688868888
4 4 A A 4 A A A A 4 4 4 6 5 5 6 S 5 8 5 8 B B 5 6 6 66 66
DATE
* i i i i n n i i i ~~\ r r t i i r
bbooooooooooooooooo
i i
i i
JASONOJFHAMJJASONOJ
UUECOEAEAPAUUUECOEA
L 9 P T V C N B R R Y N L 8 P T V C N
8888886888688888888
6 6 6 6 6 6 7 7 7 7 7 7 7 7 7 7 7 78
Figure 7.52 Simulated and observed total nitrogen (m/'/l), Pntuxi'nt Rlyor at Ihiwli:, Ml), I9H/I-H/.
-------
SECTION 8.0
REFERENCES
Anderson, E.A. 1968. Development and Testing of Snow Pack Energy Balance
Equations. Water Resour. Res. 4(l):19-37.
Anderson, E.A., and ft.H. Crawford. 1964. The Synthesis of Continuous
Snownelt Runoff-Hydrographs on a Digital Computer. Department of Civil
Engineering, Stanford University. Stanford, CA. Technical Report No.
36. 103 pp.
Ardakani, M.S. and A.D. McLaren. 1977. Absence of Local Equilibrium During
Ammonium Transport in a Soil Column. Soil Sci. Soc. Am. J., 41:877-879.
Armbruster, J.T. 1977. Flow Routing in the Susquehanna River Basin: Fart I
- Effects of Rays town Lake on the Low-Flow Frequency Characteristics of
the Juniata and Lower Susquehanna Rivers, Pennsylvania, U.S. Geological
Survey Vater Resources Investigations 77-12. ' '
Bandel, V.A., S. Dzienia, G. Stanford, and J.O. Legg. 1975. N Behavior Under
No-Till vs. Conventional Corn Culture. I. First-Year Results Using
UnlabeledN Fertilizer. Agronomy Journal, 67:782-786.
Bandel, V.A. 1965. Plant Food Removal by Field Crops. Fact Sheet 98.
University of Maryland Extension Service. Baltimore, MD.
Bandel, V.A. 1990. Instructions for Using MANURE9.WK1: An Enhanced
Spreadsheet for Calculating Application Rates of Animal Manure
-------
Donigian, A.S., Jr., and N.H. Crawford. 1976a. Modeling Pesticides and
Nutrients on Agricultural Lands. Environmental Research Laboratory,
Athens, GA. EPA 600/2-7-76-043. 317 p.
Donigian, A.S., Jr., and N.H. Crawford. 1976b. Modeling Nonpoint Pollution
from the Land Surface. Environmental Research Laboratory, Athens, GA.
EPA 600/3-76-083. 280 p.
Donigian, A.S. Jr., D.C. Beyerleln, H.H. Davis, Jr., and N.H, Crawford. 1977.
Agricultural Runoff Management (ARM) Model Version II: Refinement and
Testing. Environmental Research Laboratory, Athens, GA. EPA 600/3-77-
098. 294 pp.
Donigian, A.S., Jr., and H.H. Davis, Jr. 1978. User's Manual for
Agricultural Runoff Management (ARM) Model, U.S. Environmental
. Protection Agency, EPA-600/3-78-080.
Donigian, A.S,, Jr., J.C. Imhoff. and B.R. Bicknell. 1983. Modeling Water
Quality and the Effects of Agricultural Best Management Practices in
Four Mile Creek, Iowa, EPA Contract No. 68-03-2895, U.S. EPA
Environmental Research Laboratory, Athens, GA, (PB-83-250183).
" * -'"'. /
Donigian, Jr., A.S., B.R. Bicknell, and J.L. Kittle, Jr. 1986 Conversion of
the Chesapeake Bay Basin Model to HSPF Operation. Prepared by AQUA
TERRA Consultants for Computer Sciences Corporation, Annapolis, MD, and
U.S. EPA Chesapeake Bay Program, Annapolis, MD.
Donigian, Jr., A.S., B.R. Bicknell, L.C. Linker, J. Hannawald, C.H. Chang, and
R. Reynolds. 1990. Watershed Model Application to Calculate Bay
Nutrient Loadings: Preliminary Phase I Findings and Recommendations.
Report Prepared for U.S. EPA Chesapeake Bay Program, Annapolis, MD.
Edwards, W.M., L.B. Owens, and R.K. White. 1983. Managing Runoff from a
Small, Paved Beef Feedict. J. Environ. Qual. 12(2):281-286,
Edwards, W.M., L,B. Owens, R.K. White, and N.R. Fausey. 1985. Effects of a
Settling Basin and Tiled Infiltration 3ed on Runoff from a Paved
Feedlot. In: Agricultural Waste Utilization and Management, American
Society of Agricultural Engineers, St. Joseph, MI.
Engman, E.T. 1986. .Roughness Coefficients for Routing Surface Runoff. ASCE
J. Irr. and Drain. Eng. 112(l):39-53.
Fitter, A.H. and C.D. Sutton. 1975. The Use of the Freundlich Isotherm for
Soil Phosphate Sorption Data. J. Soil Sci. 34:902-907.
Fox, R.H. and W.P. Piekielek. 1990. Long-Term Nitrogen Fertilization of
Continuous Manured and Non-manured Corn and Corn in a 3-year Alfalfa, 3-
year Corn Rotation: The First Six Years, 1982-87. Agronomy Series
Number 110. College of Agriculture, Pennsylvania State University.
276
-------
Imhoff, J.C., B.R. Bicknell, and A.S. Donigian, Jr. 1983. Preliminary
Application of HSPF to the Iowa River Basin to Model Water Quality and
the Effects of Agricultural Best Management Practices, Office of
Research and Development, U.S., Environmental Protection Agency,'Contract
No. 68-03-2895, (PB-83-250399).
Jackson, D. 1979. Hydrologic Studies of Effects of Conowingo Dam. .
Susquehanna River Basin Commission, SRBC Tech. Report No. 1, Harrisburg,
PA.
Jaworski, N. A. and L. C. Linker. 1991. Uncertainties in Nitrogen Mass
Loadings in Coastal Watersheds. Proceedings from 1990 Chesapeake Bay
Research Conference, In press.
J chanson, R.C., J.C. Imhoff, J.L. Kittle, Jr., and A.S. Donigian, Jr. 1984.
Hydrological Simulation Program-FORTRAN (HSPF): User's Manual for the
Release 8.0. EPA600/3-14066. U.S. Environmental Protection Agency,
Athens, GA.
Knisel, W. (ed). 1980. CREAKS: A Field Scale Model for Chemicals, Runoff,
and Erosion from Agricultural Management Systems. U.S. Department of
Agriculture. Conservation Research Report No. 26. 640 pp.
Kohler, M.A., T.J. Nordenaen. and W.E. Fox. 19SS. Evaporation from pans
and lakes. Research Paper No. 38. Weather Bureau.
Krone, R.B. 1962. Fluve Studies of the Transport of Sediment in Estuarial
Shoaling Processes. Hydraulic Engineering Laboratory and Sanitary
Engineering Research Laboratory, University of California, Berkeley, CA.
Lauer, D.A., D.R. Bouldln. and S.D. Klausner. 1976. Ammonia Volatilization
from Dairy Spread on the Soil Surface. J. Environ. Qual. 5:134-141.
Leohr, R.C. 1984. Pollution Control for Agriculture. 2nd Edition. Academic
Press, Inc.. Orlando. PL. 467 p. -
Lightner, J.W.. D.B. Meiujel. and C.L. Rhykerd. 1990. Ammonia Volatilization
from Nitrogen Fertiliser Surface Applied to Orchardgrass Sod. Soil Sci.
Soc. Am. J. 54:1471-1412.
Lugbill, J. 1990. Potomac River Basin Nutrient Inventory. Metropolitan
Washington Council of Governments, Washington, DC.
Hansel1, R.S., H.M. Sell*. P. Kanachanasut, J.M. Davidson, and J.G.A. Fiskell.
1977. Experimental and Simulated Transport of Phosphorus Through Sandy
Soils. Water Resour. Res. 13(1):189-194.
Maryland Power Plant Research Program. 1988. Acid Deposition in Maryland:
The Status of Knowledge in 1987. Annapolis, MD.
Maryland Power Plant Research Program. 1989. Acid Deposition in Maryland:
Summary of Results Through 1988. Annapolis, MD. .
278
-------
Olsen, S.R. and F.E. Khasawneh. 1980.. Use and Limitations of Physical-
Chemical Criteria for Assessing the Status of Phosphorus in Soils.
Chapter 14. In: The Role of Phosphorus in Aericulture. F.E. Kwasawneh,
E.G. Sample, and E.J. Kamprath (eds.). American Society of Agronomy,
Madison, WI. pp 361-410.
Onstad, C.A., and G.R. Foster. 1975. Erosion Modeling on a Watershed.
Trans. Am. Soc. Agric. Eng. 18(2):288-292.
Partheniades, E. 1962. A Study of Erosion and Deposition of Cohesive Soils
in Salt Water. Ph.D. Thesis, University of California, Berkeley, CA.
Penman, H.L. 1948 i Natural Evaporation from Open Water, Bare Soil, and
Grass. Proc. Royal Soc. London, Ser. A 193(1032):120-145.
. ^ ' '
Phillips, R.L., and M.R. Overcash. 1976. Economics and Technology for
Controlling Dairy Feedlot Runoff. Paper No. 76-4032. ASAE, St. Joseph,
MI. . . .. '..'' . ' -
Reckhow, K.M., M.N., Beaulac. and J.T. Simpson. 1980. Modeling Phosphorus .
Loading and Lake Response Under Uncertainty: A Manual and Compilation of
Export Coefficients. U.S. Environmental Protection Agency,
EPA 440/5-08-011.
\ ซ
Reddy, K.R. 1989. Unpublished results. University of FL, Gainesville, FL.
Reddy, K.R., R.E. Jessup, and P.S.C. Rao. 1988. Nitrogen Dynamics in a
Eutrophic Lake Sediment. Hydrobiologia, 159:177-188/
Reddy, K.R. and W.H. Patrick. 1984. Nitrogen Transformations and Loss in
Flooded Soils and Sediments. CRC Critical Reviews in Environmental
Control, 13(4):273-309.
Reddy, K.R., M.R. Overcash, R. Khaleel, and P.W. Westerman. 1980. Phosphorus
Adsorption-Desorption Characteristics of Two Siols Utilized for Disposal
of Animal Wastes. J. Environ. Qual. 9(l):86-92.
Reddy, K.R., R. Khaleel, M.R. Overcash, and P.W. Westerman. 1979a. A
Nonpoint Source Model for Land Areas Receiving Animal Wastes: Ammonia
Volatilization. Trans. ASAE. 22(6):1398-1405.. .
Reddy, K.R.., R. Khaleel, M.R. Overcash, and P.W. Westerman. 1979b; A
Nonpoint Source Model for Land Areas Receiving Animal Wastes: I.
Mineralization of Organic Nitrogen. Trans. ASAE. 22(4):863-872.
/ . '
Sample, E.G., R.J. Soper, and G.J. Racz. 1980. Reactions of Phosphate
Fertilizers in Soils. Chapter 11. In: The Role of Phosphorus in
Agriculture. F.E. Kwasawneh, E.G. Sample, and E.J. Kamprath (eds.).
American Society of Agronomy, Madison, WI. pp. 263-310.
-280
-------
U.S. Army Corps of Engineers. 1956. Snow Hydrology, Summary Report of the
Snow Investigations. North Pacific Division. Portland OR. 437 pp.
U.S. Army Corps of Engineers. 1988. Water Needs Assessment of the
Susquehanna River Basin. Engineering report prepared for the U.S. Army,
Baltimore, MD.
U.S. Bureau of the Census. 1980. 1978 Census of Agriculture, Volume 1,
Geographic Area Series, U.S. Department of-Commerce.
U.S. Bureau of the Census. 1984. 1982 Census of Agriculture, Volume 1,
Geographic Area Series, U.S. Department of Commerce.
U.S. Department of Agriculture. 1978. Estimating U.S. Livestock and Poultry
Manure and Nutrient Production, Report No. ESCS-12, Economics.
Statistics, and Cooperatives Service.
U.S. Department of Agriculture. 1975. Animal Vaste Management Field Manual,
Soil Conservation Service, U.S. Government Printing Office.
U.S. Department of Agriculture, Soil Interpretations Records (SCS-SOI-5 data)
file; and the National Resources Inventory (NRI), Soil Conservation
Service. - . . .
U.S. Department of Agriculture. 1984. Usual Planting and Harvesting dates
for U.S. Field Crops. Agricultural Handbook No. 628, Statistical
Reporting Service, U.S. Department of Agriculture, Washington, DC.
U.S. Environmental Protection Agency. 1979. Animal Vaste Utilization oh
Cropland and Pastureland: A manual for Evaluating Agronomic and
Environmental Effects. 'Report No. EPA 600/2-79-059, Office of Research
and Development. '
U.S. Environmental Protection Agency. 1982. Chesapeake Bav Program Technical
Studies: A Synthesis. Washington. D.C.
U.S. Environmental Protection Agency. 1983. Chesapeake Bay: A Framework for
Action. Region 3, Philadelphia, PA;
U..S. Forest Service. 1982. Forest Statistics for New York, 1980, Resource
Bulletin NE-71, U.S. Department of Agriculture, Forest Service,
Broomall, PA.
U.S. Forest Service. 1980. Forest Statistics for Pennsylvania, 1980,
Resource Bulletin NE-65, U.S. Department of Agriculture, Forest Service,
Broomall, PA. . .
U.S. Forest Service. 1978. Forest Resources of West Virginia, Resource
Bulletin NE-56, U.S. -Department of Agriculture, Forest Service,
Broomall, PA.
282
------- |