EPA-670/2-75-022
Hay 1975
Environmental Protection Technology Series
URBAN STORMWATER MANAGEMENT
MODELING AND DECISION-MAKING
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-670/2-75-022
May 1975
URBAN STORMWATER MANAGEMENT
MODELING AND DECISION-MAKING
By
James P- Heaney, Wayne C. Huber,
Hasan Sheikh, Miguel A. Medina, J. Robert Doyle,
W. Alan Peltz, and John E. Darling
University of Florida
Gainesville, Florida 32611
Project No. 802219 (11023 GSC)
Program Element No. 1BB034
PROJECT OFFICER
Richard Field
Storm and Combined Sewer Section (Edison, N.J.)
Advanced Waste Treatment Research Laboratory
National Environmental Research Center
Cincinnati, Ohio 45268
NATIONAL ENVIRONMENTAL RESEARCH CENTER
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
REVIEW NOTICE
The National Environmental Research Center —
Cincinnati has reviewed this report and approved its
publication. Approval does not signify that the
contents necessarily reflect the views and policies
of the U. S. Environmental Protection Agency, nor
does mention of trade names or commercial products
constitute endorsement or recommendation for use.
11
-------
FOREWORD
Man and his environment must be protected from the adverse
effects of pesticides, radiation, noise and other forms of pollution,
and the unwise management of solid waste. Efforts to protect the
environment require a focus that recognizes the interplay between
the components of our physical environment—air, water, and land.
The National Environmental Research Centers provide this multi-
disciplinary focus through programs engaged in
0 studies on the effects of environmental contaminants
on man and the biosphere, and
° a search for ways to prevent contamination and to recycle
valuable resources.
This study tests, refines, and augments the capabilities of
the EPA Storm Water Management Model (SWMM). Also, decision-making
capabilities are developed for use in the study of urban storm
water runoff problems. The storm water management problem is viewed
in the broader context of urban water resources management. The
results indicate numerous areas where more comprehensive environmental
management strategies could be effectively applied.
A. W. Breidenbach, Ph.D.
Director
National Environmental
Research Center, Cincinnati
ill
-------
ABSTRACT
The purposes of this study were to test, refine and augment the capa-
bilities of the EPA Storm Water Management Model (SWMM), and to
develop decision-making capabilities, for use in the study of urban
storm water runoff problems.
With regard to the SWMM the numerous programming refinements were
introduced to provide greater reliability and flexibility. A sedi-
ment prediction capability has been incorporated directly into the
SWMM. Detailed testing was conducted in Lancaster, Pennsylvania to
demonstrate the ability of the SWMM to describe the relatively
complex phenomena occurring in an urban catchment.
A systematic procedure is presented for examining the storm water
management problem in the broader context of urban water resources
management. Related standards for flood control and drainage, street
and parking lot design, etc. are reviewed and suggestions presented
regarding modifications in practices which would ameliorate storm
water problems. An optimization procedure is described which addresses
the related problems of finding efficient and equitable control
strategies. Linear programming techniques are combined with proce-
dures from cooperative N-person game theory to examine these questions.
Lastly, the use of the SWMM for preliminary hydraulic design of sewer
systems is described. The runoff portion provides a more accurate
inlet hydrograph and the transport portion is used for refined hydraulic
analysis of phenomena such as surcharging.
This report was submitted in fulfillment of Grant No. 802219 (11023 GSC)
by the Department of Environmental Engineering Sciences, University of
Florida, under the sponsorship of the U.S. Environmental Protection
Agency. Work was completed as of June 1973.
IV
-------
CONTENTS
Page
Abstract iv
List of Figures vi
List of Tables viii
Acknowledgments x^
Sections
I Conclusions 1
II Recommendations 5
III Introduction 7
IV Refinements of the SWMM 11
V Testing of the SWMM in Lancaster, Pennsylvania 80
VI Decision-Making for Water Quantity and Quality
Control 100
VII Hydraulic Design by the SWMM 171
VIII References 178
-------
FIGURES
No.
1 Overview of Model Structure 8
2 Iso-Erodent Map (R Values for the Erosion Equation). . 22
3 Soil Erodibility Nomograph 26
4 Topographic Factor, LS 34
5 Rainfall Hyetograph 38
6 Annual Production of Total Tree Litter in Relation
to Latitude 51
7 Options Available in Revised Treatment Model 61
8 Swirl Concentrator Removal Efficiency for Settleable
Solids as a Function of Inflow and Different
Settling Velocities, V 62
s
9 Removal Efficiency by Sedimentation. Original and
Revised Relationships 65
10 Combined Inflow to North Treatment Plant Upstream of
Diversion Chamber 67
11 Combined BOD Load to North Treatment Plant Upstream
of Diversion Chamber 68
12 Combined Suspended Solids Load to North Treatment
Plant Upstream of Diversion Chamber 69
13 Hypothetical Drainage Network A 76
14 Hypothetical Drainage Network B 76
15 Spatial Analysis of Pollutant Inputs into the
Stevens Avenue Drainage Network 79
16 Drainage Districts of Lancaster, Pennsylvania and
Numbering Systems for Receiving Junctions 82
17 Sewer Layout of Stevens Avenue Drainage District ... 83
18 Flow of Treatment-Storage Options at Demonstration Site 84
19 Simulation for Stevens Avenue, Study 1, Storm 6. ... 86
VI
-------
FIGURES-continued
No. Page
20 Simulation for Stevens Avenue, Study 3 88
21 North Drainage District with Runoff-Transport
Numbering System 89
22 South Drainage District with Runoff-Transport
Numbering System 90
23 Simulation of North Drainage District, Study 3 91
24 Simulation of North and South Drainage Districts,
Study 3 92
25 Combination of SWMM Runs for Overall Lancaster
Simulation 94
26 Dissolved Oxygen Profiles Along the Conestoga River,
Study 3 95
27 BOD Profiles Along the Conestoga River, Study 3 .... 96
28 Suspended Solids Profiles Along the Conestoga River,
Study 3 97
29 Effect of Urbanization on Storm Water Hydrograph. . . . 107
30 Effect of Urbanization on Peak Discharges 108
31 Typical Chicago 10-acre Tract Drainage Basin 114
32 Capital Cost for High-Rate Filtration 138
33 Cost Function for On-Site Stormwater Control 150
34 Network Representation of Example Problem 153
35 Shadow Price for Area 1 for Assumed Values of Q . . . . 155
36 Demand for Off-site Storage 159
37 Effect of Surcharging and Conduit Alteration on Stevens
Avenue Overflow, Study 1, Storm 6 173
Vll
-------
TABLES
No. Page
1 Erosion Rates Reported for Various Sediment Sources 13
2 Factor C Values used in Computing Gross Erosion 14
3 Comparison of Data from Two Studies on Dust and Dirt
Collected in Street 16
4 Kinetic Energy of Natural Rainfall 21
5 Adjustment Factor M for estimating Monthly and Portions
of Annual Soil Loss for Locations in Maryland, New
Jersery, Delaware, Eastern Virginia and Eastern
Pennsylvania 23
6 Soil Loss Probability and Single Storm Soil Loss for
Locations in Maryland, New Jersey, Delaware,
Eastern Virginia and Eastern Pennsylvania 24
7 Soil Erodibility Index K Values for Maryland Soil Series . . 27
8 Cropping Management Factor C 36
9 Erosion Control Practice Factor P for Construction Sites . . 37
10 Erosion Printout 43
11 Estimated Amount of Tree Litter in Various Locations
in the United States 45
12 Sources of Forest Litter 46
13 Common and Scientific Names of Trees 47
14 A Comparison of Litter Production by Evergreen and
Deciduous Trees in the Northern Hemisphere 49
15 Annual Litter Production in Four Major Climatic Zones. ... 50
16 The Concentration of Nutrients in Newly Fallen
Gymnosperm and Angiosperm Tree Leaf Litter 54
17 Average Quantities of Nutrients Falling in the Litter
of Different Trees 55
Vlll
-------
TABLES-continued
No. Page
18 Beginning of Storage/Treatment 70
19 Treatment Model Input 71
20 Treatment Model Output 73
21 Printout from Pollutant Monitoring Routine 77
22 Parameters used for Simulating Receiving Water Storage . . 98
23 Estimated Monthly Quantity of Identifiable Solids from
10-Acre Urban Area, Chicago, Illinois 104
24 Control of Combined Sewer Overflows 110
25 -Effect of Housing Density on Land Use Patterns 115
26 Zoning Standard Guidelines for Parking 117
27 Recommended Parking Stall Dimensions for Various Types
of Stalls for 90° Parking 118
28 Dimensions of 1973 Vehicles Produced by Ford
Motor Company 119
29 Paving in U. S 121
30 Proposed Residential Street Design Standards Summary . . . 126
31 Allowable use of Streets for Initial Storm Runoff in
Terms of Pavement Encroachment 128
32 Major Storm Runoff Allowable Street Inundation 129
33 Allowable Cross Street Flow 130
34 Reduction Factors to Apply to Inlets 132
35 Cost of Storage in Sandusky, Ohio 135
36 Comparison of Storage Costs for Various Cities 136
37 Comparison of the Cost of Conventional Pavement and Porous
Pavement 142
IX
-------
TABLES-continued
No. Page
38 Erosion Control Costs 143
39 Unit Costs of Catchbasin Cleaning and Mechanical Street
Sweeping 144
40 Levels of Protection for Minor Storm Events 146
41 Levels of Protection for Major Storm Events 146
42 Optimization Model for the Three Study Areas 157
43 Optimal Solution with Ownership of Off-Site Facility
Prespecified 164
44 Charges for Central Facility 169
45 Initial Dimensions and Conduit Capacities 174
46 History of Surcharging and Conduit Alteration During
Storms 175
47 Final Conduit Dimensions and Capacities 177
-------
ACKNOWLEDGMENTS
This work was performed for the City of Lancaster, Pennsylvania as
part of Demonstration Grant No. 11023 GSC from the U.S. Environmental
Protection Agency (EPA). We are grateful to many people in Lancaster,
EPA, Meridian Engineering, Inc., and the American Public Works
Association for their cooperation and suggestions. Mr. Lester Andes,
Director of Public Works for the City of Lancaster, provided the leader-
ship for coordination of the study and acquisition of the data. Messrs.
Richard Field, Anthony Tafuri and Harry Torno of EPA supervised the
research activity and offered many valuable suggestions on techniques
and coordination with related EPA activities in other parts of the
country. Messrs. T. R. Darmody and Robert Travaglini of Meridian En-
gineering, consultants to the City of Lancaster, collected much of the
engineering data for the study. Lastly, Mr. Richard Sullivan of APWA
provided detailed design and related information regarding the swirl
concentrator.
The SWMM is a large and complex model which requires a significant
commitment of resources in order to use it properly. We are fortu-
nate to have a systems analyst, Mr. W. Alan Peltz, who has devoted
a considerable amount of his time debugging, simplifying and modi-
fying the model so that its usefulness will be improved. We are
indebted to Ms. Polinski for her patience and persistence in typing
numerous drafts of this report. Mr. Damon Hooten and Ms. Gina Ellis
conscientiously drafted figures and diagrams.
Computations were performed on the IBM 370/165 at the Northeast
Regional Data Center at the University of Florida.
XI
-------
SECTION I
CONCLUSIONS
Overall project goals were to test, refine and augment the capabilities
of the EPA Storm Water Management Model (SWMM), and to develop decision-
making techniques for use in the study of urban runoff problems. This
report represents the findings of the University of Florida with respect
to the specific objectives outlined below:
Objectives, Task I:
1) test present SWMM using data from Lancaster, Pennsylvania;
2) provide programming improvements to SWMM;
3) provide model refinements to SWMM; and
4) compile alternative methods of pollution control and
associated cost data.
Objectives, Task II:
1) provide detailed cost functions for pollution control
options;
2) develop evaluation and monitoring routine for assess-
ment of pollution significance;
3) develop decision-making methodology for least cost
abatement of pollution and drainage problems and
provide necessary interfaces in SWMM; and
4) develop hydraulic design capability for the SWMM.
Findings are presented in Section IV - Refinements of the SWMM, Section
V - Testing of SWMM in Lancaster, Pennsylvania, Section VI - Decision-
Making for Water Quantity and Quality Control, and Section VII -
Hydraulic Design by the SWMM.
With regard to Section IV, numerous programming refinements were
introduced to provide greater reliability and flexibility in using the
Model. Field studies in Lancaster, Pennsylvania and Gainesville, Florida
indicated the importance of soil and vegetation as sources of stormwater
pollution. Thus, special studies were conducted to obtain a better
appreciation of these phenomena. As a result of these studies, a sedi-
ment prediction capability has been incorporated directly into the SWMM.
This method uses the Universal Soil Loss Equation which was originally
developed for agricultural areas. The amount of tree and grass litter
in urban areas can be estimated using the results presented in a latter
-------
part of Section IV. This review indicates that improperly managed
collection and disposal systems for these materials can cause signifi-
cant deleterious effects on the receiving water.
Refinements were made in the Treatment Model to reflect recent develop-
ments such as swirl concentrators. Similarly, refinements in the
Receiving Model are presented which further increase the flexibility of
this Model.
Also included in Section IV is a procedure for ranking the nodes based
on their selective importance as contributors of pollution. Further-
more, the magnitude of the source is partitioned into its subcomponents.
The testing in Lancaster presented in Section V illustrates the ability
of the Model to describe the relatively complex phenomena occurring in
an urban catchment. Unfortunately, only very limited and somewhat
questionable quantity and quality verification data were available. As
a result, agreement between predicted and measured hydrographs and pollu-
tographs was only fair. Quality predictions in the Conestoga River,
the receiving waterway, were reasonably close to the limited measured
data available. The analysis indicates an improved capability to de-
scribe the various processes but also indicates the need for a better
data base to properly calibrate the model.
The Decision-Making Model for urban stormwater management is presented
in Section VI. A systematic procedure is presented for examining the
problem in the broader context of urban water resources management. The
inventory of sources of pollutants includes reference to related models
for air pollution control. Examination of the effect of urbanization on
stormwater runoff indicates that significantly higher peak flows accom-
pany urbanization. Thus, control costs increase substantially at higher
levels of urbanization.
Studies of parking practices and residential street design indicate
there has been a significant increase in paved area per household in
order to accommodate an automobile-oriented society. Residential areas
are the most significant source of storm water since they are the pre-
dominant land use. It appears that significant reductions in the amount
of paved area could be realized with little loss in value for residents.
Residential streets could be made narrower in recognition of the avail-
ability of off-street parking. Because of their low rate of utilization,
it appears feasible to temporarily pond water on many residential
streets. Curb and gutter could be eliminated in some areas which would
substantially reduce road construction costs and achieve on-site control
of storm water. Inlets could be designed to incorporate grates and pre-
vent "trash" from entering the storm sewer. Current inlet design
optimizes hydraulic efficiency to ther detriment of the receiving water.
Current standards for parking lot design provide a very high level of
service, e>g.3 shopping centers are designed with adequate parking for
-------
all but the peak 10 hours of the year. They are utilized at a much
lower rate during the balance of the year. It appears feasible to
temporarily pond water on parking lots and/or redesign them to include
pervious areas to permit percolation. The use of compact cars would
reduce the paved area by about one third.
In summary, it appears that the impervious areas associated with the
automobile are the most important source of the stormwater problems.
These developments have resulted in a significant increase in paved
area. By contrast, more innovative urban planning concepts are revers-
ing this trend by utilizing paved area much more efficiently than in
the past. The economic analysis presents a revised and expanded list
of cost functions for the various control options.
Following a review of alternative performance standards for stormwater
quality management, we have recommended that the control system should
provide acceptable performance for the "minor storm" as defined in the
Drainage Criteria Manual. To a large extent, stormwater quantity and
quality management are complementary. The suggested performance
criteria are:
1) no increase in the peak flow observed prior to
development; and
2) adequate treatment of storm water prior to release
to the receiving environment.
These standards can be satisfied on-site or off-site by forming a
coalition with others for the same purpose.
In any given city, there are a wide variety of control alternatives to
be considered. Thus, the Optimization Model was set up in a general
manner. The emphasis was on describing principles for efficient and
equitable control strategies. A procedure is included for placing a
value on off—site disposal based on alternative on-site control costs.
Much research has been done in determining the best overall solution.
However, little effort has been made to examine whether this solution
is fair to everyone. Using notions from cooperative N—person game
theory, procedures are described for solving this important equity
question. Unfortunately, the scope of the research did not allow a
detailed example of the methodology.
The purpose of the Decision-Making Model for urban stormwater quantity
and quality management is to provide a systematic and comprehensive
procedure for solving this important problem. A review of the techni-
cal aspects of the problem indicates that stormwater quantity and
quality problems seem to grow more serious as urbanization continues.
Consequently, even using sophisticated engineering controls, it does
not seem possible to devise a control strategy which will have long-
term value. These observations led to a questioning of the wisdom of
this open-ended approach.
3
-------
The proposed quasinclosed system approach solves the above problems by
internalizing the responsibility for stormwater control so that those
people causing the problem are assigned responsibility for adequate
collection and safe disposition of this water. Ideally, this principle
could be applied without discrimination. With this policy, at low
levels of urbanization there is usually a relatively large amount of
open space such as flood plains, etc., which would provide a "natural"
solution to the problem. Thus, control costs would be low. However,
as the area reaches a higher level of urbanization, many of these
natural areas are either being fully utilized or have been developed.
Thus, additional development needs to acquire their own control facili-
ties. Where stormwater quantity and quality constraints are placed on
a new development, it appears to result in the type of development which
is very much in vogue in current urban planning.
Section VII shows how the SWMM can be used for hydraulic design of sewer
systems. The Runoff portion of the SWMM permits one to generate a more
accurate inlet hydrograph. Transport permits a much improved characteri-
zation of the hydraulics of flow through sewers. For a specified system
configuration, a routine has been added which checks for surcharge con-
ditions and, if present, enlarges the downstream conduit by standard
amounts until capacity exists to accept the flow. An example illustrates
the useful ability of an undersized system to store excess flows and thus
reduce the peaks of downstream hydrographs.
-------
SECTION II
RECOMMENDATIONS
Relatively sophisticated mathematical models such as the SWMM should
be quite appealing to environmental decision-makers as a management
tool. However, the state of the art in mathematical simulation models
has reached a level of efficiency that is not being matched by the
data—gathering agencies. Invariably, a successful application of the
SWMM requires:
1) a considerable amount of input data ranging from physical
watershed information to cost functions for selection of
the most practical treatment alternatives;
2) frequent field measurements of real storm events for model
running and calibration procedures; and
3) that the data-gathering efforts to satisfy requirements
(1) and (2) should be established on as permanent a basis
as possible to enable the SWMM user to continuously update
the data base. The following recommendations are made for
furthering the continued use of the SWMM:
a) Municipalities and other governmental agencies
concerned with stormwater management should be
encouraged to gather the type of information
(such as erosion) which is vital to a complete,
meaningful and frequent application of management
tools such as SWMM in the interest of reducing the
overall impact of storm water on the urban
community.
b) It is recommended that the use of the SWMM for
analyzing real systems with real problems be
continued and expanded. Many refinements and
improvements have resulted from such interactions.
c) It is recommended that more explicit considerations
be given to the interrelationships between urban
stormwater quality management and urban water
resources management. The SWMM appears to be the
most sophisticated model for analyzing these broader
questions.
-------
d) Finally, It is recommended that the appropriate
professional organizations be encouraged to
re'-examine the wisdom of existing practices in
residential street design, parking lot design,
and urban drainage design. Design practices
for these functions appear to have evolved with
little consideration for their overall impact
on the urban area.
-------
SECTION III
INTRODUCTION
A". COMBINED AND STORM SEWER OVERFLOWS
An enormous pollution load is placed on streams and other receiving
waters by combined and separate storm sewer overflows. It has been
estimated that the total pounds of pollutants (BOD and suspended
solids) contributed yearly to receiving waters by such overflows is
of the same order of magnitude as that released by all secondary
sewage treatment facilities (Gameson and Davidson, 1964; Field and
Struzeski, 1972). The Environmental Protection Agency (EPA) has
recognized this problem and led and coordinated efforts to develop
and demonstrate pollution abatement procedures (Field and Struzeski,
1972). These procedures include not only improved treatment and
storage facilities, but also possibilities for upstream abatement
alternatives such as rooftop and parking lot retention, increased soil
infiltration, improved street sweeping, retention basins and catch-
basin cleaning or removal. The complexities and costs of proposed
abatement procedures require that care and effort be expended by
municipalities and others charged with decision-making for the
solution of these problems.
B. STORM WATER MANAGEMENT MODEL
It was recognized that an invaluable tool to decision makers would
be a comprehensive mathematical computer simulation program that
would accurately model quantity (flow) and quality (concentration),
during the total urban rainfall-runoff process. This model would
not only provide an accurate representation of the physical system,
but also provide an opportunity to determine the effect of proposed
pollution abatement procedures. Alternatives could then be tested
on the model and least cost solutions could be developed.
As a result, the University of Florida (UF), Metcalf and Eddy, Inc.,
Engineers (ME) and Water Resources Engineers (WRE) were awarded a
joint contract for the development, demonstration and verification
of the Storm Water Management Model (SWMM). The resulting model,
completed in October, 1970, has been documented (EPA, 1971 a,b,c,d),
and is presently being used by a variety of consulting firms and
universities.
The present SWMM is descriptive in nature and will model most urban
configurations and encompasses rainfall, runoff, drainage, storage-
treatment, and receiving waters. The major components of the SWMM
are illustrated in Figure 1. However, it neither defines nor deter-
mines any decisions for the system, nor does it consider alternative
methods for efficient economic comparisons.
-------
RUNOFF
(RUNOFF)
INFILTRATION
JFILTRAT
(INFIU
DRY WEATHER
FLOW
(FILTH)
DECAY
(QUAL)
TRANSPORT
(TRANS)
INTERNAL
STORAGE
(TSTROT)
COST
(TSTCST)
EXTERNAL
STORAGE
(STORAG)
k-
TREATMENT
(TREAT)
COST
(TRCOST)
RECEIVING WATER
(RECEIV)
FIGURE 1.
Overview of the Model Structure-.
Subroutine names are shown in parentheses,
8
-------
C. DECISION-MAKING
In recognition of the need for improved decision-making capabilities,
the University of Florida submitted a proposal to EPA titled "Decision-
Making Model for the Management of Storm Water Pollution Control" in
which it was intended to provide a systematic procedure which could be
applied to a wide variety of specific circumstances in support of
intelligent management decisions. The work required to obtain a least
cost solution would be considerably reduced by means of determining
the origin of the most severe pollution load, consideration of all
upstream and downstream pollution abatement procedures and associated
costs, and through the possible use of mathematical optimization
techniques.
The project was funded as part of an EPA Demonstration Grant to
Lancaster, Pennsylvania (EPA No. 802219 (formerly 11023 GSC), in which
an underground "silo," a swirl concentrator and a microstrainer were to
be installed at the outfall of the Stevens Avenue Drainage District
to control overflow into the Conestoga River (full details are given
in Chapter V). This report represents the findings of the University
of Florida with respect to specific objectives as defined below:
Objectives, Task I:
1) test present SWMM using data from Lancaster, Pennsylvania;
2) provide programming improvements to SWMM;
3) provide model refinements to SWMM; and
4) compile alternative methods of pollution control
and associated cost data.
Objectives, Task II:
1) provide detailed cost functions for pollution control
options;
2) develop evaluation and monitoring routine for
assessment of pollution significance;
3) develop decision-making methodology for least cost
abatement of pollution and drainage problems and
provide necessary interfaces in SWMM; and
4) develop a hydraulic design capability for the SWMM.
-------
Because of funding problems, this report does not cover Task III of
the research, as included in the original proposal, which xs a demon-
stration of the improved SWMM and decision-making model in Lancaster.
Thus, a detailed application of new techniques discussed herein xs not
included. Rather, the emphasis is on the results of research to date.
Model improvements, refinements and testing are discussed in Section V
and the hydraulic design capability is illustrated in Section VII. The
rest of the objectives including details of costs, monitoring and the
decision-making methodology are discussed in Section VI.
10
-------
SECTION IV
REFINEMENTS OF THE SWMM
A. PROGRAMMING IMPROVEMENTS
Utilization of the Storm Water Management Model has resulted in an
extensive array of alterations, ranging from remedies for program-
ming errors to inclusion of new routines. Major refinements are
discussed in the next section, while general improvements are con-
sidered below.
Most large computer programs have their share of "bugs" that appear
from time to time during usage, and the SWMM is no exception. Inas-
much as the model has been extensively altered since its original
formulation, simple corrections to the original program are not
easily identified and listed. Rather, users are advised to obtain
the most recent version from EPA, which will include all corrections
and refinements to date.
Input/output data-sets have been made compatible among all programs,
such that it is possible for output from a Transport run to serve as
input to another Transport run, for example. This enables large or
complicated networks to be simulated in a sequential fashion. How-
ever, a common time step is required for all models, except for input
to Receiving.
In the original Transport Model, quality routing was performed incor-
rectly during surcharge conditions. This has been corrected, and the
revisions may be seen in a listing of the current UF version of the
SWMM.
Refinements to several routines (see next section) have necessitated
some changes in program input requirements. Release 2 of the EPA
Storm Water Management Model is an improved version of the original
model that retains all of the original model's features, as well as
the refinements to be discussed subsequently. New options are used
only upon request of the user, and the default condition omits them.
B. EVALUATION AND PREDICTION OF URBAN SEDIMENT YIELD-OVERVIEW
It is estimated that naturally vegetated land areas erode at an
average of 115 tons/mile2/year (4.08 x 104 kg/km2/yr) (Ports, 1973;
Brandt, 1972). Of course, this figure varies considerably at differ-
ent locations, depending upon such things as climate, soil, terrain,
and vegetative cover. For example, natural areas of the Potomac
11
-------
2
River Basin have sediment yields of 15 to 20 tons/mile /year (5.33 to
7.10 kg/km2/yr),-as compared with yields greater than 320 tons/mile /
year (1.14 x 10 kg/km /yr) throughout geological history for the
Mississippi River Basin (Brandt, 1972). Natural erosion and sedi-
mentation are vital geological processes by which nutrients travel
from the land to the oceans. Considerations for controlling these
processes should be tempered with an understanding of the natural
system. Evaluation of alternatives to control excessive erosion should
be undertaken on a river basin or regional basis where both natural
and man-induced sediment yields can be measured or predicted.
Today, man has certainly increased the rate of erosion and the con-
sequent siltation and sedimentation of rivers in areas of his most
intense activity. The increase is a result of activities which remove
natural ground cover, thus allowing immediate soil-water contact
and increased impermeability. This increased impermeability causes
runoff of greater velocity which more readily transports the avail-
able soil. As an example of the effect of man's actions, estimates
for the Potomac River Basin indicate that the Washington DC area
produces nearly one third of the total sediment discharged into the
estuary. The Washington DC area encompasses less than 5% of the
watershed (Brandt, 1972). In the northern Mississippi River Basin,
studies indicate an increase in erosion occurs as the land use goes
from forested watershed to pasture land to cultivated land use
(Brandt, 1972). A summary of these and other pertinent data relating
erosion rates to land use is given in Table 1 (Brandt, 1972).
Sediment Production
Erosion from construction sites in and around the urban area is
probably the greatest source of sediment yield per unit land area.
Estimates of sediment yield for urban areas undergoing construction
range from 280 to 100,000 tons/mile2/year (3.55 x 107 kg/km2/yr)
(see Table 1). Sediment yields from single construction sites may
vary from 2 to 200 times as much as from naturally vegetated areas
(Brandt, 1972). In Table 2, crop management factors (C factors)
which are used in predicting gross annual erosion by the Universal
Soil Loss Equation are listed (Brandt, 1972). The erosion prediction
increases linearly with increasing value of C. These prediction
factors indicate that erosion due to construction is on the order of
100 times greater than completely developed urban sites, all else
being equal.
In addition to localized soil erosion from construction sites, open
land, and stream banks, sediment is also found rather uniformly
throughout the urban area. This sediment is often referred to as
dust and dirt and the total annual yield can be partly attributed
to localized soil erosion, but probably a greater part is made up of
12
-------
Table 1.
EROSION RATES REPORTED FOR VARIOUS SEDIMENT SOURCES
(Brandt, 1972)
Sediment
source
Natural
Agricultural
Urban
Highway construction
Erosion rate
(ton/mile2 /year)
15-20
32-192
200
320
13-83
25-100
115
12,800
13,900
1,030
10,000-70,000
200-500
320-3,840
50,000
1,000-100,000
1,000
500
146
280
690
2,300
36,000
50,000-150,000
Geographic location
Potomac River Basin
Pennsylvania and
Virginia
Mississippi River Basin
Northern Mississippi
Northwest New Jersey
Missouri Valley
Northern Mississippi
Northern Mississippi
Eastern U. S. Piedmont
Kensington, Maryland
Washington, D. C. area
Philadelphia area
Washington, D. C. area
watersheds
Fairfax County, Virginia
Georgia
Coranent
Native cover
Native, cover
Natural drainage basin
Throughout geologic
history
Forested watershed
Forest and under-
developed land
Soils eroding at the
rate they form
Loess region
Cultivated land
Pasture land
Continuous row crop
without conservation
practices
Farmland
Established as
tolerable erosion
Undergoing extensive
construction
Small urban construc-
tion area
750 mile2 area average
As urbanization
increases
Construction on 179 ac:
Cut slopes
13
-------
Table 2. FACTOR C VALUES USED IN COMPUTING GROSS EROSION
(Brandt, 1972)
Land Use Factor C Value
Cropland 0.08
Grassland 0.01
Woodland 0.005
Construction 1.0
Urban 0.01
14
-------
air pollution dust fallout, building and demolition wastes, eroded
pavement, and less discernible particulate accumulation due to wind,
traffic, and human activity. These materials are similar to natural
sediment in size, shape and composition, and therefore should be
considered as part of the total urban sediment load.
In a study made in Chicago, th.e average dust and dirt yield was
found to be 79 Ibs/curb-mile/day (22.3 kg/curb-km/day) or 14 tons/
curb--mile/year (7.9 MT/CUrb-km/yr) (American Public Works Association,
1969) . Since these data were collected in flat developed areas of
varying land use, the contribution due to heavy soil erosion from
construction sites and vacant lots can be considered negligible. A
similar study on eight "US cities indicated an average total solids
loading rate of 348 Ibs/curb^nile/day (98 kg/curb-km/day) (Sartor
and Gail, 1972). The Chicago study indicated that 75 percent of the
total solids was made up of dust and dirt having particle size less
than 3200 microns. In the second study, 75 percent of the total
solids had a particle size of less than 2000 microns and can be con-
sidered dust and dirt. The eight-city study then indicates an average
street sediment loading rate of 261 Ihs/curb-mile/day (73.53 kg/curb-
km/day). Table 3 summarizes these data for both studies.
A typical Chicago residential area has a ratio of curb length to
land area of 1 mile to .0235 miles2 or 42.5 mile-1 (26.4 km"1). This
figure can be used to convert dust and dirt data for residential Chicago
to Ibs/mile2/year for comparison with erosion data from urban areas.
Thus Chicago street sediment yield is 287 and 940 tons/mile2/year
(1.02 x 1Q5 and 3.35 x 105 kg/km2/yr) for single and multiple family
residential areas respectively. As seen from Table 1 , these dust
and dirt estimates are similar to erosion data from the Washington
DC area.
Sediment Control Alternatives
The distinctions made between erosion and street dust and dirt as
sources of urban sediment are very useful in considering sediment
control alternatives. Erosion from exposed soil surfaces can best
be controlled at the site of erosion by various methods of applying
cover to hold the soil and/or by using sediment basins or control
structures to collect the soil from the runoff. In the case of
construction site erosion, development practices also influence the
extent of erosion. Street sediment on the other hand, can be con-
trolled by street sweeping, catchbasin cleaning, or by collection
through retention of stormwater at the downstream end of the drainage
network. In order to evaluate the effectiveness of these general
control alternatives, estimates of the magnitude of sediment yielded
from each type of source must be made for both before and after
conditions.
15
-------
Table 3. COMPARISON OF DATA FROM TWO STUDIES
ON DUST AND DIRT COLLECTED IN STREETS
(Ib/curb mile/day)
Land use
Commercial
Industrial
Residential
Chicago^1)
Dust & dirt
174
243
Eight
Total solids
226
447
373
cities^2)
Dust & dirt =
. 75 total solids
170
335
280
Multi-family
Single family
Weighted average
121
37
79
(1) APWA, 1969
(2) Sartor and Gail, 1972
348
261
-------
Soil Detachment by Rainfall
Raindrop impact has been studied by C. K. Mutchler and R. A. Young
(1972) both at USDA Sedimentation Laboratory, to establish the direct
role of raindrops in the mechanics of erosion. Raindrop impact is
the primary source of energy for detaching soil from any land area
not protected by cover of some type. It also plays an important part
in sediment transport by overland flow (runoff), during which hydrau-
lic erosion develops.
Raindrops range in size from around 7.0 mm in diameter to fine mist
sizes. The relation of raindrop size to intensity of rainfall has
been established by Laws and Parsons (1943). Most rainfalls are con-
sidered to have a raindrop size distribution described by the normal
or Gaussian probability'distribution, based on raindrop volumes. Laws
and Parsons define the median drop diameter, 050, as the diameter where
the total volume of water in the larger drops equals the total volume
of water in the smaller drops and is used as a parameter to talk about
rainfall drop sizes. Obviously, from the definition of V$Q itself, it
is clear that the number of raindrops larger than H^Q is much less
than the number of raindrops smaller than the median drop size. The
choice of the term "median" is somewhat misleading. At any rate, the
median drop size is related to rainfall intensity as follows:
D5Q = 2.23 I°'182 (1)
where I = rainfall intensity, in/hr
Raindrops reach a terminal velocity dependent on their diameter. These
velocities are, for instance: 3.3 feet/second (1.01 m/sec) for a small
drop 0.25 mm in diameter, 13.1 feet/second (3.99 m/sec) for a drop 1 mm
in diameter, and 29.9 feet/second (9.13 m/sec) for a drop 5 mm in dia-
meter. Actual impact velocity, however, is often greater than the
terminal velocity due to wind effects imparting an added velocity com-
ponent. Laws and Parsons (1943) published size distributions of rain-
drops as a percent of total volume for rainfall intensities ranging
from 0.01 to 6.0 inches/hour (0.03 to 15.24 cm/hour). Gunn and Kinzer
(1949) published the terminal velocities of water drops of different
sizes in stagnant air.
o
From the relationship of kinetic energy to velocity (K.E. = 1/2 mV ),
the drop size distributions of rainfall determined by Laws and Parsons,
and the terminal velocities of water drops of different sizes deter-
mined by Gunn and Kinzer, a least-squares linear regression analysis
was performed by Wischmeier and Smith (1958) resulting in the following
relationship between kinetic energy and rainfall intensity:
17
-------
Y = 916 + 331 Iog10 I (2)
where Y = kinetic energy ±n foot tons/acre—inch.
I = rainfall intensity, in/hr
This equation and the maximum 30-minute storm intensity form the basis
of the El variable which is an integral part of the Universal Soil
Loss Equation. Units shown must be used for computation; then the
results may be converted to the metric system if desired.
Mutchler and Young presented some of their findings at the 1972
Sediment Yield Workshop, USDA Sedimentation Laboratory, Oxford,
Mississippi. From plot studies and other observations, raindrop
impact was found to be most erosive where a very thin layer of
water is present (about 1/5 drop diameter) and is relatively non-
erosive when the soil is covered with a water depth of 3-drop diameters
or greater. Several statements are valid concerning soil detachment
by rainfall:
(1) raindrop splash is the primary -agent in
soil detachment and transport from inter-
rill areas;
(2) detachment is a result of impact energy
dissipation on a saturated soil surface
not protected by cover of plant material
or sufficient water depth;
(3) the transport to rills is by splash action
alone to a limited extent and primarily
by very thin surface flow accelerated by
raindrop impact splash velocities;
(4) runoff and sediment follow the same path
down a slope from inter-rill area to rill
area to larger water courses: true sheet
flow rarely occurs;
(5) soil loss in a rill system is determined by
the supply of detached soil particles by
raindrop impact and then transport of these
particles by the small channels—rill flow*
(6) at some point, the banks or channel bed of
the watercourses dominate as a sediment
source from a watershed.
18
-------
Prediction of Erosion
The Universal Soil Loss Equation was derived from statistical
analyses of soil loss and associated data obtained in 40 years of
research, by the Agricultural Research Service (ARS) and assembled
at the AJRS runoff and soil loss data center at Purdue University.
The data include more than 250,000 runoff events at 48 research
stations in 26 states, representing about 10,000 plot-years of
erosion studies under natural rain. It was developed by Wischmeier
and Smith (1965) as an estimate of the average annual soil evasion
from rainstorms for a given upland area, expressed as the average
annual soil loss/unit area, A (tons/acre):
A = (R) (K) (LS) (C) (P) (3)
where R = the rainfall factor
K = the soil credibility factor
LS = the slope length gradient ratio
C - the cropping management factor or cover index
factor
P - the erosion control practice factor
This equation represents the most comprehensive attempt at relating
the major factors in soil erosion. It is used in the SWMM to predict
the average soil loss for a given storm or time period. It is
recognized that the Universal Soil Loss Equation was not developed for
making predictions based on specific rainfall events. There are many
random variables which tend to cancel out when computing annual time
averages which would not cancel out when predicting individual storm
yields: for example, the initial soil-moisture condition, or ante-
cedent moisture condition (AMC), is a parameter which cannot be deter-
mined directly and used reliably. It should be understood by the SWMM
user that Equation (3) enables land management planners to estimate
gross erosion rates for a wide range of rainfall, soil, slope, crop,
and management conditions.
Rainfall Factor_, R
The rainfall factor, R, is equal to the sum of the rainfall erosion
indexes for all storms during the period of prediction, £jEI. For a
single storm, R would simply equal El for that storm. If we multiply
Equation (2) by the total inches of rainfall received during each
rainfall hyetograph time interval, and we sum over all the time
19
-------
intervals, then the s.tom's total rainfall energy is given by:
R - El = £ (9.16 + 3.31 log XI)DI I (*)
where E = storm's rainfall energy (hundreds of foot-tons/
acre)
= ZLY.D. = £(9.16 + 3.31 log X^D.^
i = rainfall hyetograph time intervals
X± = rainfall intensity during time interval i, in/hr
D. = inches of rainfall during time interval i
I = maximum average 3Q^minute intensity of rainfall
To avoid confusion over the interval rainfall intensity term and the
maximum average 30-minute intensity of rainfall term in Equation (4),
variable X was substituted for I in the expression for Y. Values for
Y. are given in Table 4 (Wischmeier and Smith, 1958) for different
intensities, X^- For predicting average annual erosion, values for
the rainfall factor are shown in Figure 2 (Wischmeier and Smith,
1965) for locations east of the Rocky Mountains. It is very impor-
tant to point out that, due to the effects of hurricanes and other
tropical storms in the southeastern United States, the highest value
of the erosion rainfall factor, R, that should be used is 350
(Ports, 1974, private communication). It is also worthy to note that
the R factor does not account for soil losses due to snowmelt and
wind erosion.
Where rainfall intensity records are not available for a single storm
or a series of storms of interest, the erosion for these events can
be calculated from the predicted annual erosion. Table 5 (Ports,
1973) gives values for a new variable, M, which is the fraction of
annual erosion expected for a period of one to eleven months. Table 6
(Ports, 1973) lists values for estimating soil loss for storms of
frequency less than one year and probable annual soil loss for one
year in 2, 5, and 20 years. To obtain the expected soil loss these
factors are simply multiplied by the predicted annual soil loss.
Values from Table 5 and 6 apply for locations in Maryland, New Jersey,
Delaware, eastern Virginia, and the eastern half of Pennsylvania.
Similar tables can be developed for regions east of the Rocky Moun-
tains from data in Agriculture Handbook 282 (Wischmeier and Smith,
1965).
20
-------
Table 4. KINETIC ENERGY OF NATURAL RAINFALL*
(Hundreds of foot tons/acre Inch)
(Wischmeier and Smith, 1958)
Inten-
sity
In/hr
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
Intensity, inches/hour
0.00
0.00
5.85
6.85
7.43
7.84
8.16
8.43
8.65
8.84
9.01
0.00
9.16
10.16
10.74
11.15
11.47
11.74
11.96
12.15
12.32
0.01
2.54
5.99
6.92
7.48
7.88
8.19
8.45
8.67
8.86
9.02
0.10
9.30
10.23
10.79
11.19
11.50
11.76
11.98
12.17
12.33
0.02
3.54
6.11
6.98
7.52
7.91
8.22
8.47
8.69
8.87
9.04
0.20
9.42
10.29
10.83
11.22
11.53
11.78
12.00
12.18
12.35
0.03
4.12
6.23
7.05
7.57
7.95
8.25
8.50
8.71
8.89
9.06
0.30
9.54
10.36
10.88
11.26
11.56
11.81
12.02
12.20
12.37
0.04
4.53
6.33
7.11
7.61
7.98
8.27
8'. 52
8.73
8.91
9.07
0.40
9.64
10.42
10.92
11.29
11.58
11.83
12.04
12.22
12.38
0.05
4.85
6.43
7.17
7.65
8.01
8.30
8.54
8.75
8.93
9.09
0.50
9.74
10.48
10.96
11.32
11.61
11.85
12.06
12.24
12.40
0.06
5.12
6.53
7.22
7.69
8.04
8.33
8.56
8.77
8.94
9.10
0.60
9.84
10.53
11.00
11.35
11.64
11.87
12.08
12.25
12.41
0.07
5.34
6.61
7.28
7.73
8.07
8.35
8.58
8.78
8.96
9.12
0.70
9.92
10.59
11.04
11.38
11.66
11.89
12.09
12.27
12.43
0.08
5.53
6.69
7.33
8.07
8.10
8.38
8.61
8.80
8.98
9.13
0.80
10.00
10.64
11.08
11.41
11.69
11.92
12.11
12.29
12.44
0.09
5.70
6.77
7.38
7.81
8.14
8.40
8.63
8.82
8.99
9.15
0.90
10.08
10.69
11.12
11.44
11.71
11.94
12.13
12.30
12.46
* Example: Kinetic energy of rainfall of 0.57 in/hr = 8.35 hundreds of ft tons/acre in.
-------
NJ
EIGURE 2.
Iso-Erodent Map (R Values for the Erosion Equation).
(W±sctimeier and Sm±th, 1965)
-------
Table 5. ADJUSTMENT FACTOR M FOR ESTIMATING
MONTHLY AND PORTIONS OF ANNUAL SOIL LOSS FOR LOCATIONS IN
MARYLAND, NEW JERSEY, DELAWARE, EASTERN VIRGINIA AND EASTERN PENNSYLVANIA
(Ports, 1973)
to
to
Ending month
Starting
month
Jan.
Feb.
Mar.
Apr.
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
Jan.
0.00
0.98
0.96
0.94
0.90
0.80
0.65
0.45
0.24
0.14
0.07
0.03
Feb.
0.02
0.00
0.98
0.96
0.92
0.82
0.67
0.47
0.26
0.16
0.09
0.05
Mar.
0.04
0.02
0.00
0.98
0.94
0.84
0.69
0.49
0.28
0.18
0.11
0.07
Apr.
0.06
0.04
0.02
0.00
0.96
0.86
0.71
0.51
0.30
0.20
0.13
0.09
May
0.10
0.08
0.06
0.04
0.00
0.90
0.75
0.55
0.34
0.24
0.17
0.13
June
0.20
0.18
0.16
0.14
0.10
0.00
0.85
0.65
0.44
0.34
0.27
0.23
July
0.35
0.33
0.31
0.29
0.25
0.15
0.00
0.80
0.59
0.49
0.42
0.38
Aug.
0.55
0.53
0.51
0.49
0.45
0.35
0.20
0.00
0.79
0.69
0.62
0.58
Sept.
0.76
0.74
0.72
0.70
0.66
0.56
0.41
0.21
0.00
0.90
0.83
0.79
Oct.
0.86
0.84
0.82
0.80
0.76
0.66
0.51
0.31
0.10
0.00
0.93
0.89
Nov.
0.93
0.91
0.89
0.87
0.83
0.73
0.58
0.38
0.17
0.07
0.00
0.96
Dec.
0.97
0.95
0.93
0.91
0.87
0.77
0.62
0.42
0.21
0.11
0.04
0.00
All dates are as of the first of each month. M = 1.0 for one full year.
Example: Given an average annual soil loss of 100 tons/acre/year. The estimated soil loss for
July is 100 x 0.2 = 20 tons/acre. The estimated soil loss from May to September is
100 x 0.66 = 66 tons/acre.
-------
Table 6. SOIL LOSS PROBABILITY AND SINGLE STORM SOIL LOSS
FOR LOCATIONS IN MARYLAND, NEW JERSEY, DELAWARE,
EASTERN VIRGINIA AND EASTERN PENNSYLVANIA
(Ports, 1973)
Probability
Single storm
One
year
in
Factor
Exceeded
once in
Factor
2
5
20
0.90
1.25
1.70
1 year
2
5
10
20
0.2
0.3
0.4
0.5
0.7
24
-------
Soil Factor, K
The soil factor is a measure of the potential erodibility of a soil
and has units of tons/unit of erosion index, El. The soil erodibility
nomograph shown in Figure 3 (Wischmeier, Johnson and Cross, 1971)
is used to find the value of the soil factor once five soil parameters
have been estimated. These parameters are: percent silt plus very
fine sand (0.05-0.10 mm), percent sand greater than 0.10 mm, organic
matter content, structure, and permeability. To use the nomograph,
enter on the left vertical scale with the appropriate percent silt
plus very fine sand. Proceed horizontally to the correct percent
sand curve, then move vertically to the correct organic matter curve.
Moving horizontally to the right from this point, the first approxi-
mation of K is given on the vertical scale. For soils of fine granular
structure and moderate permeability, this first approximation value
corresponds to the final K value and the procedure is terminated. If
the soil structure and permeability are different than this, it is
necessary to continue the horizontal path to intercept the correct
structure curve, proceed vertically downward to the correct permea-
bility curve, and move left to the soil erodibility scale to find K.
This procedure is illustrated by the dotted line on the nomograph.
For a more complete discussion on this topic, see Wischmeier, Johnson
and Cross (1971).
Table 7 (Maryland Water Resources Administration, 1973) lists soil
factor values for soil types found in Maryland.
Slope Length-Gradient Ratio, LS
The slope length-gradient ratio is a function of runoff length and
slope and is given by the following equation:
LS = L1/2(0.0076 + 0.0053S + 0.00076S2) (5)
where L = the length in feet from the point of origin of
overland flow to the point where the slope
decreases to the extent that deposition begins
or to the point at which runoff enters a defined
channel
S = the average percent slope over the given runoff
length
Values for LS can be read from Figure 4 (Ports, 1973) knowing runoff
length and slope. In using the average percent slope in calculating
the LS factor, the predicted erosion will be different from the
actual erosion when the slope is not uniform. Meyer and Kramer (1969)
25
-------
2-tin« gronulor
J-mtd.
4-blocky, ploly, or mottivt
100
E! Ulth ippnprlitt d«t«. inttr lull 4. Solution: 1C • 0.11.
2-mo . to rapid
I- rapid
W. H. MlSCH<:IEB. ARS-SHC. PUROUC UMV. 2-1-71
FIGURE 3. Soil Erodibility Nomograph.
(Wischmeier, Johnson and Cross, 1971)
-------
Table 7. SOIL ERODIBILITY INDEX K VALUES FOR MARYLAND SOIL SERIES
(Maryland Water Resources Administration, 1973)
Soil series
Adelphia
Athol
Aura
Bertie
Berks
Bermudian
Bibb
Horizon
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
A
B
A
B
Texture range
Sl,fsl,l
L,scl,f si
SI, Is
Sil
Gsil.gl
Sicl.cl
G,cl
Sicl.cl
Gsl.gl
Sl,l
Gl,gsl
Ls
Scl
Gscl.gsl
Scl, si
Gsl.gcl
Ls
Sil,l
Sil,sicl,l
Stratified
SI, 1,1s
Gsl
Shsil,chsil
Sh to vshsil
Vshsil
Shattered shale
Sil.l
Fsl
Stratified silt
S
G
SI to sicl
Highly variable
K Value
0.32
0.40
0.20
0.37
0.32
0.30
0.30
0.30
0.30
0.43
0.30
0.20
0.40
0.30
0.40
0.30
0.20
0.37
0.40
0.30
0.20
0.24
0.20
0.20
0.20
0.43
0.40
0.50
0.30
0.20
0.32
0.20
27
-------
Table 7 (continued). SOIL ERODIBILITY INDEX K VALUES
FOR MARYLAND SOIL SERIES
Soil series
Birdsboro
Bucks
Chalfont
Chillum
Colemantown
Collington
Colts Neck
Grot on
Donlonton
Horizon
A
B
C
A
B
C
A
B
C
A
B
A
B
C
A
B
C
A
B
C
A
B
C
A
B
C
Texture range
Si.1,1
Sicl,cl
Sl,s,g
Sicl,l
Sil
Sicl,sil
Shsll,vshsil
Sil,vstl
Sil.sicl
Shsil.shl
Sil,sicl
Gl
Gscl,gl
Gsl
L,sl
Sc,scl
Sl,cl,scl
Sl,fsl,l
Ls
Scl,cl,sl,l
SI, Is
SI
Ls
Scl,sl,l
SI
Sil
Sil,sicl
Shsil,shsicl
Fsl,ls,sil
Sc,cl,sic
Sc,sicl,cl,ls
K Value
0.28
0.30
0.20
0.30
0.32
0.40
0.20
0.43
0.60
0.60
0.32
0.30
0.30
0.20
0.43
0.40
0.40
0.28
0.20
0.40
0.20
0.28
0.20
0.40
0.30
0.43
0.50
0.40
0.43
0.40
0.30
28
-------
Table 7 (continued). SOIL ERODIBILITY INDEX K VALUES
FOR MARYLAND SOIL SERIES
Soil series
Duf field
Edgemont
Elkton
Evesboro
Fallsington
Fort Mott
Freneau
Galestown
Howell
Keansburg
Keyport
Klej
Horizon
A
B
C
A
B
C
A
B
C
A
A
B
C
A
B
C
A
A
A
B
C
A
B
A
B
C
Sandy substratum
A
B
Texture range
Sil
Sicl
Sicl
Shsil
Chi
Chl,chscl
Chl.shsl
Sil,fsl,sl,l
Sic,c
Sic,sicl,scl
Ls,s
Sl,fsl,l
Scl,sl
S,ls,sl
S,ls
SI
S
Sl,l
Ls,s
Fsl,sil,l
Cl,sicl
C,sic,sicl
Sl,l
Sl,l
Sll.l.fsl
C ,sic,cl
Sicl, sic
Scl,sl
Ls ,f s,lf s
Ls,f s,lf s,sl
K Value
0.32
0.30
0.40
0.30
0.24
0.30
0.20
0.43
0.40
0.40
0.17
0.28
0.30
0.20
0.20
0.30
0.20
0.28
0.17
0.43
0.40
0.30
0.28
0.30
0.43
0.40
0.40
0.30
0.17
0.20
29
-------
Table 7 (continued). SOIL ERODIBILITY INDEX K VALUES
FOR MARYLAND SOIL SERIES
Soil series
Lakeland
Lansdale
Legore
Lehigh
Matapeake
Matawan
Matt apex
Monmouth
Neshaminy
Horizon
A
A
B
C
A
B
C
A
B
C
A
B
C
A
B
A
B
C
A
B
C
A
B
C
Texture range
Ls.lfs
L,sl
Scl.sl
L
Chsil.gsl
Chsl,gsl
Sil.sicl
Gl
Cl
Gcl.gl.gsicl
L.sil.sicl
Gl.vgl.gcl
Sil
Chsil
Chsicl
Chsicl,vchsil
Sil,fsl,l
Sil,sicl
S,ls,sl,l,gs
Sl,ls,fsl
Cl,scl,sc,sl
Sll.l.fsl
Sicl,sil,cl
Sl,ls,s,l,gs
Fsl,l,lfs
Sc,scl
Sl,scl,sc
Sil
Vstsil
Sicl,cl,scl,sl
Diabase bedrock
K Value
0.17
0.28
0.30
0.40
0.30
0.20
0.24
0.20
0.30
0.20
0.30
0.20
0.43
0.37
0.40
0.30
0.32
0.40
0.30
0.32
0.40
0.37
0.40
0.20
0.43
0.40
0.30
0.32
0.28
0.30
30
-------
Table 7 (continued). SOIL ERODIBILITY INDEX K VALUES
FOR MARYLAND SOIL SERIES
Soil series
Norton
Othello
Penn
Pocomoke
Raritan
Readington
Rowland
Rutlege
Sassafras
Horizon
A
B
C
A
B
C
A
B
A
B
A
B
C
A
B
C
A
B
A
B
A
B
C
Texture range
Sil.l
Sicl
Sil
Vgl,shl
Sil, l,fsl, sicl
Sicl,sil
Sl,ls,scl
L
Shsil
Sil
Shsil, sicl
Sl,l,fsl,ls,lfs
Ls,s
Sil
Cl,sicl
Stratified silt, f si
C,sil,l,g
Sil
Sil, sicl
Sil
Vshsil
Sil.l
Sicl
Stratified silt
and gravel
Sil
Ls,lf s
S,fs,ls,lfs
Fsl,l,sl,lfs
Ls
Gfsl.gsl
Scl,sl,l
Sl,ls,fsl,gsl,gls
K Value
0.32
0.40
0.40
0.30
0.37
0.40
0.30
0.32
0.28
0.40
0.30
0.28
0.20
0.43
0.30
0.20
0.30
0.43
0.40
0.40
0.30
0.43
0.40
0.30
0.40
0.17
0.20
0.28
0.20
0.24
0.30
0.20
31
-------
Table 7 (continued). SOIL ERODIBILITY INDEX K VALUES
FOR MARYLAND SOIL SERIES
Soil series
Shrewsbury
Steinsburg
Watchung
Westphalia
Woodstown
Horizon
A
B
C
A
B
C
A
B
C
A
B
C
.A
B
C
Texture range
Sl,fsl,l
Scl.sl
S,ls,sl
SI
Gsl.vgsl
Gsl
Sandstone
Sil
C,cl,sicl
Sil,sicl,l
Fsl,lfs
Fsl,lfs,vfsl
Fs,lfs,fsl
Sl,fsl,l
Ls
Scl,l,sl
S,ls,sl,gsl,gls
K Value
0.28
0.30
0.20
0.28
0.24
0.20
0.43
0.40
0.40
0.49
0.40
0.30
0.28
0.20
0.40
0.20
32
-------
USDA SOIL TEXTURE ABBREVIATIONS USED IN TABLE 7
C Clay
Ch. Channery
Cl Clay loam
Co Coarse
Fs Fine sand
Fsl Fine sandy loam
G Gravelly
Gel Gravelly clay loam
Gl Gravelly loam
Gscl Gravelly sandy clay loam
Gsl Gravelly sandy loam
L Loam
Lfs Loamy fine sand
Ls Loamy sand
S Sand
Scl Sandy clay loam
Sh. Shaly
Sic Silty clay
Sicl Silty clay loam
Sil Silt loam
SI Sandy loam
St Stony
Vfs Very fine sand
Vfsl Very fine sandy loam
33
-------
IOO
200
300
400
500
600
700
800
Slope Length (Feet)
FIGURE 4. Topographic Factor, LS.
(Wisctimeler and Smith, 1965)
-------
show that when the actual slope is convex, the average slope prediction
will underestimate the total erosion whereas for a concave slope the
prediction equation will overestimate the actual erosion. If possible,
to minimize these errors, large eroding sites should be broken up
into areas of fairly uniform slope.
Cropping Management Factor, C
The cropping management factor is dependent upon the type of ground
cover, the general management practice and the condition of the soil
over the area of concern. The C factor is set equal to one for con-
tinuous fallow ground which is defined as land that has been tilled
and kept free of vegetation and surface crusting. Values for the
cropping management factor are given in Table 8 (Maryland Water
Resources Administration, 1973).
Erosion Control Practice Factor3 P
The control practice factor is similar to the C factor except that
P accounts for the erosion-control effectiveness of superimposed
practices such as contouring, terracing, compacting, sediment basins
and control structures. Values for the control practice factor for
construction sites are given in Table 9 (Ports, 1973). Agricultural
land use P factor values can be found in Agriculture Handbook 282
(Wischmeier and Smith, 1965).
The C and P factors are the subject of much controversy among erosion
and sedimentation experts of the US Department of Agriculture (USDA)
and the Soil Conservation Service (SCS). These factors are estimates
and many have no theoretical or experimental justification. It
has been suggested that upper and lower limits be placed on these
factors by local experts to increase flexibility of the Universal
Soil Loss Equation for local conditions.
The P factors in the upper portion of Table 9 were designated as
estimates when they were originally published. SCS scientists have
found no theoretical or experimental justification for factors sig-
nificantly greater than 1.0 (Soil Conservation Service, Hyattsville,
Maryland, 1973). Surface conditions 4, 6, 7 and 9 .(P <_ 1.0), Table 9,
also are estimates with no experimental justification.
Sample Calculations
To demonstrate the use of the equation for sediment control planning
on construction sites, a hypothetical one-acre residential develop-
ment can be examined. The development site is located on the
Conestoga silt loam soil, with an average slope of 10 percent and
35
-------
Table 8. CROPPING MANAGEMENT FACTOR C
(Maryland Water Resources Administration, 1973)
LO
CTi
Type of cover C Value
None (fallow) 1.00
Temporary seedings :
First sixty days 0.40
After sixty days 0.05
Permanent seedings :
First sixty days 0.40
After sixty days 0.05
After one year 0.01
Sod (laid immediately) 0.01
Rate of
application
Mulch (tons/acre)
Hay or straw 0.5
1.0
1.5
2.0
Stone or gravel 15.0
60.0
135.0
240.0
Chemical mulches
First ninety days a
After ninety days a
Woodchips 2.0
4.0
7.0
12.0
20.0
25.0
C Value
0.35
0.20
0.10
0.05
0.80
0.20
0.10
0.05
0.50
1.00
0.80
0.30
0.20
0.10
0.06
0.05
Maximum
allowable
slope length
20 feet
30
40
50
15
80
175
200
50
50
25
50
75
100
150
200
a As recommended by manufacturer
-------
Table 9. EROSION CONTROL PRACTICE FACTOR P FOR CONSTRUCTION SITES
(Ports, 1973)
Surface condition with no cover Factor P
1. Compact, smooth, scraped with bulldozer
or scraper up and down hill 1.30
2. Same as above, except raked with bulldozer
root raked up and down hill 1.20
3. Compact, smooth, scraped with bulldozer
or scraper across the slope 1.20
4. Same as above, except raked with bulldozer
root raked across slope 0.90
5. Loose as a disced plow layer 1.00
6. Rough irregular surface, equipment
tracks in all directions 0.90
7. Loose with rough surface greater than 12" depth 0.80
8. Loose with smooth surface greater than 12" depth 0.90
Structures
1. Small sediment basins:
0.04 basin/acre 0.50
0.06 basin/acre 0.30
2. Downstream sediment basins:
with chemical flocculants 0.10
without chemical flocculants 0.20
3. Erosion control structures:
normal rate usage 0.50
high rate usage 0.40
4. Strip building 0.75
37
-------
an overland flow length of 200 feet. Froff Table 7, the K value for
the Conestoga series is. found to be Q.43. The soil loss ratio is
read from Figure 4 by locating the 10 percent curve and 20Q foot
slope length, LS =1.93. For the moment, assume that no erosion
control practice is instituted and that the soil is stripped of all
vegetative cover during the construction process. For this situa-
tion, both C and P factors should be set equal to one. To calculate
the expected annual erosion from this site, assume that the rainfall
index value is 150 as read from Figure 2 for Lancaster County,
P enns ylv ania. Then,
A = 00 (K) as) (C) (P) (3)
Annual Soil Loss = (150)(.43)(1.93)(1)CD - 124 tons/acre =
279 MT/ha
Assuming the period of construction before ground cover is applied
to the site is four months, from April through July, factor M
(Table 5) is used to calculate the soil loss.
April-July
Soil Loss
(A)(M) - (124) (.29) = 36 tons = 33 MT
For a single storm whose rainfall hyetograph is shown in Figure 5 the
rainfall factor, R, must be calculated according to Equation (4).
The single storm R value is calculated with the use of Table 4 as
shown below.
.9 •
.6
ID
I
e
.3-.
10 20 30 40 50
minutes
60
70
80
90
FIGURE 5. Rainfall Hyetograph
38
-------
The expected soil loss from this one-acre site during this storm is:
Single Storm
Soil Loss = (3.25)0 43) (1.93) = 2.7 tons - 2.45 MT
According to Table 6, this hypothetical storm produces around one-
tenth as much erosion as a one-year storm, as calculated below:
One-Year Storm , , „
Soil Loss = CR)(K)CLS)QO = (124) (.2) = 24.8 tons -
22.5 MT
By applying one ton of mulch over this one-acre construction site and
placement of a downstream sediment basin, the erosion from this single
storm event is reduced considerably.
• Single Storm
Soil Loss = (2.7) (C)OP) = (2.7) (.15) (.2) - 0.08 tons =
With Control 0.073 MT
The percent effectiveness of these control measures is calculated from
Equation (6) below:
Percent Effectiveness = (1-CP) x 100 (6)
For this example, the effectiveness is 97%.
Summary of Limitations in the Use of the Universal Soil Loss Equation
The SWMM user should understand that the Universal Soil Loss Equation
is an empirical equation based on statistical analyses of measured
sediment yield from small plots. The sediment yield from these plots
is, by definition, soil loss. The Universal Soil Loss Equation is
being used successfully by the SCS to determine soil loss and to
evaluate the effects of conservation measures on soil loss from farm
fields. However, for larger areas, the Universal Soil Loss Equation
gives only the maximum potential- sediment yield. The actual yield
will generally be less.
When the Universal Soil Loss Equation is used for computing sediment
yields from watersheds, all factors except the rainfall factor must
be modified (Williams and Berndt, 1972). Sediment yield may be
defined as the amount of soil material eroded from a watershed that
is transported and deposited in a stream either as suspended sediment
or as settled bed material* or both. It is dependent on gross erosion
in the watershed and transport and deposition by runoff into streams
or reservoirs. The percentage of sediment delivered from the erosion
source to any specified downslope location is affected by such
39
-------
factors as size and texture of erodible, material, climate, land use,
local environment, and physiographic position. The change (per
unit area) of this downstream sediment movement, from the source to
any given measuring point, is expressed by the
defined as: the ratio of sediment -yield at the measuring point to
total material eroded from the watershed and drainage system upstream
from the measuring point. Thus, adequate determination of sediment
yield from watersheds requires a modified Universal Soil Loss Equa-
tion and an equation for predicting delivery ratios*
The SWMM user is referred to Williams and Berndt (1972) for an in-depth
discussion of this topic. It is important to note that successful
use of this approach requires a multiple regression analysis to relate
delivery ratios to watershed characteristics for local conditions.
This requirement might result in considerable inconvenience to some
SWMM users. Another alternative would be to use Flaxman's (1972)
multiple regression equation. This equation relates sediment 'yield
to four independent variables: one for the climate; one for the
topography; and two for soil characteristics. However, the study
area consisted of 11 western states (Arizona, California, Colorado,
Idaho, Montana, Nevada, New Mexico, Oregon, Utah, Washington, and
Wyoming), suggesting that use of this relationship outside of the
study area would be risky. The study did demonstrate that a few
parameters can delineate most of the variation in sediment yield in
the western United States despite the great range in climate, topogra-
phy, soils and geology, and land conditions. This is encouraging
because it may also be true for other areas in the continental
United States.
Wischmeier and Smith's relationship between kinetic energy and
rainfall intensity, Equation (2), was developed for rainfall inten-
sities ranging from 0.01 to 6.0 inches/hour (0.03 to 15.2 cm/hr).
Modeling of storms having larger or smaller intensities constitutes
an extrapolation of the least-squares linear regression relationship.
Some of the other limitations of the Universal Soil Loss Equation
pertain to the five major factors and have been discussed at length
throughout the previous sections. We may summarize as follows:
1) a value of the rainfall factor R greater than
350 is not recommended;
2) the R factor does not account for soil losses
due to snowmelt and wind erosion;
3) local soils experts should be consulted before
selection of the C and P factors is made;
40
-------
4) sediment yield from the watershed may be pre-
dicted only after modification of the Universal
Soil Loss Equation and knowledge of the delivery
ratios;
5) prediction is accurate only for storms with rain-
fall intensities ranging from 0.01 to 6.0 inches/
hour (0.03 to 15.2 cm/hr); and
6) some error is incurred when computing soil loss
for an individual storm rather than using a time-
averaged (annual) value.
The Universal Soil Loss Equation, as incorporated into the SWMM, can
be applied successfully to estimate gross erosion rates from urban
subcatchments or sub-basins for a wide range of rainfall, soil, slope,
and cover conditions; provided, however, that the above limitations
are recognized and observed. The predictive capabilities of the
erosion modeling routines are enhanced by a better understanding of
their range of applicability.
Incorporation of the Universal Soil Loss Equation into the Storm
Water Management Model
An example has been shown of how the Universal Soil Loss Equation can
be used to predict the total erosion from a given land area for a
single rainfall event. Equations (3), (4) and (5) have been programmed
and incorporated into the Storm Water Management Model so that sus-
pended solids generated from eroding sites are added to the solids
flushed from streets and other impervious surfaces. In the Model,
the gross erosion for a specified subcatchment is calculated as follows;
1) the SWMM user specifies six input parameters
which are constant for each subcatchment;
a) the erosion area in acres;
b) the flow distance in feet from the point
of origin of overland flow to the point
at which runoff enters the gutter or
manhole, the runoff length;
c) the soil factor, K;
d) the cropping management factor, C;
e) the control practice factor, P;
f) the ground slope in ft/ft.
2) the SWMM user specifies the time-step varying
rainfall hyetograph, inches/hour;
41
-------
3) the SWMM user specifies the storm varying
highest average 30-minute rainfall intensity,
inches/hour;
4) the Model computes a rainfall index, at each
time-step, using the appropriate rainfall
intensity specified by the hyetograph;
5) the rainfall index, at each time-step, is then
multiplied by the specified highest average
30^ninute rainfall intensity to give the
rainfall factor, R;
6) the rainfall factor, R, is subsequently
multiplied, again at each time-step, by a
constant representing the product of (LS)
(K) (C)(P)(area) to give erosion in tons;
7) the tons of erosion are converted to a quality
concentration, suspended solids in mg/1, from
knowledge of the flow magnitude at each
time-step; and
8) the suspended solids concentration due to
erosion is added to the suspended solids con-
centration due to other sources, and both
the total suspended solids concentration and
the suspended solids concentration without
erosion are printed for each time interval
as shown in Table 10, columns 4 and 10,
respectively.
C. VEGETATION
The literature contains many studies on the quantity of stormwater
runoff. Several studies have been done that amply demonstrate the
poor quality of such urban runoff and stormwater overflows. How-
ever, little mention has been made of the sources of this water
pollution. In a Chicago study, it was demonstrated that vegetation
could account for a large percentage of street solids on an annual
basis (Heaney and Sullivan, 1971). Considering vegetation or leaf
litter as a possible source of water quality impairment, the following
section will be devoted to characterizing this material by quantifying
amounts of litter, produced, its chemical and nutrient content, and
its potential effects on water quality.
There is a wealth of data on leaf litter production. However, most
of this information is taken from forestry data and represents a
closed-canopy environment. Most urban areas are far from a closed-
42
-------
Table 10. EROSION PRINTOUT
TFJl PFMNSVI ViNll STFVFN^ 4V/E 01S T 5 IC T •»»• WBE VER5IHN *»•»
,;,? «JOV£MBER 29, 1971 OUBtTIO1- 6.5 HRS. STUDY »u (STHHM kg)
$!J«1»45V 3F .1U4NTITY 4*lD .11141. ITY RESULTS 4T LOC4TIHM J|
F'.3- ". CF3 4NO Gi.iALITY I'l fMfi/l.)
TI"E SLJ-, BOO SU3-S COLIF SET-S NIT PflU GBflSE SUS-S WITHOUT EPOS1PN
It- ao.no 0.32 fc2.»>7 638iS.M S5475.S9 0.02 n.no O.ou n.a7 a.7«
12 S.OO 0.20 4?.SB 1717. (il u O.fl2 0.00 O.Ou n.uu B]aS
15 ss.30 o.S.(.5 32u 0.00 0.10 0.00 fl.OU ol3°
(° I'.oi; C.JD ii".?1; n.ts uab«.a5 0,0'o n.in 0.00 fllou f. 53
•° ie.oo o.ja .?2 o.T! aauno n.oo A.no oloo olou o:i»
'" 0.3 o.Oi a«.70 n.T» UU26.3'' 0.00 O.no n^OO o^ou oil*
"> 25.10 0.02 u«.l? n.-,8 «aia.7? 9.00 flloo njoo o;oa n!]*
?0 50.00 0.01 U8.13 0.38 «a06.27 O.Oo 0.00 0.00 fl.OU 0.3«
?S IS.00 0.01 08.17 0.38 «3V».
-------
canopy forest situation, but with the use of some basic tree density
data, a correlation could be obtained to estimate the total leaf
litter production in a typical urban community.
Tvee. Litter Production
In their review of the literature on litter production by the world's
forests, Gorham and Bray (1964) state that "the study of the quanti-
tative aspects of litter fall remains an important part of forest
ecology, dealing with a major pathway for both energy and nutrient
transfer in this type of ecosystem." In the urban environment, this
energy and nutrient transfer might be disrupted (resulting in removal
of nutrient energy) when this organic material or fuel for the nutrient
cycle is not allowed to reach the upper soil horizons of such an
ecosystem.
In a work by Chandler (1941), he discusses the studies of Ebermayer
(1876) and others concerning the amount of dry matter in the annual
litter fall of mixed hardwood forests, composed principally of European
beech with an admixture of oak and birch. The values of dry matter
ranged from 2,300 pounds to 5,000 pounds/acre (2577 to 5602 kg/ha) with
an average of approximately 3,123 pounds (3499 kg). Such figures
appear to be nominal for this climate. Gorham and Bray (1964) indi-
cate from their review of the literature that as much as 13,500 pounds
of litter/acre-year (15,125 kg/ha/yr) might be deposited in equatorial
forests of the Congo. Annual litter production can be based on broad
climatic zones: Equatorial, Warm Temperate, Cool Temperate, and
Arctic-Alpine. The Warm Temperate group ranges between 30° and 40°
both north and south of the Equator and includes the southern part of
the United States, e.g., Florida, the Cardlinas, Tennessee, and
California. The Cool temperate areas in North America range from
Missouri and the mountains of California to Minnesota and Quebec, or
about latitude 37-47°N. Within these broad climatic regions, Gorham
and Bray (1964) have compiled annual leaf litter production data.
For some of the areas in North America, the data are represented in
Table 11. In general, it is recognized that leaf material contributes
60 to 70 percent of the origin of forest litter as shown in Table 12.
The Angiosperms are the dominant plant life of the geological era
in which we live. This grouping contains the broad-leafed hardwoods
and other flowering plants, while Gymnosperms are mainly pines or
cone-bearing trees having needles. The Gymnosperms, together with the
Angiosperms, are the seed plants. Many other forms of flora are also
included in these divisions. Some of the tree species encountered
in this study and in Table 11 are listed in Table 13 with both their
common and scientific names.
44
-------
Table 11. ESTIMATED AMOUNT OF TREE LITTER IN VARIOUS LOCATIONS IN THE UNITED STATES
(Daubenmire, 1953; Gorham and Bray, 1964; Heyward and Barnette, 1936)
Location
California
Tennessee
Tennessee
N. Florida
S. Carolina
Montreal
Minnesota
Missouri
New York
New York
Wisconsin
Connecticut
Latitude
39N
36N
36N
30N
35N
47N
47N
39N
43N
43N
43N
42N
Longitude
123W
84W
84W
83W
82W
74W
95W
92W
77W
76W
89W
73W
Plant
community
Pinus pmderosa
mixed Quercus spp.
Pinus virginiana
Queraus spp.
Pinus palustris
360 trees/acre
Finns palustris
470 trees/acre
Pinus palustris
788 trees/ acre
Pinus palustris
972 trees/acre
Finite echinata and
mixed angiosperms
Betula pcpulifolia
Pinus banksiana
Pinus banksiana and
Pinus resirosa
Pinus echinata
Tilia Americana and
mixed angiosperms
Pinus strobus
Queraus alba and
Queraus velutina
170 trees/acre
Pinus strobus
Litter fall
(Ib/acre/yr)
Leaves Other
1,873
3,390 624
3,122 892
2,408
3,032
2,408
2,943
3,390 1,427
1,516
1,873
1,784
2,765
2,765
4,103 1,338
Total
2,587
4,014
4,014
4,817
3,390
5;441
3,568
Acer eacahorum
Queraus rubra and
mixed angiosperms
1,873
45
-------
Table 12. SOURCES OF FOREST LITTER
Source % of total litter
Leaves 60-70
Branches 12-15
Bark 1-14
Fruit 1-17
46
-------
Table 13. COMMON AND SCIENTIFIC NAMES OF TREES
Common name
Scientific name
American Elm
Basswood
Red Maple
Sugar Maple
White Oak
Red Oak
White Pine
Red Pine
Ponderosa Pine
White Ash
Black Birch
Ulntus ameTi.ca.na L.
T-il-ia ameri-Qana L.
Aeer rubTum L.
Acev saoohavum Marsh
Querous alba L.
Quersus borealis
P-inus strobus L.
Pinus resi,nosa A.T.
Pinus ponderosa Dougl.
Faxi-num ameri-cana L.
Betula lent a L.
47
-------
Because of their evergreen nature, Gymnosperms might be expected to be
more productive than deciduous Angiosperm trees, although this factor
may be countered to some extent by the tendency- for Angiosperms to
occupy more fertile sites (Gorham and Bx.ay, 1964). As far as litter-
fall is concerned, Table 14 indicates that when a wide range of
sites are considered, Gymnosperms yield about one-sixth more total
litter annually than Angiosperms, the difference amounting to 446
pounds/acre (500 kg/ha). The difference for leaf litter alone is
approximately 8 percent or 178 pounds/acre.
This trend toward greater litter production from evergreen trees is
observed in other studies. Chandler calculates the amount of dry
matter deposited annually by hardwood forest tree leaf litter in
closed, second-growth stands in central New York ranged from 2,425 to
3,020 pounds/acre (2,717 to 3,384 kg/ha) (Chandler, 1941). In a
similar study, he determined the amount of dry needles from northeastern
conifers was 2,009 to 3,367 pounds/acre (2,251 to 3,772 kg/ha) (Chandler,
1944). These figures he feels are typical of healthy vigorous stands.
Another interesting point to be made from Chandler's work is that even
though conifers shed needles during the entire year the greatest drop
of needles occurs along with that of deciduous trees in autumn and
winter.
From the combined data by Gorham and Bray in their exhaustive work
they have been able to determine annual litter production in the four
major climatic zones and have correlated these production figures to
latitude. These data are represented in Table 15 and Figure 6.
We are interested in relating this type of data to an environment where
the trees do not represent a closed-canopy forest situation. If tree
density is examined, we find little effect by density differences within
a closed-canopy forest (Gorham and Bray, 1964). This can be seen in the
data by Heyward and Barnetts (1936). The results of this north Florida
study were presented in Table 11. Several studies have shown a distinct
correlation between annual fall of leaf litter and stand basal area in
both Gymnosperm and Angiosperm stands (Gorham and Bray, 1964). Basal
area is defined as the cross-sectional area of the trunk of a tree at
breast height. This basal area for a tree species is roughly proportional
to its crown size.
Studies of this type have shown that litter fall averaged about 70 to
75 kg/m over a basal area range from 8 to 40 m /ha (Gorham and Bray,
1964). This represents approximately 14.3 to 15.3 pounds/feet2 over a
basal area range from 34.9 to 174.4 feet2/acre.
Also in their literature review, Gorham and Bray have shown that in
a forest system, thinning results in a decrease of litter production
48
-------
Table 14. A COMPARISON OF LITTER PRODUCTION
BY EVERGREEN AND DECIDUOUS TREES IN THE
NORTHERN HEMISPHERE
(Gorham and Bray, 1964)
No. regions Evergreen Deciduous
averaged gymnosperms angiosperms
(Ibs/acre-yr)
Total litter
Leaf litter
by difference
8 3,300 2,854
9 2,319 2,141
981 713
Other
observed 4 624 624
49
-------
Table 15. ANNUAL LITTER PRODUCTION IN FOUR MAJOR CLIMATIC ZONES
(Gorham and Bray, 1964)
en
o
Leaves
Arctic-Alpine
Cool Temperate
Warm Temperate
Equatorial
No. regions
averaged
1
15
8
2
Ibs/acre
624
2,230
3,211
6,066
Other
No. regions
averaged
1
10
5
1
Ibs/acre
357
803
1,695
3,122
Total
No. regions
averaged
3
22
7
4
Ibs/acre
892
3,122
4,906
9,723
-------
Ul
-EQUATORIAL
A - WARM TEMPERATE
-COOL TEMPERATE
(NORTH AMERICA)
O-COOL TEMPERATE
(EUROPE)
D-ARCTIC ALPINE
u
o
m
•o
20 30 40 50 60
NORTH OR SOUTH LATITUDE (DEGREES)
FIGURE 6.
Annual Production of Total Tree Litter in Relation to Latitude.
(Gorham and Bray, 1964)
-------
that is roughly proportional to the degree of thinning. Another effect
worth mentioning is the effect of litter removal from a forested area
which causes extreme reduction of forest growth (Mayer-Krapoll, 1956).
The loss of nitrogen in the litter is believed to be significant in
lessening forest productivity. In light of this fact, litter that
is removed from an area where the nutrients can be utilized represents
a substantial amount of nutrients introduced into another system. In
an urban community, it would appear that litter not retained on the
soil and finding its way to some impervious surface could contribute
to the nutrient load of runoff. This area of potential effects from
litter will be discussed later in this section. It was also pointed
out by Gorham and Bray that the age of the stand had little effect on
total annual leaf-litter production. Other studies have confirmed
their review.
Work by Carlisle &t al. (1969) indicates similar results to those
summarized by Gorham and Bray. In this study, the wooded area was
stocked with uneven-aged (about 40 to 120 years) trees with a 90 to
95 percent closed-canopy and 64 trees/acre (158 trees/ha). The tree
species were mainly oak and birch. The mean annual rainfall of this
area was 67.5 inches (171.5 cm). Over a four-year study period, an
average of 9.56 x 10 leaves/acre/year (2.36 x 10' leaves/ha/yr) were
produced with a mean dry weight (seasonal maximum) of 3,407 pounds/
acre (3,817 kg/ha). The dry weight of the total litter fall (leaves,
twigs, flowers, acorns, etc.) was 3,441 pounds/acre (3,855 kg/ha).
The autumn leaf fall averaged 1,967 pounds/acre (2,204 kg/ha) or 57.76
percent of the total annual leaf fall. This represents approximately
53 pounds of leaves per tree annually (24 kg of leaves per tree annually).
Recognizing that there is no significant correlation between litter
fall and tree density, this figure, however, represents the higher end
of a broad spectrum of values that could be obtained from similar
calculations.
If we now can consider that we have some feel for the amount of litter
fall produced annually by trees, let us seek to determine the nutrient
content of such matter.
Nutrient Content
The literature contains a liberal quantity of data on the nutrient
content of litter. This section reviews only those findings pertinent
to our discussion and that best exemplify the trends.
Considerable interest has been directed to the study of chemical compo-
sition of tree leaves. It has been recognized for some time that the
composition of leaves varies with species, although it may be modified
to some extent by variation in site. Daubenmire (1953) concluded that
the chemical composition of tree leaves is primarily a characteristic
of the species and substantiates the general conclusion reached by
earlier workers.
52
-------
Forest liter is not wholly organic, but always contains some mineral
matter. Gorham and Bray (1964) have summarized some of this mineral
content data. They found that the Angiosperms contained more mineral
material than the Gymnosperms. The majority of the Gymnosperms had
from 2 to 5 percent ash content while the Angiosperms ranged from 4 to
14 percent ash. Mean ash content as a percentage of dry matter of
19 Gymnosperm species was 3.7 percent and of 56 species of Angiosperms,
8.4 percent. They also noted that the ash content may vary slightly
with region,'owing mainly to soil differences. A detailed analysis of
the major elements comprising mineral material in litter are numerous,
and have been reviewed extensively by Lutz and Chandler (1946). How-
ever, the scope of this paper limits this data to some broad ranges.
Daubenmire (1953) found that on his analyses for N, P, K, and Ca, values
ranged from 0.3 to 0.7 percent for N, 0.04 to 0.24 percent for P,
0.13 to 1.19 for K, and 0.19 to 2.19 percent for Ca. These percentages
represent total content of oven-dry weight of freshly fallen foliage.
In an earlier study by Coile (1936), he determined that leaves of
white oak contain 1.25 percent N. Those of red cedar, black oak, dogwood,
shortleaf pine, loblolly pine, yellow poplar, and red gum contain
progressively less to a minimum of 0.5 percent for red maple. Tables
16 and 17 give some average values for the nutrient content of newly
fallen Gymnosperm and Angiosperm foliage. Consistent with these nutrient
values were those of Carlisle et al. (1966) with N and P values of
18.9 pounds/acre/year and 1.1 pounds/acre/year (21.2 and 1.23 kg/ha/yr)
respectively.
Leaf age was considered by Morrison in a recent study. He found that
with increasing leaf age, P, K, and Mg concentrations decreased whereas
Ca, Fe, and Mn increased. N showed no significant trend with leaf age.
Other recent work was that of Kawahara (1972). He reported litter
fall and nutrients throughout a year in six forests. The amount of
total litter fall was 2,033 to 6,282 pounds/acre/year (2,277 to 7,038
kg/ha/yr). Leaf litter accounted for 54 to 78 percent in coniferous
forests and as much as 83 percent in deciduous forests. The total
litter-fall/year contained 13.6 to 64.7 pounds/acre N (15.2 to 72.5
kg/ha), 0.6 to 4.5 pounds/acre P (0.7 to 5.0 kg/ha), 3.2 to 14.6 pounds/
acre K (3.6 to 16.4 kg/ha), 19.9 to 64.8 pounds/acre Ca (22.3 to 72.6
kg/ha), and 2.4 to 12.3 pounds/acre Mg (2.7 to 13.8 kg/ha). N, P,
and K were more abundant in broad-leaved forests than in coniferous
forests. He also computed the ratios of nutrients return with total
litter fall to nutrient uptake as 33 to 52 percent for N, 36 to 69 per-
cent for P, 23 to 43 percent for K, 62 to 85 percent for Ca, and 52 to
98 percent for Mg.
In their 1969 study, Moir and Grier found pine forest floor humus
weights ranged from 11.5 to 18.5 tons/acre/year. This humus and organic
accumulation of the forest floor was determined to be almost entirely
derived from lodgepole pines. The reported nutrients from their study
53
-------
Table 16. THE CONCENTRATION OF NUTRIENTS IN NEWLY FALLEN
GYMNOSPERM AND ANGIOSPERM TREE LEAF LITTER
(% Dry Weight)
(Carlisle and White, 1966)
No.
Author Species Location
Ui
*" Gymnosperms Lutz and 9 U.S.A.
Chandler
(1946)
Angiosperms Lutz and 14 U.S.A.
Chandler
(1946)
Angiosperms Duvigneaud 25 Europe
& Denaeyer-
deSmet
(1964)
N
0.58
to
1.25
0.51
to
1.01
0.95
to
4.60
P
0.04
to
0.10
0.09
to
0.28
0.11
to
0.82
K
0.12
to
0.39
0.40
to
1.18
0.22
to
4.30
Ca
0.55
to
2.16
0.99
to
3.84
1.00
to
5.50
Mg
0.14
to
0.23
0.22
to
0.77
0.15
to
0.80
Ash
3.01
to
4.33
5.71
to
15.16
6.20
to
27.30
-------
Table 17. AVERAGE QUANTITIES OF NUTRIENTS FALLING
IN THE LITTER OF DIFFERENT TREES
(Ibs/acre/yr)
(Carlisle and White, 1966)
Trees
leaves only
Angio sperms
Gymno sperms
Author
Lutz and Chandler
(1946)
Lutz and Chandler
N P K
13.2 2.5 10.7
18.8 1.4 5.2
Ca
52.2
21.1
Mg
7.3
3.6
(1946)
-------
are as follows: total N ranged from 290 to 480 pounds/acre (325 to
538 kg/ha), total P ranged from 29 to 44 pounds/acre (33 to 49 kg/ha),
total K ranged from 62 to 151 pounds/acre (69 to 169 kg/ha) , and total
Ca ranged from 89 to 222 pounds/acre (100 to 249 kg/ha). Nutrient
quantities, like these and others, indicate substantial amounts that
could potentially contribute to the load of a receiving water. This
material could directly fall into bodies of water, it could be trans-
ported by wind or runoff, or it might have such nutrients leached from
it and then be transported in solution to a receiving body.
Normally leaves fall from trees, becoming part of the forest floor or
humus, and then begin to decay and decompose. The rates of decomposi-
tion vary among species and with varying chemical composition, tempera-
ture, moisture, and various other physical conditions. The results of
such findings can be found in works by Melin (1930) , Kucera (1959),
Hayes (1965) and others. Of major interest to this report is the
decomposition and nutrient release of this leaf litter as it affects
water quality.
Effects on Water Quality
Previous studies have demonstrated impaired water quality by the
addition of natural organic matter to small streams or pooled streams.
Of interest to this subject are studies by Slack (1964) and Slack and
Feltz (1968). From his study of the Cacapon River, Slack attributes
changes in water color and changes in water composition to the accumu-
lation of litter at low flow. He reports an increase in the concentra-
tions of some solutes with a decrease in water quality. Dissolved
oxygen decreased with the increase in water color. Chase and Ferullo
(1958) showed their leaves deplete dissolved oxygen under aerobic condi-
tions in the laboratory. After 386 days, maple leaves had consumed
a weight of oxygen equivalent to 75 percent of their initial dry weight.
Slack indicates that an increase in stream flow tends to reduce color
and improve water quality.
The results reported by Slack and Feltz in their 1968 study indicate
similar findings. In addition to their field studies, laboratory
tanks were used to evaluate the effects of various species of leaves.
They reported that the greatest effects of leaf litter on water quality
occur during chemical and biological dissolution of the submerged
leaves. Previous study has indicated that leaf litter decomposes more
rapidly in the presence of moisture (Heyward, 1936). Studies have
shown that the solute load of streams can be altered rapidly by leaf
fall. They also indicate that the degree of water quality degradation
is directly related to the weight of litter/unit volume of water and
to species composition of litter. In a 1933 study by Haupt (1933), he
reported his laboratory studies on the underwater decomposition of
leaves of various trees. He concludes his report by advising that
56
-------
possible reservoirs be protected from leaf contamination. He states
that pines are best for forests around reservoirs, and that oaks are
especially objectionable.
In the urban environment, these types of studies lend themselves to
considerations of stormwater systems and runoff water quality. If
such organic material enters a storm sewer or catchbasin where shallow
stagnant pools exist, serious water quality impairment may result.
From our previous discussion of litter, it should be recognizable that
tree litter may constitute significant fractions of organic material
in storm runoff in an urban area. We have also considered the tremen-
dous quantities of nutrients available from such litter that can be
readily dissolved. Nutrients leached from leaves and vegetation by
precipitation are either returned to the soil or removed in runoff
water. White (1973) reports the nutrient enrichment of runoff may be
enhanced during the fall and winter by the rupturing of plant cells
by freezing. In his study, he indicates that freezing increased
nutrient release if the vegetation were growing when it was frozen.
Newly fallen litter not yet decomposed might be affected in a similar
manner. These findings are of additional concern in northern climates
with snow melting and thawing contributing to runoff.
In another study, leaf leachates, simulating canopy drip during rain-
fall, were collected from southern red oak and longleaf pine (Malcolm
and McCracken, 1968). From the representative canopy drip sampled, it
was estimated that approximately 18 pounds/acre/year (20 kg/ha/yr) of
organic matter could be contributed to the soil from this source. Con-
sidering these findings, large quantities of nutrients could be added
to the organic load of urban runoff if it is permitted to fall upon an
impervious surface. With the amount of nutrients that trees potentially
return to the soil in the form of either leachate or litter capable of
being leached by precipitation, it would seem unwise to create impervious
areas immediately beneath trees or vegetation.
Grass Litter
In addition to the leaf litter that might be contributed by tree litter
fall, grass litter is also a potential source of organic material.
When we refer to grass litter, we are referring to grass clippings that
result from the production and harvesting of lawns or turf grasses. As
a lawn produces foliage, it is clipped or mowed to maintain its
attractive appearance. Well-manicured herbage is common in most urban
and residential areas and is, again, a source of civic pride and pleasure.
However, the disposal of the clippings or discharge of such litter in
such a manner as to find its way into runoff of an area could be a
source of pollution. As in the case of tree litter, grass litter is
organic material, composed of various nutrients, and readily available
57
-------
for decomposition. It has been recognized that grasses act as filters
and entrap solid materials of small sizes that fall among them. If
minimal grass clippings result from regular mowings,this litter, if
allowed to remain on the upper zone of the lawn, will soon find its way
down among the blades or leaves and gradually decompose. Among turf
specialists, it has been observed that as the clippings decompose,
their nutrients are readily incorporated into the growing.herbage. It
is also felt that a minimal amount of lateral nutrient movement occurs
in the natural filter. When this litter is not allowed to remain and
return its nutrients to the root structure but is removed and not managed
properly, nutrient enrichment of urban runoff is possible.
Once again, we consider the sweeping or dumping of this litter onto
an impervious surface where stormwater runoff can expedite decomposition
and leach nutrients into the stormwater system.
Such litter can be the source of organic material and nutrients as can
be seen by the following information. In a Florida study (Institute
of Food and Agricultural Sciences, University of Florida, April 1972),
summarized data for several years siiowed that numerous ryegrass varieties
averaged from 3,676 to 5,612 pounds of dry matter/acre/year (4,118 to
6,288 kg/ha/yr). In the same report, white clovers (primarily a forage
crop but found in some lawns) averaged between 3,805 to 5,108 pounds of
dry matter/acre/year (4,263 to 5,723 kg/ha/yr) over a .13-year study.
In a study by Ruelke and Prine (1971) , yields of various forage grasses
were determined. Several of these are also commonly used as turf grasses
for lawns. They found that Pensacola bahia grass could yield an annual
average of 8,126 pounds/acre (9,160 kg/ha) during their five-year study.
Coastal bermudagrass was found to produce an annual yield of 9,135
pounds of dry matter/acre (10,235 kg/ha) in that same study. It was
observed that yields of dry matter were proportional to levels of
fertilization. The response to amounts of nitrogen, phosphorous, potas-
sium, calcium, and magnesium removed in the harvested forage was signifi-
cantly greater as fertility levels are increased. The five-year mean
first cutting dry matter yields of Coastal bermudagrass over all fertility
levels was 2,542 pounds/acre (2,848 kg/ha). Such data probably represent
the upper limits for yields of fertilized grasses common to Florida as
forage crops and lawns. The total nitrogen content of these grasses
ranged from 1.7 percent to 2 percent of the dry matter. This percentage,
signifies again, large quantities of nutrient material that could be
derived from vegetative litter. With respect to the amounts of nutrients
applied and those removed in the cut dry matter, Ruelke and Prine (1971)
observed that approximately 50 percent of the N, 88 percent of the P,
and 79 percent of the K was recovered in the cut dry matter. Grasses
will remove differing percentages of nutrients depending on the variety
of the grass (Ruelke and Prine, 1968, 1971).
58
-------
Such studies indicate the potential quantities of nutrients available
from grass clippings or grass leaf litter. This is a brief review
of the grass litter contribution from an urban environment, but
deserves considerations of source control and urban land use manage-
ment when discussing sources of pollution in urban runoff.
Sum/nary
In concluding this section concerning litter-fall and its effects on
urban runoff and stormwater, one is reminded of the immense complexity
of the problem. It is evident that additional study is required. The
present discussion only approaches the problem. Nevertheless, one can
immediately recognize a potential problem to be considered that exists
from natural litter-fall. In a forested area where this material
becomes incorporated into the soil, a recycling of the organics and
mineral nutrients takes place. Urbanization has a great impact on
water quality and therefore requires serious land use refinements.
McGriff (1972) has pointed out that urbanization increases the sediment
load carried by streams, decreases ground water recharge, promotes
eutrophication, and causes temperature variation in streams, all of
which tend to alter water quality. Tree litter fall becomes a problem
when it is remembered that several tons/acre of this material are
deposited annually with perhaps 100 pounds (45 kg) of nitrogen enter-
ing the runoff of the area. Tree-shaded streets are a source of civic
pride; however, more refined land use systems should be investigated
to eliminate a serious water pollution and management condition from
existing.
This section has been an attempt to characterize litter-fall by quantity
and quality, as it might apply to urban water quality. The review of
the literature warrants future research and consideration on this
subject.
D. TREATMENT MODEL REFINEMENTS
Objectives and Bacl
-------
At the time of development of the original Model, biological treatment
of storm flows was precluded due to the difficulties anticipated in
maintaining large quantities of biologically active sludge needed for
the treatment. As a result, it is difficult to simulate a real event
when the associated dry weather treatment facility comprises biological
treatment. Such is the case in the City of Lancaster where both
North and South sewage treatment plants provide secondary biological
treatment. In addition, the use of biological treatment for treating
storm flows is currently being demonstrated in Kenosha
Wisconsin (Agnew, 1973). It was, therefore, considered necessary to add
this option to the SWMM in order to apply it to Lancaster and to extend
the usefulness of the Model. Furthermore, new treatment devices such
as the swirl concentrator (APWA, 1972) and high-rate disinfection
(Glover, 1972) appear quite promising. The Revised Treatment Model
is shown in Figure 7 and incorporates the above devices as well as
others.
Recent Developments
Swirl Concentrator (Treatment Option IS)—This treatment device offers
a high degree of performance in reducing the amount of settleable solids
contained in combined sewer flows. The device consists of a circular
basin in which rotary motion of the flow is induced by the kinetic energy
of the incoming flow. This rotary motion causes the flow to follow a
long spiral path through the circular chamber allowing the settleable
solids to settle rapidly and be withdrawn from the bottom for subsequent
treatment.
Modeling of this treatment unit is based on the work of APWA (1972)
which utilized mathematical analyses and laboratory scale studies for
design criteria. A full-scale, 36 foot (llm) diameter unit, is proposed
for construction at Lancaster, Pennsylvania.
Data from the APWA study have been simplified and plotted in Figure 8.
The unit has been modeled in such a way that given the flow, the size
of the swirl concentrator, the particle sizes and specific gravities,
and the fraction of particles of each size, the efficiency of suspended
solids removal can be computed.
The Model performs the following steps:
1) Particle settling velocity, V , is determined by
an iterative procedure. Initially, assumed
velocity is computed using Stokes' law:
60
-------
(BYPASS)
(02)
OVERFLOW
(BYPASS)
CH
[
(BYPA
*OV
SS) (II)
ERFLOW
DESIGN FLOW
I BAR RACKS | u't' »-|SWIRL
EM-*,DESIGN FLOW
HIGH "RATE
)ISINFECTION (BYPASS) (21)
i ' '
(BYPASS)
'
(BYPASS)
1
(32:
(3D ' _, (33)
1
DISSOLVED AIR
FLOTATION
+
1
MICRO -
WD STRAINERS
i
(BYPASS) (51)
(BYPASS) (61)
(BYPASS) (71)
(12)
1
INLET ,,,,
PUMPING ^"
1
J* — CHEM
|
FINE SCREENS |
(34)
«
'
~]
CONCENTRATOR] (13) LEVEL 1
J
LEVEL 2
(35)
1 SEDIMENTATION 1 LEVEL 3
I
CHEM
t
HIGH -RATE
FILTERS <43'
*
f
EFFLUENT ,,-,,,
SCREENS ™"
i
t
OUTLET , .
PUMPING ^ a
|
•« CHEM
CONTACT /7P, HIGH
TANK U<:' DISINF
4
(44)
BIOLOGICAL . .-.,.- .
TREATMENT Ltvtl- *>
j
LEVEL 5
LEVEL 6
"}
-RATE (y-ai i pwpi 7
ECTION U5) LtVtL '
i
. ,-t.j nrrAuntMrn 01 ITFI nw
FIGURE 7.
Options Available in Revised Treatment Model.
61
-------
CTi
NJ
10 50 100 500 IOOO
36 FT. DIA. SWIRL FLOW RATE, cf S
FIGURE 8. Swirl concentrator removal efficiency for settleable solids as a
function of inflow and settling velocity, V (APWA, 1972).
S
-------
o
where g = gravitational acceleration, cm/sec
2
v = kinematic viscosity of water, cm /sec
S = specific gravity of the particle
s
d = diameter of the particle, cm
Using this V , the Reynolds Number, R, and the drag
coefficient, C , are computed by the equations:
V d
R = -5- (8a)
0.34 (8b)
The particle settling velocity is then computed again
using the following equation :
Vo = T£- (S -1)* (cm/sec) (9)
s j LD s
Equations (8) and (9) are then used iteratively until the
values of V used in Equation (8a) and determined in
Equation (9; agree to within 0.01 cm/sec.
2) In order to use the curves developed in the APWA report, it
is necessary to convert the particle settling velocity for
the design swirl to prototype (36 ft diameter) settling velo-
city by applying the Froude scaling correction factor as
follows:
V il 1/2
Vs proto
-------
6) This procedure is repeated for each particle size to get
the total percent removal. Settling velocities shown in
Figure 8 are for settleable solids. It is estimated that
percent removal of suspended solids will be 60% of that
for settleable solids. Hence, suspended solids removal
is taken as 60% of the value computed in Steps 4 and 5.
Improvements in these assumptions will be made upon receipt
of data from actual installations.
7) As modeled by the SWMM, BOD consists of a contribution due
to suspended organic solids as well as the dissolved amount.
Hence, BOD removal is assumed at 0.4 times the suspended
solids removal, except that if incoming BOD is less than
that computed in this manner, BOD removed will be taken as
0.35 times BOD input.
Sedimentation Revisions (Treatment Option SS)—Efficiency of suspended
solids removal by primary sedimentation is computed in the original
model using the following relationship (EPA, 1972a):
. ,,, . SS Cone, x .06 . | OVFRA - 300\ f .
n = 0.656 + — 0.40 j ^00 J (12)
where OVFRA is the overflow rate in gpd/sq ft and SS concentration has
units of ppm (mg/1).
In order to check the validity of this relationship, several data points
obtained from Eckenfelder and MeCabe (1956) and Metcalf and Eddy (1935)
are plotted on Figure 9. The curve represented by the above relation-
ship is plotted on the same figure for suspended solids concentration
of 400 ppm along with a sedimentation curve as developed by Smith (1968)
based on the following relationship:
n = 0.82e-°VFRA/278° (13)
where OVFRA is equal to overflow rate in gpd/sq ft. Figure 9 shows
that Smith's curve gives a better representation of the settling
efficiency than the one computed by the equation in the original SWMM.
The Model has, therefore, been revised to incorporate Smith's equation.
Biological Treatment (Treatment Option 44)—Biological treatment as
defined herein is taken as the aeration and secondary settling portion
of the treatment. Consequently, it is necessary to use treatment option
35 in conjunction with the treatment option 44. The efficiency of
suspended solids and BOD reduction in treatment plants using sedimen-
tation and biological treatment exceeds 90 percent for flow up to a
design flow. The modeling, therefore, is formulated in such a way that
overall efficiency of SS and BOD removal will be close to 90 percent
when both treatment options 35 and 44 are provided. When the inflow
to the unit exceeds the design flow, the excess flow is bypassed.
A sample output from secondary treatment is given at the end of this
sub-section—Revised Treatment Output.
64
-------
8.
8-
8-
o.
o.
in
UJ
cc
u.
o o
CJ
o
z
UJ
o
b.
L.
U
in
K)
LEGEND
® SS CONC. 400 ppm
X SS CONC. 250 ppm
+ SS CONC. 100 ppm
H SS CONC.460ppm
< SS CONC. 790 ppm
FROM:
METCALF a
EDDY, 1935.
ECKENFELDER
a MCCABE,
1956.
REVISED CURVE
ORIGINAL CURVE
X
\
1000 2000 3000 4000
OVERFLOW RATE gpd/sq ft
5000
FIGURE 9.
Removal Efficiency by Sedimentation. Original and Revised Relationships.
Compared with Data of Eckenfelder and McCabe (1956)
and Metcalf and Eddy (1935).
65
-------
High Rate Disinfection (Treatment Option 73)—The modeling of this
unit is based on the work of Glover (1972). This treatment option
is similar to chlorination (treatment option 72) with the following
exceptions:
1) The volume of the tank is selected to provide a
two-minute detention period at design flow.
2) The chlorine demand is taken at not less than
5 ppm and not more than 10 ppm.
The unit will provide a BOD reduction computed at 2.0 times the
chlorine demand but not more than 50 percent of incoming BOD. Coli-
f orm reduction is computed in subroutine KILL in a manner similar to
option 72. This unit can also be used as a sole disinfection device
on the overflow.
Other Treatment Options—Modeling of other treatment options shown
in Figure 7 remains the same as formulated in the original SWMM. It
is anticipated that other revisions may be made when new data become
available. In addition, there are several other treatment methods
currently under investigation such as:
1) rotating biological contactors;
2) high-rate plastic media trickling filters;
3) tube settlers;
4) powdered and activated carbon absorption; and
5) chemical treatment.
Revised Treatment Output— The following example illustrates the
modeling of primary and secondary treatment given the combined sewage
passing through the North Treatment Plant, Lancaster, Pennsylvania.
Figures 10, 11, and 12 show the time-varying inflows of combined
sewage, BOD, and suspended solids to the treatment facility. Tables
18 and 19 illustrate treatment model input. The treatment scheme
selected (see Figure 7 and Table 19) includes the revised sedimenta-
tion option (treatment option 35) and the new biological treatment
option (treatment option 44):
1) the storage model is bypassed at level 0;
2) the waste water is passed through bar racks
treatment option 12) at level 1;
66
-------
O
O
-------
oo
O
K)
o
o
m
o
CJ
O .
O H 1 1 1 1 1 r
II'OO I'OO 3=00 5=00
7'00 9=00 11 = 00
TIME, HOUR OF DAY
FIGURE 11. Combined BOD Load to North Treatment Plant Upstream of Diversion Chamber.
-------
en
O
O
O
00
c
E
CO ID -
en
to
O .
O .
C4
11=00 1=00 3=00 5=00 7 = 00
TIME, HOUR OF DAY
9=00
11=00
FIGURE 12. Combined Suspended Solids Load to North Treatment Plant Upstream of Diversion Chamber.
-------
L j
UP2ATHD BY UNIVERSITY 0= FLORIDA
DECEASES 1972
STORAGE PROGRAM CALLED
LANCASTER PENNSYLVANIA NCKTH DRAINAGE DISTRICT
OUTPUT FRO* EXTERNAL ST CR AGE/TH E AT-4ENT MODELS
Ta«NS=CPT ^COEL OUTFALLS AT THE FOLLOWING ELH"ENT NUMBERS!
1
INPUT TO TREATMENT MODEL SUPPLIED FRCH TRANSPORT MODEL EXTERNAL ELEMENT NUMBER 1
NU^PE<3 OF PUNS
TIME-STEP SIZE
NC. TI^E-STEPS MOCELEO
TRIBUTARY A96A
MO. TRAMSP. MCO. OUTFALi-S
SC. O= OQLL'JTANTS
TIME ZERO
5.00 WIN,
1 44
1014.00 ACRES
1
3
39600.0 SEC
Table 18. BEGINNING OF STORAGE/TREATMENT.
-------
—— RUN NO. 1 ——
INPUT DATA FOR TREATMENT PACKAGE FOLLOWS
CHARACTISTICS CP THE TREATMENT PACKAGE ABE
LEVEL MODE PROCESS
01
12
21
NC SEP. STORAGE
BAR RACKS
(BYPASS)
SEDIMENTATION
BIOLOGICAL TREAT
(BYPASS)
(BYPASS)
3 as
4 44
5 51
6 61
7 72 CONTACT TANK
* *** WARNING *«**
THE FOLLOWING COHPINATICNS OF TREATMENT OPTIONS ARE CONSIDERED ECONOMICALLY INAOVISEABLE - SIMULATION CONTINUES
72 WITH 3S
IPRINT * 2* ICOST * 9, IRANGE * 0. I TABLE - 0
SPECIFIED TREATMENT CAPACITY USED.
DESIGN FLOttRATE = 15.47 CFS.
THEATHFMT SYSTEM INCLUDES MOOCLE UNITS
OEI3IGN FLD» IS THEfcEFCPE INCREASED TO NEXT LARGEST MODULE SIZE
ADJUSTED DESIGN FLCuRATe = 15.47 CFS.. - 10.00 MGO.
(KMQD * 2)
NO STORAGE FROM A SEPARATE STORAGE MODEL IS ASSOCIATED WITH THIS TREATMENT MODEL
PRELIMINARY TREATMENT BY MECHANICALLY CLEANED BAR RACKS (LEVFL 1)
NU^BGP OF SCPEENS * 2
CAPACITY P£R SCKtRN » 7.73 CFS
SU^MEPCED AREA = 2.sa SQ.FT. (PERPENDICULAR TO THE FLOW)
TREATMENT 8Y SEDIMfc^TATION (LEVEL 3) - (NO ASSOCIATED STORAGE)
DESIGN UVHRFLC* RATE = 1240.00 GPO/SO.FT. (IfeOO SUGGESTED*
SEO T^NK OCFTH = 10.00 FEET (8 FC£T SUGGESTED)
NUMBER HP StD TANKS =• I
SURFACE AREA « 8064.51 SQ*FT./TAh-K
NC CHLORINE ADDED
Bl CLOGICAL TREATMENT (LEVEL 4)
NO EFFLUENT SCHEfSS (LEVEL 5>
Table 19. TREATMENT MODEL INPUT.
-------
3) level 2 is bypassed;
4) primary treatment is completed in the form
of sedimentation at level 3;
5) secondary treatment is provided by biological
methods (activated sludge aeration and secondary
settling) at level 4;
6) level 5 is bypassed;
7) level 6 is bypassed;
8) chlorination is the final treatment given in
the contact tanks at level 7.
Table 20 illustrates the effectiveness of the treatment scheme selected,
Approximately 90 percent of the input BOD, 92 percent of the initial
suspended solids concentration, and 99.99 percent of the initial
coliform (MPN) concentration are removed. Of course, the treatment
options at levels 3 and 4 account for the bulk of the treatment
efficiency.
E. RECEIVING MODEL REFINEMENTS
The Receiving Model has proven to be quite versatile in the applica-
bility to different problems, and has been used at the University
of Florida for purposes other than to model the effect of stormwater
runoff on a receiving water system. For example, it has been applied
to a lake that receives input from adjacent agricultural areas
(Perez et al., 1972; Perez, 1972) and to a marsh system (Heaney and
Huber, 1972). The latter application, in particular, prompted three
modifications.
Firstly, the spatial variation of rainfall can be of considerable
importance when modeling a large marsh system. Revisions were made
to read in records from as many as 50 rain gages. Rainfall from each
gage is then apportioned among the various junctions of the Receiving
Model network. For example, 40 percent of the rainfall for, say,
junction 1 might be assigned from gage 5, 35 percent from gage 6 and
25 percent from gage 7 in a hypothetical situation. The total rain-
fall at each junction can consist of contributions from one or more
point rainfall measurements. The percentages are determined by over-
laying the Thiessen polygon networks determined by the Receiving Model
junctions and by the rain gage array. Percentages are then assigned
on the basis of the amount of junction area contributed by Thiessen
polygons of the different rain gages.
72
-------
SUMMARY OF TREATMENT EFFECTIVENESS
INPUT
OVERFLOW (BYPASS)
TREATED
RF.f CVED
RELEASED
FLOW
LEVEL 4
LFVFIL 5
LEVFL 7
TRASH:
BAR RACKS
EFFLUENT SCREENS
FLOW(M.G. )
0.000
0.032
0.011
0.0
0.0
9.5
4202.6
7C60.7
0.0
479.0
1 90 .9
7940 .5
4721.8
0.0
0.0
25.449 CU.FT (AT SO LB/CU.FT.)
o.o CU.FT (AT so LD/CU.FT.)
* BAR RACKS
* SEDIMENTATION
« BIOLOGICAL TREAT
- NO EFFL. SCREENS
n CCNTACT TANK
REMOVAL PERCENTAGES FLOW (VOL) BOO (L8)
OF CVESALL INPUTS 0.29 53.23
OF TREATED FRACTIONS 1.03 SO.14
SS COLIF (MPN)
38,45 22.00
91*59 99.99
co
CONSUMPTIONS (LB )
LEVEL 3
LEVFL 4
LEVEL 7
TOTAL
CHLCRINE
o.o
o.o
261 .9
261 .9
POLYMERS
0.0
0.0
0.0
0.0
SEDIMENTATION
a IOLOGICAL TREAT
CONTACT TANK
REPRESENTATIVE VARIATION OF TREATMENT PERFORMANCE WITH TIME (OVERALL).
E 11:20 12:30 u:40 14:50 16:0 17: i o ja:20 19:30 20:40
WATER
AV. FLOW ( CFS>
,'•' BOD
S. SQL IDS
ARRIVING < MG/L )
RELEASED (MG/L)
CCLIFOPMS
ARR ( MON/IOOWL)
REL ( MPN/ IOOML)
4.06E 03 2.91E 04 2.40E 04 3.40E 04 2.2-3E 04
59.64
53. ee
127.33
1.59E 04
289 .25 720 .49
1 00.36 77.61
5.42E 03 3.66E 03
786.76 801. 19
83.98 85.34
92.42 92.47
3.95E 03 4.02E 03
99.95 99.99
5.56
eoe.32
85.82
85.49
•
665.83
92.49 :.,
4.04E 03
3.07E-0 I
95.99 'J
OUTFLOW BY GRAVITY (NG PUMPING) (LEVEL 6)
TREATMENT BY CHLCRINE CONTACT TANK (LEVEL 7)
N'J'-1f1EP_ Qc DCS I NG L'f "TS _= 2
DOSING RATE PER UNIT
2000.00 LE/OAY
2006.17 LE/OAY
13923. CU.FT, AT 15 MIN. DETENTION TIME
Table 20. TREATMENT MODEL OUTPUT.
-------
Secondly, areas with high vegetation, such as flood plains and marshes,
exhibit a variation of Manning's roughness, n, with flow depth. For
example, measurements made by the Corps of Engineers (1956) in a
sawgrass marsh in the Florida Everglades conservation areas may be
used to account for this effect. The program allows a user-specified
depth vs n variation.
Thirdly, overall hydraulic characteristics of marshes that contain
channels change when the channels overflow. The channels have low
roughness and permit high-flow rates but have negligible available
storage compared to the adjacent overland flow regime in the marsh.
However, the latter regime is subject to high roughness, typically
small slopes and low flow rates. Hence, the concept of parallel
channels between two junctions was introduced to allow one channel
to simulate the actual channel and a parallel channel to simulate
the overland flow on the marsh. Each channel can thus have differing
hydraulic characteristics while transporting water between the same
two points.
A modification that has resulted in increased flexibility in the use
of the Receiving Water block has been the incorporation of the O'Connor-
Dobbins formula to compute reaeration coefficients and incorporation
of a linear relationship to compute C>2 saturation concentrations as
a function of chloride concentration based on data given by Clark,
Viessman and Hammer, (1971). Included in the overall modification
is the ability to vary oxygen saturation values and/or reaeration
coefficients through the use of three options:
1) a constant value at all junctions may be
used (as before), supplied by the user;
2) a different constant value may be read in
for each junction;
3) the value may be computed at each junction
at each time step.
Clearly, chlorides must be modeled when the third option is used to
compute 02 saturations. This feature computes considerably more
realistic saturation values when simulating estuaries, as long as
the chloride distribution is adequately modeled.
The third option allows computation of the reaeration coefficient,
K2, at each junction at each time-step using the O'Connor-Dobbins (1958)
formula:
K2 ~
74
H
-------
where K2 = reaeration coefficient, sec
D = molecular diffusivity of 02 in water
(a function of temperature),
U = average velocity at a junction, ft/sec
H = junction depth, ft.
Use of this formula allows more accurate K values in areas of changing
flow conditions.
F. NETWORK AGGREGATION TECHNIQUE
In order to add the capability for modeling larger areas, the Combine
Block has been added to the Storm Water Management Model. This block
has two main objectives. The first is to collate two or more different
output data-sets from Runoff, Transport, Storage/Treatment, or any
combination thereof. This new data-set could then be used as input
into any block (Transport, Storage/Treatment or Receiving Water),
except Runoff. For example (Figure 13), an output data-set from
Transport area 'A' with manhole numbers 5, 6, 12 was collated with
an output data-set from Transport area 'B' with manhole numbers 1,
3, 6, 19. Manhole number 6 is common between both output data-sets,
therefore the hydrographs and pollutographs from both manholes are
added together. The new output data-set produced from the Combine
Block has manhole numbers 1, 3, 5, 6, 12, 19. This new data-set could
then be used as input to either the Transport, Storage/Treatment, or
Receiving Water Blocks. The second objective is to combine different
data-sets and manholes into a single data set with one manhole. For
example (Figure 14), an output data-set from Transport area 'Y' with
manhole number 23 is to be used as input into the Receiving Water
Block junction* number 14. The Combine portion of the Combine Block
would be used to combine the two output data-sets into one data-set
with one manhole. This manhole number would correspond to the
junction number of the Receiving Water Block.
G. MONITORING ROUTINE FOR DETERMINATION OF LARGEST SOURCES OF FLOW
AND POLLUTANTS
In order to determine the location and feasibility of various pollutant
source control alternatives and runoff storage devices, information
in the system should be analyzed. To aid in this analysis, a tabular
printout of the total BOD, suspended solids, and flow entering each
inlet is summarized for each SWMM run as shown in Table 21. Before
discussing the applications for these summary data, a brief descrip-
tion of the table will be given.
* Junction number and manhole number are synonymous.
75
-------
AREA 'A*
AREA V
FIGURE 13.
Hypothetical Drainage Network A.
FIGURE 14.
Hypothetical Drainage Network
76
-------
-D
I
!-
;-i
TABLE 21.
PRINTOUT FROM POLLUTANT MONITORING ROUTINE
BfSLttS CP BCLLUTANT »CMTCfIf.<: RCUTIOf
PCLLUTAME ASJPC1ATFC "ITf- "ASMCLf!
IKLET PV.KCFF C.».F. S.E
» •
• OA» BOO
S HVlCf C.«.F.
4 3
4 2
4 i
4 0
3 *
t «
2 2
? :
2 :
2 2
2 2
» ?
2 r
2 1
2 1
2 1
2 2
2 1
? |
2 1
2 1
2 1
1 0
2 1
> 0
2 C
2 C
J C
i C
i C
2 0
1 C
1 C
C
1 C
c
c
e
c
c
I 0
! C
0
0
c
c
c
.: I
.96
.< 9
.44
.79
• 1 6
.C4
.»*
. C 3
. 1 7
.69
.73
.C2
.It
.93
.47
.19
.33
.C4
.31
.42
• It
f « ^
. 14
.77
.93
.93
.te
.75
.46
.'0
.3«
.31
.67
.29
.70
.?S
. !4
.20
.1C
.0
.0
.0
.0 '
.0
.C
.0
re
?4
S2
7
2 2
?c
4C
2»
3c
34
2C
;c
1 F
:c
13
;c
7
2<
20
1C
s
7
73
•
13
7
S
5
7
13
2
«
7
•
S
?
7
7
2
C
C
0
0
0
c
c
c
.93
.01
.32
.e;
.E7
.77
.70
.12
.62
.01
.93
.93
.31
.93
.01
.93
. PS
. I F
.93
.46
.2]
.es
• S4
.23
.CF
.65
.23
.23
.FS
,*)•
.62
.23
,6E
.23
.23
.2 J
.* S
.es
.42.
.c
.0
.0
.0
.c
.0
,c
.0
IL6)
TCTAL CCO
24
If
5S
e
2F
14
5?
39
39
3F
23
3 2
3J
2C
22
15
22
10
77
21
11
t
9
24
f-
13
f
t
^
f
13
i
•
e
«
E
•
2
7
2
C
0
0
0
o
0
c
0
.1 4
.97
.01
.29
.C 3
.»?
.34
.49
.45
.81
.41
• AC
.33
.71
.01
.40
.03
.49
. 99
.77
.6!
.12
.07
.77
.**
.76
.17
.19
.E7 '
. S4
.22
.S»
.19
.90
.53
.S3
.67
.99
.02
.30
.0
.0
.0
.0
.0
.0
.0
MCMFICANCE or SUSPENDED SOLIDS.
TCTAL
ISFtC» (CFI
.IC3E
. 3* BE
.5SSE
.r ME
.321?
.45flE
.33IE
.4 CfiE
.547E
.•*£?£
.5--4E
. 3 1 * C
.61 IE
.2 J6E
• SS9E
.4C3E
.249E
-23?E
.SOJF
.I9IC
. 1 C*£
.I51E
.134E
.1 CIS
.7?3E
.95C-I!
.9SFE
.1 3CE
.2C2E
. 1 2OE
.6025
.339E
.USE
.29IE
.2«6E
.F?7F
.2566
.2C9E
• 392E
.616E
.73PE
.36«E
.0
.247E
.196E
. 3 37E
.3 !7C
01
02
02
01
02
02
02
02
02
02
02
0 2
C?
02
C2
02
02
02
02
02
02
02
02
02
Cl
Cl
Cl
02
C2
02
C2
01
01
02
02
01
Cl
01
01
01
01
00
00
cc
00
cc
00
1
j:
I'i
p
1-
1.
J-j
; i
.' ;
i'!
1 !
'.I
j!
• 'i
•jj
1
|-*
•- |
!-•'
j'
;-J
;
!•
|i-j
' t
t-'
«'
I-
j;
:".
•-.
;•
•. i
•-I
'••j
-•(
;-|
• ',
-------
The first column labeled Rank indicates that the inlet numbers in
column 2 are listed in order of magnitude of suspended solids entering
the system at that inlet. The first inlet listed collects the greatest
amount of solids and the last inlet the least amount, as shown under
the column labeled Total SS. Columns 3 through 5 give a breakdown of
the total pounds of suspended solids entering the system at each inlet by
surface runoff, dry weather flow (DWF), and pipe scour. The pipe scour is
from the adjacent upstream pipes. Columns 7 thru 8 list the pounds of
5-day BOD entering each inlet from runoff (including catch-basins) and
dry weather flow, respectively. The total BOD runoff and dry weather
flow is given for each inlet under the Total BOD column. The last
column labeled Total Inflow gives the volume of water entering each
inlet from both surface runoff and dry weather flow during the storm
event. The data listed in this printout, Table 21, represent a
decomposition of the total drainage system into its localized inputs.
All information pertaining to the time sequence of events has been
eliminated since the data represent run totals.
In attempting to locate a pollution control device upstream from the
drainage systems outlet, a decision between alternate locations will
probably be required. Upstream storage tanks will often be located
at points in the network where branches or inlets with dispropor-
tionately high flows enter the system. Similarly, any treatment alter-
nate such as sedimentation tanks, catchbasin cleaning, and sewer
flushing might be systematically applied at points in the drainage
network which have significant pollutant contributions.
Transferring some of the information from Table 21 to a map of the
Stevens Avenue drainage area, Figure 15 enables the analysis of the
spatial arrangement of pollutant inputs into the sewer network. The
ten most significant suspended solids inputs account for 94 percent of
the total solids with the four eroding sites contributing more than
90 percent of the total. This indicates the possible significance of
erosion-produced solids which can now be predicted by the SWMM (see
Section IV-B). Highest BOD loads enter inlets in the two western
branches of the network. These ten inlets account for 52 percent of
the total BOD. Pipes in which significant amounts of deposited solids
are scoured during the storm event are also indicated on Figure 15.
These ten pipes contribute 68 percent of the total scoured solids
but less than 1 percent of the total suspended solids. Where pipe
scour is found to be a significant portion of the first flush effect,
an analysis of this type can indicate a "best" operating policy for
sewer flushing.
Analyzing a drainage network utilizing the data from Table 21 is at
least a first step in evaluating the spatial distribution of flow
and pollutant inputs. Data arranged in this manner become infor-
mation which can be further investigated utilizing other modeling
options.
78
-------
O O CONDUIT USED IN STUDY
Q MANHOLE
DTOP 10 INLETS IN S3
CONTRIBUTION
O INLET WITH SIGNIFICANT
PIPE SCOUR UPSTREAM
TOP 10 INLETS IN BOD
'• CONTRIBUTION
TOP 10 PIPES IN SS CONTRIBUTION
DUE TO PIPE SCOUR
HYPOTHETICAL EROSION SITE
FIGURE 15.
Spatial Analysis of Pollutant Inputs
into the Stevens Avenue Drainage Network.
79
-------
SECTION V
TESTING OF THE SWMM IN LANCASTER, PENNSYLVANIA
A. INTRODUCTION
The City of Lancaster, Pennsylvania, population 79,500,2is situated
in a drainage area of about 8.24 square miles (21.34 km ). The
receiving stream in the Lancaster area is the Conestoga River which
drains an area of approximately 473 square miles (1225 km ) into the
Susquehanna River. The average flow is 387 cfs (11 m3/sec)
with a maximum recorded flow of 22,800 cfs (638 nP/sec).
There are two sewage treatment plants within the City, both of which
discharge into the Conestoga River. The North Plant with a capacity
of 10 mgd (3.80 x 10^ m3/day) serves a population of 36,000 people,
and the South Plant, recently expanded from 6 mgd (2.28 x 102* m^/day)
to 12 mgd (4.56 x 10^ m3/day) is designed to serve 69,000 people.
Both plants provide secondary treatment. About one third of the flow
to the North Plant is derived from areas with separate sewers outside
the City serving an estimated population of 17,500 people and some
industries. The remaining two thirds of the sewage flow to the North
Plant is derived from the combined sewers serving the north part of
the City plus about 250 suburban acres estimated to have 18,500 people
and many water-using industries. In addition, most of the year the
water table is high resulting in considerable infiltration. An over-
flow line diverts excess flow to the Conestoga during wet weather. The
North Plant drainage area is estimated at 3.72 square miles (9.63 km ).
The South Plant is designed to handle a population of 34,500 served by
combined sewers and, in addition, up to an approximately equal amount
from separated sewers throughout the surrounding area. The South
Plant drainage area encompasses 4.52 square miles (11.71 km^) and is
comprised of four districts. Stevens Avenue district which is the
subject of an EPA demonstration grant is one of the four districts
connected to the South Plant. Three of the districts, including Stevens
Avenue, pump the sewage from a receiving station within the district
to the South Plant. All locations have overflow arrangements that
discharge into the Conestoga River when the capacity of the system is
exceeded.
The total drainage area of the Stevens Avenue district is 227 acres
(92 fla)which, while only about 4.3 percent of the total Lancaster drain-
age area served by North and South treatment plants, is 17 percent of
the drainage area designed to flow into the South Plant from combined
sewers. The population within the Stevens Avenue district is estimated
80
-------
at 3,900. Figure 16 illustrates various drainage districts within
the City.
B. DEMONSTRATION GRANT DESCRIPTION
In order to remedy the situation resulting from combined sewer over-
flows, the City of Lancaster decided to explore means other than
sewer separation.
Construction of several underground silos at various locations within
the City is contemplated for retention of overflow during wet periods
and subsequent pumping to the treatment plants during low flow periods.
Stevens Avenue district was selected as the demonstration site for
evaluation of the effectiveness of a silo in combating combined sewer
overflows. The sewer layout for Stevens Avenue district is shown in
Figure 17. During normal dry weather periods, the dry weather flow
is pumped to the South Treatment Plant. During wet periods, when the
incoming flow to the pump station exceeds the capacity of the station,
the overflow discharges directly into the Conestoga River through a
60-inch (1520 mm) sewer located at point 6 on Figure 16.
The City of Lancaster also authorized APWA to develop design para-
meters for a full-scale swirl concentrator for removal of solids
prior to the retention of flow in the underground silo. Location of
the demonstration site is shown in Figure 17. A flow diagram of the
proposed swirl concentrator-silo treatment is presented in Figure 18.
In order to fully evaluate this treatment, the city decided to include
chlorination and microstraining as a part of this demonstration project.
The capacity of the silo is expected to be 160,000 ft^ (4480 m^).
The tasks assigned to the University of Florida were as follows:
1) conduct further verification and testing of
the Storm Water Management Model based on
active overflow measurements on selected
storm events and to make refinements to the
Model;
2) provide results of simulations to the APWA
in order for it to develop design criteria
and sizing of the swirl concentrator;
3) simulate the effect of the swirl concentrator-
underground silo treatment; and
4) simulate the effect of combined sewer overflow
from the entire city to the Conestoga River.
81
-------
NORTH DRAINAGE DISTRICT
00
N)
JUNCTION
I = NORTH TREATMENT
PLANT
6 = STEVENS AVE
DEMONSTRATION SITE
12 = SOUTH TREATMENT £
PLANT
SOUTH DRAINAGE
DISTRICT
' tr
L'.M|T> A ! rl-ri
STEVENS AVE.
DRAINAGE DISTRICT
SCALE IN FEET
0 2000 4000
I 1 j— 1 1 1
1000 3000 5000
±
Drainage
-------
E. MIFFUN SI
3) ©
STEVENS TRADE
SCHOOL
HAND J.H.S.
o (RAIN
GAGE)
SILO
DEMONSTRATION
PROJECT
CONDUIT USED IN STUDY
© SUBCATCHMENT MANHOLE
za) MANHOLE
— — — — CONDUIT NOT USED IN STUDY
STEVENS AVENUE DRAINAGE AREA
LANCASTER,PA.
SCALE IN FEET
0 500
I i |
1000
FIGURE 17, Sewer Layout of Stevens Avenue Drainage District,
83
-------
oo
i EXIST. 60
1 SEWER
BLOCK
CONTROL GATE
GRH
EXIST. 10 SAN-SEWER
EXIST. 6" SAN-SEWER
INTERCEPTOR
r
! 1
*|
WET
WELL
I i
*JPUMP |
I I
TO SOUTH
.TREATMENT
*PLANT
i I I J
7IEW "CURST. 1972 •, NOT
PART OF DEMO. PROJECT
\ OVERFLOW
/
T
DISINFECT
MIXING
AERA-
TION
DEVICE
I
SILO TANK
(EXIST. OUTFALL 60-IN.
o-
MAIN
PUMP
I
MICRO-
STRAINER
BACKWASH
•TO RECEIVING WATER
CONESTOGA RIVER
PRELIMINARY LANCASTER FLOW DIAGRAM
FIGURE 18. Flow of Treatment-Storage Options at Demonstration Site,
-------
C. DESCRIPTION OF THE STEVENS AVENUE RUNS
A total of three studies comprising eight storms were simulated. The
City and its engineers provided input data as well as one overall set of
measurements. The Stevens Avenue district was subdivided into 41
subcatchments. A description of each study and its results are given
below:
Study No. 1— The first study was based on a series of storms between
July 29 and August 3, 1971. The six-day period deposited a record
amount of precipitation throughout the Lancaster area (variously meas-
ured between 7.3 and 9.46 inches, or 18.54 cm and 24.03 cm respectively).
During four of the six days, the storms were very intense over short
periods; in one case, being the second heaviest of record. For pur-
poses of simulation, Study No. 1 was divided into six storms. Due to
the unavailability of field data for the study, verification of the
results of computer simulations for each of these storms was not
possible. These runs do indicate that an overflow as high as 400 cfs
(11.20 m /sec) may be expected for a storm event similar to Storm No. 6.
Results of this study were used by APWA in sizing the swirl concen-
trator. A design flow of this device was established at 165 cfs
(4.62 m3/sec).
Computer simulation studies were also conducted for all six storms
to evaluate the effect of the swirl concentrator-underground silo
facilities on the combined overflow quality- The results of Storm
No. 6 are shown on Figure 19. As illustrated, the quality of all
the overflow is significantly improved through the installation of
the swirl concentrator and a chlorine contact tank. Figure 19 also
shows that as soon as the swirl concentrator operates at full capacity
(165 cfs) , excess stormwater simply overflows into the silo. The
shaded portion of the flow versus time graph indicates the volume of
silo storage, 160,000 ft3 (4530 m3). Computed flow from transport
reflects the stormwater inflow to the storage facility (which is the
same as the computed flow to the river without the silo simulation).
The silo dewatering rate is limited by existing pumping capacity
of the Stevens Avenue station (3.57 cfs or 0.10 m3/sec). For simulation
purposes, pumping was set to begin at a silo depth of 3 feet (0.91 m).
Study No. 2— This study consisted of a storm that began in the
morning of August 27, 1971 and continued almost 30 hours to the morning
of the next day. It resulted in varying amounts of rainfall through-
out the city averaging more than 3.5 inches (8.9 cm). The results of
the computer simulation were similar to those obtained from Study No. 1,
and for this reason are not included herein. Again, no measurements
were taken during this study.
Study No. 3— This study is based on a relatively minor rainfall event
of March 22, 1972. This study is of special importance, however,
85
-------
in (fi.
UJ
3:00
,COMPUTED FLOW FROM TRANSPORT
OVERFLOW FROM SWIRL TO SILO
FLOW INTO SWIRL CONCENTRATOR
.COMPUTED FLOW TO RIVER, WITH SWIRL
8 CONTACT TANK ONLY
.COMPUTED FLOW OUT OF SILO TO
SOUTH PLANT
4:00
TIME, HOUR OF DAY
0.90 cf*
5:00
g
e>
I"
•CD CM
COMPUTED BOD FROM TRANSPORT
COMPUTED BOD TO RIVER, WITH,
SWIRLS CONTACT TANK ONLY
COMPUTED BOD OUT OF SILO
/ TO SOUTH PLANT
3:00
D.W.F = 363mg/l
4--00
TIME, HOUR OF DAY
500
3:00
COMPUTED SS FROM TRANSPORT
COMPUTED SS TO RIVER, WITH SWIRL B
' CONTACT TANK ONLY
^-COMPUTED SS OUT OF SILO TO
SOUTH PLANT
4:00
TIME,HOUR OF DAY
5=00
FIGURE 19.
Simulation for Stevens Avenue, Study 1, Storm 6.
High concentrations out of silo reflect initial
flush of DWF in sewer system into the silo.
86
-------
because it is one of the types most frequently experienced in terms
of intensity of rainfall (return period of less than one year). It
is also one for which relatively complete verification data such
as rainfall, flow readings and samples were collected. The rainfall
is shown in Figure 20 along with results of the computer simulation
showing overflow quantity and quality. Shown in the same illustra-
tion are the actual quantity and quality measurements (points labeled
x) of the overflow. It can be seen that agreement between the com-
puter simulation and the actual measurements of flow is only fair.
However, considerable doubt exists as to the validity of the flow data,
since they were obtained from depths measured with a yardstick inserted
into the mouth of the outfall. The agreement between the computed and
measured quality parameters is only fair also. It should be noted that
no calibration of the SWMM was attempted due to the lack of good
verification data.
Computer simulations were also conducted on this study to determine
the effect of the swirl concentrator-underground silo system. These
results are also shown in Figure 20. With the silo system, the Model
indicates no overflow into the Conestoga River.
D. RUNS IN THE NORTH AND SOUTH DISTRICT
Limited computer simulations were also conducted for the North and
South drainage districts. The North district was subdivided into
66 catchments and the South district into 104 catchments. The sewer
layouts for the North and South districts are shown in Figure 21 and
22.
Results of computer simulation for Study No. 3 for the North district
are presented in Figure 23 and for the South district in Figure 24.
The North district outfall is located at point 1 while the South
district outfall is located at point 12 as shown in Figure 16.
An examination of these figures shows that for a rainfall event
equivalent to Study No. 3, overflow from the North district would be
about 100 cfs (2.8 m3/sec) and from the South district, 160 cfs (4.48
m3/sec) . Figure 23 shows the quality concentrations of the storm water
beTEore it reaches the North treatment plant; and, subsequently, the
quality concentrations of the combined sewer overflow and treatment
plant effluent. Note that peaks in BOD and SS concentrations correspond
to peaks in computed flow from transport. The result is that, at the
two points of highest overflow, the water quality after treatment
remains almost the same as that computed from transport without treat-
ment. The dry weather flow (DWF) contribution to the Conestoga River
from the South district outfall is 7.65 cfs (0.21 m3/sec), generating
approximately 10,500 Ibs/day ( 4763 kg/day) of BOD and 7000 Ibs/day
(3175 kg/day) of SS. The DWF from the North district outfall is 9.50
cfs (0.27 m3/sec), generating approximately 8500 Ibs/day (3856 kg/day)
87
-------
c
>
H
55
Ul
2
K m.
*i2
O
COMPUTED FLOW
•~ROM TRANSPORT
COMPUTED FLOW OUT
OF SILO TO
SOUTH PLANT
COMPUTED FLOW TO
•RIVER, WITH SWIRL a
CONTACT TANK ONLY
11:20
12:20 120 2:20
TIME, HOUR OF DAY
3'20
a
o
o
O
z-
-COMPUTED BOD FROM TRANSPORT
-COMPUTED BOD OUT OF SILO TO SOUTHPLANT
COMPUTED BOD TO RIVER.WITH SWIRL
CONTACT TANK ONLY
:20
12*20 1:20 2:20
TIME, HOUR OF DAY
3:20
4^20
m
in
-MEASURED SS
-COMPUTED SS FROM TRANSPORT
COMPUTED SS TO RIVER.WITH SWIRL
a CONTACT TANK ONLY
COMPUTED SS OUT OF SILO TO
SOUTH PLANT
It20 12^20 1:20 2:20
TIME, HOUR OF DAY
3:20
4:20
FIGURE 20. Simulation for Stevens Avenue, Study 3. High.
concentrations out of silo reflect initial flush
of DWF in sewer system into the. silo. For this
storm, sil.o would store all runoff, but results
without it are also, shown.
88
-------
00
NORTH LANCASTER DRAINAGE DISTRICT
LANCASTER, PENNSYLVANIA
LESEND
• CONDUIT USED N STUDY
(J) MANHOLE
®
SUBCATCHMENT MANHOLE
FIGURE 21. North Drainage District with Runoff-Transport Numbering System.
-------
FIGURE 22.
South Drainage District with Runoff-Transport Numbering System.
90
-------
.c
>
t-
Z
_J — -
_l
<
u.
z
<
tr
RAINFALL
-COMPUTED FLOW
FROM TRANSPORT!
|:20 12:20 1:20 2:20 3:20
TIME, HOUR OF DAY
I— COMPUTED BOD FROM TRANSPORT
4:20
O
_ o -
n
Q
00
DO O .
ro
/ /COMPUTED BOD
FROM TREATMENT
11:20 12:20 1:20 2:20 3:20 4:20
TIME, HOUR OF DAY
O
O
CO
01 O
CO
CO
COMPUTED SS FROM TRANSPORT
/ /COMPUTED SS
FROM TREATMENT
11:20 12:20 |:20 2:20 3:20 4120
TIME, HOUR OF DAY
FIGURE 23. Simulation of North Drainage District, Study 3.
91
-------
ID-
•z.
UJ
2
o:
5
o
il! o
o
RAINFALL
COMPUTED FLOW INTO
SOUTH TREATMENT PLANT
COMPUTED FLOW
INTO NORTH
TREATMENT PLANT
°:20 12:20 1:20 2:20 3:20 4:20
o
o-
_ O
V O-
Q
O
m o
o-
TIME, HOUR OF DAY
COMPUTED BOD FROM NORTH TREATMENT PLANT ONLY
COMPUTED BOD FROM
\/SOUTH TREATMENT PLANT ONLY
1:20 12:20 1:20 2:20 3:20
TIME , HOUR OF DAY
4:20
01
6
O
CO
p. o
0 21
z
UJ
a
CO
o
COMPUTED SS FROM SOUTH TREATMENT PLANT ONLY
COMPUTED SS FROM NORTH
TREATMENT PLANT ONLY
K20 12:20 1:20 2:20 3:20 4:20
TIME , HOUR OF DAY
FIGURE 24. Simulation of North and South Drainage Districts
Study 3. '
92
-------
of BOD and 8500 Ibs/day (3860 kg/day) of SS. Figure 24 compares the
stormwater flow contribution and quality concentrations of the com-
bination of treatment plant effluent and bypassed combined sewer
flow for the North and South districts.
E. EFFECT ON RECEIVING WATER
To simulate the effect of the overflow on the Conestoga River, the
Receiving Water Model was run on the entire city for Study No. 3.
The manner in which various districts were combined is shown on
Figure 25. In conducting this run, the swirl concentrator was used
at Stevens Avenue while the Refined Storage and Treatment Model, as
described in Section IV-D was utilized to simulate the existing
biological treatment at the North and South plants. The silo was
deleted in order to have an overflow at Stevens Avenue outfall since
the installation of the silo prevents any overflow for a rainfall
event equivalent to Study No. 3.
The reaeration coefficient for the Conestoga River was computed using
the formula of O'Connor and Dobbins' (1958) (Equation 14). Results
of the Receiving Water Model are shown in Figures 26 through 28.
Figure 26 shows DO profiles 3 and 24 hours after the storm inception.
A slight DO sag is noticeable 3 hours after the storm, with minimum
concentrations registered at the Stevens Avenue outfall (junction 6).
The effect of the South treatment plant effluent and overflow (junction
12) is noticeable as a slight drop in DO concentration. The profiles
for 48 hours and 72 hours after the storm are identical to the 24-hour
profile, and were not plotted. This was true for all constituents,
and it can be generally stated that the impact of the storm upon the
Conestoga River is attenuated after 24 hours. Figure 27 shows the
BOD profiles along the Conestoga River 3 hours and 24 hours after
the storm, and Figure 28 the SS profiles. The effects of reaeration
and stream self-purification are evident from examination of the
pollutant input junctions (1, 6, 11) and downstream conditions.
Initial values used to simulate the Receiving Water Model are listed
in Table 22. The reader should note that suspended solids computa-
tions do not include a contribution from erosion sites due to a lack
of data, and thus the suspended solids concentrations are somewhat low.
F. SUMMARY
Testing of the SWMM in Lancaster, Pennsylvania has revealed the im-
portance of having sufficient and accurate measured data for model
calibration. It is recommended that the SWMM user begin his sumula-
tion from a coarse analysis of the study area. Subsequent refinement
can be tailored with his knowledge of the more sensitive parameters
(impervious area, detention storage, catchbasin BOD content, infiltration,
93
-------
STEVENS
AVE.
SOUTH
B
RUNOFF ft
SOUTH
A
RUNOFF a
RUNOFF a
TRANSPORT
RUNOFF a
TRANSPORT
TRANSPORT
STEVENS
AVE.
TREATMENT
COMBINE
SOUTH
COMBINE
SOUTH a
STEVENS
SILO
SOUTH
TREATMENT
PLANT
NORTH
TREATMENT
PLANT
COLLATE
SOUTH, NORTH
a STEVENS
OVERFLOW
RECEIVING WATER
FIGURE 25.
Combination of SWMM Runs for Overall Lancaster Simulation.
94
-------
JUNCTION
12 3 45678 9 10 II 12 13 14 15
16 17 18 19 20
24 HOURS AFTER STOR
-co
o>
E
o
o
in
-3 HOURS AFTER STORM
10 20
30 40 50 60
DISTANCE IN THOUSAND FEET
70 80
FIGURE 26.
Dissolved Oxygen Profiles Along the Conestoga River. Study 3.
-------
45678
JUNCTION
10 II 12 13 14 15
16
17
18
19
20
-3 HOURS AFTER STORM
24 HOURS AFTER STORMi
10
20
30 40 50 60
DISTANCE IN THOUSAND FEET
80
FIGURE 27.
BOD Profiles Along the Conestoga River. Study 3. During the
storm reported values in the river varied between 3 and 18 mg/1.
-------
123 4 567 8 9
JUNCTION
10 II 12 13 14 15
16 17 18 19 20
in-
CO
CO
10-
'3 HOURS AFTER STORM
24 HOURS AFTER STORM
10 20
30 40 50 60
DISTANCE IN THOUSAND FEET
70 80
FIGURE 28.
Suspended Solids Profiles Along the Conestoga River. Study 3. Reported
values in the river during the storm ranged from 35 to 155 mg/1.
-------
Table 22. PARAMETERS USED FOR SIMULATING RECEIVING WATER MODEL
Initial and Saturation DO in Conestoga River
(all junctions)
BOD in Conestoga River (all junctions)
Initial Suspended Solids in Conestoga River
(all junctions)
Decay Coefficient (BOD)
Reaeration Coefficient
Flow in Conestoga River (entering junction 1)
BOD in Conestoga River (entering junction 1)
SS in Conestoga River (entering junction 1)
10.0 mg/1
5.0 mg/1
10.0 mg/1
0.20/day
1.50/day
700 cfs
0.0
0.0
98
-------
and others) and available field measurements. The model is not
fully calibrated until satisfactory results have been obtained with
more than one storm within the study area. The application to Lan-
caster suffered from lack of good measured data for several of the
storms studied, but it can be generally stated that for the Study
No. 3 storm:
1) The SWMM was able to predict only fairly the
quantity of the combined overflow for the Stevens
Avenue district; however, the only set of measured
flows are questionable.
2) Computed suspended solids were universally lower
than measured values at the overflow and in the
receiving water (Conestoga River) probably because
erosion was not modeled and appropriate upstream
concentrations were unknown.
3) Computed BOD values were higher at the overflow
than the measured values, but values predicted in the
Conestoga River three hours after the storm were
within the range of reported values.
4) The installation of the swirl concentrator and silo
complex will result in substantial improvement in
the quality of the overflow at Stevens Avenue, pro-
vided the full-scale performance of the swirl
concentrator is comparable to the results obtained
in laboratory studies by APWA. This fact should
be true for any storm comparable to those described
earlier in this report.
99
-------
SECTION VI
DECISION-MAKING' FOR WATER
QUANTITY AND QUALITY CONTROL
This section presents the results of the analysis into alternative
procedures for selecting the "optimal" overall stormwater control
strategy. The word "optimal" is used in a mathematical sense to
denote the alternative which is best for the problem as formulated.
Much use has been made of such models during the past ten years.
We have been actively involved in such developments. For example,
Heaney (1968) developed a mathematical programming' model to allocate
water supply among competing users in the. Colorado River Basin.
Pyatt r.i: at., (1969) and others have set up optimization models to
determine the least cost combination of sewage treatment and low
flow augmentation.
More recently, Kirshen, Marks, and Schaake (1972) have devised a
mathematical model for analyzing stormwater quality control alter-
natives. A linear programming model is used to determine the sizes
and operating policies for sewer pipes, storage reservoirs, and
treatment plants. The SWWM is used to generate boundary conditions-
for the optimization model and to provide more detailed information
on performance of the system.
The above optimization models assume some specified level of
performance for the single-purpose system under consideration.
Unfortunately, from a computational point of view, many of the storm-
water quality control alternatives serve other functions such as
drainage and flood control. Thus, the urban water quality problem
seems to fit well into a more general framework of urban water resources
management. An overall strategy for 'examining this more general
problem has been developed. Addressing the problem in this broader
perspective, the determination of the overall optimal solution can be
achieved using one. of several suggested procedures. A more innovative
aspect of the. analysis examines the questions of cost sharing among
the various .study areas within the urban area.
Tf one examines the various components of urban water resources, he
is made aware of how each problem has been treated independently.
Traditional concern for safe disposal of waste-water focuses on design
of treatment units. There has been little systematic effort to devise
overall strategics for drainage and flood control. The studies in
100
-------
Denver are a notable exception and they shall be discussed later.
The question of urban water supply has been dealt with as a separate
problem. Likewise, the related problems of recreation and open space
have been analyzed separately from the other problems. As a result,
it is difficult to gain insight as to how they all tie together.
Having analyzed these other problems, one begins to gain enthusiasm
that perhaps "there is a better way." More enthusiasm was gained
after participating in the layout of an 800-acre (3.24 Km^) planned
urban development which has an overall density of four units/acre
and was able to satisfy the following constraints:
1) water use ^_ safe recharge of the aquifer;
2) on-site control of storm water from a 50-year
storm;
3) no development of structures in the flood-
plain;
4) no direct discharge of storm water into
on-site recreational lakes; and
5) disposal of treated sewage effluent onto golf
course.
This area has about 86 percent of the total land in open space. The
development is for middle income groups and comprises a mix of
detached dwelling units, townhouses, apartments, and condominiums.
It appears that the development costs are competitive or lower than
existing practices.
The following sections present the methodology used in addressing the
problem. It is partitioned into the several components listed below:
A) description of the study area;
B) selection of commodities to be analyzed;
C) inventory of sources;
D) inventory of controls;
E) economic analysis of control options;
101
-------
F) specification of performance criteria;
G) statement of objectives; and
H) system optimization and examination of equity
question.
A. DESCRIPTION OF THE STUDY AREA
For the purposes of this analysis, assume the entire area under
consideration is being partitioned into N planning units. Any subset
of the N planning units will be denoted as S where ScN. In urban
areas, this planning unit might be a neighborhood or a subdivision,
a distinct political entity such as a ward of a city, or any other
partitioning of the city which seems to be appropriate.
B. SELECTION OF COMMODITIES TO BE ANALYZED
The Storm Water Management Model (SWMM) is capable of routing water
and specified conservative and nonconservative pollutants. Concep-
tually, it is possible to set up a decision model which simultaneously
determines the optimal solution for multiple commodities flowing along
the same or slightly modified network (Heaney, 1968). Such problems
are classified as multi-commodity network flow problems. Unfortunately,
multi-commodity models usually result in relatively large models since
the number of constraints for a single commodity model is multiplied
by m where m denotes the number of commodities.
It is preferable to analyze only a single commodity and incorporate
consideration of other commodities implicitly. The general idea is
to determine which commodity flow will be the "limiting factor" and
to specify the problem accordingly. For example, a stormwater reten-
tion basin can be sized not only by hydraulic constraints but also
using water quality constraints, e.g.3 allowable suspended solids
concentration in the effluent. In any given case, only one of the
constraints will be binding. Thus, it is possible to incorporate
water quality constraints implicitly by making appropriate adjustments
in the cost functions and constraints. This is the approach we shall
employ using water as the single commodity.
C. INVENTORY OF SOURCES
Early studies on the water quality aspects of the combined sewer
overflows and stormwater discharges indicate that this water may have
102
-------
a significant pollution load. Often a "first flush" phenomenon
was observed, i.e. a high initial pollutant concentration and
subsequent rapid decrease in concentration. Usually it was assumed
that the pollution emanated from material that had settled out in
the sewers during antecedent dry weather, from catchbasin sumps,
from street gutters, and/or from rooftops, parking lots, and
related areas.
Studies in Chicago provided an estimate of the type of materials
found in urban areas (Heaney and Sullivan, 1971). The results are
shown in Table 23. The identification of the potential magnitude of
the stormwater quality problem inspired subsequent efforts to deter-
mine the origin of these pollutants. A summary of these efforts is
presented below.
1. Air Pollution
Air pollutants from point and non-point sources are a significant
component of the stormwater quality problem. Precipitation in urban
areas acts as a "scrubbing" device to cleanse the atmosphere and to
flush away air pollutants which have settled in the urban area. From
an environmental control point of view, it is important to understand
the relative merits of source control of air pollution Vs• alternative
control of stormwater pollution.
Fortunately, parallel work has been done in developing air quality
management models. The model which is frequently used for this
purpose is called the Air Quality Display Model (AQDM). Like the SWMM
it is capable of estimating the transport of pollutants (particulates
and S02), the cost of control, and the impact of alternative control
strategies. This model is available from EPA (TRW, 1969). The Univer-
sity of Florida has utilized this model in studies of air pollution
control in Jacksonville, Florida, and found it to be a useful tool
(Wilson, 1971). Thus, air pollution levels can be estimated using
such approaches. Air pollution from specific industries can be
estimated using information from EPA sponsored studies. For example,
Heaney and Doughty (1971) developed a model for determining the
optimal control strategy for emissions from any type of paper and pulp
mill.
The above air pollution models have not been incorporated directly
into the SWMM since this analysis can be easily done as separate
special studies of sources of stormwater pollution.
103
-------
Table 23. ESTIMATED MONTHLY QUANTITY OF IDENTIFIABLE SOLIDS
FROM 10-ACRE URBAN AREA, CHICAGO, ILLINOIS
(Quantity of Solids Sources: tons/month)
(Heaney and Sullivan, 1971)
Exterior Solids
Refuse
Month
January
February
March
April
May
June
July
August
September
October
November
December
TOTAL3
tons/yr.
Air
pollution
0.68
0.78
0.80
0.60
0.72
0.40
0.46
0.30
0.45
0.56
0.61
0.66
7.00
Domestic
wastes
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
2.4
28.8
Garbage
1.8
1.8
2.3
2.3
2.3
2.8
2.8
2.8
2.3
2.3
2.3
1.8
27.6
Other
7.2
6.8
7.9
8.8
8.5
9.2
8.7
9.3
8.9
8.3
8.4
8.2
100.2
Street
sweepings
0.68
0.68
0.76
0.76
0.76
0.76
0.76
0.76
0.76
1.51
1.51
0.68
10.40
Catch
basins
.43
.43
.43
.43
.43
.43
.43
.43
.43
.43
.43
.43
5.20
Totalsa
tons /mo .
13.2
12.9
14.6
15.3
15.1
16.0
15.6
16.0
15.3
15.5
15.6
14.2
179.3
-------
2. -Sedimentation and Eros-ion
It became evident early in the analysis that it was essential to
better characterize the nature of the settleable and suspended
solids in stormwater runoff. That portion of the solids load
attributable to air pollution can be estimated using the AQDM.
Field observations in Lancaster, Pennsylvania? Gainesville, Florida,
and elsewhere indicate that the predominant materials in the resi-
dential areas result from localized soil erosion due to construction,
parking, etc. Thus it is possible to characterize the nature of the
solids as a function of the soils within the study area. This infor-
mation is extremely helpful in designing control facilities such as
settling basins.
Thus, a method for estimating erosion has been included in the SWMM.
This approach uses the "Universal Soil Loss Equation" which was
originally developed for agricultural areas. A detailed description
of this phase of the investigation is presented in Section IV-B.
3. Yard Litter
In addition to erosion, yard litter appears to comprise an important
fraction of the potential pollutants in stormwater. The quantity of
vegetation entering the sewer system depends primarily on local con-
ditions and practices. For example, yard waste in Gainesville, Florida
is piled in the street gutter until the next collection. If a storm
occurs during the interim, then some of this material may be flushed
into the sewers. Thus, a separate study was made of this problem.
The results, presented in Section IV-B, indicate the importance of
this component as a major source of organic material.
4. Other Sources
In addition to the above specific studies, other efforts have examined
the specific impact of highway de-icing chemicals (EPA, 1971) while
numerous efforts have been made to characterize the general magnitude
of the problem (e.g.,, AVCO, 1970). As these studies are completed,
our understanding of the origin of this material will improve accordingly.
5. Quantity of Urban Stormwater
From a hydrologic point of view, urbanization with storm sewers causes
significant increases of peak flows in the receiving waters. This is
a prime concern from the point of view of downstream flooding, in-
stream erosion, and undesirable effects on the receiving water itself.
105
-------
These effects are becoming more Important as the perceived function of
these receiving waters changes from their traditional use as drainage
canals to a growing awareness of their value as an environmental asset.
In general, urbanization causes a shift in the original hydrograph
as shown in Figure 29. Estimates of the increase in peak discharge
due to urbanization are shown in Figure 30.
Engineering design is usually based on providing a facility which can
handle some "worse case" condition. In this case, one might attempt
to size the control facilities to handle the peak flow. However, this
approach is quite expensive since the peak is so high and is of very
short duration. Because of the nature of the problem, it is logical
to place heavy emphasis on incorporating some kind of storage facilities
into the design.
Storm sewer designers typically use the Rational Method to estimate the
peak rate of flow. More sophisticated methods which incorporate routing flow
procedures are also available. One might alternatively or conjunctively
try to estimate the total volume of runoff from a given storm. As
development increases, the volume of runoff increases. Thus, the given
storm produces not only a higher peak rate of runoff, but also a larger
total volume of runoff due to the reduced opportunity for soil infiltration.
Urban hydrologists have devised techniques for estimating the total
volume of runoff which are quite useful (see Chapter 20 of Chow, 1964,
for example). The Soil Conservation Service (1972) has developed very
comprehensive procedures for analyzing the hydrology of agricultural
areas. Their procedure for estimating runoff is especially useful for
examining urbanizing watersheds. Also related manuals have been
developed by SCS for design of control facilities such as grassed
waterways, storage ponds, etc. This work has proven invaluable in
determining the magnitude of the sources of stormwater.
This section has attempted to characterize the sources of stormwater
pollution and to briefly describe the general impact of urbanization
on the rainfall-runoff relationship. With regard to the hydrologic
aspect, one can see that an ever-increasing rate of runoff and volume
of runoff accompany urbanization. Thus, controls need to be designed
for more stringent storage and treatment criteria. Studies of the
origin of pollutants in stormwater indicate that the nature of the
pollutants could be expected to change as an area undergoes urbaniza-
tion. In urban areas with a relatively large amount of pervious area,
one would expect that soil erosion and yard litter would comprise the
106
-------
Lag time
after urbanization
TIME, (hrs.)
Log time
Hydrograph
of streamflow
Center of mass
of runoff
and of rainfall
TIME, (hrs.)
FIGURE 29.
Effect of Urbanization on Storm Water Hydrograph
(Leopold, 1968)
107
-------
100
0
20 40 60 80
PERCENTAGE OF AREA IMPERVIOUS
100
FIGURE 30. Effect of Urbanization on Peak Discharges
(Leopold, 1968)
108
-------
most important sources of pollution. Correspondingly, as the urban
area is rendered impervious by further development, these sources
would be reduced and pollution emanating more from de-icing chemicals,
air pollution, grease, oil, etc., which typify highly urbanized areas,
would increase in relative importance.
D. INVENTORY OF CONTROLS
!• Water Quality Controls
It is evident, from the inventory of sources that a wide variety of
controls might be employed to reduce problems associated with specific
pollutants and increased flows. An excellent summary of alternative
water quality control devices has been presented by Field and Struzeski
(1972). A summary of the water quality controls which they discuss is
presented in Table 24. Many of these control options have been incor-
porated directly into the SWMM.
2. Water Quantity Controls
A thorough review has been made of urban storm drainage practices in
order to better understand the rationale behind present system design.
Storm sewer design is a standard procedure which engineers have rou-
tinely performed for a number of years. Referring to Figure 30, one
can see that the increase in peak flow is a function not only of the
percent impervious area but also the percent of the area served by
storm sewers. There is not necessarily a one-to-one correspondence.
As can be seen from Figure 30, it is possible to significantly reduce
the peak flows if one could discourage the use of storm sewers and/or
encourage the discharge onto pervious areas.
Subsequent subsections will examine current design criteria for streets,
parking, and roofs in urban areas. The purpose of these sections is
to examine the feasibility of modifying their design to help ameliorate
urban stormwater problems.
According to an unpublished Rand study of land use In 48 U.S. cities,
the land use proportions are as follows (Urban Land Institute, 1968):
Type of Use % of Total Land Vse
Residential 39.0
Industrial 10.9
109
-------
Table 24. CONTROL OF COMBINED SEWER OVERFLOWS
(Field and Struzeski, 1972)
A. Existing System Control
1. Maximizing treatment
2. Improved regulator maintenance
3. Infiltration and extraneous inflow control
4. Surface and sewer system "housekeeping"
5. Polymers to increase flow capacity
B. Advanced Control Systems
1. Flow regulation
2. Storage
3. Porous pavement
4. New sewer systems
C. Treatment Methods
1. Fine-mesh screening and microscreening
2. Dissolved-air flotation
3. Rotating biological contactors
4. High-rate plastic media trickling filters
5. High-rate, single-, and dual-media filtration
6. Vortex, swirl, and helical separators
7. Advanced disinfection methods, high-rate
application, on-site generation, automated
operation, ozonation, and use of combined
halogens (chlorine and iodine) and chlorine
dioxide
8. Tube settlers
9. Powdered and granular activated carbon adsorption
10. Polymer and other chemical additives for improved
settling, microscreening, filtration, and flotation
11. Chemical oxidation
12. In-line or in-sewer treatment
13. Sludge handling and treatment
14. Regeneration of carbon and coagulants
15. Reclamation and reuse
110
-------
Commercial 4.8
Road and Highway 25.7
Other Public 19.6
10Q.O
For Lancaster, Pennsylvania, the study area for this report, the
impact of 1000 new inhabitants has been estimated (Urban Land Insti-
tute, 1968). The 1000 inhabitants would comprise about 270 families
with 500 adults, 200 pre-school children, and 300 school children.
The land requirements for these new residents were estimated as shown
below:
Type of Use Acreage
Residential 90
Street 13
Public Land 20
Service 3
Retail Stores 2
128
Both of the above tables indicate that residential areas and streets
are the predominant types of land use and consequently have the
greatest impact on urban stormwater problems. Some of the newer
concepts in urban planning such as planned unit development (PUD)
and density zoning control offer hope of Improved storm drainage
practice. For example, a new PUD north of Tampa, Florida handles
all stormwater on-site and maintains over 80 percent of the land in
open space. Comparisons of density zoning control with standard sub-
division layouts are shown in the Community Builders Handbook
(Urban Land Institute, 1968). These newer land use forms appear to
offer the long-run solution to current stormwater quality management
problems.
The following subsections present a summary of current practices in
design of streets and parking areas. Also, a brief description of
rooftops is included. These sections provide a first glance at the
feasibility of stormwater control by modifying current design practices.
Ill
-------
Parking
The research into parking has provided an interesting example of how
a major urban land use category has evolved with little consideration
for its overall impact. A recent oublication by the Highway Research
Board (HRB) (1971), titled Parking Principles, provides a good summary of
the state of the art. One frequently suggested method for controlling
stormwater problems is to temporarily pond water on parking lots and
perhaps on streets. This section discusses the general present
criteria underlying design of off—street .and on—street parking facil-
ities.
A primary recommendation of the HRB report is that on-street parking
be discouraged because of the traffic hazard. Certainly there has
been a strong movement during the past 20 years towards that goal.
In residential areas, on-street parking is permitted with the following
constraints according to the 1968 edition of the Uniform Vehicle Code
as reported in the HRB report. Parking shall be prohibited in the
following locations:
1) on a sidewalk;
2) in front of a driveway;
3) within an intersection;
4) within 15 feet (4.5m) of a fire hydrant;
5) on a crosswalk;
6) within 20 feet (6.1m) of a crosswalk at an inter-
section;
7) within 30 feet (9.1m) on the approach to any
flashing beacon, stop sign, or traffic-control
signal located at the side of a roadway;
8) between a safety zone and the adjacent curb
or within 30 feet (9.1m) of points on the curb
inmediately opposite the ends of a safety zone,
unless the traffic authority indicates a
different length by signs or markings;
9) within 50 feet (15.2m) of the nearest rail of
a railroad crossing;
112
-------
10) within 20 feet (6.1m) of the driveway
^entrance to any fire station within 75
feet (23m) of said entrance (when properly
posted);
11) alongside or opposite any street excavation
or obstruction when stopping, standing, or
parking would obstruct traffic;
12) on the roadway side of any vehicle stopped or
parked at the edge or curb of a street; and
13) on any bridge or other elevated structure on a
highway or within a highway tunnel.
Let us examine street layout such as exists in Chicago wherein a
typical block is 660 feet x 330 feet (201m x 101m). A section of
this street is shown in Figure 31. Note that the section does not
have driveways and cars may be parked either on the street or in
the garage which is accessible via the alley.
For the land use shown in Figure 31, there is about 1400 feet (430 m)
of parking available on the street. The required length of parking
space is about 22 feet (6.7m) for a full-size car. Thus, there are
about 64 spaces.
This section of Chicago is typical of residential land use in many
older cities in the United States. This land use pattern can be
compared with lower densities by reducing its density. The results
of this exercise are shown in Table 25. As the density is reduced,
it is assumed that alleys are replaced by driveways and the roof area/
house increases. The percent imperviousness of the original section
is 36 percent. It is interesting to note that it is still 22 percent
at a low density of 1.2 dwelling units/acre (0.5 units/ha).
The last row of the table illustrates vividly the increase in paved
area/dwelling unit as density increases. This reflects a significant
reduction in the utilization of available parking and traffic lanes.
As urban residents make the transition to using driveways, two related
factors reduce the effective utilization of paved area: 1) the
increased paved area for the driveway; and 2) the reduction in on-
street parking spaces due to restrictions on parking in front of or
near driveways. Thus, changing land use patterns have increased
significantly the paved area/dwelling unit and reduced the effective
utilization of the paved area.
113
-------
II I I -L' ' ^
I I I r
-------
Table 25. EFFECT OF HOUSING DENSITY ON LAND USE PATTERNS
1)
2)
3)
4)
5)
Item
Land use
a) Total paved area
i> Front street
ii) Side street
ill) Alley
iv) Driveway
b) Total roof area
c) Pervious area
i) Front and
rear lawn
ii) Side lawn
d) Total of a, b and c
Per cent impervious
Per cent to sewer
system
Dwelling units
Typical lot size
Housing density in units per acre
14.40
.97
.50
.24
.23
.00
.83
3.20
3.00
.20
5.00
36
36
72
(33x124)
7.20
.97
.50
.24
.23
.00
.83
3.20
3.00
.20
5.00
36
36
36
(33x124)
3.60
.99
.50
.24
.00
.25
.62
3.39
3.19
.20
5.00
32
15-32
18
(66x132)
2.00
.88
.50
.24
.00
.14
.46
3.66
3,46
.20
5.00
27
15-27
10
(119x132)
1.20
.82
.50
.24
.00
.08
.34
3.84
3.64
.20
5.00
23
15-23
6
(198x132)
6) Impervious area/
dwelling unit
.025
.05
.089
.134
.227
-------
Off-street parking has been encouraged not only for residential
areas but also for all other types of land use. The HRB report
recommends the number of parking spaces for various types of land
use (see Table 26). For example, parking for shopping centers is
estimated to require 5.5 spaces/1000 square feet (59 spaces/lOOOm2)
of gross leased area. For a one-story center, this amounts to about
2 square feet of parking area/square foot of roof area. According
to the Community Builders Handbook (1968), this design criteria
accommodates the need for parking spaces for all but the 10 highest
hours of demand during the entire year. These peak hours occur
during three days of the year. Including cross aisles and entrances,
about 350 square feet (33 m) of paved area/parking space is needed to
accommodate full-sized American cars. Smaller spaces are required
for the more compact vehicles. Table 27 shows recommended stall
dimensions for various cars.
It is apparent from Table 27 that full-size American cars require an
additional 50 percent paved parking area/vehicle. Parking facilities
are usually designed for this worst case condition. A summary of the
dimensions of 1973 cars produced by Ford Motor Company provides a
comparison of vehicle dimensions.
Table 28 shows that the larger American cars would require a much
larger parking area and a better paving material to handle their
heavier weight (about twice the weight of a sub—compact).
The above analysis of parking in urban area illustrates the impact of
the automobile on urban land use. The desire for off-street parking
and larger vehicles has necessitated a significant increase in per
capita parking demand.
From the water quality control point of view, each square foot of
paved area increases stormwater problems so that it would be helpful
to discourage unnecessary construction of such facilities. It is
also desirable to encourage temporary on-site detention of stormwater
in parking lots. If one surveys parking facilities in newer urban
areas, he observes a relatively low overall rate of utilization. For
example, a church might have off-street parking which is actually used
only a few hours a week. It is considered good practice for a com-
mercial establishment to provide enough spaces to accommodate the
peak load of cars.
Residential areas with driveways require a significant increase in
paved areas which is not utilized too often. A value judgment is
required to evaluate whether such practices are wasteful.
116
-------
Table 26. ZONING STANDARD GUIDELINES FOR PARKING
(Highway Research Board, 1971)
Use of building site
Minimum number of
parking spaces required
Residential
Single family
Multifamilya
Efficiency
One or two bedrooms
Three or more bedrooms
Commercial
Offices3 and banks
General retail3
Shopping centers
Restaurants3
Hotels, motels3
Industrial3
Auditoriums and theatersa
Churches
College/university
Good transit access
Auto access only
Senior high school
Elementary and junior high school
Hospitals
2.0 per dwelling unit
1.0 per dwelling unit
1.5 per dwelling unit
2.0 per dwelling unit
3.3 per 1,000 sq ft GFAb
4.0 per 1,000 sq ft GFAb
4.4 per 1,000 sq ft GLAC
0.3 per seat
1.0 per rentable room plus
0.5 per employee
0.6 per employee
0.3 per seat
0.3 per seat
0.2 per student
0.5 per student
0.2 per student plus
1.0 per staff member
1.0 per classroom
1.2 per bed
a Exceptions permitted in Central Business District if adequate public
transportation is available.
b GFA = Gross Floor Area
c GLA = Gross Leased Area
117
-------
00
Table 27. RECOMMENDED PARKING STALL DIMENSIONS FOR
VARIOUS TYPES OF STALLS FOR 90° PARKING
(Highway Research Board, 1971)
Type of car
Small compact3
Larger American
compact^
Small intermediate0
Full size
Stall
width
(ft)
7.5
8.0
8.0
9.5
Stall
depth
(ft)
15.0
16.0
16.5
18.5
Aisle
width
(ft)
20.0
22.0
22.0
25.0
Wall to
Vail distance
(ft)
50.0
54.0
55.0
62.0
Paved area
per vehicle
(ft2)
187.5
216.0
242.0
294.0
a Volkswagen and Maverick, for example
° Valiant, for example
c Dart, for example
-------
Table 28. DIMENSIONS OF 1973 VEHICLES PRODUCED
BY FORD MOTOR COMPANY
Vehicle
Length
(ft)
Width
(ft)
Area/car
(ft2)
Weight
(Ibs)
1) Full-size Ford
2) Full-size Mercury
3) Full-size Lincoln
4) Intermediate size
Torino
5) Compact-size
Maverick
6) Sub-compact Pinto
18.3
18.7
19.2
17.3
15.5
13.9
6.6
6.6
6.6
6.6
5.9
5.8
121
124
127
114
92
81
4,350
4,500
5,200
3,800
2,800
2,240
119
-------
Recent studies have analyzed the feasibility of using porous pave-
ment. Included in this analysis is an estimate of the amount of
area which was paved during 1971. The results are shown in Table 29.
Note that residential streets are the most important single source
for urban areas. If much of this acreage could be paved with porous
pavement, a tremendous benefit as far as stormwater quantity and
quality control would result. Based on the low rate of utilization
of many of these facilities at present, it appears that such modi-
fications might be feasible.
According to the HRB report on parking, the following design criteria
are recommended:
1) lots parking five or more vehicles shall be
paved with a dustproof all—weather surface;
2) the lot shall be graded so that it will drain
as required by the city engineers; and
3) for adequate drainage, a minimum grade of
one percent for asphalt or concrete is desirable.
It appears that a reassessment is needed of the tendency to pave the
undeveloped portion of commercial and industrial developments. This
trend is the result not only of the ever-increasing need for parking
but also the desire for reduced on-site maintenance costs. Thus, many
service facilities are designed with large amounts of excess capacity
for parking. The increased convenience for motorists and owners of
such establishments is further aggravating off-site disposal problems.
If the appropriate groups could be made aware of these problems, then
it would be possible to reduce stormwater management problems and
increase the amount of pervious open space in urban areas.
A major transition in attitudes is needed to encourage a re-evaluation
of current parking practices. We have become accustomed to nuisance-
free parking as one of the positive indicators of quality of life.
Unfortunately, such conveniences have caused a significant increase in
the intensity of stormwater problems as cities have been paved over
in response to this desire for better parking.
Streets
Streets are a major source of sewered runoff from urban areas. Rea-
sons for not permitting drainage water to accumulate on streets are
120
-------
Table 29. PAVING IN U. S.
(Franklin Institute, 1972)
Type of pavement
Surface area paved during
1971 (106 sq. yds.)
A. Parking lots
Shopping centers
Industry
Office buildings
B. Residential streets
Low design (cities 21 50,000)
Low design (cities < 50,000)
High design
C. Business streets
Suburban
City
D. County road
Low-volume
High-volume
E. Highway
Two-lane
Four—lane
F. Miscellaneous
46.0
25.0
11.6
9.4
87.5
45.9
12.1
52.5
4.8
193.6
203.9
142.0
104.0
TOTAL
145.5
57.3
397.5
246.0
75.2
967.5
121
-------
numerous, e.g.3 avoid splashing pedestrians, interference with
traffic flow, and so on.
The previous section mentioned that provision of adequate off-street
parking is now considered accepted practice in urban areas. Since
this is the case, one would expect a re-examination of street design
principles in light of these changes. It would seem reasonable to
presume that smaller streets might be used if adequate off-street
parking is available. Surprisingly, studies of residential areas
which relate the two functions are apparently unavailable. A recent
report by the National Association of Home Builders Research Founda-
tion, Inc. (1973) addressed this question. They were unable to find
any work which related residential street design with the availability
of off-street parking. Their conclusions are listed below:
The conclusions listed below are based on results of the National
Survey on Street Development Standards. It is intended that these
conclusions be used as a basis for discussion and development of
recommendations for street development performance standards.
The Current Situation
1. There is extensive lack of uniformity
of street design standards of cities and
counties in the United States.
2. Street design standards reflect practices
of 20 years or more in the past and have not
kept pace with current knowledge.
3. Many jurisdictions have not adopted the
recomnendations of the National Committee
for Traffic Safety or the Recommended Practice
of the Institute of Traffic Engineers. If
they have used these standards as a guide for
design values such as street widths, the values
often have been increased on a presumably
arbitrary basis.
4. Current street design standards are often
incompatible with the types of streets needed
for new patterns of residential development
such as planned unit development, cluster
122
-------
housing and multi-family housing with
internal circulation. To a certain extent,
this is true even of the two previously
mentioned national standards.
5. A number of organizations concerned with street
development standards have a highway or major
street orientation.
6. The state of the art with respect to design-
ing rigid concrete and flexible asphalt pave-
ment for given traffic load and subgrade
conditions is satisfactory to obtain value
engineered street pavement sections; for
example, the thickness design methods of The
Asphalt Institute and Portland Cement Associa-
tion.
7. There is not widespread acceptance of these
design methods by cities and counties. Design
procedures, or more likely, rigid specifications
of State Highway Departments, are adopted by
cities and counties in lieu of establishing
their own requirements. These specifications
are for heavily traveled thoroughfares and major
highways.
8. Regional variations in design and, to a lesser
extent, construction standards are not a
deterrent to establishing national "model"
performance standards.
What Needs to be Done
9. The traditional hierarchy of street types,
arterial, feeder, and local streets, needs to
be revised and expanded and brought more into
line with the performance concept of Average
Daily Traffic (ADT). The concept of branching
streets should be recognized.
10. Street construction standards should be based
on performance principles using the two basic
factors of design, ADT or similar index and
123
-------
standard tests and classifications of sub-
grade materials, either California Bearing
Ratio (GBR), Resistance (R) Value or Bearing
Value Determination, in combination with
recognized design methods for concrete and
asphalt pavements. Pavement design lives should
be realistic. Traffic loadings for residential
areas should take into account lesser weights
of trucks and volume of truck traffic and
elimination of future growth factors where the
streets do not carry through traffic.
11. There is a need for wider acceptance of proven
methods of upgrading street subgrade materials
to allow for reduction in pavement surface
thickness. There is also a need for further
development and experience on new types of
economical subgrade treatment.
12. For open-space communities, consideration should
be given to distribution of a portion of sur-
face water drainage over ground rather than
through streets. This method would allow the
elimination of some curb and gutter.
13. There is a need for data on maintenance and
repair costs of different pavement types.
14. In general, there is need for an educational
program to gain better understanding of the
need for value engineered residential street
development standards and for new approaches
to meeting this objective.
Thus, it appears that current standards for residential street design
have evolved on an ad hoc basis with little or no regard for their
function. While a dearth of information exists regarding residential
street design, there is an abundance of information regarding express-
ways and major arterial streets (see Paquette et al.31972, for example).
Performance criteria are expressed in terms of traffic flows and
related parameters. In fact, benefit-cost type studies have been done
but, unfortunately, they are oriented to the traveler and not those
who are affected detrimentally by the transportation system.
124
-------
Interest in streets was nurtured by the observation that, from an
environmental point of view, current street design criteria are
causing many problems. Exploration began of the possibility of
modifying some of these practices. For example, when should
curb and gutters be utilized? From the environmental point of view,
it is desirable to avoid them if possible. Also, how much incon-
venience would result if stormwater inlet design was modified? The
trend over the past few decades has been to provide larger and
larger inlets and to remove grates. These changes have improved
hydraulic eff-Loienoy, but have aggravated water quality problems.
This research supports the view that streets should be designed on
a functional basis that examines all of the benefits and costs and
relates the design to needs other than parking and traffic flow. The
NAHB study (1973) recommends the standards for residential street
design as shown in Table 30.
By contrast with the recommendations shown in Table 30 existing
residential streets range in width from 26 feet to 40 feet (8 m to
12 m). The most typical width was 36 feet (dim) which provides for
two 10-foot (3m) traffic lanes and two 8-foot (2.5m) parking lanes.
The Denver drainage study classifies streets into the four categories
shown below (Drainage Criteria Manual, Vol. I, 1968):
Local - minor traffic carrier within a neighborhood
which usually is characterized by two moving lanes
and parking along the curb. There is no through
traffic moving from one neighborhood to another.
Traffic control is by use of stop or yield signs,
and in many cases, there is no formal traffic
control.
Collector - collect and distribute traffic between
arterial and local streets. There may be two to
four moving traffic lanes and parking is allowed
adjacent to curbs. Traffic on collectors has
right-of-way over traffic from adjacent local
streets.
Arterial - permits rapid and relatively unimpeded
traffic movement throughout a city. Four to six
lanes of traffic exist and parking adjacent to curbs
may be prohibited. The arterial traffic normally
125
-------
Table 30. PROPOSED RESIDENTIAL STREET DESIGN STANDARDS SUMMARY
(NAHB, 1973)
CTl
Service
Traffic (ADT) (1)
Pavement width
No parking
Parking 1 side
Parking 2 sides
R of W width
Street slope (4)
Sight distance
Maximum speed
Street description
Place Lane Subcollector
Very light Light Local traffic
0 - 200 201 - 500 500 - 1,000
18' 18' 26' (2)
18' 18' 28'
26' 26' 36'
24' to 30' 24' to 30' 44' to 60'
0.5% to 15% 0.5% to 10% 0.5% to 10%
75' 150' 200'
20 mph 25 mph 30 mph
Collector Arterial
Local & thru Thru only
1,001 - 3,000 3,000 +
28' (3)
36' (3)
36' (3)
44' to 60' (3)
0.5% to 8% (3)
250' (3)
35 mph (3)
(1) Average Daily Traffic (ADT) estimates are as follows: single-family detached houses - 7.0;
group or townhouses - 6.0; garden apartments 1 to 4 story - 5.0; elevator apartments over 4
stories - 4.0.
(2) Two nine-foot moving lanes plus one eight-foot emergency stopping lane.
(3) Arterial streets shall be designed for specific traffic and roadway conditions as well as
other related factors.
(4) Adequate cross slope of at least two per cent is required to prevent ponding.
-------
has the right-of-way over collector streets which
it crosses with intersections at approximately 1/4
mile intervals. Construction of an arterial will
often include a median strip with traffic channeli-
zation and signals at numerous intersections.
Freeway - permits rapid and unimpeded movement of
traffic through and around a city. Access to the
freeway is completely controlled by interchanges at
major arterial streets. There may be up to eight
lanes of traffic and parking is not permitted on the
freeway right-of-way.
The Denver study defines two storm events for which the facilities
should be designed: the initial storm (storm of 2 to 10-year recur-
rence intervals) and the major storm (storm of 50 to 100-year recur-
rence intervals). They present detailed engineering design data for
analyzing how the drainage system will perform in both of these cases.
Traditional drainage system design did not take this integrated
approach and analyze the role of the street itself as a stormwater
conveyance device. The Denver studies recognize that certain types
of streets might be used as a temporary storage areas for these
unusual events. Their criteria for minor and major storms are shown
in Tables 31, 32, and 33.
For the minor storm, temporary inundation is permitted for the less
utilized streets. For the major storm, a deeper inundation is allowed.
The allowable depths were determined based on analysis of the benefits
and the costs. The earlier section which discussed parking indicated
that many local streets are not utilized efficiently so that sections
of the pavement could be made available for temporary storage.
The Denver work is very helpful since the purposes of stormwater
quantity and quality control are complementary in this case. The
temporary ponding of water would reduce velocities and thereby dis-
courage the transport of some types of pollutants. It would also
reduce the peak rate of discharge.
The other major area of concern in this section is street inlets.
Two types of inlets are used: curb inlets and gutter inlets. Some-
times a combination inlet is also used. The design of such inlets
is discussed in detail in the Drainage Criteria Manual (1968). In
discussing the pros and cons of inlet design, they state that "The
127
-------
Table 31. ALLOWABLE USE OF STREETS FOR INITIAL STORM RUNOFF
IN TERMS OF PAVEMENT ENCROACHMENT
(Drainage Criteria Manual, 1968)
Street classification Maximum encroachment
Local No curb over-topping.3 Flow may
spread to crown of street.
Collector No curb over—topping.a Flow
spread must leave at least one
lane free of water.
Arterial No curb over-topping.3 Flow
spread must leave at least one
lane free of water in each
direction.
Freeway No encroachment is allowed on
any traffic lane.
a Where no curbing exists, encroachment shall not extend over
property lines.
128
-------
Table 32. MAJOR STORM RUNOFF ALLOWABLE STREET INUNDATION
Street classification Allowable depth and inundated areas
Local and collector Residential dwellings, public, com-
mercial, and industrial buildings,
shall not be inundated at the ground
line, unless buildings are flood
proofed. The depth of water over
the gutter flowline shall not exceed
18 inches.
Arterial and freeway Residential dwellings, public, com-
mercial, and industrial buildings
shall not be inundated at the ground
line, unless buildings are flood
proofed. Depth of water at the
street crown shall not exceed 6
inches to allow operation of emer-
gency vehicles. The depth of water
over the gutter flowline shall not
exceed 18 inches.
129
-------
Table 33.. ALLOWABLE CROSS STEEET FLOW
Street classification Initial design runoff Major design runoff
Local
Collector
Arterial
Freeway
6-inch depth at crown
or in cross pan
Where cross pans
allowed, depth of
flow shall not
exceed 6 inches
None
None
18 inches of depth
above gutter flowline
18 inches of depth
above gutter flowline
6 inches or less over
crown
6 inches or less over
crown
130
-------
curb inlet does not clog readily which is its major advantage. The
large dimensions of the clear opening compared to that of a grated
inlet allows trash to pass into the storm sewer system rather than
be trapped at the inlet."
It is at this point that conflicts arise between the desire for
greater hydraulic efficiency and the water quality management goal
of keeping the "trash" out of the storm sewer. Table 34 shows the
proportion of theoretical capacity which can be used as a function of
the type of inlet. As can be seen, the grated inlets are signifi-
cantly less efficient from a hydraulic point of view. They are more
effective from a water quality viewpoint. By adopting the design
using larger curb inlets, road engineers have reduced their costs
substantially. Unfortunately, this has further aggravated water
quality problems. For example, the current gutter design in Gaines-
ville, Florida uses an open curb gutter with a length of several
feet. Such inlets have a very high hydraulic capacity so that the
water drains rapidly. The design of inlets is an example of where
people could, by proper housekeeping, use more restrictive inlets if
they are cleaned properly. The consequences of clogged inlets are
minor for most residential streets and they accrue to people who are
responsible for the clogged inlets in the first place.
This study of streets has brought out the fact that it may be possible
to modify the design of, or eliminate curb and gutters in lower density
residential areas with well-drained .soils. A significant component
of land use in urban areas is for streets and associated right-of-way.
For example, the typical street in Chicago is 34 feet (10.4 m) wide
with a right-of-way of 66 feet (20m). Residential streets in Gaines-
ville, Florida are 28 feet (8.5m) wide with a 50 foot (15m) right-of-
way. In both of the above examples, curb and gutter is used. As
will be shown later in the economics section, the use of curb and
gutter with associated storm drains about doubles road construction
costs.
In many residential areas of Gainesville, curb and gutter is not
used. Instead, the street drains onto adjacent vegetated swales which
eventually drain to the receiving water.
From a water quality point of view, it is highly desirable to avoid
curb and gutter since the natural drainage pattern will encourage
infiltration and the vegetation serves as a water quality control
device.
131
-------
Table 34. REDUCTION FACTORS TO APPLY TO INLETS
(Drainage Criteria Manual, 1968)
Condition
Inlet type
Percentage of
theoretical capacity
allowed
(1)
Sump
Sump
Sump
Continuous grade
Continuous grade
Continuous grade
Continuous grade
Continuous grade
(2)
Curb opening
Grated
Combination
Curb opening
Deflector
Longitudinal bar grated
Transverse bar grate or
longitudinal bar grate
incorporating transverse
bar
Combination
(3)
80%
50%
65%
80%
75%
60%
50%
110% of that listed
for type of grate
utilized
132
-------
The Drainage Criteria Manual (1968) gives permissible velocities
for various types of vegetation, soil types, and slopes. This
report recommends design velocities of >_ 2 ft/sec (0.6m/sec) for
initial runoff to minimize sediment depositional problems. Recom-
mended design depths should be less than 1.5 feet (0.46m) and
preferably less than 1.0 feet (O.lrn). This report takes exception
to the minimum velocity constraint for such roadside channels. It
is the opinion of the writers that it is preferable to design the
channel such that sediment transport does not occur during the minor
storm.
In summary, it appears that one could design, in many parts of urban
areas, aesthetically appealing and yet functionally sound drainage
systems which avoid the use of storm sewers.
Rooftops
Basically, there are two types of rooftop design in common use, i.e.,
flat roof and gabled roof. In residential areas, buildings up to
three stories in height tend to have gabled roofs. Flat roofs on
the other hand, are almost entirely used for residential buildings
more than three stories in height and for non-residential construc-
tion such as warehouses, shopping centers, etc. Flat roofs are less
expensive to build than the gabled roofs.
Roofs are designed to carry a specified live load in addition to their
own weight. Many building codes, especially in the northern states,
also require roofs to be designed for a specified snow load.
From the standpoint of utilizing the roof as a potential storage
area during runoff, it is impossible to utilize gabled roofs. Flat
roofs are therefore the only type that can be considered for such a
use. However, it must be recognized that the life of a flat roof
used for storing water might be less than that of gabled roof or even
flat roof without storage.
The amount of water that can be stored on an existing flat roof will
depend on the design load. New rooftops can be designed to store
larger amounts of water.
E. ECONOMIC ANALYSIS OF CONTROL OPTIONS
This section presents cost data associated with various control
alternatives. The analysis is partitioned into three sections:
133
-------
storage options and costs, treatment options and costs, and miscel-
laneous options and costs.
1. Storage Options and Costs
Basically, there are five storage options:
1) surface storage such as reservoirs, ponds,
tanks, etc.;
2) in-system storage such as in trunk lines and
interceptors;
3) underwater storage;
4) deep tunnels; and
5) underground modules.
The main objective behind all these options is to provide temporary
storage of combined overflow during peak storm runoff periods and
to release this flow to the treatment facility during subsequent
periods of lower flow. Costs for various kinds of storage facilities
based on Sandusky, Ohio data are given in Table 35. A comparison of
storage costs for several cities is shown in Table 36. Note the wide
variation in costs which indicates the significance of local conditions.
2. Treatment Options and Costs
A variety of treatment options and costs are available in the original
SWMM. As a consequence of this research, several other treatment
options have been incorporated in the SWMM Model as in Section IV-D.
The cost functions of other treatment options are described herein.
In addition, several cost functions in the original SWMM have been
revised in the light of data recently made available.
Cost data that have been made available recently pertain to the follow-
ing units:
1) high-rate filtration;
2) microstrainers; and
3) dissolved air flotation.
134
-------
Table 35- COST OF STORAGE IN SANDUSKy, OHIO
(Rohxer and Associates, Inc., 1971)
Storage capacity Construction costa
Storage method in gallons in dollars
Reinforced concrete tank 200,000 84,930
1,000,000 358,670
Lagoons 200,000 95,820
1,000,000 351,740
Underwater tank 200,000 22,000
1,000,000 102,000
a Excluding engineering, land and acquisition costs.
135
-------
Table 36. COMPARISON OF STORAGE COSTS FOR VARIOUS CITIES
(Rotirer and Associates, Inc., 1971)
Area served
in acres
Cost
$x!06
Cost/acre
$/acre
Cost/capita
$/capita
S andu sky, Ohio
Underwater storage
(1 year storm)
Sandusky, Ohio
Underwater, underground
^J and surface
(1 year storm)
Chicago, Illinois
Deep tunnel
(100 year storms)
Saginaw, Michigan
Chippewa Falls,
Wisconsin
Lagoon
Other cities
580
2,205
192,000
90
3.57
7.75
1,270
.61
6,160
3,500
6,615
415
5,140
6,780
20
6,500
212
423
62
670
6
1,300
-------
The following new treatment options have been added to the Model:
1) swirl regulator/concentrator;
2) biological treatment; and
3) high— rate disinfection.
Other treatment options utilize the cost functions provided in the
original SWMM (EPA, 1971a) .
Revised Cost Functions
High-Rate Filtration — The original Model computes the capital cost,,
in dollars, of this unit using the following equation:
Capital Cost = 54,000 Q0-67fENR |p (15)
where Q = capacity, mgd
ENR = Engineering News Record (ENR) index for prescribed
year
F = ENR site factor for the geographic location of
treatment facilities
Recent pilot plant studies conducted by Hydrotechnic Corporation
(1972) on high-rate filtration outlined cost data which are shown on
Figure 32. On the same figure, the cost that would be computed by
the Equation (15) is also plotted for ENR of 1470 and a site factor
of 1.0. Hydrotechnic Corporation developed cost data for two filtra-
tion rates, i.e.} 16 gal/min/ft2 (65 cm/min) and 24 gal/min/ft2 (98
cm/min) .
It can be seen that Equation (15) would predict a lower cost than
estimated by Hydrotechnic Corporation (1972). Realizing that various
uncertainties exist, it appears appropriate to use the following
equation for cost purposes:
Capital Cost = 85,000 Q°-67[ENR )?
e |l058j
The curve derived from this equation for ENR of 1470 and a site factor
of 1.0 is also shown in Figure 32. It is evident that this curve
gives a better estimate than the original one. It has therefore been
incorporated in the SWMM.
137
-------
OJ
oo
CO
o
o
2 -
o
I -
ENR = 1470
F = |;0
16
85000Q°-67(ENR\F
11058 I
ORIGINAL CURVE
50
100
DESIGN CAPACITY M G D
ISO
200
FIGURE 32.
Capital Cost for High-Rate Filtration.
-------
Micros trainers — The original Model computes the capital costs of
the microstrainer facility based on one of the following equations:
Capital Cost = 30,000 Qf?§* IF for Q < 25 mgd
(13141 x — e
fFNR } <17>
= 20,000 Q |gjL-JF for Q ^ 25 mgd
Cochrane Division of Crane Company has conducted studies on micro-
strainer installations (Glover, 1972). According to Glover's estimate,
a 30 mgd (1.3 m^/sec) microstrainer installation in 1969 would cost
$217,000. This gives a cost of $7,200/mgd. A comparison of this cost
with the equations in the original Model indicates a considerably
higher cost predicted by the Model. It is proposed to use the follow-
ing relationship which corresponds to Glover's work:
f \
Capital Cost = 10,000 Q M*L p (18)
(1383)
Dissolved Air Flotation — Capital costs for this unit are computed
in the original Model by one of the following equations:
( 2.3026 ]ENR
1-35 Q exp[o.2075 + 0.0114 log Qjl^98F for Q ^ 15
., ENR _,
1.35 Q exp [2.3026(4.6032 - 0.0559 log Q>1 io98~F for 15 <_ Q <_ 100 (19)
31,000(1.35)C;te_lF for Q > 100
Full-scale units are currently in operation in Milwaukee with a
design flow of 5 mgd and in Racine, Wisconsin with design flows of
14.13 and 44.4 mgd (0.629 and 1.95 m /sec) each. The construction costs
of these facilities have been reported to be about $25,o6o/mgd in 1971
(Rex Chainbelt, 1972).
An examination of the model equations outlined above indicates that
capital costs estimated are substantially higher than reported for
Milwaukee and Racine installations. The following equation is there-
fore proposed for all capacities:
Capital Cost = 31,000 Q [ENR IF (20)
139
-------
New Treatment Options
Swirl Concentrator — A full-scale, 36 foot (llm) diameter installa-
tion is currently under construction at Lancaster, Pennsylvania.
Cost of this installation has been estimated at $100,000 and is
expected to handle about 165 cfs (4.7 m^/sec) of flow. Based on
the above data, it is proposed to estimate the cost using:
Capital Cost =100^(11^ (21)
where D = diameter of the swirl concentrator, ft.
Following construction of this and other installations, a revised
equation will be utilized.
Biological Treatment — Cost of this unit is based on the full-scale
contact stabilization plant now in operation in Kenosha, Wisconsin,
with a capacity of 20 mgd (0.88 nrVsec). Cost of this facility has
been estimated at $55,000/mgd (Agnew, 1973). Thus, the following
equation is proposed for the purpose of modeling the cost of this unit:
Capital Cost - 55,000 Q|ML-|F (22)
[1400J
High-Rate Disinfection — cost modeling of this unit is based on
studies conducted by Cochrane Division of Crane Company (Glover, 1972).
Cost of a 30 mgd (1.3 m3/sec) facility is estimated at $59,000. Thus,
the cost equation proposed for modeling this unit is as follows:
Capital Cost = 2,000 Q
ENR
1314
(23)
Other Treatment Options — Cost data for other options already in
the Model remain the same. Many other treatment options are currently
under investigation (Field and Struzeski, 1972). Enough information
is not available at this time to justify their inclusion in the Model.
3. Miscellaneous Options and Costs
Porous Pavement
The primary objective of porous pavement is to reduce runoff from
paved areas, by allowing rain to percolate through it, thereby
reducing the magnitude of outfall discharges. In its simplest form,
140
-------
the porous pavement consists of a base course of crushed gravel with
a surface course of porous asphalt concrete. In a recent publica-
tion by The Franklin Institute Research Laboratories (1972), data
are presented on the depth of the surface and base course for various
kinds of uses. In considering the use of porous pavements, consid-
eration of features such as geographical area, temperature, sub-
surface and soil conditions, and possibility of groundwater contamina-
tion are important. A cost comparison between conventional and por-
ous pavement for various uses (The Franklin Institute, 1972) is given
in Table 37.
This tabulation shows that the costs for porous pavement compare
favorably with'conventional pavements. It is pointed out that, at
present, there is only one full-scale installation near Houston,
Texas that recently started evaluating porous pavements for flow
attenuation and therefore its feasibility has yet to be demonstrated.
Erosion Control Opt-ions
Erosion from land which has been partially or completely denuded is
most often controlled with the use of various on-site methods. These
methods generally fall into two classes; namely, structures and
stabilizing treatments. In Table 38, the costs for various erosion
control alternatives are listed (Brandt et aZ.,,1972).
There are generally two types of chemicals that have been developed
for control of erosion. These two types are listed in Table 38 as
Chemical A and Chemical B. Chemical A has a manufacturer's price of
about $150/acre ($370/ha) and usually is applied with 2,000 gallons
of water/acre (18,700 1/ha). Chemical B has a price of about $250/
acre ($618/ha) and is applied with 1,000 or less gallons/acre
(9,350 1/ba) (Brandt et al.3 1972).
Table 38 gives the costs involved in procurement and application of
the control and does not include design or maintenance costs. Main-
tenance costs can vary considerably and consequently, it is diffi-
cult to give a general cost estimate. Where no real estimates are
available, a figure of $125/acre ($309/ha) for maintenance during
the developmental process has been used by some urban developers.
It has been estimated that cleaning sediment basins costs $2.50/ton
of sediment removed (Brandt &t a1.3 1972).
4. Options for Removal of Surface Pollutants
Cost estimates for alternatives which involve the removal of surface
pollutants from the drainage system such as street cleaning, catchbasin
cleaning, and sewer flushing can usually be obtained from the local
public works department for the area of concern. Where no such esti-
mates are available, the costs lis-ted in Table 39 can be -asred.
141
-------
Table 37. COMPARISON OF THE COST OF CONVENTIONAL PAVEMENT AND POROUS PAVEMENT
(Franklin Institute, 1972)
Type of use
Parking lot
Residential street
Low design
High design
Business street
Suburban development
City
County road
Highway - two-lane
Cost
$A
6.32
6.33
6.19
6.65
6.99
6.01
6.80
Porous
range
yd2
- 6.40
- 6.37
- 6.56
- 6.97
- 8.32
- 8.32
-7.92
pavement
Average cost
$/ yd2
6.36
6.35
6.38
6.78
7.66
6.44
7.36
Conventional pavement
Cost
$/
3.65
4.80
8.35
6.90
14.20
7.24
8.54
range
yd2
- 5.81
- 9.60
- 12.10
- 9.90
- 22.20
- 13.23
- 20.45
Average cost
$/ yd2
4.73
7.20
10.23
8.40
18.20
9.20
14.50
-------
Table 38. EROSION CONTROL COSTS
(Brandt et at., 1972)
A) Structures
1) Diversion berms
2) Interceptor berms
3) Sodded ditches
4) Grade stabilization structure
5) Level spreader
6) Small sediment basin
B) Stabilizing treatments
1) Seed, fertilizer, mulch
2) Seed, fertilizer, straw tack
3) Seed
4) Chemical A
5) Chemical B
6) Seed, fertilizer, chemical A
7) Seed, fertilizer, chemical B
$ 1.25/linear ft
$ 1.25/linear ft
$ 4.50/linear ft
$ 2.80/linear ft
$ 2.80/linear ft
$ 500.00/each
$l,500.00/acre
$ 493.00/acre
$ 209.00/acre
$ 300.00/acre
$ 387.00/acre
$ 480.00/acre
$ 567.00/acre
143
-------
Table 39. UNIT COSTS OF CATCHBASIN CLEANING,
SEWER CLEANING AND
MECHANICAL STREET SWEEPING
(APWA, 1969)
Catchbasin cleaning
Hand
Eductor
Orange peel
Sewer cleaning
Flushing
Scraping
Mechanical
Street sweeping
$ 3.29/catchbasin
$ 3.47/catchbasin
$ 4.38/catchbasin
$ 8.12/1,000 linear ft
$187.00/1,000 linear ft
$100.00/shift
144
-------
These costs were estimated from 1967 data collected in Chicago,
Illinois (APWA 1969).
F. SPECIFICATION OF PERFORMANCE CRITERIA
This report provides a method for examining ways to control the
water quality associated with urban stormwater. However, the water
quality problem is intimately related to the corresponding problem
of controlling water quantity. Most of the control alternatives
have a direct effect on quality as well as quantity. For example,
stormwater retention basins are growing in popularity as devices
for controlling runoff rates (Poertner, 1973). Such basins may also
provide sediment control. Unfortunately, the retention basin design
may not have considered water quality control and the result can
be eyesore for the neighbors. Recognizing the interrelationship,
the problem has been addressed from the point of view of designing
control facilities for stormwater quantity and quality management.
It appears that the "design event" for control of quantity and quality
would often differ. Typically, a higher recurrence interval would be
used for water quality control. There are several justifications for
this as sump t ion:
1) A prime concern in water quality control is
the "first flush" associated with the trans-
port of pollutants during early parts of the
storm.
2) As the frequency of the storm decreases, the
effect of pollutants being washed into the
sewer system diminishes in relative importance
and natural processes become dominant.
3) The "pollutants" associated with these natural
processes are comprised primarily of sediment
and vegetation.
4) For more significant storms, the flow in the
receiving water is increased substantially.
Studies in Denver provide a thorough examination of urban stormwater
problems. They specify design criteria for two types of storms:
minor storms and major storms. The former storms are designed around
drainage criteria whereas the latter storm is designed for flood con-
trol. Design criteria for the minor storms are shown in Table 40.
Design criteria for the major events are shown in Table 41.
145
-------
Table 40. LEVELS OF PROTECTION .FOR MINOR STORM EVENTS
(Drainage Criteria Manual, Vol. I, 1968)
Recurrence interval
Land Use (years)
Residential 2
High-value general
commercial area 5
Public building area 5
Airports 2-5
Major airport terminals 5-10
High-value downtown
businesses 5-10
Table 41. LEVELS OF PROTECTION FOR MAJOR STORM EVENTS
(Drainage Criteria Manual, Vol. I, 1968)
Recurrence interval
Land use (years)
Urban 100
Industrial Varies
Agriculture
High-value 10-25
Low-value 2-10
146
-------
The purpose of discussing both the minor storm and major storm cri-
teria together is to emphasize to the reader that these are separate
drainage systems with different purposes. The purpose of the design
for the minor system is ". . . to prevent inconvenience and frequently
recurring damage caused by frequent storm runoff" (Drainage Criteria
Manual, 1968). The purpose of the major system is to control flood-
ing with its associated significant and sometimes catastrophic con-
sequences .
Urban drainage systems have been designed based on the minor storm
with little or no consideration as to what will happen during the
major storm. It may be that designers felt that the design for the
minor storm would also provide improved flood control during the
major storm. Unfortunately, just the opposite is often the case.
The work in Denver represents an excellent start towards inducing
designers to analyze the performance of the system for both the minor
event and the major event. For example, streets play a very important
role as conveyance channels for stormwater. In the Denver report,
criteria are presented for allowable flow along various types of
streets for the minor and major storms (see Tables 31 to 33 in this
section).
With regard to stormwater quality management, the design criteria we
recommend are similar to the criteria for the minor storm. As shown
earlier, urbanization causes higher peak flows due to the increased
impervious areas which discharge directly to storm sewers, roadside
ditches, etc., and thence into the receiving water. In the follow-
ing sections, a discussion is presented regarding the selected per-
formance criteria.
These proposed criteria attempt to physically Internalize the exter-
nalities caused by stormwater runoff using a quasi-closed system
approach. It will be argued that owners of land within the urban
area are responsible for "safe" disposition of stormwater. This may
be achieved on-site by discharging stormwater onto pervious areas
and/or retention basins within each parcel of property; or by forming
a coalition with other owners to accomplish this same purpose off-site
by acquiring a common retention basin, flood plain, etc., for this
purpose. Off-site control could be on publicly owned land which could
serve also as a park, recreational area, etc. This approach has the
potential advantage of providing an explicit procedure for justifying
open spaces within urban areas by recognizing the vital function it
serves in water resources management. By physically internalizing
the externality, one has partially resolved the equity question since
the urban property owners are comparing control costs vs on-site
damage costs. Thus, they are recipients of the damages if the system
is not designed and operated properly, not someone downstream. A
147
-------
similar argument might be made for other types of developments.
Natural systems cause less off-site effects than developed lands
because of a longer detention time for stormwater which permits
additional infiltration of water. Also, the vegetation and soil
serve as water quality control devices. The approach taken in this
section assumes that the owner of land is responsible for safe
disposition of stormwater from an environmental point of view. The
design event is assumed to be the same as that shown in Table 40.
It will be argued that the stormwater from this event and more fre-
quent events will be handled such that there is:
1) no increase in the peak flow observed prior
to development; and
2) adequate treatment of stormwater prior to
release to the receiving environment.
Adequate treatment may be achieved by utilizing an appropriate control
device such as percolating the stormwater through the soil, or pro-
viding a retention basin and control device with a sufficient resi-
dence time. The responsible individual or group may provide these
facilities on his own property and/or form a coalition with others
to accomplish the same purpose at a centralized facility.
These same principles can be applied to related off-site control
alternatives. Two extensions of the on-site analysis are needed,
i,e*s the transport cost for directing the water to the control facil-
ity; and the cost of the control facility itself. The control facility
may be a pervious area such as a flood plain or a structural facility
such as a high-rate treatment device. The samp, performance criterion
can be applied as was done in the case for individual on-site treat-
ment. An advantage of this approach is that responsibility for the
stormwater can be assigned which is essential to an equitable apportion-
ment of control costs.
Similar performance criteria are being used by some cities for storm-
water quantity management. For example, the Metropolitan Sanitary
District of Chicago requires control of stormwater according to the
following criteria (personal communication, 1973):
1) A combination of storage and controlled release
of stormwater runoff is required for all com-
mercial developments exceeding 5 acres in area,
for all residential developments exceeding 10
acres in area and for all residential develop-
ments between 5 and 10 acres which have an
imperviousness of 60 percent or greater. How-
148
-------
ever, a residential development of 10 acres
or less must have an adequate outlet as signi-
fied on the District permit by the Municipal
Engineer. If the outlet is not adequate, then
detention, as determined by the Municipal
Engineer, will be required to store that portion
of the runoff exceeding the outlet capacity.
2) The release rate of stormwater from all
developments requiring detention shall not
exceed the stormwater runoff rate from the area
in its natural undeveloped state.
3) Drainage systems shall have adequate capacity to
bypass through the development the flow from all
upstream areas for a storm of design frequency
assuming that the land is in a fully developed
state under present zoning or zoning proposed
under a Comprehensive Plan. The bypass flow
rate shall be computed utilizing a runoff
coefficient of not less than 0.35. An allow-
ance will be made for upstream detention when
such upstream detention and release rate has
previously been approved by the District and
that evidence of its construction can be shown.
4) The live detention storage to be provided will
be calculated on the basis of the 100-year
frequency rainfall as published by the U. S.
Weather Bureau for this area. The detention
volume required will be that necessary to
handle the runoff of a 100-year rainfall, for
any and all durations, from the fully developed
drainage area tributary to the reservoir, less
that volume discharged during the same duration
at the approved release rate.
G. SYSTEM OPTIMIZATION AND EXAMINATION OF EQUITY QUESTION
The objective of this analysis is to devise a decision-making model
for the specific problem of urban stormwater quality management. A
fundamental question is to determine the optimal level of pollution
control. There are basically two approaches. One is to assess the
damages caused by the pollutant, convert them into an equivalent
monetary value, and thereby determine an aggregate, receptor damage
function. Knowing this function and the cost of controlling the
pollution, one can determine the optimal degree of control. The
149
-------
primary problem with, this approach, is that it is extremely diffi-
cult, if not impossible, to quantify the receptor damage function.
The other approach is based on specifying performance standards which
are based on a review by experts of the technical, economic, finan-
cial, social and political aspects of the problem. Given this set
of standards, the polluters are assumed to select the least costly
way of satisfying that standard. This is the approach we shall use.
The selected performance standards were discussed earlier.
Thus, the stated objective for this analysis is to determine the
least cost solution subject of meeting a prescribed set of performance
standards. This least cost solution might comprise a wide variety
of on-site controls as well as off-site controls. However, the
solution to the optimization problem is not the final answer. In
addition, we need to analyze whether the solution is also fair to
everyone. This question will be examined using recent developments
from cooperative N-person game theory.
There are numerous ways to control pollution using on-site control
and/or off-site control. Thus, a general framework is needed for
addressing the problem. The selected approach is based on the plan-
ning theory of zoning (Herzog, 1969). Using the planning theory of
zoning, each source of pollution calculates the cost of handling the
problem on-site as a function of the allowable rate of release from
his area. The cost function could look like the curve shown in
Figure 33.
ON-SITE
POLLUTION
CONTROL COST:
$/Q
ALLOWABLE RELEASE: Q
FIGURE 33.
Cost Function for On-site Stormwater Control
150
-------
Costs decrease as the allowable release increases. In fact, the cost
falls to zero as he is allowed to release more and more pollution.
Assume that such a curve exists for each area. Note that the abscissa
could be air pollutants, water pollutants, noise or any other "nuisance"
which is normally covered by zoning regulations.
In this case, the policy question is to determine how much pollution
can be released from each of these areas. Using the planning theory
of zoning, the answer depends on a determination of "assimilative
capacity." Assume that the required level of control can be specified.
The following example is presented in order to illustrate many of the
concepts to be discussed. Assume the city under consideration has
been partitioned into three study areas. Each study area has two
options:
1) on-site control; and/or
2) off-site control at a central storage facility.
The following notation is used:
Let Xj• = decision variable: number of units of
control j selected for area i
Xjj = upper bound on x^.
C;H = unit cost for control j in area i
DI = quantity of commodity originating in area i
Q-L = maximum allowable release of stormwater
from area i
Z-£ = total control cost to area i
TT . = ^reduction in control cost to area i if
Q-^ is increased by one unit
t. = unit cost of transporting water from area i
to the central control location
c = unit cost of central control
W = maximum available control at central
facility
151
-------
The example problem is shown in Figure 34 using a network representa-
tion. This problem is deliberately oversimplified to permit us to
understand the concepts without getting enmeshed in computational
difficulties. The results can be extended easily to more realistic
cases where multiple central control facilities exist.
The overall objective function seeks to minimize the total cost of
on— site and off-site control. The problem facing each study area is
to minimize
Zl
subject to
Qi - Qi (24)
x > 0 for all j, and
For area 1, his problem is to select x11} xi2> and x13 so as to
minimize Z-^ = lOxji + 5xi2 + 0X13 + 4Ql
subject to xll + X12 + xl 3 "*" Ql = 500
Ql 1 Ql
xn _< 100
X12 1 20°
X13 1 20°
xll> X12. X13» Ql >. 0
152
-------
D , = 5OO
FIGURE 34.
Network Representation of Example Problem.
153
-------
This is a very simple problem to solve if Q is known.
As it turns out this is a question of critical importance_in some
cases. In the context of stormwater quality management, Q-, represents
a judgment on the part of the administrator as to the "assimilative
capacity" or availability of off-site controls to area 1. Tradi-
tionally, the natural system has provided these off-site controls free
of charge. For example, the off—site area in this example could be a
swamp or flood plain.
In order to find Q-, , we need to know how much each area is willing to
pay for off-site control based on fhe-iv alternative cm-site control
costs. This can be_ done by solving the above linear program for
various values of Q. assuming on-site control is required. Thus, for
the moment, Q. is analogous to an effluent standard imposed on area i.
The problem can be solved by deleting the last term, (t^ + c)Q^, from
the objective function and finding the optimal solution for assumed
levels of Q-J^ Computationally this can be done quite easily_ using
parametric programming as explained below. Initially, set Q-^ = 0
and solve the linear programming problem. Then, as a post-optimal
procedure, one can vary Q^ continuously from 0 up to any prescribed
upper bound using as the right hand side Q^ = 0 + Gr where 0 equals
a parameter which will increase continuously from 0 to 1 and r is a
scalar, say 1000, in this case. The solution to this problem tells the
total cost to area i for any value of Q. which is of interest. This
is a very attractive feature of linear programming. Another aspect
of linear programming which is of interest is duality theory. The
solution to the dual problem is obtained when the above problem
(primaJL problem) is solved. Among other things, it tells, for a
given Q^, the reduction in cost to area i if Q-^ is increased by one
unit. This unit cost is called the "shadow price" with respect to
Q-£ and will be denoted as TT . .
Let us solve the example problem for area 1 assuming QT= 0. The
answer is simply that on-site control is used. As the constraint on
Q-, is relaxed, area 1 will substitute off-site control for its most
expensive on-site control, X-Q in this case. Thus, it is saving
$6 per unit change in Q-^ in this range. The analysis continues in
this manner until all solution possibilities have been identified.
The results are shown in Figure 35.
Assume that a similar analysis was done for areas 2 and 3. Then the
willingness to pay of each of the three areas would be known. Next
assume each area was offered as much storage as they wanted at its
cost of (t. + c) dollars per unit. In this case, the aggregate
demand would be Q^ = 300, Q2 = 300, Q3 = 250, or a total demand of
850 units. But suppose only 420 units are available. How do we
154
-------
IT,:
$/Q.
5-
0
100
200
300
ALLOWABLE OFF-SITE RELEASE: Q,
FIGURE 35. _
Shadow Price for Area 1 for Assumed Values of Qj
155
-------
allocate the available storage? If you are the coordinator or
planning authority, we assume that you will attempt to maximize
the aggregate savings to area 1, 2, and 3 from using the regional
facility. Alternatively preference will be given to those areas
whose on-site control costs are the highest.
The answer to the above question can be obtained by solving one
larger optimization model which is formulated on the following page
as Table 42. The constraints for this problem can be divided into
two categories:
1) three sets of study area constraints; and
2) a coupling constraint.
The coupling constraint is the only linkage among the three study
areas. If the values of Q-^, 0~ , and Q« are prespecified, then the
larger problem can be completely decomposed into three independent
subproblems. There are many real-world situations in which an
a pr-iori, apportionment is used, 'I.e., allocation_to each area based
on its size, population, etc. For a pollutant, W might be the
assimilative capacity of the receiving water which has been apportioned.
In the case of a pollutant, the appropriation is equivalent to pre-
scribing effluent standards.
Assume an apportionment is established such that Qi = Q = On = 420/3.
Given this apportionment, find Z-^, Z2 > an(^ Z , the least cost solutions
for the three areas. The results are:
Zi = $1360 : TI^ = $1 : 0^ = 140
Z2 = 1540 : 7T2 = 4 : Q2 = 140
Z3 = 2560 : TT3 = i : Q3 = 140
I Z± = $5460 | Qi = 420
i-1 i=l
Apportioning the water among the three areas results in a combined
cost of $5,460. Next we will examine whether, from a least cost
point of view, it would be possible to select Q-^, Q2, and Q such
that costs are reduced. This problem can be solved by running the
entire linear program with Q^, Q2, and Q3 as decision variables.
This approach is equivalent to stream standards wherein the coordi-
nator allocates the stream's assimilative capacity in an optimal
manner. This is precisely the problem to be addressed here. The
optimal solution is:
156
-------
Ul
Minimize Z •
Subject to
Area 1's
Problem
Area 2' 9
Problem
Area 3's
Problem
Coupling
Constraints
J-l l 1J
3
I *1J + Ql
*1J
Qi
0
0
*
+ I C2jx2j "*" (c+t2)Q2
J-l
0
2
I *2j + Q2
J-l
X2J
Q2
0
'+ Q2
+ 1 c3jX3j + (. 0 for J - 1,2,3
i 0
= D2
<. x2j for J - 1,2
>. 0 for J - 1,2
i 0
<_x3j for J - 1,2,3,4
>_0 f or J - 1,2,3,4
<_ W
Table 42.
OPTIMIZATION MODEL FOR THE THREE STUDY AREAS
-------
z1 = $1400 Q! = 100
Z2 = 1020 Q2 = 270
Zo = 2650 Qo = 50
3 3 3 3
Z Z. - $5070: TT = $4 £ Qn- = 420
1=1 x 1=1
We see that the solution to the overall optimization problem reduced
total costs from $5,460 to $5,070 by making more effective use of
the available capacity. As can be seen by examining the solutions,
the model took capacity from 1 and 3 and allotted it to 2 since his
net savings (T^) were higher. While this latter solution is the
least costly from an overall point of view, if the costs are assigned
as shown above, areas 1 and 3 are made worse off while area 2 is
made significantly better off. Thus, areas 1 and 3 might reasonably
object to such a solution. This is precisely the problem that has
thwarted implementation of optimal programs, i.e., while they are the
least costly, they do not seem to be "fair" to everyone. We would like
to have a procedure which is not only efficient, but also is equitable.
Perhaps one could use economic theory of demand to determine a better
solution. The aggregate demand curve for the three areas is shown in
Figure 36. Knowing the demand curve and the supply curve, then it is
possible to determine the "market price" for the central facility.
Referring to Figure 36, it is $4/Q. Thus, according to economic
theory, to achieve efficient resource allocation, a market price of
$4 per unit of storage should be used. Let ir denote the market price
(which is the same as the shadow price of central storage from the
overall optimization model), then the assessment to each study area
per unit of storage is (TT + c + t.).
Economic theory does not address the distributional questions associated
with using a market price concept. Thus, until recently, a very funda-
mental question was left unanswered as indicated below. First, if
you actually charge the "market price" then a profit of 420ir is
realized. What do you do with the money? Traditionally, public ser-
vices have been priced based on the cost of service. Thus, one could
argue that such a price cannot be charged. Economic theory uses the
notion of consumer surplus which is defined as the difference between
an individual's willingness to pay and the actual assessment levied
against him.
Let us determine the aggregate "consumer surplus" in this case if
each area pays for off-site discharge it is allotted. As can be
seen from this calculation,
158
-------
10'
0
EMAND FOR OFF-SITE DISCHARGE
SUPPLY
500
1000
Q
1500
2000
ALLOWABLE OFF-StTE RELEASE FROM AREAS 1, 2, 3
FIGURE 36.
Demand for Off-site Storage.
159
-------
Consumer Surplus if (c + t^)
Area _ is Charged _
1 6(100) = $ 600
2 4(270) = 1080
3 10 ( 50) = 500
$2180
the aggregate "consumer surplus" is $2180. In this example, the
"consumer surplus" is the savings which accrue to each of the three
areas because of the availability of 420 units of off-site disposal.
This leads us to an important principle. Using this approach, it is
possible to impute a value to the off-site facility based on the
services it is performing which otherwise would have been required
to be installed on-site. If the central facility is part of the
natural system, say a swamp, flood plain, river, lake, or estuary,
then a value of the natural system to these three areas is $2180 in
this example. Thus, we have been able to derive a way of placing a
value on the natural system based on the vital function it is pro-
viding for man! Interestingly, it is related to the notion of con-
sumer surplus in economic theory.
This ability to place a value on the natural system is of vital
importance in resource management. Traditionally, we have attempted
to justify open space retention based on its recreation and aesthetic
values. This has not provided convincing justification for preserv-
ing these areas. Using this technique one can prove, from a functional
basis, that these areas are providing other valuable services for man.
Unfortunately- the problem is not yet solved. What happens if we
attempt to implement the solution that charges only the actual cost
but retains the optimal solution? If we do so, then areas 1 and 3
will probably object since area 2 derives most of the savings. To
avoid this possibility one might charge the market price, c + t . + IT .
If this is done, then the consumer surplus is as shown below.
Consumer Surplus if (c + t.^ + TT )
Area _ is charged _
1 2(100)= $ 200
2 0(270)= 0
3 6( 50)= 300
Net Revenue to 4(420) 1680
"Off-Site Facility"
160
-------
This solution has at least two problems: (1) what to do with the
net revenue,and (2) the fact that any positive incentive to area 2
to participate has been eliminated since he is now indifferent
between on-site control and off-site control.
The above problem is amenable to attack using N-person game theory.
Luce and Raiffa (1957), Owen (1968), and Rapoport (1970) provide an
introduction to the notions of N-person game theory. A key differ-
ence between two-person zero-sum game theory and cooperative N-person
game theory is the idea of coalition formation and the allocation of
associated transfer payments among the players. The balance of this
section provides an axiomatic approach to the problem. A related
procedure for cost-sharing has recently been proposed by Loehman and
Whinston (1971).
Let x{i} denote the cost assigned to the ±th waste discharger if a
cooperative arrangement is adopted and c(N) denote the total cost for
the coordinated solution. Then, the solutions which are admissible
satisfy the following:
x{i} _> 0 for all ieN (25)
£ x{i} = c(N) (26)
In game theory jargon all such solutions are group rational since
the entire cost is partitioned among all of the waste dischargers.
The question to be examined is: How should the costs of this central
facility be apportioned among the three areas?
It is reasonable to require that
x{i} -
-------
All solutions satisfying (25) , (26) , (27) , and (28) belong to the
of the game.
Thus, the first major extension of existing procedures that is
revealed using the game theoretic approach is that the notion of
the core of an N-person game permits one to delimit more sharply
the subset of solutions which are efficient and equitable. It does
so by examining only those solutions which are rational for each
individual or subset of individuals.
Returning to the example, there are T'' - 1 or 7 possible coalitions
to be analyzed:
{1} {2} {3}
{12} {13} {23}
{123}
The next step in the analysis is to find c(S) for the above combina-
tions. c(S) is called the characteristic function f0r coalition S
and represents the total cost to that coalition. Usually, in game
theory, it is assumed that the areas not in the coalition under con-
sideration would form a directly competitive counter coalition which
would lead to a minimax solution of a two-person game. Recently,
S. Sorenson, (1972), a student of A. Charnes , generalized some of
these notions of cooperative N— person game theory to the cases where
alternative definitions of the characteristic function might be used.
He describes four possible definitions which are listed below:
c-i(S) = value to coalition if S is given preference
over N - S
C£(S) = value of coalition to S if N - S is not
present
c^(S) = value of coalition in a strictly competitive
game between coalition S and N — S
c^(S) = value of coalition to S if N - S is given
preference
As we shall see, alternative definitions can be used depending on
how the problem is defined.
We would like to have a technique which permits us to obtain a
unique solution to the problem which we can argue is equitable, or
fair , to everyone .
162
-------
The notion of the core reduces the set of feasible solutions con-
siderably. However, there is still room for negotiation among the
players as to how the costs are to be divided. Each area seeks to
minimize its costs. Assume that the most recent entrant to a
coalition pays these incremented costs. Therefore, the cost for the
±th area is defined as
Ic(s) - c(s-{i»]for all S.CN
Lacking prior knowledge regarding the likelihood of alternative
coalition formation sequences, assume that all sequences may occur
and are equilikely. Thus, there are 3! ways of forming the grand
coalition as listed below:
123 213 312
132 231 321
Under these circumstances, we contend that a fair cost-sharing pro-
cedure is to assign to each area the expected value of the cost
summed over the above coalition formation sequences. The general
relationship is shown below. Let ^ denote the cost assigned to the
ith area where
X (S-l)I(n-S)!
= - c(S - {
The assigned cost, . , is called the Shapley value for N-person games
(Shapley, 1953). Note that,
X i=
ieN
so that the entire cost is apportioned. The Shapley value is a very
useful solution notion. Unfortunately, it may happen that the
Shapley value can fall outside the core of the game (Sorenson, 1972) .
We want to avoid such cases.
We have examined several problems in cost-sharing and resource allo-
cation. It works out that different definitions of the character-
istic function are appropriate depending on how the problem is
defined. The easier case is the cost-sharing problem where ownership
has been established and the decision needs to be made whether to
control on-site only or to form a coalition with others. The more
difficult case is to determine a solution when ownership of the central
facilities is unspecified. These two cases are considered below.
163
-------
From a computational point of view, c^CS) can be determined by solv-
ing a partition of the overall optimization problem. This procedure
will be used in solving the example for various cases. Recall that
the overall objective function is to minimize the costs for all three
areas. Assume there are n goals over which we want to optimize, i.e. ,
£]», f2(x) . . . fn(x)
In our case the n separate goals represent the goals of all coalition
combinations. Then the values of the characteristic function for a
given coalition is found by determining the optimal solution to the
linear program for the specified objective function. The following
cases are considered.
Case 1: Ownership of off-site control is specified. Each area owns"
140 units of the central facility which they may use or sell to other
areas .
In this case each coalition seeks to minimize its costs given a_speci-
fied allocation of the off-site discharge W. Let [fj(x):Q^ <_ W'] denote
the objective function for the ith coalition with a specified alloca-
tion W' of the total of W units. The results_of this calculation are
shown in Table 43. Note that the coalition, 13 is inessential, i.e. ,
c(13) = c(l) + c(3).
The core of this game is delimited by
xx <_ 1360
x2 <_ 1540
x3 <_ 2560
xl + X2 — 278°
x! + X3 - 392°
x + x <_ 3830
Xl + X2 + X3 = 507°
The Shapley values are
*! = J (1360) + | (2780 - 1540) + | (3920 - 2560)
+ (5070 - 3830)
164
-------
Table 43. OPTIMAL SOLUTION WITH OWNERSHIP
OF OFF-SITE FACILITY PRESPECIFIED
Coalition
Cost
Objective
function
Resource _
allocation: W'
Optimal
solution
c(2)
c(3)
c(12)
c(13)
c(23)
c(23)
f2(x) = Z2
f3(x) = Z3
f4(x) - Z12
f5(x) - Z13
f6(x) = Z23
f?(x) =
140
140
140
280
280
280
420
$1360
1540
2560
2780
3920
3830
5070
165
-------
., = $1300
4>0 = -kl540) + ^-(2780 - 1360) + \ (3830 - 2560)
2 3 o o
+ -|(5070 - 3920)
2 = $1345
= ^-(2560) + \ (3920 - 1360) + i (3830 - 1540)
3 3 D D
+ j (5070 - 2780)
4>3 = $2425
3
Note that £ . = $5070 and that the Shapley value does fall within
the core of the game. In this case the consumer surplus or savings
from having ownership of 140 units is the difference between the
cost of on-site control and the Shapley value or $700, $755, and
$725 for areas 1, 2, and 3, respectively. This is a much more uni-
form distribution of savings than existed before.
Case 2: Ownership of off-site control is unspecified. Coordinators
own the central facility and allocate it in an efficient and equitable
manner.
It has been demonstrated earlier that the optimal solution allocates
100 units to area 1, 270 units to area 2, and 50 units to area 3.
The unit costs of providing this central facility are $4, $3, and
$5, respectively. At this cost, a larger savings accrues to area 2
than to either areas 1 and 3. Furthermore, areas 1 and 3 would be
willing to pay more than $4 to obtain additional units from 2. Thus,
they might feel that the solution is unfair.
Let us explore the possibility of using cooperative N-person game
theory to solve this problem. In Case 1, ownership of the central
facility was prespecified. Thus, the values of c(S) were unambiguously
defined. This corresponded to the second definition of the charac-
teristic function, C2(S), wherein the other coalition (N - S) could
be viewed as not being present.
Let us examine definition one wherein c-^(S) is defined as the value
if S is given preference over N - S. In this case S is allotted as
much of the off-site facility as he desires. The resulting values
of the characteristic function are shown below:
166
-------
C;L(1) - $1200; C;L(2) = $900; (^(3) - $2450
^(12) - $2280; ^(13) - $3780;^ (23) - $3480
c1(123) - $5070
Inspection reveals that this game_is not subadditive since coalitions
are never advantageous, &•$•* c(12) > c(l) + c(2). Thus, this
definition of the characteristics function would not result in
cooperation in this case since coalitions never leave the players
better off.
Let us try the opposite definition wherein N - S is given prefer-
ence over S. The resulting characteristic function values are shown
below:
c4(l) =$2000; c4(2) =$2100; c4(3) =$3150
c4(12) =$3220; c4(13) =$4230; c4(23) =$4470
c4(123) =$5070
Unlike the previous situation this game is subadditive for all coali-
tions. Thus, there is a basis for cooperation.
Indeed one can generalize that in this example coalitions are never
advantageous using c.. (S) and are never disadvantageous using c4(S).
The reason is simple. In the former case, the individual area is
allotted up to the total available storage, W, in determining c(l),
£(2), and c(3). However, as coalitions form he is forced to share
W with other areas. Thus, he is never made better off if an additional
area joins the coalition since he cannot (due to the nonnegativity
restrictions on Q. ) attain more of the central facility than he had
before. At best, ne is indifferent. On the other hand, using c, (S) ,
the area will always consider coalitions since they may thereby be
able to bid for a larger amount of the capacity of the central facility.
Note that c.. (123) = c,(123) so that the value of the grand coalition
is the same. The major difference is in the value of the subcoali-
trons. This analysis illustrates the importance of a fundamental
aspect of resource allocation — the question of ownership. In Case 1
ownership was specified a pvi-oici,. This case leads to a pervasively
cooperative situation since larger coalitions can never be harmful.
Each area has the option of using its allotted share of the off -site
facility or selling it to others. Similarly, in the situation where
(N - S) is given preference, -i.e.3 c, (S), a cooperative situation
results. However, in the case where S is given priority, a pervasively
competitive situation arises.
167
-------
This discussion can be summarized by introducing the notion of
security level which is defined as the value the player can assure
himself if he acts independently. We see that the security level
depends on how the characteristic function is defined. For the
case where ownership is not specified a priori, only definition 4
of the characteristic function guarantees a cooperative solution.
Definitions 1 and 3 lead to strict competition and definition 2 is
irrelevant since one cannot assume the other N - S players are
absent in the case where ownership of the common resource is
unspecified. Thus, we feel that definition 4 provides a reason-
able basis for an equitable cost sharing arrangement. The Shapley
values for this case are:
0 = T (2100) + \ (3220 - 2000) + \ (4470 - 3150)
Zoo o
+ j (5070 - 4230)
cj>2 = $1400
<*>., = T (3150) + \ (4230 - 2000) + \ (4470 - 2100)
job o
+ j (5070 - 3220)
3 = $2437
3
Note the Z = $5070 and the solution is in the core.
The consumer surplus for this case is $767, $700, and $713, respectively.
One could argue that, as long as the calculated consumer surpluses
are almost the same, wouldn't it be fairer to simply equalize the
consumer surplus so that each area saved an equal amount—about $727
per study area in this case? Table 44 shows the charges to each study
area for the above two cases.
In actual situationsone might decide that every area should pay at
least a minimum amount or impose other restrictions. It is straight-
forward to formulate this problem using goal programming wherein one
can minimize the sum of the positive and negative deviations of the
excesses relative to a uniform savings solution (see Charnes and
Cooper (1961) for a description of goal programming).
168
-------
Table 44. CHARGES FOR CENTRAL FACILITY
Total cost
Area
1
2
3
Totals
Allocation
of storage
100
270
50
420
Shapley
$1233
1400
2437
$5070
equal savings
$1273
1373
2424
$5070
On-site
cost
$1000
210
2400
$3610
Charge
Shapley
$2.33
4.41
0.74
__,
$/Qi
equal
savings
$2.73
4.32
0.48
169
-------
The problem of efficient and equitable urban stormwater quality
management has been addressed using recently developed notions from
cooperative N-person game theory. Numerous investigators have demon-
strated the value of optimization techniques for solving the problem
of determining the pollution control strategy which is in the best
overall interest. Unfortunately, such procedures often prescribed
solutions which did not seem fair to everyone. Whether or not the
solution was fair depends on the cost sharing arrangements which
in turn depend on who is entitled to off-site disposal of pollution.
Using a simple example, it has been shown how the optimal solution
is determined using linear programming or a simple market price
determination. Included in this analysis is an explicit procedure
for placing a value on off-site disposal which is based on savings
in alternative on-site control costs. This procedure provides a way
of quantifying a value for the natural system which is the recipient
of these wastes. Difficulties in achieving an equitable solution
were outlined.
Next, introductory concepts from cooperative N-person game theory
were presented. This section indicated how it is possible to devise
management strategies which retain the efficient solution and permit
an equitable cost sharing arrangement to be made. Then two cases
were examined. In the first case, ownership of the off-site facility
was prespecified. Polluters in this situation could utilize their
right or exchange it with others. In the second case, ownership was
not specified. Depending on how one defined the characteristic func-
tion, a competitive or cooperative situation would result. The
results of the analysis provide a way of solving this problem.
170
-------
SECTION VII
HYDRAULIC DESIGN BY THE SWMM
A. METHODS
Current practice in hydraulic design of storm and combined sewer
systems still relies heavily upon the rational method in which peak
flow rates are computed and summed at sewer junctions on the basis
of assumed runoff coefficients, times of concentration and full-flow
pipe capacities (ASCE, 1969). It has been shown that peak flow rates
in excess of those predicted on the basis of steady, uniform flow
can be accommodated by conduits when flow dynamics are included, as
is done in the SWMM (Henderson, 1966). Consequently, a design utiliz-
ing the SWMM can be expected to produce a more efficient (smaller)
conduit size than would a rational method formulation.
A further advantage accrues through the use of the RUNOFF portion
of the SWMM wherein a dynamic routine is used to generate the inlet
hydrographs. Temporal and spatial variations of surface runoff are
fully represented.
The Road Research Laboratory (RRL) method has been successfully
employed in Britain to overcome several of the deficiencies of the
rational method. However, its surface and sewer routing procedures
are less sophisticated than those of the SWMM, and it has been shown
in an EPA study to become less accurate as the percent impervious-
ness decreases (Stall and Terstriep, 1972). This is due to the fact
that the RRL model computes runoff only from directly connected
impervious areas.
A hydraulic design capability has been installed in the SWMM. Basically,
the routine checks for surcharge conditions and, if present, enlarges
the downstream conduit by standard amounts until capacity exists to
accept the flow. In its present form it is best suited for redesign
of existing systems inasmuch as the sewer layout, slopes and first
estimates of sizes must be supplied a pri,OY"i. Revisions are intended
to allow the possibility of upstream storage as an alternative to
increased conduit capacity. Eventually the program may be enlarged
to allow possible changes in slope, if constraints on invert eleva-
tions are supplied. Finally, minimization of excavation is also a
possibility if surface elevations are also supplied.
The "design storm" must be 1) chosen from the historical record; or
2) synthesized from intensity-duration curves and percentage-mass
curves as done by Stall and Terstriep (1972) ; or 3) synthesized by
other means (Sorman and Wallace, 1972). However, it should be
remembered that the return period of the runoff will not necessarily
be the same as that of the rainfall due to differences in antecedent
171
-------
conditions and other factors affecting the catchment from storm to
storm.
B. EXAMPLE
The design method is shown for Storm Study Number 6 oil Stevens
Avenue. Initial conduit dimensions are shown in Table 45. Elements
137 and 143 have been replaced by horseshoe and rectangiar shapes,
respectively, to demonstrate the methodology (see Figure 17 for
sewer layout). Currently, if additional capacity is required,
diameters of circular conduits are increased by three inches if pre-
sently less than three feet or six inches if greater. Rectangular
conduits are increased in width by six inches, under the assumption
that system constraints are more likely to limit vertical dimensions
than lateral. Non-circular or non-rectangular conduits are converted
to equivalent (equal area) circular conduits of standard sizes and
treated as such in subsequent time steps.
Table 46 illustrates the output at each time step. Dimensions are
increased until surcharging no longer exists at a given time step,
and flow routing proceeds under those conditions. The ratio of
required flow capacity to current full-flow capacity is also indicated
so that the importance of the surcharging can be evaluated. If the
ratio is near 1.0 alterations are perhaps unnecessary. For instance,
conduit number 168 appears to be definitely undersized whereas an
enlargement of conduit number 147 may be only of marginal necessity.
Table 47 illustrates the final conduit dimensions. Altered values
are indicated.
An obvious trade-off exists between lack of surcharging with increased
conduit sizes versus lower peak flow rates at the outlet when up-
stream surcharging is allowed. This is illustrated in Figure.,37 in
which the peak outlet flow rate is increased by 75 cfs (2.2 m /sec)
when conduit sizes are increased. Surcharging, or other more acceptable
upstream storage alternatives, may provide valuable relief in reduc-
tion of peak outlet flow rates.
Finally, the interrelationship between surcharging locations and
points of high-inflow volume as indicated by the monitoring routine
(Section IV-G) should be noted. Many of the surcharge locations as
indicated on Table 46 correspond to the inlet locations with high-total
inflow shown on Table 21 (although the examples are for two different
storms). The volumes indicated on either Table 46 or Table 21 could
serve as first estimates of capacities required for upstream storage
facilities.
172
-------
o
o
K)
o
z
r>
->
. o
u. o
U. CM
O
Z
5
<.
Ul
o:
to
z
o
0 o
o
SURCHARGING (REDESIGNED)
WITH SURCHARGING (EXISTING)
v\
^
;
0
6 30
45
60 75
TIME (MINUTES)
FIGURE 37.
Effect of Surcharging and Conduit Alteration on
Stevens Avenue Overflow. Study 1, Storm 6.
173
-------
D
10 YEAR STORM OF AUGUST S, 1971 DURATION 1 HOUR (STORM *6)
NUMBER OF ELEMENTS" 95
NUMBER OF TI""E INT" TO
TIME INTERVAL" 160,0 SECONDS,
|'i ELFMFNT PARAMETERS
i EXT.' TYPE DESCRIPTION
•| ElE,
•' NUM.'
lie
it'
120
1?8
i 1?9
!: 130
;•; ni
!-l ns
i'i 117
\- 138
I'l 1'JO
j>' 103
106
!: i"
1119
'i 150
IS?
153
'•; 151
•; 155
: is?
•i 1*8
•i 1",9
160
161
•i" 162
M 1*3
:.! i*,a
:• 165
••' 166
••'t 167
"1 168
"1 169
V I'l
>i 175
176
-I 179
"i 1"0
,1 181
-1 1*2
.! 103
.'.I 11! 9
•' 191
193
197
••• 198
199
1 CIRCULAR SHAPED
1 CIRCULAR SHAPED
1 CIRCULAR SHAPED
1 CIRCULAR SHAPED
l CIRCULAR SHAPED
C! = riiL»R S»APri
CIRC'JLAK SHAPED
CIRCULAR SHAPED
HORSE SHOE
CIRCULAR SHAPFD
CIRCULAR SHAPED
RECTAM1UI AR
CIRCULAR SHAPED
CIRCULAR 3HAPFO
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPFD
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHIPFD
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPFO
CIRCULAR SHAPFD
CIRCULAR SHAPED
CIRCULAR SHAPt"
CIRCULAR SHAPED
CIRCULAR SHAPFO
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR 3HADED
CIRCULAR S"APEO
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR sml'EO
CIRCULAR SHAPED
CIRCULAR SHAPF.O
CIRCULAR SHAPED
CIRCULAR SHAPED
CIRCULAR SHAPED
SLHPE
(FT/FT)
0,00050
0,03954
0,035011
0,01829
6.0?659
0,00672
0,01618
0,02000
0,01556
0,0085'j
0,03171
0,02305
0,00952
0,01558
0,02768
0.01765
0,03176
0.017P5
0,03871
0,0?35S
0,03161
0.06258
0,061 15
0,03091
0,03161
0,05332
0.01617
0.01306
0.01133
0.02130
O.OH25
0,0'1662
0 ,00781
0,01121
0,05560
0,05196
0,02215
0.02333
O.OP690
0.02172
0.01098
0,03735
0,02533
0,025"»
0,05306
0,01059
0,02000
DISTANCE
(FT)
201.00
159.00
270.00
298.00
270.00
600.00
270.00
lift .00
270.00
128,00
150.00
mo. oo
181,00
166,00
250,00
566,00
120,00
200.00
388,00
312.00
280,00
210.00
.S5B.OO
256.00
250.00
515,00
198,00
85.00
210.00
367,00
255.00
120.00
190.00
271.00
250.00
2«0.00
579,00
219.00
171.00
236.00
255.00
310.00
210.00
205.00
122.00
310.00
250.00
MANNING
ROUGHNESS
0,0130
0.0130
0.0130
0,0130
~n,013o
0.0130
O.OtJO
0.0130
0.0130
0,0130
0,0130
0.0130
0.0130
0,0130
0,0130
0,0130
0,0130
0.0130
0,0130
0.0130
0,0130
0,0130
0,0130
0,0130
0,0130
0.0130
0.0130
0.0130
0.0130
0,0130
0,0130
0,0130
0.0130
0.0130
0.0130
0.0130
0,0130
0.0130
0.0130
0,0130
0,0130
0,0130
0,0130
0,0130
0,0130
0,0130
0,0130
GEOM1 GEOM2
(FT) (FT)
5,000 0,0
5,000 0.0
5,000 0,0
5,000 0.0
~5,l'01 0,0
s.noo o.o
3,500 0,0
3,500 0,0
2,000 0,0
0,670 0.0
2.000 0,0
2,000 1.500
2,000 0,0
1,250 0,0
1,000 0,0
5,000 0,0
3,000 0,0
3,000 0,0
3,000 0,0
2,500 0,0
2.000 0,0
1,000 0,0
1,000 0,0
2,000 0,0
2,000 0,0
2,000 0,0
2,000 0,0
1,500 0,0
1,500 0,0
0,670 0,0
1.000 0.0
1.000 0,0
0,830 0,0
1.COO 0,0
1,000 0,0
1,500 0.0
1,000 0,0
-3,000 0,0
3,000 0,0
3,000 0,0
1 , c o o 0,0
1,500 0,0
1,500 0,0
l.JOO 0,0
2,000 0,0
1,000 0,0
1.500 0.0
OEOM1 NUMBER
(FT) OF
BARRELS
0.0 ,0
0.0
0,0
0.0
0.0
0.0
0.0
0.0
0,0
0,0
0,0
0,0
0,0
0,0
0.0
0.0
0.0
0,0
0.0
0,0
0,0
0,0
0.0
0,0
0.0
0.0
0,0
0.0
0.0
0.0
0.0
0.0
0,0
0,0
0.0
0,0
0,0
0,0
0,0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
.0
.0
.0
,0
.0
,0
.0
.0
.0
.0
.0
,0
.0
.0
,0
.0
.0
.»
.0
.0
,0
.0
,0
.0
.0
.0
.0
.0
.0
,0
,0
.0
.«
.0
.0
.0
.0
.0
,0
.0
.0
.0
.»
.0
.0
.0
»FIILL
(SO. FT)
19,635
19,615
19,635
19.635
19,635"
19,635
9,621
9,621
3.317
0,353
3,112
3,000
3,112
1,227
0,785
J9.635
7.069
7.069
7,069
1,909
3,112
0.785
0,785
3,112
3,112
1,102
3,112
1.767
1,767
0.35J
12,566
0,785
0,511
12,566
12,566
1.767
12,566
7,069
7,069
7,069
0.7S5
1,767
1,767
0,785
3,102
0,785
1,767
oruLL
(CFS)
525. ?«1
S19.006
184,831
511,001
111 ,556
561.157
129.505
102.667
51.610
2.706
12,271
29.671
51,177
7.518
5,913
316,938
121.687
87.837
131,581
61.087
10,329
8,937
8,831
39,899
10.308
5?, 378
29,111
IS. 037
22.176
1.914
152.767
7.713
1,925
302,810
339.710
21.692
215. 80S
102.150
109.689
98,563
T.232
20.355
16.763
S700
52.250
•J.676
11.895
OHAX
(CFS)
567,581
560,959
527,900
551,881
176, 6«l
609,611
139,865
151,080
55,616
2,922
15,651
33,616
50,515
8,152
6,119
371,693
130,662
91,861
142,108
68,134
13.555
9,652
9,511
43,091
13,576
56,569
31,110
13,000
23,950
2,067
160,988
6,330
2,078
127,067
366,919
26,668
233,069
110,322
118,463
106.448
7,810
21,984
18,104
6,156
56,430
3,970
16,087
i,
SUPER. cmncAi !
PLOW HHEN IF8S
THAN 951 FULL? i
YES
YES
YES
YtS
YES ~"
YES
NO
YES
YES
YES
YES
YES !
YES i
NO
YES i
YES
YES
NO
YES
YES
YES
YES
YES i
-YFS
YES
YES ;
NO !
NO :
YES
NO
NO
YES
NO
YES
YES
YES i
YES 1
YES
YES
YFS
Y?S
YES
YES
NO
YtS
N6
NO
EPslLON « 0,000100 "0, OP ITERATlnNS IN ROUTINO ROU1NE « 4
Table 45.
INITIAL DIMENSIONS AND CONDUIT CAPACITIES,
-------
o
U1
.«!
>'j
TI-Ea 180. SECONDS, TIME STEP» 1. ELEMENT 118 SURCHARGING. SURCHARGE OF
RATIO nF REQUIRED FLOH TO PRESENT OFULL" l.Zl
INCREASE DIAMETER OF EXT ELEMENT 138 BY 0,33 TO 1.000 FT,
TI-[. 180. SFCONDS, TIME STEPa 1. FLF"tNT 1S8 SURCHARGING.' SURCHARGE OF
RATIO nF REQUIRED FLOH TO PRESENT SFULLa i.'19
INCREASE DIA-ETER OF EXT ELEMENT 158 BY 0.25 TO 1.250 FT,
Tl-fa 360. SFCONDS, TIME STEPo ?, ELEMENT 138 SURCHARGING, SURCHARGE OF
PATIO OF REOHIBED FLOH TO PRFSENT OFULLa i.ia
INCREASE DIAMETER OF EXT ELEMENT 138 BY 0.25 TO 1,250 FT,
TI"E = 360. SECONDS, TIHF STEP= 2. ELEMENT 107 SURCHARGING. SURCHARGE OF
PATIO OF REOU19ED FLO* TO PRESENT OFUULa 1.11
INCREASE DIAMfcTtfl OF EXT ELEMENT 107 BY 0.25 TO 1,500 FT,
T]Mfi 360. SECONDS, TIME STEPa ?. ELEMENT J09 SURCHARGING,' SURCHARGE OF
BATJO OF REOUIPEO FLOH TO PRESENT OFULLa 2.59
INCREASE DIAMETER OF EXT ELEMENT 109 BY 0.25 TO 1,250 FT,
TI"E= 360. SFCONDS, TIME STEPa ?, ELEMENT 109 SURCHARGING, SURCHARGE OF
RATIO OF RFQIIIPED FLOH TO PRFSENT OFULL" 1.13
INCREASE DIAMETER of EXT ELEMENT 1U9 BY 0.25 TO 1.500 FT,
TI-ta 360. SFCONDS, TIME STEPa 2. ELEMENT 158 SURCHARGING. SURCHARGE OF
RATIO OF BEOUIIEO FLOH TO PRESENT OFULLa 1,62
INCREASE DIAMETER OF EXT ELEMENT 1S8 BY 0,25 TO I,500 FT,
TI"F» 36I>. SFCONDS, TIME STEPa >, ELEMENT 158 SURCHARGING,' SURCHARGE OF
PATT" OF BEQ»I»tD FLOH TO PPFSFNT OFULL« 1.12
TsrsFAU DIAMETER OF EXT ELEMENT 15« BY 0.25 TO 1,750 FT,
TI»Ea 360. SECONDS, TIME STEPa 2, ELEMENT 166 SURCHARGING, SURCHARGE OF
RATin HF REQUIRED FLOH TO PRFSENT OFULLa 2,21
INCBFASE DIAMETER OF EXT ELEMENT 166 BY 0.31 TO 1,000 FT,
TI-F.s 360, SECONDS, TIME STEPa ?., ELEMENT 168 SURCHARGING,
RAT;I nF BEQUIPED FLO* TO PRESENT OFULL= 2,91
INCBEASF DIAMETER OF EXT ELEMENT 168 BY 0.25 TO 1,250 FT,
TI"t= 360. SECONDS, TIME STFPa 2. CLEMENT 168 SURCHARGING. SURCHARGE OF
RATIO OF KF5HIOFD FLO* Tn PRFSENT OFULL0 1,61
INCCTASE DIAMETER OF EXT ELEMENT 161 BY 0.25 TO 1,500 FT,
Tl"ta 360, 3FCONOS, TIME STEPa ?. ELEMENT 183 SURCHARGING,'
RATIO OF SCO'IISEO FLOH TO PRESENT OFULL= I,si
TNCBFAsE DIAMETER 01- EXT ELEMENT 183 BY 0.25 TO 1,250 FT,
TI"E= 360. SFCONDS, TI"E STEPa ?, ELEMENT 198 SURCHARGING, SURCHARGE OF
PATIO PF REQ'IJPED FLOW TO PRESENT OFULLa 1.61 ,
INCBFASE DIAMETER OF EXT ELEMENT 198 BY 0.25 TO 1,250 FT,
TIMF» 360. SECONDS, TIMF STEPa 2. ELEMENT 103 SURCHARGING.'
RATIO nF REn'iiPED FLnn Tn PRFSENT OFULL1 1,70
INCREASE ^IDTH OF EXT FLFMF.NT 103 BY 0,50 To 2,000 FT,
100.03CU. FT. STORED AT UPSTREAH ELEMENT ]«
307.95CU, FT, STORED AT UPSTREAH ELEMENT 5«
19J.3TCU, FT, STORED AT UPSTREAM. ELEMENT 38
1S0.17CU. FT, STORED AT UPSTREAM ELEMENT 4T
1696,J2CU. FT. STORED AT UPSTREAM ELEMENT 4*
eae,«2:u, FT, STORED AT UPSTREAM ELEMENT
239T.25CU, FT, STORED AT UPSTREAM ELEMENT SS
S71.IOCU. FT. STORED AT UPSTREAM ELEMENT ft
OH.21CU, FT, STORED AT UPSTREAM ELEMENT 64
SURCHARGE OF 2654,93CU. FT. STORED AT UPSTREAM ELEMENT 68
IS2S.99CU, FT. STORED AT UPSTREAM ELEMENT «S
SURCHARGE OF 69S.B8CU. FT. STORED AT UPSTREAM ELEMENT 81
003.1SCU, FT. STORED AT UPSTREAM ELEMENT 98
SURCHARGE OF 3T18,61CU, FT. STORED AT UPSTREAM.ELEMENT 03
i"!
Table 46.
HISTORY OF SURCHARGING AND CONDUIT ALTERATION DURING STORM.
-------
TI"E» 360. SECONDS. TIMF STEP. f. ELEMENT HO SURCHARGING. SURCHARGE OF Z9SS.?SCU, FT, STORED AT UPSTREAM ELEMENT
RiTin nF REQUIRED FLOU TD PRESENT OFULL" I.5'
INCOF4SE DIAMETER OF EXT ELE-EHy 1«0 B» 0.25 TO 2.250 PT,
RA'ln OF BEQ'IIOEO FLn« Tf1 PSF3FNT OFULI « 1.55
OE»L»CF EXT FLE«ENT is? BY CIRCULAR cououit OF DIAMETER 2,250 FT,
T!-6e 160. SECONDS, T1"E STtPi J. ELE"ENT 1J7 SURCHARGING. SURCHARGE OF 2H56.50CU, FT, STOREO AT UPSTREAM ELEKEXT ST
OATtn nf REQUIRED FLO«» TP PPFSFNT OFMLLE 1,21
If-C^ElSE DI»UETER OF EXT ELE"E>iT 137 ^T 0.25 TO 2,500 FT,
TI"f= 360. SFCON03, Tint 3TEP» 2. ELEMENT 153 SURCHARGING, SURCHARGE OF 5a«6,T6Cll, FT, STORED AT UPSTREAM ELEMENT 55
PATln pf REOUIPFD FLPX TO PRFSENT OFULL» 1.41
INCREASE DIAMETER OF EXT ELEMENT 151 BY 0.50 TO 3,500 FT,
Table 46 (cont)
-------
FIVAL CONDUIT DIMENSIONS. ASTERISK INDICATES ALTERED CONDUITS,
•n i
FLFMENT
PARAMETERS
E»T,_TYPE DESCRIPTION SLOPE
EU.
V;*.
lie
in
1?0
i?«
1?9
n;
nt
155
137
11!
103
113
116
117
1 19
ISO
IS?
I5J
15a
1S5
1ST
1-iB
159
160
in
U2
163
\m
US
1 ftft
167
16K
1ft'
171
175
176
171)
1*0
181
1"?
1*3
!»9
1°1
1=3
197
19»
1 Of
BARRELS
0.0
0.0
0.0
0.0
0.0
0,0
0,0
0.0
0,0
0.0
0.0
0,0
0,0
0,0
0,0
0,0
0.0
0,0
0,0
0.0
0.0
0,0'
0,0
0.0
0.0
0.0
0.0
0,0
0.0
0.0
0.0
0.0
o.o
0.0
0.0
0.0
0,0
0,0
0,0
0.0
0,0
0,0
0,0
0,0
0.0
0.0
0.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
,0
.0
.'
.0
,0
.0
.0
.0
,0
.0
.0
.0
.0
.0
.0
.0
.0
.»
.0
.0
.«
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
.0
•FULL
fio.rfl
1».635
19,615
19.615
19.615
19,635
19.615
9,621
9.6P1
0,909
1.227
3,976
1,000
3,112
1.767
1.767
19.615
7,069
9,621
7,069
1,909
3,102
2.005
0,7«S
3,112
3.112
3,112
1,112
1,767
1.767
0,785
12,566
1,767
0,511
12.566
12,566
1.767
12.566
7,069
7,069
7,069
,2?7
,767
,767
,785
,102
.227
1,7*7
OFUIL
(CF3T
SZS.511
519.106
aee,«3«
511,001
141,556
560,057
129,505
112,667
87,786
1«.?72
57,»7l
01,818
50,077
12.27S
17,523
316,938
120,687
132,096
131,581
61,087
10.329
59,705
K.B30
39,099
10,308
5?,37«
29,111
1?,037
22.176
5,569
152,767
2?, 702
1,925
30?. BOO
33<>,70(l
20,692
215.805
10?, 150
109,688
91,563
11.112
Jtt.155
16.763
5,700
5?. 250
6,666
11.895
0-AX
8UPt«.C»ITICAL
ccrsl rto* KHp'tTicM '
567.580
560,959
527.900
551.8(11
676.881
609,611
139,865
1511,080
90,80*
15,110
62,501
51,171
51,515
13,256
18.925
370,691
130,662
113,095
J02.108
66,111
13,555
12,921
9,511
13,091
03,576
56,568
31,000
13,000
23,950
6,014
160,988
20,561
2,078
327,067
366,918
26,668
233,069
110,322
118,163
106,108
11,16'
21.9A1
18,101
6,156
56,010
7.199
16,087
THAN »S* FULIT
YES
YES
YES
YES
YtS
YFS
NO
YES
YES
YES
YES
YES
YES
SO
YtS
YES
YES
NO
YES
YES
YES
YES
YES
YES
YES
YES
NO
SO
YES
NO
NO
YES
NO
YtS
YES
YES
YE3
YES
YES
YES
YFS
YES
YES
NO
YES,
NO
NO
Table 47.
FINAL CONDUIT DIMENSIONS AND CAPACITIES.
-------
SECTION VIII
REFERENCES
Agnew, R., and C. A. Hansen. "Biological Treatment of Combined
Sewer Overflows at Kenosha, Wisconsin," Envirex, Inc.
Milwaukee, Wisconsin, 1973.
American Public Works Association. "Water Pollution Aspects of
Urban Runoff," Federal Water Pollution Control Administration
Contract WP-20-15, 1969.
American Society of Civil Engineers. Design and Construction of
Sanitary and Storm Seiners, ASCE Manuals and Reports on Engineer-
ing Practice No. 37, 1969.
American Soceity of Civil Engineers. Design and Construction
of Sanitary and Storm Sewers, Manual of Engineering Practice
No. 37, 1960
APWA Research Foundation. "The Swirl Concentrator as a Combined
Sewer Overflow Regulator Facility," Office of Research and Moni-
toring, US EPA, EPA-R2-72-008, 1972.
AVCO Economic System Corp. "Storm Water Pollution from Urban
Land Activity," EPA Report No. 11034 FKL 07/70, Washington, DC,
1970.
Brandt, G. H. et al. "An Economic Analysis of Erosion and Sediment
Control Methods for Watersheds Undergoing Urbanization," Final
Report, OWRR Contract No. 14-31-001-3392, Dow Chemical Co., Midland,
Michigan, 1972.
Carlisle, A., A. H. F. Brown, And E. J. White. "Litter Fall Leaf
Production and the Effects of Defoliation by Tortrix viridanal
in a Sissile Oak (Quercus Petrala) Woodland," Journal of Ecology,
Vol. 54, 1966.
Chandler, R. F. "The Amount and Mineral Nutrient Content of Freshly
Fallen Leaf Litter in the Hardwood Forests of Central New York,"
Journal of the American Society of Agronomy, Vol. 33, No. 10, 1941.
Chandler, R. F. "The Amount and Mineral Nutrient Content of Freshly
Fallen Needle Litt.er of Some Northeastern Conifers," Proceedings
of Soil Science of America, Vol. 8, 1944.
178
-------
Charnes, A., and W. W. Cooper. Management Models and Industrial
Applications of Linear Programming, Vol. I, John Wiley and Sons,
Inc., New York, 1961.
Chase, E. S., and A. F. Ferullo. "Oxygen Demand of Leaves in
Water," Water and Sewage Works, Vol. 105, No. 5, 1958.
Chow, V. T., Ed., Handbook of Applied Hydrology3 McGraw-Hill
Book Co., New York, 1964.
Clark, J. W., W. Vlessman, and M. J. Hammer. Water Supply and
Pollution Control, International Textbook Company, Scranton, Pa.,
1971.
Coile, T. S. "Composition of the Leaf Litter of Forest Trees,"
J, Elisha Mitchell Science Society, Vol. 52, 1936.
Corps of Engineers, Everglades Gaging Program, Everglades Area,
Progress Report, Jacksonville District, Report No. 6, 1956.
Dantzig, G. B., and P. Wolfe. "Decomposition Principle for Linear
Programs," Operations Research, Vol. 8, No. 1, 1960.
Daubenmire, R. "Nutrient Content of Leaf Litter of Trees in the
Northern Rocky Mountains," Ecology, Vol. 34, 1953.
Drainage Criteria Manual, Denver Regional Council of Governments,
Vol. 1, 1969.
Eckenfelder, W. W., and J. McCabe. Biological Treatment of Sewage
and Industrial Waste, Reinhold Publishing Co., New York, 1956.
Environmental Protection Agency. "Environmental Impact of Highway
De-icing, EPA Report No. 11040GKK06/71, Edison Water Quality Lab.,
Edison, NJ, 1970.
Environmental Protection Agency. "Storm Water Management Model,"
Water Pollution Control Research Series, Washington, DC, 1971:
a. "Volume I, Final Report," Rept. No. 11024DOC07/71
b. "Volume II, Verification and Testing," Rept. No. 11024DOC08/71
c. "Volume III, User's Manual," Rept. No. 11024DOC09/71
d. "Volume IV, Program Listing," Rept. No. 11024DOC10/71
Field R., and E. J. Struzeski, Jr. "Management and Control of
Combined Sewer Overflows," J. Water Pollution Control Federation,
Vol. 44, No. 7, 1972.
179
-------
Franklin Institute Research Laboratories. "Investigation of Porous
Pavements for Urban Runoff Control," EPA Office of Research Monitor-
ing, Rept. No. 11034DUY03/72, Washington, DC, 1972.
Flaxman, E. M. "Predicting Sediment Yield in Western United States,"
Journal of the Hydraulics Division, ASCE, Hy 12, 1972.
"Florida Field Crop Variety Report - 1971," Whitty, E. B. ed. ,
Agronomy Mimeo Report AG 72-51, Institute of Food and Agricultural
Sciences, University of Florida, Gainesville, 1972.
Glover, G. E. "High Rate Disinfection of Combined Sewer Overflow,"
Cochrane Division Crane Co., 1972.
Glover, G. E. "Application of Microstraining to Combined Sewer
Overflow," Cochrane Division Crane Co., 1972.
Gameson, A.L., and R. N. Davidson. "Storm Water Investigations
at Northhampton," Institute of Sewage Purification, Conference
Paper No. 5, Annual Conference, Leandudno, England, 1962.
Gorham, E., and J. R. Bray. "Litter Production in Forests of the
World," Advances in Ecological Research, Vol. 2, 1964.
Gunn, R. , and G. 0. Kinzer. "The Terminal Velocity of Fall for
Water Droplets," Journal of Meteorology, Vol. 6, pp. 243-248, 1949.
Haupt, H. "The Effect of the Decomposition of Natural Organic
Contaminating Matter on the Properties of Surface Waters,"
Nasser, Vol. 7, 1933.
Hayes, A. J. "Studies on the Decomposition of Coniferous Leaf
Litter. I. Physical and Chemical Changes," Journal of Soil
Science, Vol. 16, No. 1, 1965.
Heaney, J. P., and L. C. Doughty. "Optimal Atmospheric Emission
Control in the Wood Pulp Industry," Atmospheric Environment, Vol.
6, 1972.
Heaney, J. P., and W. C. Huber. "Simulation of the Impact of Selected
Water Management Alternatives on Stages in Water Conservation Area 3A,"
Final Report to Central and Southern Florida Flood Control District,
1972.
Heaney, J. P., and R. H. Sullivan. "Source Control of Urban Water
Pollution," Journal of Water Pollution Control Federation, Vol. 43,
No. 4, 1971.
180
-------
Heaney, J. P. "Mathematical Programming Analysis of Regional
Water-Resource Systems," Proc. National Symposium on the Analysis
of Water Resource Systems, AWRA, Denver, Colorado, 1968.
Heaney, J.P. "Mathematical Programming Model for Long-Range River
Basin Planning with Emphasis on the Colorado River Basin," Ph.D.
Dissertation, Northwestern University- Evanston, Illinois 1968.
Henderson, F. M. Open Channel Flow, MacMillan, New York, 1966.
Herzog, H. W. "The Air Diffusion Model as an Urban Planning Tool,"
Sodo-Economic Planning Sciences, Vol. 3, 1969.
Heyward, F., and R. M. Barnette. "Field Characteristics and Partial
Chemical Analyses of the Humus Layer of Longleaf Pine Forest Soils,"
Bulletin of Florida Agricultural Experiment Station, Vol. 302, 1936.
Highway Research Board. Parking Principles, Special Report No. 125,
NAS, Washington, DC, 1971.
Hydrotechnic Corporation. "High Rate Filtration of Combined Sewer
Overflows," US EPA, Rept. No. 11023EY104/72, Washington, DC, 1972.
Kawahara, Teruhiko. "The Return of Nutrients with Litter in the
Forest Ecosystems. II. The Amount of Organic Matter and Nutrients,"
Journal of Japanese Forest Society, Vol. 53, 1972.
Kirshen, P. H., D. H. Marks, and J. C. Schaake. "Mathematical
Model for Screening Storm Water Control Alternatives," Ralph M.
Parsons Lab Rept. No. 157, Dept. of Civil Engineering, MIT, Cambridge,
Massachusetts, 1972.
Kucera, C. "Weathering Characteristics of Deciduous Leaf Litter,"
Ecology, Vol. 40, No. 3, 1959.
Laws, J. 0., and D. A. Parsons. "The Relation of Rain Drop Size to
Intensity," Trans. Amer. Geophy. Union, Vol. 24, p. 452-459, 1943
Leopold, L. B. "The Hydrologic Effects of Urban Land Use," Circular
554, US Geological Survey, Washington, DC, 1968.
Leopold, L. B., F. E. Clark, B. B. Hanshaw, and J. R. Balsley,
"A Procedure for Evaluating Environmental Impact," Circular No.
645, US Geological Survey, Washington, DC, 1971.
Loehman, E., and A. Whinston. "A New Theory of Pricing and Decision-
Making for Public Investment," The Bell Journal of Economics and
Management Science, Vol. 2, No. 2, 1971.
181
-------
Luce, R., and H. Raiffa. Games and Decisions, John Wiley and Sons,
Inc., New York, 1957.
Lutz, H. J., and R. F. Chandler. Forest Soils,, John Wiley and
Sons, Inc., New York, 1946.
Malcolm, R. L., and R. J. McCracken. "Canopy Drip: A Source of
Soil Organic Matter for Mobilization of Iron and Aluminum," Soil
Science Society of America, Proceedings, Vol. 32, No. 6, 1968.
Maryland Water Resources Administration. "Technical Guide to Erosion
and Sediment Control Design," Maryland Department of Natural Resources,
September 1973.
Mayer-Krapoll, H. "The Use of Commercial Fertilizers—Particularly
Nitrogen—in Forestry," Translated and Published by Nitrogen Division,
Allied Chemical and Dye Corp., New York, 1956.
McGriff, Jr., E. C. "The Effects of Urbanization on Water Quality,"
Journal of Environmental Quality, Vol. I, No. 1, 1972.
Melin, E. "Biological Decomposition of Some Types of Litter from
North American Forests," Ecology, Vol. II, No. 1, 1930.
Metcalf, L., and H. P. Eddy. American Sewerage Practice, Vol. Ill,
McGraw-Hill Book Company, Inc., 1935.
Meyer, L. D., and L. A. Kramer. "Erosion Equations Predict Land
Slope Developments," Agricultural Engineer, Vol. 50, 1969.
Moir, W. H., and H. Grier. "Weight and Nitrogen, Phosphorus,
Pottasium, and Calcium Content of Forest Floor Humus of Lodgepole
Pine Stands in Colorado," Soil Science Society of America, Pro-
ceedings, Vol. 33, 1969.
Morrison, J. R. "Variation with Crown Postion and Leaf Age in
Content of Seven Element in Leaves of Pinus Bariksiana," Canadian
Journal of Forest Research, Vol.2, 1973.
Mutchler, C. K., and R. A. Young. "Soil Detachment by Raindrops."
Presented at Sediment Yield Workshop, USDA Sedimentation Laboratory
Oxford, Mississippi, Nov. 28-30, 1972.
National Association of Home Builders Research Fouridation, Inc.
A Manual for Residential Street Development Standards, Interim
Copy, Rockville, Maryland, 1973.
182
-------
O'Conner, D. J., and W. E. Dobbins. "Mechanism of Reaeration in
Natural Streams," Trans. ASCE, Vol. 123, 1958.
Owen, G. Game Theory, W. B. Saunders Co., Philadelphia, Pennsyl-
vania, 1968.
Paquette, R. J., N. Ashford, and P. H. Wright. Transportation
Engineering, Ronald Press Co., New York, 1972.
Perez, A. I. "A Water Quality Model for a Conjunctive Surface-
Groundwater System," Ph.D. Thesis, Department of Environmental
Engineering Sciences, University of Florida, Gainesville, 1972.
Perez, A. I., W. C. Huber, J. P. Heaney, and E. E. Pyatt. "Water
Quality Model for Conjunctive Surface-Groundwater System," Water
Resources Bulletin, Vol. 8, No. 5, 1972.
Poertner, H. G. "Better Storm Drainage Facilities - At Lower Cost,"
Civil Engineering, ASCE, October 1973.
Ports, M. A. "Use of the Universal Soil Loss Equation as a Design
Standard," ASCE Water Resources Engineering Meeting, Washington,
DC, 1973.
Pyatt, E. E., J. P. Heaney, G. R. Grantham, and B. J. Carter.
"A Model for Quantifying Flow Augmentation Benefits," Final Report
to FWPCA, 1969.
Rapoport, A. N-Person Game Theory, University of Michigan Press, 1970.
Rex Chainbelt, Milwaulkee, Wisconsin. "Screening/Flotation Treatment
of Combined Sewer Overflows," US EPA, Rept. No. 11020FPC01/72,
Washington, DC, 1972.
Rohrer Associates, Inc. "Underwater Storage of Combined Sewer
Overflows," US EPA, Rept. No 11022ECU09/71, Washington, DC, 1971.
Ruelke, 0. C., and G. M. Prine. "Performance of Six Hybrid Bermuda-
grass, Pangola Digitgrass, and Pensacola Bahiagrass at Three Fertility
Levels in North Central Florida," Soil and Crop Science Society of
Florida, Proceedings, Vol. 31, 1971.
Ruelke, 0. C., and G. M. Prine. "Preliminary Evaluation of Yield
and Protein Content of Six Hybrid Bermudagrass, Pensacola Bahia-
grass, and Pangolagrass under Three Fertilization Regimes in North
Central Florida," Soil and Crop Science Society of Florida,
Proceedings, Vol. 28, 1968.
183
-------
Sartor, J. D., and B. B. Gail. "Water Pollution Aspects of Street
Surface Contaminants," Office of Research and Monitoring, EPA,
Kept. No. EPA-R2-72-081, Washington, DC, 1972.
Slack, K. V., and H. R. Feltz. "Tree Leaf Control on Low Flow Water
Quality in a Small Virginia Stream," Environmental Science and Tech-
nology, Vol. 2, 1968.
Slack, K. V. "Effect of Tree Leaves on Water Quality in Cacapon
River, West Virginia," U. S. Geological Survey Professional Papers,
No. 475-D, 1964.
Smith, R. "Preliminary Design of Simulation of Conventional Waste
Water Renovation Systems Using the Digital Computer," U.S. Dept.
of the Interior, FWPCA, 1968.
Soil Conservation Service. SCS National Engineering Handbook,
Section 4—Hydrology, USDA, Washington, DC, 1972.
Sorenson, S. W. "A Mathematical Theory of Coalitions and Competition
in Resource Development," Ph.D. Dissertation, University of Texas
at Austin, 1972.
Sorman, V. A., and J. R. Wallace. "Digital Simulation of Thunder-
storm Rainfall," Environmental Resources Center Rept. ERC-0972,
Georgia Institute of Technology, Atlanta, 1972.
Stall, J. B., and M. L. Terstriep. "Storm Sewer Design—An Evaluation
of the RRL Method," EPA, Office of Research and Monitoring, Rept.
No. EPA-R2-72-068, Washington, DC, 1972.
TRW. Air Quality Display Model, USDHEW, PHS. Pub. No. PB 189-194,
1969.
Urban Land Institute. The Community Builders Handbook, Washington,
DC, 1968.
White, E. M. "Water-Leachable Nutrients from Frozen or Dried Prairie
Vegetation," Journal of Environmental Quality, Vol. 2, No. 1, 1973.
Williams, J. R., and H. D. Berndt. "Sediment Yield Computed with
Universal Equation," Journal of Hydraulics Division,HY 12, ASCE,
December 1972.
Wilson, A. L. "An Atmospheric Diffusion Model for Jacksonville,
Florida," M.S. Thesis, Department of Environmental Engineering
Sciences, University of Florida, Gainesville, 1971.
184
-------
Wischmeier, W. H., C. B. Johnson, and'B. U. Cross. "A Soil Erodibility
Nomograph for Farmland and Construction Sites," Journal of Soils
and Water Conservation., Vol. 26, 1971.
Wischmeier, W. H., and D. D. Smith. "Predicting Rainfall-Erosion
Losses from Cropland East of the Rocky Mountains," Agr. Handbook
282, US Dept. of Agriculture, Washington, DC, 1965.
Wischmeier, W. H., and D. D. Smith. "Rainfall Energy and Its Rela-
tionship to Soil Loss," Transactions, American Geophysical Union,
Vol. 39, No. 2, 1958.
185
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-670/2-75-022
3. RECIPIENT'S ACCESSION'NO.
4. TITLE AND SUBTITLE
Urban Storrawater Management Modeling and Decision-;
Making
5. REPORT DATE
May 1975; Issuing Date,
6. PERFORMING ORGANIZATION CODE
T.AUTHORIS) Jameg p.Heaney, Wayne C. Huber, Hasan Sheikh,
Miguel A. Medina, J. Robert Doyle, W. Alan Peltz and
John E. Darling
8. PERFORMING ORGX
9. PERFORMING ORG -\NIZATION NAME AND ADDRESS
Department of Environmental Engineering Sciences
University of Florida
Gainesville, Florida 32611
10. PROGRAM ELEMENT NO.
1BB034 ROAP:ATA Task:015
11. OONTRAOT/GRANT NO.
802219 (11023 GSC)
12. SPONSORING AGENCY NAME AND ADDRESS
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
Prepared in cooperation with the City of Lancaster, Pennsylvania.
16. ABSTRACT
The purposes of this study were to test, refine and augment the capabilities
of the EPA Storm Water Management Model (SWMM), and to develop decision-making capa-
bilities for use in the study of urban stormwater runoff problems.
With regard to the SWMM the numerous programming refinements were introduced to
provide greater reliability and flexibility. A sediment prediction capability has
been incorporated directly into the SWMM. Detailed testing was conducted in Lan-
caster, Pennsylvania to demonstrate the ability of the SWMM to describe the
relatively complex phenomena occurring in an urban catchment.
A systematic procedure is presented for examining the stormwater management
problem in the broader context of urban water resources management. Related stand-
ards for flood control and drainage, street and parking lot design, etc., are
reviewed and suggestions presented regarding modifications in practices which would
ameliorate stormwater problems. An optimization procedure is described which
addresses the related problems of finding efficient and equitable control strategies.
Lastly, the use of the SWMM for preliminary hydraulic design of sewer systems is
described.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS \C. COSATI Field/Group
Water storage, Simulation, Rainfall, Run-
off, Vegetation, Waste treatment, Erosion
control, Optimization, Cost analysis,
Storm sewers, Combined sewers, Water pollu-
tion, Computer programs, Stream pollution
Lancaster (PA), *Water
quality control, *Com-
puter model, *Urban
stormwater, Rainfall run-
off relationships, Cost
sharing analysis,
Sewerage
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (This Report)
UNCLASSIFIED
21. NO. OF PAGES
198
20. SECURITY CLASS (This page)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
186
• U.S. GOVERNMENT PRINTING OFFICE: 1975-657-5q2/"A<;
------- |