EPA-600/2-79-050C
July 1979
MAXIMUM UTILIZATION OF WATER RESOURCES
IN A PLANNED COMMUNITY
Application of the
Storm Water Management Model
Volume I
by
Elvidio V. Diniz
William H. Espey, Jr.
Espey, Huston and Associates, Inc
Austin, Texas 78704
Grant No. 802433
Project Officers
Richard Field
Anthony N. Tafuri
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed by the Municipal Environmental
Research Laboratory, U. S. Environmental Protection Agency, and
approved for publicsition. 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 recommendation
for use.
11
-------
The Environmental Protection Agency was ^eated because of in-
creasing public and government concern about the dangers or
pollution to the health and welfare of the American people.
Noxious air, foul water, and spoiled land are tragic
testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay between
Its components require a concentrated and integrated attack
on the problem.
Research and development is that necessary first step in prob-
lem solution and it involves defining the Problem, measuring
ill impact, and searching for solutions. The Municipal
Environmental Research Laboratory develops new and improved
technology and systems for the prevention, treatment and
management of wastewater and solid and hazardous waste
pollutant discharges from municipal and community sources,
for the preservation and treatment of public drinking water
supplies and to minimize the adverse economic, social, health,
and aesthetic effects of pollution. This publication^ one
of the products of that research; a most vital communications
link between the researcher and the user community.
This project focuses on methods of maximizing the^use^of water
resources in a planned urban environment,_while minimizing
?heir degradation. Particular attention is being directed
towards Ltermining the biological, chemical hydrological,
and physical characteristics of storm water runoff and its
corresponding role in the urban water cycle.
Francis T. Mayo
Director
Municipal Environmental
Research Laboratory
111
-------
PREFACE
The overall goal of this research was to evaluate the water
resource plan for The Woodlands, Texas, and to make recommend-
ations, as necessary, to maximize its effective utilization
through alterations in design and management. Any recommended
alterations were to be critically evaluated as to their compat-
ibility with the natural environment.
Collection and utilization of stormwater runoff for recrea-
tional and aesthetic purposes was a major feature of the water
resources plan at The Woodlands. Control of downstream flooding
was also of great importance and so storage reservoirs, in the
form of recreational lakes and wet weather ponds, were created
by the developers. Water quality was a concern if the impound-
ments were to be aesthetically appealing and/or suitable for re-
creation. Therefore, a major sampling and analytical program was
designed to monitor water quality and quantity at different loca-
tions in the developing area. The Storm Water Management Model
(SWMM) provided the focal point for combining the water quality
and quantity data into a predictive tool for design and manage-
ment purposes.
SWMM was oricririally developed for highly urbanized areas and,
therefore, was calibrated for this project in an urban watershed
(Hunting Bayou). Subsequently, SWMM was modified to model runoff
and water quality from natural drainage areas, such as The
Woodlands. Because of the lag in the construction schedule at The
Woodlands, the dense urban areas were not completed during the
project period. Consequently, Hunting Bayou and other urban
watersheds were sampled to provide a basis for predicting pollutant
loads at The Woodlands in the fully developed state.
Water analyses included many traditional physical, chemical
and biological parameters used in water quality surveys. Patho-
genic bacteria were also enumerated since the role of traditional
bacterial indicators in stormwater runoff was not clear. Algal
bioassay tests on stormwater were conducted to assess the eutro-
phication potential that would exist in the stormwater impound-
ments. The source, transport and fate of chlorinated hydrocarbons
in stormwater runoff was also investigated.
Several of the large Woodlands impoundments will recieve re-
claimed wastewater as the major input during dry weather. Besides
IV
-------
a concern about disinfectant toxicity to the aquatic life in the
lakes Consequently, comparative fish toxicity tests were con-
ducted wi?h ozone and chlorine, the two alternatives available
at the water reclamation plant.
Porous pavement was considered by the developers as a method
for reducing excessive runoff due to urbanization and an experi-
mental parting lot was constructed. Hydraulic data was coveted
and used to develop a model compatible with SWMM, to predict the
effects of using porous pavement in development. Water quality
chancres due to infiltration through the paving were also deter-
mined.
Hopefully, the results of this project will contribute in
a positive way to the development of techniques to utilize our
urban water resources in a manner more compatible with our cher-
ished natural environment.
v
-------
ABSTRACT
Stormwater runoff from urban areas has been recognized as
one of the major contributors to pollutant loadings in natural
rivers, lakes and estuaries. To evaluate these loadings, char-
acteristics of stormwater runoff from an urbanized area and an
undeveloped site were quantified for several key water quality
parameters at selected sites.
The Storm Water Management Model (SWMM) was modified to al-
low for modeling of 1) separate sewer systems, 2) effects of ur-
banization on baseflows, 3) performance efficiency of natural
drainage systems, 4) cost efficiency of natural drainage sys-
tems, 5) four more water quality parameters - COD, Kjeldahl
nitrogen, nitrates and phosphates, 6) hydrologic effects of
porous pavement areas. A new subroutine was programmed for each
of these objectives and included in the SWMM. All new subrou-
tines are user options.
The resulting SWMM version can model storm periods separ-
ated by zero or low rainfall, eliminate all dry weather flow re-
sulting from sewage, compute baseflow recessions, model flow in
natural nonuniform cross-sectional channels, determine costs of
natural drainage systems, model eight user selected water qual-
ity parameters for as many as 20 land uses, and evaluate the
performance of porous pavement.
The SWMM was applied to the urbanized Hunting Bayou water-
shed in Houston, Texas, and Panther Branch, an undeveloped
watershed where the new community of The Woodlands is being de-
veloped in its downstream reaches.
Because the original water quality predictions from SWMM
verification and calibration runs were too low for urbanizing
areas, a user option to input the actual loading rate and re-
moval factor for each pollutant under consideration was intro-
duced into the SWMM. This approach was verified by comparison
to observed data on Panther Branch and Hunting Bayou and then
applied to Swale 8, one of the tributary areas to Panther
Branch, which is currently being urbanized. A management strat-
egy to control storm water in a manner compatible with the
natural environment and proposed natural drainage network was
developed for Swale 8.
This report was submitted in partial fulfillment of Grant
No. 802433 by Espey, Huston and Associates through Rice Univer-
sity under the sponsorship of the U. S. Environmental Protection
Agency. This report covers the period from September 1, 1973,
to September 1, 1976, and work completed as of September 1, 1976,
vi
-------
CONTENTS
iii
Foreword iv
Preface vi
Abstract ix
Figures xi
Tables xiii
Acknowledgements
Section
" 1
1 introduction ... 3
2 Conclusions " 5
3 Recommendations g
4 General Project Information . . . . .
The Storm Water Management Model ^
Study Objectives ' 12
Study Approach | 15
5 Study Area Description 15
Hunting Bayou Watershed lg
The Woodlands Development 2Q
Swale 8 Watershed 2Q
Data Collection 24
6 ™"%£&l fepa^ sio^ «t«s^« J^ '• ' ^
interaction between ground water conditions ^
and surface drainage 33
Infiltration • 35
Costs of natural drainage systems ;•;••• *
Area-discharge data for natural sections . . . J/
Modeling of porous pavement • ^^
7 Water Quality * 53
Data analysis • • • 84
Water quality modeling 8g
8 Model Application 89
General considerations 91
Hunting Bayou 113
Panther Branch 125
r^TO »••••*******
Existing and future"development modeling for ^
Swale 8 " 162
9 Summary ".".".".... 165
References
vii
-------
CONTENTS (Continued)
Appendices
a) New subroutine source code listings
b) Revised imput coding instructions
c) Storm data summaries
d) Hydrocrraph recession data
e) Sample output
Vlll
-------
FIGURES
Page
Number
1 Master programming routine in SWMM 10
2 Study area vicinity map 17
3 Hunting Bayou Watershed 19
4 Panther Branch Watershed 21
5 Swale 8 Watershed _• ' ' ! 7fi
6 Baseflow recession curves at Station P-10 •••••• ^
7 Slopes of hydrograph recessions at Station P lu . . . ^/
8 Slopes of hydrograph recessions at Station P-30 ... ^
9 Subroutine BASFLO system logic * -,'v/ ' * "
10 Baseflow recessions at Station P-10 computed by
Subroutine BASFLO • • ' ' '
11 Baseflow recessions at Station P-30 computed by ^
Subroutine BASFLO 36
12 Subroutine CSTANL system logic
13 Normalized area - discharge curves
14 Modeling of natural cross sections
15 Porous pavement and surrounding drainage area . . . . «<*
16 Izzards dimensionless hydrograph for overland flow. . 44
17 Triangular approximation of evaporation
18 Pavement cross-section and modeled flow
19 Subroutine PORPAV system logic 56
20 Porous pavement test area ,****" co
21 Design storm rainfall and computed hydrographs ... 5b
22 Storage volumes in porous pavement - high ^
permeabilities
23 Storage volumes in porous pavement - low ^
permeabilities •
24 Water quality in porous pavements - COD ana
Kjeladahl Nitrogen - . • • • • • • ' '
25 Water quality in porous pavements - Nitrates ana ^
Phosphates 67
26 Unit area discharge relationships . . . -
27 Temporal relationships of suspended solids to ^
discharge • • - - • • • 73
. --, e- J_. -I ,-* w *-* +- -V IT n f~\ T" T* _...••• ' —'
28 Nitrate yield as a function of runoff . - - -
29 Pollutant yield as a function of peak discharge ... 74
30 Water quality relationship - Suspended Solids .... /b
31 Water quality relationship - COD . . .
32 Water quality relationship - Kjeladahl nitrogen ... 78
33 Water quality relationship - Nitrates
34 Water quality relationship - Phosphates ''•'''' R9
35 Total pollutant loadings as a function of runoff . .
IX
-------
FIGURES (Continued)
Number Page
36 Infiltration loss rates 90
37 Subcatchments and drainage network - Hunting Bayou . 95
38 Hydrographs at Stations H-10 and H-20 103
39 Hydrographs at Stations H-10 and H-20 104
40 Hydrographs and suspended solids concentrations
at Station H-20 105
41 Storm of 5/08/75 - Station H-20, Suspended Solids . . 109
42 Storm of 5/08/75 - Station H-20, COD 110
43 Storm of 5/08/75 - Station H-20, Nitrates HI
44 Storm of 5/08/75 - Station H-20, Phosphates 112
45 Subcatchments and drainage network - Panther Branch . 114
46 Hydrographs at Stations P-10 and P-30 122
47 Hydrographs at Stations P-10 and P-30 123
48 Storm of 12/05/74 - Stations P-10 and P-30,
suspended solids by original SWMM version 126
49 Storm of 12/05/74 - Station P-10, Suspended Solids. . 127
50 Storm of 12/05/74 - Station P-10, COD 128
51 Storm of 12/05/74 - Station P-10, Nitrates 129
52 Storm of 12/05/74 - Station P-10, Phosphates .... 130
53 Storm of 12/05/74 - Station P-30, Suspended Solids. . 131
54 Storm of 12/05/74 - Station P-30, COD 132
55 Storm of 12/05/74 - Station P-30, Nitrates 133
56 Storm of 12/05/74 - Station P-30, Phosphates .... 134
57 Subcatchments and drainage network - Swale 8 .... 138
58 Storm of 4/08/75 - Station D-50, hydrograph 143
59 Storm of 4/08/75 - Station D-10, hydrograph 148
60 Storm of 4/08/75 - Station D-10, Suspended Solids
by original SWMM version 149
61 Storm of 4/08/75 - Station D-10, Suspended Solids
and Phosphates 151
62 Storm of 4/08/75 - Station D-10, Nitrates and
total COD 152
63 Station D-10, future development conditions,
Suspended Solids 153
64 Station D-10, future development conditions, COD . . 154
65 Station D-10, future development conditions, Nitrates 155
66 Station D-10, future development conditions,
Phosphates 156
67 Station D-10, Runoff Hydrographs - existing and
future conditions 159
68 Water Quality from different land uses in the Swale
8 Watershed 160
x
-------
Number
TABLES
Page
1 Modeling Data Requirements by SWMM ^
2 Storm Event Hydrology Summary
3 Input Data to Porous Pavement Model ^
4 Porous Pavement Modeling Results - - • - • •
5 Outflow Hydrographs from Porous Pavement Model . . . .
6 Summary of Pollutant Potential of Dust and Dirt by ^ ^
Land Use * * . * ", '. co
7 Plotting Symbols for Unit Area Discharge Relationships b»
8 Determination of Rn and R Ratios
*7 *P
9 Nitrate Yield as a Function of Runoff ^
10 Water Quality Equations • ' *
11 Percent of Contaminants Removed from Street Surfaces
by Runoff Rate and Duration
12 SWMM Input Data for Sensitivity Analysis •>
13 Sensitivity of SWMM Modeling for Case A *
14 Sensitivity of SWMM Modeling for Case B ^
15 Rainfall Data, Hunting Bayou Watershed
16 Subcatchment Data, Hunting Bayou Watershed
17 Land Use Data, Hunting Bayou Watershed . .
18 Gutter and Pipe Data, Hunting Bayou Watershed ....
19 Transport Element Characteristics, Hunting Bayou
Watershed * * * ' '
20 Infiltration Parameters, Hunting Bayou Watershed . . .
21 Hydrograph Modeling Results for Hunting Bayou . . . .
22 Hunting Bayou - Storm of 5/08/75, Pollutant Loading
••••
23 Water Quality"Modeling for Hunting Bayou - Storm of
5/08/75 '
24 Subcatchment Data, Panther Branch Watershed
25 Land Use Data, Panther Branch Watershed
26 Gutter Data, Panther Branch Watershed . . . . - - - -
27 Transport Element Characteristics, Panther Branch ^
Watershed '
28 Rainfall Data, Panther Branch Watershed . . . . . • -
29 Infiltration Parameters, Panther Branch Watershed . .
30 Hydrograph Modeling Results for Panther Branch ....
31 Panther Branch - Storm of 12/04/74, Pollutant
Loading Rates * ' " " ", "
32 Water Qualtiy Modeling Results for Panther Branch -
Storm of 12/04/74 • - 139
33 Subcatchment Data, Swale 8 Watershed
xi
-------
TABLES (Continued)
Number Page
34 Land Use Data, Swale 8 Watershed 140
35 Gutter Data, Swale 8 Watershed 141
36 Transport Element Characteristics, Swale 8 Watershed . 142
37 Rainfall and Infiltration Data, Swale 8 Watershed . . 143
38 Land Use Data for Future Development, Swale 8 .... 145
39 Swale 8 - Storm of 4/08/75, Pollutant Loading Rates . 146
40 Water Quality Modeling Results for Swale 8 - Storm
of 4/08/75 150
41 Modeling Results for Future Development Upstream
from Station D-10 157
42 Relative Effects of Land Uses in the Swale 8
Watershed 161
Xll
-------
ACKNOWLEDGEMENTS
assistance to the Project staff in
following:
Rice University
P. B. Bedient
W. G. Characklis
F. J. Gaudet
j. D. John
F. L. Roe
J. S. Zogorski
U
S. Geological Survey, Houston
S. Johnson
E. Kamanski
R. Smith
The Woodlands Development Corporation
R. Heineman
B. Kendricks
J. Veltman
T. West
arklis, RU. llniv«»ity, tor their M.l.t.nc.
guidance.
The research reported herein was performed under the
anHssociates, Inc.. research staff included the following:
Technical Staff - E. Alexander, N. Atkisson, C. Dean,
D. Hoi Iowa y
Computer Scientists - F. S. Carl-Mitchell and
T. Sofka
-------
-------
SECTION 1
INTRODUCTION
Numerous studies and ongoing data collection programs have
provided a comprehensive understanding of the changes in the
hydrologic characteristics of a watershed during and after
urbanization. Both quantity and quality of runoff are affected.
increases in peak discharge rate and total volume of flow are
well documented, but the reduction in quality of urban storm-
water has only recently been recognized as a major problem.
Consequently, flooding from increased runoff and water quality
degradation are two major problems in urban hydrology.
The development of impervious areas such as roofs, streets
and parking lots in urban areas results in a severe limitation
or the infiltration capacity of an urban watershed Stormwater
management has generally consisted of collecting all the runoff
In gutters and discharging it into a conveyance system of storm
sewers and channels which are tributary to a nearby stream, lake
or ocean. An efficient urban drainage system has generally
implied the use of storm sewers and lined and rectified ditches.
Although local flooding problems were solved by this system,
increased time of concentration and higher peak flows which are
generated tend to create severe flood problems downstream.
Mso, impervious areas were found to have very few urban pollu-
tant assimilative properties. In fact, impervious areas tend to
generate urban pollution that is not amenable to street sweeping
and therefore, much more difficult to control (1). The increase
in flow velocities in improved channels creates a high erosion
and scour potential, thus aggravating pollution problems in
receiving waters.
An alternative drainage scheme termed natural drainage is
now being considered. The natural drainage concept is based on
?he premise that typically narrow and deep drainage ditches and
storm sewers are undesirable. Therefore, existing drainage
channels^ utilized to the fullest extent possible and any new
channels are constructed and lined with native vegetation to
funcUon similarly to the existing channels. In order not to
exceed the capacity of. natural drainage channels, runoff rates
musfbe approximately the same before and after urbanization or
the runoff rates must be reduced to runoff rates at natural
watershed conditions.
-------
The major method to reduce runoff rates is the use of minor
detention and retention ponds. In areas where a number of con-
tiguous impervious surfaces exist or are planned, the use of
porous or pervious pavements becomes a viable alternative.
Porous pavements and minor ponds provide design storage so that
they may be used to reduce runoff to preurbanization levels,
but, more importantly, they can be used to capture the initial
runoff or "first flush" volume which most studies indicate to be
the most degraded in terms of pollutant concentrations. All of
these runoff controls are utilized at The Woodlands, a new
community being developed near Houston, Texas.
In 1973, the U. S. Environmental Protection Agency (EPA)
sponsored this study to quantify the effects of the runoff
controls discussed above. The Storm Water Management Model
(SWMM), originally developed under EPA sponsorship for cities
with conventional drainage systems and combined sewerage, was
selected for this purpose because of its comprehensive modeling
capabilities. The present study was undertaken to modify the
SWMM in order to evaluate the effects of runoff controls as used
in natural drainage systems as practiced in surface drainage
design at The Woodlands and in Houston, Texas.
The SWMM was modified through inclusion of several new sub-
routines and debugging of existing subroutines. The resultant
model was then applied to Swale 8, the major watershed area at
The Woodlands where urbanization is progressing under natural
drainage concepts.
This final report summarizes the three years of research
and development effort expended in achieving the study objec-
tives.
-------
SECTION 2
CONCLUSIONS
The Storm Water Management Model (SWMM) released February 1975,
referred to in this report as the original SWMM version was exten-
sively modified by this project. The capabilities of the modified
version have been expanded to model runoff and water quality from
natural drainage areas. The study areas where the new capabi1ities
were tested are The Woodlands and Houston, Texas. During the course
of this study the following conclusions were reached:
1. After correction of errors in infiltration rate compu-
tation, the modified SWMM prediction of observed peak
x
on exact hydrograph replication, could not be modeled
for the study area.
Although runoff from natural drainage areas is of better
quality than that from area.s with conventional storm
sewer drainage, the effect of construction activity in
both types of areas could not be determined by the orig-
inal SWMM version. The predicted values were always too
low But, modeling of erosion from construction activi-
ties is now possible by use of the modified SWMM version
Laboratory methods used to determine biochemical oxygen
demand data produced inconsistent results and therefore
the biochemical oxygen demand modeling proved unsatis-
factory Data for chemical oxygen demand were more con-
sistent and subsequently used to model chemical oxygen
demand during this study.
It was determined that the functional relationship be-
tween pollutant mass and runoff volume could be linear
ized by the use of logarithmic transforms. _The resul-
tant linear equations can be used to determine loading
rates and total pollutant transport from a watershed
area.
5 The exponential pollutant removal or decay coefficient
can be considered as a constant in all geographical
areas. The modified SWMM allows for selection of the
value of this coefficient by the user.
-------
6. Water quality modeling capabilities of the SWMM have been
considerably improved. The modified SWMM can reasonably
predict mass flow rates (and pollutographs, if the hy-
draulic modeling results are accurate) for suspended
solids, chemical oxygen deamand, nitrates, phosphates,
or any other pollutant for which loading rates may be
determined. These capabilities were proved by modeling
observed events at the study areas where even transient
land use such as construction activities were succes-
sively modeled.
7. The modified SWMM can be used to model runoff events
generated by distinct periods of rainfall separated by
periods with zero rainfall. The original SWMM did not
have this capability.
8. The modified SWMM can reflect interaction between surface
water drainage and groundwater conditions by determina-
tion of recession flow rates from input recession char-
acteristics.
9. The modified SWMM can transport flow through natural
channels with a minimum of input data requirements.
Each natural channel is described by a series of coord-
inates and the program now calculates the area-discharge
curves which formerly had to be input for each natural
channel.
10. The modified SWMM can determine the cost efficiencies
in the use of natural drainage systems relative to those
for conventional drainage systems using either user
supplied or default unit cost estimates.
11. The modified SWMM can provide a detailed analysis of
storage and flow into and out of porous (pervious)
pavement systems. The drain outflow, surface runoff,
and storage volumes can be determined; but the lack of
comprehensive data precluded the modeling of water
quality in porous pavements.
12. The modeling schemes developed during this study require
considerable input data preparation and consequently the
modified SWMM, when applied to natural drainage systems,
is more user dependent.
-------
SECTION 3
RECOMMENDATIONS
litated.
areas especially with regard to
seasonal variation.
The water quality modeling approach developed should be
dology
or programming will be Identified.
for the Runoff Block the other program blocks
be modified so that all pollutants generated
from a watershed can be transported in one run of the model.
lished by this task
of the SWMM need to
The use of the modified SWMM should be Prom°^ sothat
rsSs a 53 =-s
drainage systems should be encouraged.
-------
erosion rates are reduced to pre-development or uniform levels.
Data collection should also be continued at the porous pavement
area and a sufficient number of overflow events should be sampled
so that the porous pavement model can be tested in detail under
real world conditions. Also water quality sampling of the out-
flow from the drain as well as the water in storage should be
continued during dry periods so that operation of the porous
pavement system can be properly evaluated and changes in water
quality may be understood.
Sedimentation surveys of all stormwater detention reser-
voirs should be performed at regular intervals and preferably
after every major storm event. The sediment accumulation rates
in the reservoirs would be very helpful in not only modeling
runoff and water quality but also in determining the life ex-
pectancy of the reservoir.
Records of all construction activity should be maintained
so that the location and total area under construction during a
storm event will be known. This record will prove very helpful
in modeling water quality from each watershed.
The modeling of erosion in the SWMM needs to be refined.
The coefficients of the Universal Soil Loss Equation were derived
for agricultural areas and their applicability to urban and
forested areas is limited. Possibly, a new approach may have to
be considered. The significance of pollution from construction
activity has generally been underestimated. Further study and
methods to control erosion in urbanizing areas must be developed.
A significant portion of the effort expended during this
project has been the setup and error correction of the original
SWMM version. The SWMM has several users across the country who
provide input as to improvements and corrections to the model.
The University of Florida at Gainesville, Florida has been essen-
tially a clearinq house for the updates being made to the SWMM.
In order to minimize computer compatability problems,
constant contact was maintained with the University of Florida
which has been very resoonsive to suggested improvements and has
already implemented many of the changes in order to make latter
SWMM versions compatible on all computers. This has facilitated
the setup of each new version of the model.
The problems were not limited to compatability conditions
between different computers. The reason for each new version of
the SWMM has been modifications and error corrections to the pre-
vious versions. These changes have affected the results obtained
from the previous version of the SWMM as well as the data deck
structure- and therefore require a complete reevaluation of pre-
viously completed work and the associated time and financial
-------
expenses This situation became especially critical when the
problem with infiltration computation discussed in Section 6
was discovered.
Several proaramminq anomalies and errors were identified in
S-JK «:£.".? ti: 2£ ""g-stsrsr^-ss.
include:
1 Erroneous values in the Transport Block of_the
SWMM concerning the depth of flow in certain types
of conduits. For example, a circular conduit that
would be flowing 41.5% full would be carrying only
4% of its total flow capacity.
2 In the Transport Block, input values describing a
special type of channel would be read in over ex-
isting values stored in the program describing a
trapezoidal section. This error prohibited using
the trapezoidal section whenever a special channel
was used.
3 Discovery of the existence of an undefined variable
being referenced by the program. This caused the
program to go into an infinite loop during some
calculations until the time limit was exceeded.
Considerable time was lost in tracing the variable
down so that corrective'action could be taken.
4. Erroneous graphs would appear on the rainfall
hyetoqraph when more than one rain gage was
specified.
5 As described in Section 6, amount of infiltration
was dependent on the time of start of the storm
after the start of modeling. This meant that
infiltration would be greater if the storm began
at the start of modeling time rather than a few
hours later.
6 Length of integration timestep must be shorter than
rainfall timestep or rainfall values are in error.
This is due to the program's method of averaging
rainfall intensities.
7 The SWMM Version II manual specified normalized
area increments but the program needed normalized
depth increments. This error was discovered only
after abnormal results were observed in several
runs.
-------
The rest of the errors encountered were minor format corrections.
All the corrections mentioned, except for the infiltration
and the timestep changes, were included in the February 1975 re-
lease of the SWMM. The University of Florida has subsequently
included the infiltration changes in the May 1976 release of the
SWMM.
Additional debugging of the SWMM is necessary. A signifi-
cant debugging effort and review of the program code has resulted
as an adjunct to this study; but a comprehensive correction of
program errors, a time consuming and sometimes frustrating pro-
cess, was beyond the scope of this project.
-------
SECTION 4
GENERAL PROJECT INFORMATION
THE STORM WATER MANAGEMENT MODEL
The Storm Water Management Model (SWMM) was developed in
1971 by the University of Florida and the U.S. Environmental
Protection Agency. The SWMM is composed of five integral
computation blocks as shown in Figure 1.
The Executive Block controls all activity within the^del
Intermediate ooints. The quantity and quality of flow is stored
and treated by predefined criteria in the Storage Block. ine
diversion effects of the discharge in the receiving body of
water are computed in the Receiving Water Block A more detailed
description is available in the SWMM User Manual Version II (2).
In aeneral only one or two computational blocks as well as
the Executive Block are used in a run but all blocks may be run
together The use of independent computation blocks allows for
the examination of intermediate results. Implementation of the
SWMM requires a computer having core storage capacity of at least
350K bytes which translates to high costs per run, which in turn
could limit the number of options to be analyzed.
The data requirements to model an urban watershed are
listed in Table 1. Line printer tabulations and specified
nydrographs and pollutographs are predicted as output from the
program.
Since its original release, the SWMM has undergone several
* "plosions- the most recent version became available in
version of the SWMM was used most
-------
f^
^ fl.
—-0
UJ I
UJ O
OC Q-l-
uJor t>
o x.
< o
a: o
o _i
-P
S 3
< CD
oo
c:
-M
3
o
<-
CD
s_
en
o
P-N |2|
Vv.0"'
z
Q
^
SO
U 03
10
-------
TABLE 1
MODELING REQUIREMENTS BY SWMM
Item 1. Define the Study Area
Land use, topography, population distribution, census
tract data, aerial photos, area boundaries.
Item 2. Define the System
Furnish plans of the collection system to define
branching, sizes, and slopes. Types and general loca-
tions of inlet structures.
Item 3. Define System SpecialUjss
Flow diversions, regulators, storage basins.
Item 4. Define System Maintenance
Street sweeping (description and frequency). Catch-
basin cleaning. Trouble spots (flooding).
Item 5. Define the Receiving Waters
General description (estuary, river, or lake). Mea-
sured data (flow, tides, topography, water quality).
Item 6. Define the Base Flow (DWF)
Measured directly or through sewerage facility oper-
ating data. Hourly variation and weekday vs. weekend.
DWF characteristics (composited BOD and SS>results).
Industrial flows (locations, average quantities,
quality).
Item 7. Define the Storm Flow
Daily rainfall totals over an extended period (6 months
or longer) encompassing the study events. Continuous
rainfall hyetographs, continuous runoff £ydrographs,
and combined flow quality measurements (BOD and SS) for
the study events. Discrete or composited samples as
available (describe fully when and how taken).
from: SWMM Volume 1 - Final Report, July 1971
11
-------
extensively during this study, and all modifications to the SWMM
by this project have been incorporated into this version.
All modifications are included as user options so that the
basic integrity of the SWMM is retained. Some of the new compu-
tational methods are dependent on site specific data.
STUDY OBJECTIVES
In general terms, the primary objectives of Espey, Huston
and Associates, Inc. as part of the team involved in the overall
study of the natural drainage system at The Woodlands were to
modify and expand the capabilities of the SWMM and apply it to
The Woodlands site. With the data on storm runoff quantity and
quality which were collected by Rice University and the U. S.
Geological Survey at The Woodlands, the model was to be used to
evaluate the effectiveness of natural drainage systems in mini-
mizing changes in storm runoff quantity and quality and to
assist the engineers and planners in designing the drainage
system for future phases of development at The Woodlands.
In specific terms, the study objectives of Espey, Huston
and Associates, Inc. were:
1. Modify the SWMM as follows:
a. to include a separate storm water system
b. to reflect the interaction between groundwater
conditions and surface drainage
c. to reflect natural drainage concepts
d. to include the cost of natural drainage systems
e. to include the additional water quality para-
meters COD, Kjeldahl nitrogen, nitrates, and
phosphates
f. to include the effects of porous pavement.
2. Apply the SWMM to a developed Houston watershed as a
prelude to modeling The Woodlands.
3. Apply the SWMM to Phase I of The Woodlands which is
now under construction.
4. Use the SWMM to assist in the planning and develop-
ment of the next development area at The Woodlands.
STUDY APPROACH
The scope of this study covered a period of three years.
Initially the existing SWMM was evaluated by application to the
Panther Branch and Hunting Bayou Watersheds. Hydrographs were
developed for the following storms on Panther Branch, Hunting
Bayou, and Swale 8, which is tributary to Panther Branch, in
12
-------
Phase I of The Woodlands:
Panther Branch Hunting Bayou Swale 8
10/28/74 9/08/68 4/08/75
11/10/74 9/17/68
11/24/74 11/05/68
12/05/74 10/22/70
12/10/74 11/09/70
3/26/74
5/08/75
6/30/75
Attempts to model several other storms were abandoned due to
data errors or extreme flow conditions.
The study objectives were accomplished in 12 tasks
generally divided into 3 categories as follows:
Evaluation of the SWMM
Task 1. Model observed storms on Hunting Bayou in Houston
Task 2. Model observed storms on Panther Branch in The
Woodlands
Task 3. Model Swale 8 in the developing area of The
Woodlands
Modifications to the SWMM
Task 4. Modify SWMM to be used to model areas served by
separate sewers including natural drainage
Task 5. Model the interaction between groundwater and
surface drainage
Task 6. Improve the modeling of infiltration to allow
periods of no rainfall
Task 7. Develop a subroutine to prepare area-discharge
data for natural sections
Task 8. Develop a subroutine to compare costs of natural
drainage relative to costs of conventional
drainage
Task 9. Develop a methodology to determine predictive
relationships for COD, Kjeldahl nitrogen, ni-
trates and phosphates and include modeling of
these parameters in the SWMM
13
-------
Task 10. Model the effects of porous pavements on rainfall
and runoff relationships
Testing of the Modified SWMM
Task 11. Apply the modified SWMM to Hunting Bayou, Panther
Branch and Swale 8 to observed events
Task 12. Apply the modified SWMM to Swale 8 to model
future development trends
In accomplishing these tasks, several new subroutines were
developed and incorporated into the SWMM. Each of the new
subroutines is discussed in Sections 6 and 7 and listings of the
program code are included in Appendix A. A revised set of input
data coding instructions to allow the use of the new subroutines
is also attached as Appendix B. The computer runs performed
during this study were too numerous and were therefore not
included in this report. All modeling input data are displayed
in appropriate tables and figures in Section 8. All water
quality data used in this study are summarized in tables in
Appendix C. Appendix D contains hydrograph recession data;
sample output from a SWMM run, utilizing all new computational
schemes, is shown in Appendix E.
14
-------
SECTION 5
STUDY AREA DESCRIPTION
HUNTING BAYOU WATERSHED
The Hunting Bayou Watershed is located on <*« eastern side
the ^r^oritrof-thfara'beinglrained by roadside grass-
the season of the year and maintenance schedules.
inasmuch as these characteristics are also repre sentative
The area modeled on Hunting Bayou, as shown in Figure 3,
is thfentire watershed upstream of the U S .^^cal
/ncr-cN /-rarfinrr station NO . OoO/D/oU at raJ-4-t> ouj-cci, v _,«,-,
of ThilsLtlon which has been in operation since 964
continuously records flood hydrographs and rainfall. In the
in ?973 Both of these stations are shown in Figure 3.
15
-------
Fig. 2 Study area vicinity map
16
-------
STUDY AREA BOUNDARY
r~~~T
\
\
\
V
\ 1
\
\
\
\ \
t :...
\
\ ••••
\ -..
\
\
\
\
\
r1
i
i
-I blUUT MR
It- •• SEWER P
Iz OPEN Dr
Im w GAUGING
lz
o
130
1
.1^:4'
i iv
». i 'A
I ..ix
•••« i* . \
• ' , • i i n
• ' U \ ^o-rU \-°
--- -- *• NJ \t>4QB^
" - "" "• • - V^^H-
~~ - • « f -~
v» ^ v
•"/ , ' f^\H- 20
• ».•*••>•»•••••»' \
/ / N
•••:< ! \
H| %'-»W.H-IO)
• i t /
JH./7 r-/'
: 1 i
Fig. 3 Hunting Bayou watershed
17
-------
THE WOODLANDS DEVELOPMENT AND PANTHER BRANCH WATERSHED
The Woodlands is a new community being developed in Mont-
gomery County, Texas. The community is situated in a heavily
forested tract about 45 kilometers (28 miles) north of Houston.
The Woodlands encompasses 7,200 hectares (17,800 acres) (Figure
4) and is planned to be developed over a twenty-year period,
which began September 1972. A total of 33,000 dwelling units
are programmed with a projected population of 112,000 in 1992.
A concern for nature and convenience for man were two of the
major criteria used in the development of the General Plan for
The Woodlands. The basis for all aspects of development in The
Woodlands was a unique ecological inventory conducted from 1971
to 1973.
The Woodlands site is located in the Spring Creek drainage
basin. The major stream draining The Woodlands is Panther
Branch which is a tributary of Spring Creek. Panther Branch is
an intermittent stream with major no-flow periods occurring
during the summer months. A flow gaging station, P-30 (USGS
No. 08068450) was established near the lower end of The Wood-
lands site in May, 1972. The drainage area upstream of the
gage is 54.39 square kilometers (33.8 square miles).
Panther Branch and its tributary Bear Branch form the
principal drainage system upstream of the developing areas in
The Woodlands (Figure 4). A stream stage recorder, Station P-
10, is located below the confluence of Bear and Panther Branches
and has a drainage area of 24.30 square kilometers (15.10
square miles). The drainage area of P-10 is undeveloped forest
land while the P-30 drainage area includes Phase I development
of The Woodlands.
The basic drainage system planned for The Woodlands has
been designed on the basis of what has been termed the natural
drainage concept. This concept consists of the following
principles: 1) The existing drainage system in its unimproved
state is utilized to the fullest extent possible; 2) Where
drainage channels need to be constructed, wide shallow swales
lined with existing native vegetation are used instead of
cutting narrow, deep ditches; 3) Drainage pipes and other flood
control structures are used only where the natural system is
inadequate to handle increased urban runoff, such as in high-
density urban activity centers, and 4) Flow retarding devices
such as retention ponds, recharge berms and porous pavements
are used where practical to minimize increases in runoff volume
and peak flow rates due to development. The natural drainage
concept as outlined by these four principles seeks to minimize
changes in the runoff regime due to urbanization by providing
increased infiltration and storage capacity and higher resis-
tance to flow within the channels.
18
-------
-------
SWALE 8 WATERSHED
The first area to be developed at The Woodlands is referred
to as Phase I and is located immediately upstream from Station
P-30. The largest individual drainage area in Phase I is Swale
8. With a drainage area of 195.46 hectares (483 acres), Swale
8 drains into Panther Branch through Lakes A and B.
As shown in Figure 5, the downstream third of the watershed,
west of Grogans Mill Road, is substantially urbanized primarily
in commercial and multi-family residential development.
The drainage system for Swale 8 was designed on the basis
of the natural drainage concepts described earlier. In conjunc-
tion with the natural drainage system, three major reservoirs
are located in the Swale 8 drainage area. Two reservoirs
(Lakes A and B), approximately 4.8 and 3.0 hectares in size (12
and 8 acres), respectively, are located at the Conference-
Leisure Center. The 3 hectare reservoir will empty directly
into the 4.8 hectare reservoir. The normal operating level of
the larger reservoir will be held at 1 meter (3 feet) below the
storm outflow depth thereby providing 1 meter of flood-control
storage. The effects of this storage are very significant in
reducing the impact of urbanization on the quantity and quality
of runoff from Swale 8 downstream from the lake. Lake C, a 2.8
hectare (7 acre) reservoir on the east side of Grogans Mill
Road has also been completed recently. One meter (3 feet) of
storage is provided in it for runoff control.
Two stream gaging stations are operated by the U. S.
Geological Survey in the Swale 8 watershed. Station D-50
measures the outflow from Lake A and Station D-10, located on
the east side of Grogans Mill Road, measures the inflow into
Lake B from Swale 8.
DATA COLLECTION
As described previously, all stream flow stations are
maintained by the USGS. The continuous records at these stations
proved very useful during this study; the period of record for
each station is as follows:
P-10 10/73 to present
P-30 4/72 to present
H-10 4/64 to 9/73
H-20 4/64 to present
D-10 10/74 to present
D-50 11/74 to present
Monthly grab samples for water quality were also collected by
the USGS.
20
-------
WATERSHED DIVIDE
GAUGING STATION
SCALE (feet)
500 0 500 1000
D-30
\
\
Fig. 5 Swale 8 watershed
21
-------
During the initial phases of this project, several problems
with a discharge rating curve for Station P-10 were exper-
ienced; but a satisfactory rating curve was finally developed.
The long processing time to convert gage heights to discharge
rates was substantially reduced by the development of a computer
rating curve program. The computer program, with the aid of an
input rating curve would convert gage heights to discharge rates
from which the storm hydrograph was developed.
The temporal and spatial distributions of several key
water quality parameters were defined for individual storm events
through comprehensive sampling by Rice University. Parameters
analyzed included but were not limited to suspended solids, COD,
Kjeldahl nitrogen, nitrates, and phosphates. Because of the
processing time, these data were not available immediately,
resulting in a lag of several months between data sampling and
water quality modeling. However, all the data had been processed
and were available prior to the end of the project. A summary
of all storms sampled by Rice University is listed in Table 2.
22
-------
TABLE 2. STORM EVENT HYDROLOGY SUMMARY
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
.-
Date
1/18/74
3/20/74
3/26/74
4/11/74
4/22/74
10/28/74
12/05/74
3/04/75
3/13/75
4/08/75
5/08/75
06/30/75
09/05/75
10/25/75
03/07/76
03/08/76
04/04/76
Site
Woodlands P-30
Hunting Bayou
Hunting Bayou
Hunting Bayou
Woodlands P-30
Woodlands P-30
Woodlands P-30
Woodlands P-10
Woodlands P-30
bake A
Lake B
Woodlands P-30
P-10
Laenk
=-low,CFS
1260.0
9.6
40.0
11.0
9.7
382.0
332.0
280.0
4.9
scharge
0.38
49.0
56.0
2.0
12.7
1100.0
1170.0
114.0
123.0
72.5
36.0
136
47
10.3
0.38
8.8
64
22
15
~
18.5
7.9
-
3.8
32.5
23
3. 4
6.6
64
.37
10
19
Runoff ,
acre ft
21.31.0
0.857
24. R
5.47
5.39
928.4
1238.0
822. 0
1.8
0. 135
111.0
77.0
1.77
2826.0
1610.0
93.4
93. 2
23. 8
7.16
94.0
11.23
9.48
0.822
0
1.51
117.35
57.09
18.86
12.6
10.37
6.5
. 884
59.0
29.1
6,78
2.71
45.56
0.537
14. 53
5.103
a Total Streamflow is calculated to include components of overland runoff and base flow.
b
Percentage of rainfall as runoff.
23
-------
SECTION 6
MODIFICATIONS TO THE SWMM
MODELING A SEPARATE STORM WATER SYSTEM
The SWMM was originally developed for drainage systems
which included combined sewer systems. In the case of separate
sewer systems, it was necessary to model urban storm water
systems which do not include the dry weather flow (DWF) compo-
nent of combined sewer systems. Consequently, subroutine INFIL,
to compute infiltration into the sewer, and subroutine FILTH,
which estimates DWF based on population, were not called in the
Transport Block and Card Groups 30 through 44 were omitted in
all runs, thus providing a considerable reduction in input data
requirements.
In modeling natural drainage systems, the lack of catch
basins allowed a further omission of input data. Card Group 17,
catch basin data, in the Runoff Block data deck was a blank card
The modifications to the SWMM to model a separate storm
water system are relatively simple and no difficulties were en-
countered with this phase of the project.
INTERACTION BETWEEN GROUNDWATER CONDITIONS AND SURFACE DRAINAGE
In the absence of storm sewers and their associated infil-
tration rates in a natural drainage area, hydrograph recessions
were used as indicators of the interaction between groundwater
conditions and surface drainage. Several investigators in-
cluding Holtan and Lopez (3) consider hydrograph recession rates
to be a function of total volume of water in storage in surface
depressions, vegetative and soil layers and in groundwater. The
rate at which these storage volumes are drained is determined to
be the recession rate of the hydrograph beyond the point of
inflection when all surface runoff is assumed to have ceased
(4) .
The physical process of draining water from storage can be
approximated by a linear reservoir whose outflow rate is a
function of storage. This relationship is defined by Riggs (5)
and others as follows:
24
-------
qt =
where
q is the discharge at some initial time
q is the discharge at any instant
K is the recession constant
t is the number of time units elapsed since qQ
in the above equation, the numerical value of K depends on the
time unit selected. Consequently, the equation was redefined as
follows:
Kt
in this case all units are as defined previously except K, which
is now defined as A_ln g . This equation can be linearly re-
At
presented on a In q versus t graph.
From studying a large number of observed hydrographs,
Barnes (6) determined that the different outflow rates from
water stored in depressions, surface soil, and groundwater could
be approximated by 3 separate straight line functions on a
semilogarithmic plot. Later studies (4,7) have substantiated
this assumption and Figure 6 depicts this characteristic at
Station P-10.
This approach was utilized in the inclusion of baseflow
modeling by the SWMM in subroutine BASFLO. A maximum of 5
recession rates are allowed. Each recession rate was made a
function of flow. User input data were developed by first
transforming all discharge values for all hydrographs at P-10
and P-30 to" the natural logarithmic values and then graphically
determining the slope, K, on plots of In q versus t for each
hydrograph recession segment which indicated a linear relation-
ship. Therefore, each recession is composed of a number of
straight lines which begin at a specific In q and end at another
specific in q as shown in Figure 6. The slope values were then
graphed against the beginning In q values and straight line
equations of the form K = Ko + m In g for In q versus slope were
derived by means of least squares analysis as shown in Figures 7
and 8. Because extreme data points are critical in least
squares methods, all extreme data points were subjectively
deleted prior to the derivation of the equations.
25
-------
o
,-H
I
c
o
•H
•P
U)
-P
(U
10
fd
P3
26
-------
2.0
PANTHER BRANCH
NEAR CONROE
NOTE-® INDICATES POINTS NOT CONS IDEREO
Fig. 7 Slopes of hydrograph recessions at Station P-10
27
-------
.2
STATION P-30 |
I
PANTHER BRANCH NEAR SPRING |
SLOPES OF RECESSION CONSTANTS
NOTE-® INDICATES POINTS NOT CONSIDERED!
Fig. 8 Slopes of hydrooraph recessions at Station P-30
28
-------
Subroutine BASFLO requires input values of starting In q
and the slope, m, and intercept, Ko, coefficients to compute the
recession rate, K. Five sets of data may be input.
A flow chart of all computations performed in Subroutine
BASFLO is shown in Figure 9. In general, the hydrograph at the
downstream end of the^ystem, which is output from the Transport
B^ock? is input to subroutine BASFLO which then performs the
following sequential computations.
1) Consecutive flow rates are compared to determine the
hydrograph peak.
2) For each flow rate, q._, after the hydrograph peak,
the slope with respect to the previous flow rate,
q , is compared to the slope to the next flow rate,
q^~|. As long as the absolute values of the slopes
are'increasing with each time step, the computations
proceed with no interruption A d^crea^e/^h^7
solute values of the slopes indicates that the point
of inflection has been located.
3) Using this flow rate as the beginning flow rate, sub-
routine BASFLO searches the input data table to deter-
mine during which recession interval the present re-
cession begins and selects the corresponding recession
rate coefficients.
4) The value of the recession rate, K, is computed and
the recession equation applied to determine the new
flow rate, qt-
5) If no further surface inflow to the stream occurs as
a result of a second rainfall, for example, the flow
rates are receded into the domain of a new starting
flow rate and a new set of recession rate coeffi-
cients.
6) This procedure is continued until the hydrograph
starts rising again or until the computational time
steps are exceeded.
The recession flow rates computed in subroutine BASFLO re-
Dlace all flow rates after the point of inflection on the hydro-
graph and the resulting total storm hydrograph is output and
graphed if so desired.
For streams which receive low flow contributions from
aroundwater aquifers, subroutine BASFLO includes an option to
include thes^flow rites as a constant rate, a linearly varying
rate or an exponentially varying rate. The option to use
29
-------
READ HYDROGRAPH OF SURFACE
RUNOFF COMPUTED IN TRANSPORT BLOCK
SELECT PEAK FLOW RATE
DETERMINE SLOPE OF RECESSION
BETWEEN TWO CONSECUTIVE FLOW RATES
IS THE
SLOPE
INCREASING p
lNo~"
'IS FLOW WITHIN
INPUT RECESSION
RANGE p
lYes
Yes
No
SELECT RECESSION CONSTANTS
FROM INPUT TABLE
COMPUTE RECEDING FLOW
RATES AND VOLUMES
IS VOLUME OF
RECESSION FLOWS =
AVAILABLE SUPPLY P.
lYes
No
No
PIS THERE ANY
1AINFALL DURING
THIS at /Yes
ADD WATER TABLE DISCHARGE COMPUTED
BY INPUT DISCHARGE FUNCTION
ADD COMPUTED RECESSION FLOWS TO
SURFACE RUNOFF HYDROGRAPH
Fig. 9 Subroutine BASEFLO system logic
30
-------
varying rates is in recognition of the fluctuations sometimes
observed in water tables, and especially in perched water tables
as are found at The Woodlands.
The use of subroutine BASFLO is dependent on the availabil-
ity of onsite or site compatible data for recession rates. _If
these data are not readily available the modeling accuracy is
diminished.
Changes in the recession rates resulting from urbanization
range from the elimination of depression storage by leveling and
grading, to total prevention of groundwater recharge by use of
In impervious pavement. These effects have not been quantita-
tively evaluated due to the deceleration of urbanization at The
Woodlands. At a later date these data should become available,
but until then, the user must supply recession rate reduction
factors based only on engineering judgment or data from other
sources.
The simulation of observed recessions at P-10 and P-30 for
two storms each are shown in Figures 10 and 11. These results
indicate a close correspondence with the observed events and
this should be so because the recession rate coefficients to
determine individual recessions were derived from all observed
events at the same locations. The closeness of fit indicates
that hydrograph recessions at P-10 and P-30 can indeed be approx-
imated by average recessions.
At the initiation of this project, 6 surface soil water
measuring wells were installed at different locations in The
Woodlands but unfortunately the cost of maintenance on the
sampling program was too high to be cost effective and conse-
quently no data on the accretion or depletion of surface soil
water are available.
A number of factors affect the uniformity as well as the
reliability of recession curves; some of the major factors are
listed below:
1 in areas with a marked vegetative growing season and
in agricultural areas, marked variations in evapo-
transpiration losses result in recession rate varia-
bility. Therefore, data from streams in these areas
should be carefully selected.
2 Channel and bank storage effects will affect the re-
cession rates and must be considered when inter-basin
transfer of data is necessary.
3 If more than one aquifer supplies water to a stream,
the contribution from each aquifer must be considered
separately.
31
-------
Ul
BOdVHOSIQ
LLJ
c
o
•H
-P
td
-p
o
•H
CO
CO
-------
4 Snow melt rates, if applicable, should also be eval-
uated if baseflow recessions during snowmelt periods
are being investigated.
INFILTRATION
Another aspect of the interaction between groundwater and
surface drainage IB ?he modeling of infiltration in the SWMM.
infiltration, which usually starts at a high rate and decreases
during rainfall to a lower constant rate, occurs in a three step
sequence: 1) entry of water through the soil surface, 2) trans
mission through the soil and 3) the filling of the storage
capacity of the soil, The surface entry rate may be reduced by
the inwashing of fines or other particles and by raindrop impact.
Therefore, infiltration will be limited to the lowest transmittal
rate encountered by the water. The available storage in a soil
depends on the porosity and thickness of the soil horizon. The
average infiltration capacity, f , during the time interval is
taken to be the value at the center of the time interval
(t + 0 5 At) and is calculated by Horton's (8) equation:
where
f = infiltration capacity at time t*
c
f . = maximum infiltration rate
i
f = minimum infiltration rate
o
k = decay rate
t* = time from the start of rainfall to the midpoint of the
time interval = t + 0.5 At
After calibration of the SWMM for a study watershed, the
only input data that were varied were precipitation and the coef-
ficients for Horton's infiltration equation, especially the
starting infiltration rate. Horton's equation, a function of
time only, has been found to fit a large number of experimental
infiltration capacity curves obtained from different soils with
different types of vegetal cover and in widely separated regions,
The equation was derived from experimental infiltration tests in
which the supply of water exceeds the infiltration capacity.
Therefore, a limitation of the equation is that rainfall inten-
sity must always exceed the infiltration capacity. In naturally
occurring rainfall, the rainfall intensity can be less than the
infiltration capacity and periods of zero rainfall can occur
within a storm event.
33
-------
In the infiltration model in the revised SWMM version after
the input parameters (f., f , k) are initialized, the infiltration
capacity decays as a function of the time elapsed from the start
of computations. But the decay of the infiltration rate during a
storm event of nonuniform intensity will be a function of present
rainfall intensity, past rainfall, and time. In the SWMM, if the
rainfall intensity is less than the infiltration capacity, the
capacity will decrease at the constant decay rate (k) as if the
rainfall intensity were equal to the infiltration capacity.
In applying the model it was found that increasing the
maximum infiltration capacity in some cases had no effect on the
volume of infiltration. A manual calculation through one time
interval showed that the infiltration volume should have in-
creased. Upon investigation, a programming anomaly was dis-
covered. In the SWMM, the decay of infiltration capacity started
at time zero, not at the start of rainfall. If rainfall started
at four hours after the start of modeling, the infiltration
capacity calculated by the SWMM would have decayed for four hours
and will be at or near the minimum capacity at the actual start
of rainfall. This may be the reason why previous sensitivity
studies showed that varying the maximum infiltration rate pro-
duced no effect on the runoff volume (9).
The SWMM's infiltration model was modified by EH&A to use
Horton's equation in an integrated form in association with a
temporal parameter that follows the progress of infiltration on
each subcatchment as described below. In the integrated form,
Horton's equation is:
M = fot + (fi - fo) (1 - e~kt)
where M is the accumulated volume of water, in inches, infiltra-
ted up to a time, t; and the other variables are the same as
defined previously.
During each time interval, At, the volume of water the soil
is capable of infiltrating, in inches, (M -M ) is calcu-
lated. This volume of water is compared to the total volume of
water available for infiltration. When the available volume is
greater than the infiltration volume,
Dl > (Mt + At -V
the excess is calculated and the results are comparable to the
present infiltration model in SWMM.
If the infiltration volume is greater than the available
volume,
Mt + At -V > Dl
34
-------
the time increment, tQ < t, is calculated such that the infil-
tration volume is equal to the available volume,
(Mt + ^ -Mt) = Dl
and runoff is set equal to zero. The time (t + tc) is then used
as the starting point of infiltration in the next time interval.
The University of Florida has already incorporated this
infiltration modeling scheme in the May 1976 version of the SWMM.
COSTS OF NATURAL DRAINAGE SYSTEMS
One of the major factors to be considered by engineers and
planners in designing natural drainage systems is the relative
cost of such systems with respect to conventional dramaae
systems which generally consist of sewernetworks improved and_
realianed channels. When the SWMM is being used for design pur
poses? the ability to determine these relative costs through Sub-
routine CSTANL could be very beneficial.
The description of the drainage system in the Transport
Rlock of the SWMM is in two dimensions only; elevations ana
excavation depths are not used. Therefore,.Subroutine CSTANL
uses unit area costs to determine total project costs.
Unit area costs for natural and conventional sewered drainage
systems and for right-of-way acquisition and clearing are user sup-
. . , . _r -ioui^ or the default values based on 1970 costs at
PTe'woodLndfarrused1: Tatef cost data are adjusted fc,; P«sent
conditions by use of the Engineering News-Record Cost Index These
unit costs are applied to each subcatchment drainage area to deter
mine the total costs for natural and conventional drainage sys-
tems for each subcatchment . The ratio of natural ^ainage costs
to conventional drainage costs is also determined. This ratio may
then be used to evaluate the desirability of either drainage system
for each subcatchment.
A flow chart of computations in Subroutine CSTANL is shown
if Figure 12. The computations proceed as follows:
1 Unit costs of conventional and natural drainage
systems are input to the model. If these costs
are not available, the default values of ?686/
hectare ($1700/acre) for conventional and $121/
hectare ($300/acre) for natural systems are used.
These values are based on 1970 data for The Wood-
lands .
If the cost data are not current, the Engineering
News-Record Cost Index is utilized to update the
cost data.
35
-------
COST/AC. OF CONVENTIONAL
SYSTEM KNOWN?
yes;
COST/AC. OF NATURAL
SYSTEM KNOWN?.
•yes -,
DEFAULT VALUES OF
$300/AC. NATURAL
$1700/AC. CONVENTIONAL
no
ARE THESE THE LATES
COST FIGURES?
[COST/AC. NATURAL SYSTEM] =
[COST/AC. CONVENTIONAL] * 0.176
PRESENT E.N.R. COST INDEX/E.N.R. COST INDEX OF
KNOWN COST/AC. VALUES
* COST/AC. NATURAL AND COST/AC. CONVENTIONAL
R.O.W. AND CLEARING COST =
[R.O.W. AREA]*
[LAND COST/AC. + CLEARING COST/AC.]
^
conventional
i
DRAINAGE AREA
* COST/AC.
i
TOTAL COST OF
NATURAL AND
CONVENTIONAL
DRAINAGE SYSTEM
natural
\
f
[DRAINAGE AREA * COST/AC.]
+ COST R.O.W. AND
CLEARING
COST OF CONVENTIONAL
DRAINAGE SYSTEM
COST OF NATURAL
DRAINAGE SYSTEM
RATIO OF CONVENTIONAL
TO NATURAL
DRAINAGE SYSTEMS
Fig. 12 Subroutine CSTANL system logic
36
-------
3 If total ricrht-of-way acquisition and clearing costs
are not available, they are computed by use of unit
costs for land acquisition and clearing.
4 The total costs of conventional and natural drainage
svstems for each area are computed and the ratio of
one to the other indicates the relative efficiencies
of each drainage system.
The results from Subroutine CSTANL are subject to the accuracy of
the input cost data.
AREA-DISCHARGE DATA FOR NATURAL SECTIONS
cross-sections. The normalized area-discharge curves for these
sections must be manually calculated, as shown in Figure 13, and
input to the SWMM. When modeling a natural drainage system the
user will often find that the flow is not confined within the
natural channel. Also the overbanks of a natural cross-section
Sill generally have Manning's roughness coefficient, n, values
significantly different from that of the channel. Therefore,
calculation of an area-discharge curve for a natural cross-
section mSst consider the complex relationships between depth of
'Manning's n values, one n value for each overbank and one
for the channel.
When the depth of flow is within the channel the flow cal-
culation is a direct application of Manning|s uniform flow equa-
tion If the flow depth is such that flow is not contained
wi?hin the channel, Manning's equation is applied separately to
each section as shown in Figure 14. The total discharge is
assumed to Se the sum of individual discharges. This assumption
was used by Chow (10) in his derivation of an equation for an
equivalent roughness in a channel with composite roughness.
Input to subroutine NATSEC consists of a set of progres-
sively increasing coordinate points and the desired number of
area increments. A maximum of 25 coordinate points and 50 area
increments are allowed. Also the horizontal stationing for the
left and right overbanks is required.
37
-------
N=IO
= .04
CIRCULAR PIPE
NATURAL CHANNEL
2 .4 .6 .8 1.0
A/Ac
Fig. 13 Normalized area-discharge curves
38
-------
(XI.YI)
(X2, Y2)
(X3,Y3)
(X7,Y7)
,Y6)
(X4,Y4) (X5.Y5)
Q4 = Q, * Q2 + Q3
Fig. 14 Modeling of natural cross-sections
39
-------
The storage arrays have been expanded to allow the storage
of normalized area-discharge data for ten natural cross-sections.
Further versatility is provided by allowing the user to geometri-
cally enlarge or reduce each cross-section by use of an input
proportional constant.
The subroutine calculates the full flow area by dividing
the cross-section into the triangles, rectangles or trapezoids
defined by the input coordinate system and the maximum water sur-
face elevation. The maximum water surface elevation is taken to
be the lower elevation (lower Y value) corresponding to the two
end points of the cross-section (lowest or highest X value). The
areas of these uniform geometrical shapes are then calculated and
summed to give the full flow area. Each incremental area is
determined by dividing the full flow area by the user specified
number in increments. The subroutine subtracts the area incre-
ment from the full flow area and uses an iteration process to
calculate the depth of flow corresponding to the new flow area.
With the depth of flow defined for an incremental area the sub-
routine calculates the corresponding hydraulic radius from the
known geometry.
To define the relationship between depth of flow and an
equivalent roughness value the equation developed by Lotter (11)
for a composite cross-section is used. Lotter assumed that the
total discharge is equal to the sum of the discharge of the sub-
divided areas. Thus, the equivalent roughness is:
PR
s P. R. 1.67
I -i i
= 1 ni
n = equivalent roughness
e
P = total wetted perimeter
P- = wetted perimeter of subsection
R = hydraulic radius of total section
R. = hydraulic radius of subsection
n. = roughness of subsection
s = total number of subsections
Subroutine NATSEC allows a natural cross-section to be
divided into three subsections, two overbank sections and a
channel section with Manning's n value input for each section.
Using the calculated depth-area relationship and depth-
roughness relationship the subroutine calculates the normalized
flow-area curves as follows:
40
-------
„
— = ns
Qf A R
Af f
subscript s refers to a subsection
subscript f refers to a full section
Q = discharge
A = area
R = hydraulic radius
n = roughness coefficient
In the above equation the slope terms cancel out because
channel and overbank slopes are assumed to be equal. The output
from subroutine NATSEC is a tabular version of the normalized
area-discharge curves in Block Data of the SWMM. Subroutine
NATSEC was used to great advantage in modeling runoff from The
Woodlands where all of the channel system described to the SWMM
consists of natural channels.
MODELING OF POROUS PAVEMENT
The earliest applications of porous pavements were for
nonstorage purposes; they were used on top of a regular Pavement
to provide improved drainage and reduce the possibility of skid-
ding or hydroplaning. The use of porous pavements as stormwater
management systems was initiated by Franklin Institute in Phila-
delphia, Pennsylvania under sponsorship of the U. S. Environ-
mental Protection Agency(12).
The primary benefit derived from porous pavements is an
appreciable reduction of runoff rate and volume from impervious
urban areas. If the pavement and base are designed adequately,
all of the runoff may be captured, detained and released at a
slower rate to prevent increases in flood flows. Concurrently,
the stored water may be allowed to infiltrate into the natural
ground.
Porous pavements may also be used in areas that are already
urbanized such as downtown areas of most cities as well as ex-
isting shopping centers where the storm sewer network was in-
stalled prior to excessive impervious cover development. Under
these conditions, the storm sewers may become overloaded and if
parking lot or roof storage is not a design criterion, the dis-
posal of excess runoff becomes a problem that porous pavements
could solve. This benefit is enhanced in areas with combined
sewerage because the probability of sewer overloading and the
resultant discharge of raw sewage into the receiving waters is
reduced.
41
-------
In areas of slight topography or with minimal soil depths,
the cost of installing storm sewers is very high because both
sewer size and excavation volumes are high. The use of porous
pavements in these areas reduces both sewer size and excavation
depth, thus resulting in a net savings in drainage costs.
If the stormwater requires treatment, it may be stored in
porous pavement systems isolated from the natural ground by an
impermeable membrane until the treatment plant capacity becomes
available. Thus, treatment plant capacity does not need to be
expanded. Also detention of highly polluted initial runoff by
the porous pavement and dilution by less polluted subsequent run-
off can result in reduced pollutant concentrations throughout the
storm.
Because impermeable pavements preclude the survival of any
vegetation, natural vegetation and drainage patterns can be re-
tained by the use of porous pavements. Consequently, the clear-
ing of large areas for parking lots is unnecessary and second-
ary aesthetic benefits are also derived. Other benefits in the
form of construction cost reductions, elimination of curbs and
gutters, enhanced water supply, traffic safety resulting from
skid resistance and improved visibility on wet pavements, and
possible use of solid wastes for base material are discussed by
Thelen et al. (12).
Development of Model and Theory
An extensive modeling effort was undertaken to develop a
comprehensive analysis of flow and storage in porous pavement
systems and thereby define the environmental effects of this type
of pavement. The mechanics of flow through a porous pavement
system has not received much attention; and only after a complete
understanding of porous pavement operations will the total app-
licability of this system be determined.
A deliberate attempt was made to keep the model as simple
as possible and yet to provide adequate quantification of the
hydrologic responses of a porous pavement. Also, the effects of
a variety of different pavement characteristics and configura-
tions can be evaluated. This will allow for the investigation of
various porous pavement systems to determine the optimum system
especially during planning phases of the project.
The hydrologic responses of a porous pavement may be sim-
ulated by a system of hydraulically connected control volumes for
which the inflows arid outflows are mathematically defined. The
porous pavement, the subgrade and the natural ground (or the
drain system) are considered to be sequential but internally
independent storage reservoirs.
The basic equation of continuity or conservation of mass is
applied to each reservoir:
42
-------
where
I = inflow into the reservoir
0 = outflow from the reservoir
-T|:- - change in storage volume
As shown in Figure 15, the porous pavement area would
serve to control runoff from contributing impervious areas.
Therefore, inflow to the porous pavement system, RUNOFF is
defined as:
RUNOFF = PAV + HYD
where
PAV = direct rainfall onto the porous pavement
HYD = surface runoff hydrograph from contributing areas
Contributing areas to the porous pavement will generally be de-
veloped and impervious in nature. Consequently, the surface
runoff hydrograph from contributing areas is determined by use
of the method developed by Izzard (13). This method, selected
for its programming ease, utilizes a dimensionless hydrograph
from paved areas as shown in Figure 16. The key parameters in
this method are time to equilibrium, t ; equilibrium flow, q ;
e "
equilibrium surface detention volume, V ; the intensity of rain-
fall, i; and the length of overland flow, L. The following
equations define these parameters:
q ~
e ~ 43200
where: q is in cfs, i in in/hr , and L in ft.
e 4/3 .1/3
w - k LY i '
Ve ~ 375T
where: k is an empirically derived lumped coefficient for the
effects of slope and flow retardance of the pavement, Y
is the flow depth.
V
Using t/t values based on the computation interval and Figure
16, the q/q values and the corresponding q values are determined
for the rising limb of the hydrograph. The 3 factor, defined as:
43
-------
DRAINAGE DIVIDE
•' POROUS PAVEMENT '«'
•OUTFLOW
TO DRAIN
Fig. 15 Porous pavement and surrounding drainage area
.2 .4 .6 B 9 1.0 2.0 3.0 4.0 5.0
t/L B
From: Linsley, Kohler & Paulhus, 1975
Fig. 16 Izzards dimensionless hydrograph for overland flow
44
-------
B - 6° qe *a
Vo
where: t = time after rainfall
V3 = equivalent to VG without the rainfall intensity
0 component
i<5 used to determine the q/q and corresponding q values for the
recession limb of the hydrogfaph. The dimensionless hydrograph
in Figure 16 is represented by the following equations in the
model:
0 < t/t <_ 0.15 q/qe = e
t-
where: A = 32.2 In (t/te) - 8.36
B
0.15 < t/t 1 0.35 q/qe = e
where: B = 11.64 In (t/tQ) - 5.203
0.35 < t/t <_ 0.55 q/qe = 1.992(t/te) - 0.432
ti
0.55 < t/te l 0.75 q/qe = In C
where: c = 0.533 + 2.46 e
T = t/te
0.75 < t/te i i.o q/qe = in D
where: D = 1.628 + 1.023 e
T = t/t
-3/2
t/t > i.o q/qe = (2.0 B +1.0)
7 e c
The rainfall hyetograph is input as average intensity per compu-
tation interval for all intervals during which rainfall occurs.
Runorf hydrographs are computed for each interval, successively,
and summed to determine the cumulative storm hydrograph from
each paved area. The cumulative hydrographs from all paved
areas are added to obtain the total storm hydrograph from con-
tributory areas to the porous pavement. The inflow hydrograph
is converted to units of depth based on the «ea of the porous
pavement and the computation interval. The rainfall depth onto
the porous pavement, PAV, is added to surface runoff depth, HYD,
to determine RUNOFF. An alternative user optional input table
of RUNOFF depths per computation interval may also be utilized.
This option is useful in those instances where storm hydrograph
data are available from observation or computed by other methods
45
-------
As considered in this model, the outflows from the porous
pavement system are composed of four outflow functions defined as
follows:
0total = °vert + °hor + °surf + °evap
°vert is the vertical seepage into the pavement, base, or ground.
This seepage is determined as the difference in surface water
depth at the beginning and end of each time interval. The var-
iable head permeability equation as defined by Taylor (14) is
K = 2.3 a L log h^
AAt h2
where
K = permeability of flow element
a = cross-sectional area of surface water
A = cross-sectional area of flow element
L = thickness of flow element
hj = depth of surface water at time t
h2 = depth of surface water at time t = t, + t
In a porous pavement system the cross-sectional areas of surface
water and flow element are always equal and so the equation is
reduced to
K = 2.3 L_ log hj^
At h2
This equation may be arranged to solve for h as follows:
h2 = h.
10E
where E = KAt
2.3L
Then, vertical seepage is equal to the change in water depth
during At or,
°vert = hi - h2
°hor is the lateral outflow to a drain or into the natural ground
as a result of water storage in the base and pavement. This con-
dition is analagous to bank recharge from a rising stream and for
homogenous isotropic aquifers of finite width, the influence of
each increment of rise in the stream is determined by the fol-
lowing set of equations (15):
46
-------
d2h S dh
, 2 T dt
dx
h(o,t) = 0 for t * 0
h(o,t) = ^ for t > 0
dh(L,t) = Q
dx
h(x,o) = 0
where
h = hydraulic head or water depth
x = distance from boundary
S = coefficient of storage of aquifer
T = aquifer transmissivity
H. = change in water depth at boundary =
L = total width of aquifer
Integrate the first equation and apply boundary conditions to det
ermine that the constant of integration is equal to zero and the
result is:
dh = S dh x
dx T dt
The Darcy flow equation can be extended by continuity to define
net flow rate as follows:
K A * = K h „
- v K h - T
where
V = velocity of flow
K = permeability of flow element
J~ = change in hydraulic grade
CLX
Q = total mass flow rate
A = total flow area
h = depth of flow
w = width of flow
q = flow rate per unit width
T = transmissivity of flow element
47
-------
Substitute _ _, „ ,
8h x = S ah x
^
q
T at at
define
and S is the storage coefficient of the natural ground. Then
at a distance x from the porous pavement side, the discharge
per unit width is defined as:
q =
Because the volume of flow remaining in the porous pavement
system is the only item of interest, the value of x was arbi-
trarily set equal to 1.0.
Then lateral outflow = q P At = S (hj^~h^) p At
where P is the pavement perimeter. 0 f is the surface runoff
resulting from ponding on top of the porous pavement, which
occurs either because the inflow rate is greater than the
porous pavement permeability or the total storage capacity in
the porous pavement system is exceeded. The model requires a
depth-storage relationship to determine when the storage is
exceeded. On a horizontal pavement, the model determines the
depth-storage relationship by use of input pavement and sub-
grade depths and porosities; on a sloping pavement, this rela-
tionship has to be independently computed and input to the
model.
The surface runoff from a horizontal pavement is defined
by the weir equation: _ ,_
Q = CLH3/2 Where
C = input weir coefficient
L = input weir length
H = h - h
o
h = dead surface storage on the porous pavement
h = depth of flow on the porous pavement
On a sloping porous pavement Manning's Equation is used to
determine the surface runoff.
Q = yL i y sl/2
where n
y = computed depth of flow
L = input width of flow
n = input roughness coefficient
48
-------
s = input energy slope
0 is the volume of water lost to evaporation. Either
monthly, weekly or daily evaporation rates may be input to the
model; the monthly and weekly rates are divided into average daily
rates. The daily evaporation rate is increased by 25% to allow
for heat absorption by the dark asphalt. The model only
allows for evaporation from 6 a.m. to 8 p.m. with the maximum
rate at 2 p.m. As shown in Figure 17 a triangular distribution
of evaporation is developed by the model by use of the equation:
Et
Ep= T-
where
E = peak eveporation rate, inches/hr
E = total daily evaporation, inches
for 0 < t ^ 6 E = 0
c
for 6 < t £ 14 E = ED(V^
for 14 < t £ 20 E - E
for 20 < t ^ 24 E - 0
c
where
t = clock time in hours
c
E = instantaneous evaporation rate
Model Operation
The paths of water flow through the porous pavement system
are shown in Figure 18. For each computational time interval,
all inflows and outflows are accounted for. The total runoff
hydrograph, in inches per computational time interval, is either
input"to the model or may be computed as the sum of the runoff
hydrograph from contributory areas and direct rainfall onto the
pavement as described previously.
The following sequential computational steps as illustrated
in the flow chart", Figure 19, are then performed:
49
-------
Ml'
_i ^..r_
MEASURED
EVAPORATION
0600 1200 1800
TIME OF DAY
2400
Fig. 17 Triangular approximation of evaporation
FLOW FROMBASE
°vert
EVAPORTION, Oevap
POROUS PAVEMENT
STORMWATER INFLOW, I
SURFACE RUNOFF, Osurf
I'" '•'"••'-"';"' '•'•'''••' BASE '•"'• :•' :''-'' V'V'';,'.
LATERAL OUTFLOW
^IMJ^" NATURAL GROUND -^^^
Fig. 18 Pavement cross-section and modeled flow
50
-------
READ SUBAREA WATERSHED PARAMETERS
AND RAINFALL INTENSITY
COMPUTE IZZARD'S HYDROGRAPH
PARAMETERS FOR EACH SUBAREA
READ DIMENSIONLESS
HYDROGRAPH TABLE
COMPUTE OVERLAND FLOW HYDROGRAPH
FOR EACH SUBAREA
SUM FLOW HYDROGRAPHS FROM ALL
SUBAREAS AND COMPUTE THE
INFLOW VOLUME PER AT
ADD DIRECT RAINFALL ON POROUS
PAVEMENT PER T TO DETERMINE
TOTAL INFLOW PER AT, I
Yes
°vert T0 BASE
°hor T0 DRAIN
°evap T0
°surf T0 DRAIN
o
I = 0
vert
Fig. 19 Subroutine PORPAV system logic
51
-------
Yes
No
No
COMPUTE
°vert T0 NATURAL GROUND
'hor
evap
TO DRAIN
IF NOT ALREADY COMPUTED
I = 0
vert
COMPUTE
°vert T0 DRAIN OR WATER TABLE
°evap IF NOT ALREADY COMPUTED
READ STORAGE DEPTH FUNCTION
IS WATER DEPTH GREATER
THAN STORAGE DEPTH
No
Yes
'surf
COMPUTE
TO DRAIN
0 = 0 4. + 0. +0 j- + 0
vert hor surf evap
REPEAT FOR NEXT AT
Fig. 19 Cont'd
52
-------
1. Evaporation losses in inches per computational time
interval are computed and subtracted from the sum of
the runoff depth and previous surface storage, if any.
2. The volume of runoff after allowing for infiltration
is compared to the permeability in inches per time
interval of the porous pavement. As in most cases the
permeability is much greater than the inflow runoff
rate and all of the water moves into the pavement con-
trol volume. In those cases where the permeability
has been severely reduced and is less than the inflow
rate, the inflow into the pavement is computed as the
vertical seepage into the pavement and the excess is
stored on the surface of the pavement for later compu-
tation of surface runoff from the pavement.
3. The inflow into the pavement control volume is added
to the storage volume in the pavement and then com-
pared to the permeability, in inches per computational
time interval, of the base. If the base permeability
is greater than the inflow into the pavement, then all
of the flow is transferred into the base control vol-
ume. This is true for most porous pavement systems
operating according to design. In those instances
when the base permeability is less than the inflow
volume, the inflow into the base is computed as the
vertical seepage into the base. The lateral outflow
from the pavement is also computed if an impermeable
membrane is not installed along the pavement perimeter.
The difference between the inflow into the pavement
and the outflows (vertical and lateral) from the pave-
ment is stored in the pavement.
4. The inflow into the base control volume is added to
the storage volume in the base and then compared to
the permeability in inches per computational time in-
terval, of the natural ground. If the bottom is
sealed with an impermeable membrane, then the permea-
bility is set equal to zero, and no flow is lost to
the natural ground. The flow volume remaining in the
base after vertical seepage into the natural ground,
is compared to the drain capacity, in inches per com-
putational time interval. If the natural ground per-
meability and/or drain capacity are inadequate to re-
move all of the flow in the base, the vertical seepage
into the natural ground and drains, as well as the la-
teral outflow, if any, is computed. The difference
between the inflow into the base and the outflows (ver-
tical and lateral) from the base is stored in the sub-
grade .
5. All stored volumes are compared to available volumes.
53
-------
If storages volume in the base is exceeded, the excess
is stored in the pavement; if storage volume in the
pavement is exceeded, the excess is added to the sur-
face storage on the pavement, if any exists. Surface
runoff is then computed either as broad channel flow or
weir flow from the pavement to an adjacent drainageway.
After continuity conditions are satisfied by comparing all
inflows to outflows,, this computational procedure is repeated for
each time interval in the inflow hydrograph. The surface and
drain outflows are stored in retrievable arrays. The primary
output objective at present is the surface runoff, if any. But
the other output variables allow for a thorough examination of
the hydraulic operational characteristics of the porous pavement
system, including the analysis of the desirability or adequacy of
the drains and the discharge rate from the drains.
Model Application
Subroutine PORPAV, as the porous pavement model is called,
has been applied to the porous pavement parking area in The
Woodlands. All existing data for this porous pavement area is
not sufficiently comprehensive to test all of the model capa-
bilities. No record of drain capacity is available. Also, the
available data on storage of runoff in the porous pavement system
are in terms of percent of depth that the probe is submerged.
These data have to be converted to inches of depth of water
storage to make the data useful. If the probe were exactly as
long as the base thickness then the conversion would be rela-
tively simple. But such is not the case; the probes have to be
calibrated. Consequently these data could not be used to verify
and calibrate subroutine PORPAV. A further problem arises from
the fact that due to the size of the observation wells (6 inches
diameter), whenever a water sample is withdrawn a significant
drop in water elevation in the well is observed. Therefore,
comparisons of observed and simulated events were limited to
general tests, e.g., if surface runoff was observed. Pavement
and base permeabilities had to be reduced and a design storm
applied as shown in Table 3 in order to generate surface run-
off. Contributary area boundaries were also enlarged so that the
hydrograph prediction capabilities of the model could be hypothe-
tically tested.
The porous pavement installation at The Woodlands is con-
structed as a layer of open-graded asphalt concrete underlain by
a gravel base course with appreciable storage capacity. The
whole system is isolated from the natural ground by an imperme-
able polyethylene liner. Water is removed from the base by an
artificial drain. As shown in Figure 20 the pavement area is
rectangular, approximately 64 meters (210 feet) by 39.6 meters
(130 feet) in size. Two contributory areas are identified,
1865.38 sq. meters (6120 sq ft) and 1828.8 sq. meters (6000 sq ft
54
-------
TABLE 3. INPUT DATA TO POROUS PAVEMENT MODEL
• • • POROUS PAVEMENT AT THE WOODLANDS « « « DESIGN STORM« « «
NUMBER or CONTRIBUTORY SUSAREAS = 2
NUMBER OF RAINFALL TIME INTERVALS = 7
NUMBER OF COMPUTATIONAL STEPS = 150
NUMBER OF RUNOFF TIME INTERVALS = -Q*
3UBAREA CHARACTERISTICS
SUBAREA ROUGHNESS SLOPE ft/ft LENGTH, ft WIDTH, ft
1 .050 .030 34.0
2 ..117 .010 80.0
INPUT RAINFALL DATA, in/hr
1,500000 3,000000 4,500000 6,000000 4,500000 3,000000 1,500000
COMPUTATIONAL TIME INTERVAL 2,000 MINUTES
INPUT EVAPORATION DATA
STARTING TIME 12,HR 30.^JN
EVAPORATION DEPTH FOR 30 DAYS IS 5,000 liMCnES
POROUS PAVc >[;>NT CHARACTERISTICS
PAVEMENT
P .-. R '-1 .
^JCFF
1-3.000
30.000
.500
6.000
JEfTh
in
2.31)0
12.000
36.000
3, UOC
P 0 R 0 3 -
ITYin/hr
30.00J
50.300
5.000
INIT,
0,000
0,000
0,900
MAX,
,750
6,000
1,800
NATURAL
ARTIFICIAL DRAIN
WIDTH OF PAVEMENT = 130,000
LENGTri OF PAVEMENT = 210,000
SLOPL OF PAVEMENT = (030
DQ'-'NSTREA^ FLO1 ft'lOTn Or PAVtMfcNT = 15,' 000
,'IAKNING COEFFICIENT FO^ PA;t:,Mcf,T = ,050
WEIR COEFFICIENT FOR FAVtK-NT = -Q 000
DtAD STORAGE OM PAVLMtr.T = '^QQ
INITIAL STORAGE OM PA\EhENT = -0*000
55
-------
03
4-1
W
0)
Cfl
3
O
H
O
dl
o
CN
•rH
t,
56
-------
in area, respectively.
Based on general criteria, initial testing of the model in-
dicated that storage and outflow were adequately simulated for
observed events. The design input rainfall hyetograph and cor-
responding runoff hydrograph from each contributory area are
shown in Figure 21. Flow and storage in each control volume as
predicted by the model for two different permeabilities in each
pavement element are depicted in Figures 22 and 23. Figure 22
illustrates the porous pavement system operation when pavement
and base permeabilities are 101.6 cm/hr (40 in/hr) and 203.2
cm/hr (80 in/hr), respectively. As shown in Figure 22, the de-
sign storm used (based on 100 year rainfall for the Houston area.)
did not generate any surface runoff because of the high infil-
tration rates available. Therefore, the permeabilities were re-
duced to 38.1 cm/hr (15 in/hr) and 76.2 cm/hr (30 in/hr) for
the pavement and base, respectively. The resulting model pre-
dictions include surface runoff as shown in Figure 23. Table 4
shows a sample segment, of the model output.
The environmental effects of porous pavement could not be
determined due to the lack of sufficient data. However, specific
trends were evident. Runoff which accumulated in the porous
pavement system exhibited significant (on the order of magnitude
of four) nitrate, nitrite, and Kjeldahl nitrogen concentrations
in comparison to surface runoff. Water stored in the sand sub-
base was high in orthophosphates and total phosphates. Soluble
COD concentrations seemed to be much lower in water stored
within the porous pavement system. Sample average water quality
is illustrated for the storm of 20 February 1976 in Figures 24
and 25. Suspended solids data were not compiled and consequently,
the effect of this parameter is unknown.
The fate of pollutants stored in the porous pavement system
could not be determined because no data were available for water
quality during periods between storm events. Generally, the
water would have drained out of the porous pavement system within
a short time after a storm event. The drain water also was not
sampled so that the effect of runoff retardation within the
porous pavement system could not be determined.
Subroutine PORPAV at present only allows for dilution of
pollutant concentrations by retarding runoff. In general the
drain from a porous pavement system will discharge into a re-
ceiving body of water specifically designed for this purpose,
e.g., channel, pond or wastewater treatment plant. Therefore,
surface runoff would be the only consequential flow with regard
to downstream runoff quantity and quality. Subroutine PORPAV
will determine outflow hydrographs from both the drain as well
as surface runoff as shown in Table 5.
57
-------
-------
0)
•H
U) 4J
3 -H
O H
5-1 -H
O XJ
a fd
w td
ro
CM
(saqoui)
3IAiniOA
UJ
U)
O
O
•H
Cfl
Q)
g
rH
O
CD
fd
5-1
O
C/3
CM
5-1
QJ
a.
Q)
g
Q)
>
fd
a
(saqoui
in
•H
59
-------
CO
EH
P
CO
O
H
H
Q
O
<
CM
CO
P
O
PC;
o
CM
H
PQ
EH
Q
O
O
CM
CO
P
o
o
CM
CO CO
H H
H EH
EH P
H £
^ H
H JE;
PQ
H H
K fa"
H S
CM M
EH
G
H EH
U CM
P H
Q U
PL) X
CM H
I CO
H
LO ffi
r- U
00 H
O
\ fa
^r O
fa co
O EH
CO H
z
1-4
D
W Z oOf-((Mmvfi'-iin'.or~-aocoooootDoooooooooococoooooccoo
< O •
O C "~
E--
W
Cc/l ooooooooooooooocrinoinirioccoioo«j'incTioorMr~i-ioooooo
< corj>^.o<^Lncrimv£3c^
tt tc. ocooooooooooooooooooc^moooooooooooooo
orcit ocooooooooooooooooooo-^r^-oooooooooooooo
HZ OCOOOOOOOOOOOOOOOOOOO OOOOOOOOOOOOOO r"
CO M ***
4-)
w w r3
(JO Jj
E-1 ? « jr,
z o < (L)
OJcn ocooooooooooooooooooooooooooooooooooo
N[u OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
M H s en
tt'DO OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO t^
O O K t^l
K [14 rti
i-5 S in
< o < O —-.
M b? m UD in (N r- ,-H vo 14-1 H in in m mm ^ tD
ago ^^^^^^^^"™ro-"°°^^^j^^^^!x^^ww^^^™^^^^^ "o-T a
> ^ ^ fN":rt/1t ^^^ ^.^mtN^ ^ ^
^ . ^ o
H3^ O &
z o o, C! -H
OiJ OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
t^ti4E OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO "l~l C/J 4-'
t-Hf-lO t.i... ^fk —H
ttDK ooooooooooooooooooooooooooooooooooooo ,,.. ^ *-•
o o U4 "' 4J O
^g -E
<*S O.H
u o CL 6 II
WDK r-^rH p-lr-(fNfNrNrMi-(l-( "^ ^-5 ^?
> c u. fa p-i tn
uo ^ w ^
.—c •..<
Orifn ^ ii-l rS
DO ooooooooooooooooooo ooooooooooooo -rH ri
1/10 4J r^
s d -H en
°WinrHLn m JHCU
^ 0 c
i « > O -H
QiZK OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO Q • *~-'
COE-l OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO __
< D £?
w O en ••
M-l -H W
O ^D<- mrsn-^^inM^r-^o tn in CTN m ^ m o> ^ in in a> ^ o rM r- m <^ r- ^ 04 ^ ^ oo r- ^
t, ^^"^^^^^"™™™°. ^^^^^^^^^^^ ^^^"^^^^^^^^^ .x t-,
2; —(—it-jrHf^iivir^rNjr-i £-*
M
.E OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO
60
-------
,'
1
1
\
/
1
\
\
\
1
\
\
\
\
\
1
<$
0_
_j
h-
O.
\.
\:
\\
'.
(j
./\
• |
j
'•-. /
../
m
l£2
^
in
ro ^
p £
UJ
m -— Q
— : Q. -
rO uj'oo
Q """
^ §^
CVJ _j
Hio
5 a:
in _j ^
f\l ~ *— '
in •-»• — i /^ v^ OL i i
NOIltfdlN30NOO
in
10
IO
in
^.
ro
LU
LJ
-p
-H
cu
-JJ
CO
-P
fd
IS
LT>
CN
tn
•H
4J
•H
5-1
(U
-P
fd
CM
(l/btu) NOIlVdlN30NOO
NOIlVdlN30MOO
61
-------
OCMOOOOOOOOOOOOO
O>OOOOOOOOOOOOOO
Of^OOOOOOOOOOOOO
OIT\OOOCOOOOOOOOC>
OOOOOOOOOOOOOOO
OOOOOOOOOOOOOOO
CD OOOOOOOOOOOOO
^ o o v r^ •* of\joinc\jo<\jr»io
O O O O f-v vO »-t O *W O r-l Ot3O
o
O
Q_
O
CD
O
C3C3O OOOOCDOOOC3C3OO
OOOOOOOOOOOOOOO
ooo oooooooooooo
OOOOCJOOC3OOCDOC3C3C3
OC3ODOOOCDC3OOOC3OCD
OOOOOOOOOOOOOOO
OC3O'OOOC3CDOOOOOOO
OCDC3OOC3OOOOC3OOC3C3
OC3OOOOOOOOOOOOO
oo^ooooooooooooca
COCOOCVI-OOVCVJT
C3OC3OOCDOOOOOC3OC3O
ooooooooooooooo
OOOCDC3OC3OOOC3CDOOC3
COOt3OOOOOOOOOC3C3
OOOOOOOOOOC30000
C3C3OOC3OC3OC3OOOOC3O
OOOOOOOOOOOOOOO
2:
LU
o
cr
o
Q.
OOCOOOOOOOOOOOOO
OOr^C3C3OOC3C3fOC3CDOOO
OOOC3000000OOOOC3
OC T-4O OOCDC3 C3O C3OC3C3O
OOinc3OOC3CJO-
I
OO^OOOOOOOaOOOOO
OOIAOOOOOOOOOOOO
OCDCDOOC3OC3OOOOC3OO
oooooocaooo
OOCOCJOO OC3C3C3OiC3C3OO
OC3OC3CDOOCDC3C3C3OC3C>C>
OOCNJOOOOOOOOOOOO
OOC30OOOOOOOOOO101
62
-------
SECTION 7
WATER QUALITY
DATA ANALYSIS
With regard to specific data concerning pollutant avail-
ability that may be used to define the required water quality con-
stituents to be used in the SWMM, two studies (1, 18) have been
published that provide useful information. These two studies
contain street sweeping data and determined the amount of pol-
lutants available per gram of dust and dirt swept from the street.
The results are tabulated in Table 6 for various water quality
constituents by land-use category.
Obviously there are significant differences between the re-
sults of the two studies. Sartor and Boyd report higher values
in nearly every instance. An explanation might lie in the
differences in experimental methods and geographical areas.
Specifically, APWA used only dust and dirt smaller than .32 cm
(1/8 in) in Chicago. Furthermore, in the APWA study, the mixed
samples were filtered before testing; Sartor and Boyd say only
that their samples from 10 different cities were "homogenized."
If the latter samples were not filtered, the higher values would
have resulted. The coefficients for pollutant availability
prediction equations as used in the original SWMM were based
primarily on the APWA data. It was concluded that because of the
significant variances in observed data, a different approach to
water quality prediction was necessary.
Toward this end, an extensive study of water quality data
was undertaken. A survey of available literature was made to
determine sources of water quality data for storm runoff. Spec-
ific information desired was water quality and flow data taken at
discrete time intervals during storm events for urban watersheds
with seperate sewer systems. Storm hydrographs and pollutographs
were found for two watersheds, Third Fork Creek at Durham, North
Carolina (16) and K.N.. Clapp Drainage Basin at Lubbock, Texas
(17). Point sample data collected by the U.S. Geological Survey
for the following watersheds were also analyzed:
1) Little Vince Bayou at Pasadena, Texas
2) Willow Waterhole Bayou at Houston, Texas
3) Vince Bayou at Pasadena, Texas
63
-------
TABLE 6
SUMMARY OF POLLUTANT POTENTIAL
OF DUST AND DIRT BY LAND USE
Sartor & Boyd
APWA
Average
Residential
BOD mg/g COD mg/g P04 mg/g N03 mg/g
11.9
4.3
8.1
27.75
40.
31.37
1.13
.05
.59
.064
NA
Sartor & Boyd
APWA
Average
Commercial
BOD mg/g COD mg/g P04 mg/g N03 mg/g
8.6
7.7
8.15
26.
39.
32.5
1.03
.07
.55
.6
NA
Sartor & Boyd
APWA
Average
Industrial
BOD mg/g
10.3
3.0
6.65
COD mg/g
53.
—
—
P04 mg/g
1.41
__
__
N03 mg/g
.072
_ _
64
-------
4) Plum Creek at Houston, Texas
5) Brickhouse Gully at Houston, Texas
6) Waller Creek at Austin, Texas
As described in Table 2, the data collected by Rice Univer-
sity consists of a total of 14 storms divided during the three
year project duration as follows:
Project
Year
Station
P-30
H-20
P-10
P-30
D-10
D-50
H-20
P-10
P-30
D-10
D-50
H-20
No. of
Storms
2
3
Storm Dates
1/18/74, 4/22/74
3/20/74, 3/26/74,
4/11/74
12/05/74, 3/13/75,
4/08/75
12/05/74, 3/04/75,
3/13/75, 4/08/75
3/04/75, 3/13/75,
4/08/75
3/04/75, 3/13/75,
4/08/75
5/08/75
9/05/75, 10/25/75
9/05/75, 10/25/75
9/05/75, 10/25/75
9/05/75, 10/25/75
6/30/75
W.S.* 1 6/30/75
*W.S. is station for Westbury Square drainage area.
Data summaries for all the Clapp Basin storms and for all
Rice University data are included in Appendix C.
During the first year of this project, the Lubbock and
North Carolina data, the USGS point sample data and the data
collected by Rice University for the Panther Branch gaging sta-
tion, P-30, at The Woodlands for the storm of 1/18/74 were ana-
65
-------
lyzed to determine the effect of flow on the concentration of
specific water quality constituents. Plots were made which
relate the concentration of a given constituent to the instanta-
neous unit discharge, which is defined as the flow at the time
of the sample divided by the drainage area. Plots were developed
for suspended solids, BOD, COD, and nitrates. These plots are
shown in Figure 26. Table 7 contains the legend showing the
plotting symbol used for each watershed.
The suspended solids data shown in Figure 26 show high sus-
pended solids concentration for the Third Fork Creek and for
Panther Branch. For Third Fork Creek the high concentrations
appear to be due to the fact that the samples were collected by
a submersible pump located at the bottom of the stream. Hence
during storm events the samples were taken from the lower por-
tion of the flow in the stream. For Panther Branch, the high
suspended solids concentrations apparently reflect the large
amount of upstream construction activity. In general the data
for all the watersheds indicate an increase in suspended solids
load as the flow rate increases.
For the BOD data no trend is readily apparent. However,
generally the BOD values are less than 10 mg/1. No BOD data
were available for Panther Branch. The COD data again show that
the Third Fork Creek watershed had significantly higher COD
values than other areas. Again this is probably due, at least
in part, to the sampling procedure. For the Clapp Basin in
Lubbock a relatively consistent trend is shown indicating an
increase in COD concentration with increased flow rate.
The nitrate data indicate a relatively consistent increase
in concentration as the flow rate increases. It should be
noted that the nitrate data for Panther Branch were not plotted
on this figure because the concentrations were consistently less
than 0.05 mg/1.
The general trend of increasing concentration with increas-
ing flow rate observed in Figure 26 provides an insight to water
quality conditions at any instantaneous flow rate but to apply
this method to a total hydrograph would assume an unlimited
supply of available pollutant. Such is not always the case as
evidenced by later data at Station P-10 which indicated a washout
of nitrates after a specific volume of flow had been discharged
from the watershed. The above inferences were the first indi-
cation that perhaps mass of pollutant transported during a storm
event would be a more reliable parameter than concentration.
This approach was adopted during the latter phases of this
effort and, as described at the end of this section, it proved
to be the most reliable method available.
Because no definite conclusion regarding water quality
predictions could be easily derived, further research became
66
-------
3
fi
3
o-
•
.
*
c
0
fl
** o
•
• **
,
.
*,.*
>
^
<3
' ^
•
. .
. .
•
r
i9/
^ * *
• «
• c
a
.
>
j
*. •
..
, ^
•
r
•
" '>
•.'0*
.c.
o
D *
— s
.
.
>
.
4
3« '
"**
*D
\
* S
' sv
•OG
*
»~>
«
O
O.^.
~ E
cr
V)
t>^tt
O "*
o
\
— O
O
o
o'
8
°
« E
o-
(A
O
pop o °
rd c\j —
I/BUJ '2,
CO
'sanos
«»
f e
00*
> f
o
•
* •
o
.
V C\J
Q3QN3dSnS
*».*•
•0 • •
% •
• • *O
•
Q
0
• 'e
o'
• o
•
»
•f •
« w
o
0
o
o %
Q
O • ,,
O
OOOC
O Q
no
a <^+
!"
o
0°
C
r<
8
]
-
1
O-
o
cr
•o .'
g
o
Q
O a,
E
cr
Q o
O
5
O
D
€>
a
0
Of+
00**
i °*
03 a
aoo
CO
c
0 -H
-P
rH
CU
O ^
D Q Q 00 0
0 OJ fcj]
|/6uu l Q09 !d
u
CO
O 'H
M n-(
(
G
0
-t
*
0 +
. O "*<3
. **•'
"* '
*
0 (
fi
•t
O
0)
8 "
~ 4-1
•H
D
QO)
.•= (yO
E CN
cr
«> «
\ tP
J2 -H
O o PM
> 0
5
i
0
67
-------
TABLE 7
PLOTTING SYMBOLS FOR
UNIT AREA DISCHARGE RELATIONSHIPS
Used in Figure 26
Watershed Symbol
Little Vince Bayou at Pasadena, TExas Q
Willow Waterhole Bayou at Houston, 0
Texas
Vince Bayou at Pasadena, Texas o
Plum Creek at Houston, Texas -L
Panther Branch at The Woodlands, Texas A
Brickhouse Gully at Houston, Texas ^
Waller Creek at Austin, Texas Q
Third Fork Creek at Durham, o
North Carolina
K. N. Clapp Drainage Basin at
Lubbock, Texas
68
-------
necessary. Several different approaches were attempted including
the relationships between pollutant loading and the following
parameters:
1) Rainfall a) Intensity
b) Duration
c) Total
2) Runoff a) Volume - cu. ft. and in.
b) Peak flow
c) Velocity
d) Rate
3) Land Use a) Type
b) Slope
c) Curb length
d) Population density
e) Impervious cover
f) Vegetation density
Most of these efforts proved to be fruitless and were discontinued.
Only the runoff volume relationship (2a) was judged to be reliable
and was subsequently developed in conjunction with pollutant
mass loading to the fullest extent, possible as described later
in this section.
Another approach involved the determination of temporal
relationships of pollutants to discharge. Plots of all observed
hydrographs and pollutographs on the same time axis were used to
compare time at peak and time at specific volume percentages;
a sample plot for suspended solids is shown in Figure 27. This
analysis did not provide any capability to predict pollutant
generation.
Ratios between flow at peak concentration and peak flow, R,
and also between time at peak concentration and time at peak flow,
R^, were determined for suspended solids and Kjeldahl nitrogen
during selected storms as listed in Table 8. The ratios were
not sufficiently constant so no conculsions could be derived.
A correlation between inches of runoff per day and pounds of
pollutant per acre was attempted as shown in Table 9 and Figure
28 for nitrates. Although correlation trends were indicated,
the results were inconclusive and not applicable to specific
storm event water quality predictions.
A final correlation between pounds of pollutant per inch of
runoff and peak discharge as shown in Figure 29 for Hunting
Bayou indicates the existence of a peak discharge at which the
pollutant rate per unit of runoff may be a maximum. But further
data is necessary to verify this conclusion. A similar relation-
69
-------
o
o
(uiiu/sqi )
in
MOld
SSVIfl
8
10
CJ
NOIlVdlN30NOO
O
-P
13
(U
T3
c
(U
£4
CO
M-l
O
0)
EH
I—
(N
Cn
•H
70
-------
CO
o
CM oo co r^ CM <}• o
i"^ oo r*- -
r-. oo
CO 00
CO J- -
I-H oo r^
co CM
r^ o
CM OO
4-J
CO
0000000
oo CM CM on oo r-i oo
I I I I I I I
P-I !U ffi PM PM PM PM
O O O O O
CM CM OO r-H OO
I I I I I
ffi ffi PL, fit PM
0)
4J
cd
Q
OCi
CM i— I
OO . r^-
i— I OJ OO LO LO
i— I CM CN O O
CM CM
CM
CO
CM LO LO
CM O O
-------
TABLE 9 NITRATE YIELD AS A FUNCTION
OF DAILY RUNOFF
Date
1/18/74
3/26/74
4/11/74
4/22/74
5/08/74
6/30/74
10/28/74
12/05/74
12/05/74
3/13/75
4/08/75
4/08/75
Station
P-30
H-20
H-20
P-30
H-20
H-20
P-30
P-10
P-30
P-10
P-10
P-30
Yield Rate
Ib/ac
.009
.029
.0058
.0002
.020
.0592
.0044
.0017
.0037
.0012
.0018
.053
Daily Runoff
in/day
.37
.26
.096
.003
.3
.886
.174
.183
.215
.053
.538
.610
72
-------
>s
o
TJ
o:
LU
o:
UJ
/
/
/
1/18/74
/
_ • 5/08/74
3/26/TO
12/5/74
12/5/74 "
10/28/74
/I
• 4/11/74
3/13/75
4/22/74
A STATION P-IO
• STATION H-20
• STATION P-30
0 01 .02 .03 .04 .05
NITRATE YIELD (in/day)
.06
Fig. 28 Nitrate yield as a function of runoff
73
-------
o
C\J
03
4->
00
cu
tn
.c
u
CO
•H
C
O
-H
-M
U
C
P
4-1
'O
1— I
0)
•H
4J
C
03
-P
CN
74
-------
ship between total pounds of pollutant and peak flow proved to
be unusable.
As part of this work effort, a computer program for data
processing was developed, the source code li ting for which is
also included in Appendix A. The final output from this program
was a line plot of the comparison between two parameters that
are user selected. A large number of runs were performed to
plot hydrographs, pollutographs, mass flow graphs (loadographs),
and comparisons of each pollutant against any other parameter.
The last series of runs included plots of the logarithmic trans-
forms of cumulative pounds of pollutant versus cumulative runoff
volume in cubic feet for each storm. Because this endeavor
seemed to hold promise of success, all the data for Stations P-
10, P-30, and H-20 were plotted for these two parameters.
These plots are reproduced in Figures 30 through 34. It
was found that a straight line approximation could be fitted to
these curves for suspended solids, COD, Kjeldahl nitrogen,
nitrates and phosphates. The linear equation coefficients were
determined by least squares analysis of data points that were
subjectively screened to eliminate extreme values. The corres-
ponding linear equations are also represented in Figures 34
through 34.
A further development was realized when the end point on
each cumulative pounds versus cumulative runoff curve was plotted
with respect to other similar points with the pollutant load
reduced to a unit area. In other words, plots of unit pollutant
load in pounds per acre against runoff volume in inches tend to
fall on a line whose slope is determined by land use as shown in
Figure 35.
COD and Kjeldahl nitrogen exhibit remarkably linear charac-
teristics in Figures 31 and 32. Suspended solids, nitrates and
phosphates seem to exhibit some rionlinearity but linearization
is still reasonable. As expected, cumulative load will increase
with cumulative runoff volume since M = cQ t
where
M = mass of pollutant
c = concentration of pollutant
Q = discharge
t = time interval
But application of this equation throughout a pollutograph dura-
tion would imply a constant concentration, c. This is seldom
the case in observed pollutographs.
In Figures 30 through 34 the concentration is made up of
two components, a uniform base concentration component, C , and
75
-------
q
cb
o
2
ID
a:
o
q q q
to
-------
o o
u_
o
z
ID
Q
O
U
O
•H
-P
0)
o
z
ID
cr
o
o
SdNDOd 901
ID
or
-P
•H
(U
-P
SQNnOd 901
77
-------
SONOOd 901
SdNflOd 901
o
CO
o
(£> LL
U_
O
Z
cr
a)
D^
O
^
-P
•H
OJ
•n
X
I
CO
a
•H
A
CO
G
O
-H
-P
-------
O UL
^ U_
O
(T
O
SQNflOd 901
u.
o
2
13
a:
o
o
C\
saisinod 901
5-1
-P
•H
•H
A
CO
C
O
•H
-P
U.
Lo_
O
tr
o
o
SdNnOd
901
u_
o
2
!D
a:
-p
•H
(U
-P
rti
ro
ro
in
•H
SQNnOd 901
79
-------
SdNnOd 901
901
CO
cu
D.
O
-C
I
ex
•r—
_cr
CO
O
•r—
03
IT3
oo
SQNnOd 901
80
-------
a second concentration exponent, C,, which varies exponentially
with discharge. This relationship can be expressed mathemati-
cally as
log M = C + C1 log (Qt)
The uniform base concentration is not an unreasonable concept
since most streams do tend to reach a constant concentration
during low flow periods.
The paucity of data precludes the derivation of any general
conclusions at this time/ but it is evident that the modeling of
mass flow rates (loadographs) are substantially facilitated by
this method.
In spite of the fact that logarithmic transformation is a
linearization process, it is believed that pollutant washoff, as
well as many other phenomena in nature, is an exponential func-
tion of its environmental parameters. Therefore it is not un-
expected that the logarithmic transforms would exhibit linear
tendencies.
The coefficients of the linear equations, as listed in
Table 10, are widely varying at each station and for each
parameter; but the exponent (slope) coefficient for P-10 and P-
30 is approximately unity. This would indicate a greater depen-
dence upon flow with increasing urbanization.
It should be emphasized that these equations are site spe-
cific and further data are necessary before any definite conclu-
sions can be derived.
The relationship between the slopes of the three lines
representing pollutant loadings at P-10, P-30 and H-20 is impor-
tant. P-10 and P-30 have almost similar loading rates with H-20
being greater by several orders of magnitude except in the case
of suspended solids amd COD where the differences are not as
great.
Some inferences regarding urbanization may be derived from
Figure 35:
1. The pollutant loading rate in Ib/ac per inch of runoff
is directly, but not always linearly, proportional to
increasing urbanization.
2. The variance in pollutant loading rates is roughly
similar for nitrates, phosphates and Kjeldahl nitrogen
for each type of land use. In the case of COD the
variance is not as marked and for suspended solids it
is almost negligible. The latter case may result from
severe channel erosion on Panther Branch.
81
-------
o
CO
0 I 2
INCHES OF RUNOFF
0
0 I 2
INCHES OF RUNOFF
INCHES OF RUNOFF
INCHES OF RUNOFF
o P-IO AP-30
x HUNTING BAYOU a SWALE 8
• WESTBURY SQUARE INCHES OF RUNOFF
Fig. 35 Total pollutant loadings as a function of runoff
82
-------
TABLE 10
WATER QUALITY EQUATIONS
Station Parameter Equation
P-10 SS
COD
KN
NO-
PO,
P-30 SS
COD
KN
NO
H-20 SS
COD
KN
NO-
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
log M =
0.84 log Qt
0.98 log Qt
1.01 log Qt
0.99 log Qt
1.00 log Qt
1.05 log Qt
1.08 log Qt
1.11 log Qt
0.98 log Qt
1.01 log Qt
1.47 log Qt
1.33 log Qt
1.07 log Qt
1.25 log Qt
1.08 log Qt
-.1.42
- 2.34
- 4.13
- 5.19
- 5.40
- 1.83
- 2.65
- 4.47
- 4.73
- 5.45
- 5.45
- 3.75
- 3.81
- 5.60
- 4.40
83
-------
3. Both linear and curvilinear relationships are repre-
sented indicating a nonlinear dependence of nitrates
and Kjeldahl nitrogen to runoff.
4. The one major data point for Swale 8 (D-50) shows that
pollutant generation from the Swale 8 watershed is
similar to that from the Panther Branch watershed.
The water quality predictive methods discussed in this
subsection are a novel approach to this problem. The available
data show definite trends but much more data are necessary
before general conclusions can be drawn. Consequently the
derived relationships are based on insufficient data and should
only be used after this restriction has been considered.
The modifications to the SWMM by EH&A to model water quality
are based on this new approach. As described in the following
subsection, Figures 30 through 35 facilitated the selection of
the appropriate coefficients to be used as input to the modified
SWMM.
WATER QUALITY MODELING
As stated previously, the present SWMM predicts the concen-
trations of suspended solids, BOD, total coliform, COD, settle-
able solids, nitrates, phosphates, and grease in storm runoff.
The basic theory used to predict these constituents assumes that
the amount of pollutant washed off in any time interval is pro-
portional to the amount remaining on. the ground. This results
in a first-order differential equation which integrates to the
following :
P P = P (1 . e-kt>
where
PQ is the initial amount of pollutant per unit area
P is the remaining amount of pollutant per unit area at
time, t and
k is the decay rate.
In the verification of the water quality predictive capa-
bility as described in the initial documentaiton of the SWMM
(9) , it was found necessary to add an availability factor which
yielded the following basic equation which is used with appro-
priate conversion factors to predict suspended solids and BOD:
Po - P = Ao Po (1 -e~kt>
where A is an availability factor which is defined as a per-
centage of the pollutant amount, P , that is available for
capture by storm runoff. Coliform densities are predicted
84
-------
directly by multiplying the suspended solids concentration by an
appropriate conversion factor. A more detailed discussion of the
overall procedure for predicting water quality constituents in
the SWMM is provided in the SWMM program documentation (9).
The indeterminate results of water quality data analyses
described in the previous subsection indicated that a radical
approach to water quality prediction computations in the SWMM was
required for this study. Obviously, the dust and dirt accumula-
tion rates developed for Chicago cannot be universally applicable.
Also the 4.6 value for the runoff exponent implies identical
rainfall intensities and wash off rates for all storms that are
modeled. This is not always true, especially in Texas where
rainfall intensities have a very wide range. Another compli-
cation arose from the fact that in undeveloped areas the lack of
streets and curbs makes water quality prediction difficult
because dummy curb lengths, based on average feet of curb per acre
of drainage area, had to be utilized. Also it was felt that
more than 5 land uses were necessary to adequately describe a
watershed.
A simplified approach to water quality prediction in SWMM
has been developed. Pollutant build up is not considered in the
modified SWMM. The revised model does not require input data on
dry days, street cleaning frequency, land uses, or curb length.
Instead the pollutant availability at the beginning of the storm
is input. The user can determine, external to the"model, the
effect of dry days, street cleaning frequency arid land use, while
curb length is no longer a parameter. Also the transposition of
data to different geographical areas becomes a user option.
Concurrent with the changes described above, the number of
land use options has been increased to 20. Because loading rates
for each land use are user input, any combination of land uses is
feasible. In a developed area, for example, all 20 land uses may
be of urban nature. The selection of land uses to any level
of detail is therefore possible. However, the program structure
presently only allows for transfer from the Runoff to the
Transport Block of information on only eight land uses. The
pollutant removal factor or wash off exponent, k, is now an in-
put parameter.
The water quality modeling scheme is accessed by using a
value of 2 for ISS in Card Group 9 of the Runoff Block. The
pollutant removal factor, k, is also input for each pollutant.
This allows for model flexibility in the case where pollutants
wash off a subcatchment at different rates for the same rainfall
or runoff. The loading factors in Ibs/ac are input for each
land_use specified and all pollutants being modeled. As in the
original SWMM only one land use per subcatchment is permitted;
also, the area of the subcatchment is substituted for curb length
on the land use data cards.
-------
The results from the water quality analyses as described
earlier have proved to be very helpful in selecting the pollutant
loading factors as well as the removal coefficients.
The cumulative pounds of pollutant versus cumulative discharge
curves, Figures 30 through 34, and the pounds of pollutant per
acre of contributing area versus inches of runoff, Figure 35,
can be used to select the input data for water quality predic-
tion in the new EH§A version of the SWMM. An estimated runoff
rate and the information in Table 11 can be used to determine
the removal coefficients.
The loading rates determined from Figures 30 through 34
are only approximate because of minor variability in the data
caused by factors not considered in these analyses, e.g., soil
characteristics, flora and fauna, etc. The approximate loading
rate determined from the figures can be verified by comparison
of the modified SWMM output to observed data; this procedure was
followed in all applications of the new water quality modeling
versi on.
Because the Transport Block of the SWMM can only route two
pollutants and coliform counts, the model would have to be run
at least three times at each station to develop output for sus-
pended solids, COD, nitrates, phosphates and Kjeldahl nitrogen.
In order to reduce this volume of computer operations, it was
decided that Kjeldahl nitrogen would not be routed through the
Transport Block. The modeling of Kjeldahl nitrogen, or any other
pollutant if desired, is only dependent upon the input data for
the EH&A version; for example, if the loading rate and removal
coefficient for nitrates are input as data for COD then the COD
results from the model may be interpreted as results for ni-
trates. This new capability of the SWMM enhances its scope of
application since any two pollutants (which are or may be treated
as conservatives) can now be modeled with every run. For the
purposes of this study the two pollutant pairs selected were sus-
pended solids and COD as the first pair, and nitrates and phos-
phates as the second.
As described in the following section water quality was
modeled for an observed storm at each of the three study areas.
Calibration of the new water quality model is a relatively
simple task. Modeling results based on initial loading rate and
removal coefficient estimates are used to refine subsequent
loading rates and removal coefficients until the observed pollu-
tograph is reproduced. [n applying the model it was determined
that pollutograph peak and loading rate were directly propor-
tional while pollutograph shape and removal coefficients were
inversely proportional. Consequently, increasing the loading
rate by a factor of 2 resulted in a pollutograph peak increase
to double the initial peak, and increasing the removal coeffi-
cient by a factor of 2 decreased the pollutograph duration to
one half the previous duration. The pollutograph peak was
86
-------
C/3
W
o
<:
Pn
f*
£>
C/3
H
W
S
EH
C/3
S
O
&
^
Q
W
>
0
s
co
EH
JZ;
<
S3
M
2
^
PQ
5
c
•H
cU
'H
«H
O
c
K
O
.
CD
O
m
0
o
CO
o
cJ
0
I—I
o
m
o
^^
05 ^
OS J3
,
cd
PH
e
o
CJ
0
J_(
cfl
Q)
CO
cu
ID
• •
g
o
S-i
M-i
87
-------
raised in conjunction with an increase in the removal coefficient
but a general relationship to quantify the rise could riot be
determined. Therefore the calibration runs are basically a
trial and error procedure to determine the ideal loading rate
and removal coefficient combination that would reproduce the
observed pollutograph. It must be emphasized that the loading
rate and removal coefficient thus derived are valid only for the
storm used for calibration. Application of these results to
other storms is possible only if prevailing antecedent conditions
and rainfall-runoff intensities are similar for both storms and
if the study areas are identical or at least homogeneous.
In analyzing the water quality modeling results, it became
evident that although the pollutographs could be modeled to
reproduce observed pollutographs the actual and computed mass
transport graphs did not correspond. It was determined that
this condition resulted from slight variations in the runoff
quantity model results. For example a slight difference between
observed and computed runoff rates concurrent to a high rate of
mass transport results in a large increase or decrease in pollu-
tant concentrations. Consequently pollutographs are highly
dependent on the accuracy of the hydrograph model output.
To improve the modeling of total mass loading of a pollu-
tant, the EH&A version was used to model pollutant mass flow
rates. Again the loading rate and decay factor were adjusted to
reproduce mass flow rates as determined from the observed flows
and concentrations.
The above discussion identifies the need for the user to be
completely aware of the modeling objectives. If pollutant mass
flow rates or total loadings are desired, then water quality
modeling is essentially independent of water quantity modeling;
but if pollutant concentrations are desired, then both quantity
and quality modeling results determine the accuracy of the
pollutant concentrations. Detailed analyses of each storm
modeled during this study are included in the following section.
In summary, the modified SWMM water quality model is very
capable because of its flexibility in determining pollutant
loading from various land uses. The user selected input data
determine the relative dependability of the output thus elimi-
nating some of the previous "black box" computations which
although generally applicable could not be adequately calibrated
for Stations P-10, P-30 and 11-20. It is expected that as more
water quality data become available the relationships in Figures
31 through 55 will be substantiated and verified.
-------
SECTION 8
MODEL APPLICATION
GENERAL CONSIDERATIONS
The specific types of data which are required as input to
the SWMM have been described in Table 1. This input information
is a quantified description of the watershed to provide a compu-
tational basis for the model. The basic model inputs required
are the rainfall hyetograph for the storm to be modeled, a phy-
sical description of each subcatchment to be modeled including
the drainage area, percent of impervious cover, ground slope,
Manning's roughness factors, estimated retention storage for
both the pervious and. impervious surfaces, and the coefficients
to define Horton's soil infiltration equation. Also required
are input data to define the hydraulics of the storm sewer
system for each subcatchment and for the main sewers or open
channels. These inputs include gutter length, slope, bottom
width, and roughness coefficient. For sewers and open channels
the cross-sectional area and side slopes, channel slope, and
roughness factor must be defined. For water quality modeling a
code defining the specific land use in each subcatchment as well
as the street-cleaning frequency, the number of dry days prior
to the storm event, the number of catch-basins per unit area and
the quality of their contents must also be specified.
An important portion of the input data concerns the coef-
ficients to be used in Horton's infiltration equation. This
equation is used to calculate the infiltration rate of rainfall
into the soil as a function of time by Horton's (8) relationship
as described in Section 6.
The initial and final infiltration rates, f. and f , and
the decay rate, k, were deduced from USGS rainfall-runoff re-
cords. This was done by calculating effective infiltration
rates for numerous storm events in the Houston area as listed in
Appendix C and plotting these values versus storm duration in
order to determine a graphical representation of Horton's equa-
tion as shown in Figure 36. As expected, it was found that the
initial infiltration rate was highly dependent on antecedent
soil moisture conditions.
In order to adequately describe the hydraulic efficiency of
the drainage system, it was necessary to choose Manning's rough-
89
-------
-------
ness coefficients for each drainage element. Gutters and open
channels were assigned an initial value of 0.10, while a value
of 0.03 was used for sewers. These values were adjusted during
model calibration to accomodate higher peak flows. Combined
sewers are not used in any of the study areas and therefore all
initial flow quantity and quality was set equal to zero.
As a prelude to modeling either Hunting Bayou or Panther
Branch, specific modeling criteria had to be determined. Sever-
al _ SWMM runs were made with regard to minimizing cost but re-
taining modeling validity. Data for a dummy watershed consisting
of two subcatchments, Table 12, were developed. The SWMM was run
for these data at 2, 5, 10, and 20 minute integration time
intervals, for average and extreme storm rainfall intensities.
The results of this sensitivity analysis, shown in Tables 13 and
14, indicated that there was no substantial gain in modeling
accuracy for the 2 and 5 minute integration time interval but
the cost increases for modeling at the 2 and 5 minute time
intervals was considerable. Consequently, it was determined
that a 10_minute time interval was the limit for modeling accu-
racy at minimum cost, and all further SWMM runs were made with
regard to this condition.
HUNTING BAYOU MODELING
Most of the input data concerning the Hunting Bayou drain-
age system was taken from existing engineering maps of the area,
and site inspection of the study area. A map showing the sub-
catchments and drainage network which were used as input to the
model is shown in Figure 37. Five storms were modeled initially.
The rainfall data for these storms, listed in Table 15, were
obtained from reports published by the U. S. Geological Survey
(20,21,22,23). No water quality data was available for these
storms because all five storms occurred during 1968 and 1970
prior to the initiation of this project. Consequently, only
water quantity was modeled.
As more recent storms on Hunting Bayou were sampled for
water quantity and quality, the water quality prediction capabi-
lities of the SWMM were tested. Three storm events have been
modeled. Rainfall data for these storms are also listed in
Table 15.
_The total drainage area of 799.86 hectares (1976.8 acres)
is divided into 24 subcatchments ranging in area from 10.11
hectares (25 acres) to 55.85 hectares (138 acres). Other signi-
ficant subcatchment data are listed in Table 16. Each subcatch-
ment was assigned a land use class for modeling water quality in
SWMM. These land use classes are shown in Table 17.
The drainage system includes 23 gutters and pipes in the
Runoff Block and 44 manholes and conduits in the Transport
91
-------
TABLE 12. SWMM INPUT DATA FOR SENSITIVITY ANALYSIS
SUBCATCHMENT DATA
Subcatchment No.
Width (ft)
Area (ac)
Percent Imperviousness
Slope (ft/ft)
Resistance Factor
Impervious
Pervious
Surface Storage (in)
Impervious
Pervious
Infiltration Rate (in/hr)
Maximum
Minimum
Decay Rate (in/sec)
Total Tributary Area (acres)
GUTTER AND PIPE DATA
Gutter Number
Width (ft)
Length (ft)
Slope (ft/ft)
Side Slopes
Left
Right
Manning n
Overflow (in)
WATERSHED QUALITY DEFINITIONS
Subarea Number
Land Use Classification
Total Gutter Length (100 ft)
Number of Catchbasins
Number of Constituents
Number of Dry Days
Street Cleaning Frequency
Passes Per Cleaning
STD Catchbasin Volume (cu ft)
Catchbasin Contents BOD (mg/1)
1
4000.0
400.
.5
0.15
.400
.250
.062
.184
.75
.05
.00115
2
4000.0
400.
1.0
.009
.300
.250
.062
.184
.75
.05
.00115
800.00
1
6.0
2000.
.001
3.0
3.0
.050
90.00
2
6.0
2000.
.001
3.0
3.0
.050
90.00
1
1
20.00
0.00
8
7.
0.
0
-0.
-0.
0
0
00
0
2
4
20.
0.
00
00
92
-------
- • CM CMOrHin r- rH -. o CO O
co vo vo vo in in
r^ CM
rH !->. O O OOOO mCM-^-rH
-------
PQ
W
CO
<
CJ
rt
o
En
o
S
I—I
rJ
W
Q
S
S
CO
fc
O
S
M
>
M
H
M
CO
23
W
CO
st
rH
r-3
pQ
H
O
O
o
o
m
o
m
1-1
o
o
rH
O
m
*
•H
<
CD
CO
cd
o
J_l
o
CO
rH
I
CD
-P
CJ
•H
CD
•U
P
•H
S
O
CN
P
rH
rH
cd
^H
•rH
cd
K
O
o
rH
O
o
o
o
st
O
o
st
O
O
*
O
o
•
CM
•H
PQ
CD
CO
cd
CJ
J_4
O
MH
CO
rH
I
4J
•H
CD
•P
P
CJ
•H
a
0
CM
%
rH
rH
MH
•H
cd
&4
CO
CD
-P
-H
6
o
o
V?"
w
•H
CO
B
to
rH
rH
cd
o
CD
a
•H
P
60
•H
rH
OJ
0
s
rH
cd
-P
O
st O
-^"
st st in
f^. rH O
rH .-H
.
P
MH
P
P U
a- ii"°
0 • ^
10 3
^o ° C
Sm .3
. S d
i — i cd
rH - }H
cc MH P
M- MH ,H
C 0 -H
•H CJ -H
cd 3 c
cr.
CD
0
rH
JD
>
O O CO O
vO vO i — 1 vO
ON rH rH
CM rH
st
CM r*. oo rH
St" St rH
CO rH
VO rH in O
r^ oo ON co
m st rH
CO rH
CN 00 00 CN
CO OO CM CO
r^- in rH
CO rH
oo o cy> st
oo rH r^. CN
uO rH
O
•.n
oo CN in in
St rH rH CN
00 CN
rH
uo
o co r-~ o
CJN rH CO CM
00 CN
rH
m
st st m o
vO rH
-------
LEGEND
STUDY AREA BOUNDARY
SUBCATCHMENT DIVIDE
MM SEWER PIPE
•••i OPEN DITCH
O MANHOLE
51 MANHOLE NUMBER
KXX) 500 0 1000 2000
Fig. 37 Subcatchments and drainage network - Hunting Bayou
95
-------
o
en
oo
inoooooooooooinooinoooocncN
i—lOOOCNCNOOOOr-I i—I >t i—li—(CNOOOO
O
CO
oo
LO
cfl
•H
r-l
Cfl
a
Q
W
ffi
CO
vo
CM
O
CO
oOOOOOOOOOOOOOOOOOOOOCOOO
O
>^
<
CQ
H
2
CN
CN
oo
O
g
rJ
rJ
oo
VD
m
00
r^
o o o o oo r^
rH ON
-------
„. x—x
^ rj ^" "" IJ' *• ' u> IJ">> "^ ^O u~^ lO IO LO to LO tO LO IO LO LO LO IO to lO IO
«0^ '--* rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH rH
J2cCrt, '-''-IrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrH
05 ^^ ~ oooooooooooooooooooooooo
r- C Q OOOOOOOOOOOOOOOOOOOOOOOO
iJ .
f3 /-N C rHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHrHi—IrHrH
£^ ^^ OOOOOOOOOOOOOOOOOOOOOOOO
rH ;£* *
'H C "
n
2? fcl oooooooooooooooooooooooo
to cu oooooooooooooooooooooooo
J~: W P^ •••••••••••••••«»•••»,..
W
U
< CO f-( CM CN fS CSJ (N CSJ CM CsJ CN CN CN CN CM CM CM CN CN CN CM CSJ OJ CM CN CM
>i- UH QJ ^^vOOvOOOvOvOvDvOvOvOvOvOvOvOvOvOvOvOvOvOvO
o - =* oooooooooooooooooooooooo
Oc/3 M ° ' ° .......... ......
>*
<
CQ
r5 r* oooooooooooooooooooooooo
^ ^ .^^ini-oin«niriinir)inir)inLnininininininininininin
• (U jeNCS
•H tfl r-i
rrr-r-r
g- oooooooooooooooooooooooo
........ '
25
W UQJ oooooooooooooooooooooooo
r-io* co»incocoo<^cocounr^cr>^(»cor^<^c^ONONONOsc\ON x-C
Ot r- IrHCNrHrHCNICMrHrHcSrHCNr-lrHiHrHCNJCSJCNCMCNCSICNCN CO
^ M 0)
'-rTffr rHrH r-i C3
ooooooooomoooooooooo co
•^ CS r~IP-)CNJrHiHcvJrH i— I rHrHr- IrHCMrHrHCNCNCS 3
i . 5
rC • t,
o o £
jj 2 "
co ^«N<^<^invor^ooaNO'--fcNf^^inor>>.ooQNQr_4cNifn>d- rH
W 4J rHrHrHrHrHiHrHrHrHrHCNCNjeNCMeM CO
i-Q C II
31 I
97
-------
TABLE 17
LAND USE DATA, HUNTING BAYOU WATERSHED
Subarea Land Use
Number Class.
1 1
2 1
3 1
4 1
5 3
6 1
, 7 1
8 3
9 1
10 3
11 5
12 3
13 3
14 1
15 5
16 1
17 1
18 1
19 3
20 3
21 1
22 1
23 3
24 1
Total Gutter
Length (100 ft)
59.00
64.00
24.00
79.00
96.00
155.00
61.00
30.00
90.00
35.00
22.00
31.00
20.00
45.00
68.00
55.00
134.00
60.00
99.00
35.00
41.00
113.00
86.00
78.00
Number of
Catchbasins
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Land Use Class is defined as follows:
1. Single and multi-family residential areas
3. Business and commercial activity areas
5. Undeveloped urban open land
98
-------
Block Table 18 lists all the gutter and pipe data and Table 19
shows the Transport element characteristics.
The subcatchment, gutter and pipe, and transport element
data were essentially identical for all runs. The storm related
TabL^O ?n^?faJ- ' ^^ 15' ^ infilt-tion coefficients
Table 20. Infiltration rates were estimated from the observed
rainfall and runoff. Initially, the antecedent precipitation
index was used to establish a starting infiltration rate but
this parameter proved to be unreliable as a deterministic pre-
dictor and was subsequently discarded. Therefore all infiltra-
tion rates were originally estimated and then calibrated bv
consecutive modeling runs.
Comparisons of observed and computed hydrographs for five
st°rm rtevents which were modeled are presented in Figures 38 39
and 40. A comparison of observed and computed total runoff'
volumes and peak flow rates is shown in Table 21. The overall
agreement is reasonable. The average absolute error in volume
?n run°" was 26V°f the observed value, while the average error
in peak flow prediction was 20% of the observed peak. The tem-
poral agreement of the hydrographs was very good. For instance
the times of peak flow agreed within ten minutes in four of the'
five instances. The average error was twenty-two minutes. How-
ever, the computed values tended to predict faster returns to
low flow conditions than were actually observed.
rig^al SWMM Water ^uality Predictions for the storm
u ° n0t reProduce the observed data; Figure 40 shows
the observed and computed results for suspended solids at Sta-
tion H-20. The EH&A version, with loading rates as shown in
Table 22, was run for the same storm and the predicted results
for suspended solids, COD, nitrates and total phosphorus are
compared to observed data in Figures 41 through 44 and in Table
J3. As discussed in the preceding section, even if the pollu-
tographs are adequately reproduced the pollutant mass transport
rates are not sufficiently high. Consequently the pollutant
mass transport rates were computed and compared to pollutant
mass transport rates as determined from the observed discharge
rates and concentrations. The total observed and computed
pounds of pollutant transported during the storm are also com-
pared in Table 23.
As seen in Figures 41 through 44, the rather large temporal
difference between observed and computed pollutographs is mini-
mized in the pollutant mass transport rate graphs; the reduc-
tions ranged from 75% for suspended solids to 100% for nitrates
With the exception of COD, the predictions of total pounds of '
pollutant removed were also improved by the use of mass flow
rate graphs. In the case of COD, an extremely high pollutant
removal factor seems to be necessary to reduce the total pounds
prediction. No physical explanation for this phenomenon could
99
-------
p
w
ffi
en
Prf
w
H
O
S3
M
H
H
<
O
W
PL^
M
PM
Q
r-l
O
bfl
C
•H
cn
0) -U
ooooooooooooooooooooooo
ooooooooooooooooooooooo
^^
^^N
0
o
^
o
rH
O
rH
O
o
ON
o
o
^
o
on
0
on
o
0
O ON ON
ON ON ON
ON O O
CN O 0
O rH rH
O O O
o on on
i
o o o
o on on
o o o
o o o
CNI
ON
CN
o
0
0
i
o
o
1
0
o
00 ON
on ON
ON O
CN O
O rH
0 0
o on
i
0 0
o on
i
O 0
O 0
3" >3"
ONOONONOONOONON
CNOCNCNOCNOCNCN
OrHOOrHOrHOO
OOOOOOOOO
OrHooonoonoo
1 II 1 II
OOOOOOOOO
OrHooonoonoo
i it i it
OOOOOOOOO
OOOOOOOOO
CN
"*
ON
CM
o
o
0
1
o
o
o
o
vO
on
ON
CN
O
o
o
1
o
0
1
o
o
o
VQ
ON
CN
0
O
O
1
O
o
1
0
o
ON
ON
O
O
•"^
O
VO
O
^D
0
O
CN
st
ON
CN
O
O
O
1
O
o
1
o
o
00
sfr
ON
CN
O
O
O
1
O
O
1
o
o
o --•.
rH4J
en M-I
•P X-N OOOOOOOOOOOOOOOOOOOOOOO
604J tOOOOOOOOOOOOOOOOOOOOOOO
Q)
rJ
i— llOOnCNCNOnCN|CN|CNOnonCN
>d-rHrH
CNI
CNCNCNCNJ
w
H
H
oo
rH
w
PQ
H
•U
§
•H
4-1
oomooocNoooomomoommooomo
43
O
4-1
•H
Q
43
0
4J
•H
Q
CO O
PH W
•H -H
P* Q
43
O Q)
4J £LI
•H 'H
(H P*
CO
a
•H
PM
43
a
4J
•H
Q
0)
a
•H
PM
a
4J
•H
Q
CO
PH
•r-f
PM
CO O
P- u
•H -H
PM Q
43 43
a>ococococo coo
PH 4-J ("N! Q^ p^, Q^ f"), 4-)
•H *H *H *iH -H *H *H •rH
PH Q PM PH PM PM PM Q
CO CO
PH PH
•H 'H
P-i PH
r-l r-l
CO CO
•I-1 43
4-» g
3 P
O 2
CNI on
m
00 ON O
rHrHrHrHCNlCNCNJCNJ
100
-------
r-l M-l
3
fj-i rr
O!lfO;£O^o^o^o^OTOO^OOOO--HOinOOCMinOCMOOOOOOOvl-o'
rHcooocTiinc^rH r---r^or---oooooooooo o
in m m CM CM co
0 CU CL) bO CU 0) CU
J r" r^ i i r*4 r-1 r- ' r-* r- ' /— '
CU 0) CU cO CU cO 0)
cdoc.oos'o'^'o
cO cu
3 O
0)
0
a; cu a)
rH r-H r— 1
O O O
d o d u
Od 01 CTj CU
udodcjdcjdududd
cuci3cucdcuc.cucx!oinja»cflcx!
§ s 's'ss s
101
-------
TABLE 20.
STORM
DATE
9/08/68
9/17/68
11/09/70
3/26/74
5/08/75
STATION
H-10
H-20
H-10
H-20
H-10
H-20
H-20
H-20
INFILTRATION PARAMETERS,
HUNTING BAYOU WATERSHED
INFILTRATION RATES
Initial Final Decay
in/hr in/hr /sec
1.00
1.00
0.75
0.75
2.50
2.50
0.10
0.30
0.10
0.10
0. 10
0.10
0.10
0.10
0.02
0.10
.0005
.0005
.0005
.0005
.0005
.0005
.0005
.0005
102
-------
CM
UJ
(sp) 39edVHOSIQ
C\J
OJ
(O
oo t
UJ
O
O
00
or
o
H
CO
Q Q
LJ UJ
«?
UJ ^
CO ^
00 O
o o
00
(0
u_
o
or
o
i-
co
o
CM
O
CO
a i
CJ
o
CM
CO
8 8
CM
39dVHOSId
LU
00
UJ
OvJ
o
O
O
O
00
O
CN
I
ffi
i
tc
c
o
•H
-p
fd
-p
en
03
Cn
O
CO
ro
103
-------
100
75
LJ
£ 50
O OK
10 12
TIME (hrs)
STATION
H-IO
14 16
STORM OF 11/9/70
6 8 10 12
TIME (hrs)
STATION
H-20
STORM OF 11/9/70
OBSERVED
COMPUTED
14 18 22
TIME (hrs)
STATION
H-20
STORM OF 3/26/74
26
Fig. 39 Hydrographs at Stations H-10 and H-20
104
-------
2 3 4 5 6 7 8 9 10 II 12
STREAM
FLOW
DISCHARGE
STATION
H-20
STORM OF 5/08/75
OBSERVED
COMPUTED
2100 22 23 24 I 23 456789D II 12
SUSPENDED
SOLIDS BY
ORIGINAL SWWM
VERSION
STATION
H-20
Fig. 40 Hydrographs and suspended solids concentrations
at Station H-20
105
-------
o
K"-"
CX i-l
6
o
CJ
12
0
*
w *"d
PU Q)
> .-1
M CM
Q) i— 1
0
TJvO
QJ 0
•U rH
3 X
Pi in
6 oo
O
W U rH
m in in in o
m in vo CM CM
CO i— 1 CO i— 1 CM
m <± co m i—i
CM vj- CO OO VO
CO i— 1 CO rH
VO v^ VO vO VO
o o o o o
rH rH i— 1 rH i— 1
X X X X X
ON - in
vO CM O ON vO
. . .
<± CM vO O CM
00 O
CO OO
O CO
p^ J-l
0)
0)
n
O
a
o
•H
4-)
cd
o
o
hJ
sO
0
rH
X
CO
«
-------
4-J
£
fd
-p
4-1
C
CD
•H
a
•rH
CTN O O
LO o o
rH CN CN
O O O O
m o o co
CO CO CN CN
CD
C
o
•H
4J
O
3
rH
4J
CO
O
CJ
•H rH
T3 Cd
Cd -H
O 4-J
J C
CD
CO
CD
4-J
cti
4J
O
PM
O
co
C
O
•r-
CJ
o
rH
CX
CD
Pd
P.
td
r-l
w
o
4-J
rH
rH
O
PM
CN
•
O
O
rH
CO
.
CN
O
in
oo
rH
O
o
co
.
CN
O
0
CO
CN
0
<}-
in
CO
rH
O
o
o
rH
O
o
o
CO
o
o
o
CN
O
0
O
CO
o
o
o
oo
o
0
o
CO
CO
*"O
•H
rH
0
en
"O
CD
T3
a
CD
CX
CO
p
en
Q
O
CJ
rH
fd
4-J
o
EH
CO
0)
4-J
cd
j_i
4-1
•H
2;
rj
O
rH
0
a
CO
o
PM
rH
cd
4-J
O
EH
£
c
4->
C
"C
f*
H
ft
(D
«
/n\
UJ
4-J
fd
PS
-P
o
a.
CO
fd
!H
EH
CO
rd
f-t
4->
£
fd
\_.
i
r— 1
O
(X
o
o
in
o
o
0
0
in
0
o
^'
CO
T3
•H
rH
O
Cn
T3
CD
'd
(3
CD
cx
CO
p
en
O CO CO
o o o
rH 0 0
in o o
O CO 4-J 4-J
O -H 0
EH S H
CO
CD
e
o
43
>i
rH >1
•H -P
e-H
fd >
^ -H
1 4-)
•H O
-P fd
rH
3 S3
e o
•H
T3 4-J TS
C U C
•• fd 3 fd
CO ^ rH
^ CD 4-»
O rH CO T3
rH CT> $3 CD
rH $3 O rJQ
O -H U rH
-— MH co 3
fd 43 4J
43 CO 43 4-> CO
\ ($ 4J ->-\ -rl
S— (U CO 3
$3 co fd
II -H fd CD C
MH CD H CD
^ CD M fd a.
LO T3 fd O
• CD 1 1
ID }H C
^ fd rH O 13
fd -H CD
X CO -H 4-> CX
CD 4-1 U O
— CO £ ^ rH
O 3 CD M CD
fd 13 4-J j>
\ 13 -H CO CD
45 (3 CO C T5
rH fd CD O £
— H" tf U D
• i
H
EH
0
3
107
-------
in
»
^s
oo
o
in
fa
O
§
O
EH
03
1
D
O
CQ
O
J2J
H
2
D
&
O
fa
en
^q
D
cn
C£
O
2
H
W
Q
O
^>
>H
EH
H
H ft X rH H m r^
Q CN CO
w o
r3 -u
*§ ^ CO -S
WO TO CO E
frt n fS ^ *rT * '
r^ * ^j
M^ r^ r^
1 t
^ 0 hJ
J*1^,^ •
&P vO 00
S m r*«
p r*^.
rH ^4 p^4 ^>^ O v»O vO
rH (^ ! rH ^^" vO
ft fS
^ O nJ
co C **^
OJ Q 60 •
ft O b rH rH
^T) S
>sT VO
TO G CN
•u 3 o
O O rH •
EH ft X csj O
r- i —
•a
cu
> CO
M w
<1J TO C
CO Q ^ CO -H
^ TO CO 6
O OJ TO < . .
ft s ^i T vO
CO
'O
S =
& s
• •
CO t—
rH in
• •
CN O
rH CO
• •
m T-H
m o
rH in
CN CN
rH CO
m co
O rH
m VJD
v>O p>»
rH
o
•02
II 'cT*
VO
m u
r- -^
N^ fVJ
X CN
'c x
•H
\*CQ
rH rH
w
EH
O
z
108
-------
(VI
(wiii/qi) M0~ld SSVIM
to
•H
rH
o
o
Q)
CO
CO
O
CN
I
ffl
c
o
•H
4-)
4J
CO
I
m
oo
o
in
O
e
o
4-1
C/)
NOIlVdlNBONOO
109
-------
o
o
o
h-
o
Q
O
U
O
CN
I
ffi
c
o
•H
4-1
id
-p
I
in
oo
o
in
O
g
O
CN
tn
•H
NOIJLVU1N30NOO
110
-------
o
CN
I
S3
C
O
•H
4-1
fd
4J
CO
I
in
CO
o
in
4-1
O
O
4->
CO
ro
Cn
•H
o o o o
5f: 10 cvj —:
NOIlVaiN30NOO
111
-------
-12 ^ ^ o
MCTId SSVIN
O op -i fd
QJ
n -P
M-l
CO 0)
(U iH
4->
(0 (U
o c
i—I O
CO £
CO O
E co
CO
(D
-P
fd
O
-------
be determined. The predictions of total pounds of pollutant re-
moved were improved 67% for nitrates and 75% for total phosphorus.
PANTHER BRANCH MODELING
The data input to the SWMM for the Panther Branch drainage
system was developed from existing engineering maps and numerous
site inspections of the watershed. Because of the low relief in
topography, drainage area boundaries had to be visually determined
in some areas.
The modeled drainage system for Panther Branch is shown in
Figure 45. The total drainage area of 874.40 hectares (21607
acres) is divided into 57 subcatchments ranging from 8.5 hectares
(21 acres) to 552.80 hectares (1336 acres) as shown in Table 24.
Other subcatchment data are also shown in Table 24. The input
parameter called "width of subcatchment" is dafined as the width
over which overland flow occurs. Values for this parameter were
first estimated by the method described in the SWMM User's Manual
(2). These values were subsequently reduced by approximately
40% to achieve calibration. The land use classes for each sub-
catchment used for water quality modeling are listed in Table 25.
The Panther Branch drainage system is made up of 57 autters
whose characteristics are listed in Table 26, and 61 transport
elements as described in Table 27.
Five storm events on Panther Branch have been modeled.
Similar to the data for Hunting Bayou, all subcatchment, gutter
and transport element, data for Panther Branch were identical for
all runs. The storm related data comprised of rainfall and
infiltration coefficient data are listed in Tables 28 and 29,
respectively. Infiltration rates were determined similar to those
for Station H-20.
Observed and computed total flow volumes and peak flow rates
for five storm events, which occured between 10/28/74 and 12/10/74,
are compared in Table 30. Water quality data were available for
the storms of October 28, 1974 and December 5,1974. The SWMM
was used to model both water quantity and quality for these two
storm events and only quantity of flow for the remaining three.
Comparisons of observed to computed hydrographs are presented in
Figures 46 and 47. The computed flow peaks and volumes agree
well with the observed flows; the average absolute error in the
volume of runoff was 14 percent of the observed peak. The tem-
poral distribution of runoff between observed and computed hydro-
graphs was good except for the storm events of October 28, 1974
and November 24, 1974, when the flow peaks between observed and
computed hydrographs were approximately three hours apart.
An inspection in March 1975 revealed severe erosion at the
area being cleared for Lake Woodlands on Panther Branch. A re-
taining wall that had been built to keep Panther Branch within
113
-------
Fig. 45 Subcatchments and drainage network - Panther Branch
114
-------
TABLE 24. SUBCATCHMENT DATA, PANTHER BRANCH WATERSHED
Subcatch- Width
ment No. (ft)
1 4680.0
2 2100.0
3 3240.0
4 3540.0
5 5280.0
6 1620.0
7 7800.0
8 4380.0
9 2340.0
10 6660.0
11 4380.0
12 3600.0
13 2400.0
14 3780.0
15 4140.0
16 4680.0
17 1920.0
18 3780.0
19 4980.0
20 5280.0
21 6600.0
22 2460.0
23 10560.0
24 6480.0
25 5820.0
26 2400.0
27 7320.0
28 4200.0
29 2100.0
30 2280.0
31 1920.0
32 3420.0
33 2400.0
34 900.0
35 1080.0
36 3900.0
37 1800.0
38 1980.0
39 3780.0
40 3300.0
41 2640.0
42 7620.0
43 1800.0
44 1080.0
45 540.0
46 840.0
47 2400.0
48 3600.0
49 2460.0
50 2040.0
51 1800.0
52 2040.0
53 1980.0
54 12500.0
55 2400.0
56 3780.0
57 1300.0
Area
(ac)
534.
99.
153.
405.
578.
243.
899.
338.
195.
584.
920.
248.
393.
275.
609.
353.
598.
598.
998.
612.
825.
244.
851.
564.
870.
52.6.
1366.
390.
140.
101.
256.
421.
148.
26.
301.
297.
151.
81.
257.
214.
354.
998.
96.
175.
21.
39.
169.
448.
99.
162.
266.
281.
158.
584.
99.
374.
61.
Percent
Imp er v .
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
l.'O
1.0
1.0
1.0
1.0
1.0
1.0
1.0
3.0
1.0
Slope
(ft/ft)
.009
.006
.005
.009
.009
.006
.016
.014
.015
.015
.012
.013
.013
.016
.011
.018
.020
.015
.015
.015
.010
.030
.018
.020
.008
.003
.008
.013
.013
.012
.009
.012
.008
.008
.008
.008
.011
.015
.011
.022
.013
.018
.012
.012
.012
.008
.013
.013
.025
.019
.010
.011
.028
.020
.040
.012
.013
Resistance
Factor
Imperv. Perv.
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
-200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
•200 .400
•200 .400
.200 .400
.200 .400
.200 .400
•200 .400
.200 .400
.200 .400
.200 .400
.200 .400
.200 .400
. 200 . 400
.200 .400
Surface Storage
(in)
Imperv. Perv.
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
•005 .050
•005 .050
.005 .050
.005 .050
.005 .050
.005 .050
.005 .050
•005 .050
.005 .050
•005 .050
•005 .050
.005 .050
•005 .050
•005 .050
.005 .050
.005 .050
•005 .050
•005 .050
.005 .050
•005 .050
.005 .050
.005 .050
.005 .050
•005 .050
.005 .050
•005 .050
•005 .050
.005 .050
.005 .050
Infiltration Rate
(in/hr) (in/sec)
Max. Min. Decay
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
•50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
•50 .01 .00115
•50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
.50 .01 .00115
Total Tributary Area (acres), 21606.70
115
-------
TABLE 25. LAND USE DATA,
PANTHER BRANCH WATERSHED
Subarea Land Use Total Gutter Number of
Number Class. Length (100 ft) cSSShbLins
2 I 50.00 0.00
3 | 30.00 o.OO
t I 40.'00 Q.'OO
6 ? 55.00 o.OO
7 5 22.50 o.OO
Q ? 130.00 o.OO
f * 47.50 o.OO
in I 35.00 o.OO
i? 5 100.00 o.oo
£2 f 62-50 0.00
}o 5 31.00 o.OO
14 ?- 25.00 0.00
tt 5 41.00 o.OO
i? * 4o.oo o.oo
J7 ^ 46.00 o.OO
to 5, 27.50 o.OO
if I 70.00 o.OO
on I 50-°° 0.00
2? ^ 55-°° 0.00
oo ? 71.00 o.OO
?? ? 30.00 o.OO
2? ? 85.00 o.OO
oc I 60.00 o.OO
o5 ? 62.50 o.OO
o° 5 40.00 o.OO
H \ 82.50 o.OO
OQ ? 57-50 o.OO
?? I 22.50 o.OO
?? ? 22.50 o.OO
?J f 80.00 o.OO
^ 5 35.00 o.OO
\l I 24.00 o.OO
^ I 10-00 0.00
^ 5 20.00 o.OO
?? f 60.00 o.OO
?I 5 20.00 o.OO
\l \ 30.00 o.OO
I? \ 40-oo o.oo
?? ^ 42.50 o.OO
fi f 50.00 o.OO
J? ^ 97-50 o.OO
t 5 30.00 o.OO
ft 5 15.00 o 00
II I 10'00 0.00
IS 5 10.00 o.OO
is i 15-°° °-Oo
H 5 35.00 o.OO
en c 40-°° 0.00
I? 5 35.00 o.OO
^2 I 22-50 0.00
\\ 5 30.00 o.OO
II \ 25-00 o.OO
|t 5 62.5o 0>OQ
c^ f 15.00 0>00
5567 \ 50.00 o.OO
57 5 20.00 o.OO
Land Use Class is defined as follows:
1. Single and multi-family residential areas
5. Undeveloped urban open land
116
-------
TABLE 26. GUTTER DATA, PANTHER BRANCH WATERSHED
Gutter
Number
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
Width
(ft)
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.0
6.0
5.0
5.0
5.0
5.0
5.0
5.0
5.0
6.0
6.0
5.0
7.0
8.0
5.0
8.0
8.0
5.0
8.0
5.0
5.0
8.0
Length
(ft)
5000.
3000.
4500.
4000.
5500.
2250.
13000.
4750.
3500.
10000.
6250.
3100.
2500.
4100.
4000.
4600.
2750.
7000.
5000.
5500.
7100.
3000 .
8500.
6000.
6250.
4000.
8250.
5750.
2250.
2250.
8000.
3500.
2400.
1000 .
2000.
2700.
2000.
3000.
4000.
4250.
5000.
9750.
3000.
1500.
1000.
1000.
1500.
3500.
4000.
3500.
2250.
3000.
2500.
6250.
1500.
5000.
2000.
Slope
(ft/ft)
.003
.002
.003
.004
.002
.009
.003
.003
.007
.005
.003
.003
.009
.001
.003
.002
.006
.004
.003
.005
.002
.003
.002
.003
.001
.003
.001
.001
.013
.004
.005
.003
.001
.001
.004
.001
.001
.007
.001
.006
.005
.004
.002
.004
.003
.001
.010
.001
.001
.007
.001
.001
.006
.001
.003
.005
,001
Side
L
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
5.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
Slopes
R
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
5.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
6.0
Manning
n
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
.200
Overflow
(in)
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
60.00
117
-------
TABLE 27. TRANSPORT ELEMENT CHARACTERISTICS,
PANTHER BRANCH WATERSHED
Ext.Elem
Number
4
5
8
9
12
14
161
16
56
25
27
21
23
24
29
28
30
34
134
35
37
43
44
46
47
49
50
52
551
55
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
100
93
94
95
96
97
98
99
Description
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Ditch
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Manhole
Slope
(ft/ft)
.0038
.0018
.0025
.0020
.0026
.0012
.0022
.0022
.0008
.0008
.0012
.0023
.0018
.0028
.0014
.0014
.0036
.0008
.0008
.0009
.0006
.0030
.0006
.0010
.0012
.0010
.0007
.0007
.0007
.0007
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
.1000
Distance
(ft)
4000.0
5500.0
4750.0
1500.0
3100.0
4100.0
1000.0
3600.0
1000.0
6250.0
8250.0
7100.0
8500.0
6000.0
1200.0
5750.0
2250.0
3000.0
3000.0
2000.0
4000.0
1000.0
1000.0
3500.0
4000.0
2250.0
3000.0
6250.0
1100.0
900.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
Manning
n
.1800
.1800
.1800
.1800
.1800
.1800
.1800
.1800
. ieoo
.1800
.1800
. 1800
. 1800
. 1.800
. 1800
.1800
. 1800
.1800
.0800
.0800
.0800
.0800
.0800
. 0800
.0800
.0800
.0800
.0800
.0800
.0800
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0150
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
.0130
Georal
6.0
6.0
6.0
6.0
6.0
8.0
8.0
9.0
9.0
9.0
9.0
6.0
6.0
6.0
7.0
9.0
6.0
9.0
10.0
10.0
10.0
6.0
9.0
10.0
10.0
10.0
10.0
10.0
6.0
10.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
Ceom2
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
Ceon>3
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
-0.0
« of
Barrels
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
1.0
AFull
(sq ft)
554.062
554.062
554.062
554.062
554.062
984.998
984.998
1246.639
1246.639
1246.639
1246.639
554.062
554.062
554.062
754.139
1246.639
554.062
1246.639
1539.060
1539.060
1539.060
554.062
1246.639
1539.060
1539.060
1539.060
1539.060
1539.060
554.062
1539.060
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
OFull
(cfs)
478.268
329.167
387.927
346.972
395.609
578.816
783.721
1072.922
646.996
646.996
792.405
372.086
329.167
4L0.543
437.894
855.895
465.512
646.996
1927.987
2044.939
1669.686
956.144
1260.709
2155.555
2361.292
2155.555
1803.467
1803.467
461.861
1803.467
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
QMax
(cfs)
478.268
329.167
387.927
346.972
395.609
578.816
763.721
1072.922
646.996
646.996
792.405
372.086
329.167
410.543
437.894
855.895
465.512
646.996
1927.987
2044.939
1669.686
956.144
1260.709
2155.555
2361.292
2155.555
1803.467
1803.467
461.861
1803.467
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
0.000
118
-------
a
w
w
EH
U
S3
PQ
W
ffi
EH
EH
<
a
r^i^i^OOOOOOOOOOUIi^OOOOOOOOOOOO O OOOO O
SmoooJ-ooSorioooo^mooooooooooop p oooo r*
ls~5B5£!SS22222ggggSSggggggggg§ g § § S S 5
rw *N
o
*-, r-»OOOOO''OOer*OOOOOOOOOOOOOOOOOOOOOO,*rtin »
S ^, OO oOjOOOi^OOOOOOOOOOOOOOOriO •>?
s,nssgssssgggggggggggggggggs *
^o £^t»,e^a>^i«ooooooooooooooooooo >c
~lSfMr^r-jrg«>'OOOOOOOOOCOOOOOOOOOO^l «">
S*rlOeM^»JOOOOOOOOOOOOOOOOOO ^
-s
j5StM-?!?'-'-JOOOOOO-J — —lOOOOOOOOO -^
^^H!?^—*^-*c'ooooo»— —**-COOOOOOOO "^
53
H
00
(N
W
^
PQ
<
EH
Ceo
119
-------
TABLE 29. INFILTRATION PARAMETERS,
PANTHER BRANCH WATERSHED
STORM
DATE
STATION
INFILTRATION RATES
Initial Final Decay
in/hr in/hr /sec
10/28/74
11/10/74
11/24/74
12/05/74
12/10/74
P-10
P-30
P-10
P-30
P-10
P-30
P-10
P-30
P-10
P-30
3.5
3.5
0.3
0.3
2.0
2.0
0.5
0.5
0.2
0.2
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
0.01
.0005
.0005
.00115
.00115
.00115
.00115
.00115
.00115
.00115
.00115
120
-------
TABLE 30. HYDROGRAPH MODELING RESULTS,
PANTHER BRANCH WATERSHED
Date of Storm
Total Runoff
(ft3 x 106)
Observed Computed
Peak Flow Rate
(cfs)
Observed Computed
10/28/74
11/10/74
11/24/74
12/05/74
12/10/74
P-10
P-30
P-10
P-30
P-10
P-30
P-10
P-30
P-10
P-30
24.40
39.34
64.48
72.87
52.24
73.70
36.06
45.52
44.42
51.73
29.03
36.16
53.44
73.61
57.72
78.97
32.66
48.55
33.61
43.02
342
376
979
897
680
774
273
329
464
517
i.
360
410
600
705
645
735
315
370
380
425
121
-------
(sp) 39dVHOSIQ
CVJ
888°
ro cvj —
(sp)
8888
^- ro co —
(sp) 39dVHOSId
o
CO
I
Tl
C
(d
o
iH
I
c
0
•H
-p
-p
rd
O
n
TS
>i
ffi
o o p
8 S 8
8
CVJ
39dVHOSIQ
CVJ
CVJ
(sp) 39HVHDSIQ
a Q
u ui
CD O
O O
Cn
•H
CM
122
-------
\
CM
24
rs
12
ME
8 8
8 8 8
io CM —
39dVHOSIQ
in
o
\
CM
OF
STOR
OBSERVED
COMPUTED
74
STORM OF 12
CM
CM
LU
CM
CO
OH
13
fd
O
H
I
P«
CO
£
O
•H
-P
rd
-P
W
-M
fd
CO
,C
O4
rd
h
Cr>
O
M
Ti
>i
ffi
47
tn
•H
Cn
CM —
39dVHDSia
123
-------
the original channel had been eroded away. This allowed Panther
Branch to change its course and flow through the construction
area which was being used as a sand pit. Downstream from the
area there was heavy deposition of sand that had been washed
away from the site. Erosion from consturction areas was the ma-
jor source of suspended solids observed at P-30. During the
study, peak concentrations of suspended solids declined at
Station P-30. Further storm sampling may determine whether or
not the erosion continues to be reduced.
A major drawback in applying the SWMM to The Woodlands is
that the area below Station P-10 was in a transient state due to
the development of Phase I. The continually changing land use
affects the quality of runoff. For this reason the decision was
made to regard Station P-10 as a control point. The area above
this gage is in a relatively stable condition and will give a
more accurate measurement of the pollutant loading due to the
land use rather than from a construction area. Storm water run-
off from a construction area can vary in quality from storm to
storm depending on the stage of construction. Accounting for
all construction areas and their erodibility prior to the storm
event being modeled proved to be difficult. Consequently, it is
presumed that several construction areas where the natural
ground had been disturbed and stripped of the protective vegeta-
tive cover contributed more suspended solids than SWMM could
predict from the available input data.
Investigation was also carried out on the time of occur-
rence and peak concentration of suspended solids. Except for
the storms of April. 11, 1974 and April 22, 1974, the observed
suspended solids concentration peak occurred before the observed
peak flow. The comparison would indicate that the peak concen-
tration at Stations P-10 and P-30 occurred at a flow of 0.065
times the peak flow of the storm.
During the storm of October 28, 1974, measured at Station
P-30, the rainfall caused two peaks in flow to occur, as shown
in Figure 46. In modeling the second peak the original SWMM
calculated a suspended solids peak concentration of 272.9 mg/1.
Suspended solids for this storm had an observed peak of 1000
mg/1.
The December 5, 1974 storm had a modeling advantage in that
it was the only storm analyzed where an upstream gage (P-10) and
a downstream gage (P-30) had samples taken simultaneously.
Since the entire drainage area had the same land use before
development began, most differences between the upstream gage
and the downstream gage can be attributed to the changing land
use in the Phase I development area.
Using the original SWMM version, the computed peak concen-
tration of suspended solids at Station P-10 was 142 mg/1 compared
124
-------
with an observed value of 130 mg/1. This is a good agreement and
also the time of peak concentration was the same. The falling
limb of the observed pollutograph occurred too rapidly resulting
in a difference of about 11,330 kilograms (25,000 pounds). This
is about a 40% error.
Suspended solids production at Station P-30 was about three
times greater than at Station P-10. As shown in Figure 48 the
SWMM again calculated a low value for suspended solids. The SWMM
was consistently low on the suspended solids concentration for
the Phase I development area.
The EH&A modified water quality version of the SWMM was also
used to model the 12/5/74 storm at Stations P-10 and P-30. The
observed and computed pollutographs for suspended solids, COD,
nitrates and phosphates are compared in Figures 49 through 52
for Station P-10 and in Figures 53 through 56 for Station P-30.
As was done for Station H-20, after the pollutographs were ade-
quately matched, the corresponding pollutant mass transport rates
were computed and are also shown in Figures 49 through 56. Again
the correlation between observed and computed pollutant mass
transport rates from pollutographs reproduction was unsatisfac-
tory. And so the pollutant mass transport rate predictions were
improved until the reproductions were acceptable. The loading
rates used are shown in Table 31 and the output results are
summarized in Table 32.
At Station P-10, the optimized suspended solids pollutograph
yielded very high pollutant mass transport rates (Figure 49) but
the optimized mass flow rate follows the data except for a single
peak. Suspended solids predictions at Station P-30 were more
compatible to observed data with slight differences for occur-
rence of peak pollutant mass transport rates. This condition was
also observed for COD at Station P-30 where the total pounds of
COD were also predicted too high. Modeling of phosphates at both
P-10 and P-30 proved to be difficult. The best approximations
after several attempts are shown in Figures 52 and 56. The
modeling of nitrates, especially at Station P-30 was not entirely
satisfactory. It is believed that only parts of the entire
Panther Branch watershed supply nitrates and therefore the model
which predicts nitrates from throughout the watershed would have
to be adjusted for this condition.
SWALE 8 MODELING
Existing drainage and planning maps were used to develop
the input data for Swale 8. Site inspections to determine
drainage area boundaries and extent of construction were con-
ducted on a periodic basis. This watershed is in a transition
stage. During the project term, the channel was enlarged and
construction of Lake C was underway. Lakes A and B had already
been filled.
125
-------
CO
o
_J
o
CO
Q
LJ
Q
Z
LU
O.
CO
z>
CO
STATION
P- 10
30 40 50
TIME (hrs)
60 70
80
OBSERVED
COMPUTED
o>
E
CO
Q
_J
O
CO
Q
LU
Q
•z.
UJ
Q.
CO
15
CO
STATION
P-30
20 30 40 50
TIME (hrs)
Fig. 48 Storm of 12/05/74 - Stations P-10 and P-30,
suspended solids by original SWMM version
126
-------
o
co
Q
UJ
Q
UJ
Q.
CO
ID
CO
CO
•H
rH
o
CO
-------
Q
O
O
Q
O
o
rH
I
G
O
•H
-P
rti
-P
en
in
o
(N
O
-P
en
o
m
Cn
•H
OJ
NOIlVaiN3DNOO
128
-------
LU
<
or
8 S S °
NOU.VaiN30NOO
-P -H
£ W
(U
cn nd
^_l I 1
Cfi 0)
0> ^
4J
fd dJ
M-l
U) ^
W O
w
(U
-p
fd
^
-P
•H
O
iH
I
O
-H
-P
fd
IT)
O
o
-P
CO
in
SSVW
129
-------
co
to
LU
Q.
CO
o
CL
-P
o
o
rH
I
c
o
•H
-P
fd
-P
m
o
O
^
O
-P
CN
in
o*
NOIlVyiN30NOO
130
-------
SUSPENDED SOLIDS
c
L
«
L
C
<
UDODrWCU
COMPUTED
«? •
FROM *
POLLUTOGRAPH
!
.. — •
8
04
^^
*£-£
/
/
)
-"
/
/
,.
8888
(uiuj/sq|) MOId SSVW
y
/i
C
1
I
/
1
I
i
/
2 18 0 6 0 10 20 30 40 50
AE (hrs) TIME (hrs)
* mass flow rates represent the computed concentrations
shown on the left hand side of this figure
Fig. 53 Storm of 12/05/74 - Station p-30, Suspended Solids
(l/Buj) NOIlVaiN30NOO
131
-------
Q
O
O
O
NOUVaiN30NOO
132
-------
en
0)
-P
-P
-H
Station P-30
74
05
2
Storm of
55
tp
•H
PM
133
-------
en
LJ
§
CL
CO
o
Q.
H
O
m
0)
-P
fd
-p
O
EH
O
m
I
o
•H
-p
IT)
o
CN
M-l
O
O
-P
in
tn
•H
(I/flu'
NOI1VHJLN30NOO
134
-------
rd
>
o
£
Q)
&
£
fd
icient
1 1 i
O O CO CO
r^ co LO to
CO
co
CO
s
CJ
23
PQ
Pi
W
P-i
CO
W
rJ
PQ
CU
PH
O
rH
CU
CU
o p
cd
•^ P!
t-P O
rH -H
v~-' p
o
CO 3
CU rH-
P P
cd co
Pi P!
o
txO CJ
P!
cd
•H
p
PJ
CU
T>
•H
CO
CU
Pi
P
P!
cd
p
o
PH
o
to
PH!
cu
Pi
p.
cd
to
txQ
o
P
LO
co
O CM
O O O O
M-l -rH
1 p
•rH O
P cd
rH
2 £
e o
•rH
T3 P T3
a o a
• • cd 3 cd
CO J-l rH
£ CU P
O rH CO T3
rH bO £ (U
rH £ O JQ
O 'rH 0 to
x-sM-l CO 3
cd X P
,£ CO ,111 4-> CO
-^ cd P -H -rH
W3 T-H J^ TI)
4^1315 {U
^-^ CU CO ^)
p! co cd
II -H cd 0) p!
MH cu }-i cu
•rr cu to cd PH
•J T3 cd O
1 1
_J a) i i
rH1 Pi
^" Cd rH O T)
cd -rH CU
X CO -H P P,
CU P O O
x-N CO P! £ rH
o 3 cu to cu
cd U p >
\Td -H CO CU
,n p! co pi -d
rH cd
-------
CN CM
^ tO cd fi CO
r^ pq -u 3 O
m AJ H P-t X
fN y
CN cn o
§ "Ij
fa £ a -
cn o « cn §
|y| cn ^-i ci) cO "^"^
£2 ^a Q* P-* ffi o
0 Ik, q «-<
HPM
_! J
CO g
co •
1 ^ v! CJ t— 3
3 co C ""-^
^" rH PJ O te
Ui i p^4 ^^ ^
O
2 P*
S
« cn
rH *"O
Pi! cO £ co
w -P 3 o
ffi O O rH
H j- c ^^ ><
5 -cLS
o _2 ^ cn -H o
5* 4-1 'Q 13 W g rH
C3 J3 O QJ .cO **^. i
fa rH J-l PH !>| O pL|
rH CX rH
^~: M
r-3 . H
P ^ q ,j *
ro P-I O e£
w
PQ
H
co m
..3s.
r^- vo
CN VO
m co
o co
rH ^tf
. . . .
m co
rH
r^ o^
CO CN O
O CN CO
1
OO i^^ O (
<1" CO
B
O> CN •<
. . °- ". ti
T^ O
CO v£5
rH
m
-^" ON
0 rH
CN v£)
CN
ON O
O CN
O O
CO vO
rH
cn
0)
cn .u
CU CO
•u X
CO ^,
M cn
" 8 3 JS
co O Z P-.
m ON
O
....
rH S ""^
00 rH
ON CN
co r*^
....
•^f vO
C) l^*>
CN
OO v±
vO CN
o m
co' in
co co
CN
a
CN CN
00* CO CO
oo r^ x-s
m 0
0)
cn
•* —
txO
^^
S £ " ^
ON ON vD
CN rH LPI ||
r»> X-N
s_xcN
•y^ rvi
CO 00 X™
rH rH v~^
0 co ^ x
CN *d
6 ^^
^. cn
^rO
^o '"'^
P*- CO ^/^-/
CN m
o co
l^- ^O td
"° H
O
J2
cn
cu
cn -u
0) co
j_> f~|
cO d-
M cn
Q -U O
^ q -H #
co O 23 K
136
-------
The drainage system for Swale 8 described in the SWMM is
shown in Figure 57. The total drainage area of 185.87 hectares
(459.3 acres) was divided into 10 subcatchments ranging from
9.30 hectares (23 acres) to 26.71 hectares (66 acres). All sub-
catchment data is listed in Table 33. Land uses for the upstream
subcatchments were classified as open space, whereas the last
three downstream subcatchments were designated as multi-family
residential and commercial as shown in Table 34. Gutter data
for all subcatchments are listed in Table 35. Seventeen drain-
age system elements, Table 36, were used to model the entire
area. Of these, two elements were storage units, Lakes A and B,
and all 6 channels were trapezoidal in shape as a result of
channel enlargement. Table 36 lists all transport system ele-
ment characteristics.
One storm event on Swale 8, that of 4/08/75, was modeled
because the only other observed storm event, 3/13/75, had a peak
inflow into Lake B of 0.06 cubic meters per second (2.0 cfs)
from 2.06 centimeters (0.81 inches) of rainfall. The storm
related temporal data for rainfall and infiltration are listed
in Table 37.
The transitional phase of development in Swale 8 gave rise
to several problems in modeling runoff. The most severe problem
is the total lack of lake volume data. The topographic maps
prior to lake construction show the natural ground contours, but
the reservoir areas were used as borrow pits for fill material
for the dams as well as other construction at The Woodlands.
Consequently, the original storage capacity of the reservoirs
was not known and no subsequent reservoir surveys have been
conducted; therefore, the elevation-area-capacity data for these
lakes was necessarily only approximate. Also groundwater was
being pumped into the Lake A and B system and again the pumpage
rate was not recorded.
A further complication arose from the fact that the outflow
structure for Lake A is controlled by different outlets at dif-
ferent water surface elevations. The outflow rating curve (dis-
charge as a function of water surface elevation) is composed of
three segments, one controlled by the low flow orifice, the
second controlled by weir flow through the flood discharge out-
let which in turn is limited at extreme flows by the capacity of
the outfall conduit and resulting in the third segment of the
rating curve. The SWMM is not capable of modeling this complex
outflow scheme.
Under the conditions described above, the modeling of
runoff storage in the lakes proved to be difficult. Several
attempts to model the outflow from Lake A (Station D-50) for the
storm of 4/08/75 proved to be unsuccessful as shown in Figure 58
137
-------
LEGEND
SUBCATCHMENT DIVIDE
WATERSHED DIVIDE
—— OPEN DITCH
(^] SUBCATCHMENT NUMBER
^~y]
O MANHOLE
26 MANHOLE NUMBER
SCALE IN FEET
^ii
500 0 500 1000
Fig. 57 Subcatchments and drainage network - Swale 8
138
-------
^->s ^ r^ i—I i—I r—li—4 i—l T—li—li—I
C3 03 O i—( »—I '""I ^ t—I T—l rH i—^ <-H (••<
si-^u oooooooooo
so oooooooooo
•H
JJ
cj /-s 2 in m m in in m m m in in
UUt^ OOOOOOOOOO
ij .i 2
r-1 ••>.
—1C- OOOOOOOOOO
«^~(5< OOOOOOOOOO
c ^ 2
0) > r>?«3-»r<>-S'<'
CO Ul COCCCOCOCOCOC3C3Q9CO
55 y f.^^^r^r.ir^t—|f-^»—<^i—I
u c-
o
''c
o °^ >
*£ '* £. OOOOOOOOOC
3 = E
e-5 > OOOOOOOOOO
^ oooooooooo
^ ? °
^-|C3^ OOOOOOOOOO
^ 03fc-OJ OOOOOOOOOO
p; 0) C. cvieMCNjescMescNCNCscM
a M
£ 0) iJ
^ »-l iJ OOOOOOOOOO
< *~^
S3
OT C >
y o
^ C- M
•]
^ ^co cM
-------
TABLE 34. LAND USE DATA, SWALE 8 WATERSHED
Subarea
Number
1
2
3
4
5
6
7
8
9
10
Land Use
Class .
5
5
5
5
5
5
5
2
2
3
Total Gutter
Length (100 ft) (
649.20
1320.80
646.40
1860.90
1222.90
778.00
1755.70
1100.00
1550.30
1455.20
Number c
^atchbasi
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
NOTE: Land Use Class is defined as follows:
1. Single and multi-family residential areas
3. Business and commercial activity areas
5. Undeveloped urban open land
140
-------
Q
w
IS
o o o o
o o o o
OO CM CO
CO
W)
C
•H
c
O O O O
OO
-------
o
o
0
o
o
o
o
oo
<}•
vD
^
OO
^J-
^
o
o
o
0
o
o
o
CM
CM
^
VO
^
CM
CM
O
0
O
O
0
o
o
F^. O
ON O
O> O
O O
CM
,— 1
r^- o
ON O
ON O
r^- co
**O CO
ON O
rH 0
rH
rH
r^ co
^D O
ON O
CM
O
O
rH
CM
O
O
O
O
0
O
O
0
CO
o
r-.
vO
00
rH
co
O
r--
O
O
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
0
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o vo o —I
ooooooooooooooooo
o
o
1
o
o
o
rH
CU
1— 1
o
d
CO
s
o
o
o
CM
rH
O
rH
O
o
"O
•iH
O
N
CU
CX
CO
H
O
O
1
o
o
o
rH
0>
rH
O
JC
c
s
o
o
o
CM
CM
O
i — i
O
o
-o
•H
O
CU
CX
CO
)H
H
o
o
1
o
o
o
rH
CU
rH
o
c
cO
£
o
o
o
CM
rH
m
o
o
o
-V
•H
o
N
CU
(X
CO
J.J
H
o
o
1
o
o
o
rH
CU
rH
o
rC
c
CO
o
o
o
o
rH
o
m
o
o
T)
•H
o
N
CU
a
CO
^
H
o
o
1
o
o
o
rH
CU
rH
o
c
CO
s
o
o
o
CO
o
m
o
o
T5
•r-l
o
N
CU
a
CO
J_l
H
o
o
1
o
o
o
rH
CU
rH
o
_f*t
d
CO
s
o
o
o
-------
TABLE 37. RAINFALL AND INFILTRATION DATA, SWALE 8 WATERSHED
RAINFALL DATA
in inches at 20 minute intervals
.02
.08
.00
.00
.82
.12
03
00
00
00
12
12
05
24
00
,00
,12
, 12
.05
.80
.00
.00
.12
. 12
INFILTRATION RATES
Initial Rate, in/hr 0.75
Final Rate, in/hr 0.01
Decay Rate, /sec 0.00115
200
12 13 14 15 16 17 18
4 5 6 7 8 9 10
STATION
D-50
OBSERVED
COMPUTED
Ficr. 58 Storm of 4/08/75 - Station D-50, hydrograph
143
-------
The extent of assumed data was just too large in magnitude to
even approximate the proper operation of Lakes A and B.
Consequently all further modeling was conducted only on
that drainage area of Swale 8 upstream from Lake B (Station
D-10). The results of this modeling effort are discussed in
the following subsection.
EXISTING AND FUTURE DEVELOPMENT MODELING FOR SWALE 8
Due to various external influences, urban development at
The Woodlands did not proceed as rapidly as had been expected.
Consequently, site development plans were available for Phase I
(Stage 1) only. In early 1976 a major portion of the Swale 8
watershed was being platted for development. Therefore all
future development modeling was conducted for Swale 8.
The continued operation of Station D-10 promised to be
beneficial in evaluating model predictions but unfortunately
Lake C was constructed upstream of this gage. Data for Lake C
has all the same problems associated with data for Lakes A and
B. Therefore the model predictions were biased by approximate
input data. Nevertheless, the SWMM was run for a phased devel-
opment scheme for Swale 8. Three development scenarios were
evaluated; existing conditions (development in Subcatchment 8
only and construction in Subcatchment 7), immediately develop-
ing conditions (development in Subcatchments 7 and 8 and con-
struction in Subcatchments 3, 4, and 5), and future but not
ultimate conditions (development in Subcatchments 3, 4, 5, 7 and
8 with no construction areas) . Ultimate conditions would assume
100 percent urbanization and no development plans are available
for these conditions.
Water quality predictions by both original and EH&A ver-
sions were attempted. Using plat maps provided by The Woodlands
Development Corporation, the proposed urbanization area and curb
lengths were measured. For Subcatchments 3,4, and 5 the proposed
area to be urbanized amounted to 76, 73, and 72 percent, respec-
tively. The average curb length per urbanized area was deter-
mined for these three Subcatchments and applied to all other
areas where curb lengths were not available. These computed and
proposed curb lengths were used for pollutant generation by the
original SWMM model. The changes in land use and increase in
imperviousness were also computed and input to the SWMM. All
of these data are listed in Table 38.
As described earlier, the EH&A quality prediction version
required the input of loading rates for each pollutant. The
curves shown in Figures 31 through 55 were used to derive each
of the desired loading rates listed in Table 39. Land use and
imperviousness data was the same as that for the original ver-
sion (Table 38) .
144
-------
TABLE 38. LAND USE DATA FOR FUTURE DEVELOPMENT, SWALE 8
Sub- Total Urban Urban Curb
catchment Area Area Area Length
(Acres) (Acres) (%) (Feet)
1 23 0 0 2760
2 47 6 13 8480
3 23 17 76 3540
4 65 39 73 10720
5 41 30 72 14480
6 29 4 15 -
7 81 27 37 12440
8 25 14 56 4320
145
-------
4J
o
Q:
(U
M-!
O
O O O
o
^-^ Clj
•H
W 4-J
QJ C
-P O)
cd T?
CKl -H
w
tuO Q)
cd CD
O CL
nJ O
,—1
CD
>
a)
OJ
O a)
-U
}-<
O
CM
CO
a
^D on
o o
o o
m
CN
\o o
r-l Osl
CN O O
O
CN
c^ oo
o o
moo
on
O
on
W
c
cd
4-1
CU
CX
co Q
2 O
CO CJ
PL,
w
H
O
146
-------
Observed and computed hydrographs for the storm of 4/08/75
are compared in Figure 59. Observed and computed mass flow
rates as determined by the original SWMM version for all three
development conditions are compared in Figure 60.
As shown in Figure 59, the total hydrograph for the storm
of 4/08/75 is composed of two hydrographs resulting from two
distinct periods of rainfall separated by 3 hours and 20 minutes
of no rainfall. Only the first period of rainfall, runoff and
water quality was modeled. Based on previously described exper-
ience with pollutograph differences resulting from computed hy-
drographs, it was decided that only mass flow rates would be
modeled.
Loading rates determined from the results of modeling at
Station P-30 were used in the first run. It became evident that
the initial loading rate estimates for developed areas were too
low indicating the extreme effects of lake and golf course con-
struction, as well as channel improvement. These activities
were concentrated in the Swale 8 watershed and well diluted in
the Panther Branch watershed; for example, the observed peak
mass flow of suspended solids at Station D-10 was three times
the peak mass flow computed from loading rates derived at Sta-
tion P-30. Also, the areas already developed have not been
stabilized - when rainfall intensities are sufficiently high,
even the freshly sodded areas will erode severely; consequently,
the loading rates for developed and construction areas in the
Swale 8 watershed were much closer than expected. At Station
D-10 the suspended solids loading rates from developed areas
were 82% of the rate from construction areas. The same ratio at
Station P-30 was 78%. The relatively high rates for developed
areas indicate the severe erosion potential from recently devel-
oped areas. The results of modeling the storm of 4/08/75 for
Swale 8 are shown in Figures 61 and 62 and in Table 40. Predic-
tions based on loading rates determined for the 12/05/74 storm
at Station P-30 are also presented.
The EH&A version of the SWMM was also run for the two devel-
opment conditions described earlier. The storm of 4/08/75 was
used to provide a basis for comparison between existing and
future conditions. The pollutant mass transport rates predicted
from future development are shown in Figures 61 through 66. As
anticipated, the modeling of Subcatchment areas 3, 4 and 5 as
construction areas changes the pollutant loads considerably; the
changes range from an increase of 77% for suspended solids to a
decrease of 8% for nitrates. After the construction phase of
development has been completed, the peak pollutant loads do not
decrease as may be expected, but the total pounds of pollutant
does decrease. These dramatic environmental effects of con-
struction activities are graphically illustrated in Figures 61
through 66 and listed in Table 41.
147
-------
120
100
80
UJ
e?
cr
erv
60
§40
20
STATION
D-IO
2 3456789 10 II
TIME (hrs)
__L i J
12 13 14 15 16
OBSERVED
COMPUTED
Fig. 59 Storm of 4/08/75 - Station D-10 Hydrograph
148
-------
o
LU
>
cc
LJ
Cn
•H
149
-------
LT>
-^
00
0
CU
CO
0
0
£j
O
•H
O
13
0
r-l
fX
CU
cO
Q
CO
rH 13
CO £
•P 3
O O
H PM
a
,J*5 CO -rH
cO co £3
CU CO -^
PM ^"! ,O
rH
4*5 0 rJ
cO C ^
CU O 00
PM CJ g
CO
rH 13
cO C
4-> 3
0 0
H p.,
a
^ CO -r-l
cO co g
CU CO \
PMS43
rH
r^ O (-J
CO £ ^
CU O OJO
PM U g
O ON
ON CO
CM LD
LO CO
v£5
CO
ON i — 1
O co
0
co r^-
•<|- rH
CO CM
N.O
« •
rH 00
ON
ON ON
OO ON
O CM
•
CO CM
ON CO
CM
CO rH
r^ CM
• •
CO O
rH CO
1 1
i — 1 r- H
Lfi O^
o in
rH CO
•
CM
CO
CU
CO 4-)
CU CO
4-J rC
cO (X
J-l CO
4-J O
•H ^
& PM
O
CU
CO
^^
W)
v_^
II txO
X
s-^^-s
vO
LO ||
r-^ s-^
v_xcM
v_x
^ X
•H
B^
•^ co
rQ r&
rH rH
v—^s-x
H
O
S3
150
-------
00
ui
I-
<
X
0.
CO
o
I
0-
\-
o
h-
SSVW
en
o
^
CM
en
o
o
to
o
z
UJ
Q.
(O
LJ
(muj/qi) MO"1J SSVW
o
Q Q fO
UJ UJ i
uj o. 5
en 5 o
en o o:
o o u.
I
Q
c
o
•H
-M
I
LO
00
o
O
g
n
o
4J
en
•H
CM
151
-------
Q
O
O
<
o
(wuj/qi) MOld SSVIAI
CM O
Q
O
U
-P
O
0)
-P
•H
53
I
Q
O
•H
-P
fd
-P
CO
UJ
MOld SSVI^I
o
Q
LU
K-
O
O
I
Q
UJ i
> a-
85 ?l
(/) O
CQ cr
o u-
LT)
CO
o
O
4J
CO
04
tn
•H
152
-------
CO
O
-I
O
CO
O
LU
a
UJ
a.
CO
15
CO
z
g
\-
o
15
cr
o
ID
o:
80
O' - —
w~ UJ
a.
o
'>—
o
*-«z " a.
ujP uj O
|x< »
o-iii Q. Q
I
Mcru ssvw
•H
rH
o
w
a
o
•H
-P
•H
o
u
(U
I
o
rH
CD
a
Q)
M
s
4J
^3
Cn
O
rH
I
Q
C
O
-H
-P
(0
-P
CO
fO
Cn
•H
NOIlVdlN30NOD
153
-------
NOI1VH1N30NOO
154
-------
LU
<
tr
H
o o
tr
\-
(Ujiu/qi) MOld SSVW
QC
S g
? M o
XLU x u
o.::;. a. Q
I
en
0)
-P
fd
M
-P
•H
c
o
•H
-P
•H
n0
C
O
U
-P
C
CU
a
o
iH
0)
>
0)
Q
00 CO
-------
to
CVJ
o
CVJ
If)
Moid ssvw
0)
Q)
-P
U)
c
o
-H
-P
•H
o y
en
LU
0.
CO
o
X
Q.
ID
a:
i-
X^ X UJ
Q-—.• Q_ Q
o
U
-P
c
(U
g
a
o
rH
0)
>
CL>
Q
-P
CVJ
I
Q
O
•H
-P
(0
-P
NOIlVdlN3DNOO
156
-------
23
o
M
H
<
H
in
co
r-l n3
cd p!
4-i ^
O O
H PH
o cr.
c^ co
CM LO
LD CO
ON
co CM T—I CM
vo co i—i t—i
i—I
W
P
c
^ co T-I
cd co g
"> i S _n
r-l
CO
23
O
CO
23
O
I-l
H
M
P
23
0
U
O
H
H
CO
M
X!
W
I-l
H
M
P
23
O
00 CT> O
O CM
r-l rH S
0
i— 1 M
O CO H
vO CJ
£3
(V]
H
CO
23
o
U
CO
^
o
r-l
H
M
p
ON CO 23
CT> CO O
0 r-l U
.
O in P
oo co W
O P-i
rH O
i_3
W
£>
w
p
t
CO P
^ 0
CO
d)
4->
cd
^_i
4-1
•H
CO
CU
4-1
cd
411
ex
CO
0
,£
CO CJ 23 P-i
CO
O
CO
CD
13
CD
CX
CO P
3 O
en U
co
CD
4J
cd
4-1
•r-l
23
CO
CD
4-1
cd
CX
CO
O
157
-------
One reason for the increase in peak pollutant mass transport
rates is the change in the runoff hydrograph. As seen in Figure
67, the hydrograph peak is increased by approximately 40%.
Another reason is the increase in the input loading rates as
listed in Table 39 for developed areas, which result in a
doubling of peak pollutant mass transport rates for COD, ni-
trates and phosphates. The 20% increase in the suspended solids
pollutant mass transport rate is a result of hydrograph modifi-
cation due to urbanization.
To determine the relative magnitude of pollutants at Station
D-10, a further investigation of water quality in Swale 8 was
also conducted. For existing conditions, equal pollutant loading
rates were applied to each land use area individually and then
together. This analysis provided an insight into the relative
pollutant generative capacity of each land use and also the ef-
fects of flow and pollutant routing in the SWMM. Both polluto-
graphs and pollutant mass transport rates were compared and as
shown in Figure 68 and Table 42, it is evident that any pollu-
tant is transported from the residential area at the highest
concentration and unit pollutant mass transport rate. The in-
crease in imperviousness in an urban area is a major reason for
this increase because the runoff intensity is increased. Due to
the effect of drainage area and routing characteristics, the
results presented in Figure 68 and Table 42 apply to the Swale 8
watershed only. A similar analysis of another watershed should
be performed if this type of information is desired for that
watershed.
In summary, the EH&A water quality modeling version greatly
improved the capabilities of the SWMM. Water quality modeling
results are much more dependable and observed events can be ade-
quately simulated. Each of the storms used to test the new SWMM
version, and as described in this section, was selected to pre-
sent a range of flow, water quality and land use data; thus the
model was tested over a range of different conditions.
158
-------
RUNOFF
HYDR06RAPHS
UJ
a:
o
o
8 10 12 14 16
- EXISTING
- CONSTRUCTION
•• DEVELOPED
Fig. 67 Station D-10, Runoff Hydrographs - Existing and
Future Conditions
159
-------
SSVW
Q
»-
Q
O
CJ
a co
>
LU
Q CO
2 UJ O
Z3 a: o
X=)
U
Q
-12
CO
QJ
rH
W
0)
-P
NOIlVdlN30NOO
-p
c
(D
M
(U
4-1
-------
o
w
ffi
en
tf
w
H
oo
w
>J
g
CO
w
tu
H
CO
w
CO
p
En
O
CO
H
CJ
W
w
w
M
W
CN
W
PQ
H
o
cd
^ 4_) •> —
O -H CJ
£ S'B
,0
I—I
0) CJ
4-> CO ^ -H
cd co cd ^
J-4 cd Q) ~~^-
bu * — i
•H
cd
0
,— 1
4-1
cd ^ o i-l
4-) cd 51! "^
3 CU O bO
t-H PLH CJ B
^ . J
T^^
O
.._ i
n^
cr
- ,n cJ 1-1
cd ^ o 1-1
p?; |~i rj> <^
161
-------
This study has indicated the magnitude of environmental
impact by construction activity. The SWMM was used in an attempt
to model this impact and the resulting effort was only partially
successful. The need for further data accumulation was highlighted.
The need for specific data requirements was identified and estab-
lished, and a thorough review of the methodology and programming
for the SWMM was conducted.
Runoff quantity and quality from areas not served by combined
sewers can now be modeled by SWMM with a significant reduction in
input data requirements. The SWMM can now be used to determine
baseflow recessions, as well as to quantify the effects of ground-
water depletion by urbanization. Generally the increase in base-
flow recession rates is directly proportional to groundwater
storage depletion by urbanization or other activity by man. The
SWMM can also be used to design and evaluate natural drainage
systems, and again the input data is significantly reduced. Only
a coordinate cross section description and the channel character-
istics are required.
A further development of the SWMM also allows for an evalu-
ation of the relative efficiency of economies between natural
and conventional drainage systems.
The porous pavement system model can assist in the planning
of urban development by determining the hydrologic response to a
design storm. The runoff rate and flow volume reduction can be
evaluated. Also, to quantify the effect of pavement clogging by
extraneous sources, the permeability and porosity of the pavement
or base may be reduced accordingly. The status of flow and
storage in the porous pavement system at all times, as shown in
Table 3, will indicate the efficiency of operation as well as
provide guidance in sizing the pavement and base thickness and
areal extent.
The model output can facilitate decisions to be made regarding
stormwater quality control. Porous pavement systems should be
designed to retain, as a minimum, the intial 30 percent of
runoff. Physical, chemical or biological treatment may be utilized
within the system or the stored stormwater can be pumped or
drained to a treatment plant when treatment capacity becomes avail-
able. The possibility of evaluating dilution effects on pollutant
concentrations in the stored stormwater needs to be investigated.
162
-------
A novel approach to water quality prediction has been devel-
oped. As more data becomes available, the predictive capabilities
of this method can be refined and specific pollutional char-
acteristics can be accentuated. The limited data at The Woodlands
and at Hunting Bayou indicate the strong potential for this type
of approach.
If the water quality relationships derived in this study are
comprehensively developed, and the data used are sufficient to
cover all types of land uses as well as geographical areas, then
the universalization of site specific data will be simplified and
the applicability of the SWMM will have been tremendously improved
This remains as the best avenue for further research because of
the dual functions it serves - to improve definition of nonpoint
source water quality characteristics as well as to indicate the
transferability of data, it is suggested that new methods of water
quality prediction, such as the one developed during this project^
be given immediate attention.
The intensive construction activity throughout The Woodlands
for the duration of this project provided an insight to the ex-
treme pollutional loading, especially in suspended solids, that
a receiving water body experiences when soil is moved and drain-
age systems are altered even though the changes are minimal.
The desirability of porous pavement usage and natural drainage
system implementation has been established and the hydrologic
characteristics of these innovations have been sufficiently
quantified to allow their detailed design.
Water quantity and quality from undeveloped areas (Station
P-10) were modeled very well by use of the SWMM. The relatively
low pollutant concentrations were well duplicated for most obser-
ved events even though the curb length for pollutant generation
is a calibrated parameter. The new natural section routine
facilitated input data preparation for the Transport Block.
The marked increase in suspended solids in the runoff from
developing areas proved to be difficult to model. The coeffic-
ients of the Universal Soil Loss Equation have to be calibrated
to model erosion; and more importantly, erosion from agricultural
areas is different when compared to erosion from construction
areas. Consequently, the SWMM could not be calibrated to model
erosion. The relative sucess on modeling the magnitude of
erosion by the modified method is inherent in the input data for
this method. The user selected loading rates are necessarily
high.
The most severely polluted runoff comes from urbanized areas
(Station H-20). The modified (EH§A) method was more accurate than
original SWMM to generate the high concentrations in this case.
163
-------
The lakes constructed in the Swale 8 watershed have performed
well their design function to retard excess runoff from urbanized
areas and to act as clarifiers for suspended sediment removal.
Although the lack of comprehensive data prevented a detailed
analysis of Lakes A and B, the inflow and outflow hydrographs
and pollutographs show how truly effective the lakes can be. It
is believed that Lake C behaves similarly, thereby compounding
the modeling problem but alleviating the problem of receiving
water pollution from construction activity.
Future development in Swale 8 watershed is not expected to
significantly increase the pollution transport rate from the
watershed. Of course the assumption is made that only natural
drainage systems will be constructed and the development plans
will not be altered.
The SWMM has undergone extensive evaluation and modification.
It has proved to be applicable in most areas; the only limit-
ations being areas with a transient land use and other areas
where extremely high suspended solids concentrations are generated
The modifications to the SWMM have improved its capabilities
considerably. The model can be applied universally but the model-
ing results are highly dependent on the availability of local
data.
164
-------
REFERENCES
1. Sartor, J. D., and G. B. Boyd. Water Pollution Aspects of
Street Surface Contaminants, USEPA Report No. EPA-R2-72-081,
November 1972.
2. Huber, W. C., e t a 1 . , Storm Water Management Model User's
Manual, Version 11~7 USEPA Report No. EPA-6 70/2-75-0 1 7 , 1975.
3. Holtan, H. N. and N. C. Lopez, USDAHL-73 Revised Model of
Watershed Hydrology, USDA Agricultural Research Service,
Plant Physiology Institute Report No. 1, 1973.
4. Linsley, R. K., M. A. Kohler, and J. L. H. Paulhus, Hydrology
for Engineers, McGraw-Hill, 1975, p. 152.
5. Riggs, H. C., The Baseflow Recession Curve as an Indicator of
Ground Water, International Association of Scientific Hy-
drology, Publication No. 63, 1963, pp. 352-363.
6. Barnes, B. S. Discussion of Analysis of Runoff Characteris-
tics, Transactions ASCE Vol. 105, 1940, p. 106.
7. Chow, V. T., Handbook of Applied Hydrology, McGraw-Hill,
1964, 14-10.
8. Horton, R. E. An Approach Towards a Physical Interpretation
of Infiltration Capacity, Proceedings Soil Science Society
of America, Vol. 5, 1940, pp. 399-417.
9. Metcalf and Eddy, Inc., University of Florida, and Water Re-
sources Engineers, Inc. Storm Water Management Model, Vol.
1, USEPA Report No. 11024 DOC 07/71, 1971.
10. Chow, V. T., Open Channel Hydraulics, McGraw-Hill, 1959,
p. 136.
11. Letter, G. K., Considerations on Hydraulic Design of Channels
with Different Roughness of Walls, Referenced by Chow, Open
Channel Hydraulics, p. 136.
12. Thelen, E., et al., Investigation of Porous Pavements for
Urban Runoff Control, USEPA Report No. 11034 DUY 03/72,
March 1972.
165
-------
REFERENCES (Cont'd)
13. Izzard, C. F. Hydraulics of Runoff from Developed Surfaces,
Proceedings Highway Research Board, Vol. 26, 1946,
pp. 129-150.
14. Taylor, D. W. Fundamentals of Soil Mechanics, Chapter 6.
Permeability, John Wiley & Sons, 1965.
15. Pinder, G. F., J. D. Bredehoeft, and H. H. Cooper, Jr.,
Determination of Aquifer Diffusivity from Aquifer Response
to Fluctuations in River Stage, Water Resources Research,
Vol. 5, No. 4, August 1969.
16. Colston, N. V. Characterization and Treatment of Urban Land
Runoff, USEPA Report No. EPA-670/2-74-096, December 1974.
17. Wells, D. M., J. F. Anderson, R. M. Sweazy, and B. J. Cla-
born. Variation of Urban Runoff Quality with Duration and
Intensity of Storms — Phase II, Office of Water Resources
Research, August 1973.
18. American Public Works Association, Water Pollution Aspects
of Urban Runoff, Federal Water Pollution Control Authority,
Publication WP 20-15; 1969.
19. Amy, G . , e t a1., Water Quality Management Planning for Urban
Runoff, USEPA Report No. EPA-440/9-75-004, Dec. 1974.
20. Johnson, S. L. Urban Hydrology, Houston Metropolitan Area,
Texas 1968, U. S. Geological Survey.
21. Johnson, S. L. Annual Compilation and Analysis of Hydrolo-
gic Data for Urban Studies in the Houston, Texas Metropoli-
tan Area, 1969, U. S. Geological Survey.
22. Ferguson, D. E. 1970. Annual Compilation and Analysis of
Hydrologic Data for Urban Studies in the Houston, Texas
Metropolitan Area, 1970, U. S. Geological Survey.
23. Ferguson, D. E. Annul Compilation and Analysis of Hydro-
logic Data for Urban Studies in the Houston, Texas Metro-
politan Area, 1971, U. S. Geological Survey.
166
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/2-79-050C
4. TITLE AND SUBTITLE
MAXIMUM UTILIZATION OF WATER RESOURCES IN A PLANNED
COMMUNITY - Application of the Storm Water Management
Model; Volume I
7. AUTHOR(S)
Elvidio V. Diniz
William H. Espey, Jr.
3. RECIPIENT'S ACCESSION NO.
5. REPORT DATE
July 1979 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Espey, Huston and Associates, Inc.
3010 S. Lamar Blvd.
Austin, Texas 78704
10. PROGRAM ELEMENT NO.
1BC822, SOS #2, Task 02
11. CONTRACT/GRANT NO.
802433
12. SPONSORING AGENCY NAME AND ADDRESS
Municipal Environmental Research Laboratory, Cin., OH
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final 1973-1976
14. SPONSORING AGENCY CODE
EPA/600/14
15. SUPPLEMENTARY NOTES
One in a series of volumes of one report. Project Officers: Richard Field and
Anthony N. Tafuri, Storm and Combined Sewer Section, FTS 340-6674, (201) 321-6674
16. ABSTRACT
A Management strategy for utilization of water resources in the planned community of
The Woodlands, near Houston, Texas, was developed by modification and application of
the EPA Storm Water Management Model (SWMM). Selected sites on Panther Branch, which
flows through The Woodlands, and on Hunting Bayou, a completely developed watershed
within the city limits of Houston, Texas were modeled for testing and verification
of the modifications to the
The capacity of the SWMM to model urban runoff quantity has been improved to include
the "natural" drainage concepts of The Woodlands and the infiltration computation
model in the SWMM is now capable of operating with a rainfall record which includes
periods of zero rainfall. Three new subroutines generate normalized area-discharge
curves for natural sections, model baseflow conditions, and model the operation
of porous pavements, respectively. Verification of the SWMM with regard to suspended
solids and BODr was attempted and modifications to predict COD, Kjeldahl nitrogen,
nitrates and phosphates were performed. This innovative water quality modeling
scheme has proved very successful in predicting future effects of urbanization.
17.
a.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
Mathematical models, Water pollution,
Surface water runoff, Drainage, Water
resources, Water quality
b. IDENTIFIERS/OPEN ENDEDTERMS
Woodlands Project,
Urban stormwater analyses
Natural drainage, Porous
pavements, Baseflow,
Urbanization effects,
Planned urban development
Infiltration
c. COSATl Field/Group
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
181
SS (This page)
22. PRICE
EPA Form 2220-1 (Rev. 4-77)
167
-------
------- |