vxEPA
United States
Environmental Protection
Agency
Environmental Research
Laboratory
Athens GA 30605
EPA-600/3-79-054
August 1979
Research and Development
Nonpoint Source
Model Calibration in
Honey Creek
Watershed
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency, have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR) •
7 Interagency Energy-Environment Research and Development
8 'Special" Reports
9 Miscellaneous Reports
This report has been assigned to the ECOLOGICAL RESEARCH series This series
describes research on the effects of pollution on humans, plant and animal spe-
cies, and materials Problems are assessed for their long- and short-term influ-
ences Investigations include formation, transport, and pathway studies to deter-
mine the fate of pollutants and their effects This work provides the technical basis
for setting standards to minimize undesirable changes in living organisms in the
aquatic, terrestrial, and atmospheric environments
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/3-79-054
August 1979
NONPOINT SOURCE MODEL CALIBRATION
IN HONEY CREEK WATERSHED
by
T. H. Cahill, R. W. Pierson, Jr., and B. R. Cphen
Resources Management Associates
West Chester, Pennsylvania 19380
Grant No. R-805421-01
Project Officer
Thomas 0. Barnwell, Jr.
Technology Development and Applications Branch
Environmental Research Laboratory
Athens, Georgia 30605
U.S. Environmental Protection Agency
Begion 5, Library (5PL-16)
290 S. Dearborn Street, Room 16cu
Chicago, IL 60604
ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
ATHENS, GEORGIA 30605
-------
DISCLAIMER
This report has been reviewed by the Environmental Research
Laboratory, U.S. Environmental Protection Agency, Athens, GA.,
and approved for publication. Approval does not signify that
the contents necessarily reflect the views and policies of the
U.S. Environmental Protection Agency, nor does mention of trade
names or commercial products constitute endorsement or recommen-
dation for use.
11
-------
FOREWORD
As environmental controls become more costly to implement
and the penalties of judgement errors become more severe,
environmental quality management requires more efficient
analytical tools based on greater knowledge of the environmental
phenomena to be managed. As part of this Laboratory's research
on the occurrence, movement, transformation, impact, and control
of environmental contaminants, the Technology Development and
Applications Branch develops management or engineering tools to
help pollution control officials achieve water quality goals
through watershed management.
These efforts include programs to provide state-of-the-art
models for analyzing agricultural nonpoint pollution and evaluate
their effectiveness as land management tools. One tool to eval-
uate the magnitude of diffuse pollutant load reductions is the
EPA Non-Point Source Model. This report applies the model to a
large watershed in the Great Lakes region in an evaluation of
selected best management practices.
David W. Duttweiler
Director
Environmental Research Laboratory
Athens, Georgia
111
-------
ABSTRACT
The U.S. EPA Non-Point Source Model has been applied and
calibrated to a fairly large (187 sq. mi.) agricultural water-
shed in the Lake Erie drainage basin of north central Ohio.
Hydrologic and chemical routing algorithms have been developed.
The model is evaluated for suitability as a land management tool
in estimating the reduction in sediment associated phosphorus
from cultivated land. Results indicate a very different impact
for selected best management practices (BMP) estimated using
the Universal Soil Loss Equation (USLE) gross erosion approach.
Modifications to the model structure are suggested to improve
BMP evaluation.
This report was submitted in fulfillment of Grant No.
R-805421-01'by Resources Management Associates under the
sponsorship of the U.S. Environmental Protection Agency. The
report covers the period 17 October 1977 to 16 October 1978,
and work was completed as of 16 October 1978.
IV
-------
CONTENTS
Foreword iii
Abstract iv
List of Figures vii
List of Tables ix
List of Abbreviations and Symbols xi
Metric Conversion Table xi
Acknowledgment xi i
1. INTRODUCTION 1
2 . CONCLUSIONS 3
3 . RECOMMENDATIONS 8
4 . BACKGROUND 12
1 7
Physiographic Setting ^
Stream Flow Record ^
Weather Record ^
Chemical Record "'
5 . BASIN CONSIDERATIONS ] 9
Surface and Sub-Surface Flow System 19
Land Resource Data Base 23
Use of Imagery 23
5 . MODEL CONSIDERATIONS 28
Cover Factor (COVER) 28
Moisture Accounting in the Soil Mantle 38
Evapotranspiration 39
Hydrologic Parameters (LANDS) 44
Index Infiltration (INFIL) 44
Interflow (INTER) 44
Pollutant Generation Parameters (QUAL) 45
Rainfall Kinetic Energy (JREP) 45
Erodability (KRER) 46
Soil Fines Transport (JSER and KSER) 46
v
-------
page no,
7 . MODEL CALIBRATION 48
1976-Record of Hydrologic Events 48
Input Data Adjustments 55
Snowmelt 55
Precipitation Record 55
The NPS Model "Scale" Problem 56
Pertinent Knowledge 56
Initial NPS Model Calibration 58
Hydrologic Calibration
8. HYDROLOGIC ROUTING SUB-ROUTINE 66
Prior Hydraulic Routing 57
Time of Travel (TT) 57
Results of Routing 71
Time of Concentration (Tc) 75
Intrabasin Routing (IBR) 77
Chemograph Routing 79
9. MODEL VERIFICATION 86
1977 - Record of Hydrologic Events 86
Model Verification 89
10.EVALUATION OF BEST MANAGEMENT PRACTICES 94
NPS Model Parameter Adjustment 95
NPS Model Simulation 98
Suitability for No-Till 98
Estimation of No-Till Parameters 100
Simulated Erosion Control by No-Till 100
Evaluation of BMP's by USLE 104
REFERENCES 109
APPENDIX A Land Resource Information 111
System
APPENDIX B Precipitation Records 128
VI
-------
FIGURES
Number Page
1 Hydrogeology of Honey Creek Basin 13
2 Honey Creek Basin Sampling Stations 15
3 Climatological Stations 16
4 Particulate Phosphorus - Suspended Solids Curve 18
5 Flow Pathways for NFS Model 20
6 Flow Pathways for Honey Creek Basin 21
7 Daily Flow of Honey Creek at Melmore, 1976 22
8 Example of Cell-Point Data Coding System 24
9 Political Sub-Divisions in Honey Creek Showing Venice
Twp 25
10 Estimation of Monthly Cover Conditions for Crops in
Honey Creek 37
11 Palmer Analysis of Soil Moisture 40
12 Hydrograph - Storms "A", "B", "C" and "D" (semi-log
plot), including precipitation, Snowmelt and Soil
Temperature 49
13 Particulate Phosphate - February Runoff, 1976 50
14 Suspended Solids - February Runoff, 1976 51
15 Nitrate and Nitrite - February Runoff, 1976 ........... 52
16 Hydrograph Simulation without Routing 65
17 Channel Cross-sections used in Routing ................ 68
18 Channel Segments and Sub-Basins used in Routing 70
19 Schematic Representation of Time of Travel Routing .... 73
vii
-------
LIST OF FIGURES (contd.)
Number Page
20 Simulated Storm D Routed by Time of Travel 74
21 Schematic Representation of Time of Travel plus
Time of Concentration 75
22 Simulated Storm D Routed by TT and Tc 76
23 Schematic Representation of Time of Travel plus
Interbasin Routing I 77
24 Schematic Representation of Time of Travel plus
Interbasin Routing II 78
25 Simulated Storm D Routed by TT and IBR 80
26 Simulated Flow of February - March, 1976, Routed by
TT and IBR 81
27 Chemograph Precedence of Hydrograph - 1976 82
28 Routed Chemograph, Storm D, 1976. (SS, mg/1) 84
29 Routed Chemograph, Spring, 1976. (SS, mg/1) 85
30 Late Winter Runoff Period, 1977 87
31 Actual Hydrograph and Chemograph, Spring, 1977 90
32 Simulated Flow, TT plus IBR, Spring, 1977 91
33 No-till Cover Values 101
34 No-till ys Conventional Cover Values 102
35 Conventional Cover Values - Row and Field Crops 103
Vlll
-------
TABLES
Number Page
1 Water Quality Record Summary 17
2 Land Resource Information System 26
3 Land Use above Melmore 27
4 Crop Allocation by County 30
5 Crop Rotation Sequences 31
6 Frequency of Crop Sequence 31
1 Monthly Cover Conditions by Crop 32
8 Land Conditions preceding and Following Major Crops .. 33
9 Estimation of Monthly Cover Conditions - Row Crops ... 34
10 Estimation of Monthly Cover Conditions - Field Crops . 35
11 Synthesis of Monthly Cover Conditions 36
12 Water Balance Analysis - Feb. 29 to July 3, 1976 41
13 Annual Moisture Balance - 1976 42
14 Annual Moisture Balance - 1977 43
15 Major Storm Events - 1976 53
16 Comparison of Mass Transport 54
17 Precipitation - Runoff Comparison 57
18 Initial Hydrologic Calibration of NPS Model, 1976 59
19 Mass Transport Estimates - 1976 62
20 NPS Model Calibration Summary - 1976 64
21 Land Use Composition by Subbasins 72
ix
-------
LIST OF TABLES (contd.)
Number Page
22 Major Storm Events, 1977 88
23 Verification Summary 92
24 Land Management Practice vs NPS Parameters 96
25 Suitability of Cropland for No-Till 99
26 Erosion Control by No-Till Crop Production 105
27 Effect of No-Till as Estimated by USLE 107
28 Comparison of Methods for BMP Evaluation, NPS Modeling
vs USLE Analysis 108
x
-------
LIST OF ABBREVIATIONS
Abbreviation
sq. mi.
ha.
MSS
D.A.
cfs/sq. mi.
PE
•- square mile
-- hectare
-- multispectral scanner
-- drainage area
-- cubic feet per second per square mile
-- potential evapotransportation
ENGLISH UNIT
acre (acre)
cubic foot (cf)
cubic feet per second (cfs)
degree Fahrenheit (deg F)
feet per second (fps)
foot (feet) (ft)
gallon(s) (gal.)
parts per million (ppm)
pounds per acre (Ib/acre)
square foot (sq ft)
ton (short) (ton)
ton per acre (tons/acre)
yard
METRIC CONVERSION TABLE
MULTIPLIER METRIC UNIT
0.405
28.32
28.32
0.555(°F-32)
0.305
0.305
3.785
1.0
1.121
0.0929
907.2
0.3674
0.914
hectare (ha)
liter (1)
liters per second (I/sec)
degree Celsius (deg C)
meters per second (m/sec)
meter(s) (m)
liter(s) (1)
milligrams per liter (mg/1)
kilograms per hectare (ka/ha)
square meter (sq m)
kilogram (kg)metric ton
metric tons per hectare
(metric tons/ha)
meter (m)
XI
-------
ACKNOWLEDGMENT
This applied research on the NFS model application could
not have been possible without extensive chemical sampling and
analysis over a two year period by Dr. David Baker and his staff
at Heidelberg College, supported in large part by the Buffalo
District, U.S. Army Corps of Engineers.
Xll
-------
SECTION 1
INTRODUCTION
The Honey Creek basin is a 187 sq. mi. tributary of the
Sandusky River (1430 sq. mi.) draining from north central Ohio
into the western edge of the central basin of Lake Erie. Like
much of the Lake Erie direct drainage area, the physiography is
glacial in origin and is characterized by flat, high water table
soils which are very productive if draine.d. The storm runoff
from similar agricultural lands around the basin is estimated
to contribute about half of the annual total phosphorus input to
the lake (COE, 1975). Even with projected point source reduc-
tions in phosphorus loading, it is anticipated that this agri-
cultural non-point phosphorus load must be reduced by 35% to
achieve desired water quality in the lake. Land management
strategies are being evaluated, but the source areas must be
defined, as well as erosion and transport mechanisms of the sedi-
ment associated phosphorus understood, to rationalize the appli-
cation of feasible strategies.
One of the management tools considered to evaluate the mag-
nitude of diffuse pollutant load reduction is the EPA Non-Point
Source Model (EPA, 1976). The Honey Creek basin is substantial!}
larger than the test plots and small catchments used for initial
model development, but represents a good practical scale for
watershed planning. Successful use of the model for input to de-
cision making by the agricultural technical community and farmers
will greatly enhance the selection of those tillage practices
which will achieve the greatest pollutant load reduction.
The available chemical transport record at the major gage
station of Melmore (D.A. 149 sq. mi.) is excellent, much better
than most stream systems of this size. On the other hand, the
meteorologic record does not include any well-located continuous
precipitation recorders. While the rainfall record is poor by
comparison to prior research applications of the NFS model, it is
representative of the data availability in most watersheds of
this size across the country. The combined effect of good chem-
istry and precipitation data should still allow satisfactory
calibration of a model such as NPS, especially in terms of total
mass transport estimates.
One major limitation of the NPS model is the "scale" prob-
lem, with the original equations written for small catchments
and lacking stream transport and routing subroutines in its
-------
original form. For the Honey Creek application, routing from I •»
subbasins which average 10 sq. mi. in size to the Me lino re gage
is necessary for hydrograph and anemograph simulation.
A second consideration for application of the NFS model in
Honey Creek is that the flow pathways are different from that
envisioned in the basic model structure. While modification of
the algorithms would improve the simulation of moisture storage
and movement, it was felt that the model should be applied in
its original form and all adjustments made by the calibration
process, to allow full evaluation of the model. That is, the
model can be improved for this particular basin, but then its
transferability to other basins will be limited.
This study is interesting in that it attempts to transfer
the NFS model based on small test plot research to large water-
shed land management in one giant step; in retrospect, it might
better have been accomplished by building up from a series of
small subbasin applications, with corresponding model evolution,
However, the usual constraints of time, money and data necessi-
tated that the work proceed in its present form. 'The reader
should carefully distinguish between problems and limitations
associated with the model itself and those inherent in the
Honey Creek data base, especially the rainfall data.
The NFS model is part of a set of models under development
by EPA1S'Athens Environmental Research Laboratory. In addition
to NFS, the Agricultural Runoff Management (ARM) model is a
process-oriented model describing the fate of pesticides and
nutrients on the land surface. The ARM hydrology module is
similar to NFS. A system is currently under development which
links the hydrology and pollutant generation modules from the
ARM and NFS models with channel process algorithms and a data
management system. This system, called the Hydrologic Simula-
tion Program-FORTRAN (HSPF), is based on the proprietary Hydro-
camp Simulation Program and is expected to be generally avail-
able in early 1979.
-------
SECTION 2
CONCLUSIONS
The work described in this report is applied research; al-
though the NFS model is used as the basic tool for analysis and
subsequently modified for that purpose, many of the deficiencies
identified in the version used for this application have since
been corrected, are presently being addressed or could have been
better considered under the "sister" ARM model. The important
question answered in this study is what will be necessary to use
models of this general type as operational tools for agricultur-
al land management analysis. It is our conclusion that such an
objective is possible within the existing model structure and
framework, given certain basic data requirements and model re-
finements. The existence of detailed hydrologic, chemical and
meteorologic data for a watershed to be calibrated cannot be
assumed without understanding the limitations and requirements
of the model itself. Routinely gathered stream chemistry and
weather data as presently collected by the various agencies are
inadequate for a model such as NFS. The investment in stream
sampling and chemical analysis for the data base used in this
study (1000 samples, storm and non-storm over two years, 1 sta-
tion) represents a substantial cost, but it is obvious that a
much greater investment would have to be made to produce a
calibration that is truely satisfactory. For a basin of this
size, no less than five stream stations should be used, plus
at least three continuous rainfall recording stations.
One of the complimentary features of this watershed is the
existence of a computerized land resource data base, which
greatly facilitated the sub-basin analysis for routing (Section
9) and the spatial variability of selected land management prac-
tices (Section 10). Given certain further refinements in the
basic NFS algorithms, such an integration of the land data
system with a hydrologic and chemical model would be even more
powerful a tool for analysis of alternatives. This line of in-
vestigation warrants further research.
Original NFS Model Deficiencies
Many of the limitations experienced with initial model cal-
ibration were recognized by the original model developers (EPA,
1977 a) , but it is important to describe the experience in this
watershed. The "scale" problem associated with the NFS model
-------
in its original form cannot be overestimated, as the application
to basins greater than 50 hectares requires more than just sub-
routines that estimate fluid wave movement and particulate
transport through channel systems, as have been developed in
this study. The model in its present form must be revised to
consider the time delay of the subsurface flow component, both
interflow and groundwater, beyond the adjustment allowed by the
parameters IRC and KK24. Conceptually, the model simulates an
infinitely small basin in which the discharge of interflow and
groundwater occurs far too quickly, producing a hydrologic re-
sponse that is too rapid for even the smallest of basins. The
runoff response must be delayed temporally by a "time of satura-
tion" for all land areas beyond the stream channel and
wetland areas. A spatially variable UZSN might be one way to
simulate this effect, based on a classification of drainage
types or something similar. In addition, the "Interflow" path-
way should draw from the lower zone storage, rather than the
upper zone (Fig. 3, LANDS simulation,NFS Model,EPA 1977a).
For a larger watershed (/- 50 ha), the movement of fluid
wave, fluid parcels and suspended particles through the channel
network and»the integration of a number of sub-basins must be
considered before a broad application of the NFS model is poss-
ible. This study offers some rather simplistic approaches to
this set of problems and produces a satisfactory result. The key
question raised in the pursuit of this solution is how important
is the exact shape and timing of a hydrograph, a chemograph, or
both, if the total simulated quantities are correct. The answer
is fundamental to all mathematical modeling analysis: do we op-
timize the simulation, or try to understand the process that
produced the actual data. For example, the chemograph peak pre-
ceding the hydrograph peak implies that the suspended solids
come from only part of the basin; whereas the water comes from
a much larger area, but with much of it delayed in movement
through the soil. This is reinforced by the knowledge that in
the open channel, the fluid wave will move faster than the sus-
pended solids. The spatial differentiation of these chemical
response areas is essential to management application, and so
the use of this model as a decision-making tool must go beyond
the "lumping" of key parameters such as soil and land cover
characteristics. This applies to both the LANDS and especially
the QUAL sub-routines, where exponents such as KSER are far too
general.
The precedence of the hydrograph peak by the chemograph is
considered by some investigatiors (Verhoff, unpub.) as evidence
that the chemograph is produced almost entirely by in-channel
processes and is not the direct result of land runoff during
a storm event. There is little question that channel scour-
ing and deposition processes take place during runoff, result-
ing in meanders, flood plains and other fluvial characteristics.
It is unlikely, however, that any major portion of the pollu"
-------
tants carried out of the watershed originated in the channel it-
self. In all probability, the channel storage of fines is a
temporary sink of materials, eroded from the land surface, in
transit out of the basin. In either case, for the NFS model to
satisfactorily deal with this question some consideration must
be given to include a "channel processes" algorithm in the chem-
ograph routing analysis, (as presently included in HSPF) a step
that was not taken in this initial study.
Potency factors should be held constant for variables such
as particulate phosphorus, which have a strong relationship with
suspended solids. This is not to suggest that these factors do
not vary slightly with different storms and seasons, but
by allowing them to be varied on a monthly basis as has been
done in prior work with the NFS (the result of initial focus on
urban watersheds), they tend to be misleading and may mask
hydrologic variability as chemical variability. Other elements
of the QUAL sub-routine are too coarse, at the present, to war-
rant a degree of refinement such as monthly potency factors.
The evapotranspiration process, as discussed in the report,
is the key to sucessful moisture accounting in a basin. A
monthly value of K3 (as presently included in ARM) would not
only be an improvement but also is supported by the comparison
of adjusted pan evaporation data with various evapotranspiration
estimates. Variable K3 values were used in this modeling study
for model calibration runs of 1976 data, but not for model veri-
fication with 1977 data, although a better result could be anti-
cipated if K3 had been varied. The potential use of Palmer In-
dex data, available throughout the country, should also be re-
cognized as input in the absence of pan evaporation data, or
even in lieu of it.
A great deal of improvement and refinement of the QUAL al-
gorithms is recommended, with correlation studies of JRER with
rainfall kinetic energy and antecedent rainfall (implying an
event or seasonally based value) and analysis of KRER with "K"
(erodability) values in comparative watersheds under identical
or similar hydrologic and cover conditions. The KSER exponent,
as discussed in the report, simply contains too many distinct
variables to be lumped together, except in instances where spa-
tial definition is not possible. Two alternatives are possible:
rewrite the equation to distinguish the variable or develop "co-
occurrence" statistics on the combination of slope, soil parti-
cle size distribution, cover and drainage density in respective
sub-basins from the data base (envisioned as a four-way matrix)
and calibrate KSER using a sub-basin data set. In either case,
the coefficient must be refined to allow any type of management
analysis of the basin.
Although the preceding discussion focuses on inadequacies
of the NPS model in its unmodified form, in all fairness it
-------
must be recognized that the hydrologic simulation achieved by
run 33 was fairly good, except for the exaggerated peaks and
time displacement. Also, the suspended solids simulated was
very close to the actual mass transport, although the simulated
transport came off in a flash with each storm.
NFS Model Extensions
The extensions to the basic model that were made for the
Honey Creek calibration are listed here, with pertinent conclus-
ions .
In-stream Time of Travel (TT). The time of travel for both
fluid waves and particles was estimated for main channel seg-
ments and added to respective sub-basin simulated hydrographs.
Kinematic velocity was used for fluid waves and fluid velocity
was used for particles, both estimated from channel geometry and
hydraulic equations of open channel flow.
Intra-basLn Time of Concentration (Tc). Sub-basin simulat-
ed hydrographs were displaced temporally by an estimate of time
required for fluid and chemical wave production, based on sub-
basin size, cover and average slope.
Sub-basin routing (IBR). Simulated hydrographs for sub-
basins were redistributed temporally, or routed, to approximate
the time lag of actual hydrograph development, over the period
of Tc (as a function of Tc).
Sub-basin integration. A routine was developed to inte-
grate the modified and displaced hydrograph and chemograph simu-
lations at a downstream station (Melmore).
Output. An output plotting capability to generate both
simulated and actual output on a CALCOMP plotter, at variable
scales, was developed and used.
Other Program Changes. A few mechanical changes were made
in the program over the research study to improve output format,
data processing or general usability.
The initial conclusion reached with respect to the exten-
sions listed is that they are essential for application of the
NFS model to any watershed larger than a few hectares, and a
basin the size of Honey Creek (47,000 hectares) should be cali-
brated for sub-basins individually, rather than attempting to
calibrate the entire watershed at only one station (Melmore).
Given the data available for this effort, much remains to be
learned in taking such a large step at once. Considering the
simplicity of routing equations used and the actual complexity
-------
of fluid wave and particle movement through the basin's infinite
number of pathways, the simulations produced are satisfactory.
This sense of satisfaction, however, should not allow compla-
cency with the results. The simulations mask many important but
subtle processes that must be understood to improve land manage-
ment techniques. The NPS model, even with the modifications
made herein, is not yet able to completely simulate and distin-
guish all the effects of possible changes in agricultural land
management (other than variations in COWEC and other factors
as discussed in Section 10) potentially suitable for this water-
shed.
Finally, it is concluded that the use of NPS and other
hydrologic-chemical models for land management analysis will
result in more accurate estimates of erosion reduction than
methods such as the USLE. Continuous accounting of soil mois-
ture is the key to predicting the magnitude of hydrologic and
chemical response in a watershed, all other things being equal.
There is also the very real danger of blanket application of
management practices that could be as effective if applied on a
very selective basis. To avoid this, such models must be sensi-
tive to spatial variability.
-------
SECTION 3
RECOMMENDATIONS
The following recommendations are intended to accomplish
two purposes: improve the NFS model in general terms to make
it more applicable as a management tool and evaluate the poten-
tial role of modeling in improving agricultural land management
in a demonstration watershed. On one hand, many technical steps
can be taken to improve this particular model, representing
state-of-the-art technology for analyzing the problem of diffuse
pollutants. One needs to evaluate the overall benefits of de-
veloping a "better" model, however, to determine whether the
additional investments are warranted.
ROUTING SUB-ROUTINE
A real question exists as to how much more refinement need
be added to the procedure developed in this study for management
analysis. Better routing routines can certainly be written
given the intrabasin hydrographs and chemographs for calibra-
tion. With actual sub-basin chemographs and hydrographs, the
routing coefficients and hydraulic storage can be estimated,
with incremental hydrograph and chemograph routing. This re-
sult could then be compared with the more simplistic approach,
using the same data sets.
MODEL REFINEMENT
Even with the addition of routing sub-routines, the NFS
model should be refined in several important ways.
Spatial Differentiation
For the application of this model to land management ana-
lysis, it will be necessary to distinguish among those elements
or combinations of factors within a hydrologic unit or sub-ba-
sin that will vary both hydrologic response and/or pollutant
generation. This is not to suggest that the responses are
mutally exclusive, but rather that a given land area may allow
a major portion of incident precipitation to contribute to run-
off without contributing significant pollutants. Certainly it
would be impractical to analyze each unit area, or cell, within
a basin individually, but the aggregation of cells into clusters
or response groupings within hydrologic units would be entirely
-------
feasible, given the type of data base that exists in Honey Creek.
From a modeling point of view this step would not add greatly to
the model's complexity, but should greatly improve its ability
to predict the highly variable response to storms that are a
function of soil moisture storage. More importantly, the eval-
uation of various land management techniques, which are almost
entirely dependent on spatial variables within a sub-basin
(soils, cover, slope), could be based on the simulated hydro-
logic and chemical responses.
The initial analysis of "hydrochemical response" groups
could be done within the existing LRIS framework, applied to
the present NFS algorithms, and evaluated. With calibration,
the selected management practices could be evaluated for appli-
cation and suitability in a given hydrochemical response group
and the respective algorithm modified, if necessary. ' It is
suggested, however, that the basic structure of the existing
NFS model be left intact until it is shown necessary for the
sake of evaluating sensitivity to a given practice.
QUAL Algorithms
As discussed previously, several of the equations and para-
meters used in the QUAL sub-routine require refinement. The
KSER parameter includes too many real variables to allow any
kind of calibration to the real world, and its refinement should
be a separate task in itself. The apparent sensitivity of this
parameter to land surface conditions makes it an important man-
agement variable, but lumping this factor together with fixed
watershed variables such as sediment particle size and slope
only makes the comparison of practices more difficult. Drain-
age or grading modifications, such as terracing or berms, that
attempt to alter the overland flow pathway will also be impos-
sible to distinguish. The wide range of values experienced in
various calibration efforts reflects the uncertainty in the use of
this parameter in its present form. The exponent JSER is held
to a very limited range of variability. The reformulation of
this equation could significantly improve the applicability of
the NFS model for management analysis.
The foundation for development of the detachment co-effi-
cient (KRER) and exponent (JRER) is much more firmly based on
prior research. it is recommended however, that their exact
relationship with rainfall kinetic energy (E) or erodability (R)
be better defined, as well as the relationship between KSER and
the "K" factor of the USLE. A wealth of experimental data
exist with respect to the first problem, whereas the second
may be solved by analysis of multiple sub-basins with all vari-
ables held constant except soil type. Certain storm events in
the Watkinsville data set might meet the criteria, or analysis
of much larger sub-basins such as in Honey Creek or the full
Sandusky. Refinement of these two relationships will allow
-------
most of the calibration process to take place in the transport
algorithm.
Cover Factor (COWEC)
The most obvious variable controlling sediment production
in a watershed is the vegetative land cover; the most straight-
forward management practice is to modify the density and timing
of this cover on cultivated fields. Given these two facts, it
would seem obvious that the COWEC variable would be the most
important basin characteristic to be measured prior to applica-
tion of the NPS model. Ground cover in an agricultural water-
shed such as Honey Creek, however, varies both temporally and
spatially in a given year, and from one year to the next. Data
collecting efforts for COWEC during 1976 in Honey Creek - con-
sisting of photointerpretation and computer coding of IR imag-
ery and farmer surveys - were exhaustive for a basin of its
size, allowing the NPS model to be finely tuned for this para-
meter. With the passage of time, the quality of this factor in
the data file will deteriorate without some simplified method of
updating. Prior research with NASA in this basi*n has suggested
that low altitude multispectral scanner (MSS) data collection is
prohibitive in cost and of limited value, as is high altitude
photography. The question then is whether a cheaper, lower res-
olution imagery source such as LANDSAT can serve this purpose.
A large NFS calibrated basin such as Honey Creek should be ana-
lyzed using a series of available LANDSAT tapes to determine
whether bare soil, live vegetation, dead vegetation, snow and
combinations of these can be distinguished at critical times of
the year. Again, this is a study subject in itself, but its
potential importance for widespread acceptance of the NFS model
should not be overlooked.
In spite of the various deficiencies experienced and iden-
tified in the NFS model, or perhaps because of them, the re-
search summarized in the preceding paragraphs must be conducted.
Model development is an evolutionary process, with each succeed-
ing generation or product more useful, if the underlying struc-
ture and theory are sound. In this respect, the NFS model of-
fers the greatest potential of presently developed tools because
of its focus on continuously accounting for soil moisture, the
key determinant in both hydrologic and chemical response in a
watershed. Only by knowing what the antecedent moisture condi-
tions are in a watershed with respect to precipitation events
can we hope to understand and intervene in the process of agri-
cultural erosion. Any analytical approach that neglects this
consideration, no matter how finely divided the data set that
describes the spatial determinants of the watershed, will prove
inadequate. At the same time, the combining of sensitive para-
meters for a watershed as is presently done in NFS parameters
such as KSER will not give sufficient definition for management
analysis. The most useful model must be structured with the
10
-------
strong features of both approaches; continuous moisture account-
ing with spatial differentiation of variables.
Research Issues
Several fundamental research questions having tremendous
implications for the success of land management techniques to
reduce soil associated pollutants are only partially answered
for diffuse sources. First and foremost is the question of
source areas: classic soil erosion theory as well as test plot
data tells us that, all other things being equal, the more
steeply sloped areas, usually in the headwaters of a basin, will
contribute more erosion. On the other hand, partial area hydro-
logic theory (Dunne, 1976) suggests that overland flow occurs on
only a limited portion of the basin: those low lying,high water
table soils contiguous to the stream channel. Alternative con-
clusions will lead to very different management strategies. If
most of the sediment being carried from an agricultural water-
shed is coming from only 15 to 20% of the basin land area, and
that area is the less permeable, poorly drained soil, the im-
proved drainage and/or reduced cultivation technique might be
more successful than those strategies which attempt to minimize
rainfall erosive energy by improving cover conditions, or re-
ducing velocity of overland flow by slope reduction.
The sediment transport question is also important. Does
material take many years to move out of a watershed, with sub-
sequent scouring and deposition in a number of temporary sinks,
such as swales, ditches and stream bottoms, or does the majority
of material observed in passage from a basin during a storm
event originate on the land surface during that storm? This
question also weighs heavily on the selection of suitable land
management techniques. The time required to measure any success
achieved by such alternatives as fertilizer reduction is very
much dependent on the answer to this question.
It is recommended that in our pursuit of a better model to
simulate the mechanisms of diffuse pollutant production and
transport that we not ignore these basic research issues. The
improvements in water quality necessary over the next decade
must be at the lowest possible cost, necessitating a better
understanding of the processes involved than we presently pos-
sess .
11
-------
SECTION 4
BACKGROUND
In order to understand the logic behind parameter selection
and modifications to existing model algorithms, it is necessary
to understand the unique characteristics of the Honey Creek
basin and both the potential and limitations of existing hydro-
logic and chemical records.
PHYSIOGRAPHIC SETTING
The Honey Creek basin (D.A. 179 sq. mi.) is a tributary to
the Sandusky River (D.A. 1430 sq. mi), which drains from north
central Ohio into the western edge of the central basin of Lake
Erie (GLBC, 1975). The surface physiography of Honey Creek is
characterized by glacial deposition: 57% of the land form is
ground moraine, 21% is end moraine and the balance is comprised
of lacustrine deposition, outwash, flood plain and glacial drift.
Basins closer to the lake shore lying to the west of the study
basin (Honey Creek is 30 miles from Lake Erie) are comprised pre-
dominantly of lacustrine materials and are even flatter; the
Portage is a good example. Within Honey Creek, the land is char-
acterized as being poorly drained with a portion of the head-
water area comprised of a swampy wetland, the remains of glacial
Lake Willard. A major geologic feature of the regional subsur-
face strata is the axis of the Cincinnati Arch, a "slightly
north-plunging regional anticline whose long axis extends from
north to south across western Ohio from Cincinnati to the vicin-
ity of Toledo." Running generally parallel and to the west of
the Sandusky River, this feature defines the sedimentary beds
that comprise a carbonate aquifer that many geologists believe
constitutes the major ground water flow system (Norris, 1974).
The recharge area for much of this aquifer is in the headwaters
of the Scioto River basin, lying south of the Lake Erie basin
divide. (Fig. 1). This movement of ground water north toward
the Lake shore is an important factor in the Honey Creek basin,
controlling the high water table conditions that exist in most
of the watershed.
The surface and sub-surface conditions in the Honey Creek
basin suggest that little, if any, surface water moves into deep
or inactive groundwater storage. In fact, even the active
groundwater storage may be minimal because the base flow values
average 0.01 cfs/sq.mi. and less. The watershed is extensively
12
-------
HONEY
CREEK
BASIN
HIGH YIELD AREA
WESTERN EDGE OF
CARBONATE AQUIFER
AXIS OF CINCINNATI ARCH
Figure 1. Hydrogeology of Honey Creek basin,
13
-------
underdrained with artifical tile, a necessary feature for culti-
vation usage. The majority of drain tiles lie at a depth of 3
to 5 feet below surface, forming a subsurface drainage network
not unlike a first order stream network. The flow of water
through the soil mantle to the tile, the gradient, size and age
of tile, the elevation of the outlet and possible submergence
during storm runoff all contribute to the interaction of this
flow pathway in the basin hydrograph. For the initial NFS model
calibration, the tile drain system will be conceptualized as
very rapid interflow (Fig. 3, LANDS simulation NFS), but this
will be modified in future model development. The existing
hydrographs for the Melmore gage do not show any distinct in-
flection point on the semilog plot of recession curves,
suggesting the separation of true surface runoff. This
interflow and groundwater flow may be impossible to calculate
accurately on the basis of hydrographic analysis alone. The
chemistry data, however, used in conjunction with the reces-
sion curve do offer insight as to the tile drain flow; this is
discussed in a later section.
STREAM FLOW RECORD
The U.S.G.S. established a permanent stage recording sta-
tion at Melmore (D.A. 149 sq. mi.) in the Honey Creek basin
in January 1976 to record hourly stage readings. A rating
curve, with adjustments, has been in development until
the present. Other flow measurements in the basin have been
irregular and are inadequate for the present model calibration
effort. Thus the Melmore data set will constitute the primary
information for model development, based on a 21 month record.
(Fig. 2).
WEATHER RECORD
No continuous precipitation record exists within Honey
Creek proper, but two observation stations exist at Tiffin (Fig.
3), just at the mouth of the basin. One record is hourly values,
non-recording, and the other is a daily record. The basin is
bracketed by two other daily record stations at Plymouth and
Bucyrus, which are equidistant from the centroid of the basin
with the Tiffin gage, (each s* 14 miles). The elevation of both
these stations is several hundred feet higher than Tiffin. Re-
view of major storm events over the period of record shows a
fairly good consistency between these three stations,
with the exception of brief summer thunderstorms, which for the
most part, produce little runoff at the Melmore gage. The snow-
fall and snow-on-ground record is also from Tiffin, whereas
soil temperature is taken from the Ohio Agricultural Research
Center at Wooster, approximately 55 miles east southeast of
Honey Creek. Although not ideal, the available weather data ap-
pear adequate for initial calibration. Of course, the time
factor is significant, and the hydrographs will be evciluated
14
-------
to
c
o
•H
4J
(0
-P
co
C
•H
rH
I
(d
CO
CO
Q)
u
0)
c
o
w
CN
Q)
M
d
Cn
•H
15
-------
CO
c
O
•H
-P
O
•H
(ji
O
rH
O
-P
(C
E
-H
i-l
O
ro
(D
S-i
3
tn
•H
16
-------
primarily by shape rather than time, because a time displacement
could easily be expected.
CHEMICAL RECORD
The Melmore data set consists of 967 samples through July
10, 1977, with measured parameters as shown in Table 1. For
model calibration, the parameter " Particulate Phosphorus"
( TP ,), defined as TP-OP, will be used in comparison with
suspended solids. This relationship is excellent for the data
set and gives a " potency factor" of 0.002,(Fig. 4). Neither
the OP nor the NOo records lend themselves to so simplistic an
analysis, however, and the flow and seasonality influences are
predominant. Thus the initial calibration of the QUAL sub-
routine will use the SS and TPpart parameters only.
Table 1.
WATER QUALITY RECORD SUMMARY
HONEY CREEK BASIN
Melmore Gage : 1969-1975 Irregular grab sampling
Ortho Phosphate (OP)
Total Phosphate (TP)
Suspended Solids (SS)
1976-77 Daily plus 6 hr. intervals
During runoff
OP, TP, SS, NO2+NO3, NH3
Bi-weekly - pH, CL, Si, TKN
Irregularly-conductivity
Secondary Stations from 8-6-76 to present
Grab sampling over runoff
plus weekly.
(Same parameters as Melmore)
17
-------
|/6ui (d -O snuiiu d-i) 84Dndsoi|j
18
-------
SECTION 5
BASIN CONSIDERATIONS
SURFACE AND SUB-SURFACE FLOW SYSTEM
The conditions that exist in the Honey Creek basin with
respect to water movement through the soil mantle are very
different from that conceptualized in the general development
of the NFS model. For this reason, simulation of the flow
pathways will require the development of different algorithms
for certain portions of the model. First, the Honey Creek flow
network will be compared diagramatically with that envisioned
in the NFS model; next, each model parameter will be reviewed
and modified, where necessary.
Figure 5 shows a simpified diagram of the flow pathway
envisioned in the NFS model, with the movement of interflow
along the saturated zone above the water table dependent upon
the soil permeability and hydraulic gradient, just as active
groundwater storage will contribute to streamflow as stream
level recedes. In Honey Creek, much of the cultivated land has
been extensively artifically drained with tile networks (of
varying dimension, design and condition), producing a system as
simplified in Figure 6. The tile system provides a flow pathway
which can be thought of as first order stream input, very rapid
interflow or a short term reservoir of soil moisture, because in
many soils the intervening soil layers have low permeability
and the time of movement throughout the lower zone by infiltra-
tion and perculation may be significant. Upon reaching the
tile system the movement can be very rapid if there are not
hydraulic restrictions to flow (i.e., flooding of outlets).
The zone between tiles (Fig. 6) will comprise the active ground-
water storage zone, varying in capacity ,as infiltration drains
into the tile and eventually decreasing to zero as the water
table returns to the invert elevation of the tile network. The
secondary active groundwater storage zone situated below the tile
system and comprising long term base flow is quite small, as re-
flected by the annual hydrograph for 1976 (Fig. 7) which shows
base flows decreasing close to zero for the full basin. In ef-
fect, the introduction of an artifical drainage network under
much of the basin has lowered the water table sufficiently so
that the basin is almost totally dependent on precipitation in-
put for maintenance of stream flow. This pattern is character-
istic of most of the region in northwestern Ohio and southeast-
19
-------
o
p
& tf
& w
as
J ,
ij»
%
h-^
IU
I
^
s
•
>ui
i
19
IS
QJ
T3
O
e
c/3
o
m
w
>i
«d
o
in
(U
in
•H
20
-------
UPPER
ZONE
LOWER
ZONE
HONEY CREEK
- FLOW PATHWAYS
TILS' INT£KJ=LQW
o o o
GROUND WATETR
STORAae CMlNiMAt.)
\
LATERAL INFLOW
TO TILES
TABLE
SECTION ACROSS
Figure 6. Flow pathways for Honey Creek basin.
21
-------
_D
i
D)
•I
o
*
LU
H | MTTTT 1 1
- o o §
39HVHDSIQ
G
0
•H
.p
to
CD
M
O
iH
0)
g
0)
M
u
>1
a)
c
o
M
m
o
5
o
Q
cu
5-1
22
-------
ern Michigan, as well as other parts of the country where tile
systems are an integral part of intensive cultivation systems.
LAND RESOURCE DATA BASE
A computerized data base has been developed for the Honey
Creek Basin. The data base contains spatially referenced in-
formation for those characteristics listed in Table 2. The
data are encoded for a randomly located point in each of 11,700
four-hectare (approx. 9.88 acres) cells. The application of
this method (Bliss et al. , 1974) has proven very successful
for the characterization of spatial features in land-water
systems analysis, providing a statistically reliable set of
information on the combination of ingredients that exist in a
watershed. Figure 8 shows the location of points and cells,
as encoded, in a portion of the Honey Creek basin near Attica,
superimposed on a 1:24,000 quad sheet base. For each point,
some 11,500 covering the basin, the data listed in Table 2
are stored in a file, which varies in size and complexity de-
pending on the parameter. The soil file, for example, is
taken directly from the SCS-5 data and includes complete in-
formation on the physical and chemical properties of each soil
horizon, or layer. Other files, such as land cover, consist
of a coding number for a particular coding scheme. Appendix A
provides a more detailed description of the Land Resource In-
formation System (LRIS) as it exists in the Honey Creek basin.
Of particular interest in this research study is the
question as to how such a spatial data base can be integrated
with a non-point source model such as NFS. The initial prob-
lem to be considered is that of estimating the type and spatial
distribution of alternative flow pathways, as described in the
preceding section, with the existence of tile drainage systems
in much of the cropland. As shown in Table 3, the land use dis-
tribution in the basin above the Melmore gage (Fig. 2) is pre-
dominently agricultural, (82%) with the proportion of row crop
to field crop about 1.7 to 1, as of June 1976. What the com-
puter file does not contain, at present, is information concern-
ing which cropland is presently artifically drained.
USE OF IMAGERY
Several pieces of information have been considered in order
to evaluate the question of tile systems. Most townships in
Honey Creek keep records of ditch construction and maintenance,
many of which show up on the U.S.G.S. quad maps, but some on
which are not clearly defined. Township maps show all major
ditches, but do not show sub-surface tile systems. In Venice
Township (Fig. 9), a survey of farms was conducted during the
summer of 1976, with about a 25% sampling. One of the pieces
of information gathered was the location, if known, of tile
drains in the fields under cultivation bv the farmer interview-
23
-------
*—'
+•1 ^
Figure 8. Example of cell- point data coding system,
Honey Creek basin.
24
-------
-------
Table 2. LAND RESOURCE INFORMATION SYSTEM
DATA FILES
For each of the 11,500 point/cells (4 ha) covering the
Honey Creek watershed in a UTM grid, the following
information is coded:
Land Use by both point and cell : 8 major and 37
minor classes. Dominant land use in
cell and specific land use at random
point in cell.
Soil Phase by soil series and slope category.
Soil file data from SCS computer file
for "A" horizon with suitability classes,
"K" factors, "T" factors and other para-
meters .
Watershed in 26 sub-areas.
Slope Magnitude in tenths of a percent.
Slope Direction by azimuth from 0° to 360°.
Elevation of point in tenths of foot.
Physiographic Region by soil series groupings.
Political Sub-Division
ed. These were mapped, where possible. Finally, color infra-
red imagery gathered in June 1976 was studied to determine pos-
sible surface moisture patterns as an indication of sub-surface
tile systems. The imagery revealed the patterns of tile systems
identified in the survey quite accurately. Black and white
photographs from August 1970 were also studied, but they revealed
that the change in surface soil moisture offered only a mottled
appearance in tiled areas, with the vegetation obscuring many
drainage features. Thus it was concluded that, a combination of
IR imagery, photointerpretation and field survey can produce a
fairly good picture of the spatial extent of tile systems. This,
of course, is a level of detail that would be difficult to re-
plicate in most modelling studies. Therefore, it was decided
that for this initial model calibration effort, the spatial
variation in sub-surface drainage would not be. considered.
26
-------
Table 3. LAND USE IN HONEY CREEK ABOVE MELMORE
Land Use Catagory Area(ha.) Percent of Basin
Row crop
Field crop
Other agriculture
Woodlot
Wetlands
Water
Urban
Missing Data
19,820
11,992
240
4096
840
392
1848
0
50.5
30.5
0.6
10.5
2.2
1.0
4.7
TOTAL
39,228
27
-------
SECTION 6
MODEL CONSIDERATIONS
COVER FACTOR (COVER)
The QUAL subroutine of the NFS model requires specifica-
tion of land cover values by month. These values represent
the percentage of the soil surface covered by crop or crop
residue that act to absorp the energy of raindrops. In agri-
cultural regions, the simplest case is the cultivation of a
single crop year after year using the same plowing, planting
and harvesting schedule. In the Honey Creek watershed, sev-
eral crops are grown, each with its own cover characteristics.
Furthermore> certain crop rotation sequences are used so that
the description of cover conditions of soy beans preceded by
wheat, for example, differs from soy beans preceded by corn.
Finally, two distinct land preparation schedules are used in
the watershed: fall plowing, in which the soil is bare all
winter, and spring plowing, in which residue from the prior
crop remains on the field over the winter and early spring
months.
Studies performed by RMA for the Corps of Engineers (RMA,1976)
provided data useful in the development of monthly cover values.
From a survey of 28 farmers in Venice Township in August and
September 1976, it was learned that corn, soybeans, wheat and
oats account for nearly all of the crops grown in the area.
The percentages of crop land in the survey devoted to each of
these four crops are similar to percentages for the three prin-
cipal counties in which the Honey Creek watershed is found (see
Table 4) . The crop rotation sequences used by the farmers sur-
veyed is shown in Table 5. Counting two-crop sequences in this
table reveals the frequency by which each of the major crops is
preceded or is followed by the other crops. These frequencies
are shown in Table 6. For example, when corn is planted, 58%
of the time it is preceded by wheat, 24% by corn, and 16% by
soybeans. Explicit use of this information will be made below
in developing the aggregate cover conditions for row crops and
small grains.
Each crop was examined month by month for the cover it
provides to the soil (See Table 7). The numerical assignments
made are based on discussions with Seneca County SCS and SWCD
28
-------
personnel and on descriptions of farm management practices
found in Ohio CES literature, particularly Bulletin 594 (Ohio
Erosion Control and Sediment Pollution Control Guide). Row
crops (corn and soybeans) are described for both fall plowing
and spring plowing schedules. About half of the land covered
by the Survey is spring plowed with the prior crop residue left
on the field as an erosion control practice. Typical corn
yields of approximately 125 BU/acre produce about 7000 Ibs of
residue per acre or 80% cover. In spring plowed schedules,this
80% residue cover was assumed to persist until February or March
when slight decreases occur because of decomposition, wash off
and blow off of the crop residues. Soybeans typically produce
about 2000 Ibs residue/acre which can cover about 40% of the
soil surface. Wheat and oats produce 5000 and 10,000 Ibs of
residue per acre, respectively. In both cases, it is assumed
that this residue is left on the field giving an 80% cover.
Wheat is planted in the early fall, grows slightly before
winter dormancy, then reemerges in early spring. After harvest
in July, wheat is followed by a cover crop - clover, alfalfa or
other legume. This crop is often planted before wheat harvest
and can quickly establish a good cover before its growth is
slowed by winter temperatures.
Oats are planted in the early spring and harvested in July.
Oats are nearly always followed by wheat.
Based on the above considerations and assumptions the
monthly cover assignments were made for each crop as shown in
Table 7 .
Next, each major crop was listed with the prior crop and
post-harvest events. Whenever several alternatives exist, the
fraction of the time each alternative occurs is designated
based on information in Table 6 . This crop event list is
shown in Table 8 .
Finally, a synthesis of the percent cover information
(Table 7) and the crop frequency information (Table 6 ) was
achieved by straightforward weighing of percent cover informa-
tion, as shown in Tables 9 and 10 . Monthly values for row
crops and for small grains are listed separately because the
Honey Creek land use file spatially distinguishes these two
crop categories. The weighted average monthly cover values for
row crops (corn and soybeans) and small grains (wheat and oats)
are summarized in Table 11 . Figure 10 shows a plot of the final
percent cover values for row crops and small grains.
29
-------
ft
H
K
to
Q.
O
o
EH
H
U
H •
s <
W EH
> <
Q
fe
O W
Q.
O
O EH
to i-q
H ID
PH U
< H
ft C^
S O
O <
U
O vO fO
CN — irv
— co cvi
EH CO
O U
U
ffi
>i EH
CQ H
O U3
H EH
EH ^
< D
U CO
W
ft >
O K
& D
u to
CD
30
-------
Table 5. CROP ROTATION SEQUENCES USED BY 26 FARMERS
IN HONEY CREEK BAS^N ( VENICE TOWNSHIP)3
3 crop sequences
ccw
cbw (5)
cow
4 crop sequences
5 crop sequences
ccco
ccbw(2)
ccow
cbcb
cbcw
cbbw
cbow(6)
cowb
cccbw
ccbow
cbcbw
cbbow
cwbob
a. The number in parentheses is the number of farmers
using that rotation sequence for those rotations
reported more than once. c=corn, b-soybeans, w=
wheat followed by winter crop, o=oats.
Crop
Table 6. FREQUENCY OF CROPS PRECEDING AND FOLLOWING
A GIVEN CROP a
Percent frequency of a crop pair
crop previous year crop year after
o
o
C/l
c
TO
0)
-Q
>-
O
in
ro
OJ
TO
O
O
O
1/1
c
03
0!
-Q
TO
O
Corn
Soybeans
Wheat
Oats
24 16 58 2
82 7 7 4
12 43 0 '4.S
31 69 0 0
24 58 8 10
22 7 34 37
92 8 0 0
7 7 85 0
a. The frequency of all crop pairs in the rotation
sequences shown in Table 5 were determined in
order to infer and weigh cover conditions.
31
-------
o
UJ
o
o
•z.
ro t-
§ g
H
o H-
CJ UJ
CO
D
O
D
2 <-3
H 3
EH <
o -i
O i/>
^^ ~O
0
0,
§ > o
f-M ^ .—
U -SL 4-"
u
>H *"
« i
1
co «:
2 Q-
O <
H
EH
Q •
§ 1
U Z
w
CO m'
8 s:
M
EH z
S 4fi^
o ->
s
.
r--
Q)
•H
X!
rt
E"1
O O O O O
00 -3-00
o o o o o
OO -3-OO
0 O O 0 O
-3" CO CM -T CO
0 O O O 0
vO vO LA LA OO
O O O O O
oo oo -3- -a- f*\
• • • • I
o o o o o
-CT -3" CM CM LA
o o
— — LA LA LA
CTi
O O O O O
oo
•
o o o o o
LA
"
O O O O O
\O rA CM
O O O LA O
r^» CA CM
0 O O O 0
OO J- CM
.
.
• • • • • • r—j
(/} ,_) flJiHWi— ICQQJr-l" O*
&i Oj iHCUC(ifl'HCi)M m!>
O (S (d O O
M r-H • • d) rH Q) • • -PO
C_) EH r~l CH W CT^ .O r~? Xt W CT^ I — 1 f$
^H f^ ^-J Q) CU K^ t^ K*^ Q^ Qi rH d) ^1
O U U CO CO g S
P4 CO
o
CM
0
LA
0
o
0
oo
•
o
OO
"
LA
cn
0
-3-
.
o
1
1
1
•
.H
o
w
4J
o '
32
-------
Table 8. LAND CONDITIONS PRECEDING AND FOLLOWING MAJOR
CROPS IN HONEY CREEK WATERSHED
Previous Crop
(Jan. thru Mar.)
Crop land prep.
This Crop
(Apr. thru harvest)
Next Crop
(Harvest thru Dec.)
Crop land prep.
corn
fa 11 p1ow
spr. plow
soybeans fall plow
spr. p1ow
wheat
oats
winter cov
er phase
fa 11 p1ow
spr. plow
12
12
8
8
- 58
corn
(46% of row crop)
row crop fall plow
(and oats) spr. plow
wheat
fal1 plow
and plant
46
corn
fall p1ow
spr. p1ow
soybeans fall plow
spr. plow
wheat
oats
winter cov-
er phase
fa 11 p1ow
spr. p1ow
42.5
42.5
3.8
3.9 soybeans
7.7 (54% of row crop)
1.9
1.9
row crop fal] plow
(and oats) spr. piow
wheat
fall piow
and plant
wheat
corn
planted
previous
fall
fal1 plow
spr. plow
soybeans fall plow
spr. plow
100
15
16
35
34
wheat
(75% of small
grain crop)
oats
(25% of small
grain crop)
alfalfa, interseed
clover or in wheat
other or plant
legume in residue
row crop fal1 plow
spr. p1ow
wheat
fa 11 p1ow
and plant
iOO
Percentage indicates the proportion of the subject crop area (this crop)
that is found in the specified preplant or post-harvest condition.
33
-------
cn >
O
Is Li
s °
p i.
CO
o •
t"! I
o
t**-"* >.
a;
*^,
Rf1
c
c
ll ?•
^
•->
6
S
o.
*•» V.
TIMATT
Feb. M
co
•T J5
o»
.q
rj
corn
*~t
ft
ji
£
o
C"
\ _J
c
o
1_
CT
— i
C
L.
C
u
prior tc
O !r— o |— co
oio °=r ":c
-=" -cr o
u <-> 3 -C
O 0' , - C, L-S
OO 1 -— O i O LT 1 1 (* ., 1-3"
0 0 -1 -I | !
— ~ OD
*NI CN C LTi
U U 2
°L °
0
0 1 O
03 to CO
1 '
O
°L °
O
o|o o
™n "~
oio o
o
o|o o
o
V *3 O -3*
(N C^l CO 0 ^ ~
— — o • • .
CO ( Q,-*
O LA (N OkC Oio O
CIO r- olo r-C «(3; o|o co|g £
esj c^j co co co — —
o o y\ o Q
C lj -D ^j X Q Q
vO (>J CO i
coo c. o °t5 S° S
o|o • o|o -y °^i- o|o ..
CN rx CO CO CO —
— — O O u^ O o
fd
i. J -0 ^ 3 "5 g
(L'
"
T:
10
c
a>
p
0
Ci
_
c
IT;
"O
c
i soybeans
o
olo .... .
•*- y 5 -C \
0 C u-M— OJO ! —
olo ' " '
LTi LTl LT
LT, (Ni rsl
-'.r. -* o
0
01
~- U -
OiO O W -D X
^.j^ i ^. U T (T
y, a a- •*•
c s; ^ " -,
— C. O a. c
OJO o1 'm J-: -C
• • c o ^ t;
u -- - o c
0 £, o -C
" ^^ C TO U
oio o £ S *"^ -
I/" 0 L *D C C1
C O ft) y- T
S -' t §• 15 a^
olo ; o S u o o £
"~| o >. o y _> ^
0) , C •— o O ^-"S^ H1*
L- o kc ^"^ C" CO O f-> ¥ j:- c t-
d vO I r\i OIO •• •• • pi~/ OCj^-
|0|0 ' ' un LT. " — Z o ^
j- -a- o Co uls'DC
6 i H „
o o -2 -C "m if ""• O
^ r~,
O O ^^ O CO O r^ "§ .- O l-
olor^f^iOio • . . if i- '" c'
LA LA '* £ t: C c.
: ! : ! ? i- si!
oUr 0— c NO — S^""-1^
oio ^r ol° " ° co° J iz k^i
Q ..2 air
r--\ r<> r^\ f~\ r- (~ i 0 *• '~
,r .=- 0 0 o £-b£|
-' urj^ 3 o t +- , „
y fD •*• — ' — *
IT --
C. !"•
-------
iz;
H
en
tf
O
O
H
EH
H
Q
2
O
U
O W
U >
o
>H U
J
ffi 1-1
EH H
S O
O W
O
s o
O H
H EH
EH U
H
o
rH
C/3
a
<
?
c,
dJ
U
OJ
o
L>
r o
CO
o
CO
OD
O
CO
f*\
*
1
0
1
un
C
0)
U 03 C >
0) M O
4_i — (J (_)
C 0)
0
CO
c
CT
OJ
JO
3
O
L.
r
03
CJ
-C
2
o
CNi
O
o
o
_;
o
o
—
C ft)
<13 1) O
— -C
CL 3 0
__'
OJ • — O
0) 3 — CO
E -a o
D — tS\
1/1 D C
TJ ^- O "~
o
•
o
o
o
g
m r-- oio
*-i O^ -3" CO
Jl ...
1) 1/1
> *J LA LA
u ro •
2
Q
o
3
O
w.
-T
LT\ LTV
o
o o
-3" -3"
o
o
.
9-
Q. 4-J
C t/1
ra
Q Q. O
olo
o
o o , o
OlO vO — O|O f-\
2
.
3
-o
1- D
aj -•
c
ol— o|o
olo
- ,o|o
•o
o
U "^~.
fU (U o
1- 3 X
a. -a CD
.- o -a
-c c
.- T3
Q- —••
C in
O 0)
01 —
i— — fO
O O 1-
— 3
O
.
a. a. a.
o o o
w O >
000
CL II
Oi CQ
Q.-0
t/i O
3
-Q O
>- —
o a.
••II !! II
(U
rH
X!
n3
EH
35
-------
CO
O
O Csl
sO O
o —
N£> O
CO
u
IS
§
OO -*D
vD —
o
CM
CO
§
Q
iz;
O
U
-3-
O
O
U
ffi
EH
a
o
g
CO
H
CO
w
ffi
CO
fO
EH
36
-------
a:
LU
o
o
Jan. Feb. Mar. Apr. May Jun. Jul. Aug.Sep. Oct. Nov. Dec.
75-
50-
a:
LU
25-
Small Grains
Jan. Feb. Mar. Apr. May Jun. Jul. Aug. Sep. Oct. Nov. Dec.
Figure 10. Estimation of monthly cover conditions for
crops in Honey Creek, (see table 12)
37
-------
MOISTURE ACCOUNTING IN THE SOIL MANTLE
The hydrologic (LANDS) simulation algorithms for the NFS
model are driven by the process of infiltration, the rate for
which is highly dependent on changes of moisture content in
the soil profile. The concepualization of a surface layer of
soil and an underlying soil mantle as two moisture storage
compartments or zones, with variable moisture contents, is
reasonable and consistent with existing theory and experience.
In reality, the descending soil horizons are usually a. grada-
tion of moisture contents, complicated by varying permeabili-
ties and textures. In general, however, the surface layer
usually loses moisture much more rapidly than the underlying
soils in cultivated fields, with the "plow depth" as the def-
inition of upper layer.
The initial task in any hydrologic calibration is to
develop a water balance for the test plot or watershed, pre-
ferably on an annual basis, (Eq. 30, p. 178, NFS). The most
difficult aspect of this process is estimating the moisture
that enters deep percolation, a storage component that is
impossible to measure, except by inference. For the Honey
Creek basin, however it is not unreasonable to assume that
the amount of precipitation that eventually contributes to
deep percolation is zero, based on the underlying aquifer con-
ditions, the low permeability of many soils and the extensive
under drainage. The changes in soil moisture storage can be
measured reasonably well for small plots as were used in ini-
tial ARM model calibration, but for a watershed the size of
Honey Creek this becomes a formidable task. At present, no
field measurements of soil moisture are being made within the
basiru
In agricultural regions of the country, a great deal of
thought has gone into evaluating changes in soil moisture in
order to optimize planting of crops and anticipate drought
cycles as they begin to develop. One such method or model used
by the U.S. Weather Service (NOAA) is the Palmer Index Analysis
(Palmer, 1965), which calculates soil moisture changes based on
climatological conditions and evaluates drought trends and
severity. The analysis is done on a regional basis (there are
10 regions in Ohio) for the 35 week growing season from late
February to early November. In this analysis, soil moisture in
both the upper soil layer and underlying soil is estimated, as
well as potential evapotranspiration (Thornthwaite Method) and
actual evapotranspiration.
It was believed that this soil moisture modeling data would
be of great value in applying the NFS model in various regions
across the country, if it correlated reasonably well in predict-
ing actual runoff from larger watersheds, such as Honey Creek.
The Palmer data for the 18-week spring runoff period in 1976,
38
-------
from February 29 to July 3, were evaluated in terms of the mois-
ture budget equation, with the runoff from the Honey Creek basin
estimated to be 3.16 inches, (Table 12). The weather data used
in the analysis included 11 stations in north-central Ohio (Div.
02). The average precipitation compared favorably with the
Tiffin daily record. The evapotranspiration and changes in soil
moisture were calculated on a weekly basis, and showed a net
loss of 1.82 inches of soil moisture to runoff. During the same
period, the stream gage at Melmore recorded a total flow of 3.16
inches out of the Honey Creek basin. This perfect balance is
certainly only a coincidence, but does reinforce the assumption
of zero deep percolation in the basin. More importantly, it
gives credence to the moisture accounting methods used in the
Palmer model. The soil moisture changes as estimated for both
surface and underlying soil are shown in Fig. 11.
An important question with respect to the NPS model cali-
bration is how well the evapotranspiration predicted in the
Palmer model compares with adjusted pan evaporation data, since
both are used in the Honey Creek calibration. These pan evapo-
ration data are gathered at the Hoyville Agricultural Research
Station (Fig. 3), but only for the 7-month period April through
October. A comparison of values for the period April 4, 1976,
through July 25, 1976, showed poor correlation. The ratio of
evapotranspiration to adjusted pan evaporation varies consider-
ably, from 0.38 to 1.02. The estimation of this factor in the
NPS model (K3) as the fraction of the watershed with deep rooted
vegetation appears to be invalid in the Honey Creek basin, which
has only 10.5% of the basin in woodland. In fact, it seems rea-
sonable to apply a monthly value of K3 to better estimate the
changing ratio between actual and pan evaporation.
EVAPOTRANSPIRATION
Accurate estimation of this component of the moisture
accounting process is the key to successful calibration of the
NPS Model, at least in a hydrologic sense. As discussed pre-
viously, the Palmer Drought Index Model, developed by NOAA,
uses a modified version of the Thornthwaite Equation to calcu-
late evapotranspiration and associated soil moisture changes in
upper and lower soil layers. In the Palmer Model-, the potential
evapotranspiration (PE) is calculated on a weekly basis for a
growing season (March through October) and on a monthly basis
for the full year. A comparison of these estimated values of
evapotranspiration are shown in Table 13 and 14 for the Honey
Creek basin during 1976 and 1977. Also shown are average
annual values from lysimeter measurements at the Coshocton agri-
cultural research station, some 70 miles eastsoutheast of Honey
Creek. These various values can be compared to the net loss,
assuming no deep groundwater storage, or loss of soil moisture.
For the summer months of June, July and August, all methods
estimate a higher evapotranspiration than net loss, suggesting
39
-------
\
\
s1
I
£
D)
_X
3
I
i
CTi
CO
(0
0)
cu
M
u
(U
c
o
K
OJ
-p
tn
•H
i
r-1
•H
O
w
•H
S-l
0)
g
H
(ti
oo
O
(saipui) nos NI aamsiow
40
-------
vo
rH
ro
>H
o
EH
CN
M
w
2
H
CO
pa
— OO_
oooooooooooo
-
L. }— O">
•>l— ro
l? fe
w
U
>H
C C
re O
OOOOOOOOOOOO
O
a
o
CM
CO
H
— o
. .c
. ro o
O-ir-^r^ivDf-^ — rsiCOvDOfNlo
— oooooooo — o—
v£> i— ro o f"~ -=r
O ^- O ^ — C (N
II II II 0)
w
a
PQ
fNI
rH
(1)
H
X)
EH
— CO
0 -
UO —
41
-------
en
rH
H
Cd
O
O
u
w
u
CO
w
H
§
z
C6
a.
03
w
0)
rH
XI
(0
EH
tn ch «g-
vc O f*l
O rH f-l
*r r~ r-
o ^* ^
o o o
fN
rH C
fN fN
rH
r- m
fN n
CO
co en
fN fN
< rH n
o mm
SO VO O
co n so
o n m
CO "V m
O m n
In >i
S? £ rH IH
— o a -H a
C C -H « 0 —
0 -H O T3 J= C
•H a> —
4J -HE iH^CC 3
(3 i-H -H d 0) tH -H C
JJ (TJ U VU VM r-
CX 3 ffl O -rl -H E
•rt 0» •UCE-'EH (C
O C IB O -H 1
1) W — EH — V.
U 4.
0- r» r~ ^
O rH rH fN *M
vo in Tf r*
O VO fN rH fN
. • • • VD
o > Cb
d) >,rH C >1
>H rH kl O O T3
O "H 3 -H E (U
E in n) o -u JJ
rH cn T3 JC nj au u)
OJ o ^ 3
S iJ O O 0 C 1-1
ii a n n
raj jj a. a r: a <
41 >
z w
c
0
Evapotranspirati
OS
rH
CO
rH
00
fN
VO
m
vo
*•"
en
irt
in
o
m
Ch
n
rH
m
m
fN
01
•H
•~ ,_,
15
II
Average Annual
Coshocton Lys
i
o
0 LT
—4 >H
o c'
m r- f»
m tN «
rH rH rH
CO CO C
O O rH
n^ r*i ro
21 £ £"
n f*j a
•»«r m
CO VO Is-
in in ^'
r» co ts
0 0 rH
in in «*
00 CTv V3
fN fN tN
ro ^* fN
Ov P- C"
rH rH r-l
CO CO <=
O O rH
fN V
o d
0 1
M
0.
Palmer Index (
Weekly
Monthly
Ttiomwaite Equation
U = 1.6 (lot)"'
•rc
42
-------
u
w
Q
O
Z
L)
O
a,
w
en
CO
rH
£
IN
CO
in
o
r-
o
CTi
3
O
g
J
EH
H
w
S3
O
w
w
u
I
<:
CQ
g
tD
EH
CO
H
s
J
rtj
!D
S3
<
QJ
J
H
K
£0
W
O
3
O ra cn • -u O CX
C rt • Q O E c
U} ~^ m O H ~ -H
CM O ^
o . r-
IN' ° r-:
c^ o "I -co
IN a\ ^ c
CM rH O CO
In
^i
m fc CM
vo o m t; ix)
in v£> ,
• • _j ^ ^,
ro m •"! CM ^
en o "-1
CO LO "^
TT *T fsj
KS ~
rH 1-5 -H
CO O rH
O C^ rH
•HO o
>i -~- V
>irH • r-
rH r-1 C -H
•H 3 -H -H
IB O ^- — >
Q S O >, 4J
II — (U >,rH C >1
~CC3S-I rHl-lOO
in -H -H o O — i 3 -H ^
(1)IWM-I rH E W r3O -^j
jju-iiw in rH in Q— to G.
O -H -H E <1) 0 rH
CE-I5-1 ni S J OO Oi C
•H (1) 1 1 Oil 1C
^" rHBjt .15 10,
-P O >i
to z wi
!"""> ^*. °
^T , ^T
in' "^ •*'
VD _ IN
l£) ° rH
i^ ^n -^
en _ cc
CM <^ 0
rn m CM'
, m
fH CO
rn c'
"1 o
CM 0
cr; o
rH 0
B
7H
r- c ^ £
° -° -1 gj a
J-J 13 t g4
>, [^ 3 -in -^J
^ C K . ,
C" -rH ^ ^~t X
S a < ^ A1 .
ft. tnl c )= •>, ^
•q 5 &,£ q3§
S i 2| glB
.3 o a,' •§ 5S-
^ a. > o n
s, * < o ft,
w
43
-------
that soil moisture is being given up from storage; for the bal-
ance of the year the Palmer and Thornthwaite equations appear
far more representative than the adjusted pan data or the Coshoc-
ton data. The pan data for February, March, November and Decem-
ber used in the NFS model were based on the Palmer values.
HYDROLOGIC PARAMETERS (LANDS)
Index Infiltration (INFIL)
Recognizing that infiltration is both spatially and tem-
porally variable in a watershed, the NPS model uses an input
parameter (INFIL) to establish an index infiltration level,
equal to the mean infiltration capacity ( £~) when the current
storage of soil moisture in the lower zone (L3S) is equal to the
index level of moisture storage (LSSN) for the lower zone, or
L3S/L8SN =1 . This parameter is a function of soil character-
istics, such as cohesiveness and permeability, and is estimated
to vary by two orders of magnitude, (0.01 to 1.0 in./hr).
With a cellular data base such as exists in Honey Creek, it
is possible to apply specific data on infiltration capacity vari-
ability by soil series distribution within the watershed, or even
evaluate varying infiltration by soil horizon. The computer
files show that the majority of soils in Honey'Creek have a sur-
face permeability with a lower range of from 0.2 to 0.6 in./hr
and that for many soils, the "B" and "C" horizons have substan-
tially lower values. It is recommended that a classification
system be devised for sorting out the major groupings of soils by
infiltration rates through the full profile, with consideration
given to cover conditions, because infiltration can be much
greater with certain cover types.
Interflow (INTER)
As described earlier in the discussion of types of drainage
systems in Honey Creek, the existence and efficiency of tile
underdrains in much of the cropland of the basin will radically
alter the movement of water through the soil mantle as interflow.
The use of one value for interflow (such as 3.0, recommended in
Figure 35, NPS) is not suitable for this basin, and in fact var-
ious drainage classes should probably have different values. A
time dependency should also be included, such as time since
start of rainfall.
The concept of partial area hydrology, as originally pro-
posed by Betson (1964) and developed by others (Dunne, 1969,
1976) may have great applicability in developing better algo-
rithms for interflow. Generally recognized by most research
hydrologists as a more accurate description of runoff mechanisms
in humid regions, this theory proposes that only a small portion
of some watersheds actually produce runoff from a storm, as des-
44
-------
cribed by Dunne(1975).
Although the Koney Creek basin is a humid region, it is also
highly disturbed and is dissimilar from the wooded catchments
Dunne and others have studied. With the extensive underdrain
system, however, it is likely that Honey Creek produces varying
amounts of surface runoff (and interflow) from distinct and
spatially definable portions of the watershed. It is likely that
under all but the most severe storms, the hydrograph produced at
Melmore (and the pollutants carried with it) is generated by run-
off from only a fraction of the land surface in the basin.
The task at hand is to recognize where and how saturated
zones occur within the watershed, because the runoff-producing
zone will vary both during and between storms. In fact, it has
been found that the saturated area will expand during.a storm,
as the water table rises to intersect the surface. The occur-
rence of this condition will be a function of the factors that
distinguish the various drainage classes, plus soil moisture and
precipitation.
Consider the erosional mechanisms that might exist at the
land surface under saturation conditions. The soil particles
disturbed would move with overland flow as soon as saturation
occurred. With continuing precipitation, the saturated zone
would increase and a greater depth of water would exist over much
of the saturated zone, contributing directly to runoff but insu-
lating the land surface from the erosional impact of the rain-
drops. The moving film would, of course, possess the energy to
suspend and transport material, but the border area at the ex-
panding saturated zone wouH undergo substantially greater erosion
mechanisms. Thus the volume of runoff would continue to increase
over the period of continuing precipitation, but the pollutant
concentration of the runoff would increase and then decrease.
This is exactly the chemograph-hydrograph relationship that has
been observed in the Honey Creek basin and many others, (Cahill,
1976) .
POLLUTANT GENERATION PARAMETERS (QUAL)
The QUAL subroutine algorithms utilize several parameters
that reflect the pollutant generating mechanisms of erosion and
transport. It is believed that some of these could be modified
and improved upon for the Honey Creek basin.
Rainfall Kinetic Energy (JREP)
This exponent of the soil splash equation is intended to
approximate the relationship between rainfall intensity and in-
cident energy to the land surface for the production of soil
fines. Based on various investigations, NPS authors concluded
that soil splash is proportional to the square of the rainfall
45
-------
intensity, so the a value of 2.0 would be reasonable for JRER in
equation (9), NFS. Because the major sediment producing storms
in Honey Creek are in the early spring and not the intense summer
thunderstorms, it is felt that the actual kinetic energy associ-
ated with a given storm should be considered in evaluating the
sediment generation mechanism. The estimated annual R for this
part of Ohio is 130 (AR, 1975) but it is obvious from the chemo-
graphs, hydrographs and mass transport estimates developed for
Honey Creek that the erosive force associated with the storms of
late winter and early spring are far more important in soil part-
icle transport. In fact, although storms "B" and "M" are almost
equal in R values and total storm energy, storm "B" produced over
14 tons of suspended solids whereas storm "M" produced a tiny
fraction of that amount, (see Table 15,page 53).
The reason, quite simply, is that not only has the soil par-
ticle been displaced and brought into suspension by the rainfall
impact on bare soil, but the ground surface is sufficiently sat-
urated in wetter portions of the basin to allow immediate trans-
port with the overland film flow (or in microrelief channels).
Thus a storm such as "B", in which 9 hours of precipitation,
ranging from 0.1 to 0.3 in./hr, was followed by 0.5 in./hr inten-
sity during the last hour, resulted in what was the highest con-
centration of suspended solids (and total phosphorus) observed
during the entire year. What part the variation in precipita-
tion intensity during the storm played as compared with the
saturated ground conditions is yet to be determined, but it would
appear that antecedent moisture is a far more significant deter-
minent than kinetic energy of rainfall. Cover appears to be a
secondary factor, since this major runoff producing event oc-
curred prior to spring plowing, with row crop cover averaging
about 50% far more than subsequent periods of precipitation.
Erodability (KRER)
Because this parameter is intended to include consideration
of the intrinsic erodability of a given soil, it is related to
the "K" factor in the U.S.L.E.. The K factor, of course, is
estimated based on texture properties of a soil, as well as
structure, organic matter content, permeability and clay mineral-
ogy. For Honey Creek, the spatial distribution of K factors are
available and can be calculated exactly for any sub-basin or the
basin as a whole. For the various drainage classes, appropriate
K values can also be developed from the LRIS data file. It is
suggested that the KRER term be considered as (A K), to allow
for parameter calibration while maintaining the exact K value.
Soil Fines Transport
This is perhaps the weakest link in the NFS model, with the
development of the JSER exponent and the KSER coefficient re-
quiring substantial improvement. Limited experience in model
46
-------
calibration thus far suggests a narrow range of values for JSER
(1.6 to 2.0), with tremendous variation found in KSER (0.01 to
5.0) .
It is believed that too many variables are being included in
this one basic coefficient, making the application of the NFS
model to large, more complex watersheds something of a "black
box" affair. Within Honey Creek, it is possible to develop a
set of statistics on combinations of slope, length, particle size
and surface roughness (using land use and cover factors), but the
weighing of these ingredients is a task that requires a great
deal of thought. The coefficient KSER represents, in effect, the
potential of a particle which has been brought into suspension to
remain in the fluid and move to a channel where it will be trans-
ported.
Given the applicability of partial area hydrology, however,
with an expanding zone of runoff producing saturated land and
an expanding "erosion zone" at the edge, the KSER term could re-
present a function of the areal increase.
47
-------
SECTION 7
MODEL CALIBRATION
The period of record that exists for Honey Creek at Melmore
includes some 20 months over 2 years with some severe ice inter-
ference during the winter of 1977. The NFS model was calibrated
on 1976 data and verified with the 1977 data. These two years
are very different in terms of hydrologic events and their dis-
tribution. These differences will be apparent in the subsequent
discussion. In spite of the limitations of the two data sets,
the calibration will proceed as outlined.
1976 - RECORD OF HYDROLOGIC EVENTS
The peculiar feature of this year is that after the late
winter-early spring runoff period, (February 10 to March 10)
there was only one runoff event of any consequence throughout the
balance of the entire year. (Table 15). This 30-day period of
runoff comprised the major sediment and nutrient transport period
for the year. 81% of the total phosphorus and 89% of the total
sediment moved out of the Honey Creek basin (Table 16). For this
reason, special emphasis was given to this period. Although it
is characteristic of this region to have the major runoff events
during the early spring with summer storms having little runoff,
this year is an extreme case.
Figure 12 shows a semi-log plot of flow during the period
in question, with the initial runoff event (A) caused by snow
melt on frozen ground. The second event (B) was caused by heavy
rains (1.81 in.) on ground heavily saturated by the preceding
snowmelt, compounded by thawing ground (see Soil Temperature,
Fig, 13) and warming weather. The resulting runoff contained the
highest suspended solids concentration ( 1131 mg/1) and total
phosphorus (1.82 mg/1) of the year. The patterns of suspended
solids and particulate phosphorus are almost identical, with both
peaking slightly before the hydrograph peak (Figures 13 and 14).
The nitrate concentration does not vary dramatically over the
storm hydrographs, with slight dilution effects at the peak run-
off. The nitrate increase, on the recession curve of storm "B"
suggests tile flow input. The highest nitrate (Fig. 15) values
occur later in the year during the drier periods. Ammonia peaks
are observed during the snow melt event with subsequent peaks on
the recession curve of storm "D", possibly as a result of tile
flow dominance. It should be noted that the ammonia values
48
-------
OIHO 'Nlddll
Q
C
rS
•r-i
,-H
d
r- •
CTl (i)
iH M
•* J. J
fl ^;
•H 0!
d. fi'
M 1)
4 J
fl - !
^ "~"-
S-i —i
o o
5 W
i—i
CD rd
4J
d)
o •>
c -P
O -H
ffi a.
•H
4-! O
o a)
i-*<
w ;
03
rH
QJ
DOT
49
-------
rH
<4-t
O
-S
- in
ft
0)
Cn
0)
- o
rH O
0)
4-)
-a
u
-H
-P
M
n
rH
31VHdSOHd
50
-------
i
i
I
i
§
_e
M-l
M-l
O
Pi
>i
K
ffl
c w
•H -"H
rH ,H
d o
•a
0)
01
en
SQHOS
51
-------
1 i ,L 1
i I I
in
1C
O
I
»-$ "
fO
0)
-P
•H
M
-P
•H
IS
QJ
-P
-P
•H
52
-------
10
r-
/T\
O i
H
»*
Z
H
CO
^
<;
PQ
(.v
P5
w
CM
u
;*
2
o
w
2
H
CO
EH
3
H
M
S
O
EH
CO
r£
(-H
O
i~3
S
T3
M
O
O
0)
PH
(U
CH
(0
0
(U
!H
r\
i
rH
0)
5
jj
t/i C
-H
S3
n
ro rH
5- (Tl
ri£< nj
4J
C 0
CD EH
CU
•>
4J 01
CU >i
HI CO
T3
T3
O y-i
•H 0
rH
(11 •
tu •
a. o
z
c/]
3
M
>1
rj
— CQ
>i
CO Si
•a -u
\ 3
C S
•H >i
-^ rH
CM
rH
rH C
4 CNrHCN rH
CN OP-y3CNin t^i^rvjVOP- ^rOOlrH
co P-P-CNCNOS cNcNin^^0 TTincrsin
rH OOoOO OOO°° OOOO
p~rHin os as ^«O3 rom » ^ QS ^r oos os oo ko os
•TTOCNP-T rOkOCOrHr^mfj, P-OSOOSTTO
OSrHrHCN^rH rH rH rH
^J* rH (H
rO rHrHrHrHCN ^J—J—jrHCN iHrHrHCNI
O
4J
Cs
VI3
H
o
C^l
C
-H
CN
cs
rH
ro
II
rH
rH
CO
C
•H
cd
^
rH
TS
AJ
Q
4J
kD
P-
CT.
rH
• T3
in
rH
CD
rH
rQ
(0
EH
Z X
—
0]
Cl;
-U
10
Q
M-4
M-l
0
3
a
c
CO
•H >.
rH IT!
3 Q
r->
E
O D
•U "Z.
'Jl 1
P-
CN as ^r
X, \ v^
CN ro ^0
33 3
l-i IH ^
js £ x:
4J 4J 4J
O O CO OS rH
rH T CN CN rH rO
\ \ \ \ \ ^
CN ro ro ro in in
rH ^T O ro O CN
^3. \O CO C23 -5T in
rH rH
4J
OS «r
rH CN
\ \
kO U3
CN r-
p- p-
rH —1
M -J
kO kD
P- P-
^
\
r^-
3
VH
£
JJ
O
ro
\,
kD
CN
03
rH
^
O
p-
•q'
CN
•v.
p-
3
SH
s:
4J
O ro
rH CN
\ \
r- r-
CN in
OS O
-t CN
j ^
kO O
p~ p~
o
rH
"X.
CO
3
H
.C
-^
CO
\
en
rH
CN
CN
2;
kO
p-
LO
rH
\
CO
CO
CN
CN
O
kd
p-
ro
V,
OS
3
)H
JT
4J
0
ro
\
CO
ro
T
CN
^
kO
r^
OS
CN
\
OS
3
V-l
£
^J
CO
CN
~\
OS
CN
P*
Ck|
O
kO
P-
53
-------
VO
r~-
cr\
rH
EH
PH
0
CO
J2
f^K
PH
EH
CO
co
I!
fc
o
o\°
0\° I rH
CO
rH rH
3 4-1
C O
T3
O
-H
S-l
0) 0
PL) rH
G •
M ^
O
ft (D
w g
r-J >. •
CO
o
•H
4-1
id
M
_p
d
OJ
u
f^ .
0 rH
O "^x
t
d tr>
td g
a) —
g
^-^
r~>
^ o
O rH
i — |
PH X
r-TO
td g
4-> —
O
EH
54
-------
shown are extremely high for a natural stream system (± 2 mg/1),
especially during periods of greater flow. The conductivity is
a mirror image of the log flow, indicating little or no input of
groundwater into this flow, with slightly higher dissolved solids
levels in the snowmelt runoff. The orthophosphate shows peaks
during the snowmelt of event A and on the recession curve of
storm C., but otherwise remains fairly constant over the period.
INPUT DATz\ ADJUSTMENTS
As iri any model calibration study, the necessary input data
are seldom in the exact form desired. There are also data gaps
which must be estimated or ignored. In the Honey Creek basin the
available meteorologic record leaves much to be desired, espe-
cially during 1976. The adjustments made in this model calibra-
tion are described here for future reference, so that the reader
is aware of what limitations may apply to the calibration results.
Snowmelt
No record of solar radiation wichin a reasonable distance of
the _ Honey Creek basin for use in the NPS snowmelt algorithm is
available. Because the snowmelt-ground thaw period is
so critical in terms of runoff and sediment transport, the calcu-
lation of snowmelt was done "by hand," based on the record of
snowfall, snow on ground and air temperature at Tiffin. During
winter months, the total accumulation of snowmelt was calculated,
based on 0.10 inches of precipitation per inch of snowfall as
equivalent moisture value. This total moisture storage was in-
put as precipitation during the period of snowmelt, in proportion
to the daily snow- on-ground reduction. The actual observation
of this moisture as runoff from the basin appeared to have been
delayed until actual ground thaw for both years, but it should
be noted that the stream flow stage record was subject to ice
interference during the initial thaw period, and so the correla-
tion between snowmelt and runoff should be viewed with caution.
It. would be ideal to have three or more well distributed
continuous recording rainfall gages distributed throughout a
basin the size of Honey Creek for NPS model input, data. Unfor-
tunately, no such network yet exists, although one is planned,
For the period of record, three daily tecording stations that
bracket the watershed—Tiffin, Bucyrus arK1 "|jlymouth--have been
compared and indicate that the Tiffin daily gage is representa-
tive for most major storm events, with the exception of several
summer storms. Tiffin also has an hourly record produced, at a
separate station (Water Works) , but this record is not contin-
uous and is rounded off to tenths of an '.nch. This is the only
hourly precipitation record avaiJab e for model ca.libiat.ion itom
February j j <".- through .April 1977. At that time a continuous re-
corciinq • \\r^, \;i ^luorri. was establishes.; -y the Heidelberg CoDJo-^c
-------
River Laboratory at Tiffin. Also, the hourly record at Tiffin
is absent for most of March 1976 and had to be synthesized from
the Fremont hourly record for distribution and the Tiffin daily
for total amount.
THE NFS MODEL "SCALE" PROBLEM
Simply put, the QUAL sub-routine of the NFS model is design-
ed to simulate pollutant washoff from land areas in small catch-
ments, where the subsequent time of pollutant travel in the
channel system to the point of flow measurement, as well as chem-
ical sampling, is quite small as compared to the time of overland
flow and/or interflow. The application of the full model (and
not just the hydrologic or LANDS sub-routine) to large basin
systems requires the development of methods by which this "time
of travel" can be accounted for in analyzing downstream chemo-
graphs and hydrographs, or conversly in using these data for
model calibration and verfication.
Before a discussion of alternatives, let us consider several
pieces of knowledge that will have a bearing on these alterna-
tives. Alsov the thrust of the current work effort should be
acknowledged as the transition of detailed research modeling
methodologies to broader management evaluation of water quality
improvement strategies. This perspective will also have a
bearing on alternatives.
PERTINENT KNOWLEDGE
The relationship of chemograph to hydrograph for pollutants
such as suspended solids and total phosphorus has been observed
on various watersheds (Cahill et al. , 1974; Baker, 1973), with
the same general relationship holding true: the peak of the
chemograph always precedes the hydrograph peak.
The movement of storm runoff in natural channels has been
studied intensively during the past 50 years, under varying stornv
channel, gradient and precipitation conditions, with most inves-
tigators concluding that the flood wave, or hydrograph, moves
through the channel system faster than the fluid velocity.
The theory of "partial area" hydrology leads to the conclus-
ion that the runoff that comprises the hydrograph (and its re-
sulting pollutants) has come from only a portion of the watershed.
The land area producing this surface runoff is a function of soil
moisture, permeability and proximity to channel; it is not de-
fined by the same considerations (texture, slope,) as would nor-
mally be used in defining "critical erosion areas."
There is little question that points one and two hold true
for the Honey Creek basin; point three is only hypothetical at
this time and might be evaluated on an initial basis by comparing
precipitation and runoff for different storm events (Table 17).
56
-------
H
o\°
O rH
K
U
J>H
H
j^
O
ffi
O
en
H
§
g 2
o 3
u o
EH
p4 CO
CM
0 vo
3 CTi
03 rH
1
1-^
ti-4
0
H
EH
v__>^
|5 . ro ^ O O O ro
OdrHrHrHCNrHO
rH -H • • • • • •
pq • — ' O O O O O O
4J
3 *—•
PH • ro CO Ch CTi r~» >~D
-t-'Cr-^ocriroo
3 -H ...•••
O'-'OCNrH^OO
-P -^
^i . o o o o o in
Q^rjJ 0>j , — | \£) ^t ^ f\]
a -H
H^OCNOrHOO
ro
r~~
•
in
m
oo
in
0)
rH
A
(0
EH
rl
O
-p
Ul
« U Q W
FH
57
-------
Assuming uniform precipitation over trie basin the r.'.rsr t ix
storm events for 1976 are compared, (with uase i'lov; ent iaiated,
indicate that during the initial runoff period in February
(storms A, B and. C) virtually all the precipitation ';Ju-it fell in
the watershed ran off shortly thereafter iron all l^pd arearj,
As the year progresses (and saturated or near saturated surface
areas decrease rapidly) the volume of runoff to rainfall decreas-
es remarkably, to a smal? fraction of the basin possibly contri-
buting runoff. By the time of. the late spring and summer storms,.
many of which are equal or greater in magnitude thai, the February
storms, the input of one to two inches of precipitation causes
barely a ripple in the basin. If the theory of diminishing re-
sponse area does not hold, true, then one would conclude that the
basin functions as a giant sponge, soaking up precipitation and
returning it as evapotranspiration over a long period of time.
Although this mechanism cf moisture storage is certainly opera-
tional in the basin, it is likely that something similar to a
"partial area" hydrologic response is primarily responsible for
the seasonal variability in Honey Creek basin hydrology.
INITIAL NFS MODEL CALIBRATION
Although recognising the specific characteristics of hydro-
logic response and runoff characteristics of the Honey Creek
basin, as well as the limitations of the NFS model, it was be-
lieved important to first follow the procedures recoirnaended for
model calibration and evaluate the model in its present form.
This section documents those steps and the results of that cali-
bration.
The 4 month spring-summer period of 1976, (April-July) was
initially selected for hydrologic calibration, because it. repre-
sented a period of substantial precipitation (13.4 iu.) and very
little runoff (0.83 in.), a characteristic of this type of water-
shed that must be understood. Pan evaporation data were also
available for this period, with data collection beginning ah the
Hoytville station (Fig. 3) on April ]„ Remember that; during cal-
endar year 1976, most of the runoff arid sediment transport out of
the basin occurred during February and March, preceding this ini-
tial calibration period. The focus during these initial, runs weis
on moisture accounting, especially the evapotranspiration loss.
This approach, however, proved unworkable, as the fo-Llowing para-
graph describes.
The NFS model trial run results are shown in Table 18 , with
the monthly precipitation and runoff at the Me lino re gage shown
for comparison. Initial parameter selection was based on recom-
mended values from the manual or basin characteristics, "ft was
also assumed during these initial runs that the only land use in
the basin was agriculture (Runs 3-11) . Trie initial runs indicat-
ed that the evaporation input data were expressed .improperly, with
Table 24 (NFS) showing values expressed in hundredth.'? of an i.nohs
58
-------
r-
i — i
^r
rH
ro
CN
rH
rH
rH
o
rH
cn
CO
r^.
m
i
i
i
i
i
>,
C
o
ropland
CJ
i
rH
rrt ^.
OJ 3 5 •
a^+j o c
!Q O rH -H
u
>f
rH
i-j
^
•< &4 — '
rH p| ^->
rd -H '
3 U C
4J 0) H
U W -^
< &H
Month
CO CO
1— 1 r_l l^Q
. CN
ro rH ^
O^ r-4 ^>
O
rH
r-
o
rH
n
rH
rH
O
r-
rH
O
rH
O
CTi
rH
O
s
o
r-f
m
rH
00
H
rH
rH
CO
O
CN
DUT
CO
rH
O
o
ro
""
1
( — ,
in
o
CO
in
o
in
to
o
en
o
in
in
o
rH
O
o
rH
O
*3*
rH
C
CN
ro
0
r-
o
rH
CO
rH
CN
cn
rH
O
rH
1
) J
3
o
o
ro
0)
Cn O ^O rH r- rH
^3* «3* ro CN O CN
O O O O O O
01 CT1 0 rH r- rH
«3* ro ro CN O CM
0 O O O O 0
co r- in rr ^r ^r
ro CN CN rH O O
O 0 O O O O
CN -sr co ro co
o cn cn '*o CN
rH O O O O
o
ID
o
ro
CN
o
cn
o
o
in
rH
O
CN
O
O
rH
rH
CN
rH
in
in
~
CN
CN
1
UD
r^ 1 1 1 1 1 1
cn
rN
^ 1 1 1 1 1 1
CO
^ 1 1 1 1 1 1
1 1 1 1 1 1
1 1 1 1 1
in
•"! i i i i i i
o
°i 1 1 1 1 1 1
0
CO
o
rH 1 1 1 1 1 1
in
° 1 1 1 1 1 1
CM
CN
"1 1 1 1 1 1 1
r-~
" i i I i I I
in
1-1 — 1 rH
^ • I 1 1 • 1
(^ ro o
,— m
~Z LT) O
M. .1)1-1
r^ o o
•^ CN
O ^ ON O rH O
CO O O rH O rH
i
CN! CO Lfl -rr •*? CN
H 0 0 o 0 Og
o o o o o o
o o o o r~- ^
vo cri ro o T f"
'-O CN ro o-i O rH
CO
m
r-
CN
n
o eg
ro sx
I
>, p jj , . • 1j y Q ^
^ < '-/5 O 2 a t-' ^. i r2
in
I I ! I I 1 1 I I • in I
CN rH
in
11)11 1 1 1 1 rH >
O CN
1 1 1 • 1 1 1 1 1 1 rH 1
rH CO
1 1 1 1 1 1 1 1 1 1 CN 1
rH
I 1 1 ! 1 1 1 1 1 1*1
*T
1 1 1 1 1 CN 1 1 1 1 III
rH
in
1 1 1 1 1 -III III
o
o
1 1 1 I 1 1 1 1 1 1 • 1
1 1 1 1 1 1 1 1 1 III
1 1 1 1 1 III- III
0
^r m
I i I 1 I -ill II-
o o
in CN
1 1 1 • 1 -III III
0 O
o o
I 1 1 • 1 1 1 1 1 1 1 •
-H O
LTj CN CN
inoomoo-S'OOr-i oom
OCO"*rOOCOOOOO rHCOO
^SS«J?mS3Sg
H ^ ^ £j
59
-------
m
VO
£ S
H
*.
i-l S
Q
i s
w
ION OF NONPOINT SOU
!4 » .26 27 28
i—i "
ffl
H
. "1 CM
<
u
r!
:AL HYDROLOGIC
17 18 20 :
i— i
EH
H
5 £ ,
Ce ss-< a
.H »»4 Q -«4
O w U 9 O
OIK 5 *) 0)
~t Q 0 U
^ >• a *o.
t
00
rH
(I)
rH r
•9 =
nj o
pj Z
men(sr»flh«Dr*rOf<4ma9
OiH^innrsiMiN^o^
•vrMoooooeoo©
ehmr^or^ovvor^^'^r
(NiH^tnmmmroctiiNeo
mcsooooooooo
n(*4flB
miNoooopoooo
CO
r- i-*
**i rj
o
^- r>*
0 *H
VD m
•V N
ass
^ fN
(N{Ninvpm^3'(Noo
v r^i
cnrnMOO«>f>jr-«>r-m
(H^rmr^vcm *4 r* eft r-
vo M rM r- in m fM
m r4 d d d d d
c rnupfMfiOepiNSBtQ*'1"''''
^ ."^ .-4H-jo»p-o>q
— '•H*oeeooooo
i * ^";SS88S85R
>*1'*'NvWrifMrifMd-4
r "^ 2 •
{IlilftHfi
2 ° mo
CD « 0 «
S g ^
pa
CM ^
Ch O
in
vo .
CO »H
S ^ I i i i ° •» i i i ^ , S o , , ^ c o in
tn
i i i i i t t i i i • i i i i 4 r lit
o
1 1 t 1 1 1 1 1 1 1 1 1 1 1 1 t 1 ' ' o
m
i •** i i i * i i i \ i i i i i i i ill
i *j i i i i i i i i i i ) i i t i i t t
S 23
s ° °
o or* co o
d iilli^^llllliill^ m
•* o m o
_, m <*
I i i i I I 1 I 1 1 I 1 1 1 I I 1 I ill
So oo
d -II ' ' MM '
•-< rH
w fM in in fM fM ^
ow
-------
but the model reading thousandths. Thus in Run 6, the evapora-
tion data weremultiplied by 10. It readily became apparent that
it would be necessary to allow for much greater evapotranspiration
in the basin than the model was providing, and so in Run 7 the
parameter K3 was increased to 0.5. Next, in Run 8, the raw pan
evaporation data were artifically multiplied by 1.5, and in Runs
9 and 10 the parameter INFIL was first increased with decreased
LZS, and then reduced. For Run 11, the evaporation data were used
in their correct form (pan evap. x 0.7) and the LZSN was increased
to facilitate greater uptake by root systems. The total runoff
for the period, as well as the monthly runoff figures, was not
within an acceptable range, but virtually all of it occurred as
base flow. The total runoff for the 4 months (1.15 in.) was
still higher than actual runoff (0.83 in.), but it was felt that
the full calendar year, with all land use categories included,
should be run. This was done beginning with Run 12, and it be-
came apparent that the heavy runoff experienced in February and
March could not be effectively simulated unless the initial stor-
age in both the lower and upper zones was significantly increased
By Run 16, the annual runoff was correct, but the runoff in late
winter period was still low and that of the summer too high.
The precipitation data, especially snowmelt, were reexamined
and compared with the daily records for the three stations sur-
rounding the watershed. Some improvement was made for February
and August, but it was thought that the only way to produce the
combined snowmelt, thaw and precipitation of February was to
"load up" the initial UZS. This produced Run 18, with February
increased to a satisfactory level but the balance of the year
too high.
Because 89% of the suspended solids for the year are flush-
ed from the basin during the 30 day period of February 10 to
March 10, it was decided to focus on refining this period, com-
prised of the four runoff events described previously. The ma.ss
transport occurring for these events is summarized in Table 19 ,
with storm "A" the "byhand" simulated snowmelt runoff. Refine-
ments were made to the shape of the recession curves (IRC and KK
24), and the overland flow component increased by significantly
decreasing the Manning friction factors, a reasonable adjustment
for February and March, but not the balance of the year. By Run
24, it was believed that the initial hydrologic calibration was
acceptable for the spring runoff period, although the annual
total (11.2 inches) was still too high. The QUAL sub-routine
was run using the LANDS inputs for Run 24, with the three storms
in February and one in March used for mass transport comparisons.
All four storms were low in terms of sediment production,
and so in Run 25 the overland flow component was increased by
reducing the Manning "N" for pervious surfaces to 0.5 (frozen
ground, untilled). This resulted in a very minor increase in
suspended solids, so Run 251 increased KSER from 1 to 3, pro-
61
-------
c
5
to'
3'
™ rH .JO
o £ S
EH
CO
c
VD -H
CTi '^j >
^ ^ jj|
g ££co
CO
H a;s
EH Jjrj
§ fc<
H
EH
CO
w c
• H
HI ">
5 5
0 ^
CO ||
S "!
EH ^
CO
< 31
'^ c^ ^
OD 01
t~ ON
^ CO
^r
o
rH
O
X rsi
0 \
o
ro ^
rH -H
C C
f*~i ot
r — *j
C (N
C 0
CV (N
cr. ON
(N ON
O r-
CN CN)
fH ^
m
i — i
VD
i-l
O
O
O
in
rsj
C
ON
0
rH
c
kD
ro
«T
O
"C"
(M
f^
o
(N
O
ON
o
o
CO
01
o.
tn
o
oo
Vi) tN
O O
ro
o
r-
c
o
o
ON
CTi
0)
rH
XJ
EH
tJ
a
O
g
o
o
B
CO
5|
62
-------
ducing good calibration for storm "3", the dominant event, but
poor calibration for the other three storms. This run simulated
72% of the actual suspended solids runoff for the period, with
the simulated flow for Ktorp "D" still too high.
In Runs 26x 21 and 28, chjngeo were made in LANDS parameters
{NN,UiiSN) to increase the overland flow component. Storm "B"
increased sediment delivery greatly, but the other storms showed
little sensitivity. Runs 2:) through 33 reflect final adjustments
to both LANDS cm 3 QU?\.L parameters simutaneously, in order to
balance both the runoff and sediment production.
The final calibration summary is shown in Table 20, and
gives very good overall mass transport simulation. However,
while the total runoff arid pollutant mass transport simulations
were being refined in Runs 24 through 33, it was apparent from
CALCOMP plots of the output that the timing of both runoff and
sediments was far too rapid. Figure 16 shows the hydrograph
simulation without routing for February-March, 1976, with the
extreme peaks representing the actual NFS model output for Run 33
at Melmore, done as one single basin. It is obvious that the
time delay of both fluid and particles in moving through the
watershed must be considered in order to replicate the chemo-
graphs and hydrographs observed at Melmore.
63
-------
Table 20. NFS MODEL CALIBRATION SUMMARY - 1976
1976
Actual Data
1976 (Run 3311
Simulated Data
Flow Suspended Solids
(inches) (tons)
JAN .
FEB.
MAR .
APRIL
MAY
JUNE
JULY
AUG.
SEPT.
OCT ,
NOV .
DEC.
_
4.
1.
0.
0.
0.
0.
0.
0.
0.
0.
0.
39
69
29
18
18
12
08
05
04
04
02
—
16,965.
3,115.
55.
155.
194.
166.
62.
30.
4.
2.
1.
Flow
(inches)
4.
2.
0.
0.
0.
0.
0.
0.
0.
0.
0.
05
19
42
57
39
28
23
20
12
05
18
—
16, -806 .
3,009
428
499
106
58.
11.
11.
13.
8.
12.
3
8
6
6
2
2
TOTAL
7.38
20,822
8.68
20,963
64
-------
350
300
250
200.
X
w
4-)
150
c
rH
fa
100 .
50
s-simulated
a-actual
0.0
25 35 45 55 6<;
Figure 16. Hydrograph simulation without routing, February
and March, 1976 (done as one basin at Melmore).
65
-------
SECTION 8
HYDROLOGIC ROUTING SUB-ROQ'L'li-JE
There is little question that the successful application of
any hydrologic and chemical transport model to a watershed
greater than a few square miles must include a sub-routine for
routing the various sub-basin hydrcgraphs through the network of
channels to the measured station. For the purposes of this dis-
cussion, it is important to distinguish between the movement of
fluid waves or hydrographs, and the movement of actual parcels
of water and the suspended sediment, particles associated with
them. First, the NFS model simulation of flow will be routed as
a fluid wave in an effort to replicate the observed hydrographs
at the Melmore gage. Then, the same routing-routine will be
applied to the NFS model simulation of suspended solids and sedi-
ment-associated phosphate, but. i»t particle velocity. Conserva-
tion of mass will be maintained, with the ceil j braced volumes
shown in Table 20 redistributed in time,
Using storm D (Runoff fioni March, ^-11) of 1976, as an ex-
ample, development of the routing is presented as a three step
process:
1. Time of travel (TT) from the subbasin mouLhs.
2. Time of travel -f sub-basin time of concentration (Tc)
3. Time of travel -(- intrabasin ^outing (IBR),
Only the (TT +IBR) step is applied to the entire February -
March 1976 period and the February - March - April 1977 period,
All routing results are presented as hydroaraphs or chemographs
and are compared directly with actual data for the p^;. i od rout-
ed.
Almost all existing techniques for hydrologic routing have
evolved through the need to predict the maximum stage and fining
of flood conditions, a problem that neeessarijy is concerned
with the movement of the maxiiau'ti f 1 aid wave, or flood crest
through the main channel and the oveibank flooding which results,
The techniques used in routing analysis are generally considered
to be either hydraulic, based on the equations of unsteady non-
uniform flow in open channel and continuity (Chow, 1959) or
hydrologic, based on the movement of observed hydrcy rr'phs •'-r^'ouch
the channel system. The method utilise;1 ir- selectee ';ared •"•:•',
the data and computational resources ^tv-ji.l -:'bJ e a- VP: J r-7 ':c'-. 6*--
66
-------
gree of detail required.
PRIOR HYDRAULIC ROUTING IN HONEY CREEK
A 1967 study by SCS in the upper Honey Creek involved the
calibration of the TR-20 runoff model (Soil cover complex method)
and routing of flood waves by the hydraulic "method of character-
istics" (Massau, 1900; Lin, 1952) focused on very localized
flood conditions, because parts of the main channel had been
cleared and diked to protect contiguous farm fields. The primary
concern was the maximum flood stage reached and its frequency of
occurrence, so that structural modifications could be made to the
channel and dike system. As part of this study, a great deal of
detailed information was gathered on channel cross-sections,
bridge structures, historical high water marks and other topo-
graphic features, as well as channel roughness or friction
factors (Manning's "n"). Although the TR-20 study for Honey-
Creek never entered its second phase, some routing analysis was
made for selected storms occurring in the upper watershed, with
time of travel estimates to the juncture with Aichholtz Ditch-
No actual rated stations were developed, but theoretical dis-
charge velocity curves were calculated for selected cross-sec-
tions based on Manning's equation with estimated values of "n"
and channel slope "S". The cross-sections for which these theo-
retical discharge velocity tables were estimated are shown in
Fig. 17.
TIME OF TRAVEL FROM HONEY CREEK SUBBASIN TO MELMORE
The problem faced in the present study is to utilize the
TR-20 study results, other available information and the NFS
model output to simulate the actual hydrographs occurring at the
Melmore gage during 1976 and 1977. Because no other detailed
hydrographs were measured in the basin above Melmore during any
of the recorded events, hydrologic methodologies for routing are
of no value. Also, because the localized flooding is of no
immediate concern in this analysis, a detailed procedure such as
the one TR-20 model uses would be a case of overkill, considering
all of the simplifing assumptions made in the NFS model itself
and the input parameters utilized (uniform precipitation, Tiffin
hourly records, infiltration rates, etc). After several efforts
at more complex procedures, a simplistic routing methology was
applied based on the time of travel (TT) of flood waves from
individual sub-basins moving through the channel systems at the
kinetic wave velocity (1.5 x fluid velocity).
Routing models that rely on kinematic wave assumptions have
been successfully developed on large watersheds (Quick and Pipes,
1975), where a series of actual hydrograph records exist in a
river network. In Honey Creek, of course, this is not the case
at present, and so a complete hydrograph routing did not appear
warranted. Rather, it was decided to move the flow in each
67
-------
•H
-P
3
o
fi
•H
0)
to
CO
fi
o
•H
-P
u
CD
w
i
to
CO
0
M
U
(U
C
u
0)
M
3
tn
-H
68
-------
simulation interval of the subbasin hydrograph at the velocity
occurring at some fixed fraction of the simulated flow for that
simulation interval. Ideally, the velocity occurring at the
actual flow should be used but because actual flow is only avail-
able at Melmore during the calibration period, some other flow
source must be selected to determine velocity. Use of simulated
flow (Qsim) to determine velocity (V) would lead to an underesti-
mation of travel time peak. In order to select a better stand-in
for actual flow, the simulated peak flows (Run 33) for the Febru-
ary and March 1976 storms were compared to the actual peak flows,
excluding the snow melt event.
For these settings of the model parameters, the ratio of
actual to simulated maximum flow (Qmax.act./Qmax.sim.) clusters
around 0.25. Therefore, to determine the velocity of flow dur-
ing a simulation interval, Qsim is multiplied by the ratio and
the resulting adjusted flow (adjusted Qsim) is converted to velo-
city in the QV table for that stream reach. The Qmax values
were selected for comparison since they are quasi-independent of
time and thus easily identified. Such a procedure to determine
Q for conversion to V and time of travel is a good approximation
only for peak simulation value. At low flows, use of the correc-
tion ratio will seriously underestimate actual flow velocity.
However, there are two reasons why this shortcoming should not
seriously affect the routing results. First, the velocity is
fairly insensitive to flow in the lower flow ranges. Second,
the relative amount of water possibly misrouted by this error is
minor compared to the total flow of the storm peaks.
The movement of full hydrographs by kinematic wave in full
models (such as the UBC model) assumes linearity in the flow-
velocity relationship (Q-V), which is not true in Honey Creek
channels. In fact, one of the key assumptions made in this pre-
sent routing procedure is that during overbank flooding condi-
tions, the flood wave celerity is a function of main channel
velocity rather than the average velocity of the full section.
Because only one flow is routed in this procedure, the actual QV
table for the identified segment (Fig. 18) is used, avoiding the
assumptions of linearity. It is also assumed that the NFS model
is correctly calibrated if the simulated total flow for a given
storm period is equal to the actual total flow, and only the tim-
ing of water parcels is incorrect.
The steps of the routing procedure that consider time of
travel from subbasin mouths to Melmore are summarized:
1. Select storm period to be routed.
2. Run unit area (1 sq. mi.) NFS simulation for all sub-
basins (Fig. 18).
3. Identify stream segments through which each subbasin
hydrograph must be routed.
4. Identify QV curve that applies to each stream segment.
5. Sum the drainage area above each segment.
69
-------
Cn
C
•H
4-)
3
o
cfi
C
-H
CO
rtf
10
-o
<1>
O)
0)
I/I
.c
*/>
O
_Q
-O
00
V) 10
0) en
E -^
9 S
"> 6
c Cn
C f\\
n **
§ tn
.fc
«« rH
n\
P
1
K
i
O
M/
c
c
(0
X!
u
CO
•
i °°
I ""*
-------
6. Multiply unit Qsim by adjustment ratio and by drainage
area above each segment to get adjusted Qsim.
7. Calculate V for each segment using adjusted Qsim and
the appropriate Q-V curve.
8. Calculate time of travel (TT) in hours for each segment
(TT - 0.67L(ft)/V (3600) ) .
9. Sum times of travel for all streams segments through
which flow from each subbasin is routed.
10. Multiply unit area simulated hydrograph of each sub-
basin by its area.
11. Integrate subbasin simulated hydrographs at Melmore,
each displaced by time of travel determined in step 9.
Preparation of the Calibrated NFS Model Output for Routing—
As documented above, the NFS model was calibrated for the
area of Honey Creek watershed above Melmore with NLAND (land
uses) set at five: row crops, small grains, woodlot, water/wet-
land, and urban. The opportunity to route the results from the
calibrated model, however, allowed a fuller use of the available
land use data. It was desirable, though, to reduce the number of
individual NFS model simulations required for the 14 subbasins to
be routed. In order to achieve this, the NFS model was run five
times, once each for the five "pure" land uses types listed above
with NLAND = 1 and ARFRAC =1.0. In each case, a unit area of 1
square mile was applied (AREA - 640). The simulated flow from
each of the 14 subbasins was then reconstructed by applying the
percentage of each land use in a subbasin to the simulated flow
from the unit area of the corresponding land use type and summing
over all land use types present in the subbasin.
The result of the above calculation represents the simulated
flow from 1 square mile of the subbasin. This unit area flow was
then multiplied by the area of the subbasin to obtain its full
simulated flow. This calculation was performed and the result-
ing flow was routed for every 15 minute simulation interval
occurring in the period selected for routing.
In Table 21 are shown the percentages of the five land uses
in each of the 14 subbasins that were used to obtain the input
flows to the routing routine.
RESULTS OF ROUTING USING SUBBASIN TIME OF TRAVEL (TT)
A schematic representation of the renting achieved by appli-
cation of this procedure is shown in Fiq. 19 for five consecu-
tive simulation intervals, all with the same time of travel of 10
hours. Routing of storm D (March 2-11, 1976)
71
-------
M
W
&
U
W
S
o
in
£j
H
CO
£3 O
C
"3 '\)—I^-^-O-H*--?T)—itJ-M.xf^--"!
SH '-> -> O (-> O '^ -> O ~i o O r> b rS
i
i
^x d ^ **— *-* ^o <\! c** c^ -0 "^ * i^;' r\ -^
S-t(t3 '^> O —-* --* f'^^^J-H-H H^-^"^vJ"
(DrH OOOO(MJOOO">0-3i'~lO
i i ] i • • * •
td o>
-P
O_ _ i _ i
t" f*^ *r} "^ +• r*"- N«* L/> Q\ rjN "J : "~> ^\J ',^\
I ^» '
^ ' ^OCy^r^O 0!f^J***C^*00'N40NlaOl-nOs
T3 '•"> O O rH O !-H rH O <~< -H O I—4 f-4 O
o .....: . .
O i i
I i
I
"L, w ^ -, • cxj ^U -a a* -ti^i ^ u->
r-t.^J f>--£>!>-ocr»;'HNOOOor~'''TO(M
nj (« f\lc-rt"n^-f\4rorMf^ror
-------
5IMUIAHOH INTERIM.
me: HOU«
TT,
Figure 19. Schematic representation of time of travel
routing. A simplified hydrograph occurring in 5 sim-
ulation intervals is shown by time of travel. In this
example TT is constant.
by the above 11 step procedure is shown in Fig. 20. This routed
result has reduced the sudden intense response and the prematuri-
ty of the unrouted simulation, and the time of travel displace-
ment has significantly improved the fit to the actual hydrograph.
However, the simulated hydrograph is highly fluctuating and
still precedes the actual.
The times of travel used in this preliminary routing exam-
ple were calculated for each of the stream segments shown in Fig.
18. Interpolation and extrapolation algorithms were used to
determine V from possible intermediate and out-of-range adjust-
ed Qsim values from the respective Q- V curves. The full section
velocities during overbank flooding conditions were not used,
because it is assumed that both the fluid wave motion and the
suspended solids velocity are best represented by the main
channel velocity. It is recognized that this is not strictly
the case, but a reasonable approximation that will lead to a
slight overestimation of waves and particle velocity and a sub-
sequent underestimation of travel time during high flow periods.
Time of travel values developed in this routing example are used
in the next example with the addition of intrabasin time of con-
centration.
73
-------
74
-------
TIME OF CONCENTRATION (Tc) WITHIN SUBBASINS OF HONEY CREEK
PLUS TIME OF TRAVEL
Consideration was given to various methods for estimating
the time required for each of the simulated hydrographs to actu-
ally develop in the 14 sub-basins, ranging in size from 3.38 sq.
mi. (sub-basin 4) to 20.49 sq. mi. (sub-basin 8). Evaluation of
the NFS model LANDS algorithms indicated that adjustment of the
recession parameters (IRC, KK24) and INTER were inadequate to
simulate the time delay of interflow experienced in a tile under-
drained basin such as Honey Creek. Various schemes were consid-
ered which would utilize the Honey Creek cellular data base
LRIS) and drainage network geometry, but this approach was defer-
red for later research.
Simplistic methods that estimate the time of concentration
as a function of basin geometry (length of basin, slope or eleva-
tion differential) and surface condition (surface roughness co-
efficient) were compared with the Tc estimated in the SCS TR-20
model calibration in 1968 for Upper Honey Creek. These included
two forms of the Kirpich equation (1940) and the general equa-
tion for overland flow (Chow,1959). Although none of these can
adequately describe the complex elements of a hydrograph in a
sub-basin such as Broken Knife Creek or Achholtz Ditch, they
approximate the delay in runoff within sub-basins. The times of
concentration calculated for the TR-20 model application to the
upper Honey Creek subbasins were selected for use in the routing
routine. Values of Tc for subbasins 10-14 were estimated based
on similarity with the subbasins for which TR-20 times were
available. The time displacement of the simulated flow by apply-
ing both TT and Tc is shown schematically in Fig. 21.
SIMULATION
TIME: HOURS
Figure 21. A schematic representation of time of
travel plus time of concentration routing for 5
consecutive simulation intervals.
75
-------
oo O
1JD U
MH >i
O rH
•P
O C
E CU
-H T3
-P C
1}
CO
£
^ O
O -rH
-P -P
CM '
G d
•H o;
co u
rH
(N
CN
0)
SH
os;
OST
(TOI x sjo)
DOT
76
-------
Application of the time of travel procedure as outlined in the
previous section plus the estimated time of concentration to the
NFS simulation of storm D produced the hydrograph shown in Fig.
22. It can be seen that the routed hydrograph is shifted slight-
ly ahead of the actual hydrograph and that it still fluctuates
significantly.
INTRABASIN ROUTING PLUS TIME OF TRAVEL
As mentioned above, the TT + Tc routing procedure yielded
a fluctuating hydrograph appearing slightly earlier than the
actual hydrograph. Routing the NFS model output in this way
shifts the entire volume of water in a simulation interval from
one simulation interval to a single future one. The simulated
hydrographs from the 14 subbasins arrive at Melmore over a num-
ber of simulation intervals determined by the TT + Tc-displace-
ment. The variable time displacement tends to spread the simu-
lated hydrograph over a period as long as that of the real hydro-
graph. The spikey nature of the unrouted simulated hydrograph,
however, is sustained in the routed hydrograph as a result of
assigning all the volume of water in a simulation interval to
just one other in the routed hydrograph (see Fig. 22). It was
desirable to improve the routing routine in a way that would re-
duce the spikey nature of the hydrograph, would bring better
agreement between the routed and real hydrographs, and would
conform more closely to reality than the TT plus Tc concept.
To these ends, a simplified device was applied that first dis-
places the flow from each simulation interval by its TT (as above)
then divides that flow equally among all the simulation inter-
vals in Tc for that subbasin. Routing flow from one simulation
interval by such intrabasin routing (IBR) is shown schematically
in Fig. 23.
SIMULATION INTERVAL r ITTT^T 4 r^ fc 17 I 6 [ 91101 u 11111311*1 is i ibl
TIME. HOURS
TTi
Figure 23. Schematic representation of time of travel
plus intrabasin routing. Discharge from one simulation
interval is displaced TT , then divided equally among
all simulation intervals in Tc.
11
-------
Physically interpreted, this routing modification means that flow
from one simulation interval appears as stream flow in all parts
of the subbasin simultaneously but that these parts take differ-
ent lengths of time to deliver their flow to the subbasin mouth,
the longest taking the interval Tc to appear. The Tc routing
assumes that the flow from each simulation interval originated
from the furthest point in the subbasin and none from interven-
ing points. The result obtained when intrabasin routing is ap-
plied to a five simulation interval hydrograph is shown schemati-
cally in Fig. 24. As can be seen by comparison of Fig. 24 with
21, the routed hydrograph appears earlier and is considerably re-
duced in amplitude by applying intrabasin routing.
SIMULATION
TIME: HOURS
91 IOI 111121IJII41
TTT
Figure 24. Schematic representation of time of travel
plus intrabasin routing II. Discharge is from 5 simu-
lation intervals which are each displaced TT, then
divided equally among all simulation intervals. The
various textures indicate disposition of discharge
from each simulation interval.
78
-------
When IBR and TT routing proceduies are applied to storm D,
the result shown in Fig. 25 in obtained. The spikinens of the
TT (Fig. 20) and TT + Tc (Fig. 22) hydrographs has been elj.rair.it
ed. The TT + 1BR hydrograph, however,, appears approximately v_i?-
day earlier than the actual hydrograph. The cause of this dis-
placement could be anyone or combination of the following
1. Estimation of TT is too low.
2. Estimation of Tc is too low.
3. A major component of peak flow, such as contribution
from tile drains, is not modeled.
4. Error in precipitation record.
TT-IBR Routing for storms in early__.19J76_
The TT-IBR routing procedure as illustrated for storm D,
was next applied to the entire February-March 1976 period
(Julian day 32-60) . The results of the routing e.re shown in
Fig. 26. It is seen that the calibrated NFS mode1 input, witn
TT-IBR routing, produces storm hydrographs that appear earlier
with higher maximum flows than actually occurs. The earliest
simulated peak in Fig. 26, is actually not in response to re- ir -
fall but is caused by the high initial setting of UZS and L?:S,
Note also that between Julian days 70 and 90, the moJeL as c.-.!_'
brated shows three hydrologic responses to rainfall that ve-re
not actually recorded at Melmore, This could aiso b<"- • '<'• -v^-ti .-:•.•
deficiencies in the rainfall record,
CHEMOGRAPH ROUTING
Attention next turned to the question of routing of sus-
pended sediments f a different problem from the fluid ^'ave iwye-
ment simulated in the preceding sections. As discussed previ-
ously, anemographs almost invariably precede hydrograph peaks,
and frequently diminish in magnitude vith subsequent increases
in flow. This phenomena is well illustrated by i'176 storms, ar
shown in the expanded scale Fig, ?*7. The D hylrograph has -we
distinct peaks, obviously one runcff occurred shortly after t hi-
first had peaked. The chortle-graph, however, shows three distinct
peaks and possibly a fourth, all occurring out of Fyirh.rr,ni ra-
tion with the hydrograph, except for the second. The hour],'/
rainfall record is shown in pig. 1 <, with three major awd on-.-
minor bursts, possible corresponding wit'i the chemograph peak;.-,,
The NFS model in its present form produces chemograph re-
sponse preceding the hydr-,^grc,.ph peak by exhausting the fines
pool with the first flush of overland runoff simulated. If the
available pool of sedinevit is quite largo the simulated overland
flow input carries with it the suspended sediment in direct pro-
portion, producing a chemical peak identical with the flow peak,
This limitation is identifiable in prior NFS mode] calibration
effects,, even on the smallest of watersheds, such as P2 water-
79
-------
(TOT x sjo)
80
-------
i «
'1
01 x sjo)
81
-------
tn
'w
o
0) i
C Cr
CJ E
'£>
i
.r.
CL
13
CD
C
0)
M
Q.
o
' \i
r-j
LTi
CXI
o
o
s
82
-------
shed, Watkinville, GA., and Third Fork Creek, Durham, NC (EPA,
1977). This precedence of hydrograph by chemograph has been
observed almost universally in all such studies of stream trans-
port (Cahill, 1974; Baker, 1973) . The problem faced in this cal-
ibration of the NFS model to a large agricultural watershed is
that the fines pool is seldom exhausted, unlike an impervious
basin. Variable response areas can be approximated with the
ARM model algorithms which divide a watershed into zones which
vary the cumulative frequency function, but that flexibility is
not available in NFS.
Given the limitation in the NFS model, the chemograph cal-
ibration will consider the same two time displacements of indivi-
ual chemographs, TT + IBR, as were applied to sub-basin hydro-
graphs, with one exception; the fluid velocity in channel reaches
will be used rather than the wave velocity. The result of this
routed chemograph are shown in Fig. 28 for storm D. Because the
hydrographs were out of synchronization, the chemograph, routed
at the slower velocities actually shows a better correlation.
The full February and March chemographs simulation is shown in
Fig. 29, with the simulated runoff tending to occur in bursts
with extreme peaks, rather than the smoother curves which actual-
ly occur. The overall mass transport simulation for the routed
model is good for the various storms in terms of total load, but
the long, drawnout transport cannot be simulated, at present,
with the NFS model.
One major shortcoming of the routing scheme used is that
since this NFS program lacks algorithms describing in-channel
flow, several very important phenomena are neglected, including
channel storage, sediment scour or deposition and dispersion.
The use of such algorithms, of course, will require very exten-
sive and detailed data on channel geometry, bed condition and
sediment particle size distribution which are seldom available
except for small catchment studies.
83
-------
•H
C/3
GOT
rd
tn
E
•H
rH
o
T3
OJ
w
VD
CTi
•H
1-3 s
S |
O
4J
i
0)
4d
U
Q)
00
CN
CD
5-1
3
tn
•H
( OT x
papuadsns
84
-------
papuadsns
85
-------
SECTION 9
MODEL VERIFICATION
1977-RECORD OF HYDROLOGIC EVENTS
As previously mentioned, calendar year 1977 had a very dif-
ferent distribution of hydrologic event response and chemical
runoff from 1976, although the rainfall distribution was not
all that different with the exception of a very wet March. The
late February thaw period still produced the highest runoff valu-
es for the first nine months of the year, but at a peak less than
a third of the 1976 value. Corresponding concentrations of sus-
pended solids and total particulate phosphate were also appreci-
ably lower (Table 22). Storm "B" was particularly curious, pro-
ducing a runoff of 1.3 inches with a recorded precipitation of
0.05 to 0.10 inches, obviously a situation where ground thaw con-
tributed significantly to runoff. The actual snowmelt had oc-
curred a week earlier, but apparently been stored in the upper
soil layer and not released until ground thaw, shown in Fig. 30.
The pollutant concentrations were quite high, suggesting a signi-
ficant buildup of sediment from the winter period. March was a
particularly wet month with 4^ inches of rainfall, but runoff
volume and pollutant concentrations were lesser in magnitude
than in 1976. This heavier rainfall in the early spring allowed
storms through April and May to produce significant runoff,
whereas in 1976 this was not the case.
The runoff period from February 22 through March warrants
detailed consideration. As shown in Fig. 30, the 12 inches of
snow on the ground as of February 8 (equal to 1.61 inches rain)
melted to less than 3 inches by February 12, with the ground
thawing just enough to hold this melt but not sufficiently to
allow percolation to the tiles. With temperatures dropping well
below freezing from February 15 to 21, the major portion of this
moisture remained trapped, at or near the soil surface. The
warming period which began on February 22 freed this soil ice
and began heavy runoff, much of it being overland flow. The
small precipitation of the 24th (0.1 in.) occurred on near sat-
urated conditions, thus the dramatic suspended solids peak
following rainfall by about seven hours. The major rainfall for
the period occurred late on February 26 and early on February 27
(0.9 inches at Tiffin, only 0.3 inches at Plymouth and Bucyrus),
but produced only a slight peak on the recession hydrograph and
almost imperceptable chemical increases.
86
-------
(0,1} ' 'dl
(01
iT
vo
o o
o co
J (d s^ '•
(0
(U
s
<
\
1
\ -
\
\ \
\
1
1
1
1
k
1
m
CN
\
(N
CN
CN
\
\
CN
i-<
CN
\
CN
1
0
CN
\
^
T7~f
O
-H
0)
U-l
O
c
3
S-l
Q)
4-1
C
•H
^
QJ
4-1
O
tn
•H
87
-------
f^
CFi
, — I
^
t-pf
^*
H
cn
rtj
CQ
H
("?
U
>H
M
>^»
|C— 1
O
w
«*
w
EH
jxj
H
W
S
O
EH
U)
O
I'D
g
CN
CN
d)
rH
nj
EH
s— v
r*~l
V-i
O
U
O
cu
Cr>
O
CD
O
e
rH
0
•fi
to c
•H^
k-i rs c
O K-H
rrj rHu^,
S rrj-H
•UEH
C 0 —
CD EH
s
-o en
CD >i
M to
•O
O HH
CD •
CM O
z
o
^ CO
rrj £
tf -P
•X 3
• o
c e
•H ^
CM
rH C
14H H
C <*H
'H -H
(0 EH
K
CD
4J
f8
Q
,
X »—
10 rH
rri
co E
t/3 "^
.
X ^-,
(0 r-l
S \
&-
X "m
to >w
^.
O
in
A!
0
d >,
Z B3
-~
CO
01
4J
Q
VM
O
D
K
C
rd
•rH rV)
rH CO
3 D
O O
•U 2
CO
' — 1 -H
rH 4J O O
S S
0 t>
S rH
' £ r-i
' X
5 2 fN
•H C
rH
VO
rH
II
in ro in o
fN
3
^
•^.
,M in oo
UH ro o
0
in
o
ro in
fNl O
o
in CN
X, fNl
fN !
1 rH
ro fN
X, fNl
CN
O O
CN CT\
«T
co ro
VP O>
ro rr
o
rH rH
rH fN
1 1
ro CN
rH fN
CN fNl
O ro
T in
< G2
O O
*T VD
C^ CO
0 0
rH CO
TP fNl
0 0
T rH
O VO
rH 0
fN
x, in
CN X,
1 ro
•w 1
fNl ^
fN ro
in rH
o "«r
rH
vo ro
ro fN
in in
tNi *r
f-H rH
ro cr.
O O
1 1
CN O
CN ro
in ro
in VD
U Q
OOOOCOCNC^ r- .^_
o ro tN c o AT
OCNCOrHOo*1*
CS.1
roOTro^vDinfNjfNir^Oco
r—t in CM
o CN oo fNi o°i cri r^* ro ro fN c^ f-j
COrrr^r^COCOfNCNIOrHOrM
OrHOOrHrHi-HOrHOfNrH
invocNr—iroONt — ror^roroin
rovor-c^or^rH^«o>r-voo
OrHOOfNrHrHrHOOlHfNl
invoro'«j'fN]TjifNrHinocoin
vovocoofNcnorHcNr-r^in
OrHOr-irH'— ICNpHOOrHfNl
2 i — I ro cyi vo
rt fNl CN CN fN) fN
X, X. X, X, VO X, VD in OOrH
ro ro ro ro "\ ^j1 x^ ^x ^x *x
fN CO CN CO 1 rH 1 1 II
i— lrHtNCNfN(NCNr-H'*rCOLnC3N
X. X, — v \ X, \^ \ X, \ X, \ X.
rorororO'rf^'inr^r*r^coco
r--CTiOOCNONfN^J1VOOO
^3* VO r^ ON CO VD ^f O fN f~H
rH rH O VD T
rH f-H
vor~ror~-rHfN]o>cororo
rHrorofNvofNrofjN^r^
rH O
rH CT* rH rH ONONminfN^J'OO LO rH
CN CO 00 in COCOrOfNVCVDfN VO ro
rHrHrHrHfNlfNfNlrHrHfN]
rH fN CO rH
X, fN fN i-H C O\ O
ro\.x,x.cox,orrr-i-H
1 roro^x.inx.x.x.x.
ro i I I ^ 1 in r^ r^ r~
rH CO CN CO 1 CN 1 1 1 I
XrHrMfNirMfNlfOfH^r-
^rSy^-^^-^-^-^-^-^
ffm^
U-l — .
CN f^ <-H CO CN CN ro fN U~l £04-14-1
r^ p* co oo 01 i— i CN co co coo0-!
iH r-{ rH rH rH C Q
lg
•ti*-
WtnUEw^Wt-^EZO ^
88
-------
The stormsthat followed in March, April and early May, some
eight events in all (Fig. 31), conform to the classic pattern of
chemograph preceding hydrograph by several hours, with the par-
ticulate phosphorus tracking the suspended solids. The most in-
teresting aspect of these 1977 spring data is the fact that storms
of the saroc•general magnitude and duration as 1976 spring storms
did produce significant runoff and pollutants, whereas in 1976
the April and May storms did not. The contrast between the two
spring periods is striking and appears to be almost entirely a
function of soil moisture storage.
MODEL VERIFICATION
In spite of the significant differences between 1976 and
1977 in terms of both hydrologic response and pollutant produc-
tion in Honey Creek, it was decided to run the 1977 data using
the final LANDS and QUAL input parameters-for Run 333, with LZS,
UZS, and SGW set as of December 31, 1976. The comparison of
actual and simulated monthly values for flow and suspended solids
mass transport is shown in Table 23. The table indicates that
there are several major problems with the verification and they
are apparently very different in nature.
First, the delayed snowmelt period of late February could
not initially be predicted, because the rule used in accounting
for snowmelt as runoff was to proportion it to the reduction of
snow-on-ground measurement, resulting in simulated runoff almost
two weeks too early. Obviously, the ground thaw must be consid-
ered in estimating the full snowmelt effect, but without solar
radiation data such observation is hindsight. Realizing this,
the snow-melt runoff was moved from the 8th to the 22nd, result-
ing in the simulation as shown. Even with the snow melt consid-
ered properly, the precipitation record and the stream flow re-
cord are incompatible, with two distinct chemograph peaks on the
rising stage of February 23 and 24. The rainfall record only co-
incides with the second peak, and the major rainfall of late on
the 26th and early on the 27th appears to have no relation to
the hydrograph peak. It would appear that the problems of in-
consistency with this event are poor rainfall data and improper
snow melt input. From the model point of view, the simulation of
even the precipitation of the 26th failed to produce any signifi-
cant sediment production.
The hydrograph simulation is good for the February, March
and April period, as shown in Fig. 32. With the exception of the
snow melt event, all storms are simulated well, with the peaks
still occurring too rapidly. The sediment simulation is not too
good, however, with the model showing a marked lack of overland
flow.
The largest single problem with the 1977 data used in model
varification occurs in the period of early July. During the 4th
89
-------
OSc
TTtT
GG-c
ub'f
OST
001
r-
LO
r-H
( OT x
tr>
O
g
0)
45
o
CO
T3
•H
•H
O
U3
'O
O
M
U
ro
CU
S-l
•H
( 01 x T/&UI) spTjos pspuedsns
90
-------
( oi x sjo)
91
-------
Table 23. VERIFICATION SUMMARY
1977
Actual Data
1977 (Run 33377)
Simulated Data
Flow Suspended Flow
(inches) solids (tons) (inches)
JAN. 0.02 0.82
FEB. 1.07 993
MAR. 2.47 1245
APRIL 1.59 692
MAY 0.63 475
JUNE 0.042 5.5
JULY 0.632 29582
AUG .
SEPT.
0.005
0.65
2.24
0.86
0.44
0.14
0.102
0.32
0.56
Suspended
solids (tons)
0
243
272
1421
11.4
0
O2
0.97
3653
1. Fortran error message, underflow, encountered during simu-
lation interval for April 21, 12:45, standard Fortran fix-
up taken and simulation continued.
2. Actual record complete through July 10, 1977. Large dis-
crepancy in flow and suspended solids between actual and
simulation is due to a large local storm in the eastern
Honey Creek basin not recorded at Tiffin gage that was
used as precipitation input to NPS model.
3. Simulation complete through Sept. 18, 1977.
92
-------
and 5th, heavy rains produced a major runoff event, with the
highest suspended solids and phosphorus concentrations of the
year. The Tiffin gages, however, do not record any major rain-
fall, and so the model fails to simulate this event altogether.
Local observers reported severe, highly localized thundershowers
during this period, tracking across the middle of the basin.
It was also apparent that the moisture retained in the soil
was greater the first half of 1977, a fact that is reflected by
higher base flows and greater sediment production during May and
June. The use of a variable K3 might compensate for this condi-
tion in part.
93
-------
SECTION 10
EVALUATION OF BEST MANAGEMENT PRACTICES
The magnitude of diffuse pollutants from agricultural land
runoff in the Great Lakes, especially Lake Erie, has been well
documented in several recent studies (COE, 1975; GLBC, 1976) .
The tributary input of phosphorus from diffuse sources was
estimated at 6600 metric tons during 1975, with most of the phos-
phate mass transport occurring in association with soil particles.
It has also been estimated that most of this material originates
from agricultural land within the Lake Erie basin, much of which
is intensively cultivated, especially in the west'ern portion.
Because of physiographic features much of the agricultural land
in the western Ohio portion of the basin is flat and poorly
drained, but extremely productive under the necessary condition
of artifical drainage. The Honey Creek basin is highly represen-
tative of the major soil types and drainage conditions, and as
shown in Table 3 is 83% agricultural land, with only one small
municipal point source. Melmore record shows that the 151 sq.
mi. drainage area above the gage discharges 38 M.T./yr. (0.98
kg/ha) of total phosphorus, a value consistent with the major
agricultural basins in Lake Erie, which range from 1.1 to 1.4
kg/ha/yr. output.
Given the existence of the land resource data base, one of
the tasks accomplished in prior studies was to calculate the po-
tential gross erosion in the basin, by sub-areas, using the land
cover, soils and slope file combinations for each of the 12,000
cells, based on the Universal Soil Loss Equation (Stewart, 1975).
From this work during 1976 came a prototype report(RMA,1977),
which focused on land management techniques and sediment runoff
from cultivated lands. It was apparent that in order to influ-
ence the local farmers in implementing these practices several
steps were necessary; first, in the actual source areas, transport
and impact of these pollutants had to be better defined and mea-
sured and second, the degree of success in improving water qua-
lity (while maintaining productivity) had to be proven, either
on an estimated basis by modeling studies or in actual practices
by demonstration.
With the calibration of the NFS model for 1976, it was felt
that a comparison of the estimated pollutant load reduction by
the two methods would be of interest. One potentially successful
94
-------
practice, that of no-till planting, was selected and the NPS
model run to reflect the impact of this practice. In the final
calibration of NPS for the 1976 data, it was found that very
careful specification of land cover, fines accumulation, fines
removal and initial fines pool were critical. As calibrated, the
output is supply-limited during the high water table period of
the year and overland-flow limited during most of the growing
season.
NPS MODEL PARAMETER ADJUSTMENT
The NPS model calibrated to Honey Creek for 1976 represents
what, is hereafter refer red to as conventional (cropland) manage-
ment. As presently formulated, it does not contain explicit pro-
visions for tenting alternative farm management scenarios. When
calibrated to a watershed as large as Honey Creek, the model re--
presents at a point, the variety of natural processes and manage-
ment practices that are occurring over a 150 square mile area.
Insofar as the fines generation equation of the QUAL subroutine
represents a distinct reality, the COWEC parameter, set on a
monthly basis, .IK the only model parameter that is directly
measurable and used as measured. Other model parameters attempt
to approximate measurable physical characteristics or process
rates but are subject to the model calibration process which can
result in large discrepancies between the measured and the cali-
brated parameter veilue. The difference in real world values for
the parameters and the calibrated values is a rough measure of
the discrepancy between actual and model processes.
For example, the parameter INFIL represents an index of the
mean soil infiltration rate on the watershed. Its use in NPS
attempts to represent a distribution of soil permeability in
space but actually results in the simulated infiltration rate
being distributed in time as a function of soil wetness. The
mean soil infiltration rate of Honey Creek soils is approximately
0.4 in./hr, a value assigned to ILJP1L tor the first calibra-
tion run. However, calibration efforts quickly revealed that
INFIL must be radically lower than the real world value of 0.4 in./
hr: a final value of 0,02 in./hr was eventually selected. Even
such an artificially low value was not sufficient to produce the
overland flow from pervious areas needed for sediment washoff .in
the last half of calibration year 1976.
With these observations in mind, it is possible nonetheless
to estiiaaLe conceptually the direction of change to various NPS
model parameters that can be anticipated if a change in cropland
management occurs. Table 24 presents a preliminary attempt to
do this. Across the top are listed the NPR model parameters in
three groups: LANDS (hydrology), QUAL (sediment production), and
land use variables. Candidate best management practices are
shown down the right. The estimated direction of change in a
model parameter due to implementation of a practice is indicated
95
-------
TO
£4
Ed
FI
q
"1
S
<
A4
O
5s
H
Q
i
E-i
53
H
O
Oi
I
S
o
P-,
01
EH!
"r '
^\
p!
a
<
j
u
M
S?
(J
g
Q
O O O
SZT
'/tdX3
£H
1
SS
a a
«« oa
a
OH I
H31NI
NSZfl
aa
ooa
a
a
a
aoa
•»*
(N
CD
nd
EH
Q
tJ
K H
U M
•--1 " . I
H 71
u s!
< O!
3
H
03
Q
Z
D
S3
>
EH
H
K
W
b-
Z
<
C
D
2
0
K
EH U
RIDGE PLA!*
DIRECTION OF
-
e<
o
&
c
u
&
H
K
COHTOUP. ST
TERRACE
^ KH LO Z U
*H uJ ^K O f-i
,_-> f^ Z+ L_ ^ ry^
t— 1 !*•» <. t— ' O W
C-i 2; to £ E^ C- 12
D^2 W
-------
by an arrow. For example, the BMP selected for detailed analysis,
no-till, may increase the upper soil zone storage capacity (UZSto
by increasing the organic content of the surface horizon. Infil-
tration capacity (INFIL) may improve due to minimum disturbance
of soil aggregation processes. Surface roughness (NN) will in-
crease due to accumulation of crop residues. The index to actual
evapotranspiration (K3) may decrease due to light colored crop
and weed residues retarding soil drying. The same residues in-
crease interception storage (EXPM). Depending on the time of
year the NFS model simulation is begun, initial upper zone stor-
age may be slightly higher. No-till may have an influence on
intrinsic soil erodibility (KRER) but its effect is likely to be
complex. No-till may lower the basic capacity of overland flow
to carry sediment (KSER) due to its effects on sediment particle
size and on surface roughness. Its most obvious and measurable
effect is on cover (COWEC) which is anticipated to increase ovar
conventional tillage every month. Potency factor (PMPMAT) for
phosphate is likely to increase due to the limited fertilizer
placement options of no-till. Since no-till minimizes soil dis-
turbance, accumulation of fines (ACUPV) during normal plowing
seasons will be nearly zero. Undisturbed soil aggregation pro-
cesses may increase the natural rate of fines removal (REPERV)
during non-storm periods. Finally, the initial supply of fines
for no-till is likely to be a small fraction of that present
under coventional farm management (SRERI).
NFS model parameters can be similarly evaluated for ail on-
field practices. However, applicability of NFS model to simulate
BMP becomes tenuous for mixed field/channel effects such as tile
drains or permanent diversions; the NFS model was not designed
to simulate BMP's located in the stream channel, although such
algorithms are under development. Grassed waterways, stream bank
and outlet protection could be treated as practices on critical
areas instead of channel practices, but this would require sep-
arate NPS model calibration of these partial areas, separate
flow and water quality data for which are not likely to be avail-
able.
Although it is a fairly straightforward exercise to antici-
pate the direction of change of a model parameter due to imple-
mentation of a BMP, it entirely another matter to select a rea-
sonable value. The first reason for this difficulty is that
very few specific field tests have been made of these practices
with the NPS model parameters in mind; second, even if such mea-
surements were made and significant changes noted betweem con-
ventional practice and BMP, one would be reluctant to use them
as such since most model parameters are "calibrated" and as such
bear a fragile and tenuous connection to field measurements. The
only notable exception to this is the cover parameter COWEC
which, as mentioned above, is directly measurable in real space
and used in the model without modification during calibration.
97
-------
NPS_ model simulation
No-till crop production is selected to demon strat--- tvo ap-
plication of the NFS model to estimation of sediment lead redac-
tion by implementation of a potential BMP for several, reasons:
1. No-till is a field practice,
2. No-till significantly increases cover (COWEC) .
3. No-till is suitable on signigicant areas of Honey Creek,,
No-till crop production in this discussion is .limited to row
crops, corn and soybeans, which, for Honey Creek in 1976, ac-
counted for 52% of the watershed area. COVVEC, ACUPV anr1 SRERL
values for no-till corn and soybeans were estimated. The steps
used in the development of no-till scenarios are to determine
new model parameter values for no-till, then to run i-.no model i.'or
all land uses except row crops, for row crops under conventional
management and finally for row corps under no-bill management-.
The monthly suspended solids loads for three no-till scenarios
are then calculated by an area-weighted addition of loads from
the 3 above NFS model runs.
Suitability for JlOjtil_l:_Three_
The Ohio Agricultural Research and Development Center has
classified major Ohio soil series according to their suitability
for no-till corn production (OARDC, 1973). Economic criteria
are Considered in this classification: corn yields of no-till
suitable groups are better, the same, or only slightly lower than
yield? from conventionally cultivated corn. A computerized count
of Honey Creek soils in cropland as they fall into the nc-till
suitability groups is shown in Table 25. Less than 10% of the
watershed is suitable without drainage improvements. Another 25%
is suitable if drainage i? improved. About fifteen percent, of
che watershed that is cropland is classified as unsuitable (!<:;, 5-a)
or insufficiently studied to classify (4.4%). Nineteen percent
of the watershed is not cropland. The remaining third of the
watershed (34.4%) is cropland that can be planted using lo-tili
met nods for row crops if excessive soil moisture d^lr.y- conven-
tional plowing until May 10th or later.
However, due to crop rotation schedules, a fielr: suitable
for no-till in any on of the three suitable categories is nor
always available for no-till row crop production, bcinq corom: t te."i
to wheat or oats during some seasons. Thus the oercenr.oge 01 r.oe
watershed potentially available to no-tili ma^.naem. TLT. if: the^e
categories is reduced by approximately three~ej ghts w
proportion of cropland in 1976 devoted to snail grajx
no-till suitability figures are summarized In Table '£
The three no-till scenarios tor Honey Creek to ":•
in detail are derived from Table 25, with tho crop ic
98
-------
Table 25. SUITABILITY OF HONEY CREEK CROPLAND
FOR NO-TILL ROW CROP PRODUCTION
No-till suitability
class
Cropland
Percent of Honey Creek Watershed
Independent of
crop rotation
schedule
Restriction by
crop rotation
schedule
Suitable without
drainage improve-
ments
Suitable only with
drainage improvements
Suitable only if plant-
ing delayed beyond May
10.
Not suitable or not yet
classified
Other Land Uses
Urban, woodland, water
or wetland
6.7
25.0
34.4
14.9
4.2
15.6
21.5
19.0
99
-------
striction, as follows:
No-till Scenario 1: no-till row crop on 4.2% of watershed.
No-till Scenario 2: no-till row crop on 19.8% (4.2%+15.6%)
No-till Scenario 3: no-till row crop on 41.3% (4.2%+15.6%+
21.5%)
Estimation of no-till parameters
No-till COWEC values were estimated by the same bookkeeping
procedure used for the calibration of the NFS model to all of
Honey Creek. Crop rotation effects, prior crop residues, crop
area ratios, and conventional practices for small grains were
taken into account. Crop residue carryover from previous years
and lack of ground disturbance in April and May were the main
determinants in the increased COWEC values estimated for corn
and soybeans as shown in Fig. 33. The COWEC curve for no-till
corn is compared to COWEC values for conventionally tilled row
crops (corn and soybeans) in Fig. 34. Although there is some
difference in COWEC between no-till corn and soybean in the last
half of the year (Fig. 35) , the no-till corn COWEC values were
those used in the no-till NFS scenarios. This choice will not
affect the results obtained since there is no overland flow and
thus no suspended solids loads simulated from crop land during
the last half of the year.
The NFS model permits specification of monthly accumulation
rates for soil fines. The ACUPV parameter allows the user to re-
present soil disturbance, such as plowing, as a daily increase in
the pool of soil fines available for washoff. Calibration of the
NFS model for Honey Creek in 1976 set this rate at 10 Ibs/acre/day for
row crops during April and May and during October and November to
represent pool increases due to spring and fall plowing activi-
ties. Since no-till crop production does not significantly dis-
turb the soil, ACUPV values were reduced to zero for the no-till
run of the NFS model.
Similarly, the size of the fines pool (SRERI) in February
was reduced from 250 to 25 Jbs/acre for the no-till simulation.
(Had this reduction not been made, the best no-till scenario
would have been nearly indistinguishable from the 1976 conven-
tional simulation).
Simulated erosion control by no-till
In order to estimate suspended solids loads for the three
no-till scenarios, the NFS model was run separately for three
situations.First,four land uses (small grains, woodland, water/
wetland, and urban) with calibration parameter settings were run.
Second a one square mile area of row crop, also with calibration
parameter settings, was run. Finally, a one square mile area of
row crop with cover (COWEC), fines accumulation rates (ACUPV),
100
-------
I 1
CD
CJ
.90
.80
.70
.60
.50
.40
.30
.20
.10
NFS INPUT COVER VALUES
no-till corn
no-till soybeans
JAM FEB MAR APR MAY iUN JUL AUG SEP OCT NOV DEC
Fiqure 33. No-till cover values.
101
-------
CJ
C3
52
H-
CJ>
O no-till corn
£ conventional row crop
90 - NPS INPUT COVER VALUES
.60 -
.70
.60
.50
40
.30
.20
.10
0 -
JAN FCB MAR APR MAY JON JUL AUG SEP OCT NOV DEC
Figure 34. No-till vs. conventional cover values.
102
-------
•a.
oc
.90
.60
.70
.60
.50
.40
30
,20
10
NPS INPUT
COVER VALUES
C ON VENT IONAL MGM1:
row crops
Q small grains
JAN FEB MAR APR MAY JUM JUL AUC SEP OCT NOV DEC
Figure 35. Conventional cover values - row and field
crops.
103
-------
and initial fines (SRERI) modified for no-till (see above) was
run. The monthly and annual loads for each scenario were calcu-
lated from these three NFS model runs by adding together appro-
priate proportions of the separate simulations as required by
each scenario. The NFS-model-simulated results, summarized in
Table 26, predicts that no-till row crop production on well-
drained cropland in 1976 could have reduced suspended solids
loads by 3%; by 15% if artificial drainage had been improved to
the extent possible, and by 33% if row crop planting had been
delayed beyond May 10th.
In considering these results, it must be remembered that
they apply only to the hydrologic situation of 1976, with its
unusual snowmelt-rain event in mid-February. Examination of the
monthly results in Table 26 reveals that the three scenarios
differ for February, March and April only. Since there was over-
land flow (simulated) from May through July, the identical re-
sults in these months means thatthe loads generated were not
limited by the supply of fines available for washoff but by the
depth of overland flow, which was the same in all scenarios.
This result is reasonable in light of the relatively mild rain-
fall experienced during this period; there was no severe late
spring or early summer storm to deplete the no-till available
fines and thus distinguish it from the conventional management
results. The months from August to December were lacking any
NFS model overland flow component from pervious surfaces and thus
could not deliver sediment to the streams from cropland. The
suspended solids simulated during these months is solely from
impervious urban surfaces.
Implementation of scenario three is dependent upon a wet
spring to delay planting to May 19th. Since this condition does
not occur every year in Honey Creek, the long term potential re-
duction in sediment loads, as predicted by the NFS model for no-
till crop production, probably lies somewhere between 15% and 33%.
Although the 15% reduction in sediment load from Honey Creek
estimated by the NFS model for no-till row crop production under
scenario 2 appears modest, it must be remembered that only 20% of
the land area is involved. If a unit area of row crops under
conventional management is compared to the same unit area under
no-till crop production, it is seen that the simulated delivery
of suspended solids from the no-till area is 30% of that from
conventionally grown row crops,or a 70% reduction. Other BMP
candidates to complement no-till should be evaluated for cropland
not suitable for no-till or not in a row crop rotation, possibly
by application of the NFS model as illustrated here.
Evaluation of BMP's by the Universal Soil Loss Equation
The Universal Soil Loss Equation (USLE) is another planning
tool used to predict soil loss. "C" factors were estimated for
104
-------
w
H
Pi
U
X
w
(-7
»^-I
o
K
2
H
^3
? — 1
o
H
EH
O
;D
P
0
Pi
ft
(^
O
Pi
O
J
J
H
EH
1
O
s
>H
m
j
O
Pi
EH
!Z
O
o
iz;
o
H
cn
0
Pi
H
•
U3
CM
CU
X!
rrt
fo
EH
^,
CO
J3 T3
O -H
H rH
EH 0
x) co in
0 g ^T rvj"
3 0 H
0\°
i nif^
*"• •
•^
4-1
i
O fl
O O O O
O ^ CM O
CM H
C rH
O -H
O EH
O O O O
O O CO O
co o ^ in
*. K.
U3 CO
rH
.-H
A-3
4J
a
o
a
• • •
X! M M >,
O
O
ro
o
CM
0\°
oo
o
O
O
O
(ti
4->
O
EH
o
C!
O
•H
4->
O ^
3 td
T3 O
CD J
Pi
•P
-P a
C 0)
CU g
0-H
M -O
(D CU
P-I cn
105
-------
conventional and no-till crop production for Honey Creek (D. Ur-
ban, 1978) and applied to the calculation of potential gross
erosion rates. While the gross erosion estimate is substantial-
ly greater than the actual net transport of sediment from the
basin, the relative reduction in USLE and NFS model as results
of no-till crop production can be compared (see Table 27).
Each of the 11,500 four hectare cells were calculated using
the USLE, based on a "C" factor of 0.245, "L" of 300 feet, "R" of
125 and using the "K" and slope from the computer data base.
The no-till improvements to "C" were applied only on that water-
shed land area where no-till is suitable, (see Table 25). As-
suming that the relationship between estimated gross erosion and
actual net erosion (i.e., the delivery ratio) holds constant
with spatially variable BMPs such as no-till, the percentage re-
duction comparison shows that the two methodologies of evaluat-
BMPs would give very different answers (Table 28). The NFS
model indicates far less of a reduction in sediment delivery
from the basin due to no-till application on suitable lands than
does the USLE. It will be an interesting future research effort
to compare relative reductions in sediment yields' estimated by
the NFS model and the USLE for other BMP scenarios.
It should be noted that the USLE estimates a long term
gross erosion value and so to be compared directly with the NFS
simulation, several years of data should be analyzed.. This, of
course is not yet possible for the Honey Creek basin., but the
comparison shown is informative in terms of the variability to
be expected.
106
-------
CO
O
O
-rH
•H
O
-P
o
O
O
oo
*.
o
CN
CO
CO
r-
CM
CO
CTi
rH
ID
in
in
SH O -H
-P -H
rH C
d -P -H
U O
•H c td
J_l I i
(d
m
w
r3
CO
0)
X!
-P
^
O
^
imated
-P
to
ni
VL/
CO
td
_
2
' *
i — i
EH
1
O
"
O
EH
W
H
*
(N
0)
rH
X!
td
EH
£
n
~J o\o o\°
T) rH 0 CO
n _ ...
3 ^ ^ n m
rg *, rH 10 t-
0)
ftf
CO 0)
0 U
" ^ CTi LO T CO
I .A • ...
-H C ^ ^ H °
f— 1 j_j
0 0
to
to
O
i 1 — 00 CPi f~-
fj rn O rH rH
"o — . ^ ^ ^ "^
w w o^ ^ r~^ ^
rH n ^ H ^ ^
td EH ^ ™ --1
-P ^""^
O
EH
H H
H -
tO H
d IT) c3 M
O *3«
4-J-HtN HH rHH Hr-i
C -P • rH rH rH
(U -H 0 -H tO -H CO -H CO
CO rO -P CO -P CO -P CO
D C n i id I fd i (d
SHOU OH OrH OrH
PuU 2U KU 2iO
0 Q)
)H • nC
O T3 -P
g C 0
•H td
0) rH M
S 0
Q) CO -
^ ^ QJ
O SH 3
X! X! rH
td td
td fl
0) td =
SH U
td G =
td
C! SH Q)
td 0 3
rH ^ rH
-P id
rH CO >
•H QJ
td M =
O W
1 1 i H i •-
O
*>" r*
OO O -H
CO T3 rM
td O
to 0) id
QJ g rH
d ^ co
rH H3 -H
C td rO
H rH QJ
a §
H
EH
O
^
107
-------
13
(0
O
o
H
EH
<
D
C
0)
e
H
W
U
H
EH
U
EH
3
W
S
W
O
EH
CO
W
ffl
O
CM
CO
Q
O
K
EH
W
S
CM
O
O
CO
I
CJ
OJ
CO
CO
H
CO
w
A
CO
D
tn
H
Q
O
CO
4J
O
3
•O
tfl
W
0)
Q) O
O
•H
EH
CO
n
in
o\o
n
ro
r-
0)
13
O
CO
ft
ro
in
n
rH H
rH
•H 01
4J W
I 03
O rH
3 O
H
H
rH H
rH
•H W
4J W
1 fO
O rH
3 U
H
H
•H W
4-> W
I (tl
O rH
3 U
00
CM
0)
H
.Q
(d
EH
108
-------
REFERENCES
Baker, D.B. and Kramer, J.W., 1973. "Phosphorus Sources and
Transport in an Agricultural Basin of Lake Erie." Proc. 16th
Conf. Great Lakes Res., Ann Arbor, Mich.
Bliss, N., T.H. Cahill, E.B. MacDougall and C.A. Staub, 1974.
"Land Resource Measurement for Water Quality Analysis." Tri-
County Conservancy, Chadds Ford, Pa.
Betson, R.P., 1964. What is Watershed Runoff? J. Geophys.
Res. 69, No. 8, 1541-1552.
Cahill. T.H., P. Imperato and F.H. Verhoff, 1974. "Evaluation of
Phosphorus Dynamics in a Watershed," Jour. Env. Eng. Div.,
A.S.C.E., Vol. 100, No. EE2, Proc. Paper 10445.
Cahill, T.H. and T.R. Hammer, 1976. "Phosphate Transport in
River Basins," I.J.C. Fluvial Transport Workshop, Kitchener,
Ontario.
Corps of Engineers, 1975. "Lake Erie Wastewater Management Study:
Preliminary Feasibility Report."
Buffalo Dist., U.S.C.O.E., Buffalo, N.Y.
Donigian, A.S., Jr. and N.H. Crawford, 1977. Simulation of
Nutrient Loadings in Surface Runoff with the NPS Model. Hydro-
comp, Inc. Palt Alto, CA. Prepared for U.S. Environmental Pro-
tection Agency, Athens, GA. Publication No. EPA-600/3-77-065.
Donigian, AS., Jr., D.C. Beyerlein, H.H. Davis, Jr. and N.H.
Crawford, 1977. Agricultural Runoff Management (ARM) Model,
Version II: Refinement and Testing. Hydrocomp, Inc. Palo Alto,
CA. Prepared for U.S. Environmental Protection Agency, Athens,
GA. Publication No. EPA- 600/3-77-098.
Dunne, T., 1975. Recognition and prediction of Runoff Producing
Zones in Humid Regions. Hydrological Sciences, Bull. XX, 3,
9-1975.
Great Lakes Basin Com.., 1975. Great Lakes Framework Study:
App. 2, Surface Water Hydrology. Ann Arbor, Mich.
109
-------
Great Lakes Basin Comm., 1976. Maumee River Basin, Level B Study
Analysis of Non-point Sources, by Resource Management Associates,
GLBC, Ann Arbor, Mich.
Huggins, L.F., 1977, (unpublished). Manual for Small Catchment
Distributed Parameter Hydrology Model, version 9. Dept. of Ag.
Eng., Purdue Univ., W. Lafayette, Ind.
Kirpich, Z.P., 1940. Time of Concentration of Small Agricultural
Watersheds. Civil Eng. (NY), Vol. 10, No. 6, p. 362.
Lin, P.N., 1952. Numercial Analysis of Continuous Unsteady Flow
in Open Channels, Trans. Am. Geophys. Union, Vol. 33, No. 2, pp.
227-234.
Massau, J., 1900. Graphical Interfration of Partial Differential
Equations with Special Applications to Unsteady Flow in Open
Channels. Anneles de 1'Association des Ingenieurs sortis des
Ecoles Speciales de Grand, Vol. 23, Ghent, Belgium.
Norris, S.E., 1974. Regional Flow System and Grqund Water Qua-
lity in Western Ohio. Jour. Res. U.S. Geo. Surv., Vol. 2, No. 5,
1974.
OARDC,1973. An Evaluation of Ohio Soils in Relationship to No-
Till Corn Production, by G.B. Triplett,Jr, D.M. Van Dorn Jr.
and S.W. Bone. Ohio Agricultural Research and Development Center,
Wooster, Ohio.
Palmer, W.C., 1965. Meteorological Drought. Res. Pap. No. 45,
U.S. Dept. of Comm., Weather Bureau, Off of Climatology.
Wash., D.C.
Phillip, J.R., 1957. The Theory of Infiltration: The Infiltra-
tion Equation and Its Solution. Soil Science 83:345-475, 1975.
Quick, M.C. and A. Pipes, 1975. Nonlinear Channel Routing by
Computer.Jour, of Hyd. Div., A.S.C.E. Hy 6, June, 1975.
Resource Management Associates, 1977. "Honey Creek Report,"
for U.S.C.O.E., Buffalo, N.Y. (in publication).
110
-------
APPENDIX A
HONEY CREEK BASIN
LAND RESOURCE INFORMATION SYSTEM
The land resource information system (LRIS) that exists for
Honey Creek watershed consists of land use, soil, and slope data
recorded and computer-stored for each of 11,600 data points or
cells. Other data files include political and hydrologic bound-
aries in which each of the 11,600 data points is assigned a
number corresponding to its location within a political unit or
a subbasin in Honey Creek watershed. RMA has developed computer
programs that can characterize the whole basin or any political
or hydrologic subunit in terms of land use, soil and slope.
The utility of the soil information is enchanced by an ancillary
soil characteristic data file. With this soil data file any
characteristic listed can be selected to characterize any or all
spatial subsets of the watershed. Furthermore, other programs
convert the statistical characterization of the watershed into
a spatial characterization, by sending the stored information to
a CALCOMP plotter that produces maps reflecting the character-
ization selected. To illustrate the flexibility of LRIS and its
backup statistical and mapping computer capabilities, Resource
Management Associates produced a series of thirteen tables. In
this series various characteristics of Honey Creek watershed are
summarized for each of 18 subbasins. The subbasins are arranged
in a hierarchy from the headwaters to the mouth but the data
presented are not cummulative; statistics for a main stem reach
refer to that reach only and do not represent the total area
above that reach. The subbasins described in the tables are
shown on Figure 18. The subbasin numbering on the map and ta-
les correspond. For each characteristic, listed across the top
of the table, two entries are shown. For every subbasin, on the
left, is given the area, in hectares, that characteristic or co-
occurrence of characteristics occurs. On the right is given the
percentage of the whole basin that is found in the subbasin and
having the listed characteristic. Footnotes are used to amplify
certain aspects of the tables.
Table Al. Presents land use in Honey Creek by subbasin. In the
table, seven land use categories are shown and represent one
aggregation of the 21 land use categories coded for Honey Creek.
The summary shows that 50 and 30% of the whole basin are devoted
111
-------
to row and field (small grain) crops respectively.
.? • Displays the distribution of physiographic regions in
the subbasins. Ground and end moraine account for nearly 80% of
the watershed.
_e __A3 1. Divides the subbasins into slope categories. The
basin is quite flat: 77.3% has a slope of 0-2% and another
15.3% has a slope of 2-4%.
Table A 4 . Shows how the various subbasins occur with respect to
county jurisdictions. Overall, more than half of Honey Creek
(57.5%) is in Seneca Co.
Tab_le_A5. For this table and most of the rest, the ancillary
soil characteristic data file was used. Here, each subbasin is
statistically described by the drainage characteristics of its
soils. Thus, Ackerman Ditch is somewhat to very poorly drained
in 1124 of its 1136 hectares. (This representation assumes no
artificial drainage) .
Tab 1 e A6 . Shows that most (82.1%) of Honey Creek has a seasonal
water table that is quite close to the surface (0-0.5 ft.:
category 1 + category 2) . Not all subbasins are similar how-
ever: note subbasins 5 and 6 in water table categories 1 and 2.
Table A7_. Illustrates the intrinsic erodibility of Honey Creek
soils according to an index (k) developed by SCS. About two
thirds of Honey Creek soils have a moderate to moderately high
erosion potential (k=0.37 or 0.43).
A8_ . Shows the permeability of cropland soils by subbasin.
_
The tabTe entries show the hectares of land having the stated
permeability and the percentage of the total basin area repre-
sented.
Table_A£. Describes the subbasins soils in terms of their fine
ne~ss"oir~texture. The table entries reveal the hectares of land
in a subbasin that have soils (plow layer) capable of passing
stated percentages of their particles through a number 200 sieve.
Table A10. Shows the distribution of plow layer soil textures
in the various subbasins. Over 60% of the watershed is covered
with silty loam soils.
Table All . Departs from the format in the previous tables. It
ITsl;s~TTectares of crop types by physiographic regions. All the
physiographic regions appear similar in the ratio of land devot
ed to field vs row crops with the exception of glacial drift.
112
-------
Table A12. OARDC has classified Ohio soils according to their
suitability for no-till corpping of continuous corn. Table 1?
presents this classification for Honey Creek watershed. It is
seen that only 7% is suitable outright and another 28.7% suit-
able if artificial drainage is provided.
Table Al3. Presents the estimation of soil loss in Honey Creek
watershed by application of the Universal Soil Loss Equation.
The equation was applied to each of the 11,600 points in the
data file. The areas subject to levels of erosion in one of
five categories were added for each subbasin. Thus Ackerman
Ditch (1) had 444 hectares that lost 2-4 tons per acre per year,
and 372 hectares losing 4-8 tons/acre/yr. The various assump-
tions made in the calculation are listed in footnote 3 to the
table.
Subbasin code used in Tables A-1--A-3 and A-6--A-10
_!, ACKERMAN-Dl.TCH ___
, 2 HONEY— CR.-3 _RI - 10 3 .
_i--HONEY-.AB BROKNKNF.
_4-TRIB.AT SCOTT ..RD .
5 TR13-XT-HEI S_ R 0 AD-
__6_HONEY_AB_ AICHHOLZL
_7_ HONEY.. CR. 3.RT 4 .. .
_a_BRflKNKNF_a_LINE..R.-
- 9 A1CHHLZ. 3 CO R 49
10. HONEY CR.A8 SILVR.
1L..SI LVR. _C R_ AB _ M AR SH_
li .SILVR CR .BL. MARSH .
13 SILVR. CR. AT HONEY.
15_ BUCKEYE C 3 RT_.67.
16 HONEY CR a RT 331.
lt_MOHAWK..3 Rt_213 ____
18 HONEY CR a MOUTH
19 MISSING DATA
ZO TOTAL ALL BASINS
113
-------
rxi>
g
H
CO
ffl
fft
* ^
j ^
co
r*
PQ Q)
hr^
W EH
2 1)
u w
* D
W T!
3 C
O nj1
ffl >4
2
H
CO
D
O
I—,
*J
1
rH
1
rJ.4
Q)
,
"fl
EH
UJ
_J I/I UJ
< 3) CL
t- <
0" C _J »-
t- _J 0
UJ
I
*™~ "™~ ~~"
| l/V UJ
*— d o:
ir> ? < i_
_) o
C~
X
C Q
CO a
O UJ ^
0- C O H-
UJ
X
— —
*>
QC UJ
UJ O 3^
X
Z, U4 U.
^ 2
2 ' ^ S
CU ;
13 -1 ft
^ f
-O -i U"!
f\J sO C\
r^. r- —
GO CO sO
t\J (Nl
O O C
o o c
1
0 0 C
PO 0 ^H
coo
>-C tM (N.
(T) ~4, ^
t— *
1
— * o —
ooo
cc >^- sC
(N f"i
C 0 rr,
C O --
•4" (V ^C1
i— » C^
LT
*± >
1 01 U
i Z -2.
a c
x «
1 1 "
1
-
-
cc
ol
M
in
O
o
o
cc
0
o
^
rr>
(XJ
O
IN
r—
o
o
rg
•— i
r-l
^
CO
o
in
o
S
- 0
X 01 3
! ^H O ^
1 «* T «
I cr- o
1 , -1 -
og ,$• 1*^ CO i^- r\'
CM gj c\ih cc r: ^»-
m m cr -j- o co
—i O rn j- j- rt-.
1
0 C 0 0 0 0
000,000
O O 0 c 'J J-
O C O o 0 c
gj go
1
0-1 00 — 0
o o o o o o
'; g; -1 IT
r^ ,-. -i !^ --.
1
j
-r o ^ i (xj ^ a
co IT. rvi K- i~ r
J- c* r~ r- IT r--
1 '"^
f * ci , £ > 5
• > > UJ & IU *
J -J _1 Z O Z 3
< i-i i-i 0 3 0 C
110 (Si x «> x a
! i '
oj ro ^f JA ^? ^
^ «-* fH •"**"* -* ^
ro' 0
0
-s
<—• .,y -^
xO
• ^
O O '~
CO O
o o ; oo
j
i
j
r^ o i gr
c r j -J-
1
i
o o i ^
"xj ! m
>-4 •>•<
j ->-< _J
UJ in ! <
7s S\ * v-
) O *- ! O
: 3 »r 1 »-
j to O1* ' O
^ HU>4 f-f ) l^J
114
-------
-4 rg
OO -i UJ
< < ex
CO t- <
O1* CD O l&— «O
l/> ]UJ -4
o
o o
Z UJ
>-• «x or
10 1- <
CO - o
t- O «• UJ
1/1 o u 'o:
3 O <
i*- o o o*. h- co
H-J OO O UJ CM
x
c/3 ;x
« ~u~ w °
ffl z o
rn z < ex
UJ uj a: :<
•O O '»- CM
'X s: o -H
« %
H u,»- :w "
SO Z 0 -4
Z -< UJ ;UJ
, . 13 t tfl a!
U o * — <
in a a a *- CM
"s_j r-t O 31 Z o P-
"r? « a ,iu co
HO ,i
O D> - o
C£ Q Z LU O
,_ "^ Z ^ f- LU
i*-i ^ < u a.
H CJ a a: uj <
>r. <*• cc a w) '»- o
F r\ f* & 51 *~* *— >
UJ r*H O ILL)
2 ft •!
O (0
H J-i 0
w c 2is£ L °
« -H o*c* ,oc
U; < t- ot ,<
, > s. en _l :D UJ (- O
w ft o o i- o
H rC LU
ffi CL l-^
H < Lb
1 O
- O
0 o
0) 'l^
5 ,
(3 z" 2
r i ^_ uj
§ -j
i
JJ <0 -4
o o o
o o o
.0*0
~4
-1 -< O
CM >f O
rsj •-< cj<
CM ~4 O
•O O CO
moo
o «r *
CP 0 0
o o o
CM o o
-< rg rM
,-« in >-
-4 m o
— < ro o
u^ *Q ff)
000
o o o
000
r- ui o
o o o
— i in
en rg
0 O 0
O 0 O
o o o
v1 >J ^
UJ UJ CD
Q O oi
2 re ^**
rJ <*1 ^
I ' 1
LTV O -^ **1 O
lf\ ^| 00 <-4 CO
CM r\j o co «o
•-4 ij* ^ O ""•
•/Q t— 4 CD ' r*i r>j
O -4 O 0 O
o o o o o
•»• •*•«!• O *
^ ' '"'J
1
^^ co rn I ao in
en o ^ o o
^ ^ <^ O ^)
~» in -o * CM
f— —1 JJ !rg O
"^ i/\ en CM o
rg •<• o o o
— < CM O "O
00 >t Is- O
CM -^ rH
^ 00 ^ O^ *-4
o -r en i* oo
*t o rg o rg
o -t in rg -i
rg rg -*• en so
rg r-< irg en
o o o m o
*•• O
-> in ^
O -^ O O O
o o o 10 o
o o o o o
j
o o o o o
o o o o o
o o o ' o o
o o o o o
.5 . mia-A o o,
.6^HONEY. 0 0
J- HONEY^ 0 0
_fl_BR.QKNK 0 0
9 JMCHHL 0 0
i i i I i
~t -o -t -t ,o -14-
tn : »o CM -4* , en en f—
t
^•oo rg *o rg oo co
in —» r-H o 'n sj- -*•
O O -4 O O O O
o o o o o o o
1
•»• -O *t O 03 O *
!-H CM
O O -4 _< O 0 O
O 0 O O O O O
o o o ^ o o o
tfN in o o r* o ~<
-4 — ' O O O O O
o * o o Q 30 IA CO -^
o ten -H CM O •-* -^
1
y> ( o CM oo >o -o ^o
•M j o m o ' -n m o CM
n co in
H CO
i
I i ' | X
K a> a! t£ v UJ >
jj > > > ai *: uj
Z _J -1 _l Z (J Z
3 M »4 M O 3 O
r cq m ifl x 03 x
3 «4 rg rn •£• irv -» o
rg en O
-r co •»•
CO O
o o o
o o o
o o o
-* CM
o o o
o o o
o o o
moo
O 0 O
o o o
CM
0^0
000
O CO O
•o
rM O O
(N *4 O
o •* -^
O r.
-4
0 O O
o o o
o o o
O O 0
o o o
O O 0
rg O 0
p •-* o
50 J5 0
co r-
f—
v: z
T Z trt
o a —
TC -X- T
1
f^j CO ^
o
o
o
-»•
i
^.
o
03
CO
v^
0>
•0
o
CM
^4
fM
o
00
in
•0
—4
•o
r4
CM
CO
-r
CM
o
rg
CM
(M
CM
CM
o
-t
cr
t ^
o
k-
?,
115
-------
t/1 -I
< <
CO t-
O> CO O
CO
o
Z
^^ ^
T O
r-- a
OQ
UJ1
<*l
2i
o
C/J
H
w
W -P QJ
rO U-
SUM
H QJ
(U Pi
Q ft
i
in -4
co
I
W n!
W
CU
o
J
en
ro
<
a
"?
o
•
I
I
I O
o!
UJ
X
U'
¥1
UJ
x
XI
131
o
01
Mi
I
UJ
o;
UJ;
4-
-4
IN
o
•a
0
^
•0
&
CO
r-A
V
u.
Z
f-
a
IN
^ o vn
rg .0 rg
r- r- -4
co 3d -o
IN IN
OOO
ooo
0*0
ooo
ooo
ooo
ooo
ooo
CO •*• O
ooo
ooo
O 00 O
INI
000
ooo
O ~* CO
IN
rg — i — «
ooo
0 CO O
o * •*•
*"*
o> ro in
ooo
1
z 3 *~
I ^* ?
j 1
1 H "*i *
O — * -^ O -J* *Q •?
•-4 GO -4 co in 10 rg
!
'
fN O CO *O * *O IN
3s *J" O -4 >t IN r^t
•-4 co f^ rg »n <—4 ^4
lf\ (*^ ^ -J/ IN (*1 r— 1
O O O O O O O
O O 0 O O O O
o * o o o •*• o
1
0 O 0 O -4 O O
O 0 O O O O 0
O O O rg co 00 o
-4 IN
O O 0 o 0 O O
o o o o o o o
* ^ o rg o * O
O O O -. rt O O
o o o o o o o
oj ^o rg *o ^" O1 co
-. -i -i n -0 rg
•H O -4 O -4 -< O
o o o o o o, o
•O O ^' •-*
o o o o o o o
rg co 051 rg -*• o IN
•-i o ^"' o* ~o 'N ^n
rg -4 rg ^
O rg ro' 33 in m ro
rg -< -< o -< r-i o
i
CO vf- O -O CO »N O
rg ^ o (^ -N —4 ^
CT> in 0; rr, |^ o, -•
^ 03 0 0> <> <> 0
oo -3 CT* f* rg -*-, (\j
O -^ -^ ^M O ^Ni O
co co r\j f1- h» ^«4 rsi
^ ^H U> O f*l i^ <7»
fi r^ ••J' ^ *~* f*J
. • ^ j ;
V >' ?: X >- ad at
o o S M o d zi
I T (D o (N
UA i/1
O Wl
IN — «
O O
O 0
•$• ^
0 0
o o
* rg
""*
0 O
o o
0 rg
-< rH
0 0
•O ^
rO IN
0 0
O O
•a o
-4 rg
•-4 r-4
0 O
IN (N
m m
in s^
O 0
-J- *
IN O
IN IN
*O f"l
n (M
0 -*
IN O
f- O
—* «-4
e£ >
> uu
_1 2
i_i a
—1
*
rg
0
pg
-4
»-4
5-
UJ
*
0
3
co
•r o
i- rg
CO sf
O CO
•^ ro
rO ~4
O O
o o
o o
rn o
O 0
CO -t
rg
-< o
o o
co rg
•Q -<
rg o
0 0
-.r co
o
-4 O
O O
0 0
•o
in -4
0 O
CO -0
IN ^
IN
m ro
-4 0
CO •O
rg in
•O -4
co in
•»• rg
IN 20
in «o
r>j -H
rg »-4
3- 3
LLJ r
J5 5
X Z
->•
r<1
CO
O
-0
o
o
o
r-t
o
m
_
0
•o
rg
O
•o
**"
—4
O
rg
in
n
o
PS*
CO
—4
^
o
o
IN
f|
O
"^
-0
•-4
a>
V
UJ
•^
a
X
o
o
^
0
o
o
o
o
o
o
o
o
o
o
.»•
o
o
o
o
o
o
o
o
o
o
0
o
•z.
12
jr
o
o
I O
•
o
o
rg
O
00
rg
O
rg
o
rg
in
rg
in
o
•
•*•
CO
o
CO
o
116
-------
— -
_
l/> _J UJ
< < nc
co K <
•O CO O *~"
^3 *— O
I/) UJ
I
— _
o
Z UJ
rj I/I < 0
£ XO uj
0 x
u — —
>^ 0 >- **
PQ Q 1— d
Z Z it
< ID <
UJ -t > a r-
2 3 o o
HUJ
I
fj
r-i UJ
M i
U a »»
a >•
Sj O (- UJ
fl u. z or
W J-=) <
£2 M < O h-
X a: 0 0
*— ' i_j UJ
K i
O »»
„ (J 1- UJ
r*-i uj z a:
O z o <
1_| -H UJ D »-
r1 -o o u
EH LU
^ j-
U _
0
A
\
0)
rH
i-Q
ITJ
EH
* rj
* a
* •
sO CO
o-l CT*
-4 <\
0 J
0 O
o •$•
-
o o
o o
o o
0 0
o' o
0 0
J- CM
(M -O
X> >»•
r^i O
rH &
"-* r-j
O O
O 0
o o
I
T.
U H
rr 0
O fc
1
kfl § 2
z 5
r-* OC V
10 UJ U-
00 ^ *
3 1 "
-H a- m
-O -H in
IM -O fM
f*~ N> r^
00 CO -C
i-~*
o o -i
0 O O
>-i r
ill
^ ^ tt
- n «
IS* -* -O
co m -o
o •* co
^•4 •& (\
rvj uS •— (
•4" c\j ro
O -^ O
o o o
^ r\4 ^5
"
o o o
0 0 O
O O 0
o o o
o o o
0 O O
O O ro
o o in
O O oc x
•i- > ct
05 rH <
i/i a;
o
o 3 „
a: oj
NOv.
5- > «!
X UJ >
O Z _|
-( 0 rH
-, X ^
1CT- O r-
»-l H
.*••»• O
oj •* m
rg -O IM
r^ in ^*
-H O 1*1
rH (M rH
O rH O
O O O
•o -o •«•
rH in
0 O O
0 O 0
0 O O
o o o
o o o
0 O O
h- O O
o o o
o o o
(M
fl
h- r>J U*
rH sf IM
JJ 0 00
O< 0 CO
N- o ro
IM rH
l/> UJ o
OC Z SE
< O _|
z: x uj
sr
-i i-
co < ta
or e£ oc
o u i-
> > LL
-J _J Z
•->•_< o
I/I I/I X
rvj m ^
rH ^f 0-*
ro r- (M
CO CO -^
^ >o 03
^ s^ ro
-H CO rH
rH CM O
o o o
•O - y?
UJ >- 3T
X 01 f
0 Z T
=> 0 0
00 X Z
1ft -O t*
*»• o
m o
00 -t
0
•0
rH
IM O
O O
o o
o
rH
0 0
0 0
0 0
0 0
o o
0 0
o o
o o
o o
M O
ro O
00 -^
o
m
^H
t-
o
-C «I
ra <
<±r
o o
z
>- r-1
UJ I/*
a —
x sr
oo o»
o
o
o
rH
^
^
rH
^
rH
rH
O
o
in
^
o
CO
o
rn
-O
CO
•O
rn
o
*
m
IM
pft
(M
rH
(M
in
rH
in
^
in
CD
CO
o
^w
z
CO
-J
^
r-
0
o
117
-------
z
V) _J
m K
co o
UJ
CL
u
UJ
X
00 <
in 1-1 o
VJ
CQ
CQ
P
CO
PQ
pf^l
W
w
rt
u
W
J-*^i
o
ffi
13
H
W
o
rf]
2
o
4/>
1
-r ^
o
> o
>- ti
—1 UJ >-
fY "^ _ I
0 or
m o c o
G. 2: Q
< a
r-
•? > O
T _) U..
S Q" .2
U' O •—
4-N1 2_ o <
t i Q_ rr*
^ Q
CC • f
0 Q Q
0
UJ
a:
i—
x
M
UJ
a.
I—
o
u
•^
r
uJ
cf
l
r-4
— *
O
o
0
0
o
o
-f
-1
•o
If*
-0
0
p-4
cc
NiJ
-r
0
0
fM
I
i
)
X
u
a
i
z;
«t
i
UJ
O
1
CM
,*
CO
CM
O-
CM
-H
O
^
CM
o
o
-r
o
eg
CM
c<"i
0
~o
CM
CO
CM
rM
•O
O
^>
r-
i
tr)
O
I
r-i
<*
•1
oi
o
>!
U)
g
-|
NO
CM
r>.
CO
CM
•*
0
CM
in
o
0
o
m
CM
O
o
-
h-
CM
CO
Nj
a
i
H
ON
-<
-O
r—
CO
0
0
o
o
o
o
r*-
0
CM
r-4
m
o-
o
a>
CM
^>
o
-o
ro
t-4
i
Q
a:
1
r—
h-
o
u
,
r-
1
m
a:
1
i
in
in
CNJ
^
•o
CM
m
o
^
i
o
0
o
in
ro
CO
*$•
-o
r-
->
CM
-H
CO
o
o
CO
1
a
a
fy*
uj
UJ
3
|
Kj
01
a
o
f-4
(M
O-
r-4
in
o
o
^
o
0
o
CM
—4
O
00
in
o
ON
cn
-J-
rsj
CC
a
o
*o
m
M
a
X
X
o
5
CD
^
UJ
Z
O
X
1
•a
i
-1
CO
0
.t
00
m
r*t
o
CO
CM
1-4
O
o
o
in
CM
CM
ON
"^
CM
(A
-O
CO
sj-
CM
0
CO
j
J.'
r—
CC
ra
a:
u
UJ
0
H
rO
.-
cn
o
ro
m
I
CM
O
^
CO
o
o
o
CC.
m
-o
r—
[^
m
•£)
CO
NO
o
o
o
CO
m
i
a:
uJ
2
^
J
z
7.
1
1
ON
CO
NO
f-4
CM
•*•
O
o
^~
n
o
-J-
fM
r*
no
-t
00
c
ro
*
O
CO
C?1
CM
O
CM
0*
1
a:
0
u
ra
i>j
_i
X
«H
C7.
J-
in
J-
-t
in
CM
CM
o
CM
f-
0
o
fM
"^
m
0
^
sj-
"
f^-
c^
^
•r
r-
-4
r-
in
aC
-i
f-4
*"
CO
ad
o
UJ
z
o
X
43
•-I
•O
NO
|
CO
CM
f-4
. f\
1
O
0
NO
f-4
0
o
o
;
i-
\ f-4
1
0
o
.00
1
CM
*
O
CO
'J*
o
fM
f*1
m
•
Tr"
CC
*
03
Q^
O
cti
J
J
-t
CM
CM
m
-^
o
0
>0
I—I
0
o
o
in
0
0
^J-
CM
CO
f-4
•M
in
aj
0
CM
X
t/1
a:
^
^
CO
cc
o
a!
^
1/1
f\J
-J-
-*•
NO
m
o
CM
CM
O
rM
f-4
r-4
O
O
^
NG
0
o
00
CM
CM
^
CM
ON
-T
o
CO
J3
-1
^
UJ
o
r—
&
O
a:
^
ro
O
l"1
j
CM
'o-
-<
!
O
,0
CO
o
,0
o
i
im
o
1 -r
•a
,""
r-
-1
^
o
to
o
~-t
•*"
1
CC
o
S3
UJ
5"
cJ
o
UJ
z
o
J
f-4
m
CO
»$•
-J-
f-4
r-4
0
NO
in
0
o
o
•o
o
•o
UN
CM
r-
f-4
CO
CO
f-
o
CO
o
c<1
f-
O
r— '
a:
re
o
UJ
X
UJ
o
m
in
•4-
i-
co
•o
•1-
fl
r\j
o
fM
O>
O
o
0
o
-1
NO
in
^
in
*
CO
C.I
f-4
CM
CXI
—4
X
PH4
fM
h-
QC
ra
Q.
O
UJ
O
X
O-
CM
^«
CO
d
r-4
0
o
CO
o
o
o
•fr
o
CO
00
r-4
CN;
CM
CM
u\
O
— 4
O
"•"N.
-<
j
I
r— 4
IN
1
Cu
ra
y
T
O
J
•*
ro
3>
O
NO
r-4
CM
O
0
O
rH
O
o
o
rM
O
CO
CO
r-
"1
CO
0
CO
1— 1
CM
r-l
NO
T
r—
^
0
ra
a:
U
UJ
6
X
CO
o
0
.*•
o
0
o
o
o
o
0
o
o
0
o
J.
0
0
0
<
r-
o
c>
2
Irt
f-4
o
o
•f
•*
o
CM
•J-
-3-
-O
a
I/O
CQ
i S
118
-------
Z i »#
01
DC
CO O, <
3
I/I
111
a:
CO' (/) < K
»•* O O
1 I S
UJ
ac
i O
I 2-
U
Ol
»• iN
M -«S
.•%!
D 0
—4
O
-M
r^
CO
\i
O
0
-'
•0
r—
CO
0
iA o
>A ^-4
.—4
IM rj
*-•, O
SO — *
o*
—4 CO
—4
CO -0
o •-«
,y% i-\j
lA ^
O O
^r
A
o
.M
O
O
"1
(V*
u-
ro
— *
O
f-t
,-1
n
-i-
-t
— '
o
vt IM ^
r~ rg; rn
t> ^r ~o
^ CO O
*t ra ^j
ro -• -.
O O O
o
°
J.
o
300
Q O
O O O O
oooooooomoooooooooo
oaooooooooooooooooo
O O O
rvj O O
o
•
o
ooooaooooooooooooo
H
J
CQ
<
EH
ffi
O
H
ffi
X CD
O 0)
H 4-1
CO '
a,
o cu
EH Q
ffi
EH
CM
W
Q
CD
iH
X!
m
EH
I
-t •
-o
•
o
o
I
UJ
a
u
ai
»»'
a!
tu'
3:1
o
01!
CD
00
•
0
o
0
0
0
0
o
0
o
0
o
CM
o
-1
o
•o
-*•
*
—1
•a
IA
0
UJ
v;
o
1
oooooooooooo
^MOOOOOOOOOOO
—4
1
(NI—lOoOO-- O>i> -^rvJOO'Ni'NrM
-OOf*1v.t'CT*fS«Jtr> lO^-HLO
*O ^O rH LO r~4 N- lf> ^-4 rTT|
00>o-Jocr,_Mr-rs4c-,^
^-.OOT-ro^^^^^^
1 ' 1
sO-Qnor-Jsoorsioo OfMO
^*-*fNJfv"stp COf^-CO%,f OQ «lT\ CO
(^ (^ ^ r^J (N t^ co O* r** 0^ co •$"
~4 i -t ~< rs| rvj — * *^ — •
! 1 1
O ^ t*" l-f^ ("^ VO QO fO rO P^ l^ ^
f\,( (^ O ro "~* t^J r*i f^ o *H O O
1 ! i '
•^ O f^J CO O to ^O *^ ® O O <^J
fs4 O *^ ^ to CO ^ CO -^ O -^" ^
CT1— — l>J<1>^^M^vc(ia;K:
ujujcocnujuj^:xuj>>>
si2'-i-(zznoz_J-i_i
§o:oiDo2i-io.-i.-i.-.
«-;H3;rcq^T<.l«rti«
1 i ' i i I ' 1 : •
rJrq^irt>Or^cqc'Cl'-l'Mm
I 1 1 II ,^-4^rt
O O O O O O
0 0 O 0 O 0
1 I
O — < tVJ O *o vo o !•
CM «^ ' f*>
rf> ^O f^ r\l NO O
o o o o o o
O CO -O O CO O
-O» O ro £?> >O
— < CM m "^
h^j t*- m csji r*- o
r— ^ *— 4 -^ r\| ^-4 o
1 i
i 1
O CO O IA O
COl t^ r-4. O CO
CM ~*
•&< «o O ^ CM O
O O -< 0, 0 0
co o -o co t> o
«O O IA CO- CO
*-4 ^M ^ t-4
UJ Jit ot ^ UJ *^
O 3 O O O ~
X CO X Z I jr
! : '
.$• l/\ O M CO O»
CO
o
o
1-
•
•1-
1A
CO
-t
•t
o
•M
119
-------
CO
H
O
CO
Q
13
3
QJ
r-H
o
U &-?
EM >i
O -P
•H
>H rH
EH -H
Ho
rM
h-l *H
H t!
m o
H J-|
Q W
O
PS u
W -H
CO
u c
H -H
CO M
3 -P
H C
Pn H
EH
S
M
•
r~~
1
0^
rH
rQ
rd
EH
2
— _/
VI «J
^ h*
CO N_?
7> CO l-i
3
1/1
_-__
O
Z
-H <•
XI
-H -1 O O
« _l CC
— UJ
IN
LU
cC. '
< !
H. NQ
O ro
LU -H
X rH
0
rH
LU
CC
1- 0
O >0
LU >$•
I
fn
o
LU i
ac i
<
1- -J-
o a
LU rH
*
i o
o
LU
CC
<
N. O
N->
'-
rH
* O
UJ
IX
>~ NT
U rvi
LU
X
~~ Nj-
o
aj
CC
r— rg
U -H
LU fNj
X
o
O
LU
oc
' ^
1- 0
o
'-U
i
'_
•o
1 °
1 OJ
! X
t- -a
o r-
LU IN
X
0
o
LU
CC
t- O
o
rU
, X
if
i UJ
3
-i
i
IN
•O
CO
CM
ON
ON
0
NO
rH
Nj-
O
^
33
CO
,$•
O
o
o
is.
.H
>"N|
in
IN
IN
J-
IN
O
rH
o
o
IN
03
0
O
-0
ro
IN
o
o
UJ
a
i
-.
NO
fNJJ
r»»
S3
IN
CO
o
NO
ON
ro
CD
0
0
co
ro
o
0
o
NO
„,
rvj
£
-O
rH
"g
r-
r-
o
o
-o
rH
,H
O
ON
^
f>4
o
Q
>f
Z
\
1
0
NO
r*»
CO
-«
o
,j.
>O
ro
O
xf
r\j
rH
O
o
0
•*•
o
•a
rH
CO
O
•0
•0
o
o
-o
rH
-1
O
CO
•o
o
o
M
rH
i
m
3
r-
|
j
in
m
IN
«H
•a
ro
_^
-1-
O
CO
_^
^
NO
CO
0
o
o
•*•
rH
0
NO
NO
ro
O
•o
rH
NO
0
NT
NO
IN
-<
0
_j.
N*-
o
(-,
f-%
i
•
o
•H
IN
N^
—
O
O
^.
0
r
CO
r»
o
CO
IN
ro
IN.
^
CO
CO
f«r
O
o
^.
o
IN
O
cr>
ro
rH
-0
rH
NO
rH
O
CO
IN
rH
O
•a
~j-
ro
IN
CO
CO
O
rH
UJ
o
r
H
i
ro
^H
CO
o
ro
in
-r
_J
IN
r-
•o
-"
IN
O
co
(7s
O
0
0
rH
_j
NO
in
ON
IN
CO
-t
m
rH
0
o
o
O-
rH
CO
o
ON
ON
rH
^.
X
CO
RDKNK
1
1
1
NT-
CO
NO
f-4
IN
O
o
,,
•0
^
IN
rO
^
O
o
o
o
o
CO
rH
rsl
o
CO
O-
0
O
CO
o
o
CO
ro.
u~\
•O
r-
j*
IN
1CHHL
»
i
^«
iTt
^
^
m
(NJ
-.
0
IN
ro
_
^
o
^
in
0
o
^.
•j-
rt
NO
in
•0
CO
O
IN
n
ro
o
Nl-
rH
"I"
O
o
o
IN
ro
rH
NO
m
Ul
0
i
0
NO
NO
CO
IN
—
-f
O
^.
o
IN
ON
O
IN
rO
N*-
0
0
O
NO
o
IN
S
-J-
M
^
rH
rH
O
o
rH
rNj
rH
O
CO
IA
0
^
o
0
^
ILVR--
J
•H
Nl-
IN
ro
—
ro
o
CO
IN
rH
NO
0
o
NO
IN
O
O
o
o
o
CO
ro
O
o
NO
O
o
•J-
IN
r-N.
O
CO
o
•H
IN
ON,
*J-
_J
rH
N
sr
>O
O
1
V)
mi
*"'
o
ro
IN
ro
O
O
^
in
o
NO
IN
O
O
3
r-
o
,r
ro
f
0
O)
IN
•0
IN
O
5
ro
0
j.
-0
f^
•0
o
CO
NO
IN
V
ai
o
J
r-
CO
Nf
NT
O
00
CO
rH
ro
O
CO
^
rH
O
0
o
NO
o
o
NO
IN
C7N
O
3
J-
O
O
CO
LO
r^
NO
rH
CNJ
Nt
o
CO
CO
rH
UJ
co
1A
Nj-
CO
NO
N^
ro
O
NO
ro
rH
in
•H
^f
0
o
o
o
CO
_,
(N
m
CO
•o
rH
rg
-0
r-
O
IN
ro
0
._,
IN
in
NT
CO
0
o
CO
ro
UJ
O
X
NO
ON
rg
NT
CO
ro
0
0
CO
.)•
o
0
IN
O
0
o
NO
-H
IN
ro
r-
Nj-
o
o
CO
-H
rg
Z
LU t/>
0 —
X f
CO CT*
p
a
o
-4
Nt
•r
•H
NT
-H
•a
k
h-
co
r
L
CO
h1
i
-^>
NO
33
CO
O
S5
rg
; -
rvi
JN
rH
CO
•Nt
O
V
o
IN
^
-0
33
C?N
CO
IN
NO
ro
•O
( •-*
0>
ON
; -O
; ,0
•Nt
; ^
0
rsj
0
IN
>n
C7>
O
0
120
-------
H
CO
ff)
m
D
co
>H .
CQ
CO Q
H ;
O W
CO
Q O
a c
U
UJ
X
i^ < H
— Q O
I < uj
z _i a:
o Q. <
• «£ O r-
I at u
O UJ
• UJ
Of/- -il QC
•O O l-l
O
•
rg
UJ
oc
ft
tu
X
vQi -t -3- Oi -* -t
-O rg -j- r«x m r»-
»J* f'J «-4 O iA O —« rO ^ *.f
CM ^a4J^*iA»-*flO--tCOtA
o ccirvj.jotsirstOcO'O-*1
o r^ *-* o rg
O -t O O
-4- roiM-r-*-Oft-OOO-O-O^CQ-O
i*M *NI *O t<> f*" iO M (<>
•in vi^ *si o* rsi ff\
^0 O pg
-* O en
O
ir>
oo co o
rj o
o
o
-*•
o
•
o
(M
•
o-
rg
ro
O
CO
*
o
rg
in
~O
•t
o
Pn -P
U -H
(0
(1)
-o
•
o
O
•a p-
rg • »•
O
i rg «H
n> o o
>»• rg
•o
r-
IM
irt
0
Oj
CD
! >!
UJ UJ
s *
1
uj CD en
z « -1
X >J
UJ UJ
3 §
o
1
X
u
*:
^
s
*\
co 01 o
VR_
i
o
3
KEY
HONEY
UJ *^*
Z ifl
o •-
x ar
00 O
I <
I I-
o
121
-------
z »«
«•» _J
3 O
I/) UJ
I
_____
34
O
Z ' 01
*" 4
— Q 0
z uj
I
o, **
Zi
1 < LU
Q z -J ^
K 00. <
"rj r- x. a H-
rij QC V_>
1-3 O UJ
Ou ^"
o ' — :- -
K CD
U > £
0) -• l/l LJ
2-H . '"-J J
0 Wo ' ~n £
-H
& £
W i ;S
S 1 — 1 "^" ' —
H-H r1 H
. -. Is- UJ
fa O x
tn
o T!
C f !
a to £
OrH 1 5
H CX01 *~
EH 0 3 tt
H) M x
CQ U
H |~
EH 0 * ^
W ^
H 4-> " Cr^-jO^iw-Jp Or- * vj-
M -fl -O — 1 LA -~4 "O ^^ CO lO O 'NJ "^ ."^ "O '**•
""* r-H
'JD "O (VJ N^J i%j \| o ~O *O ^ ' S3 M j") rsj -^ CQ
<^1 fNJ ^^ f*~ •— * C^ *4" O ^^ ^J" (N fO IA Q* •>!*• •O
•— * J* 30 CO sO rH 'O CO f\) LA r— i »_>4 o rO >^ ^
-H r>j fNJ r\j in NJ •-* — 4 i~O
I
oooooooooooooooo
ooooo ooooooooooo
O OOOO OOOxfOOOOOOO
mo^ma30r--H^,-g3>-001Ln-tLn
O-^OO.'-IINJ— irg-4— 100— 4O3-H
-^-^f\l'j'^t'O>>goot>g^-r\jO'4'O^O-4"
O ^N/'^J'O •OJ>"3 ^^"iO O-Or\JP-fNJ
-^vl-^^--^lOJ^^-C7lr--A'^•r^J>J''^J— *r—
coLT-Alm-f o^pgoo-orgoom-i
~4-4— "O^IOO^JOOOOOOOO
-OOTOJ^-OOXirgooo'NJOor^J-
"nOrosQ^o OLA *Q r-t mj-
^-4 •
m o^ o* pg ^ o ^ ^ o o o* m o o TO ^f
o."Mm— i—io-H.t- o^-^t-oo-Nj .0000-0
rg-Ovi"co,x) -O-O IALA p >4"r~
^-*n^"OLAr^ OO CO--4 cO'-t
-H •— * ^ —4
oooooooooooooooo
oooooooooooooooo
o -o o o o o o-i- o o -o o 0,0 o o
n —4
OOOOOOOOOOOOOOOO
oooooooooooooooo
1
o -j- -t o "o oopgooicoooioooo
(M -<
,
i
O-4OOOOO:— lOOOOOOOO
o oooo 00000,000000
i , 1
•Mpgooii o>i-cooo:cooooo-l-
!
i i
o pg •-* o,-o o -$• LA. -^ cg,rg m PI'LO -o m
ooooo cj^^pgr-^rHi-irolrgOLn
'
^ ' -o o o -o ^NJ pg o o o pg ^ *^J ' *%! ^j o
r*-t* ,p— fOp"-i-0coOLAOp"-^-oApg
'pg pg o'—* ^f o LA *O >r< ^H pg LA
i , «j-pg1— *mcg rH!^-4 pg
Sv>JJ>J^l>.a!aiaJ>Ju;>-
ujuJuJa9cnujuj^:xuj>>>uj^uj
S2z-i-(zza'JZ_|-J_jzoz
OOOCaiOOcct""(O<-**"«»'"'O^O
^2 a;i-jM3;roa^>^^icoT
r4rs f^i^Jrtfi^Of^cctO^O^Nfn^tAO
1 1 ! I 1 i 1 ^ i __ __i I . ^
1 1 1 1 i 1 1 ,«•(•-< «-l iH •-«««•**
122
3> •*• O O
rg ro o O
,—4
CO O si"
O .Q —4
—4 i— < r^
-J-
O O O O
O O 0 O
-TOO 05
|
>T o o pg
o -H o c^*
^4
O -^O O OJ
O ^0 -^
r-; -r o
c>
o o o x
0 0 O 0
—4
000 J-
co
o
O o O pg
O O o rg
rg
O O O ^3
iT\
>J-
o
O O O — i
o o o o
o o o o
! ^
1
!
i
O O O 0
o o o o
o o o o
LA
o o o pg
o o o ; o
1 '
o o o -o
; LA j- o >t
rg rg o r-
-t
0 0 •* -0
co pg LA
-< -H m
, -< -H pg
Si! Z
?:>--« • -1
«T OJ *^ <
5Z 1.0 t-
o •- o
Sis- t-
1
SJ CO O O
-------
u
>H
W
O
ffi
O
w
o
u
D
X
o
I
0)
rH
X!
Ck -J <
a C <
O v/> O
I I
002
2 2 -J
<-z r
O — <:
N- ^j li L/l
^01-
O < uj
-J- X. O-
Is*
. <
o <
rM _j o -J t~
, UJ
<
S
O
i/>
•^ (Nji-IC*ir\O-^fOOs<*
iN-O^-m-CO-coin
sO^Of^J-OfNjrxiOOC •£>*&
n~i f\j P"- ^** i»4 o*1 "4* o *™4 *^
t— * , Q"- CD co -o ~< °o r*^ ix in
CO— <0N^a?Of*^<>JNO'N
O "*^ O O ""^ fNJ *"^ t^J *** •—*
v$>
ooooooooOo
oooooooooo
aooo^-orsj'ooo
«~4
o 000,0— * o o o o
OOoOOOO'OOO
"M ~H
o'-ooooo'-o-
O:OOOOOOOOO
IN^.O.O^OOO
0 , - 0 0 , .0 0 0 ' 0 0 C
oooooooooo
1 i i
1 m -H r-
i
o -i o o o in -. i o -< • P-
OtOOOOOOOoO
IN ro
i
t
•^ ' -H •— * o "~* O *~* i *~* •"•* O
j 1
«O IM O >O ! -O O O i CO CC CO
1
, 1
1 i
§; i ^ *± i ; ^ _J
>; >J IJ>->-IT:X>
,-ic4c«('*i«i«ori'«qoiO
1 I i ; l 1 <-i
*Q ~t >*•
^)
-. 0 -"
CO O a;
^- f^ o
o o o
O 0 O
o o o
O o O
000
0 o -O
0 0 -«
O 0 0
CO O >!
000
o o o
000
— t O ff\
o o o*
-t -1- -»
ro IM r*1
0 0 O
-f •»• O
(NJ CO ^
*_< f-4
l^t >O vO
in •- IM
1L-S1LVR- 2504
1Z.-SILVR. 768
13..SILVR_ 1204
0 - *
•r, ro ^
(N CO CO
CT1 * >O
in »r r^
0 0 -«
o ^- o
f\j o 0^
o o o
o o o
o o o
000
Odd
000
coo
000
o *t o
C1 O O
o o o
O O IN
•t O o
O O _|
\0 O o
O> IN CO
(N *< J-
000
o M rg
o in — i
-H IN
cr in rvj
- M -5-
14_HONEY_. 876
15.BUCKEY 1168
16 HONEY 1968
0> .*• 0
IN m o
m
c-
o
0
•o
^
0
o
•o
*
o
CO
X
CC
o
CO
*c
tN
•*
o
•o
a-
•a
2
rg
cr
•O
-.
-o
CO
-------
Table A-ll. DISTRIBUTION OF ROW AND FIELD CROPS BY PHYSIOGRAPHIC REGION
Other
Row crops Field crops land uses
Total
Missing for
data region
Region Hectare % Hectare % Hectare % Hectare % Hectare
1. Glacial
drift
624 1.3 672 1.4 644 1.4 0 0.0 1940 4.1
2. Flood
plain
alluvium
888 1.9 688 1.5 416 0.9
0 0.0 1992 4.2
3. Glacial
outwash
terrace
560 1.2 332 0.7 332 0.7
0 0.0 1224 2.6
4. Ground
moraine
disected
5240 11.1 3148 6.7 1860 3.9
0 0.0 10248 21.7
5. Ground
moraine
undisec-
ted
9048 19.2 5052 10.7 2816 6.0
0 0.0 L6916 35.9
6. End
moraine
4940 10.5 3408 7.2 1632 3.5
7. Lacustrn 2160 4.6 1224 2.6
sed. and
org. De-
posits
1272 2.7
!. Missing
data
100 0.2 48 0.1 32 0.1
9. Subbasin 23560 50.0 14572 30.9
total
9004 19.1
0 0.0 9980 21.2
0 0.0 4656 9.9
8 0.0 188 0.4
8 0.0 47144 100.0
124
-------
^~ — •• •— •
E«
to — )
o z o
rV *-)
s
K 5 u.
O 2"
U -jE
in — <
J — — '
T i/> O
J
H
EH
O -it- z
X CD < —
f-\ < _, 1-
Q^ xt >— ' U- *I
^•N Z3 *— ' _l
V—) l/l C-
CM
Q
<~j j
•^ TD
pt O t—
*-^ ^
CH i/^
U
CM
0 So
CT U-
>H < u. z:
EH f^, ^~ "" ^
H - ^
i_3 on o
H
pq
^-
|_ fxj
xj •— '
JJ
x
1
a
I
5
1 1
5 S
CD 4
co !
r\j -^ cr
xO xO —1
CO fSJ xO
CM i- r-
0- 00 °=
Cxi (NJ
-4-10
OOO
xt x}- O
rsl r\j
O CO CO
— o o
CO fxl sf
oo a- M
xt rn -i
xO ro o
ooo
xt oo r\j
co -4- —
CM —*
fx. CNj xT
rt -<. O
00 CM O
CO LT* O
r- in CM
- oo .->,
OOO
o
5 i •-]
in o •"*
m ^4 co
r-4
CM r\4 o
— ' O x^
xO — * °^
CM in f^i
ooo
ooo
O ^ CD
co o r-
^ (M r^
st O (NJ
^O sO O
CO ^ Is-
CO ^J" r^
ooo
^- og o
•O CT1 r\j
ff, — t T-*
4~ O O
-"<""» r,
O fNJ sO
.4-
m xc ^
-4 xt -•
rg vO ^
tr ir —<
in -< c^
CM
>VJ — 1 Pxi
000
xT sO O
o in r~i
^-j — *
J '
a M ;
004-
1 i _
f- m
r<*\ O ^
0 0, 0
00 xt O
xt 00
,— )
o in ro
•vt in ^J
o 0
^ ^
0 ^ N
r-* *-* Q
f\J O
— 1 V)
_l O
J * Z
P 3 ^
it X. UJ
So z
«-l O
LU
a: o: z
< < o
SB Z X
t 1 1
ea -J ^-
< CD <
oJ oc o:
ooo
> ^ 5
J -J -1
l-t t-4 t-t
1 "i "l
*^ ^ ro
H -1 -H
O -* *
ro rri h-
CM CO CO
&> xt xO
CO xt1 xt
—4 i— t CO
O -" CM
ooo
O CO CM
CM h-
in -*• in
00-*
0 -0 xt
M r- -. LU >-
ai ^ uJ
z o z
030
*! *, r
^ lA O
O1 "J1 O
f\i <^> O
•^ °° «*
00 0
n ^
^-4 >— *
o — • o
ooo
CO -3- O
•4'
•4- o o
0 — 0
O TC O
O CC'
r\j sj-
-< r" 0
ooo
CM xO O
f. IT,
•— J
-4-0
ooo
xO C^ O
CO xt
CO O O
ooo
xt O O
CM —
»-4
O ^ O
CM —4 O
cc rj
-------
,J- CM
— r\j ^>
OO _J UJ
< < a:
CO h- <
co 3J o H- ,0 co
!D h- O r^ fM
CO UJ <—t 0s
X — i fsj
i O ^
o , ~> o
Z < UJ
u*J < <
PN- uo Q t— o >o
1-1 O NO — '
(N s: uj .t Nt
^ ^
w . — _
£3O en o
0 'I i * o -
U 1 < OJ
H z_j a:
n. ,. 0 Q. <
L1 NO z o ^- Nt co
>H - or o NO co
EH O o -u -• •»•
O x
co m —
W II ! - o
U P-H - o oj
H uj
EH co a
UA <
LA >— CM CO
ri] ^-^ O m oj
LJL Lj ^ N>
ft (0 ~.
0) —
Q >i „ "»• °
13 \ o-.
e£ d) ^
3 S ? 5
ft 0 -, 4- £ -J- Nt1
O (Tj 0 o co
K \ ^ CM Nt
U to T
••§""-; - -
Q -P • o -
(V N— ' "J
< •* <*
tN] CO en CM i— NO CM
« . .
H -H oo
wo „ • S
o to 7 <
fV OJ -1 >-< O O
rn o-t <-> M
W TS UJ1
0) , X
3 2 < i Z1 ° °
r~"1 C i ( P* * •
EH-H ! u : o o
X -P - ! oj i
W CO l i < . !
EH W -10 i ^ , °' °
o 'IS!
— ! ! ,M
.! o
(0 -"
ro x
_, U H-
1—1 H- ac
1 ~
gf t O ^9
"^S
Z Of
cy < o
r-H "• Z
^s z , a >
jQ — LU UJ
fTj N« N,f Z
r_i < ' U O
bH co < X
eo ' :
:D -. oj
^ ' !
-1
NO
OJ
f-
CO
OJ
CO
O
>0
fJN
ro
CO
O
o
CO
m
en
^
O
O
NO
OJ
CM
-t
CM
O
r-t
0
^
OJ
r«-
*
o
o
0
0
o
o
^
z
z
:>£
O
oc
CO
CO
UJ
z
c
X
m
tj>
_4
NO
CO
-4
O
CO
NO
m
o
OJ
— •
LO.
O
o
CM
OJ
r-
o
CO
CM
ro
m
O
CO
OJ
"•
o
o
CO
o
0
o
£3
CL.
t-
h-
O
o
to
^
CO
_M
at
~t
LA
lA
OJ
NO
OJ
rr,
_J
v4-
OJ
-O
CO
_4
NO
CO
r—
o
-t
m
o
r-
Nt
NO
O
-t
o
°
0
o
o
o
o
o
^
o
at:
to
•— i
UJ
;j
^
•c
^i
(K
IA
0 -i
—1 CO
H
oj O
0> Nt
-4 CO
LA C^
O f-
0 0
!
Nt CO
CM
en
1
o r-
oj1 -<
oi co
NO tO
c> r^
co «t
OJ -1
oj o
m Nt
cnj NO
-"
*4 CM
m oj
o o
NO NO
-t o
-1 -4
O OJ
en CM
0 «t
O OJ
•T' O
j :
C3 0
1
NO O
Q 0
c3 o
J o
ISi
^j!
o .»•
X
X I-
01 a:
< ra
CO Ct
< u
> >
LU 01
Z Z
0 O
X] X
•O r^-
CO
^
^
CO
O
en
IA
Nt
i-4
$•
•o
rn
"
0
IA
""
-
rt
o
-t
^
-:
•
o
CM
m
o
o
0
oc
UJ
z
(_,
_ _
ra
u.
z
*
o
Of
CO
CO
o>
CO
-O
CM
*
O
0
Nt
NO
~i
OJ
l«-
NO
O
NO
CM
NO
CM
r^
OJ
-1
CM
«t
CO
-O
£
o
o
o
0
o
o
er
Nt
cc
o
o
ra
r^j
_l
X
X
O
o
*
IA
-t
LA
oj
-H
O
CM
fn
r-4
•H
0
IA
-.
CM
OJ
fN»
NT-
en
CM
_i
^3
CO
O
CO
CO
en
0
o
0
0
o
o
CL
>
_J
•™»
to
CO
«7
a:
u
UJ
z
O
0
NO
NO
CO
OJ
m
Nt
O
CO
Q
OJ
0>
o
CM
Nt
NO
-I
O
NO
^N>
-•M
}
O
•*
-t
_4
O
NO
NO
^
O
^
CM
O
O
o
X
L/l
Q^
*a^
X
CO
^
a:
o
a
j
~
•*
Nt
OJ
OJ
en
r-4
-"
en
0
co
CM
NO
O
o
CM
CM
o
CM
»^
-O
0
CO
CO
CM
r-
o
Nt
CM
rn
O
O
O
CM
0
O
X
10
<
3!
_.
CD
ev
0
at:
_
i/
CM
-1
Nt 0
Nt cn
NO OJ
iA O*1
O >n
OJ — •
-i O
O O
O 03
o
O LA
-4 O
CM NO
LA -J
-t CM
.! r.
O 0
r\j co
iT\ O
og m
r\j r^>
<-4 •-*
-O -! *
UJ O
Z JE
a _j
X LL
i jr
h«i
UJ
-J Z
•- o
i/> x
en' -t
-1
en
CO
Nt
-t
-1
sr
o
CO
CO
_4
en
o
CO
-4
r^
o
CO
ev
en
O
CO
Nt
NT
r~
o
o
CNJ
en
o
o
NO
°,
o
o
K
NO
H-
a:
(5
L)
UJ
U.
o
-5
CO
IA
*
r-
m
-0
Nj"
«
^
C)
o
f)
—4
in
— t
S
N-
CM
0,1
^J.
,t
O
rt
O
OJ
CO
c>
en
_,
-o
en
NO
o
o
o
o
o
o
f~t
m
CM
1-
or
fB
ct
0
>-
LU
Z
0
NO
N>
OJ
Nt
X)
rn
"^
o
O
CO
«.
o
o
o
OJ
r-
0
CO
0
en
CM
Nt
CO
IA
-O
O
o
NO
OJ
_4
O
^
CM
,
0
0
m
«-4
CM
^_
or
re
~JL
^
T
a
i^
-<•
m
CO
o
NO
-*
CM
O
O
o
^4
cr
o
e\
Nt
^
_,
NO
m
IA
LA
0
~t
CM
NO
O
CO
CO
CM
o
o
CO
0
0
o
X
f—
—
o
r
re
^
UJ
z
i
CO
0
0
^t
o
o
o
o
o
o
o
o
J.
o
o
o
o
0
o
o
o
o
o
o
0
^
^-
«T
O
O
(/)
o^
5C
O-
O
0
o
—4
^
,t
r-4
5
-4
CO
o
Nt
CO
en
CO
CO
_4
-O
CO
CO
LA
O
CM
O
NO
NO
C-
0-
CM
O
NO
—t
2
CM
^
CM
^.
CM
•T
-t
0
NO
o>
^^
0
o
0
l/>
z
to
«I
CO
^J
1
«r
^
>_
O
o
CM
126
-------
in
•^
(N
II
U
EH O
O
en rn
W II
U J
H
EH
U
(d
0)
Q
<, Q)
^ H
P-i U
O fd
oa i—
CO CO O
z <
to o
o
z
I <
a o.
>o z o
co
A
CO
I
•J- -t
<
oj
X
uJ
a:
<
h—
o
LU
X
»*
UJ
•^ ! -t o
i "i V:
00 fM
ro (V h-
-i CM! CM
C
-o
-*•,
ro O oo
O rJ O
-1- CO O
-O OO 00
•-*>}• ro
CM CO
-I OO
in
3 -^ in
' i
i !
J 'O fSi
i— i
«-4,
1
*•
CO <1| -H CO
in i-o
i 1
O; —i ^
roi r\J in
in! xj-: CM
CM, ro
in (70 .4- \o
o ro' *t *r
CM -^; ^ ro
oo o
ro O
•4- 00
*o CM 'CM
ro
-H; o! o
CM' ,4- (M
f~ I CO,
o o; o
O CM O
o o o
CO CO
O CM
CM rf
CO CO -O
oo ro
co o o
o
ro oo o r-
o H fM
a-
o
o in'
•4- -r o °°
CM ^O «O CO
•-i CO 0^ ^~
O (M O
oo ro •*
CT> r- in
CM o
co o
•t CM
r-l O O ^^
OO -4-
^ o
^* r^ O
o o o
O CM O
O ro
CM -T
o m ro -^
o o o o o o o
•4- O
o o
*H CM CM O
o o o o
-J- -3- •-) cc
sD ""O ^ CO
CM
•O rg OD
i-t in rM
~* ro
CM O
"- "M
vO ~O CM OO
in r- —i o
o o — o
*O CO O CM
ro rM •*> ro
-J CM 0s *J-
O O
-O O CM
Qi O f~
ro CM in
O fi
r-> O
ro ^ •)• -t
O O O _i
fM -1-
-«•
•4" CO *4" O
•a •£> oo -t
~4 ~* ~* m
-o o
O O o
o
o
I 00
I co
CO
•
00
50
CO
.o
cr-
o
o
a-
U W _
•• 0
Q 4-)
< 0)
ffi O
•r
i
ro CM
»f
UJ
a:
o
UJ
•o
\f\
CM
fM
O
(VI
o
O
CM
ro
O
-O
CM
ro
O
ro
CM
CT
CM
in
o
CM
•0
CM
M
O
-O
o
in
o
00
O
in
o
fM
IN
in
CM
a-
in
-o
O
o
o
O H
H -H
cn o 7
O 10 fM-<
W t!
CU
-P —•
H g
EH -H
S H->
W W
EH H
O
04
0)
rH
X!
fd
EH
«
cc
LU
X
LU
a:
LU
I
30
§
a
o
§
UJ
v:
u
Ul
^
ro
O
O
1
in
ro
in
-4-
00
O
co in
o -*
CM «O fM
^o in in
ro
CD
-4
CM
t-
a.
ra
*?
O
^)
o
ro M
O
O
X
h-
O
3T
ra
a;
o
a
i
o
o
ej
Z
CM
r-
ro
CM
•I
03
o
t-
o
127
-------
'--
r-
•H
J
H
Q
Q
S
>H
Qi
D
O
ffi
H
fe
H
^
12
w
S
w
PQ
0
cn
H
^
n
&
0
u
£.
o
H
!
H
M
U
§
PM
, |
1
pa
0)
rH
EH
•H
M
<
U
r— 1
Jd
>i
rd
p
&-t
XI
0)
PT.
H-4
>1
J_l
(0
C
rO
^
Q
>1
rH
t )
rj
0
K
">i
rH
1
•M
r^J
a
rH
^
O
K
*
>1
•H
rd
n
>1
r-H
o
*
I>1
rH
'id
Q
rH
3
O
K
•X
>i
rH
rd
D
rH
rHrOrH 0 i-HrH rHCO
• •• • •• ••
OOO 00 O O
in in 'S1 rH cr> CMCM '3<'«3r-00
-------
>1
|H.
O
ffi
-P
w
3
tn
< >i
rH
•H
(0
Q
n
3
0
ffi
3
>i
rH
•H
rd
P
rH
3
0
ffi
Q)
d
^ >i
rH
•H
(0
P
^~. rH
I \ ""i
-P r->
£ o
o w
o
"" rt
s
rH
1 ^
ffl f-H
•H
(U (d
rH P
ft
EH
H
rd
PI
oo yD ^i* uo
o o o o
rH <* VD •<* CN •=!<
o cTi oo o o r~
. EH • • • • •
o o o o o o
CN ^ CO
. • .
0 O O
rH rH UO
00 CN 00
• . . £H
00 rH
in rH
• •
O 0
oo r~ CN
H ^ O
EH •
0 O 0
rH rH rH rH
O O rH O
O1! O rH
rH O H rH O
. . EH • • EH
O O O rH O
rHCNOO^in^r^COC^OrHCNOO
rH rH rH rH
rH rH 'tf' CTi
O O O CN
00 rH CN rH CO
o o o r- oo
• EH EH • • • •
O O O O CN
"31 CNOOU3«3< CNrH VD
• .... *• >
0 OOOO 00 00
in vo v,o in ^j* ^i*
rHrH CNOCT^'CJ' i— IO CO
EH . . . • . '. £H • 'EH
00 OOOO OO 00
"^'tfOO rHCTi CNin ^
..* t» •• •
OOO OO OO 00
OOrHCTlrH VDCOOO O V£> CTl
rHOO^DO CNrHU3 CNO O
. . EH • • •
OOOO OOO OO 00
i— IrHCN1^ rHV£>CNCNOO
OOOrH OOOO'*
CN oo oo ^c CTI co in
CNOr~-(Ti O O^CNCN
EH • ... EH • EH ...
oooo o ooo^r
uy
^
EH
*3r--cocriOrHcNoo'3'ink£>r^cocriOrH O
rHrHrHrHrHrHCNCNCNCNCNCNrvlCNCNCNOOOO EH
130
-------
NO DATA
O O rH
EH ... EH
000
00
CO
ro
O
CN <0
CN O
O O
00 rH (Ti
• • *
O O O
^ CN rH O O\
rH O O rH O
EH • • • EH •
O O O O rH
rH
!H
o
NO DATA
EH EH
o
o
CN
O
EH EH EH
rH CO O
O rH CO
• • •
O O O
O
K
-P
U
O
•H
Q
co
o
rH O
| j Kj1
O O
CN
O
co -^
• •
o o
rH CN
• •
O O
o o
o
CN
CM
CN CN
** CN
O O
LT) V£>
CN O
O O
O 04
*sT rH
• •
O CN
-p
a
o
u
t
-
i
M
(U
o
ffi
-p
04
0)
en
<—> <_J
t^i
rH
•H
Q
CTi
0
0
0 O
IT)
O O
O
^3* O
O O
in CN
o o
O
o
CN
O
O O
CO rH 00
rH LO O
• • •
o o o
co
CO
r~-
•
CN
iHrHrHrHiHrHiHrHrHCNCNCNCNCNCNCNCNCNCNrOCO
rd
-P
O
131
-------
en
rH
K. M
r-J
H 0
p-H , M-f
Q "^
U O
Q {H
^ S >,
S-j r~~'
rf rt
g Q
O
is
H .
H ^1
* 0
a H
W o
S "S
EH ^
[y
rH
i^ -H
?~\ "J
S Q
CO
H
o 7|
u ^
2; °
CD «
H *
^ |
M
^^ r~H
1 S
CN
|
« ^
0) Q
rH
oo^ inrH cNoocNr~- oom in
• • •• •••• •• ,
OO OO OrHOO OO rr
VDin rH'J1 CNrHOOCNrH ^TOCTi
OOCN IJ3O CNCNCNinoO CTlCNCO
00 OO rHOOOO OO^j.
oo CN rH rH in in r-
0 0 O 0 O O rH
EHEH-'-EH -En-EH EnEn ...En
OOO OO OOO rH
rH OO
• • i — 1 i — 1 i — 1 CT>
0 0 • . . .
O OOO
i— 1 OO rH CO r~-inrH CNCNCN OO CNCNOOCOOO CO
OOO O CN-^OO O OO O OOOrHO O
OOO O OOOO O OO O OOOOO rH
rH
rd
rHrHrHiHrHrHrHrHrHrHCNCNCNCNCNCNCNCNCNCNOO O
EH
132
-------
Table B-3. COMPARISION OF PRECIPITATION RECORDS
FOR MARCH, 1976, BETWEEN TIFFIN
DAILY AND FREMONT HOURLY RECORDS
Tiffin Fremont
March 1 .05
2 T .30
3 0.22 .07
4 0.79 .91
5 0.68 .30
6
7
8
9
10 0.13 .12
11 T
12 0.14 .05
13 0.01
14
15
16 0.17 .24
17 0.14
18 T
19 T
20 T .17
21 0.35
22 T
23
24
25 0.09 .15
26
27 .17
28 0.25
29 .17
30 T
31 0.15 .15
3.12 2.85
133
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1 REPORT NO.
EPA-600/3-79-054
3. RECIPIENT'S ACCESSI ON- NlO.
4. TITLE AND SUBTITLE
Nonpoint Source Model Calibration in Honey
Creek Watershed
5. REPORT DATE
August 1979
6. PERFORMING ORGANIZATION CODE
7 AUTHOR(S)
T. H. Cahill, R. W. Pierson, Jr., and
B. R. Cohen
8. PERFORMING ORGANIZATION REPORT NO.
9 PERFORMING ORGANIZATION NAME AND ADDRESS
Resources Management Associates
P. 0. Box 740
West Chester, Pennsylvania 19380
10. PROGRAM ELEMENT NO
1BA609
11. CONTRACT/GRANT NO.
R-805421-01
12. SPONSORING AGENCY NAME AND ADDRESS
Environmental Research Laboratory--Athens, GA
Office of Research and Development
U.S. Environmental Protection Agency
Athens, Georgia 30605
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA/600/01
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The U.S. EPA Non-Point Source Model has been applied and
calibrated to a fairly large (187 sq. mi.) agricultural watershed in
the Lake Erie Drainage basin of north central Ohio. Hydrologic and
chemical routing algorithms have been developed. The model is
evaluated for suitability as a land management tool in estimating
the reduction in sediment associated phosphorus from cultivated land.
Results indicate a very different impact for selected best management
practices (BMP) estimated using the Universal Soil Loss Equation
(USLE) gross erosion approach. Modifications to the model structure
are suggested to improve BMP evaluation.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Simulation
Runoff
Water Quality
Planning
Land Use
b.IDENTIFIERS/OPEN ENDED TERMS
Nonpoint pollution
Model studies
COSAT! 1'ield/Group
48G
68D
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
20. SECURITY CLASS (This page)
UNCLASSIFIED
21. NO OF PAGES
__
22 PRICE
EPA Form 2220-t (9-73)
134
------- |