EPA-670/2-75-065
June 1975
Environmental Protection Technology Series
Short Course Proceedings
WJCAT30NS CF
STORMWATER MANAGEMENT MODELS
National Environmental Research Center
Office of Research and Development
U.S. Environmental Protection Agency
Cincinnati, Ohio 45268
-------
EPA-670/2-75-065
June 1975
Short Course Proceedings
APPLICATIONS OF STORMWATER MANAGEMENT MODELS
August 19-23, 1974
University of Massachusetts/Amherst
Edited By
Francis A. DiGiano
Peter A. Mangarella
University of Massachusetts
Amherst, Massachusetts 01002
Project No. 803069
Program Element No. 1BB034
Project Officer
Chi-Yuan Fan
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.
-------
FOREWORD
Man and his environment must be protected from the adverse
effects of pesticides, radiation, noise and other 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 multidisciplinary focus through programs engaged in
o studies on the effects of environmental contaminants on
man and the biosphere, and
o a search for ways to prevent contamination and to recycle
valuable resources.
This five day short course was designed for engineers and
technical personnel associated with consulting, regulatory, govern-
mental, and educational organizations with an interest and responsi-
bility in the control of storm and combined sewer overflows. The
Short Course format provided for lecture and workshop sessions on the
development and use of computer assisted models (with special emphasis
on the EPA Stormwater Management Model) for the prediction and control
of urban stonnwater quantity and qualjty.
A. W. Breidenbach, Ph.D.
Director
National Environmental
Research Center, Cincinnati
iii
-------
ABSTRACT
This Short Course was sponsored by the U. S. Environmental
Protection Agency. The specific objectives were to encourage the
consulting profession to implement stormwater models in solving the
problem of storm and combined sewer overflows and to make state
pollution control agencies aware of this tool in their pollution
abatement efforts. Emphasis was placed on presentations of various
types of models, their data requirements and case studies of their
use. The EPA Stormwater Management Model (SWMM) was highlighted.
It is hoped that this compilation of instructional papers, prepared
by the Short Course faculty, will enable practicing engineers to
broaden their use of stormwater management models.
The Short Course was held at the University of Massachusetts
from August 19 to August 23, 1974. Registration totaled 81 with
representation by consultants; Federal, State and Municipal engineers,
including the Canadian government; and University researchers.
This report was submitted in partial fulfillment of Project
Number 803069 by the Department of Civil Engineering at the
University of Massachusetts, under the sponsorship of the Environmental
Protection Agency.
iv
-------
ACKNOWLEDGEMENTS
We would especially like to thank Mr. Richard Field, Chief and
Mr. Chi-Yuan Fan, Project Officer of the Storm and Combined Sewer
Section and Mr. Harry Torno, Staff Engineer of the Municipal Pollution
Control Division of the EPA for their cooperation throughout the
planning and execution stages of this Short Course. Their help in
establishing the program and topics of the Short Course has been of
immeasurable value.
We would also like to thank those speakers who participated in
this Short Course. Special recognition is to be given to
Mr. Thomas K. Jewell, graduate research assistant, and Mr. David Gaboury,
undergraduate research assistant, for their tireless effort in applying
the EPA SWMM to the UMASS computer system and in developing a case
study for Greenfield, Massachusetts.
We are also grateful to the staff of the Continuing Education
Conference Management Center for handling all of the Short Course
details. Finally, we acknowledge the help of our Environmental
Engineering secretarial staff, Miss Dorothy A. Blasko and
Ms. Phyllis L. Mayberg, for their help in proofing the lecture materials.
Francis A. DiGiano
Short Course Co-Chairman
Peter A. Mangarella
Short Course Co-Chairman
University of Massachusetts/Amherst
-------
CONTENTS
Page
Abstract iv
Acknowledgements v
Short Course Faculty viii
Keynote Address 1
Bernard B. Berger
Introduction 6
Richard Field
Environmental Impact of Urban Stormwater Runoff 11
Murray B. McPherson
Quantity Aspects of Urban Stormwater Runoff 83
David E. Winslow
William H. Espey
Quality Aspects of Urban Stormwater Runoff 138
Larry A. Roesner
Impact of Stormwater Runoff on Receiving Water Quality . . 159
Larry A. Roesner
Introduction to Urban Stormwater Runoff Models 177
Robert P. Shubinski
Simplified Methods of Computing the Quantity of Urban
Runoff 200
Robert P. Shubinski
The WRE Model 214
Robert P. Shubinski
The EPA Stormwater Management Model 242
Wayne C. Huber
Decision-Making for Water Quantity and Quality Control . . 247
James P. Heaney
VI
-------
Page
Selected Case Studies Using Stormwater Management
Models - Quantity Aspects 250
Gerald T. Qrlob
Planning a Study Using Stormwater Management
Models 269
Gerald T. Orlob
Three Case Studies on the Application of the
Stormwater Management Models 280
Jekabs P. Vittands
Comparative Analysis of Urban Stormwater Models , . , 333
A!bin Branstetter
List of Participants ,,,,,,,,,,,,,,,«, 421
vii
-------
SHORT COURSE FACULTY
Bernard B. Berger, Director
Water Resources Research Center
University of Massachusetts
Albin Brandstetter, Ph.D.
Environmental Resources Planning
Section
Battelle Pacific Northwest Laboratories
Richland, Washington
Francis A. DiGiano, Ph.D.
Short Course Co-Chairman
Assistant Professor of Civil Engineering
University of Massachusetts
Richard Field, Chief
Storm § Combined Sewer Section
U. S. Environmental Protection Agency
James P. Heaney, Ph.D.
Associated Professor of
Environmental Engineering Sciences
University of Florida
Wayne C. Huber, Ph.D.
Associate Professor of
Environmental Engineering Sciences
University of Florida
Thomas K. Jewell
Graduate Research Assistant
Department of Civil Engineering
University of Massachusetts
Peter A. Mangarella, Ph.D.
Short Course Co-Chairman
Assistant Professor of Civil Engineering
University of Massachusetts
Murray B. McPherson, Director
ASCE Urban Water Resources Program
Gerald T. Orlob, Ph.D
Senior Partner
G. T. Orlob § Associates
Orinda, California
Larry A. Roesner, Ph.D.
Principal Engineer
Water Resources Enginners, Inc.
Walnut Creek, California
Robert P. Shubinski, Ph.D.
Vice President
Water Resources Engineers, Inc.
Springfield, Virginia
Jekabs P. Vittands, Project Mgr.
Metcalf § Eddy Consulting
Engineers
Boston, Massachusetts
David E. Winslow, Manager
Espey-Huston and Associates, Inc.
Houston, Texas
viii
-------
A GENERAL COMMENT ON STORMWATER MANAGEMENT MODELS
By
Bernard B. Berger*
He shall examine this week two sets of procedures concerned with urban
storm flow. I refer first to procedures for predicting the quantity
of urban runoff, given data on frequency, intensity and duration of
storms traversing the urban area; and, second, to procedures for
predicting the mass of specific pollutants conveyed by the resulting
runoff to points of disposal. The first set, with key word "quantity",
represents a general methodology that has long been familiar to the
civil engineer-hydrologist. The second, with key word "quality",
represents a technique that is still experimental and uncertain reflec-
ting as it does (1) community values and habits in terms of unrecorded
and often undetected wastes in addition to street debris and dirt;
(2) transport properties of sheets of surface water in terms of pollu-
tant pickup and conveyance; and, (3) mass transport of pollutants
contained in mixtures of sewage and storm runoff that constitute com-
bined sewer overflow. Management of "quality" requires in some degree
management of the storm flow.
Were we interested in flow per se, our effort would obviously be directed
only to flood prediction and control - historically and traditionally,
a major area of civil engineering concern. But our interest is, of
course, not so limited. We proceed under a compulsion to cleanse our
streams and keep them clean.
The problem of water quality deterioration produced by storm runoff
*Director, Water Resources Research Center, University of Massachusetts/
Amherst.
-------
is widely experienced, and it is severe. It applies to rural areas and
to urban areas, to lands developed by man and to undeveloped lands.
We recognize, however, that the urban area presents a special case
because its impervious surfaces and its conduits increase, quicken,
and concentrate storm flow, and because such flow encounters and picks
up a variety of substances that may be grossly pollutional.
As engineers and environmentalists we seek methods for protecting our
water resources against such damaging substances. In so doing we are
bound to ask such questions as:
- What is the nature of these materials and where and how do they
originate?
- How do their effects vary with frequency, intensity and duration
of storms?
- When is it desirable to capture storm runoff so that controls may
be applied to protect the quality of receiving waters?
- What types of controls are feasible and what conditions apply to
their use?
- To what extent may such materials be controlled through improved
programs of community cleanliness, and through land use regulation
and zoning?
- When is a sewer separation program a reasonable solution for the
combined sewer overflow problem?
These and similar questions are of compelling importance in view of the
goals of the Water Quality Act of 1972 and its requirements for highly
effective end-of-pipe waste treatment technology. However, answers to
these questions are useful only if such information can be applied
dependably to predictions on the resulting quality of the receiving
water. It is, therefore, appropriate that we examine cause and effect
relationships in this context of water quality management. In doing
this we must also ask: can we with confidence, determine the magnitude
of improvement in water quality that would result from a specific
-------
Investment in control of an urban non-point source of pollution? Is
it possible to design a program of control that is optimally balanced
between economic and environmental consideration?
We may ask many additional questions, all of them important. To
answer them we must have available a predictive system which encompasses
all significant variables; fitting and weighting them in a way that
provides a reliable picture of consequences of an intricate set of
pollutional stresses and correctives. The development of such a pre-
dictive model would appear to be as challenging an assignment as any
yet faced by the civil-environmental engineering profession. Such
systems have, in fact, been proposed. We shall examine these systems
during this week's short course.
It is surprising to me that we have been so slow in recognizing the
need for a reliable storm water quality prediction and management model.
Perhaps the reason lay in the assumption that pollution by storm waters
hardly seemed important in view of existing, extensive discharges of
raw sewage and primary plant effluents. But the fact is we were slow
to comprehend the full extent of pollution by combined sewer overflows.
This was so in the face of our knowledge that major interceptors
installed three-quarters of a century or longer ago were fixed in
carrying capacity while the contributing areas and population expanded
enormously. We have also been slow in recognizing the limitation of
separate storm sewers in protecting the quality of receiving waters.
The frequency of combined sewer overflows came to our attention quite
late, actually in 1947 when McKee reported the results of his analysis
of Boston's sewerage system in the Journal of the Boston Society of
Civil Engineers. According to McKee, overflows occurred 5-7 times
monthly or whenever an intensity of precipitation greater than 0.03"
per hour was experienced, each increment of 0.01" per hour being equiva-
lent to the normal dry weather flow. This finding was important, of
course, and it stimulated general interest. But even so, the connection
-------
between overflow and Its pollutional impact appears not to have been
fully recognized at that time.
This association was made a few years later when studies in Detroit
showed that large quantities of organics deposited in the sewer at times
of dry weather flow were scoured and discharged in overflows during
periods of storm. The Detroit study was found to reflect conditions
1n many other cities served by combined sewer systems. In 1951,
Thomas Camp summarized the situation prevailing in most U.S. cities
as follows: "Overflows of mixed sewer and storm water occur nearly every
time it rains, or about 5 to 7 times per month on the average. But
in those 5 to 7 times per month, as much as 1/3 of the sewage solids
produced by the tributary population over a year's time is discharged
without treatment into the receiving waters, carrying large numbers of
intestinal bacteria and viruses while discharging only about 3% of a
year's production of liquid sanitary sewage. The reason for this
apparent anomaly is that the flow capacity of combined sewers ranges
from about 50 to 200 times the average flow of sanitary sewage. So
during dry weather the velocity is so small that sewage solids settle in
the sewer and very small rain storms will suspend these materials."
Perhaps I belabor the obvious in this citation. We are now, of course,
thoroughly informed of the importance of combined sewer overflows and
the need for their control.
There is still one additional question we might ask: why were we so
slow in recognizing the potential pollutional importance of urban storm
runoff apart from combined sewer overflows? It was not until the late
1950's that investigations in the U.S. established the substantial con-
tribution of pollution by surface wash from urban areas. Of special
interest was the observation that pollutants occasionally persisted
without appreciable diminution during the course of the storm. In
passing we may note that similar studies were undertaken in Moscow
and Leningrad in the 1930's.
-------
An Increasing number of investigators are now contributing to our
understanding of the problem. We believe we may now see the picture in
reasonable perspective. However, it is important that we ask what, if
anything, have we still failed to recognize? The magnitude of invest-
ments that will be made to control pollution by urban storm runoff
makes this question urgent. Perhaps during this week of evaluation we
may detect areas of useful inquiry that remain to be explored,
I wish to reemphasize my conviction, shared by many if not all of you,
that without an urban stormwater quality prediction model our efforts
to keep our waters clean can be only partially successful at best, and
at worst - ineffective and highly costly. We would all also agree,
I'm sure, that models proposed for general use should be tested
thoroughly in demanding field situations. Each of us has an opportunity
during this week to examine the evidence, to put probing questions to
those who have developed and tested the models, and to identify flaws
in system design and testing procedure. This short course, therefore,
has a dual purpose: (1) to communicate to you important information
relative to a new engineering technology, and (2) to derive a feedback
to help strengthen this technology. Let us now listen with care, probe
and comment constructively, and learn.
-------
INTRODUCTION
By
Richard Field*
This is the first U.S. Environmental Protection Agency (EPA) sponsored
course on urban stormwater modeling; and a good way to start would be
to briefly introduce you to the EPA Storm and Combined Sewer Technology
Program, and how mathematical modeling is an important integral part of
that Program.
Storm-generated discharges have come to be recognized as one of the major
water pollution problems in urban areas. In an early U.S. PHS report
published in 1964, tfie nationwide significance of pollution caused by
storm generated discharges was first identified. This problem was
acknowledged ten years ago by the Federal government when Congress
authorized funds under the Water Quality Act of 1965, PL89-234, for the
research, development and demonstration of methods for stormwater manage-
ment. The Storm and Combined Sewer Overflow Pollution Control Program,
then under the U.S.. Public Health Service and now part of the EPA, was
designated to carry out this legislative mandate. The effort is now directed
by our Section located in Edison, New Jersey.
Up to the present time more than 140 projects totaling approximatley $90
million have been awarded. EPA's share is about $46 million. Our present
FY 75 extramural program budget is $1.2 million.
*Chief, Storm & Combined Sewer Section, Advanced Waste Treatment Research
Laboratory, National Environmental Research Center-Cincinnati, U.S.
Environmental Protection Agency, Edison, NJ 08817.
-------
EPA STORM WATER MANAGEMENT MODEL
One of the major objectives of the EPA Program is the further development
and refinement of mathematical models for the total management of sewer
systems which consider the full range of wet- and dry-weather flow pollu-
tion control.
In recent years, engineers have found that the traditional analytical
methods are inadequate to evaluate the dynamic reaction of the urban drain-
age system to storm events. It was recognized that a mathematical computer
simulation program to model urban runoff quantity and quality during the
storm process would provide an invaluable tool for engineers. Such a
simulation system was completed in 1971. It is known as the EPA Storm
Water Management Model having the acronym SWMM, pronounced "swim".
SWMM is capable of being used in detailed master planning as it enables the
decision maker to learn the consequences of alternative courses of action.
For example, if storage facilities are being considered, the planner can
calculate the hydraulic and the pollutant load that the facility will receive.
Similarly, the pollutional load imposed on streams by storms can be calcu-
lated. SWMM is also useful in predictingand planning required changes in
the hydraulic conveyance capacities of the collection system, brought
about by local overloading. For design purposes, the model is useful in
determining the anticipated flowrates that a specific sewer needs to convey.
The size and slope of the sewer can then be specified. Although the model's
operational capability has not yet been applied, there is a likelihood it,
-------
or other versions, will be used when storage and treatment works are
more widely used in storm and combined sewer systems. The model's
operational capability lies in its ability to provide the system operator
with information of what to expect at various points in the system in
terms of flow and quality from a rainfall event.
The specific purpose of the course was to introduce the concepts and
modeling techniques associated with the prediction, control, and manage-
ment of stormwater discharges and combined sewer overflows to participants
from government agencies and consulting engineering firms. The participants
were specialists in water quality systems planning, design, and management.
Specialists who participated in the development of SWMM and other persons
who have been active in applying the model or in developing additional
options to extend the range of applicability for SWMM or other stormwater
models, presented lectures at the course. Participants in the course are
expected to improve their planning, design, and management abilities to
apply stormwater management models upon completion of the course.
The program is now in the initial phase of demonstrating the ability of
computer assisted mathematical models to enhance urban stormwater manage-
ment by analyzing a major combined sewer system, by selecting and designing
control and treatment approaches based on cost effectiveness, and by
designing a computerized means of overall management of the system during
storm flows. It is the eventual goal to handle all wet- and dry-weather
flows in this manner.
8
-------
EPA FILM "STORMWATER POLLUTION CONTROL: A NEW TECHNOLOGY"
The film1 portrays a complete overview of the U.S. EPA's involvement in
developing countermeasures for pollution from combined sewer overflows
and stormwater discharges.
The film starts out with a background of sewer construction which logically
leads into the pollution problems caused by urban runoff. Early drainage
plans made no provisions for storm flow pollution impacts. Various
countermeasures are shown that are ready for implementation by munici-
palities. The majority are full-scale operations. They include:
(a) control techniques, e.g. infiltration prevention, street cleaning,
porous pavement (for runoff attenuation, in-sewer routing and storage,
improved overflow regulators, and off-line storage; (b) physical, physical-
chemical, disinfection and biological treatment shown as adjuncts to the
sanitary plant or as remote "satellite" facilities at the outfall; (c) and
various schemes which reclaim stormwater for beneficial purposes including
enhancement of visual aesthetics, recreation and water supply.
A couple of years ago we felt that the time was right to make an evaluation
of previous program efforts, and we implemented a contract to develop the
state-of-the-art and assess techniques available to manage and treat
combined sewer overflow and stormwater. A text^ on this subject has been
"Stormwater Pollution Control: A New Technology," U.S.Environental Protection
Agency. A 16mm sound-color film available from the National Audivisual
Center (GSA), Washington,DC 20409.(Prices: Purchase - $119.50; Rent - $12.50)
2
"Urban Stormwater Management and Technology: An Assessment," EPA-670/2-
74-040, U.S.Environmental Protection Agency. Available from the Storm &
Combined Sewer Section, Advanced Waste Treatment Research Laboratory,
Edison, NJ 08817.
-------
completed and will be available shortly. Expanding on the cliche "a
picture is worth a 1,000 words," we felt a film would be worth 100,000—
so a film was included as part of the work.
The specific reasons for this film were to alert the environmental
engineering communityto the stormwater problem and the immediate need to
apply available technology to counteract it.
10
-------
ENVIRONMENTAL IMPACT OF URBAN STORMWATER RUNOFF
By
Murray B. McPherson*
INTRODUCTION**
All the basic engineering methods for in-stream flood regulation
are applicable to flood mitigation works within local urban catchments,
but generally on a smaller scale: acceleration of flood flows by
canalization through threatened reaches with consequent reduction in
stages; isolation of contiguous land in threatened reaches from
flood flows by means of embankments, such as levees and flood walls;
and the attenuation of flood peaks by means of storage, located
either upstream from threatened reaches or at lateral points fed by
diverted flows. Similarly, possibilities also exist for social-
economic-administrative-legal controls to preclude some degree of
damages, such as flood-plain zoning, flood-proofing of structures,
flood insurance and related schemes.
Flow in urban drainage conduits is principally by gravity. As for
natural drainage basins, smaller sewer branches unite with larger
branches, and so on, until a main sewer is reached. The smallest
catchment area, on the order of a fraction of a hectare in size, is
that tributary to a street inlet. For most smaller tributary areas
in the upper reaches of an urban drainage system, the time required
to reach peak runoff after the beginning of a storm is a matter of
only a very few minutes. Hence, high-intensity, short-duration convec-
tional rainfall is normally the main type of precipitation contributing
to the largest runoff rates (in the majority of U.S. metropolitan areas).
However, magnitudes of pollution loads are also a function of land-
management practices and antecedent precipitation, and maximum loads
might not coincide with maximum runoff. Further, because all pollution
loads are of concern, cyclonic storms are also relevant.
*Director, Urban Water Resources Program, ASCE
**From "Urban Runoff," by Murray B. McPherson, Chapter 1, Part III, "Illu-
strative Special Topic Studies," in Hydrological Effects of Urbanization:
Environmental Impact, an IHO report to be published by Unesco in August, 1974.
11
-------
Rainfall interception by vegetation seldom has an important effect
on urban runoff magnitude. Most soil infiltration-capacity curves approach
a steady, minimum rate after one or two hours, and the capacity of a
vegetal cover may be several times as great as for a bare soil. Antecedent
precipitation can affect soil infiltration capacity, but very few quantita-
tive data are available for evaluating this factor.
Some of the precipitation which reaches roofs, pavements and pervious
surfaces is trapped in the many shallow depressions of varying size and
depth present on practically all urban surfaces. There have been no field
measurements of depression storage because of the obvious difficulties in
obtaining meaningful data.
The residual or excess precipitation remaining after infiltration and
depression storage have been abstracted is available to detention, the
storage effect due to overland flow in transit. Overland flow refers to
surface runoff across a sloping-plane surface at unsteady state, which
occurs over extensive portions of an urban area in conjunction with storm
occurrences. Overland flow from the land is usually collected in street
gutters or ditches, which in turn are drained by street inlets. Additional
storage effects occur in the gutters or ditches. Considerable research
attention has been accorded the development of inlet hydrographs, but
progress has been handicapped by a scarcity of suitable field data.
Flow from the ground surface enters underground systems of conduits at
street inlets. The volume of detention in a conduit can effect a reduction
in the peak rate of flow of an input hydrograph in the same basic way as
any detention storage attenuates an inflow hydrograph, and storage routing
hydrographs are used. Advances in conduit routing capability are outpacing
available reliable data for monitoring the suitability of techniques de-
veloped. This is not to say that the advances have not been significant,
but only to emphasize a disadvantage that is encountered universally.
Shown schematically in Figure r ' is a simplified description of the major
12
-------
INPUT
SURFACE
WATER
DRAINAGB
OJ
GROUNDWATER
SUPPLY
PRECIPITATION
1
J*
•
I
OUTSIDE
FLOWS
.
FLOW
OVER
LAND TO
STREET
INLETS ON
SEWERED
CATCH-
MENTS
•i
-*
i
FLOW OVER UNSEWERED LAND
J 1
UNDER-
GROUND
COLLECTION
SYSTEMS
(SEWERS
AND
UNDER-
GROUND
STORAGE)
i
SURFACE I MAN-MADE]
STORAGE j OPEN j
(PONDS, | DRAlivAGE j
LAKES, | CHANNELS j
DELIBERATE! AND
1
^j USE OF I^J IMPROVED j ^
PARKS, 1 NATURAL |
PLAYGROUNDS! CHANNELS 1
i i
AND OTHER j
NORMALLY j
DRY AREAS) I
i I i
. T~~~l
1
I
1
1
I
1
r
RECEIVING
BODIES
OF
WATER
(STREAMS,
RIVERS,
LAKES,
OCEAN)
4
I
GROUNDWATER
QUANTITY
AND
QUALITY
(* EVAPOTRANSPIRATION NEGLIGIBLE DURING PRECIPITATION)
FIGURE 1-URBAN STORMWATER DISPOSAL PHYSICAL SUBSYSTEM.
CD
-------
components of the urban stormwater disposal physical subsystem. In a
given instance, either or both of the two components indicated by dashed
lines might be absent.
Whereas there is a continuum between the subsystems of water supply,
water use and wastewater reclamation, stormwater has historically been
regarded as a purely negative good or nuisance and its subsystem has
seldom been deliberately connected to the other urban water subsystems.
Historically, urban settlements have been drained by underground systems
of sewers that were intentionally designed to remove stormwater as rapidly
as possible from occupied areas. Substantial departures from that
tradition are required by new national priorities: enhancement of urban
environments; conservation of water resources; and reduction in water
pollution.
The greatest public concern will increasingly be on the quality of water.
This concern is intimately related to acknowledged imperatives of aesthetic
enhancement, expansion of recreational opportunities and more extensive
availability of waterfronts for public uses. Runoff is a carrier of
wastes, either as harvested for water supplies and converted to water-
borne sewage or as an urban ground-surface ablution. Thus, public health
considerations can transcend or temper economic considerations. In
addition, comprehensive approaches for managing water pollution problems
require that other water uses, planning, and guiding sound development, also
be considered.^ For example, utilization of the "blue-green" development
concept, which employs ponds with open space, for stormwater detention
and recreation, can enhance urban property values and decrease property
depreciation rates, thereby increasing long-term local government revenues.* ^
On the other hand, peak drainage runoff rates can be reduced by means of
le •
(4)
* —
proper land-development design/3^ The guiding principle is to reduce
the liabilities and increase the assets of urban runoff.
For historical reasons, about one-fifth of the nation's population (or,
about three-tenths of the population served by public wastewater collection
14
-------
systems) is served by combined systems of sewerage. ' Overflows from
combined sewers are thought to comprise a significant source of stream
pollution. Combined sewers have the dual functions of removing epher-
merally occurring storm water from urban surfaces and conveying wastewater
on a perennial basis. Conduit size is governed by storm drainage require-
ments, capacity requirements for wastewater being comparatively small.
For all but the relatively few days a year that storm runoff occurs, the
continuously flowing wastewater is intercepted near the combined sewer
outfall by means of a regulating device which diverts it to the treatment
plant via an "interceptor" sewer. Most storm flows are much too great to
be accommodated by interceptors, and almost all the wastewater burden is
discharge through the outfall to the watercourse when rainstorms are heavy
and prolonged. At the same time, sludge and debris that have been stranded
in combined sewers during relatively low rates of flow in preceding dry-
weather periods are scoured from the laterals and trunk sewers of combined
systems and are lifted or otherwise transported by the augmented flows
and eventually discharged to the receiving waters. It is estimated that,
in consequence, as much as 5 percent of the annual flow of sewage, and
20 to 30 percent of the annual volume of solids, are discharged to receiving
watercourses from combined systems.^ '
/0\
Pollution from storm sewer discharges^ ' may be almost as severe as that
from combined sewer overflows. The U. S. Environmental Protection Agency
advised in 1971 that requirements for control of pollution from combined
sewer overflows were rapidly becoming more stringent and that control of
f Q\
pollution caused by urban stormwater discharges was on the horizon. ;
More recently, the U. S. Council on Environmental Quality has noted that
the contribution of pollution from runoff sources is even greater than had
been suspected, from both diffused urban and non-urban sources. ' When
abatement of pollution from storm water conveyed by separate systems of
storm drains is attempted, the difficulties to be overcome may be more
severe because all such storm water must be passed through new and special
treatment facilities, there being no interceptor sewers in separate stormwater
15
-------
systems to divert small storm occurrence flows to perennially operated
wastewater treatment plants as in combined systems.
In sum, there is widespread interest in multi-purpose drainage facilities
that exploit opportunities for add-on water-based recreation, that
provide more effective protection of buildings from flooding, and admit
use and reuse of storm water for water supply. These all require storage
utilization for effectuation, and some require special treatment plants,
necessitating employment of some kind of control system for managing the
sudden and brief impact of stormflows. The scale of the problem is almost
overwhelming: most of the larger metropolises have well over a hundred
catchment areas and cumulative drain lengths of several thousand kilometers
in length. The difficulties to be overcome were compounded by the 1972
Amendments to the Federal Water Pollution Control Act which established a
national goal of zero-pollution.
LAND-USE CHANGES
Little of the land occupied by metropolitan areas has not been drastically
altered by urbanization, particularly the land in large central cities.
For example, the distribution of land use in the City of San Francisco is
approximately as follows:^ '
Percentage of
Classification Total Area of City
Residential 30%
Streets 25%
Publ.ic (about half recreational) 23%
Vacant 8%
Commercial 5%
Industrial 5%
Utility 3%
Institutional 1%
Distribution of land use between a central city and its contiguous metro-
politan area can differ appreciably. For example, the proportion of land
16
-------
occupied by residences 1s commonly higher in the suburbs. Furthermore,
there is considerable variation in land use from one metropolis to another.
Contributing to this variability are dissimilarities in growth rates.
Ranges in land use among seven of the largest U. S. metropolises are as
follows:(12)
Minimum Maximum
Land Use Portion Portion
Residential 1/3 1/2
Roads and Streets 1/6 1/3
Open Space and Recreation 1/20 1/4
Commerical, Industrial,
Institutional and Mass
Transportation 1/8 3/10
Urban expansion absorbs an estimated 170,000-hectares of land each year
in the U.S.^ ' Metropolitan suburbs will be particularly hard-pressed
to expand all public facilities in the face of a doubling in residential
construction across the country anticipated for this decade. ' Nationally,
suburban populations of metropolitan areas had surpassed those of the sur-
rounded central cities by 1970, and this sprawling trend is expected to
continue. Of considerable importance to urban runoff management is the
usual rapid diminishment in population density with distance from the
densest centers. For example, in 1960 the urban population around the
City of St. Louis was one-fifth larger than that of the City, but distri-
buted on over four times as much land area. '
Buildings, streets and other urban land coverings inhibit the access of
precipitation to the soil. The presence of extensive impervious surfaces is
thought to be the main reason why the total volume of direct storm runoff
from urban areas is generally greater than for comparable non-urban catch-
ments. ' For example, greatly increased volumes of direct runoff associated
with urbanization growth have been documented for some streams in Long
Island, New York. ' Although an indirect indication, evidence has been
documented of a correlation between the degree of imperviousness and the
17
-------
post-urban enlargement of stream channels for a number of small water-
(18)
sheds in the Philadelphia metropolitan area. ;
Estimates of imperviousness ranges for typical urban development are as
follows:(19)
Percent
Land Use Imperviousness
Low Density Residential 20-30
Medium Density Residential 25-35
High Density Residential 30-40
Business-Commercial 40-90
Light Industrial 45-65
Heavy Industrial 50-70
Another estimate of imperviousness for typical urban development has
also been offered. '
Percent Imperviousness
Land Use Low Intermediate High
Single Family Residential 12 25 40
Multiple-Family Residential 60 70 80
Commercial 80 90 100
Industrial 40 70 90
Public and Quasi-Public 50 60 75
Thus, it is not uncommon to find individual downtown city catchments that
are almost entirely covered by impervious surfaces and outlying suburban
catchments with as little as a tenth as much impervious cover. It follows
that the effect of imperviousness on direct storm runoff volume can vary
over a considerable range from one urban catchment to another.
For the State of New Jersey, it has been found that urban and suburban
land-use characteristics are closely related to population density. '
MORPHOLOGICAL CHANGES IN DRAINAGE
(21}
Figure 2V ' indicates the diminishment of tributary channels over a 53-
p
year period in a 68-km. section of Maryland of the Rock Creek watershed,
which is now a heavily populated suburban area adjacent to Washington, D.C.
18
-------
FIGURE 2-DIMINISHMENT OF TRIBUTARY CHANNELS WITH URBANIZATION(21)
-------
In 1913 this section was rural, with many small tributaries fed by springs
and seeps, most of which were covered over as storm sewers were installed.
Of the 103-km. of natural flowing stream channels that existed in 1913,
only 42% could be found aboveground in this section in 1966.
The function of underground drainage conduits is to remove storm water
from urban surfaces (except combined sewers, which in addition convey
wastewater on a perennial basis). The smallest catchment area (on the
order of a fraction of a hectare in size) is that tributary to a street
inlet. Flow in storm and combined sewer systems is principally by gravity.
Like natural drainage basins, smaller sewer branches unite with larger
branches, and so on, until a main sewer is reached. Thus, a main sewer
not only transmits upper reach flow to a receiving watercourse but serves
as a collector of surface runoff all along its route. There are over
300,000-km. of storm and combined sewers in the U.S.
From the foregoing, it is therefore not surprising to find that the total
lengths of underground drainage conduits dwarf those of open watercourses
p
in major cities. For example, total lengths in the 246-km. of the City of
(22}
Milwaukee as of the beginning of 1970 were as follows. '
Lakefront Length - 13-km.
River Lengths - 60-km.
Combined Sewers - 885-km.
Storm Sewers - 1320-km.
(In addition, there were 1105-km. of sanitary sewers).
Examples of underground conduit drainage density in some major cities
are given in Table 1. Ratios of total sewer length in km. to city
2
area in km. range only between 8 and 18, except for Los Angeles which
makes much more extensive use of open channel drainage than the other
cities cited. Also shown in Table 1 are ratios of total sewer length in
2
km. to the 0.6-power of the city area in km. , ranging between 81 and
136, except for the low value of 20 for Los Angeles. The corresponding
20
-------
TABLE 1
DENSITY OF UNDERGROUND DRAINAGE CONDUITS
IN SOME MAJOR CITIES
Area of City,
A, Km2
124
580
360
1 1,191
246
829
337
160
l.a 114
158
Total Length
of Storm and/or
Combined Sewers,
L, Km.
2,190
5,786
4,656
1,389
2,205
6,650
4,023
1,796
1,400
2,816
L/A
Km. /Km.2
17.6
10.0
13.0
1.2
9.0
8.0
12.0
11.2
12.3
17.8
L/A°'6
Km./CKm.2)0'6
121
127
136
20
81
118
122
85
82
135
City
Boston, Mass.3
Chicago,111.3
Detroit,Mich.a
Los Angeles,Cal.
Milwaukee,Wis.c
New York,N.Y.d
Philadelphia,Pae
Saint Louis,Moa
San Francisco,Cal
Washington,D.C.3
a: Hallmark, Dasel E., and John G. Hendrickson, Jr., "Study of Approxi-
mate Lengths and Sizes of Combined Sewers in Major Metropolitan Centers,"
ASCE Combined Sewer Separation Project, Technical Memorandum No. 4,
ASCE, N.Y., N.Y., May 1, 1967. (NTIS Id. No. PB 185 999).
b: Bauer, W. J., "Economics of Urban Drainage Design," Journal Hydraulics
Division, ASCE Proc., Vol. 88, No. HY6, Paper 3321, November, 1962.
c: Prawdzik, T. B., "Environmental and Technical Factors for Open Drainage
Channels in Milwaukee," ASCE Urban Water Resources Research Program,
Technical Memorandum No. 12, ASCE, N.Y., N.Y., February, 1970
(NTIS Id. No. PB 191 710).
d: (Unpublished notes of ASCE Combined Sewer Separation Project,
November 15, 1967).
e: Radziul, J. V., C. F. Guarino and W. L. Greene, "Combined Sewer
Considerations by Philadelphia," Journal Sanitary Engineering Division,
ASCE Proc., Vol. 96, No. SA1, Paper 7050, February, 1970.
21
-------
ratio would be only about 2 for river and stream channels, ' tending
to support the notion that underground storm drainage systems replace
mostly the smallest natural channels.
The catchment sizes of sewered drainage areas are generally much smaller
than those of the natural streams passing through major urban areas, as
(24)
implied by the examples in Table 2. ' The small median sizes further
substantiate the contention that underground drainage systems replace
mostly the smallest natural channels, because natural catchment boundaries
mostly tend to be preserved when underground drainage systems are provided.
A survey of combined sewer sizes in major cities revealed that an average
of seven-tenths of cumulative city system lengths are 60-cm. (24-in.) or
(25)
smaller, and that this ratio increases for smaller cities. ' Because
design capacity criteria for combined and storm sewers are very similar, it
appears safe to conclude that, for both types collectively, an average of
seven-tenths or more of total system lengths are 60-cm. or smaller in size.
(26)
The typical design capacityv ' of a 60-cm. storm or combined sewer is not
much more than about 0.3-cu.m./sec., indicating that the bulk of storm
drainage systems convey very modest rates of flow, thus adding further sub-
stantiation to the surmise that because the majority of catchments are of
small size and therefore collect mostly storm sheetflow, they replace mainly
fringe tributary natural channels.
CHANGES IN FLOOD CHARACTERISTICS
The small size of tributary catchment area served by the majority of storm
drainage conduits and their small hydraulic design capacity are heavily
countered by the large number of such catchments and the fact that they,
together with the larger catchments, commonly drain collectively as much
as three-fourths or more of metropolitan surface areas. For example, from
seven-tenths to nine-tenths of the total areas of the cities of Milwaukee,
Washington, San Francisco and Philadelphia are drained by sewered catch-
(24)
ments, compared with only about three-tenths of Houston/ ' Because very
22
-------
TABLE 2
STORMWATER DRAINAGE-CATCHMENT AREA-SIZE DISTRIBUTIONS
FOR SOME MAJOR CITIES
Total Area
of City,
City Sq.Km.
San Francisco,
Cal.
Washington ,D.C.
Milwaukee, Wise.
Houston, Tex
114
158
246
1150
Number
of
Catchments
42
93
465
1283
Largest
Drainage
Area,
Hectares
1750
2500
740
1030
Average
Drainage
Area,
Hectares
227
152
39
26
Median
Drainage
Area,
Hectares
77
26
10
2
few flow gagings have been conducted at conduit outfalls over hydro-
logical ly significant periods, there are few, if any, trustworthy
generalizations that can be made specifically about sewered catchment
flooding. Historically useful data essentially has been secured from
stream gages on watersheds that have undergone varying degrees of urban-
ization. Examination of available data and of the results of various
studies has indicated that the most dramatic hydro!ogic impact of urban
development is that on peak flows, where the basin lag time (or time of
concentration) "is reduced as an area becomes urbanized, and the storm-
flow often is concentrated in sharper, shorter, higher peaks than those
for natural runoff.
,,(27)
The very rapid response of a sewered catchment is illustrated, in Figure
3.
(28)
The hydrograph shown was obtained at the outfall sewer exit of
a 19-hectare drainage area using a Parshall flume. Response time is so
short for such a small but typical catchment that an observation interval
no longer than one-minute is required to obtain sufficiently detailed
data. Rapid response and serious difficulties in making flow measurements
(2g\
within sewer systems, ' underground, have presented severe obstacles in
the acquisition of basic data.
23
-------
3
o
LU
CO
O
U.
U.
O
a:
a
•
S
2
Z
RUNOFF
a:
RAINFALL
A
/\
/ l
M
I
n 1
|L
40
TIME FROM BEGINNING OF RAINFALL, MINUTES
FIGURE 3- HYETOGRAPH AND HYDROGRAPH, STORM OF
4 AUGUST 1965, .19-HECTARE NORTHWOOD
DRAINAGE AREAC28)
24
-------
Over a period of two decades the portion covered by Impervious surfaces
2
grew from 17% to 31% in an urbanized 356-km. stream watershed located 1n
the Atlanta, Georgia, metropolitan area. Analysis of rainfall and runoff
records for the period imply that storm runoff volume has increased in dry
months, base flow has decreased in wet months, and peak runoff from
summer storms has increased significantly/ ' The latter implication
was based on observed changes over time in unit hydrograph characteristics
and further substantiates earlier observations in this regard by other
researchers.' '
Man-made underground conduits and the gutters of streets have much lower
frictlonal resistance than the natural minor rills that formed definite
watercourses and all the ephemeral channels in the furthermost catchment
reaches that they mostly replace. In addition, nearly all major cities
often use improved natural channels or excavated special channels for
collection of storm water from drainage conduit outlets, such as in
(29}
Milwaukee/ ; which also expedite the collection and traversing speeds
of surface runoff. Because dramatic increases in peak runoff have been docu-
mented for some only partially and even lightly sewered catchments, it
appears likely that the increased peaks could be caused by the acceleration
of flow afforded by the lesser frictional resistance of streets and conduits,
although the role of changed aggregate channel lengths is clearly uncertain.
Comparative studies of urban versus rural drainage basins indicate that
as the relative magnitude of flood peaks increases the ratio of urban peak
rate to rural peak rate declines, the effect of urbanization being more
(27)
pronounced for the more frequent occurrences. ' Apparently, grossly
overburdened surface and underground collection systems relapse towards the
carrying capacity of the natural systems they replace.
Further generalizations are probably not warranted because of the throttling
effect of various encroachments, such as loss of natural storage from flood-
plain development, and the wide variety of drainage and flood mitigation
development patterns commonly encountered.' '
25
-------
FLOOD MITGATION VERSUS THE AMENITIES OF DRAINAGE
Geohydrologic processes have formed natural drainage channels that convey
storm waters to the seas. Because of an early dependence by commerce on
water transport, most large metropolitan areas originated as urban centers
on or near streams, lakes, estuaries or seacoasts. Intrusion of urban
development on natural flood plains has resulted in damages to occupying
structures and sometimes loss of life. However, as argued in previous
sections, the bulk of urban development in most regions of the nation is
located on land that was originally subjected to sheet-flow runoff collected
locally in minor channels. Subterranean systems of conduits facilititate
human occupancy by draining sheet-flow runoff from the land surface. That
is, fluvial drainage areas contributing to urban flood-plain inundation
are often of gigantic size compared with individual underground conduit
catchment areas which rarely exceed 2600-hectares in extent. Thus, under-
ground drainage systems are more of a convenience or amenity than a preserver"
of public safety. Structural means for mitigating flood-plain inundation
are designed to provide a much higher level of protection than that for
storm drainage systems because of the much greater threat to human life and
more community-wide economic implications.
Urban drainage facilities are generally owned, operated and maintained by
local governments, and designed and constructed by local governments and
private land developers. Human life is seldom threatened by the flooding
of these facilities. Because the principal local detrimental effects of
flooding are damage to the below-ground sections of buildings and hindrance
of traffic, the consequences of flooding range from clearly assessable
property destruction to annoying inconvenience. It follows that provision
of complete protection from flooding canonly rarely be justified. Instead
facilities are designed which will be overtaxed infrequently. However,
because of the marginal level of protection afforded, storm drainage
flooding damages are also of considerable magnitude, probably exceeding
those in urban flood plains. In addition, intangible damages from flooding
26
-------
are much more extensive than for stream flooding and generally recur more
often, and direct damages are usually much more widely dispersed throughout
a community.
The safety of people and properties occupying flood-prone areas is the
concern of every level of government, and National, State, special
district and local governments are involved in the development of an ever-
greater arsenal of remedial or mitigative measures and policies. However,
most stream flood-management is undertaken by national agencies. Protection
against "upstream" floods experienced on creek and headwater areas is
afforded primarily by the Soil Conservation Service, whereas the Corps of
Engineers concentrates more on "downstream" floods on mainstems and major
tributaries. A notable exception is the Tennessee Valley Authority, which
has primary responsibility for both aspects in its region. "A complete
inventory of nationwide flood damages has never been undertaken Current
efforts to manage flood-plain use and development on a national basis are
Increasing."*32^
Actual national average stream flooding damages are currently estimated at
more than $1000-million annually for urban and non-urban flooding collectively,
of which about half occurs in urban areas. ' In comparison, the estimated
actual national average annual loss from the flooding of urban areas served
by underground drainage (storm and combined sewer) systems is at least
$1000-million. ' That is, flood damages assignable to sewered catchments
may exceed substantially those in urban flood plains. Further, national
investment for storm drainage conduit facilities appears to be at least
three times as great as that for flood plain protection works benefiting
urban areas. The dominance of storm drainage investment over stream
protection costs is expected to continue. While the need for more stream
protection will rise as greater urban sprawl leads to inevitable new en-
croachments on flood plains, the same urbanization is expected to require
as much as a doubling in the extent of storm drainage systems by the end
of this century.
27
-------
Perhaps only about one-sixth of metropolitan territory is within 100-
(33}
year flood plainsv ' versus well over half of that territory being
served by systems of underground drainage conduits. Thus, flood damages
per unit of area are clearly greater in flood plains than in sewered
sectors of metropolitan areas.
Increased volumes of direct runoff from underground drainage conduits
clearly can aggravate flooding of urban flood plains. On the other hand,
increased receiving-stream stages can cause or induce flooding of under-
ground drainage systems, because of the intimate hydraulic linkage between
them. Extensive programs of integrated flood plain management are
particularly crucial for metropolitan areas that have little topographic
/ pC \
relief, such as greater Chicago. ' There are large cities that early
dedicated most of their flood plains over to parkways,such as Philadelphia,
or developed extensive systems of major drainage channels in step with
urban development, such as Los Angeles. ' ' Regardless, new suburbs
have more than occasionally aggravated flooding with subsequent induced
damages in many major cities located on large streams or bodies of water
because their low topological elevation makes them hydro!ogically sub-
servient.
There are several non-structural means for mitigating flooding damages
resulting from the tenacious inclination of people to occupy the flood
plains of urban streams, but these require changes in the activities of
occupants and are therefore socio-political in character. The proclivity
of home-owners to remain on flood plains has been demonstrated in a pilot
study of occupant motivation, hazard perception and attitude, and
receptivity towards various forms of communication designed to discourage
(38 39)
further encroachment. ' '
In sum, discharges from conventional storm drainage facilities and flood-
plain intrusion by structures both tend to aggravate flooding, and thereby
jointly tend to raise the potential for flooding damages. Revising
storm sewering criteria, such as by including much more in-system storage,
28
-------
can be an effective adjunct in flood plain management. While there is
universal agreement that planning and development of drainage systems
and flood plain management programs should be coordinated and integrated,
prospects for accommodation diminish rather than improve because of an
increasing concern over water quality considerations in sewered systems
and a contemporary neutrality or indifference on water quality matters by
agencies dealing predominantly with flood plains. It is ironical that
much of the flood plain flooding problem as well as the land runoff
water quality problem could possibly be more effectively countered on the
land feeding urban watercourses.
Lastly, while the principle of the use of local detention storage is
often championed in lieu of main channel improvements for metropolitan
areas, sharp differences of opinion have been known to arise, partly or
perhaps principally because implementation of the two methods commonly
falls within different jurisdictions of authority/40^
29
-------
POLLUTION EFFECTS*
Wide recognition of the deteriorating quality of natural waters and
large-scale efforts to control waste discharges in the U.S. date
(41)
essentially from the end of World War IK
"The concept that management of the quality of water is quite as important
as its physical management is now widely recognized and advocated in the
United States. However, it is of such recent origin that no precise
definition of the term has been established and widely accepted, principally
because the jurisdiction in which the responsibility for management falls
is exceedingly fragmented."^ The most recent philosophy of public
objectives relates to the nature of water as a resource, with an emphasis
(42)
on its capability for meeting a variety of social objectives/ Such
orientation moves away from an almost exclusive emphasis on control of
quality-depreciating agents as an end in itself. There has been a definite
shift from "pollution control" to "quality of the environment" as the
objective of national policy.^ The Federal Water Pollution Control Act
Amendments of late 1972 set forth a national goal of zero pollution.
Using broad, collective, subjective indices, estimated U.S. stream pollution
causes divide as follows: industrial wastes (trade wastes), one-third;
municipal wastes, one-third; agriculture, one-fifth; water management
practices, construction, navigation and recreation, three-fiftieths; mining,
one-twentieth; and other urban wastes and power generation the remaining
close to one fiftieth.^44' The division of pollution load among these
sources varies widely with location. The same assessment of the prevalence
of pollution found that almost a third of U.S. stream lengths are
characteristically polluted.^ Incidence of polluted streams is highly
*From "Hydrological Effects of Urbanization in the United States", by
M. B. McPherson, Chapter 6, Part II, "National Case Studies for Selected
Countries", in Hydrological Effects of Urbanization: Environmental Impact,
an IHD report to be published by Unesco in August, 1974:
30
-------
correlated with major urban population concentrations, ' as the
preceding estimated allocation of causation implies. Attention to
(47)
subsurface water pollution has not been neglected. '
Slightly more than two-thirds of the nation's population is served by
public systems of wastewater sewerage, with about 46 percent of these
people connected to systems with treatment plants that are overloaded or in
need of major upgrading and with 7 percent residing in communities with no
treatment facilities whatever/48'' However, four-fifths of the total SMSA
population is served by wastewater sewerage. ' While about 55 percent
of the volume of wastewater processed by municipal plants comes from
homes and commercial establishments and the other 45% is from industries, '
the latter is only a small part of the total urban industrial burden.
"Industries discharge the largest volume and most toxic of pollutants ...
Major water-using industries are believed to discharge, on the average,
about three times the amount of (settleable and suspended solids and
oxygen demanding organic materials) as is discharged by all of the sewered
persons in the United States The volume of industrial wastes is
growing several times as fast as that of sanitary sewage "^ The
fact that potential water supply sources are evaluated with regard to levels
of quality as well as adequacy of quantity means that the choice of source
ultimately can be dictated by quality considerations, and thus quality
rather than quantity alone can determine the location of source impact on
the hydro!ogic balance of a subregion.
The surface and underground waters of the nation, the present focus
of water pollution abatement activities, are the "raw" sources for drinking-
water supplies. The most demanding quality requirements are for the trans-
formation of raw water into drinking water in purification plants and the
subsequent delivery of safe water to users. Stream quality objectives
have tended to emphasize dissolved oxygen levels and micro-organism contents.
Water purification viability is little affected by raw water dissolved
oxygen levels, and micro-organisms can be rendered harmless as is evidenced
31
-------
by the essentially unblemished public health record of water treatment and
distribution over the past several decades. Rising amounts in raw waters
of toxic chemicals, including heavy metals and exotic organics, uncer-
tainties on the instantaneous and cumulative human health hazards of
these substances, and resultant ambiguities on treatment removal objectives
and standards, are sources of considerable management concern. That water
(51)
purification must increasingly be more meticulous is widely accepted. '
About one-fifth of the nation's population (or three-tenths of those
served by community wastewater sewerage) is served by combined systems of
f5«iS
seweragev ' and of the fourteen largest cities, ten have combined systems
in whole or in part!52) The first definitive identification of the pollu-
tion problem from combined sewer overflows on a national scale appeared in
1964. ' It has been estimated that, on the average, where dry-weather
flow receives secondary treatment, during storm events overflows from com-
bined sewers represent a load equivalent to about one-third of the treat-
ment plant effluent load, which is therefore about one-fourth of the total
(53)
load discharged to receiving waters. ' The average annual BOD load
from combined sewer overflows is estimated at roughly 550-kilograms per
hectare served, not including additional loads due to treatment plant
upset or flows bypassed at plants as a result of storm events. '
"Requirements for control of pollution from combined sewer overflows are
rapidly becoming more stringent. Control of pollution caused by urban
storm water discharges is on the horizon."^9J Documentation of the
potential storm drainage (rainwater sewers) pollution problem started in
1964* ' A disproportionate amount of street surface pollution load is
associated with fine dust and dirt particles, and a significant fact is
that heavy metals and toxic material concentrations are much higher in
urban runoff than in domestic sewage/54' Preliminary findings indicate
that polychlorobiphenol concentrations of stormwater range between 0.4
and 2 parts per million; BOD ranges between 1/2- and 17-kilograms per curb
kilometer; COD ranges between 4- and 110-kilograms per curb kilometer;
and mercury and lead have been found to range between 0.0055- and 0.085-,
32
-------
and between 0.034- and 0.522-kilograms per curb kilometer, respectively.^)
(Tetraethyl lead presently is an additive in most U.S. petrol.)
Because very few sewered catchments have been gaged, and nearly all
have been gaged only at outfalls, very little is presently known about
rainfall-runoff-quality processes. Much more investigation must be under-
taken. (54)
Salting of streets and highways, used for the facilitation of traffic
in ice and snow, results in local soil concentrations of leachates. 56,57^
In 1966, pesticide use in the conterminuous U.S. totaled almost one-third
of a million metric tons of active ingredients, half of which was used by
fanners and the remainder by government, industry and homeowners.^56'
Pesticides and herbicides have diffused and concentrated throughout the
land-air-water phases and pose direct and indirect threats to the health
/ \
of urban dwellers and the survival of numerous animal species. 58' One
of the first documented studies of the carriage in runoff of pesticides
f \
from agricultural catchments was reported in 1966.159J
Disposal of the residue from wastewater treatment^60^ is fast becoming
a major problem,'61J with sludge quantities increasing, while use of nearby
potential land disposal sites is progressively foreclosed by development,
and sea disposal is challenged as a suspected cause of environmental
degradation.
Total municipal wastewater solids, most of which are removed in conventional
treatment, average on the order of 90-metric tons per day per million persons
served, dry weight. However, sludge is often transported to disposal sites
as a slurry. Costs of transportation and handling can be reduced by
slurry volume-reduction through concentration or "thickening" of sludge
solids. Based upon the maximum concentration that can be hydraulically
transported, the minimum slurry load is about 360-metric tons per day per
million persons served. However, normal practice, which reflects reliability
of operation as well as cost-attractiveness, results in a slurry of about
900-metric tons per day per million persons served. The magnitude of the
33
-------
sludge disposal problem can be appreciated by noting that the preceding ton-
nage is in the neighborhood of half the weight of solid wastes per million
persons accumulated per day. As opposed to solid waste disposal, waste-
water sludge disposal encounters two monumental liabilities: an
extremely large water content and a very high organic composition. As a
result, the feasible avenues of disposal generally differ from, and often
are more restricted than, those for solid wastes.
Since mid-1971, a special district serving most of the Chicago metropolitan
area has been barging by river about half of its daily sludge production in
slurry form to an area about 320- kilometers from Chicago'62' Expected
increases in sludge load from the Chicago area will be disposed of in
strip mine and farmland reclamation.
Several major coastal cities dispose of sewage sludge by barging it
out to sea. Almost 4-million cubic meters of sludge slurry from the New
York metropolitan area are disposed of annually in the New York Bight of
the Atlantic Ocean !63>64)
As opposed to the more or less steady regularity of solids content for a
given wastewater system, solids in surface water drawn for domestic use
often exhibit large seasonal variations (largely from episodic suspended
sediments) and can range from such small amounts that no treatment reduction
is warranted to concentrations well above those of normal wastewater.
Hence, the seriousness of the water treatment sludge disposal problem
varies among different sections and localities'65^ On the average, the
amount of sludge produced annually from all U.S. water treatment plants is
on the order of about 25-metric tons of slurry per million persons
served.
Land disposal of wastewater treatment plant effluents as an alternative to
conventional water disposal is receiving greater attention as concern for
control of surface water quality intensifies'67*68^
Wherever solid wastes are deposited on land, there will be opportunities
for these wastes to intermingle with surface water and groundwater. The
34
-------
extent of pollution of groundwater by the leaching of surface water through
deposits of solid wastes is largely dependent upon the character of
the immediate geologic environment, **' but groundwater can be protected by
modifying landfill sites.(|70'71 '72)
Estimated sources of total solid wastes produced in the U.S. in 1969 divide
as follows:(73)
Source of Wastes Million Metric Tons
Residential, commercial and institutional - 230
Industrial - 100
Mineral - 1,500
Agricultural - 2,100
Concentrations of mineral and agricultural wastes are spread widely over the
land and are mostly isolated from population concentrations, although some
deleterious effects on urban residents are nevertheless suspected, and have
been documented in isolated instances. The bulk of the other wastes are
disposed of in or near urban areas and represent an ever-present threat to
public health. Space available for municipal solid waste land-disposal is
fast disappearing. Integrated management of urban residuals in air and
water and on the land is regarded as a necessity for adequate environmental
protection in the future.(74)
Urbanization and associated human activities have accelerated geological
(751
processes of land erosion. ' Damage is inflicted at the scene where
soil is eroded, where it is washed downstream, and where sediment remains
suspended or comes to rest.
"Efforts to control sediment pollution must be concerned not only with
treatment of water which is already affected by sediment, but also with
controlling the massive amounts of soil being washed into waterways each
year."^ Three broad categories of land disturbance are appropriate foci
35
-------
in control efforts: erosion of "natural" or geological character;
erosion resulting from agricultural, forestry and mining activities;
and erosion associated with suburban development.
(76)
"Evidence being
made available by current research suggests that sediment yields in
areas undergoing suburban development can be as much as 5 to 500 times
greater than in rural areas."^ ' An estimated 170,000-hectares of land
per year are taken over by urban expansion. '
The sequential qualitative effects of various land-uses on relative
sediment yield and channel stability have been described as follows: ^
Land Use
Sediment Yield
Channel Stability
Natural Forest or
Grassland
Heavily Grazed Areas
Cropping
Retirement of Land
from Cropping
Urban Construction
Stabilization
Stable Urban
Low
Low to moderate
Moderate to heavy
Low to moderate
Very heavy
Moderate
Low to moderate
Relatively stable with
some bank erosion.
Somewhat less stable
than the preceding.
Some aggradation and
increased bank erosion.
Increasing stability.
Rapid aggradation and
some bank erosion.
Degradation and severe
bank erosion.
Relatively stable.
Stabilized urban development sediment yields tend to approach presettlement
levels, 'as noted above.
Sediment from erosion impairs water quality, reduces the volume of available
water supplies, and damages recreational waters, quantitatively and
36
-------
qualitatively, such as by accelerating eutrophication. Effectiveness
of mosquito-control works can be seriously reduced by erosion-deposition
occurrences. The biological viability of aquatic biota is inhibited by
suspended sediments, resulting in reduced waste-assimilation capacity of
water-bodies and other deleterious effects. ' The "importance of
sediments as pollutants is increasing, particularly in view of the ability
of soil to absorb pesticides and other organics including oily substances
and to release these materials in the water resource. Indeed, this facility
of sediments to receive chemicals from the solution phase, when properly
understood, may be subject to manipulation and made to serve effectively
to trap and immobilize harmful wastes moving in our environment."^ '
In terms of public safety: main-channel deposits can aggravate main-
stream flooding; sediment deposited on and in property on the occasion
of overbank flooding can cause considerable damage over and above that of
inundation; structural integrity of water conveyance facilities, such as
culverts and bridge piers, can be imparied or destroyed; and buildings
can be badly damaged or destroyed and channel flooding can be aggravated
by landslides induced by heavy rains.^82'83'84^
Removal of sediment deposits from public thoroughfares and other public
areas adds to the cost of maintenance. Turbidity from mineral and other
sediments can tax water treatment facilities.
In sum, few beneficial effects, if any, can be claimed presently for
erosion in the urban environment, and the cost of uncontrolled sediment
movement is indeed large.
The significant amount of water withdrawn in metropolitan areas for power-
generation cooling has been noted in Section 6. If the present growth
rate of total energy consumption in the U.S. oersists, in about fourteen
(85^
years it will be double the present amount. ' Generation of electricity
(oc\
accounts for less than one-fourth of all U.S. energy consumption/ ;
Waste heat from electric power generation is one of the most serious
emerging sources of water pollution, accounting currently for four-fifths
37
-------
of all the thermal heat entering the nation's waters* and heat discharges
(86)
increase as generating capacity doubles every decade. ' The total
quantity of waste heat discharged to condenser cooling waters will more
(87)
than double from the year 1967 to the year 1980.v ' The trend towards
larger, nuclear plants, which discharge three-halves as much thermal load
in water per unit of power as fossil-fuel plants, pose an increased threat
to aquatic systems. (86) Among several alternatives to the use of water
bodies as heat sinks is the use of cooling towers, already commonly
deployed in Europe but little used in the U.S., much to the ire of conserva-
tionists, particularly for nuclear sites. 8' Dissipation of residual heat
is regarded as a serious problem for future generating plants.(89,90,91)
Considerable attention is being given to the environmental problems of
(92)
plant siting/ ' but the root source of these and related problems is
(93)
considered to be the absence of a national energy policy. '
National energy and water policies will be shaped by overall growth policy.
The National Government in 1971 was attempting to resolve the final deter-
(94)
mination of a national growth policy/ ' and in 1972 it was asserted that
"the time has come for the United States to adopt a deliberate population
(95)
policy."^ ' Because most land-use decisions are made by private interests
and governmental control over land-use is fragmented at the local level,
better land-use planning at all levels of government is advocated as a
part of overall national grow policy. * Also needed are regional bodies
with authority to plan and control those facets of land use that transcend
(97)
local boundaries, such as pollution abatement. '
38
-------
BETTER DESIGN OF STORMWATER DRAINAGE SYSTEMS
DESIGN OBJECTIVES
Much has happened since APWA presented its report on urban drainage in
1966^34 ' and on combined sewers in 1967/98 , anc| even s-jnce the report
on storm runoff quality in 1969^" '. Historically, urban settlements
have been drained by underground systems of sewers that were intentionally
designed to remove storm water as rapidly as possible from occupied
areas. Substantial departures from that tradition are required by new
national priorities: enhancement of urban environment; conservation of
water resources; and reduction in water pollution.
The greatest public concern will increasingly be on the quality of water.
This concern is intimately related to acknowledged imperatives of
aesthetic enhancement, expansion of recreational opportunities and more
extensive availability of waterfronts for public uses. Runoff is a carrier
of wastes, either as harvested for water supplies and converted to water-
borne sewage or as an urban ground-surface wash. Thus, public health
considerations can transcend or temper economic considerations. In
addition, comprehensive approaches for managing water pollution problems
require that other water uses, planning, and guiding sound development,
also be considered. ^ For example, utilization of the "blue-green"
development concept, which employs ponds with open space, for stormwater
detention and recreation, can enhance urban property values and decrease
property depreciation rates, thereby increasing long-term local government
revenues/ 3 ' On the other hand, peak drainage runoff rates can be
reduced by means of proper land-development design/ 3 / Tne guiding
principle is to reduce the liabilities and increase the assets of urban
runoff/ 4 )
Capturing and/or diverting storm water for recreational and aesthetic
enhancement, for use as a water supply supplement, and for meeting the
zero-pollution goal of Public Law 92-500 (October 18, 1972), all require
or imply the employment of some degree of detention and/or retention
39
-------
storage. Such storage would have to be added to existing systems and
included in new systems, at substantial cost.
The Environmental Protection Agency advised in 1971 that requirements
for control of pollution from combined sewer overflows were rapidly
becoming more stringent and that control of pollution caused by urban
stormwater discharges was on the horizon. 9 ) in 1972 the Council on
Environmental Quality noted that the contribution of pollution from runoff
sources is even greater than had been suspected. 1° )
Discharges frcrnconventional storm drainage facilities and floodplain
instrusion by structures both tend to aggravate flooding, and thereby
jointly tend to raise the potential for flooding damages. Revising storm
sewering criteria, such as by including much more in-system storage, can
be an effective adjunct in flood plain management. While there is universal
agreement that the planning and development of drainage systems and flood
plain management programs should be coordinated and integrated, prospects
for accommodation diminish rather than improve because of an increasing
concern over water quality considerations in sewered systems and a con-
temporary neutrality or indifference on water quality matters by agencies
dealing predominantly with flood plains/!00) It is ironical that much of
the flood plain flooding problem as well as the land runoff water quality
problem could possibly be more effectively countered on the land feeding
urban watercourses.
Techniques for hydrologic analysis of rivers and urban streams are at a
stage of development far beyond those for local drainage, mostly because
of a much broader data base.
The procedure used in nearly all current storm sewer design is the "rational
method"'26 ' the numerous inadequacies of which have been discussed else-
where^0^ It is suffucient to emphasize that the method yields only an
estimated peak flow but none of the other attributes of a hydrograph. A
full hydrograph is needed for the design of detention storage, for evalua-
tion of pollutant burdens, for design of stormwater pollution abatement
-------
facilities, for designing local protection works along streams, such as
pumping stations for passing local drainage flows over levees and dikes, and
as inputs for design of stream and river development works. Also,
quantification of the effects of urbanization on the hydrologic regimen 1s
dependent in many cases on the availability of sewer outlet hydrographs.
Further, as urban water management problems become increasingly acute, the
need for multiple-use of water becomes more evident. In exchanging one
use for another, for example by using storm water as a source of water
supply, knowledge of the time-histories of flows and water qualities is
essential for reliable design of transfer facilities. In sum, there is
essentially world-wide agreement that improved methods for design based
on field observations are sorely needed.
Analytical requirements for planning are less rigorous and permit less
detail than for design because investigation of a range of broad alterna-
tives is at issue. What are sought for planning tools are general parameters
or indicators for large-scale evaluation of various alternative schemes.
Hence, the degree of computational detail required in metropolitan planning
is much less than for design. However, a certain amount of intensive de-
tailed analysis is needed to establish parameters and indicators and to
provide an underlying understanding of the governing hydrologic processes,
so that simplified expedients are not inadvertently misused.
The ultimate solution for the problem of abating pollution from urban storm
sewer discharges and combined sewer overflows is the treatment of such flows
prior to their release into receiving waters. Outflow surface-water
hydrographs of sewered catchments exhibit very large peaks, on the order of
two or more times those of equivalent non-sewered areas. ^°2 Capturing,
transporting and treating all discharges/overflows, unattenuated, would
require gigantic collection sewers, pumping stations and treatment facilities
all' of which would be used less than the equivalent of about an hour a day,
on the average, over a typical year. Further, there will be periods running
into several weeks or even months where the discharge/overflow control works
will not operate at all, as a consequence of zero or light precipitation.
41
-------
Therefore, schemes for system-wide discharge/overflow collection and
treatment incorporate some form of auxiliary storage for the purpose of
attenuating collection and treatment facilities inrushes, with the
objective of scaling down the size of such facilities to reasonable and
manageable proportions. In addition, the complexity added by converting a
facility with simple gravity flow into a multiple-component interacting
scheme, where practically instantaneous response to flow incidences must
be made, requires incorporation of some degree of automatic operational
control. Combined sewer system applications are exemplified by plans
for the Chicago metropolitan area^103*104^ and for the City of San
Francisco!11>105)
Compared with design and planning, central service objectives stand out
in bolder relief in operations: minimized flooding of properties, minimized
quality degradation of receiving water bodies, maximized recreational
and aesthetic opportunity or maximized water supply use. Whereas process
mastery is vital for flexibility in design and planning of project alternatives,
the control mode can be, and probably must be, approached in a much more
pragmatic fashion.' °6' The real test of performance effectiveness occurs
in basements of buildings and in contaminant levels of receiving water
bodies.
SEWER DESIGN
The "rational method" applies to a very unique and restricted set of assumed
conditions. Once a system is designed using the fixed features inherent
in the method there is no logical way to analyze modifications, such as
provisions for relief, reusing the method. Also, because there is no direct
way to verify the method in the field, even the adequacy of the original
design cannot be checked. The simplistic scope of the method permits and
requires a wide latitude of subjective jdugment in its application.
Because the method is based on hypothetical conjecture unsupported by
runoff data, generally speaking there can be no "wrong" storm sewer designs
when the rational method is employed. An.outstanding limitation of the
42
-------
method 1s that the product is restricted to a peak flow; and although by
heaping more assumptions on the method a hypothetical hydrograph can be
computed, such as by arbitan'ly proportioning the contributing area with
time, few desgners are willing to elect this frivolous option.
If the method Is so inadequate, why has it seemed to "work" all these
years? Three factors affect the magnitude of a design flow in using
the rational method: C-value, inlet time and the frequency for the rainfall
Intensity-duration curve used. Computed peak flows are larger as the 0-
value is raised, as the inlet time is shortened and as a curve for a
rarer rainfall frequency is used. Intelligent, though arbitrary, selection
of values for these three variables has been found to give ostensibly
"satisfactory" results in a number of places. There are several factors
which may contribute towards conservatively "safe" designs, such as the
usual practice of designing sewers to accommodate at least the design
flow rate at a flowing-full condition, whereas some degree of surcharge
might be sustained without flooding. Also, the probability of all design
assumptions being satisfied simultaneously is less than the probability of
occurrence of the rainfall rate used in the design, contributing in effect
to a "safe" design with a built-in factor of safety.
Probably the most important influence is the narrow range of input values:
the range of short duration rainfall intensities provided by nature. For
example, a 20% variation in design rainfall Intensities about 5-year values
covers a range of about a 2-year to about a 15-year frequency. ^'°^ Stated
another way, shifting the C-value up or down by 20% has the effect of
changing a 5-year design frequency to about a 15-year or 2-year frequency,
respectively.
If. all that is wanted is a design peak flow, and 1f_ a public works department
is not going to make any field rainfall-runoff or rainfall-runoff-quality
measurements, the rational method is as good an arbitrary procedure to use
as any, considering the primitive state-of-the-art of urban hydrology.
If. a design hydrograph is wanted but a public works department is not going
43
-------
to make any field measurements it is recommended that the unit hydro-
graph procedure synthesized from Louisville data by Eagleson^" ancj
expanded to include other data by Espey^ ' be used, such as in its
adaptation for metropolitan Denver.(109»110) The reliability of the
computed peak flow would likely be rather poor and not much better than
that obtained by the rational method. However, if alternative designs
are at issue the unit hydrograph approach is more likely to yield a more
reliable difference between the designs because it takes more land-use
characteristics into account. Many departments are faced with questions
which require immediate answers, on the effect of sub-catchment land-use
changes on total catchment drainage system adequacy, such as the extent
of new relief sewer construction that would be required to convert an
existing residential sector into a shopping center or other drastic
change. There is no reason why the unit hydrograph method should not be
used for such applications. Indeed, there is no logical exuse for not
using that method to examine the merits of one alternative against another
in any application.
J/f a design hydrograph is wanted and i_f_ a department js_ going to make some
field measurements, use of a range of tools becomes feasible. These tools
have been calibrated using most of the limited amount of data that is avail-
able. Outstanding among these, in order of simplicity of interpretation
and application, is the unit hydrograph, already cited, methods tested at
Purdue^ ' and the Road Research Laboratory method. ' However, the
transferability or applicability of these methods to catchments other than
those on which these methods were calibrated would be completely speculative
unless some local measurements were used to check their approximate local
relevance and to sharpen the parameter values involved for the purpose of
obtaining more consistent and reliable results.
It is important to note at this juncture that we have gone from the
simplistic rational method, which can be intelligently applied by sub-
professionals, to methods which require a good understanding of basic
hydrological principles and some mathematical transformations plus an
44
-------
ability to conduct field research and interpret the results of the
findings for use in design. Even the best of tools in unskilled hands
can be badly abused, if a department has a serious commitment to improved
or more flexible design it should have at least one well-qualified hydro-
logist on its staff. Qualifications can be achieved by a good engineer
with an initially sparse competence in about a month of full-time intensive
self-study. As the discussion in this paper continues, techniques will
be mentioned that require still greater capabilities that must be mastered
by staff personnel.
Because complex processes, such as in the hydro!ogical response of a
sewered catchment to a precipitation occurrence, can never be fully
replicated in a computation due to incomplete technical understanding of
the processes and the infeasibility of detailing the myriad pieces involved,
resort is made to simulation of response of a conceptually equivalent
system. The simulation package is commonly called a "model", and this term
is even applicable in principle to the primitive "rational method". As
you know, models can become quite sophisticated and involved. There are
instances of departments that assumed that the more complicated the
procedure the better the results. That is, applying a larger number of
guesses to a larger number of unknowns is somehow better than a smaller
number. Too often the claims of model builders about the universality of
their models has been accepted without question. Model builders are to be
listened to but not fully believed for they have a vested interest in the
survival of their progeny.
Use of models 1s a function of what you want to use them for, how you
want to use them, how much you are willing to invest, how often they would
be used, what levels of precision you want, what kind of outputs you
want, how much time you can spend in getting the model to work, how much
you can commit to calibrate and verify a model, etc., etc. While outside advice
can be obtained on the answers to these many questions, in the long run
only Department personnel can decide these matters if Department needs
are to be satisfied as the Department sees them.
45
-------
There are several elaborate models that are undergoing continuous up-
grading and improvement. Among these are the Stanford Watershed
Model^113»1l^and the EPA model.^115-118^ in the same complexity category
are the models developed by Dorsch of the Federal Republic of Germany and
by S06REAH of France, both of which are being promoted in the U.S. Early
in 1973, EPA formed a "users group" to create a forum for the exchange of
ideas and information relating to the EPA model and its use, and to enable
users to benefit from the experience of others in applying or modifying
the model to fit particular situations. ^19'
Occasionally the assertion is made that the totality of a jurisdiction
should be subjected to detailed modeling using a model of the type mentioned
immediately above. Most of our major cities have dozens and even hundreds
of stormwater drainage catchments with cumulative conduit lengths often
exceeding 1000-miles. For example, Milwaukee has 1,370-miles of separate
storm drains and combined sewers within the 97-sq. mi. of the City. 22 '
These conduits are distributed over 465 drainage catchments in Milwaukee
having a maximum size of 1,820-acres and a median size of 25-acres.'24 '
Additional examples are given in Tables 3 and 4. These examples should
suffice to indicate that an extremely compelling argument would be required
to justify an extremely detailed modeling for the existing facilities
of jurisdictions of these large sizes. However, under restrictions already
cited, detailed modeling is justified and completely appropriate in the
hydro!ogic-hydraulic design of extensions to existing systems and for
altogether new systems.
rf a department is make field measurements and is using comparatively
short-cut methods (such as the unit hydrograph, Purdue or RRL methods),
overall reliability and the physiological meaning of these methods in terms
of gaged observations can best be monitored and explained by the occasional
use of the more elaborate methods such as the Stanford Watershed or EPA
models. Department staff must be faimilar with functional model details,
however, or subtleties and aberations will be easily overlooked.
46
-------
TABLE 3
EXTENT OF UNDERGROUND DRAINAGE CONDUITS IN
SOME MAJOR CITIES**
Boston
Chi cago
Detroit
Los Angeles
Milwaukee
New York
Philadelphia
St. Louis
San Francisco
Washington, D.C.
Area of City,
Square Miles
48
224
139
460
95
320
130
62
44
61
Total Length
And/or Combined
1,360
3,600
2,900
860
1,370
4,130
2 ,500
1,120
870
1,750
of Storm
Sewers, Miles
(*Note: Data are not available for total metropolitan areas, but it may
be assumed that in most of the above instances the length of sewers in
the areas surrounding these cities are much greater).
TABLE 4
STORMWATER DRAINAGE-CATCHMENT AREA-SIZE
DISTRIBUTIONS FOR SOME MAJOR CITIES^ 24
City
San Francisco
Washington
Milwaukee
Houston
Total Area Number
of City, of
Sq. Mi. Catchments
44
61
95
444
42
93
465
1283
Largest
Drainage Area,
Acres
4330
6180
1820
2550
Median
Drainage Area
Acres
190
65
25
6
47
-------
In sum, elaborate models are of limited practical use in massive applica-
tions, such as in the preliminary design phase of a master plan, but
are invaluable in monitoring and explaining simpler methods and may be
useful in detailed final design vf department staff are skilled in their
use and i_f local data is available for their calibration. Continual im-
provements are being made in the more elaborate models, all of which
enhance their application potential.
Parenthetically, hybridization of models also may be useful. For example,
the water computer sub-routine of the EPA model can be "piggy-backed" on a
computer program for the simpler rainfall-runoff models such as the
unit hydroaraph, Purdue or RRL methods. (Researchers at Purdue added
water quality samplers at their rainfall-runoff field measurement stations
in 1972, and future data analyses will incorporate quality aspects, perhaps
including previously investigated models.) Hydrologic modeling of systems
of underground drainage which does not or cannot accommodate water quality
considerations must be judged of limited utility under contemporary
imperatives. On the other hand, for comprehensive capture and treatment
schemes, estimates of pollution loads may be needed only for design in the
few occasions when discharges/overflows would occur, and when all occurrences
are wanted - as in metropolitan-scale problem assessments - the simplest
techniques should be employed because of the enormous scale of the assess-
ment problem. To reiterate, limited and periodic use should be made of
more elaborate models to monitor the simpler methods, provided some local
data is available for calibration.
CHALLENGES IN THE USE OF MODELS FOR DESIGN
A survey has been made of existing raingage networks.'120»121 ' in 1969, 120''
fifteen of the largest metropolitan areas had a network of recording rain-
gages of between about 5 and 192 instruments, with records spanning periods
of 2- to about 50-years. Since 1970, an urban-weather research program^! 22'
with instrumentation that includes 250 recording raingages, the world's
first major field-laboratory program aimed at assessing the effect of a
48
-------
major urban-industrial complex on precipitation/ has been in progress
in the St.Louis metropolitan area. The most advanced automatic network
was installed in San Francisco in 1971, where rainfall rates from 30 gages
are transmitted at a ]5-second interval to a control data acquisition and
recording station/ '
Despite the not inconsequential amount of raingage network data accumulated
thus far, advances in the characterization of time-spatial characteristics
of storms over urban catchments have been slow in coining.
Human life is seldom threatened by the flooding of urban drainage facilities.
Because the principal local detrimental effects of flooding are damage to
the belowground sections of buildings and hindrance of traffic, the
consequences of flooding range from clearly assessable property destruction
to annoying inconvenience. It follows that provision of complete protection
from flooding can only rarely be justified. Instead, facilities are designed
which will be overtaxed infrequently. A monumental question in the analysis/
design of drainage systems is the choice of storms to be used. Storm
definitions used for deriving river basin extremes such as "reservoir design
floods" and "spillway design floods" are irrelevant because urban sewer
systems are expected to be overtaxed much more frequently than major river
structures whose failures could be catastrophic. From this standpoint,
the mean frequencies of occurrence of flow peaks and volumes and quality
constituent amounts is the issue, not the frequencies of the input rainfall,
and were it possible to arrive at statistical series for discharge-quality
independently of rainfall we could vastly simplify the storm characterization
issue.
Furthermore, because there are inherent non-linearities in most methods for
processing inputs for linear models, the dynamic models are non-linear by
definition, the statistics of the rainfall input array may differ appreciably
from the statistics of some or all of the arrays for runoff-quality charac-
teristics. That is, attempting to assign a mean frequency of probable
occurrence to a "design storm" is meaningless because of statistical non-
homogeneity of rainfall, runoff and quality. Also, such an approach neglects
the effect of prior storms on the runoff from a given storm/ '
49
-------
In a given application, it is more reasonable to route rainfall data of
local record through a model to arrive at rate-volume-constituent
frequencies than to synthesize a storm of specious probability. For
example, in developing the Master Plan for combined sewer overflow pollu-
tion abatement in San Francisco, preliminary indications of overflow
frequency were obtained by using hourly rainfalls of the 62-years of record
from the City's U.S. Weather Service gage. '
Over and above the storm definition problem is the inherent difficulty with
any runoff model in the necessarily subjective separation of abstractions
(infiltration, depression storage, etc.) from total rainfall to resolve
rainfall excess (amount and pattern), which is the input from which an
equal volume of direct runoff is generated by models of one kind or
another. This problem is greatly aggravated by the extreme scarcity of
field data for calibrating and verifying models of all types.
In the operating mode, any control system must not only respond almost
instantaneously to the actual occurrence of rainfall but must anticipate
the probable character of subsequent time and spatial changes barely
before they occur. A desirable adjunct would be an incipient storm-
occurrence forecasting capacility. As in design/analysis, separation of
abstractions from total rainfall to derive rainfall excess is necessarily
highly subjective, and delimits the relability of affected control-response
components. While detailed rainfall-runoff-quality modeling based on
real field data is both desirable and logical for development of control
criteria, existing models suffer substantially from a severe shortage of
data for their calibration, verification and, particularly, for their
realistic application to non-gaged catchments. As noted earlier, the
control mode can be, and probably must be, approached more pragmatically
than design/analysis, and the ratio of art to science required for the
automatic operational control of urban water services will probably be
highest for storm sewer systems.' To illustrate the expected
pragmatic emphasis, simplified flow-routing methods^ ' based on the
storage continuity equation will probably find much greater use than complel
50
-------
methods based on simultaneous solution of the partial differential equations
of flow. That is, control logic will tend to incorporate simplified flow-
routing methods while the validity, sensitivity and reliability of such
short-cuts will be monitored by selective use of more complete methods.
MASTER PLAN IMPLICATIONS
Exploration of metropolitan-wide pollution-control alternatives should
employ the simplest of mass-balance and mass-routing techniques, because
of the scale of the problem and the sparsity of suitable data for more
refined analysis. Total metropolitan water-balance inventories' 8^
can serve as an excellent base for such an approach, but very few have
been made or even started. While addressed to automatic control, the
approach used to develop the preliminary master plan for combined sewer
overflow control in San Francisco' ' is an excellent example of the
basic approach recommended. Modeling, in the traditional hydrological
sense, is of extremely limited value in the total metropolitan overview.
While zero-pollution was a concept that was not even given serious considera-
tion not too long ago, our new national policy has defined this as a goal.
Therefore, all metropolitan-wide alternatives should consider this goal
as the upper limit of objectives in developing any plans, together with
partial-accomplishment stages that should be adequately flexible to
permit approaching the ultimate solution over a series of compatible
phases of development.
For estimates of total rainfall-total runoff relations, the results
obtained for Gray Haven» Northwood, Oakdale and Boneyard Creek could be
very helpful.^129' These could be generated very simply using a local
Weather Service gage record. A possible extension of this information
would be in obtaining rainfall excess for unit hydrograph applications.
The writer was asked, by the City Engineer of a large community that was
about to initiate a master plan for combined sewer overflow abatement,
for suggestions on the development of a suitable basic data base for
meeting immediate subcatchment planning needs and for long-range master
51
-------
planning. Recommendations are summarized in Figure 4, divided into
four parts.
Item 3 in Figure 1, raingage network, refers to the advanced type
San Francisco, where (as noted above) signals from thirty raingages (and)
over a hundred water-level gages) are transmitted as frequently as every
15-seconds to a central data acquisition and recording station as a part of
a system developed for the City. HJ05.130 J ^ computer program
("SYMAP"), operated off-line, interpolates between raingage network readings
and prints out rainfall depths for a fine grid covering the entire City of
San Francisco.
The main reason for urging early installation for the raingage network,
Figure 4, was to permit ascertaining as soon as possible the predominant
spatial variations of rainfall peculiar to that metropolitan area, whatever
they might be. At present, there is only one first-order Weather Service
gage in that particular entire metropolis.
Items 1, 2 and 4 have been discussed earlier in this paper, but to reiterate
somewhat, of crucial importance is the fact that entirely different models
may be needed for assessing alternative land-use plans, for near-term
replacement of the rational method for design, for a preliminary master plan,
for implementing a master plan, and for other purposes that occur to a de-
partment as it proceeds to embrance all of its responsibilities in a compre-
hensive framework.
The City Engineer was advised that the writer suspected he would wish to
include, at some time, waterquality parameters (perhaps both concentrations
and event amounts), that peak flows will not be nearly as important as
occurrence volumes in some applications, and because not more than one element
can be maximized or minimized or optimized in any given calculation, either
a not now available very flexibile model or a group of different models will
be required. Further, because new storage will be required for almost any
overflow-capture scheme, a long-term objective would be acquisition and
mastery of models that would admit that feature for any point or points in
a catchment.
52
-------
DEVELOP
UNIT
HYDROGRAPHS
FOR COMPUTER
ANALYSES
2
, ESTIMATE
FLOW-QUALITY
MODELING
NEEDS
[APPLY IN URGENT
IAND-DSE DKJiAXSS, TO
INDICATE DIFFERENCES
BETWEEN ALTERNATIVE
[DEVELOPMENT BISECTIONS
'REVIEW DEVELOP ONE
'ALTERNATIVE OF MODELS
'MODELS ADOPTED REMOVE
'FOR EXPECTED COMPUTATIONAL BUGS,
JUSSS APPLT TO CASES
CONTINUE TO APPLY, WHILE
REFINING ASSUMED PARAMETER
VALUES VIA OBSERVED FLOW AMD
RAINFALL DATA, ITEM © AMD
POSSIBLY ITEM ®
AS FLOW-QUALITY DATA FROM (§) ACCUMULATES,
FIRST CALIBRATE, THEN VERIFY MODEL
( PRELIMINARY)* REVIEW SUITABILITY 07
CHOSEN MODEL TO OVERALL MASTER
PLAN OBJECTIVES
SHARPEN FURTHER
OR ABANDON IN ^*
FAVOR OF (Continued 7)
OTHER
METHODS FROM (j)
REFINE MODEL
OR ADD NEW
MODELS OE — - .....^»
START ALL (Continued)
OVER
STUDY | PREPARE
RAINGACE [SPECS. FOR
NETWORK (RAINGAGE
REQUIREMENTS (NETWORK
(EQUIPMENT
4
ESTABLISH j PREPARE
FLOW-QUALITY DATA SPEC. FOR
REQUIREMENTS ; .FIELD
LOCATE SUITABLE FLOW-QUAL.
CATCHMENTS STATIONS
AWARD CONTRACT AND
INSTALL RAINGAGE
NETWORK, TELEMETRY
AMD DATA ACQUISITION
AMD REDUCTION FACILITIES
RECORD, REDUCE AND ANALYZE RAINFALL
DATA FROM RAINGAGE NETWORK;
INPUT RESULTS TO © AND
POSSIBLY TO (p . (AREAL PATTERN
RELATIONSHIPS ARE CRUCIAL FOR MUSTER FLAN)
, ... , v
r
(Continued)
INSTALL AND RECORD AND REDUCE FLOW-QUALITY AND
TEST FIELD LOCAL RAINFALL CATCHMENT DATA; INPUT
FLOW-QUALITY RESULTS TO 0 AMD ©
STATION
EQUIPMENT
. .. _ _^.
(Continued)
en
CO
j I
J I
I I I I
1 I
j I
j I
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
MONTHS
FIGURE 4 A SUGGESTED BASIC DATA BASE FOR MEETING IMMEDIATE SUBCATCHMENT
PLANNING NEEDS AND FOR DEVELOPMENT OF A MASTER PLAN
-------
For the preliminary master plan evaluation, it was suggested that the
department consider seeking access to the simple but excellent rainfall-
runoff-quality model developed recently for the Corps of Engineers.^' '
Lastly, it was argued that for inputs for all models nothing short of an
approximation of the full rainfall record should be used. That is, around
20-years of U.S. Weather Service "excessive precipitation" data for the
City at 5-minute intervals (shortest period available) should be routed
through the model used, each time it is used, until some indication of the
use of a lesser, selected set of storm data becomes apparent. It was
pointed out that the City Engineer would be hard-put to explain an
artificial, synthetic storm's frequency to irate citizens who have been
flooded or to a State official regulating overflows. Defense against storms
of record is rather direct, and in the writer's opinion the only realistic
option open to a-public official. The temptation to use synthetic, phoney,
artificial input data should be resisted. The credibility of an actual
Weather Service gage record is far greater in a practical sense than con-
fections such as a "design storm".
SUMMARY AND CONCLUSION
For preliminary analysis and design of specific projects, the use of
the simplest models is recommended because of the tremendous detail
otherwise required and because of the dearth of calibration-verification
data available for more sophisticated models. Rather than embark on a
completely empirical adventure, it would be prudent to make ad hoc temporary
rainfall-runoff-quality measurements, to calibrate and verify the crude
models used. If actual field data is available, the use of detailed
modeling techniques would be justified, but only to give a better insight
on the underlying processes, and to develop a more rational understanding
of the performance of the simpler models used, although opportunities for
design application may arise. The guiding rule should be a compatibility
of tool sophistication with scale of use and quality of input data.
Artificialities, such as the intensity-duration-frequency curves used in
the"rational method", should be avoided and historical storm data should be
54
-------
used as inputs for all types of models.
All but the rational method require use of a computer for efficient cal-
culation. This point has been reserved until last in order to avoid over-
emphasizing the role of computers.
The importance of drainage design is implied by some findings of the
National Water Commission: the estimated replacement value of existing
storm and combined sewer systems alone is $36-billion; and the collective
costs of urban facilities over the past have represented seven-tenths of
total national expenditures for all water resource facilities, and this
proportion is expected to increase considerably in the future.(132) When
and if abatement of pollution from urban storm runoff becomes a national
reaility, these investments will be raised almost astronomically.
In closing, it appears appropriate to refer the reader to a recent biblio-
graphy, with abstracts, covering over six-hundred papers dealing with the
subject of urban hydrology/133' Lastly, use of unit hydrographs by the
Corps of Engineers for estimating flows from local drainage in conjunction
with river protection works is illustrated in the proceedings of a 1970
seminar.(134)
ACKNOWLEDGMENTS
The contents of this section of the paper have drawn heavily on an ASCE
Program report^ ' and a recent co-authored paper.' ' Preparation of
both of these and the present paper was supported by means of a contract
between ASCE and the Office of Water Resources Research, U. S. Department
of the Interior.
POSTSCRIPT
There are supplementary reports to Footnote 25^ ' and Footnote 26. ^ '
Features of both the San Francisco and Chicago master plans are outlined
(138}
in a recent ASCE Program Technical Memorandum/ ' Two other recent
(33 139 )
issues are relevant to the subject of urban streamf1ows.v ' '
55
-------
SOME MANAGEMENT POSSIBILITIES
In the past, drainage conduits were deliberately designed to accelerate
the movement of collected storm water to receiving water bodies via gravity
flow, and seldom was detention storage incorporated in systems because
of the preoccupation with rapid surface water removal. However, in-
system storage is being provided in a number of new systems and is being
added to some existing systems. Interest in in-system storage has come
about because of broader integrated facilities-planning attitudes. As
noted earlier, there is widespread interest in multi-purpose drainage
facilities that exploit opportunities for provision of add-on water-based
recreation, substantial increases in levels of flooding protection for
buildings at small increments of cost, use and reuse of storm water for
water supply via artificial groundwater recharge and related means, and
reduced and/or programmed pollution burdens at outfalls. These all require
utilization of storage.
Requirements for assessment of storm drainage flooding damage have been
studied, and alternatives to the direct disposal of stormwater runoff have
been explored.^ '
Peak flows of storm drainage facilities can be attenuated by a number of
structural and land-use control means, such as by the use of designed
pondingM42' and encouragement of land development site grading patterns
that will increase flow distances over unpaved areas. ' The turf areas
separating pavements have long been exploited to produce ponding of overland
flow, with consequent savings in drain size and reduction of peak outflows,
(144) (145) (146)
at airports, ' highway interchangesv ' and for shopping centers/ '
Also, recreational areas have been utilized for temporary detention
storage: '
Storm drainage damages are the consequence of storing water in the wrong
places, such as in the basements of homes. Hence, greater use of designed
storage in lieu of undersirable storage depends for justification on
relative protection from flooding afforded by different system plans at
56
-------
corresponding differences In cost. That 1s, trade-offs between benefits
and disbeneflts should be resolved by choosing among various mixes of flow-
acceleration and storage components. Unfortunately, the contemporary
absence of a satisfactory body of hydrologic and economic field data on
urban storm drainage system floods constitutes a monumental liability in
the assessment of those floods and their associated damages.
The cost of alleviating pollution from combined sewer overflows, by means
other than separation into independent storm and wastewater systems, would
be greater than the combined amount required to eliminate existing deficiencies
in wastewater treatment facilities and to provide new treatment capacities
to meet replacement and population growth needs between 1971 and 1974.l148'
Because twice as many people are served by separate storm sewer systems
as combined sewer systems, it is evident that the cost of alleviating pol-
lution from stormwater discharges would be even much more substantial.
Stated another way, the cost of alleviating pollution from combined sewer
overflows and storm sewer discharges would be at least twice the replacement
value of all such existing sewers.
The U. S. Government has been involved for several years in the development
of measures for countering pollution from combined sewer overflows. T49,15(W
The ultimate solution for the problem of pollution abatement from both urban
storm sewer discharges and combined sewer overflows is the treatment of
such flows prior to their release into receiving waters. Capturing, trans-
porting and treating all discharges/overflows at unattenuated flow rates
would require gigantic collection sewers, pumping stations and treatment
facilities - all of which would be used a very few hours per year. There-
fore, practically all schemes for systemwide discharge/overflow collection
and treatment incorporate some form of auxiliary storage for the purpose of
attenuating sudden collection and treatment facilities inflows, with the
objective of scaling down the size of such facilities to reasonable and
manageable proportions.
On the basis of limited indications,it appeared that complete capture of
57
-------
o
of combined sewer stormwater flows would require perhaps 300-m of
storage per hectare of drainage area. ' However, more recent estimates
3
place storage requirements nearer to about 600-m /ha. The amount of
storage required can be reduced by taking advantage of the fact that,
because of the area! and temporal varability of rainfall intensity, the
proportion of main sewer capacity in use at a given time varies between
various main sewers throughout a rainstorm.
Of three basic approaches involved in the most advanced comprehensive
attacks on the combined sewer overflow problem, one utilizes exploitation
of ambient storage in existing trunk sewers by manipulating add-on con-
strictions (movable dams or gates) in outfall sewers. The constrictions
are placed in the vicinity of existing regulators, and all flows other than
from rare rainfalls are released thereby at attenuated rates to existing
interceptors and thence to existing treatment facilities. That is, by
raising the constructions in those main sewers where the potential storage
volume is underused at a given time during a storm, and lowering the con-
strictions in advance of local inundating flow inrushes, the flows into
the interceptors can be modulated for more efficient use of the interceptors
reflected in a greater diversion of flows to the treatment plant via the
interceptors and consequent reduced overflows to the receiving waters. This
basic approach is essentially exemplified at Minneapolis-St. Paul/152r156'
Detroit,(157-160) Seattle,(161-164)and cleveland.(165)
A second basic approach incorporates new storage located at elevations
well below all street sewers, in the form of tunnels or vaults, and
new or auxiliary discharge/overflow treatment works. It might or might not
have ancillary features such as pretreatment basins or pumped-storage
power generation to offset the substantial energy requirements for eventual
lifting of flood waters from underground storage to the ground surface. A
bonus of this approach is that street trunk sewers would be converted into
manifolds of diversion diffusers, with the result that their effective
hydraulic capacity would be raised. This basic approach is exemplified in
the pioneering plan developed in Chicago(166,170) and 1n analagous plan
(113,114")
58
-------
alternatives under consideration elsewhere*
The third basic approach is embodied in a plan for San Francisor ^ '
which includes: a single, new combined-flow treatment plant; a number of
new detention reservoirs located immediately below the streets in the
upstream portions of a majority of catchments; a number of new shore-line
detention reservoirs; a deep cross-system storage and transmission tunnel;
and achievement of a total system fully automatic operational control
capability.
All three basic approaches incorporate plans for some degree of automatic
control because of their complexity/ 10^
Because combined sewer overflows occur very suddenly, any facilities
provided for treatment of potential overflows must be put on line almost
instantaneously. This means that they would have to be activated immediately
with the occurrence of any stormwater flow that would exceed interceptor
sewer capacity. Further, such plants might be idle more than about nine-
tenths of the year. Effectiveness of overflow pollution abatement using
treatment facilities designed specifically for that purpose therefore will
require some form of automatic operational control. Remote supervisory
control would quite likely not be adequately responsive. The control logic
required has yet to be developed, and it is possible that different metro-
politan sewer systems will require their own fairly unique logic development.
Pollution abatement of storm water from separate systems of storm drains by
means of treatment faces very nearly the same difficulties. In one sense,
requirements are more severe because all storm water must pass through new
and special treatment facilities, there being no interceptor sewers in such
systems to divert small storm occurrence flows to perennially operated
wastewater treatment plants.
Consideration may be given to the incorporation of stormwater conduits in
multiple-service tunnels. There are numerous existing utility tunnel systems
in Europe, Asia and North America.'172' Most utility tunnel concepts^1972)
deal with structures located fairly close to the ground surface, and include
59
-------
possibilities for limited incorporated of storm sewers/17^ As an
adjunct or alternative, consideration could be given to much deeper
locations, at least for main feeders and arteries, for power transportation,
surface water and wastewater, and for other services.
Lastly, possibilities exist in rainfall-deficient regions for deliberate
capture of urban storm water to augment water supplies, provided the
quality is suitable, a requirement that might be met by various managerial
i \
schemes/176' Manipulated recharge of groundwater supplies with surplus
surface water has been practiced for some time in some parts of such
regions. (177,178)
RESEARCH STATUS AND NEEDS
"The field of urban hydrology is almost devoid of modern research invest-
ment. 179 ' "More research is needed to develop understanding of the
whole stormwater pollution problem. This research should cover the hydrology,
the hydraulics, the treatment, the effects on receiving waters, and related
factors. "(l80) " ____ too few data have been collected to describe the
effect of urban and suburban development on flood runoff. As a conse-
quence, little is known about the rainfall -runoff process and still less
about rainfall -runoff-quality, particularly for sewered catchments. " ----
considering the huge relative investment in sewered systems and the almost
complete absence of hydrologic data on sewered systems, it has been obvious
for quite some time that this is the sector most woefully needing research
attention."^181^ Suitable data collected with properly coordinated instru-
mentation in networks representing a variety of climatic, topographic and
land-use conditions are virtually non-existent. Further, while all metro-
politan areas have underground systems of storm drainage, flooding of
waterways is a serious problem in some metropolises but negligible or non-
existent in others. Additionally, flood amelioration is preoccupied with
rainfall -runoff whereas rainfall -runoff-quality processes are of more
vital concern nationally.
A national program for acquisition of needed sewered catchment data has been
60
-------
proposed together with a documentation of its justification,'1' but
progress in implementation has been painfully slow. In anticipation of
the possible future availability of new data, considerations for modeling
sewered urban catchment rainfall-runoff-quality processes were detailedM82'
and considerations for characterizing rainfall time and spatial distributions
in future research were explored. *°*' In support of the latter, existing
raingage networks were surveyed.^184»185' The U.S. Geological Survey is
developing a program of data collection and studies to serve national needs
in urban hydrology*186*187' The Office of Water Resources Research has
developed a broad program of projected urban water resources research,-89'
including urban hydrology analyses.
The procedure used for design of storm sewers in the United States is almost
exclusively the "rational method.M^10^ The method yields only an estimated
peak flow but none of the other attributes of a hydrograph. Its many
limitations have been reviewed elsewhere. 101' A full hydrograph is needed
for the design of detention storage, for evaluation of pollutant burdens,
for design of stormwater pollution abatement facilities, for designing local
protection works along streams, such as pumping stations for passing local
drainage flows over levees and dikes, and as inputs for design of stream
and river development works. Also, quantification of the effects of urbaniza-
tion on the hydro!ogic regimen is dependent in many cases upon the avail-
ability of sewer outlet hydrographs. Further, as urban water management
problems become increasingly acute, the need for multiple-use of water
becomes more evident. In exchanging one use for another, for example by
using storm water as a source of water supply, knowledge of the time-histories
of flows and water qualities is essential for reliable design of transfer
facilities.
By 1969, only a very few sewered catchments had been gaged; records seldom
extended over more than a very few seasons; although rainfall-runoff had
been measured, for still fewer of the catchments had rainfall-runoff-quality
measurements been made; runoff and runoff-quality measurements had been
made at outfalls but not in-system; and data reliability was often suspect
61
-------
because of known instrument errors, lack of observations synchronization,
and too crude signal resolution.^106' Proportionally more rainfall-
runoff data was available for partially sewered areas / but ^ amount
was nevertheless pitifully small. Since then, field research sponsored by
the Environmental Protection Agency^9J an(j others has extended the data
base somewhat. 1^0 > However, not much new data has been acquired on
completely sewered catchments using flumes or weirs rather than stage-gages
for determining discharge.
All hydrologic models require the application of some degree of subjective
judgment in the assignment of numerical values to model component parameters.
Reliability of judgment is improved when parameter values are based upon
fits to observed events for the catchment being modeled, and is maximized
when parameter values can be generalized for a number of catchments in
terms of physical features such as degree of imperviousness, type of land
use and channel density. The usual purpose of modeling is to project perfor-
mance for future events. For urban catchments, this usually involves physical
features such as degree of imperviousness, type of land use and channel
density. The usual purpose of modeling is to project performance for
future events. For urban catchments, this usually involves physical
changes over time such as revised land use, new detention storage and
modified land-management practices. Independently from any consideration
of catchment changes, recorded precipitation must be adapted in some way
to represent expected future precipitation. Thus, projections of runoff-
quality events inevitably require the application of some degree of judg-
ment. The principal reason for testing or calibrating a model with
observed data is to enhance confidence in its use, and this approach of
going from the present to the future is employed in all water resource
computations involving risk and uncertainty.
On top of all of the above is the inherent difficulty with any runoff
model in the necessarily subjective separation of abstractions (infiltration,
depression storage, etc.) from total rainfall to resolve rainfall excess
(amount and pattern), which is the input from which an equal volume of direct
62
-------
runoff is generated by models of one kind or another.
Nine projects were identified in 1970 where the urban rainfall-runoff
Process was being modeled for sewered and partially sewered catchments'190'
A Project completed since'115'117^ has developed the only existing model
that readily accommodates water quality parameters. The model was
eveloped as an assessment technique for comparing alternative solutions
y means of a comprehensive computer program capable of "representing urban
stormwater runoff phenomena, both quantity and quality, from the onset of
Precipitation on the basin, through collection, conveyance (both combined
and separate systems), storage, and treatment systems to points downstream
from outfalls which are significantly affected by storm discharges."^91)
An 1mProved version of the model has been tested for a 1500-ha. San
ranciscoM9l) sewered catchment.^118' These and any other models purportedly
^eloped for sewer application suffer substantially from a severe shortage
of data for their calibration, verification and, particularly, for their
realistic application to non-gaged catchments.
n default of suitable data, some astute planners have applied necessarily
weakly-founded generaiizeci unit hydrograph parameters^ 107»108,22Pirical indications for urbanized catchments for which suitable data
Were nonexistent.^193^ A unit hydrograph has benn defined for use in
linage planning for the Denver metropolitan area, Colorado. 010,109,33)
Se of unit hydrographs by the Corps of Engineers for estimating flows from
.Ocal linage in conjunction with river protection works is illustrated
ln the proceedings of a 1970 seminar.<194)
sewered catchment runoff generation technique developed in the United
Kln9dom has been tested using some U.S. data.(l40'112)
63
-------
A dearth of historic runoff records on urban streams, compounded by the
rapid changes that have taken place in urban watersheds, has thwarted
effective use of existing simulation models for planning.
A mathematical watershed simulation model that operates with continuous,
multi-seasonal hydrologic data was used in a study of the effects of
rainfall data time-interval and raingage density on simulated hydrographs
compared with observed hydrographs for four sewered catchments. '
Experimentation was severely hampered by the almost total absence of
appropriate data.
An empirical total rainfall-total runoff relation has been developed that
approximates the runoff volumes measured at the outlets of four sewered
(i ?Q)
urban catchments. '
A recently developed model requires as computational inputs: the ratio of
direct runoff volume to the volume of rainfall subsequent to the beginning
of direct runoff; and the associated duration of rainfall excess/195'
For predictive application these two factors must be guessed where no
runoff records are available or generalized in some way in the rather
unique instances where some runoff data exist.
An elegant mathematical model has been developed for routing intercepted
stormwater flows in an interceptor sewer of a combined system,''^) and
was verified in part using unsteady flow laboratory test findings. ^'
ACKNOWLEDGEMENT
This material was taken from a chapter prepared for the IHD and supported
by means of a contract between the American Society of Civil Engineers
and the Office of Water Resources Research, U. S. Department, of the
Interior. The first draft of this chapter, distributed for U.S. review,
was also supported entirely by OWRR, ' Supplementary considerations
were reported in a later paper. '
64
-------
REFERENCES
(Note: The notation "NTIS" In some footnotes refers to the National
Technical Information Service, U.S. Department of Commerce,
Springfield, Virginia 22151, where copies of the item cited may
be ordered).
1. American Society of Civil Engineers, Basic Information Needs in
Urban Hydrology, a report to the U.S. Geological Survey, New York,
N.Y., April, 1969. (NTIS Id. No. PB 185 442).
2. Advisory Commission on Intergovernmental Relations, Intergovernmental
Responsibilities for Water Supply and Sewage Disposal in Metropolitan
Areas. Report A-13, Washington, D. C., October, 1962, p. 40.
3. Jones, D. Earl, Jr., "Urban Hydrology - A Redirection," Civil Engineer!ngT
Vol. 37, No. 8, August, 1967.
4. Thomas, H. E., and W. J. Schneider, Water as an Urban Resource and
Nuisance, U.S. Geological Survey, circular 601-D, GPO, Washington, D.C.,
1970.
5. Sullivan, Richard H., "Inventory of Combined Sewer Facilities," ^ivil
Engineering, Vol. 38, No. 11, November, 1968.
6- U. S. Public Health Service, Pollutional Effects of Stormwater and
Overflows from Combined Sewer Systems - A Preliminary Appraisal.
Publication No. 1246, GPO, Washington, D. C., 1964.
7. American Society of Civil Engineers, Combined Sewer Separation Using
Pressure Sewers. Federal Water Pollution Control Administration, U.S.
Dept. of the Interior, Water Pollution Control Research Series ORD-4,
New York, N.Y., October, 1969, p. 9.
8. Weibel, S. R., R. J. Anderson and R. L. Woodward, "Urban Land Runoff
as a Factor in Stream Pollution," J.WPCF, Vol. 36, No. 7, July, 1964.
9- Cywin, Allen, and William A. Rosenkranz, Advances in Storm and
Combined Sewer Pollution Control Abatement, Office of Research and
Monitoring, Environmental Protection Agency, Washington, D.C., 27 pp.
October, 1971.
65
-------
10. Council on Environmental Quality, Environmental Quality, Third Annual
Report, GPO, Washington, D.C., August, 1972, p. 16.
11. Department of Public Works, "San Francisco Master Plan for Waste Water
Management - Preliminary Comprehensive Report," City and County of San
Francisco, September 15, 1971, p. VIII-2.
12. Abrams, Charles, "The Uses of Land in Cities," Scientific American.
Vol. 213, No. 3, September, 1965, p.154.
13. Department of State, U.S. National Report on the Human Environment. Dept.
of State Publication 8588, GPO, Washington, D.C., 1971, p. 9.
14. The Editors, "The Seventies: Super but Seething," A Special Report,
Business Week. December 6, 1969, pp. 98 and 211.
15. Advisory Commission on Intergovernmental Relations, Urban American and the
Federal System, Report M-47, Washington, D.C., October, 1969, p.85.
16. Task Force, "Effect of Urban Development on Flood Discharges - Current
Knowledge and Future Needs," Journal Hydraulics Division. ASCE Proc.,
Vol. 95, No. HY1, Paper 6355, January, 1969.
17. Franke, 0. L., and N. E. McClymonds, Summary of the Hydro!ogic Situation on
Long Is!and. New York, as a Guide to Water-Management Alternatives. U.S.
Geological Survey Professional Paper 627-F, GPO, Washington, D.C., 1972.
18. Hammer, Thomas R., "Stream Channel Enlargement Due to Urbanization," Water
Resources Research, Vol. 8, No. 6, December, 1972.
19. Soil Conservation Service, Section 4, "Hydrology," in SCS National Engineer-
ing Handbook, U.S.Dept. of Agriculture,GPO,Washington,D.C., January, 1971,
p. 15-11.
20. Stankowski, Stephen J., Population Density as an Indirect Indicator of
Urban and Suburban Land-Surface Modifications. U.S. Geological Survey
Professional Paper 800-B, GPO, Washington, D.C., 1972.
21. National Park Service, Man-Nature-City, U.S.Dept. of the Interior, GPO,
Washington, D.C., 1971, pp. 6 and 7.
22. Prawdzik, Ted B., "Environmental and Technical Factors for Open Drainage
Channels in Milwaukee," ASCE Urban Water Resources Research Program,
Technical Memorandum No. 12, ASCE, New York, N.Y., February, 1970.
(NTIS Id. No. PB 191 710).
66
-------
23. Leopold, Luna B., M. Gordon Wolman and John P. filler, Fluvial
Processes In Geomorpholoqy, W. H. Freeman and Co., San Francisco,
California, 1964, p. 145.
24. Tucker, L. S., "Sewered Drainage Catchments in Major Cities," ASCE
Urban Water Resources Research Program, Technical Memorandum No. 10,
ASCE, New York, N.Y., March 31, 1969. (NTIS Id. No. PB 184 705).
25. Hallmark, Dasel E., and John G. Hendrickson, Jr., "Study of Approxi-
mate Lengths and Sizes of Combined Sewers in Major Metropolitan
Centers," ASCE Combined Sewer Separation Project, Technical Memorandum
No. 4, ASCE, New York, N.Y., May 1, 1967. (NTIS Id. No. PB 185 999).
26. American Society of Civil Engineers, Design and Construction of
Sanitary and Storm Sewers, Manuals and Reports on Engineering Practice
No. 37, New York, N.Y., 1969.
27. Waananen, Arvi 0., "Urban Effects on Water Yield," pp. 169-182, in
Effects of Watershed Changes on Streamflow. edited by W. L. Moore and
C. W. Morgan, University of Texas Press, Austin and London, 1969.
28. Tucker, L. S., "Northwood Gaging Installation, Baltimore-Instrumentation
and Data," ASCE Urban Water Resources Research Program, Technical
Memorandum No. 1, New York, N.Y., August 1, 1968, p. a9. (NTIS Id.
No. PB 182 786).
29. Wenzel, H. G., Jr., "A Critical Review of Methods of Measuring Discharge
Within a Sewer Pipe," ASCE Urban Water Resources Research Program,
Technical Memorandum No. 4., New York* N.Y., September, 1968.
(NTIS Id. No. PB 182 789).
30. Wallace, James R., "The Effects of Land Use Change on the Hydrology of
an Urban Watershed," School of Civil Engineering, Report ERC-0871,
Georgia Institute of Technology, Atlanta, Georgia, 66-pp., October,
1971. (NTIS Id, No. PB 206 426).
31. Riley, J. Paul, and V. V. Dhruva Narayana, "Modeling the Runoff
Characteristics of an Urban Watershed by Means of an Analog Computer";
E. F. Brater and Suresh Sangal, "Effects of Urbanization on Peak Flows";
William H. Espey, Jr., David E. Winslow and Carl W. Morgan, "Urban
67
-------
Effects on the Unit Hydrograph"; and Donald VanSickle, "Experience
with the Evaluation of Urban Effects for Drainage Design": all in
Section 3, "Urban Watersheds," pp. 183-254, Effects of Watershed
Changes on Streamflow. 1969, op. cit.
32. Water Resources Council, The Nation's Mater Resources. GPO, Washington,
D.C., 1968. pp. 5-2-2 and 5-2-7.
33. Goddard, James E., "An Evaluation of Urban Flood Plains," ASCE Urban
Water Resources Research Program, Technical Memorandum No. 19,
New York, N.Y., December* 1973. (NTIS Id, No, PB 227 337).
34. Poertner, H. G., R. L. Anderson and K. W. Wolf, Urban Drainage Practices
Procedures» and Needs. American Public Works Association, Special
Report, Project 119, Chicago, 111., December, 1966, p. 3.
35, Sheaffer, J. R., D. W. Ellis and A. M. Spieker, Flood-Hazard Mapping in
Metropolitan Chicago, U. S. Geological Survey, Circular 601-C, GPO,
Washington, D. C., 1970.
36. Wood* Walter J., "Los Angeles County Flood Control System and the
Early 1969 Storms," Civil Engineering, Vol. 40, No. 1, January, 1970.
37. Rantz, S. E., Urban Sprawl and Flooding In Southern California. U.S.
Geological Survey, Circular 601-B, GPO, Washington, D. C., 1970.
38. James, L. Douglas, Eugene A. Laurent and Duane W. Hill, "The Flood
Plain as a Residential Choice: Resident Attitudes and Perceptions and
Their Implications to Flood Plain Management Policy," Environmental
Resources Center, Report ERC-0671, Georgia Institute of Technology,
Atlanta, Georgia, 288-pp., October, 1971. (NTIS Id. No. PB 206 424).
39. James, L. Douglas, "A Decision-Sequence Model of Residential Occupancy
of Urban Flood Plains," pp. 331-340, Vol. 2, Proceedings of Interna-
tional Symposium on Modelling Techniques in Water Resources Systems.
Hay 9-12, 1972, Ottawa, Ontario, Canada.
40. McPherson, M. B., "Management Problems in Metropolitan Water Resource
Operations," ASCE Urban Water Resources Research Program, Technical
Memorandum No. 14, Mew York* N.Y., September, 1971, p. 8. (NTIS Id.
No. PB 206 087).
68
-------
41. Kneese, Allen V., and Blair T. Bower, Managing Water Quality:
Economics, Technology, Institutions, Resources for the Future,
Inc., via The Johns Hopkins Press, Baltimore, Maryland 1968, p. 3.
42. McGauhey, P. H., Engineering Management of Water Quality, McGraw-
Hill Book Company, New York, N. Y., 1968, pp. 277-278.
43. National Academy of Engineering - National Academy of Sciences,
Policies for Solid Waste Disposal, Bureau of Solid Waste Management,
U. S. Dept. of H.E.W., Public Health Service Publication No.
2018, GPO, Washington, D. C., 1970, p. 12.
44. U. S. Congress, House of Representatives, The Cost of Clean Water.
Vols. I and II, House Doc. No. 92-23, GPO, Washington, D.C.,
1971, pp. 64-65.
45. U. S. Congress, The Cost of Clean Water, 1971, ibid., p. 48.
46. U. S. Congress, The Cost of Clean Water, 1971, ibid., p. 53.
47. U. S. Environmental Protection Agency, "Subsurface Water Pollution,
A Selected Annotated Bibliography," Office of Water Programs,
Washington, D. C., March, 1972: Part I, "Subsurface Waste
Injection" (NTIS Id. No. PB 211 340); Part II, "Saline Water
Intrusion" (NTIS Id. No. PB 211 341); and Part III, Percolation from
Surface Sources" (NTIS Id. No. PB 211 342).
48. Federal Water Quality Administration, Clean Water for the 1970's.
A Status Report, U.S. Dept. of the Interior, GPO, Washington,D.C.,
June, 1970, pp. 4-5
49. U.S. Congress, The Cost of Clean Water, 1971, op, cit., p. 54.
50. The Council of State Governments, Environmental Quality and State
Governments, Lexington, Kentucky, December, 1970, p. 7.
51. Graeser, Henry J., "America's Drinking Water Is...Is Not...Safe,"
The American City. June, 1970, p. 146.
52. McPherson, M.B., "ASCE Combined Sewer Separation Project Progress,"
Civil Engineering, ASCE, Vol. 37, No. 12, December, 1967.
53. Guarino, C. F., and J. V. Radziul, "Combined Sewers & Stormwater
Drainage - Today & 2050 A.D.," in Clean Water for Our Future
Environment. ASCE, New York, N.Y., 1971, p. 352.
69
-------
54. Rosenkranz, William A., and Francis J. Condon, (Research and Monitoring,
EPA, Washington, D.C. 20460), "Technology Improvements Related to
Storm and Combined Sewer Pollution Control," a paper presented at the
ASCE National Meeting, Atlanta, Georgia, January 25, 1972.
56* The Franklin Institute Research Laboratories, Selected Urban Storm
Water Runoff Abstracts, July 1968-June 1970, Environmental Protection
Agency, Water Pollution Control Research Series No. 11024EJC07/70,
GPO, Washington, D. C., 375 pp., July, 1970.
56. Syracuse University, "Street Sal ting - Urban Water Quality Workshop',"
Proceedings, Edited by R. H. Hawkins, State University College of
Forestry, Syracuse, N.Y., July, 1971.
57. Edison Water Quality Laboratory, Environmental Impact of Highway
Deicing, EPA Report 11040GKK06/71, GPO, Washington, D. C., June, 1971.
58. Pimental, David, Ecological Effects of Pesticides on Non-Target Species,
Office of Science and Technology, Executive Office of the President,
GPO, Washington, D. C., June, 1971.
59. Weibel, S. R., R. B. Weidner and J. M. Cohen, "Pesticides and Other
Contaminants in Rainfall and Runoff," J.AWWA, Vol. 58, No. 8, August, 1966,
60. Burd, R. S., "A Study of Sludge Handling and Disposal," Federal Water
Pollution Control Administration (now in EPA), Washington, D.C., 1968.
61. Anonymous, "Chicago Ships Its Sludge to the Farm!and," Engineering
News-Record, Vol. 186, No. 5, February 4, 1971, pp. 22 and 23.
62. Anonymous, "Chicago Reclaiming Strip Mines with Sewage Sludge," Cjvil
Engineering, ASCE, Vol. 42, No. 9, September, 1972, p. 98.
63. Ketchum, Bostwick H., "Ecological Effects of Sewage Sludge Disposal
at Sea," a paper presented at the WPCF Annual Conference, Boston,
Mass., October 7, 1970.
64. Anonymous, "Open Dumping: Pollution or Fertilization?," Water and
Wastes Engineering, Vol. 8, No. 4, April, 1971, pp. 53 and 54.
65. Committee, "Disposal of Wastes from Water Treatment Plants," in four
parts, J.AWWA: Vol. 61, No. 10, October, 1969; Vol. 61, No. 11,
November, 1969; Vol. 61, No. 12, December, 1969; and Vol. 62, No. 1,
January, 1970. (Available from NTIS, Id. No. PB 186 157).
70
-------
66. Krasauskas, John W., "Review of Sludge Disposal Practices," J.AWWA,
Vol. 61, No. 5, May, 1969, p. 225.
67. Evans, James 0., "The Soil as a Resource Renovator," Environmental
Science and Technology. Vol. 4, No. 9, pp. 732-735, September, 1970.
68. Lewicke, C. K., "Recycling Sludge and Sewage Effluents by Land
Disposal," Environmental Science and Technology, Vol. 6, No, 10,
pp. 871-873, October, 1972.
69, Schneider, William J., Hvdrologi'c Implications of Solid-Waste Disposal,
U. S. Geological Survey, Circular 601-F, GPO, Washington, D.C., 1970.
70. Coe, Jack J., "Effect of Solid Waste Disposal on Ground Water Quality,"
J.AWWA. Vol. 62, No. 12, December, 1970.
71. Zononi, A. E., "Ground-water Pollution and Sanitary Landfills - A
Critical Review," Ground Water. Vol. 10, No. 1, Jan.-Feb., 1972.
72. Hughes, George M., and Keros Cartwright, "Scientific and Administrative
Criteria for Shallow Waste Disposal," Civil Engineering, ASCE, Vol. 42,
No. 3, March, 1972.
73. Council on Environmental Quality, August, 1970, op. cit., p. 107.
74. Council on Environmental Quality, August, 1970, op. cit.. p. 58.
75. Office of Water Resources Research, "Urbanization and Sedimentation,
A Bibliography," U. S. Dept. of the Interior, Washington, D.C.,
October, 1971. (NTIS Id. No. PB 203 IBS).
76. Powell, M. D., U. C. Winter and W. P. Bodwitch, Community Acffon_
Guidebook for Soil Erosion and Sediment Control, National Association
of Counties Research Foundation, Washington, D. C., March, 1970, p. 6-3.
77. Department of State, 1971, OP. cit., p. 9.
78. Guy, Harold P., Sediment Problems in Urban Areas, U.S. Geological
Survey, Circular 601-E, GPO, Washington, D. C., 1970, p. E-2.
79. Wolman, M. Gordon, "The Nation's Rivers," J..MPCF, Vol. 44, No. 5,
May, 1972.
80. Task Committee, "Influences of Sedimentation on Water Quality: An
Inventroy of Research Needs," Journal'Hydraulics Division, ASCE Proc.,
Vol. 97, No. HY8, Paper 8325, August, 1971.
81, Office of Water Resources Research, 1970 "Annual Report, U.S.Dept. of
the Interior, GPO, Washington, D. C., 1971, p. 20.
71
-------
82. Rantz, S. E., Urban Sprawl and Flooding In Southern California,
U.S.Geological Survey, Circular 601-B, GPO, Washington, D.C., 1970.
83. Guy, Harold P., "Flood Flow Downstream from Slide," Journal Hydraulics
Division, ASCE,Proc., Vol. 97, No. HY4, Paper 8018, Technical Notes,
April, 1971.
84. Guy, Harold P., and D. Earl Jones, Jr., "Urban Sedimentation in
Perspective," Journal Hydraulics Division, ASCE Proc., Vol. 98, No. 12,
Paper 9420, December, 1972.
85. Ehrlich, Paul R., and John P. Holdren, "The Energy Crisis," Saturday
Review, August 7, 1971, p. 50.
86. Council on Environmental Quality, August, 1970, op. sit., pp. 52-53.
87.m Levin, A. A., T. J. Birch, R. E. Hillman and G. E. Raines, "Thermal
Discharges: Ecological Effects," Environmental Science and Technology.
Vol. 6, No. 3, March, 1972, p. 225~
88. Anonymous, "The High Price of Supplying Power," Business Week, June 12,
1971, p. 64.
89. Brown, F. Stewart, "Waste Heat Disposal from Power Generating Stations "
Journal Power Division, ASCE Proc., Vol. 96, No. P03, Paper 7317
June, 1970.
90. Dallaire, Eugene E., "Thermal-Pollution Threat Draws Nearer," Civil
Engineering, Vol. 40, No. 10, pp. 67-71, October, 1970.
91. Hauser, Louis G., "Cooling Water Sources for Power Generation,"
Journal Power Division, ASCE Proc., Vol. 97, No. P01, Paper 7804,
January 1971.
92. Committee on Power Plant Siting, Engineering for Resolution of th»
Energy-Environment Dilemma: A Summary. National Academy of
Engineering, Washington, D. C., 1971.
93. Anonymous, "The Crisis of Turning Energy Into Power.", Business VtePk,
January 8, 1972, pp. 58 and 60. "'
94, Advisory Commission on Intergovernmental Relations, Federalism in 1971-
The Crisis Continues, Report M-73, Washington, D.C., February, 1972"
9B. Commission on Population Growth and the American Future, Themes and'
Hiqhliqhts (of the Commission's Final Report), GPO, Washington D C™
1972, p. 7. '
72
-------
96. Dreyfus, Daniel A., "Needed: A National Land Use Policy," CJvil
Engineering. Vol. 42» No. 5, pp. 52-54» May* 1972.
97. Citizen's Advisory Council on Environmental Quality, Annual Report to
The President and to the Council on Environmental Quality* GPO,
Washington, D.C., 1972, p. 13.
98. APWA, Problems of Combined Sewer Facilities and Overflows. FWPCA,
Water Pollution Control Research Series, WP-20-11, GPO, Washington,
D.C., December 1, 1967.
99. APWA, Water Pollution Aspects of Urban Runoff. FWPCA, Water Pollution
Control Research Series, WP-20-15, GPO, Washington, D.C., January, 1969,
100. McPherson, M. B., "Urban Runoff," ASCE Urban Water Resources Research
Program, Technical Memorandum No. 18, ASCE, New York, N.Y,, August,
1972. {Available from NTIS as PB 212 580).
101. McPherson, M. B., "Some Notes on the Rational Method of Storm Drainage
Design," ASCE Urban Water Resources Research Program, Technical
Memorandum No. 6, ASCE, New York, N.Y., January 22, 1969. (Available
from NTIS as PB T84 701).
702. Leopold, Luna B., Hydrology for Urban Lanj Planning - A Guidebook on
the Hydrologic Effects of Urban Land Use, U.S.Geological Survey,
Circular 554, GPO, Washington, D.C., 1968.
!03- Koelzer, Victor A., William J. Bauer and Frank E. Dalton, "The Chicago
Area Deep Tunnel Project - Use of the Underground Storage Resource,"
J.WPCF, Vol. 41, No. 4, April, 1969.
104. F1ood Control Coordinating Committee, Development of a Flood and
Pollution Control Plan for the Chicagoland Area, Sumnarv of Technical
Reports. Chicago, Illinois, 111 pp. and 27 exhibits, 1972.
105. Anonymous, "Computerized Monitoring System Helps San Francisco Control
Overflow," Water and Sewage Works. Vol. 119, No. B, August, 1972.
106. McPherson, M. B., "Feasibility of the Metropolitan Water Intelligence
System Concept (Integrated Automatic Operational Control)," ASCE Urban
Water Resources Research Program, Technical Memorandum No. 15, ASCE,
New York, N.Y., December, 1971. (Available from NTIS as PB 207 301).
73
-------
107. Eagleson, Peter S., "Unit Hydrograph Characteristics for Sewered Areas,"
Journal Hydraulics Division, ASCE Proc., Vol. 88, No. HY2, Paper 3069,
March, 1962. Author's Closure to Discussion, Vol. 89, No. HY4, Paper
3580, July, 1963.
103 Espey, W. H., Jr., and D. E. Winslow, "The Effects of Urbanization on
Unit Hydrographs for Small Watersheds, Houston, Texas," A Report to
OWRR, TRACOR Document No. 68-975-U, Austin, Texas, September 25, 1968.
109 Claycomb, Elmer L., "Urban Storm Drainage'Criteria Manual from Denver,"
Civil Engineering, Vol. 40, No. 7, July, 1970.
110. Wright-Mclaughlin Engineers, "Urban Storm Drainage Criteria Manual,"
Denver Regional Council of Governments, Denver, Colorado, 1969.
jl-l Rao, A. R., J. W. Delleur and B. S. P. Sarma, "Conceptual Hydrologic
Models for Urbanizing Basins," Journal Hydraulics Division, ASCE Proc.,
Vol. 98, No. HY7, Paper 9024, July, 1972.
112. Terstriep, M. L., and J. B. Stall, "Urban Runoff by Road Research
Laboratory Method," Journal Hydraulics Division, ASCE Proc., Vol. 95,
No. HY6, November, 1969. Authors' closure to discussion, Vol. 97,
No. HY4, Paper 8012, April, 1971.
113, Crawford, Norman H., "Studies in the Application of Digital Simulation
to Urban Hydrology," Hydrocomp International, Palo Alto, California,
100 pp., September, 1971. (NTIS Id. No. PB 206 715).
114. Crawford, Norman H., "Computer Simulation for Design Criteria for
Urban Flow Storage Systems," Hydrocomp International, Palo Alto,
California, 109 pp., January, 1973. (NTIS Id. No. PB 218 827).
115. Water Quality Office, Storm Water Management Model, Environmental
Protection Agency, Water Pollution Control Research Series, 1n four
volumes (GPO, Washington, D.C.):
Vol. I - "Final Report," 11024DOC07/71, 352 pp., July, 1971.
Vol. II - "Verification and Testing," 11024DOC08/71,172 pp., August,1971
Vol. Ill - "User's Manual," 11024DOC09/71, 359 pp., September, 1971.
Vol. IV - "Program Listing," 11024DOC10/71, 249 pp., October, 1971.
116. Chen, Carl W., and Robert P. Shubinski, "Computer Simulation of Urban
Storm Water Runoff," Journal Hydraulics Division. ASCE Proc., Vol. 97,
No. HY2, Paper 7924, February, 1971. Authors' closure to discussion,
Vol. 98, No. HY5, Paper 8859, May, 1972.
74
-------
117. Lager, John A., Robert P. Shubinski and Larry W. Russell, "Development
of Simulation Model for Stormwater Management," J.WPCF, Vol. 43,
No. 12, December, 1971.
118. Roesner, Larry A., David F. Klblerand John R. Monser, "Use of Storm
Drainage Models in Urban Planning," National Symposium on Watersheds
in Transition, June 19-22, 1972, Colorado State University, Proceedings,,
American Water Resources Association, Urbana, Illinois, pp. 400-405,
1973.
119. "Storm Water Management Modeler," Harry C. Torno, Editor, an Information
bulletin, Municipal Technology Branch, Technology Division, Office of
Research and Monitoring, EPA, Washington, D. C. 20460. (First issue was
•
March 5, 1973, and the first users' conference was held 1n Arlington,
Virginia, April 2, 1973, where one user noted that his firm had modified
the model to account for the first-flush phenomenon and to accommodate
surcharged and interconnected sewers
120. Tucker, L. S., "Ralngage Networks in the Largest Cities," ASCE Urban
Water Resources Research Program, Technical Memorandum No. 9, ASCE
New York, N.Y., March 17, 1969. (Available from NTIS as PB 184 704).
121. Tucker, L. S., "Non-Metropolitan Dense Raingage Networks," ASCE Urban
Water Resources Research Program, Technical Memorandum No. 11, ASCE,
New York, N.Y., January, 1970. (Available from NTIS as PB 191 709).
122. Changnon, Stanley A., Jr., Floyd A. Huff and Richard G. Semonin,
"METROMEX: an Investigation of Inadvertent Weather Modification",
Bulletin American Meteorological Society, Vol. 52, No. 10, October, 1971,
123, Changnon, Stanley A., Jr., Richard G. Semonin and William P. Lowry,
"Results from METROMEX," Preprint Volume of the Conference on Urban
Environment and the Second Conference on Biometeorology, October 31-
November 2, 1972, Philadelphia, published by AMS, Boston. (Illinois
State Water Survey Reprint Series, No. 219, Urbana, 111.).
124. Linsley, Ray K., Discussion of "Urban Runoff by Road Research Laboratory
Method," J. Hyd. Div., ASCE Proc., Vol. 96, No. HY4, Paper 7189, April,
1970.
125. Sauer, Vernon B., "Unit-Response Method of Open-Channel Flow Routing,"
J. Hyd. Div.. ASCE Proc., Vol. 99, No. HY1, Paper 9499, January, 1973.
75
-------
126. Koch, C. Thomas, "Methodology for Managing Resources," J.AWWA.
Vol. 64, No. 2, February, 1972.
127. Koch, C. Thomas, "Systems Description for Urban Water Resources,"
pp. 19-41 in Proceedings. 16th Annual Conference on Water for
Texas, "Urban Water Resource Planning and Management," Texas A. and
M. University, College Station, September, 1971. (NTIS Id. No.
PB 210 325).
128. McPherson, M. B., "Need for Metropolitan Water Balance Inventories,"
J. Hyd. Div., ASCE Proc., Vol. 99, No. HY10, Paper 10078,
October, 1973.
129. Miller, Clayton R., and Warren Viessman, Jr., "Runoff Volumes from
Small Urban Water Sheds," Water Resources Research. Vol. 8, No. 2,
April, 1972.
130. Control System Industries, "Hydrologic and Hydraulic Data Acquisition
and Control System," Bulletins SB-103 and PB-101, Mountain View,
Cal., and Waltham, Mass., 1971.
131. Hydrologic Engineering Center, Corps of Engineers, "Urban Storm
Water Runoff: 'STORM111, Generalized Computer Program 723-S8-L2520,
Davis, California, May, 1974.
132. Koelzer, Victor A., "Urban Water Management," J.AWWA. Vol. 64,
No. 9, September, 1972.
133. Knapp, G. L., and J. P. Glasby, Urban Hydrology; A Select^
Bibliography with Abstracts, U.S. Geological Survey, Water
Resources Investigation 3-72, 211 pp., GPO, Washington, D. C., -1973.
134. The Hydrologic Engineering Center, "Proceedings of a Seminar on
Urban Hydrology," 1-3 September 1970, Corps of Engineers, U.S.Army,
Davis, California.
135. McPherson, Murray B., and William J. Schneider, "Problems in
Modeling Urban Watersheds," Paper No. H-46,presented at the American
Geophysical Union Spring Meeting, April 16-20, 1973, as part of
"A Symposium on Models for Urban Hydrology", (in press, Water
Resources Research).
136. Flood Control Coordinating Committee, "The Chicago Underflow Plan
Development of a Flood and Pollution Control Plan for the Chicagoland
Area," State of Illinois, County of Cook MSDGC, City of Chicago,
29 pp., 1972.
76
-------
137. Department of Public Works, "San Francisco Master Plan for Waste Water
Management, Supplement I," City and County of San Francisco,
May 15, 1973.
138. McPherson, M. B., "Innovation: A Case Study," ASCE Urban Water Resources
Research Program Technical Memoradnum No. 21, ASCE, New York, N.Y.,
February, 1974. (Available from NTIS as PB 232 166).
139. Lanyon, R, F., and 0. P. Jackson, Metropolitan Sanitary District of
Greater Chicago, "A Streamflow Model for Metropolitan Planning and
Design," ASCE Urban Water Resources Research Program Technical
Memorandum No. 20, ASCE, New York, N.Y., January, 1974. (Available
from NTIS as PB 232 181).
140. Stall, John B., and Michael L. Terstriep, Storm Sewer Design - An
Evaluation of the RRL Method, Environmental Protection Technology
Series, EPA-R2-72-068, GPO, Washington, D. C., October, 1972.
HI. Laursen, E. M., W. J. Bauer, B. R. Hanke, M. M. Hufschmidt, L. D.
James, J. W. Knapp, and P. H. McGauhey, "Findings of the Damage and
Storage Task Committee," Appendix C in Urban Water Resources Research,
ASCE, New York, N.Y., September, 1968. (NTIS Id. No. PB 184 318).
142. Rice, Leonard, "Reduction of Urban Peak Flows by Ponding," Journal
Irrigation and Drainage Division, ASCE Proc., Vol. 97, No. IR3,
Paper 8351, September, 1971.
143. Jones, D. Earl, Jr., "Where is Urban Hydrology Practice Today?,"
Journal Hydraulics Division, ASCE Proc., Vol. 97, No. HY2, Paper 7917,
February, 1971.
144. Hathaway, G. A., "Military Airfields: Design of Drainage Facilities,"
Transactions, ASCE, Vol. 110, 1945.
145. Forest, E., and H. G. Aronson, "Highway Ramp Areas Become Flood Control
Reservoirs," Civil Engineering, Vol. 29, No. 2, February, 1959.
146. Anonymous, "Storm Water Control for a Shopping Center," Civil Engineering.
Vol. 30, No. 2, February, 1960.
147. Daily, Eugene J., "Storm Water Detention in Urban Areas," Public Works.
Vol. 92, No. 1, January, 1961.
148. Council on Environmental Quality, Environmental Quality. Second Annual
Report, GPO, Washington, D.C., August, 1971, p. 145.
77
-------
149. Field, Richard, and Edmund J. Struzeski, Jr., "Management and Control
of Combined Sewer Overflows," J.WPCF, Vol. 44, No. 7, July, 1972.
150. Office of Research and Monitoring, Projects of the Municipal Technology
Branch Through June 1972, U.S. Environmental Protection Agency,
Environmental Protection Technology Series, EPA-R2-72-080, GPO,
Washington, D. C., September, 1972.
151. Ure, James E., "Reducing Pollution from Combined Sewer Overflows,"
APWA Reporter, Vol. 39, No. 5, May, 1972.
152. Anderson, J. J., "teal-Time Computer Control of Urban Runoff,"
Journal Hydraulics Division, ASCE Proc., Vol. 96, No. HY1, Paper 7028,
January, 1970. Discussions, Vol. 96, No. HY9, Paper 7497, September, 1970.
153. Callery, R. L., "Dispatching System for Control of Combined Sewer Losses,"
Metropolitan Sewer Board, St. Paul, Minn., Report for EPA, No. 11020
FAQ, March, 1971
154. Anonymous, "Relieving Combined Sewer Pollution," Water and Sewage
Works, V01. 118, No. 10, October, 1971.
155. Tucker, L. S., "Control of Combined Sewer Overflows in Minneapolis-
St. Paul," Technical Report No. 3, Metropolitan Water Intelligence
System Project, Colorado State University, Ft. Collins, October, 1971.
,«-r Anderson, J. J., R. L. Gallery and D. J. Anderson, "How to Control
1 DO.
Sewer Outflows," Water and Wastes Engineering. Vol. 9, No. 9,
September, 1972.
157. Brown, John W., and Darrel Suhre, "Sewer Monitoring and Remote Control-
Detroit," a paper presented at the ASCE Annual and Environmental
Meeting, October 13-17, 1969, Chicago, ASCE Meeting Preprint 1035,
New York, N.Y.
158. Detroit Metropolitan Water Services, "Detroit Sewer Monitoring and
Remote Control," in Combined Sewer Overflow Abatement Technology.
FWQA, U.S. Dept. of the Interior, Water Pollution Control Research
Series 11024, GPO, Washington, D. C., Section 6, pp. 219-290, June, 1970.
159, Remus, Gerald, "Storm-Water Retention Can Work ... And Prevent the
Heavily Polluted 'First Flush1 from Overflowing to Damage the
Receiving River," American City, Vol. 85, No. 10, October, 1970.
160. Anonymous, "Regional Sewer System is No Dream in Detroit," Engineering
News-Record, Vol. 185, No. 24, December, 1970.
78
-------
161. Gibbs, Charles V., and Stuart M. Alexander, "CATAD System Controls for
Regulation of Combined Sewage Flows," Hater and Wastes Engineering,
Vol. 6, No. 8, August, 1969.
162. Municipality of Metropolitan Seattle, Maximizing Storage in Combined
Sewer Systems, U. S. Environmental Protection Agency, Water Pollution
Control Research Series, No. 11022 ELK, GPO, Washington, D. C.,
December, 1971
163. Gibbs, C. V., C. J. Henry and F. J. Kersnar, "How Seattle Beat Pollution,"
Water and Wastes Engineering, Vol. 9, No. 2, February, 1972.
164. Gibbs, C. V., S. M. Alexander and C. P. Leiser, "System for Regulation
of Combined Sewage Flows," Journal Sanitary Engineering Division. ASCE
Proc., Vol. 98, No. SA6, Paper 9402, December, 1972.
165. Pew, K. A., R. L. Callery, A. Brandstetter and J. J. Anderson, "The
Design and Operation of a Real-Time Data Acquisition System and
Combined Sewer Controls in the City of Cleveland, Ohio," a paper
presented at the WPCF Annual Conference, October 10, 1972, Atlanta,
Georgia.
166. Pikarsky, Milton, and Clint J. Keifer, "Underflow Sewers for Chicago,"
Civil Engineering, ASCE, Vol. 37, No. 5, May, 1967.
167. Anonymous, "Deep Tunnel Storage May Solve City Storm Water Problem,"
Environmental Science and Technology. Vol. 3, No. 3, March, 1969.
168. City of Chicago, et al, "Underflow Plan for Pollution and Flood Control
in the Chicago Metropolitan Area," in Combined Sewer Overflow Abatement
Technology, June, 1970, op. cit., Section 5, pp. 139-218.
169. Pikarsky, Milton, "Sixty Years of Rock Tunnelling in Chicago," Journal
Construction Division, ASCE Proc., Vol. 97, No. C02, Paper 8504,
November, 1971.
170. University of Wisconsin, Deep Tunnels in Hard Rock, Proceedings of An
Engineering Institute, Milwaukee, Wisconsin, November 9-10, 1970, Water
Pollution Control Research Series, U.S. Environmental Protection Agency,
GPO, Washington, D.C.
171. Parthum, Charles A., "Building for the Future - The Boston Deep Tunnel
Plan," J.WPCF, Vol. 42, No. 4, April, 1970.
79
-------
172. Corey, John B. W., "Feasibility of Utility Tunnels," J.AWWA, Vol. 64,
No. 4, April, 1972.
173. American Public Works Association, Feasilibity of Utility Tunnels in
Urban Areas, Research Foundation Report APWA-SR-39, Chicago, February,
1971.
174. McPherson, M. B., "Utility Tunnel Concepts for Sewers," pp. 54-58 in
Proceedings: Conference on Engineering Utility Tunnels in Urban Areas.
American Public Works Association, Special Report No. 41, Chicago,
February, 1972.
175. Sorensen, Kenneth E., "Fourth Dimension for Urban Development,"
Journal Urban Planning and Development Division, ASCE Proc., Vol. 97,
No. UP1, Paper 8064, April, 1971.
176. Angino, Ernest E., Larry M. Magnuson and Gary F. Stewart, "Effects of
Urbanization on Storm Water Runoff Quality: A Limited Experiment,
Naismith Ditch, Lawrence, Kansas, Water Resources Research. Vol. 8,
No. 1, February, 1972.
177. Task Group, "Design and Operation of Recharge Basins," J.AWWA. Vol. 55,
No. 6, June, 1963.
178. Knapp, George L., ed., Artificial Recharge of Groundwater: A
Bibliography, Water Resources Scientific Information Center, Office of
Water Resources Research U.S. Dept. of the Interior, WRSIC 73-202,
GPO, Washington, D. C., February, 1973.
179. Ackermann, William C., "Research Problems in Hydrology and Engineering,"
in Water Research^ edited by A. V. Kneese and S. C. Smith, The Johns
Hopkins Press, Baltimore.Maryland, 1966, p. 499.
180. National Research Council, Waste Management and Control. National
Academy of Sciences, Publication 1400, GPO, Washington, D.C., 1966,
p. 170.
181• ASCE. Basic Information Needs in Urban Hydrology. April, 1969, op. cit.,
p. ix. (Ref. 1).
182. Dawdy, David R., Robert L. Smith, Norman H. Crawford, Peter S. Eagleson
and Warren Viessman, Jr., "Considerations for Modeling Urban Rainfall-
Runoff-Quality Processes," Appendix A in Urban Water Resources Research.
ASCE, September, 1968, op. cit.
80
-------
183. Thomasell, Albert,-Jr., "Considerations for Characterizing the Time
and Space Distributions of Metropolitan Storm Rainfall," Aooendix R
in Urban Waster Resources Research, ASCE, September, 1968, op. cit.
184. Tucker, L. S., "Raingage Networks in the Largest Cities," ASCE Urban
Water Resources Research Program, Technical Memorandum No. 9, ASCE,
New York, N.Y., March 17, 1969. (NTIS Id. No. PB 184 704).
185. Tucker, L. S., "Non-Metropolitan Dense Rainagage Networks," ASCE Urban
Water Resources Research Program, Technical Memorandum No. 11, ASCE,
New York, N.Y., January, 1970. (NTIS Id. No. PB 191 709).
186. Schneider, William J., "The U.S. Geological Survey Urban Water Program,"
in Effects of Watershed Changes on Stream Flow, 1969, op. cit.
187. Water Resources Division, "Projects Related to WRD Urban Water Program,
FY72," U. S. Geological Survey, Washington, D. C., October, 1972.
188. Office of Water Resources Research, A National Urban Water Resources
Research Program. U.S. Dept. of the Interior, GPO, Washington, D.C.,
1971.
189. Tucker, L. S., "Availability of Rainfall-Runoff Data for Sewered Drain-
age Catchments," ASCE Urban Water Resources Research Program, Technical
Memorandum No. 8, ASCE, New York, N.Y., March 3, 1969. (NTIS Id. No.
PB 184 703).
190. Tucker, L. S., "Availability of Rainfall-Runoff Data For Partly
Sewered Urban Drainage Catchments," ASCE Urban Water Resources Research
Program, Technical Memorandum No. 13, ASCE, New York, N.Y., March, 1970.
(NTIS Id. No. PB 191 755).
191. Lager, John A., "A Simulation Technique for Assessing Storm and
Combined Sewer Systems," in Combined Sewer Overflow Seminar Papers,
FWPCA, U.S. Dept. of the Interior, Water Pollution Control Research
Series, DAST 37, 11020-03/70, GPO, Washington, D.C., November, 1969,
p. 151.
192. Soil Conservation Service, Section 4, "Hydrology," in SCS National
Engineering Handbook, January, 1971, op. cit.. Chapter 16. (Ref. 19).
193. Rantz, S. E., "Suggested Criteria for Hydrologic Design of Storm-
Drainage Facilities in the San Francisco Bay Region, California,"
U.S. Geological Survey, Open File Report, Menlo Park, California,
November 24, 1971.
81
-------
194. The Hydrologic Engineering Center, "Proceedings of a Seminar on Urba
Hydrology, 1-3 September, 1970, Corps of Engineers, U.S. Army, Davis
California.
195. Rao, A. R., J. W. Delleur and B. S. P. Sarma, "Conceptual Hydrologic
Models for Urbanizing Basins," Journal Hydraulics Division, ASCE
Proc., Vol. 98, No. HY7, Paper 9024, July, 1972.
196. Harris, Garth S., "Real Time Routing of Flood Hydrographs in Storm
Sewers," Journal Hydraulics Division. ASCE Proc., Vol. 96, No. HY6,
Paper 7327, June, 1970.
197. Pinkayan, Subin, "Routing Storm Water Through a Drainage System,"
Journal Hydraulics Division, ASCE Proc., Vol. 98, No. HY1, Paper
8642, January, 1972.
198. McPherson, M. B., "Better Design of Storm Water Drainage Systems,"
APWA Reporter, Vol. 41, No. 2, May, 1974.
82
-------
QUANTITY ASPECTS OF URBAN STORMWATER RUNOFF
By
William H. Espey *
David E. Winslow
1.0 RAINFALL/RUNOFF
In order to properly design urban drainage facilities,
consideration must be given to the required design storm rainfall,
the rainfall-runoff relationships as affected by the surface
and geometric characteristics of the watershed. In general,
the design of an urban storm drainage system should be guided
by the following six criteria**:
1) The system must adequately dispose of all sur-
face runoff, resulting from the selected design
storm, without causing serious damage to physi-
cal facilities or serious interruption of nor-
mal traffic.
2) Runoff resulting from storms exceeding the
design storm must be disposed of with the
least amount of damage to physical facilities
and interruption of normal traffic.
3) The storm drainage system must have a maximum
reliability of operation.
4) The construction costs of the system must be
reasonable with relationship to the importance
of the facilities it protects.
5) The storm drainage system must require minimum
maintenance.
6) The storm drainage system must be adaptable to
future expansion with minimum additional cost.
President, Manager, respectively, Espey, Huston & Associates, Inc., Austin, TX.
**Master Drainage Manual - City of Austin, 1974, by URS/Porrest
& Cotton, Inc., and Espey, Huston & Associates, Inc.
83
-------
1.1 The Concept of Rainfall Excess
The two parts of the runoff process have been called
the land phase and the channel phase. This section discusses
the land phase problem or methods of determining rainfall excess
from observed basin rainfall given information about basin
characteristics and previous rainfalls. There are many problems
involved in successfully analyzing each part of the runoff
process; however, the land phase problem is possibly the more
difficult process to analyze.
1.1.1 Relation Between Precipitation and Runoff - According
to Chow (1964), there are two major groups of interrelated
factors which influence the runoff process:
a) Climatic factors
Precipitation: Form, type, intensity,
duration, time distribution, areal distribution
frequency of occurrence, direction of storm *
movement, antecedent precipitation, and soil
moisture.
Interception: Vegetation species, composition,
age, and density of stands, season of year.
Evaporation: Temperature, wind, atmospheric
pressure, nature and shape of evaporative
surface.
Transpiration: Temperature, solar radiation
wind, humidity, soil moisture, kinds of '
vegetation.
b) Physiographic factors
Basin characteristics: Size, shape, slope
orientation, elevation and stream density, land
84
-------
use and cover, surface infiltration condition,
soil type, geological and topographical
conditions.
Channel characteristics: Size and shape of
channel cross-section, slope, roughness, and
length.
1.2 Review of Methods for Determining Rainfall Excess
This section reviews some of the recent technology which
is available for modeling land phase problems for the purpose of
rainfall excess prediction. In particular, infiltration models,
API-type models, and functional rainfall models will be discussed.
1.2.1 Infiltration Models - The principle of the infiltration
approach is straightforward, i.e. runoff occurs when the precipi-
tation intensity exceeds the rate of infiltration. A simple,
empirical method based on the infiltration concept described by
Linsley et. al. (1949) is the so-called $-index method. This
method, which ignores time and space variation of infiltration
phenomena, is a simple expression for the infiltration capacity
of the basin as shown in Fig. 1. If a time-intensity graph of
rainfall is constructed, the $-index is the average rainfall in-
tensity above which the volume of rainfall equals the volume of
observed runoff. Thus $ is the constant rate at which water must
be extracted from the rainfall to yield the observed runoff. In
some applications the $-index is related to an antecedent preci-
pitation parameter or to season of the year in order to improve
results.
Obviously, the $-index method is a simplified
model for rainfall excess determination since it ignores
temporal and spatial variation of infiltration rates. However
the method may yield acceptable results in some cases,
particularly where homogeneous conditions exist on the watershed.
85
-------
For example, Sauer (1970) after examining other methods (to be
discussed below) used the 5-index method and obtained useable
results. Fig. 2 from Sauer shows the median 1-index, and
standard error bands for each month of the year based on about 500
storms from 21 basins in Northern Louisiana. The use of such a
relation is justified in this case since mean basin rainfall
determination is poor due to inadequate rain-gage coverage.
ui
.—J Runoff
j pf Loss
Losses
Rainfall Hyetograph
TIME
Fig. 1 Schematic of $-Index
(after Linsley ejt. al. , 1949)
1.4
1.2
*
mf
.6
.4
.2
0
Dashed lines include 2/3 of values
-\
JFMAMJ JASOND
MONTH
Fig. 2 Seasonal Variation of Infiltration Index, $
(after Sauer, 1970)
86
-------
Grayman and Eagleson (1969) developed a model of catch-
ment dynamics using a variation of the 3E-index method to determine
rainfall excess. These authors reason that only a fraction,C ,of
the basin rainfall will result in runoff. The parameter ,C,
represents the net effect of the interaction of the complicated
factors described above in Section 1.1, that is ie = Ci where ie
is total rainfall excess and i is total basin rainfall. The
percentage of rainfall that appears as runoff will generally
increase during wet periods due to saturation of the soil and
other factors; therefore, C is modeled as an exponential decay
function,
C = 1 - exp (B4AP30)
(1)
where B^ is a curve-fitting parameter which reflects the influence
of the soil-vegetative cover and AP30 is a 30-day total antecedent
precipitation.
Horton (1939) found from runoff plot experiments that
the soil infiltration capacity rate as a function of time decreases
during the infiltration process by the relation,
if • -k(f-fc) (2)
where f is the actual infiltration capacity, fc the minimum infil-
tration capacity which is reached if the infiltration process
continues indefinitely, and k is a constant. A solution to this
differential equation yields the well known Horton formula,
f - tc + (fo-Ve-kt (3)
where f is the initial infiltration capacity at the beginning of
o
the storm. A representation of equation (3) is shown in Fig. 3.
87
-------
z
o
tc.
\-
TIME
Fig. 3 Horton Infiltration Formula
Three drawbacks of the Horton formula have been identified:
(1) The formula assumes an adequate and continuous supply of water
available for infiltration. Thus, a model based on the Horton
formula is only approximately true for storms where the precipi-
tation rate drops below the infiltration capacity rate. (2) The
method requires an estimate of f0 at the beginning of the storm.
This requirement is the main difficulty of the method although
estimates of fo have been made using correlation techniques.
(3) The rainfall excess value determined using equation (3) refers
only to overland flow.
In spite of the above drawbacks, the Horton formula
allows a reasonable representation of the distribution of rainfall
excess during a storm period and there are numerous models based
on the Horton formula. Betson (1964) used the formula to develop
a model for a watershed infiltration capacity function. The EPA
Storm Water Management Model (Metcalf and Eddy, Inc.,et al (1971))
uses this formula to determine soil infiltration losses.
If the area under the Horton infiltration capacity curve
for time t is considered as a volumetric loss equal to /D (c+be~nt)d
where D is storm duration, c, b, and n are coefficients which
correspond to fc, (fQ-fc) and k respectively; then, this volumetric
loss, L should equal rainfall, R, minus runoff, RO, or,
L = R-RO=/D (c+be"nt)dt
88
-------
Because for a given storm, the Initial infiltration capacity may
not be the maximum value, the integration of equation (4) should
be over an interval t(sm)-(D+t(sm)) as shown in Fig, 4 where t(sm)
is the time which corresponds to an initial capacity rate, f(sm).
0 t(sm) Time in hours D+t(sm)
Fig. 4 Graphical Representation Of Infiltration Capacity Function
This initial rate may then be related to a soil moisture index, sm,
After additional modification of the basic Horton formula, Betson
(1964) integrates the formula to obtain,
L = ct-(b/n)e~nt |D+t(sm)
|t(sm) (5)
When equation (4) was fitted to watershed data using regression
methods, a useable expression for determining rainfall excess re-
sulted. Shown in Fig. 5 are infiltration curves developed by
Horner and Jens (1942) based on studies in St. Louis. The upper
curve was assumed for residential areas with soils having high
infiltration capacities.
In a recent application (EPA/Rice University Grant No.
802433) of the SWMM to a small urban watershed In Houston, Texas and
to The Woodlands New Town development near Houston, Texas the Horton
infiltration coefficients were adjusted based on U. S. Geological
Survey (USGS) rainfall-runoff data in the Houston area. This was
89
-------
o:
x
UJ
<
cc
g
h-
<
cc
RESIDENTIAL AREAS
SANDY SOIL AREAS
NDUSTRIAL AND COMMERCIAL AREAS
*_i - .1 . • . A »...._... ^. T. t'..'.. . . .1 . t ... v_i_j_-i'_i_n^
1
i
0
20
40
60 80 100
TIME, MINUTES
120
RGURE 5
PERVIOUS SURFACE INFILTRATION-CAPACITY CURVES.
(After Homer and Jens, 1942)
140
(3)
-------
accomplished by calculating "effective infiltration rates" for a
number of observed storm events and plotting these values versus
storm duration in order to determine graphically the coefficients
for Horton's equation. It was found that the initial infiltration
rate, fo, varied from 0.75 to 3.0 inches per hour depending on
antecedent soil moisture conditions. Further, 0.12 inches per
hour was found to be an acceptable final infiltration rate, fc,
while the decay rate, k was found to be .00005 sec
1.2.2 API-Type Models - Several types of models can be
generally classed as API-type models. In this section, some of
these models are considered including graphical coaxial models
and regression models.
1.2.2.1 Graphical and Numerical Coaxial Models - Kohler and
Linsley (1951) describe a graphical coaxial model for determining
direct runoff volume; such a relation is shown schematically in
Fig. 6. The first requirement for use of such a model is the
UASON
OUAD4ANT
PtlOPITATION
OU ADI ANT
Fig. 6 Standard API Type Rainfall-Runoff Relation
91
-------
determination of an index of the moisture condition of the drainage
basin at the moment precipitation starts. The most common index
with regard to this initial moisture condition is the antecedent
precipitation index API.
The API may be defined in many different ways. For example,
Grayman and Eagleson (1969) defined it as the total precipitation
over a preceding N-day period. Another API model has the formula
API = P
if
where b is a constant less than unity which is a function of time.
k
If a day by day value of the index is required, b may be assumed
K t
to decrease exponentially with time as b = K , where K is a
recession constant. Thus
API.. = API K*1 (7)
L O
where APIQ and API are initial and t-day values of the index
respectively. The value of the index after any day is related to
the index the day before since API2 = K^API^) where t *= 1. Accord-
ing to Chow (1964), the value of K normally ranges between 0.85
and 0.98 and depends on potential evapotranspiration of the region
as well as time of year.
In application, the API-type model shown in Fig. 6 is
entered with an API value for the given season of the year; then
(as indicated by the arrows in the figure) is adjusted with a storm
duration value and finally is adjusted using a value of storm
precipitation to obtain direct runoff. The model as given determines
the total direct runoff for an event of any duration in terms of
total precipitation. In practice, such a graphical model is devel-
oped by trial and error using observed data until good agreement
between observed and predicted runoff occurs.
The relationships illustrated in Fig. 6 are strictly
applicable to the storm event only; if continuous prediction of
92
-------
runoff for say 5-hour periods Is required, some modification of
the graphical API-type model must be made. Sittner et. al. (1969)
made such a modification in order to construct a model useful for
continuous flow forecasting.
1.2.2.2 Regression Models - Moore (1968) developed an API-type
model using graphical linear regression methods, which is similar
to a coaxial model. In the method, rainfall excess is related to
mean basin rainfall, rainfall duration, API, and season of the year
as in the U. S. Weather Bureau coaxial model. The API used by
Moore is a simplified 6-day index whereby API = 6/6 Pj+5/6P2+
+1/6P,; other data for the model is obtained from observed rainfall
and runoff records. In Moore's method, storm runoff is graphically
related to the above parameters by the method of deviations of
graphical multiple regression rather than by the coaxial method.
Storm runoff is plotted against precipitation and deviations are
then plotted against storm durations and the second deviations are
plotted against API. The last plot has season of year as a para-
meter and the trial curve relating storm rainfall and runoff. A
second trial curve of storm rainfall and runoff is then used and
the process is repeated. After a few iterations of the process,
acceptable results can be obtained. Moore (1968) used the method
to obtain synthetic daily direct runoff estimates and obtained good
results.
In order to evaluate the effects of increasing impervious
cover on the runoff from a small urban watershed (Waller Creek—4.13
square miles), Espey et. al. (1965) developed a rainfall-runoff
regression relationship which relates total runoff to impervious
cover, antecedent-precipitation index, amount of rainfall, and
Duration of rainfall. Winslow and Espey (1972-73) developed a rain-
fall runoff regression model for The Woodlands a HUD New Town
located near Houston, Texas. A relationship was obtained by
93
-------
multiple regression analysis of 142 storms for 21 Houston area
watersheds. The following equation was obtained to predict the
runoff for a given rainfall and watershed condition:
RUN = 0.325 R1"23 M0'" I0'067 Si'0'12 (8)
where RUN is the total runoff in inches,
R is the total rainfall in inches,
M is the soil moisture index,
I is the percent impervious cover, and
SI is the soil index.
The data used to develop this relationship, except for the total
runoff, were obtained from the USGS in Houston. The total runoff
for each storm was obtained from the annual compilations of Houston
urban watershed data from 1965-70, and from unpublished records
of the Houston office of the USGS. The soil moisture index, M,
is computed by essentially the same procedure as the antecedent
precipitation index described by Linsley, et. al. (1958). The soil
index, SI, is the maximum permeability of the soil in inches per
hour.
1.2.3 Functional Models - A functional model proposed by the
Soil Conservation Service and described by Kent (1968) is widely
used. The relation between direct runoff and storm rainfall is
determined by assuming an initial abstraction, Ia which includes
surface storage, interception, and infiltration prior to the runoff
event and a potential maximum retention, S. It is assumed that the
ratio of actual retention (P-Ia-Q) and potential maximum retention
S equals the ratio of actual runoff, Q and potential maximum runoff
(P-I ) as
el
p-yq __2_
94
-------
where P and Q are accumulated rainfall (potential maximum runoff
if I. = 0) and accumulated direct runoff respectively. A rear-
a
rangement of equation (9) gives
(p-ia;+s (10)
which is the basic SCS rainfall-runoff equation. On the basis of
experience, I = 0.20S and therefore,
a
(P-0.2S)
P - U.8S
2
(11)
Equation (11) was developed from experimental plots for numerous
soils and vegetative cover conditions and is mainly applicable to
small watersheds and for direct runoff from 24-hour storm rainfalls.
The value of S is related to the so-called runoff curve
number according to the equation,
rw 1000
CN " TUT? (12)
Research data obtained by the Soil Conservation Service has
provided a relation between CN's and various hydrologic soil-cover
complexes for more than 3,000 soils. In addition, soils are
divided into four hydrologic groups, depending on infiltration rate,
and antecedent moisture is accounted for by division into three
classes according to runoff potential. In practice, a composite CN
for the watershed is used to enter Fig. 7, the rainfall-runoff
chart which represents a graphical solution to equation 10.
Mills and Viesman (1972) developed empirical relation-
ships based in part upon the Soil Conservation Service soil cover
complex method to estimate the runoff from small watersheds. For
rainfalls less than 1.5 inches, the runoff was predicted by a
relationship between the percent impervious area on the watershed
and the percent excess rainfall. An adjustment was made for initial
abstraction combined with initial surface depression storage.
95
-------
10
Rainfall (P)
fwith Pi I0; SJtI0+F;:
RUNOFF (0)
4- P-I0+S
and F« P-T0-0
Curves on this sheet are for the
t-l-H-4-l
•case I0.0.2S, so
h
(P-0.2S)Z-
p + o.es
MnflHratlon
curve 1
abstraction T I
4 5 C 7
RAINFALL (P) IN INCHES
Fig. 7
(P - 0 23)*
—Solution of the runoff equation, Q • p . 'Ra—
after Kent (1968J • + °-Qs
-------
For rainfall amounts greater than 1.5 inches, an additional incre-
ment of runoff was added due to the pervious areas by using hydro-
logic soil classes, vegetation cover, and other controlling .
parameters.
1.3 Overland Flow Analysis
One of the methods commonly used in estimating rates of
runoff from small drainage areas such as air fields and urban
impervious areas of uniform slope and surface characteristics is
the hydraulics of overland flow. The following section will
present development of the basic overland flow equations based on
theoretical considerations and experimental data.
Two of the more important and classical studies of
overland flow characteristics were by Horton (1935) and by Izzard
(1946). Horton presented equations for overland flow based on
turbulent flow considerations. Izzard's overland flow equations
considered laminar flow conditions and were presented in the form
of dimensionless hydrographs (Figures). Izzard's dimensionless
hydrographs were verified in various laboratory tests and resulted
in acceptable agreement between experimental and laboratory results.
Horton's overland flow equations apply to a uniform rainfall supply
°f unlimited duration. In addition, the Horton equation for the
Recession portion of the overland flow hydrograph has no experimental
Verification. Izzard's dimensionless overland flow hydrograph is
Relatively simple and reflects non-uniform rainfall supply.
Equations of overland flow can be developed based on
the assumption of laminar flow. If the depth of flow in the
°verland sheet is quite small, the quantity of water temporarily
stored in this sheet surface attention is relatively large. For
conditions of laminar flow the following equation can be derived
(12a)
97
-------
oo
-------
.where f> is density, g is gravitational acceleration, and // is
absolute viscosity. 'The assumption is made that the slope, s,
is so small that the sine and tangent are equal. Since///? is
equal to kinematic viscosity /",
dv = fis (D-y)dy (13)
Integrating the above equation with the boundary conditions v «
0, y - 0
Integrating from y = 0 to y = D and dividing by D, the mean
velocity is
gsD2
vm = TP'
and the discharge per unit width, q, equals vD or
q = bD3 (16)
where b is a coefficient involving slope and viscosity.
Based on Izzard's experiments the time to equilibrium
can be expressed as a dimensionless overland flow hydrograph.
t = 2Ve
50qe (17)
where t is defined as the time in minutes when flow is 97 percent
e
of the supply rate and V is the volume of water in surface deten-
tion at equilibrium. From a strip of unit width the equilibrium
flow q can be expressed as follows:
q - 1L
e 43,200 (18)
where i is the rainfall rate (or the rate of rainfall excess if
the surface is pervious) and L is the distance of overland flow.
99
-------
,The constant 43,200 gives q in cubic feet per second when i is
in inches per hour. Substituting average depth (Dg) on the strip,
where D = V /L, for outflow depth, Eq. 16 becomes
Dfi . Ve -
Substituting Eq. (18) for qg gives
V = KL 4/V/3
e J5TT (20)
where V is the volume of detention (cu ft) on the strip at
equilibrium. Izzard found experimentally that b could be expressed
as follows:
b = 0.0007i+c
(21)
o
where s is the surface slope and the retardance coefficient c is
as given in Table I.
TABLE I
RETARDANCE COEFFICIENT c IN EQ. 21
Very smooth asphalt pavement 0.007
Tar and sand pavement. 0.0075
Crushed-slate roofing paper 0.0082
Concrete 0.012
Tar and gravel pavement 0.017
Closely clipped sod 0.046
Dense bluegrass turf 0.060
Izzard's overland-flow hydrograph in a dimensionless form is
shown in Figure 8. A dimensionless recession curve defines the
shape of the receding limb. At any time t after the end of rain,
cl
the factor 6 is
VG (22)
where Vg is the detention given by Eq. 20.
100
-------
1.4 Runoff Hydrograph
Water which constitutes the surface stream flow reaches
the urban drainage system by various paths from the original point
at which the rainfall first fell on the watershed. One of the
principle paths of water entering the drainage system is overland
flow. Overland flow is the flow across the surface of the water-
shed in which its flow properties are not dictated by the flow
characteristics of the surface drainage system. In addition, other
waters infiltrate through the soil and flow beneath the surface to
the stream channels. The actual paths of movement of water to a
surface drainage canal can be summarized as follows:
(1) Overland flow
(2) Inner flow
(3) Ground water flow
In urban watersheds the distance for overland flow is
relatively short. The component of overland flow is an important
element in the formation of the total runoff hydrograph. Inner
flow or sub-surface flow represents the water which infiltrates
the soil surface and moves laterally through upper soil strata
and eventually enters the stream channel. The third element of
runoff originates from precipitation which percolates into the
water table. If the ground water table is penetrated by the
stream channel, ground water flow may enter the stream
channel. This flow has been characterized as base flow or dry
weather flow.
Shown in Fig.9 is a typical runoff hydrograph for an
urban watershed. The hydrograph consists of a rising limb, crest
stage, and falling limb or recession. On urban watersheds the
rising limb is a function mainly of the characteristics of the
storm, and the hydraulic conveyance characteristics of the improved
101
-------
o
ro
UJ
o
ct
o
CO
o
RECESSION
INFLECTION POINT
TIME
FIGURE 9. RUNOFF HYDROGRAPH
-------
drainage system. The point of inflection on the falling side of
the runoff hydrograph is commonly assumed to mark the time at
which surface inflow into the main channel system has ceased.
1.5 Urban Drainage Systems
The effects of urbanization on the hydrologic response
of a previously undeveloped watershed are becoming more widely
known as continuing data collection programs yield more informa-
tion concerning runoff quantity and quality. The effects of urban-
ization on peak runoff rates have been evaluated in many studies
over the past several years. Carter (1961), in relating peak flow
to degree of imperviousness showed that 50 to 100 percent increases
may be expected. Studies by Espey e£ al. (1965) and Espey and Winslow
(1968; 1969) have shown that urbanization may increase flow rates
as much as four times those found for undeveloped areas. Numerous
other studies (Wiitala, 1961; Van Sickle, 1962 (see Fig. 10);
Crawford and Linsley, 1962; and Sawyer, 1961) have shown similarly
large increases in peak flow rates. The changes in flow rates
mentioned above are caused by several factors. Impervious areas
such as parking lots, streets, and roofs, result in increased
total runoff volumes and peak flow rates. Storm drainage and
collection systems decrease the time of concentration and increase
peak flow rates. The lining and rectifying of drainage ditches
also may generate increased runoff rates. Such increased runoff
rates due to development often result in substantial downstream
flooding and property damage.
The Task Force on Effect of Urban Development on Flood
Discharges, American Society of Civil Engineers has summarized
these effects as follows:
A. Where the urban development encompasses the major
part of a drainage area, the following effects may
be felt:
103
-------
10-12 JULY.I939
22-25 NOV.1940
— 23-25 SEPT.I94I
12-19 MAY,I953
6-11 APRILJ959
23-2? JUNE.I960
10
15 20 25 30 35 40 45 50 55 60 65 70 75 80 85 90 95
TIME AFTER START OF RAINFALL EXCESS,IN HOURS
FIGURE 10. BRAYS BAYOU UNIT HYDROGRAPHS (AFTER VAN SICKLE, 1962).
-------
(a) Change in total volume runoff
(1) Annual
(2) Monthly
(3) Short-term (daily or storm)
(b) Change in distribution of total runoff (con-
centration in storms; less base flow)
(c) Change in peak rates of
runoff
(d) Change in frequency of peak rates
(e) Change in shape of short-term hydrograph
(f) Change in sediment content of stream
(g) Throttling action by enclosed drainage
system
B. Where urban drainage area is tributary to a river
with a larger undisturbed drainage area:
(a) Change in total runoff will depend in part
on ratio of urban drainage to drainage area
of river.
(b) Change in flood peaks (magnitude and timing)
will depend on ratio of areas and differences
in character of areas.
(1) Floods in river may have two or more
peaks.
(2) If urban area dominates, peak rates and
frequencies may be changed.
(c) Quick runoff from urban area may cause "back-
water" effects and even reverse flow on main
river.
(d) Rapid rises and falls may have destructive
effects on river banks.
An urban drainage system can be characterized by a
»
high percentage of impervious cover. The drainage system will
normally include man-made impervious pathways for collecting or
guiding the flow of water over the surface (curbs, gutters, lined
channels, paved parking areas, streets, etc.) and underground
(storm, wastewater and combined sewers). The system includes all
105
-------
appurtenances such as catch basins, storage basins, inlets, manholes,
sediment traps, ditches, wiers and outfall structures that guide,
control, or otherwise modify either the quantity, rate of flow or
quality of runoff from urban drainage areas. Shown in Fig. 11
is a simplified urban drainage system. The system consists of sub-
systems dealing with
a. surface runoff,
b. transport of flow, and
c. the receiving water.
106
-------
SURFACE RUNOFF SUBSYSTEM
TRANSPORT
SUBSYSTEM
RECEIVING WATER
SUBSYSTEM
\
FIGURE I I . THE URBAN DRAINAGE SUBSYSTEM (After
Orlob,l974)
107
-------
1.6
Open Channel Flow
Open channel flow consists of flow in natural streams,
rivers, canals, sewers, tunnels, and pipes partially full. Prob-
lems are more difficult to solve as shape can be practically
anything and roughness is often difficult to determine.
Uniform flow is established in open channels which
are sufficiently long, have constant slope and cross-section,
and are unaffected by backwater conditions. For any channel of
given roughness, cross section, and slope, there is only one depth
of flow YQ at which the flow is uniform for a given flow rate.
1.6.1 Slopes - Piezometric head = slope of water surface = S :
M^^MBOM^^M* »• y
Hydraulic Slope , ~ hf/L, where h^ is the energy loss for the
horizontal distance, L; Slope of the bed = S .
For uniform flow only
= s.
(23)
Uniform Flow - Chezy Equation
V
Summing forces along the direction of flow in the channel schematic
shown above
trQPwL = Wsin°<= ALsin* (24)
where TQ is the shear stress, PW the wetted perimeter and W is the
weight of the water. For small angles sincx = S
108
-------
ToPwL = YALSo <25>
<26>
where R is the hydraulic radius.
By analogy with pipe flow
-=0"! V /* (27)
where f is the friction factor.
Equating the two expressions for "tr
J2 =YRSn (28)
v = lay /RsT - /a* [RSO (29)
or C «
fT
The Chezy equation is
V ^ vj V -Ko _ /** /\ \
o (30)
1.6.2 Equations for Open Channel Flow - The most common
equation is the Manning Equation:
v = L*§i R2/3 S1/2 (3D
n
where n is a measure of the boundary roughness. From the Chezy
Equation
Cfis - IdM R2/3 sl/2 (32)
f o n
r - X-486 *2/3 SQ1/2 (33)
C n~~ rr/I TT7T- '
K. O
109
-------
C . Rl/6
Other formulas for open channel flow, evaluation of the Chezy
coefficient from Darcy Equation
(35)
From Manning Equation
From Hazin
C = Rl/6 (36)
r = 157. 6
c
where m is a measure of absolute roughness and
m * .11 for smooth concrete; m = 0.83 for rubble masonery;
m - 2.36 for earth channel.
From Kutter Equation
0.00281 1.811
r . 41.65 + g ........ + "" H — (38)
l + £(41.65 + U-°y*l)
VR s
C =
-------
1. 7 Methods of Routing
Many simplified methods of flood routing have been
developed through the years, some of which have general appli-
cation, while others are designed for a specific situation or
problem. Most of the simplified methods are storage-routing methods
and are based on a simplified form of the continuity equation for
unsteady flows
""5F + "-0 (40)
where V is the average velocity
A is the cross sectional area
B is the top width of the channel
h is the water surface elevation from a given datum
x is a distance in the direction of flow
t is time
q is lateral inflow in cfs Per unit length of channel
The well-known Muskingum method is a storage-routing
method and is discussed in detail in paragraph 1.7.2. The unit-
response method is also a storage-routing method which utilizes
the principle of superposition. Other simplified methods include
lag methods and mechanical and electronic machines designed for
specific situations. In addition the kinematic wave theory is
discussed with respect to the SWMM.
1.7.1 Basic Concepts - Change-in-storage is the basic under-
lying principle of most simplified methods of flow routing in
natural channels. Equation (40) is the partial differential
equation which contains the elements of storage routing and is
sometimes called the storage-continuity equation. This equation
is usually simplified to fit the type of routing to be accomplished.
For instance, in reservoir routing, the concept that outflow equals
inflow minus change in storage is used. Change-in-storage is
111
-------
usually related only to prism storage for reservoirs, whereas in
open-channel routing it is usually necessary to consider wedge
storage as well as prism storage.
Wedge storage and prism storage is illustrated in Fig.
12. It is assumed, of course, that a uniform, unbroken, water
surface exists. This is not the case in most natural river
channels but the computational methods tend to average out the
differences. In Fig. 12, it can be seen that wedge storage is
larger on a rising stage than on a falling stage because the water
surface slope is steeper on a rising stage. In this sense,
accounting for wedge storage is an indirect way of accounting for
changing water-surface slope.
Rising stage
Falling stage ....
. \ ^ Wedge storage
Steady . \ \- '
flow
V i
-------
occurs, storage decreases until outflow ceases. The other lim-
iting condition of open-channel routing is the idealized, uniform
channel as shown in Fig.13b In this case, a flood wave moves
downstream with no change in shape. Actual river channels are
usually somewhere between these limiting conditions. Pools and
riffles, and expansions and contractions, are in effect series of
partial reservoirs which have the same attenuating effects as
reservoir storage.
Inflow
Storage
Outflow
Reservoir storage Idealized channel.uniform throughout
(a) (b)
Actual channels are
somewhere in between
Fig. 13 Relation of Inflow, Outflow, and Storage
for Flow in Open-Channels
Muskingum Method - The Muskingum method is based on the
storage-continuity equation as in reservoir routing. The basic
difference between the two methods is in the relation of storage
to discharge. In reservoir routing, storage is usually considered
a unique function of outflow. In the Muskingum method, storage
is related to both inflow and outflow to account for wedge stor-
age that results from changing slopes of the water surface. The
relation for storage, S, is defined as follows:
G
S = K xl + (1-x) 0
(41)
113
-------
where K the slope of the discharge- storage relation, has the
dimension of time, and represents the time required for the center
of mass of the flood wave to traverse the reach, x is a dimension-
less constant which weights inflow and outflow, I is the inflow
rate, and 0 is the outflow rate.
Combining equation (41) with the storage-continuity
equation results in the following routing equation:
(42>
<43>
and
°2 =
fl —
Cl =
r -
2
Co h + Cl h •
K - Kx + 0.5i
Kx -1- 0.5 At
K - Kx + 0.5
K - Kx - 0.5
K. - Kx + 0.5
f C2°l
\t
At '
At
At *
Application of equation (42) to a routing problem is
a simple process because CQ, GI, and C2 are constants. In most
cases, K is considered constant for the full range of discharges
to be routed, however it can be varied if significant differences
are defined. The constant x is always used as a non- variable.
The range of x is 0 to 0.5, with x = 0 being reservoir conditions
and x = 0.5 being idealized uniform channel conditions resulting
in pure translation (no change in shape) . Substitution of x = 0
in equation (43), (44), and (45) will result in the identical
equation defined for reservoir routing. Most natural channels
have values of x in the range of 0.2 to 0.3. Techniques of
estimating K and x from actual records are given by Chow (1964).
A critical factor in the use of the Muskingum method
is the selection of At, the computation interval. This time
114
-------
interval must be greater than 2Kx to avoid negative values of C ,
but for some reaches this criteria will result in a At too long
to adequately define the hydrograph. In these cases, the reach
must be subdivided into shorter subreaches and a stepwise routing
performed. The outflow from the first subreach becomes the inflow
to the next subreach, and so on until the downstream site is
reached. The value of K for each subreach is made proportional
to the length of the subreach.
1*7,3 Kinematic Waves and Their Limited Application in
the SWMM - A kinematic wave is one in which Q, the flow rate, in
a given channel is a function of depth, y, alone as is assumed in
uniform flow computations.
Using the Ch6zy equation to elaborate:
Q = CA \J RSf Sf = friction slope (46)
rewriting and expanding S^ into components
_ \| /^ jjyv 3v _ 1 3v \ (6.7}
V ~~ OA \l li\."_ "* «\« ~ TV" <\» "" „ r\ 4- / Vt / /
^ • O OX g OX. g Ot
where S = bottom slope
= water surface slope (48)
|X = velocity head slope (49)
oX
= acceleration term (unsteady flow) (50)
g 3t
If the last three terms are small compared to SQ the assumption
that S - Sf can be made. Henderson (1966) offers typical values
to be 26 || . \ £ . | || . and fa g . respectively. Hence
it seems that the S ~ S* statement is generally applicable.
o r
115
-------
Although it has been shown that Q can reasonably be
assumed to be a function of y alone, it remains to be seen that
a flood wave can be described in this manner. Let the continuity
equation be written in the form:
+ B £ = 0 where B = width (51)
a t
introducing the total derivative
° (52)
eleminating B and dQ
al • at % + If ' ° <53>
rearranging
dx _ r _ 1 dQ /CAX
3t - C ~ B dy (54)
where C = wave celerity.
From this it is seen that an observer moving with velocity C will
see Q as a function of y alone.
Due to the simplistic nature of the equations, kinematic
theory can be easily implemented in computer formulations.
Accordingly, an attempt was made by the initial developers of
the SWMM to implement such a scheme into the TRANSPORT routing
technique .
This was attempted by developing dimensionless parameters
a and ty which are defined respectively as j— and *=• where the
subscript f denotes "full". Now if Mannings' equation is intro-
duced using the assumption S - Sf
Q = . AR SQ 2 (55)
116
-------
If the above equation is normalized by the full-flow values of
Q, A, and R, there results
2/3
* s §- - ^T75- ' f <»> (56)
Xf A R '
t Af Kf
The normalized flow rate will be a function only of the normalized
area since the hydraulic radius is a known function of the flow
area for a given geometry.
According to the Final Report, this method produced
reasonable hydrographs in conduits with steep slope (> ,001)
but not for mild slopes.
For this reason, it was found necessary to deviate
from the Sf = S approximation by adding two dynamic terms |£ and
' obviously> the method no longer produces a strictly
g
kinematic result on mild slopes. The method of implementation
in the computer program allows for the invert slope SQ to be used
alone if S is of sufficient magnitude.
o
117
-------
2.0 URBAN HYDROLOGY - STATE OF THE ART
In general, urban hydrology literature consists of a
maze of empirical, statistical and theoretical methods which in
many cases are more art than science. Many urban design techniques
consist of a combination of these methods. For the purposes of
this review, the urban hydrology literature has been segregated
into the following categories:
1) Empirical/Statistical Formulas,
2) Flood Frequency Analysis, and
3) Simulation Methods
2 .1 Empirical/Statistical Formulas
Empirical formulas are, in general, based on some form
of hydrologic hypothesis with resulting relationships developed
by empirical derivation of equations and/or coefficients. Empir-
ical formulas are by far the oldest and still the most widely
used method of predicting peak flows from urban watersheds. Of
the many empirical formulas (Hawksley, 1857; Adams, 1880; McMath,
1887; Parmely, 1905; Gregory, 1907; and historical review by
Chow, 1962) that have evolved since the early 1800*s, the so-
called "rational" method has continued to be the most widely used
method in the United States for estimating peak flows from small
urban watersheds (Ardis, Dueker, and Lent, 1969). The "rational"
method was first proposed in Great Britain by Lloyd-Davies in 1905;
however, this same method appears to have been applied in Ireland
as early as 1850 by Mulvaney and in the United States by Kuichling
in 1889. Because of the wide use of the "rational" method for
storm sewer design, the literature abounds with papers on this
method. Various hydrologists, such as Horner and Jens (1942),
Potter (1950), and Watkins (1962), have pointed out the limita-
tions of the "rational" method. Meek (1928) found the runoff
118
-------
coefficient used in the "rational" method is not a constant, but
varies with the characteristics of the storm and previous surface
moisture conditions of the watershed. MacDonald and Mehn (1963)
investigated the rational runoff coefficients for residential
areas in New Orleans. Summarized in Tables III and IV are ra-
tional runoff coefficients reported by various investigators
and the ASCE. More recently, Schaake, Geyer, and Knapp (1964),
as part of the Storm Drainage Research Project at Johns Hopkins
University, evaluated the application of the "rational" method
to 20 urban watersheds. Results from study were, in general,
inconclusive.
The basic theory of the unit hydrograph appears to have
been suggested first by Folse (1929). The Boston Society of Civil
Engineers (1930) stated, "the base of the flood hydrograph appears
to be approximately constant for different floods, and peak flow
tends to vary directly with the total volume of runoff". Three
years later, in 1932, Sherman introduced the basic concept of
the unit hydrograph. The unit hydrograph has become widely
accepted as one of the outstanding contributions to the science
of hydrology, The general concept of the unit hydrograph can be
summarized as follows (Morgan and Johnson, 1962):
1) For a given drainage area, the time base of
surface-runoff hydrographs resulting from
similar storms of equal duration are the same
regardless of the intensity of rainfall.
2) For a given drainage area, the ordinates of
the surface-runoff hydrographs from similar
storms of equal duration are proportional to
the volume of surface runoff.
3) For a given drainage area, the time distri-
bution of surface runoff from a particular
storm period is independent of that produced
by any other st(5rm period.
Because the unit hydrograph is such a valuable hydrologic tool,
many investigations have been undertaken to develop synthetic
119
-------
TABLE III
COMPARISON OF RATIONAL FORMULA RUNOFF COEFFICIENTS
(after MacDonald and Mehn, 1963)
Reference
H. E. Babbitt (a)
E. E. Seelye (b)
L C. Urquhart (c)
Fruhling (d)
ASCE (e)
Commercial
0.70-0.90
0.60-0.75
0.60-0.95
0.70-0.90
See Note (f)
0.45-0.70
0.35-0.55
0.25-0.40
0.20-0.50
0.25-0.40
Typt of Area
Farm or Cleared Land
0.05-0.25
0.10-0.35
0.05-0.25
0.10-0.25
Wooded Area
0.01-0.20
0.01-0.20
0.10-030
(a) Babbitt, 1956
(b) Seelye, 1960
(c) Urquhart, 1959
(d) Metcalf and Eddy
(e) ASCE, 1960
(f) 0.70-0.95
0.50-0.70
for Downtown areas;
for.Suburban areas.
TABLE IV
RUNOFF COEFFICIENT VALUES RECOMMENDED BY THE ASCE
(after MacDonald and Mehn, 1963)
Character of Surface
Streets:
Asphaltic
Concrete
Drives and Walks
Roofs
Lawns:
Sandy Soil, Flat 2%
Heavy Soil, Flat 2%
Parks
Run-off Coefficients
(0.83) 0.70-0.95
(0.88) 0.80-0.95
(0.80)
(0.85)
0.75-0.85
0.75-0.95
(0.08) 0.05-0.10
(0.15) 0.13-0.17
(0.18) 0.10-0.25
120
-------
unit graphs from relationships based on watershed features. The
watershed features commonly used in development of these relation-
ships are watershed area, shape, topography, channel slope, stream
density, and channel storage. Of the various synthetic unit
hydrograph methods that have been developed, probably the work
by Snyder (1938), USSCS (1957), Commons (1942), and Mitchell
(1948) are best known and most widely used. In particular, the
Snyder method has been an integral part of the hydrologic design
procedures of the USCE.
Chow (1962) developed a procedure to generate the runoff
hydrograph based on the unit hydrograph. Using data for Illinois
watersheds, he developed a procedure to define the unit hydrograph
which, in combination with a rainfall-runoff relationship based
on the SCS technique, yields the total runoff hydrograph.
Application of the unit hydrograph concept to urban
sewer systems was first reported by Horner and Flynt (1936).
Predicted hydrographs were in good agreement with observed data;
however, the data base was small, consisting of only three small
(less than 5 acres) urban watersheds. Tippetts, Abbett, McCarthy
and Stratton, Consulting Engineers, successfully applied the
unit hydrograph to runoff calculations for the city of Philadel-
phia in 1947. Van Sickle (1962) used the unit hydrograph as an
integral part of the development of urban design criteria for
the city of Houston. Eagleson (1962), using USCE hydrologic
data for five small (0.22 to 7.52 sq.mi.) urban watersheds lo-
cated in Louisville, Kentucky, investigated the application of
the unit hydrograph concept to these watersheds. Espey, Morgan,
and Masch (1965), while studying the various effects of urbaniza-
tion on the hydrologic characteristics of a small urban watershed
located in Austin, Texas, developed a set of synthetic unit
hydrograph equations (Table V) based on data from 24 urban water-
sheds located throughout the United States. Based on additional
hydrologic data from 11 urban watersheds located in Houston, Texas,
121
-------
TABLE V
URBAN EQUATIONS
(Espey, Morgan, and Masch, 1965)
- 1.93x10* A0'91
= 4.44xl05A1J7Gf1<19
and
W=1.34xl04A°'92Q-°-94
'BU
122
-------
Espey and Winslow (1968) tested and modified the unit hydrograph
equations developed by Espey et^ al. (1965). The modified equa-
tions are shown in Table VI.
Watkins (1962) developed a method (RRL) in England de-
scribing the runoff from urban sewer systems. The RRL method, in
general, is based on the same principles as the unit hydrograph
method. The RRL method assumes that pervious area does not pro-
duce runoff, and that impervious area yields 100% runoff. A time-
area diagram of directly connected impervious area of the basin
is constructed and the resulting runoff is routed through a linear
reservoir. Terstriep and Stall (1969) report testing on three
urban watersheds located in Baltimore, Maryland, Chicago, and
Champaign, Illinois, using the RRL method with good results. For 286
storms measured on 12 watersheds located in England, Watkins (1962)
also evaluated the application of the following urban design methods:
1) Rational (Lloyd-Davies) formula,
2) Tangent method,
3) Modified Tangent method, and
4) Unit Hydrograph method;
Catkins' evaluation of these methods can be summarized as follows:
1) "The rational method is unsatisfactory for all
but the smallest areas. For sewer systems con-
taining pipes not larger than 24 inches in
diameter, the range of available pipe sizes is
such that the errors introduced by the use of
this method are tolerable. It was found that
when the "rational" formula gives an accurate
estimate of the peak rate of runoff, it does so
because there is, for the particular area con-
cerned, a rough balance between the error intro-
duced by neglecting the shape of the area/time
diagram and the error introduced by neglecting
the reservoir retention in the sewer system.
2) "The tangent and modified tangent methods were
found to be normally no more accurate than the
"rational" formula, and it was also found that
in some circumstances the use of these methods
could increase the errors."
123
-------
TABLE VI
URBAN EQUATIONS
(Espey and.Winslow, 1968)
R
Q
T
B
W
50
W
75
16.4 $ L°'315 s-0.0488j-0.490
3.54x 104 T -1'10 A1'00
K
3.67 x 105 A1'14 -1'15
4.14 x 104 A1'03 Q-1'
1.34 x 104A°'92 -°'
124
-------
3) The unit hydrograph method, in the form usually
employed by hydrologists for calculating stream-
flow in natural catchments, was shown to be
unsuitable for designing urban sewer systems
owing to the difficulty of determining the shapes
of the unit hydrographs."
Other empirical equations have been developed for
overland flow by Izzard (1946), Kerky (1959), Brater and Woo
(1962), Morgali (1963), Henderson and Wooding (1964), Grace and
Eagleson (1966), and Woolhiser and Liggett (1967). These methods
all permit better definitions of the time to equilibrium and the
peak flow.
At Johns Hopkins University, Viessman and Geyer (1962)
developed equations which relate the peak flow and time of rise
to the rainfall rate, drainage area, and other physiographic
parameters. Equations were also developed to describe the con-
centration and recession curves for the storm hydrograph. The
equations were developed from data for very small and mostly
impervious areas.
2.2 Flood Frequency Analysis
Carter (1961) developed an empirical equation relating
the mean annual flood to the lag time, drainage area and percent
°f impervious cover to determine the effect of urbanization on
the mean annual flood in the vicinity of Washington, D. C. This
equation is:
Q = 223 K A0'8V0-45 (57)
where Q is the mean annual flood in cubic feet per second and is
equivalent to the flood having a recurrence interval of 2.33
^ars, A is the drainage area in square miles, T is the lag time
expressed in hours, and K is an adjustment factor based upon the
125
-------
degree of imperviousness of the area. The factor K is expressed
as:
„ _ 0.30 + 0.0045 I /cox
K (37315 ' * }
where I is the percent of impervious cover.
Martens (1961) further investigated the application of
Carter's equations to some small urban watersheds located in
Charlotte, North Carolina. Anderson (1968) also applied Carter's
equations to selected urban watersheds located on the east coast.
For a small urban watershed (4.1 sq.mi.) located in Austin,
Texas, Espey e£ aJL (1965) found good agreement between Carter's
predicted mean annual flood and measured data.
Putnam (1972) developed relationships between peak dis-
charges for various recurrence intervals and physiographic and
urban parameters for approximately 200 watersheds in the North
Carolina Piedmont Province of which 42 are urban. The flood fre-
quency equations relate the flood peak discharge for various recur-
rence intervals with lag time and the drainage area. The flood
frequency equations for the 50 and 100-year flood were derived by
extrapolating the regression coefficients, based on the regression
coefficients for the 2, 5, 10, and 25-year flood frequency equa-
tions. The equations are as follows:
P2 = 221 A0'87 T"°*60 (59)
P5 = 405 A0'80 T-°'52 (60)
P1Q = 560 A0'76 T-°'48 (61>
P25 = 790 A0'71 T-°-42 (62)
P50 = 990 A0'67 T"°'37 (63)
126
-------
P1QO = 1200 A0'63 T'°'33 (64)
where,
P. is the peak discharge in cubic feet per second
for the flood having the recurrence interval
indicated by the subscript,
A is the drainage area in square miles, and
T is the lag time in hours.
The lag time is obtained from the following equation:
T — n 50 -0 57
T = 0 49 ( I J&} U*-?UTU'-)/ .,
^ ' v •->/ * \ 65)
where,
T is the lag time in hours,
L is the length of the main water course in miles,
S is the stream bed slope of the main water course
in feet per mile, and
I is the ratio of the area of impervious cover to the
total drainage area.
The flood frequency equations developed by Putnam (1972) are
based on limited urban data. Because of the limited historical
data used in this analysis, extrapolation to the 50 and 100-year
flood should be exercised cautiously.
Benson (1964) studied the relationship between flood
peaks and physiographic and hydrographic factors for primarily
rural basin in the Western Gulf of Mexico Basin. The study area
comprised most of Texas and New Mexico, and small parts of Loui-
siana and Colorado. Multiple regression analysis was used to
develop relationships between peak discharge for various recur-
rence intervals and topographic hydrologic units; first, the area
where the annual peaks were caused primarily by local thunder-
storms or widespread tropical storms, and second, the relatively
area where the annual flood peaks were caused almost entirely
snowmelt. Peak discharges within the rainstorm-flood area
127
-------
were found to be significantly related to seven factors: drainage
area, rainfall intensity for a given duration and frequency, main-
channel slope, basin length, surface area of lakes and ponds, the
ratio of runoff to rainfall during the months of annual peak
discharge, and the annual number of thunderstorm days. The last
two factors, although statistically significant, play only a small
part in the variability of flood peaks. It may be expected that
some index of rainfall that involved a short duration and an
element of frequency would be highly correlated with the peak
discharge of a given frequency.
Flood frequency equations were developed by Espey and
Winslow (1972) which predict the peak flow for urban basins for
selected recurrence intervals. The equations were based on Log-
Pearson Type III flood frequency analysis from which the 2.33, 5,
10, 20, and 50-year recurrence interval peak flows for each water-
shed were determined. Data for sixty watersheds located in Texas,
Virginia, Maryland, Delaware, Michigan, Illinois, and Mississippi'
were used in the multiple regression analysis. The independent
variables used in the analysis were as follows: 1) drainage area
(A) sq. mi., 2) slope of the main channel (S), ft./ft. 3)
impervious cover (I), percent, 4) channel urbanization'factor (I),
dimensionless 5) rainfall (R), inches for 6-hour duration for the'
approximate recurrence interval (U. s. Weather Bureau, 1961).
The resulting flood frequency equations obtained by Espey and
Winslow (1972) are shown in Table VII. The correlation coefficients
shown are based upon the logarithms of the data, while the average
percent errors are based on the actual flow data rather than the
logs. It should be noted that the channel urbanization factor
becomes more significant and the percent impervious cover becomes
slightly less significant as the recurrence interval increases
128
-------
TABLE VII
DERIVED FLOOD FREQUENCY EQUATIONS
Urban Watersheds
Correlation Average %
Coeff (Logs) Error
Q2.33 3 169Au'''I^S^Rl-fO*-1-17 0.97 30
31
On - 172A0.80|0-27S0.43R1.734>-1.21 0>97
QIO
Q20
Q50
Q5
QIO
Q20
°50
8 "]73A 1 S
J Rib *
= 243A°-84I°-24S°-48R1-62 * '1-38
= 297A0-85I°-22S°-50R1'574>-1-61
50
83 1.13Qj;03
= 1.24Ql;05
= 1.34Ql;0|
» 1.47Q1-08
0.96
0.96
0.96
0.99
0.99
0.99
0.97
31
32
34
8
16
22
28
129
-------
2.3 Simulation Analysis
Simulation methods are generally characterized by their
attempt to describe in a somewhat analytical fashion the various
physical processes involved in the generation of runoff. The
methods usually consist of a series of "loss" functions, which
describe the amount of infiltration and the amount of interception
storage, and routing procedures which describe overland and chan-
nel flow. Presented in the following paragraphs is a general
summary of most of the current simulation methods.
One of the earliest simulation methods was the Stanford
Watershed Model (Linsley and Crawford, 1960). The model, devel-
oped for digital computer use, simulates both surface and ground-
water phenomenon, overland flow, channel routing, and snowmelt.
The model employs 21 parameters for rainfall runoff and 9 addi-
tional parameters for snowmelt. Several of the input parameters
are determined by trial for each watershed to be simulated. Many
modifications of the model have been made (James 1970* Liou
1970; Claborn and Moore, 1970; and Anderson, 1968) and will not
be discussed here.
The Soil Conservation Service (1971) has developed a run-
off model which is primarily applicable to rural watersheds. The
model is essentially a linear reservoir routing routine. The
runoff processes are defined by a series of 9 hypothetical re-
servoirs through which the flow is routed. The reservoirs
represent the various processes involved, such as infiltration,
groundwater recession, and storage.
An electric analog computer model has been developed
at Utah State University (Narayana et al., 1969; Evelyn et, a^.,
1970) . The model is operated for a series of sub-watershe"ds~for
which rainfall excess is determined by calculated infiltration,
interception and depression storage. Overland and channel flow
are determined by a linear routing scheme.
130
-------
A digital computer runoff model has been developed by
the University of Cincinnati (1970). The model consists of a
series of computational schemes to determine rainfall excess,
based on extensive literature review, a simplified version of the
two-dimensional overland flow equations, and Mannings equations
for pipe and channel flow.
One of the newest simulation models is the Storm Water
Management Model developed by Metcalf and Eddy et al. (1971) for
EPA. The model is designed to be a comprehensive tool to model
both the quantity and quality of runoff, the impact of the run-
off on the receiving water quality, and the effect of treating
the stormwater. The runoff part of the model utilizes an
infiltration loss function, Mannings' equation, and the continuity
equation to determine the flow rate both spatially and temporally.
The model is currently being used in several EPA projects, as out-
lined by Fields and Struzeski (1972), and is also being used in
various regions of the country for stormwater design purposes.
The Chicago Hydrograph Method developed by Tholin and
Keifer (1959) is essentially the overland flow routing method
developed by Izzard (1946) and a storage routing method for storm
sewers. The method considers infiltration losses and depression
storage in determining rainfall excess.
131
-------
3.0 REFERENCES
Adams, J. W. 1880. "Sewers and Drains for Populous Districts",
New York: D. Van Nostrand.
American Society of Civil Engineers. 1960. "Design and Construc-
tion of Sanitary and Storm Sewers", Manual of Engineering
Practice.
Anderson, E. A. 1968. "Development and Testing of Snowpack
Energy Balance Equations", Water Resources Research, Vol. 4,
No. 1, 19 -.37.
Ardis, C. V., K. J. Dueker, and A. T. Lenze. 1969. "Storm
Drainage Practices of Thirty-two Cities", Hydraulic Division,
ASCE HY1, pp. 383 - 407.
Benson, M. A. 1964. "Factors Affecting the Occurence of Floods
in the Southwest", U. S. Geological Survey Water Supply
Paper 1580 - D, 72 pages.
Betson, R. P. 1964. "What is Watershed Runoff?", Journal of
Geophys. Res., Vol. 69, No. 8. p. 1541-1552.
, R. L. Tucker, and F. M. Haller. 1969. "Using Analytical
Methods to Develop a Surface-Runoff Model", Water Resources
Res., Vol. 5, No. 1, p. 103.
Boston Society of Civil Engineers. 1930. "Report of the Committee
on Floods", Journal of the Boston Society of Civil Engineers,
Vol. 17, No. 7 (September), pp. 285 -464.
Brater, E. F., and D. C. Woo. 1962. "Spatially Varied Flow from
Controlled Rainfall", Journal of the Hydraulics Division,
ASCE, Vol. 88, HY6, pp. 31 - 56.
Carter, R. W. 1961. "Magnitude and Frequency of Floods in
Suburban Areas", Geological Survey Research, Prof. Paper 424-B,
U.S. Geological Survey, Washington, D.C.
Chow, Ven Te. 1962. "Hydrologic Determination of Waterway. Areas
for the Design of Drainage Structures in Small Drainage
Basins", University of Illinois Engineering Experiment
Station Bulletin No. 462.
1964. "Handbook of Applied Hydrology", McGraw - H.V.
Book Co., New York.
132
-------
Claborn, B. J. and W. Moore. 1970. "Numerical Simulation in
Watershed Hydrology", Hydraulic Engineering Lab., University
of Texas, Austin, Technical Report HYD 14-7001.
Commons, G. C. 1942. "Flood Hydrograph", Civil Engineering, Vol.
12, pp. 571 - 572.
Eagleson, P. E. 1962. "Unit Hydrograph Characteristics for
Sewered Areas", American Society of Civil Engineers, Pro-
ceedings, Vol. 88, No. HY2 (March).
Espey, Jr. W. H., C, W. Morgan and F. D. Masch. 1965. "A Study
of Some Effects of Urbanization on Storm Runoff from a Small
Watershed", Center for Research in Water Resources, Univer-
sity of Texas, CRWR-2, July.
— and D. E. Winslow. 1968. "The Effects of Urbanization
on Unit Hydrographs for Small Watersheds", Houston, Texas,
1964 - 1967, TRACOR Doc. No. 68-975-0.
1972. "Urban Parameters Affecting the Runoff Response
of Small Urban Watersheds", Paper presented at the 1972 ASCE
Annual and National Environmental Engineering Meeting in
Houston, Texas.
and C. W. Morgan. "The Effects of Urbanization on Peak
Discharge", Paper presented at the Water Resources Symposium
No. 2, on the Effects of Watershed Changes on Streamflow,
University of Texas.
Evelyn, J. B., V. V. D. Narayana, J. P. Riley and E, K. Israelsen.
1970. "Hydrograph Synthesis for Watershed Subzones from
Measured Urban Parameters", Utah Water Research Lab., Utah
State University, Logan, Utah.
Field, R. and E. J. Struzeski. 1972. "Management and Control of
Combined Sewer Overflows, JWPCE.
polse, J. A. 1929. "A New Method of Estimating Stream Flow",
Carnegie Institution of Washington, Publication 400.
Grace, R. A. and P. S. Eagleson. 1966. "The Modeling of Overland
Flow", Water Resources Research, Vol. 2, No. 3, pp. 393 - 403.
Dayman, W. M. and P. S. Eagleson. 1969. "Streamflow Record Length
for Modeling Catchment Dynamics", Hydrodynamics Lab. Report No.
114, M.I.T., 137 p.
Gregory, C. E. 1907. "Rainfall and Runoff in Storm Water Sewers",
Transactions, American Society of Civil Engineers, Vol. 58
(1907), p. 475.
133
-------
Hawksley. 1857. "Report of Commission of Metropolitan Drainages",
London.
Henderson, F. M. and R. A. Wooding. 1964. "Overland Flow and
Groundwater Flow from a Steady Rainfall of Finite Duration",
Journal of Geophysical Research, Vo. 69, No. 8, pp. 1531 -
1540.
Horner, W. W. and F. L. Flynt. 1942. "Relation Between Rainfall
and Runoff from Small Urban Areas", Trans. Amer. Soc. of
Civil Engineers, 60, 1135 - 78.
and S. W. Jens. 1942. "Surface Runoff Determination from
Rainfall Without Using Coefficients", Transactions, ASCE,
Vol. 107, pp. 1039 - 1075.
Horton, R. E. 1935. "Surface Runoff Phenomena, PT I, Analysis
of the Hydrograph", Publication No. 101, Edwards Bros.,
Ann Arbor, Mich.
. 1939. "Analysis of Runoff Plot Experiments with Varying
Infiltration Capacity", Trans. Amer. Geophy. Union, Part IV.,
p. 693.
Izzard, C. F. 1946. "Hydraulics of Runoff from Developed Surfaces"
Proc. Highway Research Board, Vol. 26, pp. 129 -145.
James, L. Douglas. 1970. "An Evaluation of Relationships Between
Streamflow Patterns and Watershed Characteristics Through
the use of OPSET: A Self Calibrating Version of the Stanford
Watershed Model", Lexington, University of Kentucky, Water
Resources List, Research Report No. 36.
Kent, K. M. 1968. "A Method for Estimating Volume and Rate of
Runoff in Small Watersheds", SCS-TP-149, 19 p.
Kerky, W. S. 1959. "Time of Concentration for Overland Flow",
Civil Engineering, Vol. 29, p. 60.
Kohler, M. A. and R. K. Linsley. "Predicting the Runoff from
Storm Rainfall", U. S. Weather Bur. Res. Paper No. 34,
Washington, D.C.
Kuichling, E. 1889. "The Relation Between the Rainfall and the
Discharge of Sewers in Populous Districts", Trans. Amer.
Soc. Civil Engineers, Vol. 20, (January), pp. 1 - 56:
Discussion, 57-60.
134
-------
Linsley, R. K., Jr., M. A. Kohler and J. L. H. Paulhus, 1949.
"Applied Hydrology", New York, McGraw-Hill Book Co.
and N. H. Crawford. 1960. "Computation of a Synthetic
Streamflow Record on a Digital Computer", Pub. No. 51, In-
ternational Association of Scientific Hydrology, pp. 526 -
538.
Liou, E. Y. 1970. "OPSET: Program for Computerized Selection of
Watershed Parameter Values for the Stanford Watershed Model",
Lexington, University of Kentucky, Water Resources Inst. Re-
search Report No. 34.
Lloyd-Davies, D. E. 1905. "The Elimination of Storm-Water from
Sewerage Systems", Min. Proc. Inst. Civ. Engrs., pp. 41 - 67.
MacDonard, F. W. and A. Mehn. 1963. "Determination of Runoff
Coefficients", Public Works, November, pp. 74 - 76.
Martens, L. A. 1961. "Flood Inundation and Effects of Urbanization
in Metropolitan Charlotte, North Carolina".
McMath, R. E. 1887. "Determination of the Size of Sewers",
Transactions, American Society of Civil Engineers, Vol. 16,
p. 183.
Meek, J. B. L. 1928. "Sewerage, With Special Relation to Runoff",
Institution of Civil Engineers, Engineering Conference 1928,
Report of Discussions, 162 - 8, Discussion, 168 - 74,
Metcalf and Eddy, Inc. 1971. "Storm Water Management Model, Vol.
1 .Final .Report", University of Florida, Gainseville, and
Water Resources Engineers, Inc., for EPA.
Miller, C. R. and W. Viessman, Jr. 1972. "Runoff Volumes from
Small Urban Watersheds", Water Resources Research, Vol. 8,
No. 2.
Mitchell, W. D. 1948. "Unit Hydrograph in Illinois", State of
Illinois, Division of Waterways, Springfield, Illinois.
, D. 0. 1968. "Synthesizing Daily Discharge from Rainfall
Records", ASCE Proc. Paper 6119, HY5, Vol. 94.
, J. R. 1963. "Hydraulic Behavior of Small Drainage Basin",
Department of Civil Engineering, Technical Report 30, Stan-
ford University.
135
-------
Morgan, P. E. and S. M. Johnson. 1962. "Analysis of Synthetic
Unit-Graph Methods", Journal of the Hyd. Div. ASCE. HYS.
pp. 199 - 220.
Mulvaney, T. J. 1850. "On the Use of Self Registering Rain and
Flood Gauges in Making Observation of the Relation of Rain-
fall and of Flood Discharges in a Given Catchment", Trans.
Instn. Civ. Engrs. Ire., 4(2), IB.
Narayana, V. V. D., J. P. Riley and E. K. Israelsen. 1969.
"Analog Computer Simulation of the Runoff Characteristics of
an Urban Watershed", Utah Water Research Lab., Utah State
University, Logan, Utah.
Orlob, G. T. 1974. "Urban Storm Drainage, an Overview", Tech-
nical Memorandum No. 24, Management of Urban Storm Runoff,
ASCE Urban Water Resources Research Program.
Parmely, W. C. 1905. "The Walworth Sewer Cleveland, Ohio",
Transactions, American Society of Civil Engineers, Vol. 55,
p. 3456.
Potter, W. D. 1950. "Surface Runoff from Agricultural Watershed",
Surface Drainage, Highway Research Board Report No. 11-B,
pp. 21 - 35.
Putnam, A. L. 1972. "Effect of Urban Development on Floods in
the Piedmont Province of North Carolina", U.S. Geological
Survey, Open File Report, Raleigh, North Carolina.
Sauer, S. P. and G. F. Pinder. 1970. "Numerical Simulation of
Flood-wave Modification Due to Bank Storage", Paper No. H-9,
AGU Fall National Meeting, San Francisco, California.
Schaake, J. C., J. C. Geyer and J. W. Knapp. 1964. "Runoff
Coefficients in the Rational Method", Technical Report No. 1,
The Storm Drainage Research Project, The Johns Hopkins
University, Baltimore, Maryland.
Sherman, L. K. 1932. "Streamflow from Rainfall by the Unit-Graph
Method", Engineering News - Record, Vol. 108, pp. 501 - 505.
Sittner, W. T., C. E. Schauss and J. C. Monro. 1969. "Continuous
Hydrograph Synthesis with an API - Type Hydrologic Model",
Water Resources Res. Vol. 5, no. 5.
Snyder, F. F. 1938. "Synthetic Unit Graphs", Transactions, Amer-
ican Geophysical Union, Vol. 19, pp. 447 - 454.
136
-------
Soil Conservation Service. 1957. "Engineering Handbook", U. S.
Department of Agriculture, Washington, D.C., Sec. 4, Hydrology,
Supplement A.
Soil Conservation Service. 1971. "National Engineering Handbook
Section 4, Hydrology", USDA.
Terstriep, M. L. and J. B. Stall. 1969. "Urban Runoff by Road
Research Laboratory Method", Journal Hydraulics Division,
ASCE, Vol. 95, pp. 1809 - 1834.
Tholin A. L. and C. J. Keifer. 1959. "The Hydrology of Urban
Runoff", Journal Sanitary Engineering Division, ASCE, Vol.
85, pp. 47 - 106.
University of Cincinnati. 1970. "Urban Runoff Characteristics",
11024 DQU 10/70, Cincinnati, Ohio, Phase I Interim Report
for EPA.
U.S. Geological Survey. 1971. "Seminar on Digital Modeling of
Stream Systems. Water Resources Division Training Center,
Denver, Colorado, May 17-28, 1971.
Van Sickle, Donald. 1962. "The Effects of Urban Development on
Storm Runoff", The Texas Engineer, Vol. 32, No. 12.
Viessman, Jr. Warren and John C. Geyer. 1962. "Characteristics
of the Inlet Hydrograph", ASCE, HY5, (September).
Watkins, L. H. 1962. "The Design of Urban Sewer Systems", Road
Research Technical Paper No. 55.
Wiitala, S. W. 1967. "Some Aspects of the Effect of Urbanization
on Floods in Jackson, Mississippi", Prof. Paper 575-D, D259-
D261, USGS, Dept. of Interior, Washington, D.C.
Woolhiser, D. A. and J. A. Liggett^ 1967. "Unsteady, One-
dimensional Flow Over A Plane: The Rising Hydrograph", Water
Resources Research, Vol. 3, No. 3, pp. 753 - 771.
137
-------
QUALITY ASPECTS OF URBAN STORMWATER RUNOFF
BY
Larry A. Roesner*
BACKGROUND
It has only been about ten years since sanitary engineers began
to realize the significance of urban runoff as a source of pollution in
receiving waters. Historically [l], the earliest sewers were built for the
collection and disposal of storm runoff, For convenience, these sewers
discharged to the nearest watercourse. In later years, domestic and
industrial wastewaters were discharged into these sewers, thereby
converting them to the combined sewers. As the significance of the
pollutional effects of discharging raw sewage to the watercourses became
recognized, the major cities embarked upon programs of interceptor
sewers to divert some multiple (generally 1.5-5) of the average dry
weather flow to a central location for treatment prior to disposal.
Even with interceptors however, stormwater overflows from
these combined systems were still observed to carry significant pollution
loads. This fact caused water pollution control agencies to begin thinking
that separation of storm runoff and sanitary sewage was the answer to
the pollution problem and for several years (about the mid-19601 s) there
were many studies on alternative methods of separating sanitary sewage
and storm water in existing combined systems. Sewer separation was found
to be very expensive, however [2], and so, while a few cities undertook
separation programs, most cities began to look for alternative methods of
dealing with the problem. Perhaps the most significant result to come out
of this push for separation is that combined sewer systems are no longer
designed for new developments; separate systems are installed.
While the push for sewer separation was on, the federal
government was sponsoring research on the quality characteristics of
urban runoff per se. Tables 1 and 2 (taken from Reference 1) were
compiled from these studies and show the comparison between the
characteristics of combined sewer overflow and urban stormwater.
It is apparent that the BOD and solids loads in urban runoff are signi-
ficantly higher than those found in combined overflows which represent
a mixture of sanitary sewage and stormwater.
As a result of these findings, the emphasis on stormwater
pollution control is now being placed 013. controlling the pollutant load
of all stormwater discharges whether they emanate from a combined
system or from a separate storm sewer system.
t
Principal Engineer, Water Resources Engineers, Walnut Creek, California.
138
-------
TABLE 1
Characteristics of Combined Sewer Overflows*
ChmcurUtk IU»t*«* Value*
TS3 (
TS (lUg/l)
Volatile TS (lllg/l)
pll
Scttleable solids (ml/1)
Organic X (nig/I)
NIliN (tiik'/D
Soluble rO,(niK'/D
Total coHforma (iw./\M rti!)
Feral conforms (no./HH) ml)
Fecal jtrc|»lo«-«Ti (nu./IWl nil)
JSO-2..UW
15-320
4.9-8.7
2-1,550
1.5-3.U
O.t-12.5
0.1-6.2
20,(«M»-OI) X Mi«
! 20,000-17 X I0«
20,000-2 X 10'
'Selected data.
TABLE 2
Characteristics of Urban Stormwater*
BODt(mg/l)
coo(mg/l)
Tss(inu/l)
VolatileTS(m£/l)
Settletble wlids ( ml/1)
Oivanic N (mg/1)
NIliN (mt/l)
Soluble P04700
5-3,100
2-U.JOO
450-14,600
12-1,600
0.5-5,400
0.1-16
0.1-2.5
0.1-10
0.1-125
2-J5,OOOf
0-110
0-0.2
0-1.9
200-146 * 10«
55-112 X 10«-
200-1.2 X 10«
•Selected data.
t With bitfhway tlcicing.
SOURCES OF POLLUTANTS
Basically, pollutant loads are introduced into urban runoff from
three sources:
1, The land surface itself, primarily impervious surfaces;
2, Catch basins; and
3. The sewers in combined systems.
these three sources, the land is the most important.
139
-------
Catch basins can be a source of first-flush or shock pollution.
An American Public Works Association (APWA) study [3] in Chicago
found that:
". . . the liquid remaining in a basin between runoff
events tends to become septic and that the solids trapped
in the basin take on the general characteristics of
septic or anaerobic sludge. The liquid in catch basins
is displaced by fresh runoff water in the ratio of one-
half the volume for every equal volume of added liquid.
During even minor rainfall or thaw this displacement
factor can release the major amount of the retained
liquid and some solids, The catch basin liquid was found
to have a BOD content of 60 ppm in a residential area,
For even minor storms, the BOD of the catch basin liquid
would be seven-and-one-half (7-1/2) times that of the
runoff which had been in contact with street litter.
Improved design of catch basins, and better operational
and maintenance practices, could reduce this first-flush
pollutional effect. "
In combined sewer systems, wastewater is incorporated into
the storm runoff. In addition, the storm runoff, as it passes through
large sewers, scours sediment deposited by wastewater flows
during preceding dry-weather periods. Figures 1 and 2 illustrate the
effects of wastewater sewage and of catch basins and storm sewer scour
on the quality of stormwater overflows [4].
As stated above, the most important contributor of pollutants
to urban runoff is the land surface itself, primarily the streets and gutters
and other impervious areas directly connected to streets or storm sewers.
Pollutants accumulate on these surfaces in a variety of ways. There is,
for example: debris dropped or scattered by individuals; sidewalk
sweepings; debris and pollutants deposited on or washed into streets from
yards and other indigenous open areas; wastes and dirt from building and
demolition; fecal dropping from dogs, birds and other animals; remnants
of household refuse dropped during collection or scattered by animals or
wind; dirt, oil, tire and exhaust residue contributed by automobiles; and
fallout of air pollution particles. The list could go on and on. Irrespective
of the way in which pollutants accumulate on the urban watershed, they
are generally associated with one of the following forms of street litter:
1. Rags,
2. Paper,
3. Dust and dirt,
4. Vegetation, or
5. Inorganics.
Table 3, which gives estimated street litter components for a residential
area in Chicago, provides a rough measure of the relative importance
of these components.
140
-------
1500
o<
31
'"2'clOOO
"
500
o
D
_ 400
8
300
200-
100
300
S— 20°
SH^
05 p>
mW^ 100
UJ
o
800
600
^400
200
1100
11:00
12^00
1300
14^00
combined sanitary and
stormwater overflow
12:00
13=00
11=00 12=00 I3--00
CLOCK TIME
stormwater overflow only
14^00
14=00
FIGURE 1
Comparison of Stormwater and Combined Overflows
(Selby Street, San Francisco)
141
-------
>
0
300
100
30Oi
WOO
contribution from catchbasins
and storm sewer scouring
combined overflow without
catchbasin or scour effects
itoo
1200
1300
14=00
1100
1*00 13=00
CLOCK TIME
14=00
FIGURE 2
First-Flush Pollutional Effects of Catch Basins
And Sewer Scour on Combined Sewer Overflow
(Selby Street, San Francisco)
142
-------
TABLE 3
Monthly Summary of Estimated Street Litter Components,
From a 10-acre (4 ha) Residential Area, Chicago**
Street Refute Components
(Toni/MontM
Rag.
Paper
Dutt&
Dirt
Vegetation
Jan.
Feb.
March
Apr"
May
June
July
Aug.
Sept.
Oct.
Nov.
Dec.
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.036
.036
.036
.036
.036
.036
.036
.036
.036
.036
.036
.036
.55
.55
.55
.55
.55
.55
.55
.55
.55
.55
.55
.55
.00
.00
.00
.08
.08
.08
.08
.08
.08
.83
.83
.00
TOTAL*
.0180
.432
6.60
2.22
•Sew* Mali have been rounded off.
Inorganic
.09
.09
.09
.09
.09
.09
.09
.09
.09
.09
.09
.09
1.08
Total*
.68
,6b
.76
.76
.76
.76
.76
.76
1.56
1.56
.68
10.48
It is readily apparent that the most significant component is dust and
dirt except during the fall of the year when vegetation (primarily leaves)
becomes the dominant component.
TYPES OF POLLUTANTS AND LOADING RATES
Nearly all of the pollutants found in urban runoff are associated
with the dust and dirt component of street litter. By type, COD, BOD,
and solids (suspended and settleables) are found in the greatest quantity.
Nitrogen and phosphorus are also found in significant quantities. In areas
where street deicing by salting is practiced, winter runoff contains very
high chloride concentrations. Other pollutants found in urban runoff
include pesticides, herbicides, fertilizers and other chemical additives,
heavy metals, and many other known and unknown pollutants.
Data on the rate at which pollutants accumulate on an urban
Watershed is very scanty. In fact, it is almost non-existent. A lot
of data has been collected on the quality of combined sewer overflows
and stormwater discharges for various cities in the United States (Tulsa,
Oklahoma; Washington, D.C.; Atlanta, Georgia; San Francisco,
California; Sacramento, California; and Roanoke, Virginia) as a result
°f the U.S. Environmental Protection Agency's demonstration grants
**This table is a reproduction of Table 4 in Reference 3,
143
-------
program for abatement of stormwater pollution. The studies are reported
in EPA's Water Pollution Research Series; however, the dissimilar forms
in which the data is reported makes it difficult and in some cases
impossible to generalize. The problem is that data is often presented
as average concentrations or as pounds of pollutant runoff per inch of
rain, and the reported values may be for combined wastewater and storm
runoff rather than for storm runoff alone.
Even for a given watershed, there is no apt description of
"typical" stormwater runoff characteristics because of the variablility
of rainfall-runoff patterns, Thus, reports of "mean concentration" or
pounds per inch of rainfall are meaningless as generalized variables and
they show poor correlation with any runoff parameters of interest.
Results from a demonstration project conducted in Tulsa,
Oklahoma [5], were summarized in terms of pounds of loading per day
per mile of street for each of 15 areas sampled in the study. These
results, presented in Table 4, give an indication of the magnitude of
pollutant buildup for different land uses. These findings must be viewed
with caution, however, because they were computed by taking the
"average" concentration of the pollutant for all events monitored, which
when multiplied by the total storm runoff gave "total annual mass
emissions" converted into a rate per day per mile. A much more useful
way to have developed this data would have been to sum the products
of discharge and concentrations over each of the observed events and then
to sum all such total storm emissions over a year.
One of the best existing sources of information on the rate
of accumulation of pollutants on urban watersheds is that data collected
in a field study in Chicago by APWA [3], This study determined the rate
of buildup of dust and dirt in the streets on a number of different test
areas and then related the concentrations of various pollutants to the dust
and dirt. A summary of their findings is contained in Table 5.
To convert the data contained in Table 5 into a form comparable
with that of Table 4, the Dust and Dirt is multiplied by 2 (gutters per
street) x 52.8 (100's of feet of gutter per mile) x (constituent concen-
tration/1000). E. g. , the rate of BOD accumulation on an urban area that
is single family residential is: 0.7 x (2 x 52.8) x 5.0/1000 = 0.36
Ibs/day/mile. Rates of pollutant buildup in pounds per day per mile are
given in Table 6.
Comparison of the values in Tables 4 and 6 indicates that
while the Chicago data is consistently lower than the Tulsa data, the
rates of buildup are similar in commercial areas except forPO4. Multiple
family residential values for the Chicago area also compare well with
the Tulsa data except for PO4. The principal reason for the lower values
in the Chicago data is that the reported values are for the soluble portion
of the constituents and do not include that portion found in suspended and
settleable material. Also, the Tulsa data are approximations based
144
-------
TABLE 4
Average Daily Loads Per Mile of Street*
(Tulsa, Oklahoma)
Total
Test Street
Area Miles
Residential
3 14.87
5 16.32
7 6.84
8 6.97
9 3.11
11 49.05
13 5.58
15 2.06
Commerical
2 7,41
10 12.99
12 3.39
Industrial
1 11.46
4 28.40
6 12.24
Average Values
Residential
Commercial
Industrial
Average Load: Ibs/day/mile of street
BOD
1.41
2.80
1.20
2.72
1. 12
1.60
2.58
2.47
2.54
2.10
4.53
4.85
3.98
1.70
1.98
3.06
3.51
COD
11.46
21.43
7.20
20.89
13.09
13.29
15.16
8.67
15. 12
20.44
25.47
41. 10
29.29
12.73
13.9
20.3
27.7
Total
Solids
120
43
63
69
47
66
81
56
92
82
113
838
175
49
63. 1
95.7
354.
Organic
Kjeldahl
Nitrogen
0.26
0.11
0.12
0.12
0.07
0.08
0.25
0.07
0.32
0. 16
0.22
0.41
0."28
0.09
0.14
0.23
0.26
Soluble
O r th o phosphate
0.34
0. 13
0.10
0.21
0. 11
0. 15
0.20
0.17
0.29
0.13
0.30
1.30
0.30
0. 13
0.18
0.24
0.57
^Reproduced from Reference 5.
145
-------
on over-simplified computations. Thus, of the two sets of reported data
the Chicago data is probably better, given that the amount of constituents
contained in the solids can be determined.
TABLE 5
APWA Findings on Rate of Pollutant Buildup
On Urban Watersheds*
Amount of Dinf ond 0
Lend Uie
Commercial
Indutrrial
Mul'ipl* family
Single family retidenee
Aitumed weight** average
Amovni
Item
Water Soluble (mg/g)
Volatile Water Soluble (mg/g)
BOO (mg/9)
COO (mg/9>
tQ* <<"9/fl}
N (mg/9,1
Total plate count«/«. (* 1000)
Confirmed c«l!feWfl (x 1000)
Fecal enterecacci/g
rf and Strength of BOO
Amt , cf 0/0
by lond -If
by Land Uie
BOO of D/D
Icy Joy' 100 ft of curb mg/g
3.3
4.6
2.3
0.7
1.5
7.7
3
3.6
_5
5
of Pollutant by Type of land Uie
Single Family
6.0
3.3
5.0
40
.05
.48
10,900
1.300
645
Multiple Family
5.6
3.4
3.6
40
.05
.61
18,000
2,700
518
Co/nmerc ial
12.4
6.9
7.7
39
.07
.41
11,700
1,700
319
TABLE 6
Average Daily Loads Per Mile of Street
(Chicago, Illinois)
Average
Land Use
BOD
Single Family Residential
Multiple Family Residential
Commercial
Industrial
0,
0.
2.
1.
36
87
70
45
Load: Ibs/ day /mile of
COD
2,
9>
13,
95
70
6
0
0
0
N
.03
. 15
.14
0
0
0
street
PO4 '
.004
.012
.024
One fact that is quite evident from both the Chicago and Tulsa
data is that the rate of buildup of pollutants on an urban watershed varies
significantly with land use. Intuition would tell us this is true. Both
sets of data indicate that industrial and commercial areas are much dirtier
than residential areas. This would be expected since there is higher
*See Reference 3.
146
-------
pedestrian and vehicular traffic densities in these areas-, The data shows
that pollutant accumulation rates are approximately one and one-half to
five times as great in commercial and industrial areas as they are in
residential areas.
Additional information of pollutant accumulation rates can be
found in Reference 6.
ENTRY OF POLLUTANTS INTO URBAN RUNOFF
The first raindrops that fall on an urban watershed simply wet
the land surface. As additional rain falls, the impervious surface will
become wet enough that some of the water begins to form puddles, filling
the depression storage. This initial rain begins to dissolve the pollu-
tants in the gutters, streets, and on other impervious surfaces and
eventually, as this water actually begins to flow off the watershed, it
carries the dissolved material in it.
As rainfall intensity increases, overland flow velocities
become sufficient to pick up solids. Suspended solids are, of course,
picked up at smaller velocities than settleable solids. The settleable solids
are carried off the watershed in two ways. If the velocity is sufficiently
high, the settleable solids may be suspended in the overland flow. At
lower velocities, particles may simply be rolled along the bottom surface
toward the stormwater inlet.
The rain that initially falls on pervious surfaces infiltrates
into the ground. If the rainfall is sufficiently intense, the infiltration
capacity may be exceeded and the excess rainfall begins to fill the
depression storage on the pervious surfaces. Finally, if the rainfall
is of sufficient intensity and duration, runoff will begin to flow off the
pervious areas, onto the impervious areas and thence into the stormwater
inlets. Present experience, however, indicates that the amount of runoff
and hence the pollution loads contributed from pervious surfaces in urban
areas , 1s small compared to that coming from the impervious areas
and can be neglected in determining the quality of surface runoff. This is
especially true of surfaces covered with vegetation such as lawns and
gardens. Figure 3 illustrates the differences in runoff and pollution
load from a watershed that would occur if it was converted from a park
(90% pervious) into a multiple residential area (20% pervious).
ESTIMATION OF THE RATE OF POLLUTANT BUILDUP
ON URBAN WATERSHEDS
Since dirt is the major component of street litter and is the
Primary source of pollutants in urban runoff, the most basic approach
f°r estimating pollutant buildup rates would be to relate them to the dust
and dirt accumulation rates.
147
-------
ALTERNATIVE LAND USE
(Multiple Residential)
EXISTING LAND USE
(Park)
10:00
11:00 12:00
CLOCK TIME
13:00
FIGURE 3
Effect of Changed Land Use on Characteristics
f Subcatchment Runoff, Selby Street
148
-------
Using APWA units (Table 5) for the rate of dust and dirt
accumulation, the rate of buildup DD for a given land use L can be
expressed as:
= ddLx(GL/100)xAL
where
DD_ = rate of dust and dirt accumulation on a watershed
of land use L, in Ibs/day;
dd.. = rate of dust and dirt accumulation on watershed L
in Ibs/day/ 100 feet of gutter;
GT = feet of gutter per acre in watershed L; and
j_i
A T = area of watershed L in acres.
j_j
The initial quality of a pollutant p on watershed L, at the
beginning of a storm can then be computed as:
Pp = (Fp x DDL x ND) + Ppo (2)
where
P = total pounds of pollutant p on the watershed L. at the
" beginning of the storm;
F = pounds of pollutant p per pound of dust and dirt;
ND = number of dry days since the last storm, and
P = total pounds of pollutant remaining on watershed L
" at the end of the last storm.
In practice, P is usually limited to the amount that would be
accumulated in a 90-day dry period. The reason for this is that the
efficacy of extrapolating daily buildup rates beyond this point (which
Was arbitrarily selected) is uncertain. Moreover, if equation (2) is used
repetitively over long periods of time, positive errors could tend to
accumulate in P resulting in overly large values of P .
If street sweeping is practiced on the watershed, the number of
days since the last storm must be modified to account for the number
°f street sweepings that occurred since the last rainfall. The correct
expression to use for P is then:
P
Pp = Ppo(1-E)n+NS XDDL xFp
DD
,LxFp (N -nNs) (4,
149
-------
where
Nq = number of days between street sweepings;
n = number of times the street was swept since the
last storm; and
E = efficiency of street sweeping (0.6 to 0. 95).
DETERMINATION OF URBAN RUNOFF POLLUTION LOADS
To compute the amount of pollutant washed off the watershed
during a storm, it is assumed that the amount of pollutant removed at
any time t is proportional to the amount remaining:
dP
-at* = -KPP <4>
We stated earlier that the runoff rate Q also affects rate of pollutant
removal, therefore K must be functionally dependent upon Q, However,
given two identical watersheds except for their area size, for the same
rainfall rate r on both watersheds a higher runoff rate would occur from
the larger watershed. This area effect can be eliminated by dividing the
runoff Q by the impervious area of the watershed. The impervious area
is used because only a negligible amount of the runoff comes from the
pervious area. Since cfs per acre are equivalent to inches per hour,
we can say that K is functionally dependent on the runoff rate R from
the impervious area, where R is in inches per hour. Finally, assuming
that K is directly proportional to R and that a uniform rainfall of 1/2
inch per hour would wash away 90 percent of the pollutant in one hour
(a somewhat arbitrary assumption), we can say that K = 4.6R. Making
this substitution into equation (4) and integrating over a time interval
At (during which R is held constant) gives:
Pp(t+ At) = Pp(t)e-4'6RAt (5)
Equation (5) is the basic form of the overland flow quality model developed
by Metcalf & Eddy, Inc. , as part of the EPA Stormwater Management
Model [7]. Although it is simplistic and contains many assumptions,
it is the best overland flow water quality predictor or simulation model
that presently exists. Moreover, experience with that model (See
Reference 7) has shown it to give fairly good results.
Some idea of how equation (5) behaves can be gained by exami-
nation of Figures 4 and 5. Figure 4 shows that for a constant runoff
rate R, the amount of pollutant remaining on the watershed decays
exponentially. Under a time varying R the picture is quite different, as
illustrated by the upper graph Figure 5. From the curve of P vs. t
it can be seen that the amount of pollutant removed during an interval
150
-------
TIME, t
FIGURE 4
Basic Form of the Overland Flow Quality Model
151
-------
PO
Pt
O
Z
UJ *
HX
zc/>
s
If At
V
\
-4-J--
L(t)-At
t t+At
TIME,!
_
LJL
O
cr
o
UJ
X
U. U.
I t0
pt,H
TIME,!
FIGURE 5
Development of Pollutograph (M vs, t)
from Time History of P ^
152
-------
At is P(t) -P(t+At), The rate of removal of mass from the watershed M
is simply [P(t)-P(t+ At)]/ A t, which can be expressed as: P
M = P(t) x(l-e-4'6RAt)/At (6)
The variation of M with time for the associated hydrograph
is plotted in the lower graph of Figure 5. A plot of M versus t is termed
a pollutograph, one of the most informative methods for expressing the
pollutant load carried by urban runoff. To determine the concentration
of a pollutant in the runoff as a function of time, one simply divides the
pollutograph value M by Q (with appropriate conversion factors),
THE ROLE OF THE TRANSPORT SYSTEM
All of the preceding discussion has dealt with determination of
the pollutant loads and quality of water washed off of the urban watersheds.
Thus, the pollutographs shown in Figure 5, describe the rate of mass
transport of a pollutant into the storm sewer system from a single water-
shed in an urban drainage area. For the watershed shown in Figure 6,
there would be twenty such pollutographs formed, These individual
pollutographs are then routed through the storm sewer or transport system
using mass continuity to develop an outfall pollutograph at the lower end
of the system. Depending upon the travel time in the transport system
and the time to peak for the individual pollutographs, the resultant
pollutograph at the outfall may have a high peak due to compounding
of individual peaks from the tributary watersheds, or it may have a lower
peak and broader base if the travel time in the sewer system is long
compared to travel time on the individual watersheds.
The compounding effect is observed for stationary storms on
steeper watersheds while the low-peak broad-base outfall pollutographs
are observedin very flat systems. The compounding effect can also occur
in a flat system however if the storm is moving along the fall line,
EFFECT OF LAND USE CHANGES ON
STORM RUNOFF CHARACTERISTICS
To demonstrate the effect of land use changes on storm runoff
characteristics, assume the park area in the Selby Street system (see
figure 6) was changed to a multiple residence area. The major
significance of this change is to increase the impervious area of the sub-
Catchment from 1 1 to 75 percent and to generate a dry weather sanitary
flow from the area.
The effect of the modified land use on the characteristics of
runoff from the subcatchment are illustrated in Figure 7a. The
ct of the increased imperviousness on the outflow hydrograph is
rarnatic, producing a seven-fold increase in peak runoff, The increased
153
-------
OUTFALL
LAND Use LEGEND
SINGLE FAMILY RESIDENCE
COMMERCIAL
INDUSTRIAL
OPEN OR PARK
FIGURE 6
Watershed Sub catchments and Drainage System
Selby Street Drainage System
154
-------
10:00
1400
1100 I2:OO
CLOCK TIME
IS:00
1400
A. Effect of Changed Land Use on
Characteristics of Watershed Runoff,
Selby Street
zooo
1*00
ALTERNATIVE LAND USE
Mutt** ftnltmttal
STORM OF 6 NOVEMBER.
I96S
11-00
1200 IJ'OO
CLOCK TIME
B. Effect of Changed Land Use on
Characteristics of Combined Sewer
Overflow, Selby Street
FIGURE 7
Effects of Changed Land Use on Runoff Characteristics
From Selby Street Watersheds
155
-------
runoff rate scours additional surface pollutants off the watershed resulting
in increased pollutant concentrations and mass emission rates. Notice
that the mass emission rates for BOD and suspended solids decrease
after the peak of the storm and remain approximately constant after 12:15.
The concentrations of these pollutants in the watershed outflow, however,
increase. This behavior is due to the influence of the dry weather flow
and is explained as follows. By the time the surface runoff peaks, most
of the surface pollutants have been washed off the watershed. Surface
runoff after that time is rather clean and dilutes the dry weather sanitary
flow coming off the subcatchment. Then as the surface runoff subsides
further, the dilution effect disappears and the subcatchment outflow con-
centrations approach those of the dry weather flow.
The effect of the changed land use on the characteristics of
the combined sewer overflow at the Selby Street outfall is illustrated in
Figure 7b. As expected, the overflow discharge is increased as a result
of the added runoff. The spike at the peak of the discharge curve is an
interesting phenomena that results from the rapid peaking of the runoff
from the multiple residential area. Recall that the imperviousness of this
area is 1.7 times (75 percent versus 42 percent on the average) that of
the remainder of the watershed; hence, the outflow from this subcatchment
responds much more rapidly to changes in rainfall rates than to the
other watersheds. The spike in the discharge overflow curve is caused
by the rapid response of the multiple residential area to the rainfall
increase at 11:30.
The indicated effect of the changed land use on the mass
emission rates of BOD and suspended solids is expected. Notice that
the mass emission rates for both of these pollutants reflect the effect of
the surge in the overflow discharge. It is interesting to note that the
BOD concentration in the overflow is essentially unchanged for the two
cases even though the BOD load on the system is increased, 'as illustrated
by the mass emission rate curve. This would indicate that the BOD con-
centration in the watershed outflow is not dramatically different for the
case of the park and the multiple residential land use alternatives.
Inspection of Figure 7b shows this to be true during the high runoff
period,
The suspended solids concentrations for the multiple resi-
dential land use case are higher than for the park, and the peak concen-
tration is lagged. The increased concentration is due to the increased
wash off of surface pollutants as the result of the increased runoff. The
reason for the lag in the peak concentration is not apparent.
CONCEPTS OF POLLUTION CONTROL OF URBAN RUNOFF
We will not deal with this subject at any length in this lecture
since it will be dealt with in later lectures. However, some conceptual
ideas should come to mind on the basis of the material presented here.
To start with, we have seen that the major contribution to urban runoff
156
-------
pollution comes from the land surface itself. Hence, control of the
pollution problem should start here. Many cities have begun anti-litter
campaigns in commercial areas and have installed litter baskets at regular
intervals at the curbside. These projects have been fairly successful
and one need only look at the amount of trash in the containers to realize
how much litter is being kept off the street as a result of these programs.
We have seen that dust and dirt is the principal component of
street litter. Evaluation of pollution control alternatives should surely
include a look at more frequent street sweeping, but not street flushing
as is practiced in many cities since this simply places the pollution load
farther down the system.
We have also seen that the principal source of runoff and
pollution is from impervious areas, Figure 3 is a vivid example. One
possible way to reduce runoff and pollution loads is to drain as much
impervious area as possible to pervious areas. House gutters for example,
should drain to lawns rather than the street or storm sewers, Parking
lots could be made to drain into a gravel bed underneath the pavement
before draining to the storm sewers system, This would provide
temporary storage and a chance for the runoff to infiltrate (due regard
must be taken in these cases, however, for possible contamination of
groundwaters),
Another alternative is to collect stormwater in storage
facilities and to treat it between storms. The proper combination of
treatment and storage combined with some of the ideas outlined above
would result in a very effective urban runoff management program from
the standpoint of both quantity and quality,
157
-------
REFERENCES
1. Field, Richard and Edmund J, Struzeski, Jr, , "Management and
Control of Combined Sewer Overflows," Journal WPCF, 44, No. 7,
pp. 1393-1415, 1972.
2. American Public Works Association, Problems of Combined Sewer
Facilities and Overflows, Water Pollution Control Research Series,
Federal Water Pollution Control Administration, Report No.
WP-20-11, 1967, Section IX, pp. 74-86.
3. American Public Works Association, Water Pollution Aspects of
Urban Runoff, Water Pollution Control Research Series, Federal
Water Pollution Control Administration, Report No. WP-20-15,
January, 1969.
4. Roesner, L. A, , D.F. Kibler, and J.R. Monser, "Use of Storm
Drainage Models in Urban Planning," Presented at the AWRA Sym-
posium on Watersheds in Transition, Colorado State University, Fort
Collins, Colorado, June 1972.
5. AVCO, Economic Systems Corporation, Storm Water Pollution from
Urban Land Activity, Water Pollution Control Research Series,
Federal Water Quality Administration, Report NO. 11034 FKL 07/70.
6. Sartor, James D., and Gail B. Boyd, Water Pollution Aspects of
Street Surface Contaminants, Environmental Protection Technology
Series, Environmental Protection Agency, Report No. EPA-R2-
72-081, November 1972.
7. Metcalf & Eddy, Inc. , University of Florida and Water Resources
Engineers, Inc., Storm Water Management Model, Water Pollution
Control Research Series, Environmental Protection Agency, Report
No. 11024DOCO7/71, Volume I, Final Report, July 1971.
158
-------
IMPACT OF STORMWATER RUNOFF ON
RECEIVING WATER" QUALITY
BY
Larry A. Roesner*
INTRODUCTION
At the present time, every city in the United States ultimately
disposes of most of its stormwater runoff in a natural surface watercourse.
The impact of these discharges on the quality of the receiving waters
depends principally on three items:
1. The quantity and quality of the stormwater discharge(s).
2. The type of waterbody into which the discharge occurs.
3. Location of the outfall(s).
In the following section we will show the pollutional potential of stormwater
and present some examples of its impact on different receiving waters.
The last section contains general discussion on considerations involved in
assessing the impact of stormwater discharges on receiving water quality.
POLLUTON POTENTIAL OF STORMWATER
Nearly every receiving waterbody has a set of water quality
standards specified for it. These standards have generally been set on
the basis of the "natural" quality of the water plus the beneficial uses
identified for it. Table 1 lists water quality standards for three beneficial
uses: drinking water supply, recreational use, and propagation of aquatic
life. Comparison of these criteria show that the water quality standards
vary significantly for different uses.
Some idea of the pollutional potential of stormwater runoff can
be gained by examining Tables 2 and 3 which show measured concentrations
of stormwater overflows in San Francisco, California and in Portland,
Oregon, respectively. Note that in both cases the overflow qualities
are for combined sewer overflows. Where applicable, water quality
standards from Table 1 are also shown. It is evident from these data
that a large pollution potential exists for untreated stormwater overflows
with respect to such pollutants as suspended solids, COD, BOD, nitrogen
phosphorus, i.e. our standard pollutants. What is not shown, however
are the metals which are incorporated in the runoff and which pose a
potential for toxicity effects to aquatic life in receiving waters. Table 4
shows recent combined sewerage data collected in Seattle, Washington.
Here it can be seen that most of the metals concentrations approach the
limit of the standards or exceed them. Iron and lead in particular both
exceed the limit of the standards by nearly an order of magnitude.
*
it
Principal Engineer, Water Resources Engineers, Waltfut Creek, California.
159
-------
TABLE 1
Water Quality Criteria
fn ••
.«•<*•«
"TEE-** fJ^kTe**
4»*llty *MU» tiatta,* BE/llter Coneeetretiona,
•(/liter
AH (d*c*Zf*Bt) O.S
irionja 0*01 O.09
tarii" 5*«
Cadaiiin] o.in
Careen chloroform extra**
(exotic oraenic chaertiila) 0.1
QOerlde ISO.
ChroadUB O.OS
Ceepor J.O
Cyanide 0.01 0.02
nuerlde »•' *•*
IfOB. plW Me.t|eVMC*t 0*)
lion ^"^
»ltt»t« *»•
Ikeao!* 0.001
MMU. 0.01
••lfa.ce ISO*
total MaMlMd felU* {») SOO.
ZUc S.
*Ce*ceBtretteu In weter abeuld not be U ewee* ef then llndU. whe>
•or* tultable aupvltea eem ke Bade available. ___j
rejection ef ripely.
CM)
HUM twiuiT IM a*ytne un
^hreeheld eneeatntlem*
Mtemuecte* fretkweter laltveter
Ibtel dleeelved eelld* (IM). •I/liter lOOOt
Ileetrlcel. conductivity, |uhe*/eej 9 23VC lOOOt
Huimm for .alamo Id tick 23 &
DtMolvod oxy§*» (D.O.), "Iniw* »»/lit«r 5.0» 9.0>
riOettetbl« Oil »B-1 I'**", *«/! t«« 0 W
Ae-ettot*
-------
TABLE 2
Treatment of Combined Sewer Overflows, San Francisco (4)
Wet Weather Monitoring Results
Laguna Street Outfall* 15 March 1967
SAMPLE
TIME
FLOW, cfs
COLIFORM MPK - Con f,/ Fecal, 10*/«1
CONDUCTIVITY, uraho/cm
ALKALINITY. Bg/1 as CaCOj
SUSPENDED SOLIDS, Bg/1
VOLATILE SUSPENDED SOLIDS, fflg/1
CREASE, mg/1
COD, mg/1
BOD, mg/1
FLOATABLE PARTICIPATES, mg/1
SETTLEABLE SOLIDS 9 30 Min., 01/1
TOTAL KJELDAHL NITROGEN, mg/1 -K
AMMONIA NITROGEN, tag/1 -N
TOTAL PHOSPHATE, mg/1 P04
SULFATE, ng/1
CHLORIDE, Bg/1
SODIUM, mg/1
POTASSIUM, ng/1
CALCIUM, mg/1
MAGNESIUM, mg/1
1'.
2020
7.0
70/13
338
82
304
234
63.4
458
252
2.3
40
19.95
10.50
3.20
29
32
29.5
6.70
11.2
4.9
2
^030
12.5
70/2.3
220
43.4
442
264
'18.7
425
169
1.5
35
15.75
4.20
2.12
18
21
18.5
3.65
10.4
2.9
3
2045
11.4
24/6.2
200.8
27.6
194
149
28.4
210
97
1.1
13
13.30
2.80
1.47
15
21
19.0
2.90
8.0
1.9
4
2100
7.2
2.3/2.3
178
35.5
130
86
16.2
198
104
1.0
7
s.o:
3.5(
1.4
18
14
L4.5
3.0C
8.4
3.2
5
.211.5
4.6
6.2/6.2
206
46.4
104
81
18-. 5
169
92
2.2
12
8.75
4.20
1.86
17
19.5
18.0
3.35
8.0
3.2
6
2110
2.0
6.2/0.46
219.8
51.3
73
56
18.3
251
108
0.4
10
10.85
4.55
1.40
19
18
16.0
3.85
8.8
3.2
7
44
5/2.3
158.3
29.6
187
92
20.5
210
51
0.8
5
7.00
1.40
1.17
17
14
11.0
2.75
7.2
3.4
8
2230
108
70/0.6
76.8
16.8
237
119
17.9
165
41
1.1
15
3.85
2.45
1.13
10
5.5
6.0
1.15
4.8
2.4
Standard from
Table 1
.OT
3000. !
20. 2
.2 '
0. 2»J
250. l
CTk
*Land Use is Multiple Family Residential
1Chemical standards for Drinking Water
2Water Quality for RecreatTonal Use
3Water Quality for Aquatic Life
-------
o>
ro
TABLE 3
Summary of Characterization of Combined Sewage
Sullivan Gulch Pump Station (5) Portland, Oregon*
February - April 1969
CHARACTERISTIC
PH
TEMfERATUHE.'F
DISSOLVED OXYGEN. MG/L
SETTLEABLE SOLIDS. ML/L
TOTAL SUSPENDED SOLIDS. MG/L
VOLATILE SUSPENDED SOLIDS. MG/L
* VOLATILE SUSPENDED SOLIDS
B.O.D.. MG/L
C.O.O. MG/L
B.O.D./C.O.D.
AMMONIA NITROGEN. MG/L
ORGANIC NITROGEN. MG/L
TOTAL NITROGEN. MG/L
NUMBER OF
OBSERVATIONS
26
26
IS
26
28
28
28
14
24
14
7
7
7
MEAN
5.0
49.7
8JD
3.1
146
90
67
106
190
Ml
6.1
8.2
1X3
STANDARD
DEVIATION
+ A
+ 6,6
* 13
* J.O
+ »
+ 26
+ 17
+ 38
+ to
» M
* 1.4
+ 3.1
* 4.3
MINIMUM
4i
34.0
3.7
1^
70
57
36
67
138
J6
3.7
6.10
».S
MAXIMUM
6.0
56.0
104
SJ)
325
186
83
166
324
.64
7A
14.0
21.0
Standard from
Table 1
5.*
20.*
*Land use is Single Family Residential
Chemical standards for Drinking Water
2Water quality for Recreational Use
3Vtater quality for Aquatic Life
-------
TABLE 4
Urban Runoff Characteristics
Viewridge Two Area (6)
Seattle, Washington**
Parameter
Temp. C°
pH
Cond . umho/cra
Turbidity, JTU
DO, ng/1
BOD, ng/1
COD, ng/1
Hexane Ext., mg/1
Chlor ide , mg/1
Sulfate, ng/1
Organic N, mg/1
Ammonia N, mg/1
Nitrite N, mg/1
Nitrate N. mg/1
-Jlydrolyzable P, mg/1
Ortho P, mg/1
Copper, mg/1
Lead, mg/1
Iron, mg/1
Mercury, mg/1
Chromium, my/1
Cadmium, ng/1
7.inc, mg/1
Sett. Solids mg/1
Susp. Sol ids;, m
-------
STORMWATER POLLUTION OF RECEIVING WATERS
The actual degree of pollution that occurs in a receiving water
due to a stormwater discharge depends to a large extent upon the type
of receiving water and the location of the outfall. Consider, for example,
the La gun a Street confined sewer outfall in San Francisco. Figure 1
shows the drainage area (primarily middle class multiple family housing)
plus an enlarged plan view of the municipal marina into which the Laguna
Street outfall discharges. The marina waters are designated as recreational.
Figure 2 shows confirmed coliform MPN's measured within the harbor and
outside the harbor during the period 11-18 March 1967. It is readily
apparent from the figure that there is significant pollution both inside and
outside the harbor as a result of the Laguna Street stormwater overflows.
To illustrate the effect of outfall location on the dispersion
characteristics of the receiving waters, consider Figure 3 which identifies
the location of the Baker Street combined sewer outfall in San Francisco.
This outfall although located less than one mile east of the Laguna Street
outfall, produces an entirely different receiving water response because
the outfall discharges to the open shoreline rather than an enclosed marina.
Thus, concentrations around the outfall are significantly affected by the
strong tidal currents through the Golden Gate. For the storm of 19-20
December 1969, Figure 4 shows the rainfall hyetograph, and the hydrograph
and pollutographs at the Baker Street outfall. Figure 5 shows the mathe-
matically simulated boundary of the pollutant field at 3, 4, and 5 hours
after the start of rainfall. The effect of tidal influence on movement of
the pollutant field is clearly evident in the figure. The effect of mixing
and dispersion can be seen in Figure 6 which shows time histories of
suspended solids in the vicinity of the outfall, Comparison of Figures 2
and 6 show that, on the average, Baker Street discharges are diluted two
to five times more than those in the enclosed Laguna Street Marina,
As a last example, we consider a discharge of a combined
overflow to a stream, Figure 7 shows the Kingman Lake drainage area
in Washington, D. C. The Anacostia-Potomac Rivers were modeled by
the 47 node system shown in the figure. An appropriate tide was imposed
at node 32, freshwater inflows were imposed at nodes 1 and 47, and the
Kingman Lake basin discharge for the storm of 22 July 1969 (see Table 5)
was imposed at node 15. The computed oxygen balance in the receiving
water system for 5 and 25 hours following the storm of 22 July is shown
in Figure 8. The maximum deficit occurred at node 17, 3,000 feet
from the point of release and 25 hours after the start of the storm. The
oxygen deficit was continuing to increase and move seaward at the end of
the simulation. Similarly, the computed travel of suspended solids is
shown in Figure 8, also.
It should be noted that these changes were the direct result
of one outfall discharging for one very large storm. The residual effects
from earlier storms and pollution releases from coincident discharges
would have to be evaluated to determine the full impact on the river
system.
164
-------
SAN FRANCISCO BAY
cn
DRY WEATHER
FLOW
MONITORINa
STATION
TRACER —
INJECTION
SAN FRANCISCO BAY
DRY WEATHER MONITORING
• 000 1000
S4UPUHS
STATION
LAGUNA STREET OUTFALL DRY WEATHER MONITORING
LAGUNA STREET
RECEIVING WATER SAMPLING STATIONS
FIGURE 1 Laguna Street Drainage Area and Receiving Water Marina
(Reproduced from Reference 4)
-------
.
.
r
* DATA FROM WEATHER BUREAU, F.O.B.. S/
.
BATA JNCRAOEO OVER
4 HOUR MTERVALS
mF
j
r1
y
~n, , „
kN FRANCISCO
,
Jl, , . n
£
**
ftM £
i
S
ao4 z
d
£
. i
HMM BMM a MM MMI
DATE
MMM ffMM Ml
1
&
Q.
5
• STATION
• STATION 2 !
A STATWN 3 ^ £££
A STATION 4 ' HarD-r
SmTKW
L
r\
DATA FROM WEATHER BUREAU, F.O. a, SAN FRANCISCO
DATA AVCRAeEO OVER
4 HOUR MTERWALS
-n «
- OM
SL ,
n
_l
2
<
oc
HMHMHMNMNMNMN
HMM It MM a MM MMM B MM !• MM IT MM
DATE
• MM
a MM
MMM
MMM
N
IT MM
II M N
MMM ItMAR
e STATION «| _
. STATOW 61 °Ulif
* • STATION 7
DATE
RAINFALL INTENSITY AND CONFIRMED COLIFORM MPI\fc IN
LAGUNA OUTFALL RECEIVING WATERS-MARCH 1967
RAINFALL INTENSITY AND CONFIRMED COLIFORM MPN^ IN
LACUNA OUTFALL RECEIVING WATERS-MARCH 1967
FIGURE 2 Coliform Counts Measured in Laguna Street Marina
(Reproduced from Reference 4)
-------
cr>
SAN FRANCISCO »AV
SAN FRANCISCO
FIGURE 3 Baker Street Receiving Water Grid System
(Reproduced from Reference 7)
-------
Ktll DO I
U>U IMIlt |W« MMCItCII U tVt-MIt Hilt*
!• / M
I.MO-
t.tqo
t.lM
»••«
•••'•""""I™"— •••I™—— ••t——~-|.~««»»»»«»»»»»M»«»»«
»•» »».« H.« »».! K.t M.I ll.t
I.I
TIM In HOUCt
i«* «.i ».» *»"
TIM I* NOUM
KIKC'Ct UMW t • •
U 1UCUI> l»r|H
•Mil MMC? .( U.I |«!«"
11*1 ID ICWI
I.I t.«
«.l t.«
III* In M*t*t
»•••*'
II.*
FIGURE 4
Baker Street Combined Sewer Overflow Results
Storm of December 19-20, 1969
(Reproduced from Reference 7)
168
-------
SAN FRANCISCO BAY
BAKER STREET
COMBINED SEWER
SAN FRANCISCO
30CO 4000
FIGURE 5
Baker Street Receiving Water BOD Movement
(Reproduced from Reference 1)
169
-------
10
M
CO
BAKER STREET
RECEIVING WATER
DEC: 19,1969
STORM
NODE 40 (OVER OUTFALL)
NODE 34
TIME
FIGURE 6
Concentration History at Three Junction Points
(Reproduced from Reference 7)
170
-------
APPROXIMATE LIMITS
OF DRAINAGE BASIN
WASHINGTON D.C.
SCALJE W MILES
FIGURE 7
Kingman Lake Receiving Water System
(Reproduced from Reference 7)
171
-------
TABLE 5
Computed Time Variation of Overflow Quality
(Reproduced from Reference 7)
Time from Start
of Overflow, min
0
30
60
90
120
150
180
210
240
270
300
330
360
390
420
450
480
Storm of July
BOD, mg/L
220
135
79
54
48
40
14
4
6
6
10
18
27
37
42
63
90
22, 1969
SS, mg/L
261
425
921
755
798
698
252
25
9
7
13
20
28
38
44
60
80
Storm of Aug.
BOD, mg/L
245
239
218
150
87
70
60
54
55
47
43
47
58
67
72
96
125
20, 1969
SS, mg/L
321
327
289
277
506
675
779
576
518
450
410
325
241
201
179
175
180
172
-------
9.01-
zoo i—
ao
§
TO
ISO -
23 HOURS
CO
6.0
1
I
I
20 O IS 5
NODE POINTS ALONG ANAOOSTIA
100 -
•0» MINTS ALOM* AMACOITU
KZHGHAN LAKE RECEIVING HATER
DISSOLVED OXYGEN PROFILE
KINGMAN LAKE RECEIVING WATER SS PROFILE
FIGURE 8
Kingman JLake Receiving Water Response to Storm of 22 July 19^9
(Reproduced from Reference 7)
-------
ENVIRONMENTAL ASSESSMENT CONSIDERATIONS
Receiving waters serve multiple uses,; thus procedures
established for the planning of multiple purpose projects apply. We cannot
simply chlorinate wastewater to maintain low coliform counts on a beach
and, in turn, introduce a level of toxicity that will kill fish. Nor can we
channelize the river for navigation or flood control without giving due
consideration to the habitat of fish and fish-food organisms,
Figure 9 shows a conceptual diagram of an ecologic-water quality
model. The interactions shown by the arrows are ecologic processes that
transform chemicals such as carbon, nitrogen and phosphorus between
their abiotic state and the successions of organic biomass. It is apparent
from this figure that if the water quality is changed as the result of a
waste stormwater discharge, there will be a resultant shift in the ecologic
balance of the system. The severity and duration of the sh.ift can be
directly related to the severity and duration of the discharge.
The bases for environmental assessment are computations and
value judgments. Computations are performed according to established
knowledge or theory. Value judgment is then applied to extend the assess-
ment beyond the state of the art of current computational technology.
Computation analysis should be carried out to depict effects
with time and spatial detail. Along the time axis, there are long-term and
short-term effects of urban runoff. In the spatial scale, the effect may
be regional, i.e., the effect maybe detected elsewhere downstream. Thus,
stormwater runoff may create a transient increase in suspended solids
and bacterial counts. Bacteria may die off rapidly but suspended solids
may settle,exert a long-term effect on the ecosystem.
The extent to which value judgment must be applied to the
assessment of ecologic impact depends on the degree of sophistication used
in the computational analysis. The version of RECEIV documented in the
EPA STORMWATER MANAGEMENT MODEL (8) simulates quality effects
of BOD, dissolved oxygen, and suspended solids only. All other ecologic-
water quality impacts must be inferred by value judgments. This model
has recently been expanded, however, in the EPA's River Basins Modeling
Projects program to include nine water quality parameters plus algae (9).
Use of this model requires value judgments for evaluation of impact on
higher trophic levels plus benthos.
Perhaps the most comprehensive computational tool presently
available for ecologic-water quality assessment is the so-called Ecologic
Model (10), which describes the interrelationship between some 23 water
quality constituents and four biotic trophic levels. It too, however, requires
the use of value judgments for interpretation of computed results and impli-
cation of these results on factors and processes that are not adequately
accounted for in the models.
174
-------
Man-Induced
Waste Loads
Natural Inputs
Water
Quality
\>Nutrients\
Fish
\
Nutnehts
Zooplankton
Benthic Animal <}
FIGURE 9
Conceptual Diagram of Ecologic-Water Quality System
175
-------
REFERENCES
1. Public Health Service Drinking Water Standards, U.S. Public Health
Service, revised 1961.
2. McGauhey, P.H., Engineering Management of Water Quality, McGraw-
Hill Book Company, New York, 19687
3. McKee J,E. and H. W. Wolf, Water Quality Criteria, 2nd ed, , Report
to California State Water Quality Control Board, SWPCB Publication
3A, 1963.
4. Engineering Science, Inc. , Characterization and Treatment of Combined
Sewer Overflows, Final report submitted by Department of Public Works,
City and County of San Francisco to FWPCA on Grant WPD-112-01-66,
November 1967.
5. Marske, D.M., "Rotary Vibratory Fine Screening of Combined Sewer
Overflows," Section 2 of Combined Sewer Overflow Abatement Technology,
Water Pollution Control .Research Series 11024-16/70, June 1970.
6. Farris, G. , R. G, Swartz, and N. R. Wells, "Water Quality and Quantity
Monitoring for the Ribco Urban Runoff and Basin Drainage Study, " Draft
Report, Municipality of Metropolitan Seattle, Seattle, Washington,
April 1974.
7. "Storm Water Management Model, Vol. II, Verification and Testing,"
11024 DOC 08/71, Metcalf and Eddy, Inc., Univ. of Florida, Gainesville,
and Water Resources Engineers, Inc. , Report for EPA, August 1971.
8. "Storm Water Management Model, Vol. I, Final Report, " 11024 DOC
07/71, Metcalf and Eddy, Inc., Univ. of Florida, Gainesville, and
Water Resources Engineers, Inc., Report for EPA, July 1971.
9. Water Resources Engineers, Computer Program Documentation for
the Dynamic Estuary Model, An Intermediate Technical Report,
EPA Contract NO. 68-01-1800, May 1974.
10. Chen, Carl W. and G. T. Orlob, "Ecologic Simulation for Aquatic
Environments," Final Report to the Office of Water Resources Research,
Water Resources Engineers, December 1972,
176
-------
INTRODUCTION TO URBAN STORMWATER RUNOFF MODELS
By
Robert F. Shubinski*
DEFINITION OF THE URBAN DRAINAGE SYSTEM
The succeeding sections will be concerned with an environment
that is predominantly urban. That is, it will be characterized by a high
proportion of impervious (or nearly impervious) surfaces. Moreover, it is
a system that will include man-made impervious pathways for guiding the
flow of water over the surface (curbs, gutters, lined channels, paved
parking areas, streets, etc.) and underground (storm, wastewater and
combined sewers). The system includes all appurtenances that guide, control,
or otherwise modify either the quantity, rate of flow or quality of runoff
from urban drainage such as catch basins, storage basins, inlets, manholes,
sediment traps, wlers and outfall structures.
Figure 1 illustrates a somewhat simplified urban drainage system.
In this case, we view the system as an assemblage of subsystems dealing with
a. surface runoff
b. transport of flow and quality, and
c. the receiving water.
SURFACE RUNOFF SUBSYSTEM
The Surface Runoff Subsystem is illustrated for our example in
Figure 2 which depicts the drainage area tributary to a sewer inlet as a
system of surface elements (the rectangles), gutters (the dotted line) and
drainage ditches (the dashed line). Each drainage subarea is characterized
by its area, a degree of imperviousness, its slope, and certain coefficients
that relate to its production of quality constituents that may be transported
to the inlet by overland flow.
Input to the subsystem is comprised of rainfall that may be described
in terms of an intensity-time graph derived from direct measurements in the
Watershed. The inset in the upper left of Figure 2 illustrates a typical
rainfall hyetograph.
Within the subsystem a certain mass of a quality constituent
(pollutant) may exist at the outset of the storm and be delivered by the
^low at mass rates and concentrations that may depend on the nature of the
storm, the character of the surface, and the sources of the pollutant (upper
rl8ht insert, Figure 2).
Principal Engineer, Water Resources Engineers, Springfield, Virginia.
177
-------
SURFACE RUNOFF SUBSYSTEM
TRANSPORT
SUBSYSTEM
.RECEIVING WATER . ->__
SUBSYSTEM
FIGURE 1. THE URBAN DRAINAGE SUBSYSTEM
178
-------
precip.
intensity
time
OVERLAND
FLOW
SYSTEM
I
o-
O-
INPUT n
pollutant
load
b inlet
time
OUTPUT
time
FIGURE 2. SURFACE RUNOFF SUBSYSTEM
time
179
-------
The overland flow process modifies the rainfall hyetograph by
infiltration, surface retention and transient storage, so that at the inlet
one observes a much modified Inlet hydrograph. a temporal description of
inlet flow (lower left insert, Figure 2). In addition, the combined flow
and quality processes produce an inlet pollutograph. a time-concentration
graph of a particular pollutant as it leaves the Surface Runoff Subsystem
and enters the waste water conveyance system (lower right insert, Figure 2).
These two graphs, one of flow and the other of quality, comprise the output
of the Surface Runoff Subsystem and are input to the Transport Subsystem.
TRANSPORT SUBSYSTEM
The Transport Subsystem is comprised of the physical works for
conveying storm waters and their associated pollutant loads from all of the
inlets in the system through a network of underground conduits to a point
(or points) of disposal. Enroute, flow and quality are both modified by
accretions to the system from other tributary areas and/or point sources of
pollution. In addition, flows and pollutant concentrations are attenuated
in passing through the system, the degree of modification depending on such
factors as system storage> "off channel" storage, phase relationships of
inflow hyetographs and pollutographs and certain hydraulic properties of the
system. The two lower inserts in Figure 3 illustrate a typical set of
outputs from the Transport Subsystem, a hydrograph and a pollutograph, that
in turn become inputs to the receiving water.
RECEIVING WATER SUBSYSTEM
The Receiving Water Subsystem may be a stream, a lake, an estuary
or a coast. Discharge into an estuary will be used for illustration.
The impact of the discharge on the estuary will probably be
assessed in terms of the concentration of a particular quality constituent:
its distribution in space, its persistence in time, and its frequency of
exceedance of a certain critical level. For a given hydrological event,
the system may bt observed synoptically (at the same instant in time) or
temporally (at the same point in space).* One gives the distribution in
space, the other the persistence in time. From the standpoint of quality
management both viewpoints are usually required for each hydrological event.
To obtain frequency of exceedance of a critical level the impact on the
receiving water must be observed a "statistically significant" number of
times. Figure 4 illustrates typical responses ("impacts") for an estuary.
*These are sometimes referred to as the Lagranglan and Eulerian viewpoints,
respectively.
180
-------
INPUT
time
to storage
by pass
OUTPUT
time
time time
FIGURE 3. TRANSPORT SUBSYSTEM
181
-------
INPUT
time
RECEIVING WATER
SUBSYSTEM
time
ISO-CONCENTRATION
LINES
OUTPUT
critical
time
FIGURE A. RECEIVING WATER SUBSYSTEM
182
-------
HYDROLOGIC EFFECTS OF URBANIZATION
The term "urbanization" refers generally to a condition in which
a natural watershed has been developed for residential, commercial, and
industrial purposes. The principal hydrological factor associated with
urbanization is an increase in imperviousness of the watershed surface with
attendant reductions in infiltration and other abstractions from storm
rainfall. A summary of potential hydrological effects of urbanization is
presented in Table 1. In addition to major changes in hydrology, there
is evidence indicating that certain changes in the micro-climate of an
urban area can be expected, Table 2.
While urbanization always increases the imperviousness of a natural
watershed, the hydrological significance of this change is not always clearly
evident in contrast to the postulated effects.in Table 3. For example,
increases in total runoff and peak flow may be offset by the use of building
practices which provide for detention near the site where it occurs of
rainfall excess on flat roofs or in small depressions or ponds. Another
factor to be considered is the manner by which collected rainfall is dis-
charged from roof tops. In the San Francisco Bay region, for example,
collected rainfall may be disposed of in four different ways, each having a
pronounced effect on the volumes and peaks of local hydrographs: discharge
through a downspout into a dry well and thence into the underlying ground-
w«ter; surface storage and eventual evaporation; discharge through a downspout
to a spatter block and thence over a lawn; or discharge through a downspout
and pipe drain into a street gutter. Other factors to be considered in
analyzing runoff from impervious areas are: the location of an impervious
area with respect to the total watershed; and the proximity and geometry of
the street gutter and underground drainage system.
CHARACTERISTICS OF THE URBAN WATERSHED
There is an extreme variability in the hydrological characteristics
°* local urban drainage systems. Figure 5 illustrates the configuration of
a 47.4-acre urban watershed located in the Northwood section of Baltimore,
which has some features in common with many other urban drainage basins. It
is approximately three-fifths residential and two-fifths commercial, with the
latter represented primarily by a large shopping center and parking lot.
Average ground slope is 3 per cent and the degree of imperviousness ranges
between 0.15 and 1.00 with an average value of 0.68. As indicated in Figure
*» the Northwood watershed is composed of 32 subbasins, each of which has
its own land use, infiltration and runoff characteristics. The underground
drainage system of Figure 5 is shown in greater detail in Figurefi.
183
-------
TABLE 1. POTENTIAL HYDROLOGIC EFFECTS OF URBANIZATION
Urbanizing Influence
Potential Hvdrologic Response
Removal of trees and vegetation
Initial construction of houses,
streets and culverts
Complete development of residential,
commercial and industrial areas
Construction of storm drains and
channel improvements
Decrease in evapotranspiration and
interception. Increase in
stream sedimentation.
Decreased infiltration and lowered
groundwater table. Increased storm
flows and decreased base flows
during dry periods.
Increased imperviousness reduces time
of runoff concentration thereby
increasing peak discharges and
compressing the time distribution of
flow. Volume of runoff and flood
damage potential greatly increased.
Local relief from flooding.
Concentration of floodwaters may
aggravate flood problems downstream.
TABLE 2. CLIMATIC EFFECTS OF
Ratio, City
Climatic Variable to Environs .
Solar radiation (insolation) in horizontal surfaces
Ultraviolet radiation, summer
Ultraviolet radiation, winter
Mean annual temperature greater in the city by 1 to 1.3°F
Annual mean relative humidity
Annual mean wind speed
Speed of extreme wind gusts
Frequency of calms
Frequency and amount of cloudiness
Frequency of fog, summer
Frequency of fog, winter
Total annual precipitation
Days with less than 2/10-inch of precipitation
0.85
0.95
0.70
-
0.94
0.75
0.85
1.15
1.10
1.30
2.00
1.10
1.10
184
-------
Rein Gog* *2
b
RAIN GAGE if 1 - WEIGHING
BUCKET TYPE
RAIN GAGE *2 - TIPPING BUCKET TYPE
•—• STORM CONDUIT
• INLET
£ MANHOLE
_ ..._ DRAINAGE AREA BOUNDARY
• —.- SUB DRAINAGE AREA BOUNDARY
(?) SUB AREA NUMBER
FIGURE 5 LAYOUT OF NORTH WOOD DRAINAGE BASIN
(2)
185
-------
SUBAREA TRIBUTARY TO
INDICATED POINT-'
435' CONDUIT, 15"
DIAMETER, 3.80% SUDPE:
435'-l5"at 3.80%
455'-15"at 5.07%
£40-10"ot
*:^:L-
298-15 at 4.0%
C*225-2r'af
< i/i r
225-18" at 3.17*
307'-2rat 1.27o
165'-15"at 3.80%
150-18" at 3.0%
165-15 at 3.457.
OUTFALL
PARSHALL
FLUME
FIGURE 6 SCHEMATIC DRAWING OF DRAINAGE
CONDUITS.(2)
186
-------
A representative rainfall hyetograph and runoff hydrograph for the
Northwood watershed is shown in Figure 7 *, Figure 7 is notable in at
least two respects. First, it illustrates the incidence of a rapid response
of urban runoff to storm rainfall after initial losses are satisfied. Second,
it demonstrates that losses from rainfall can be substantial when total
rainfall is compared with total runoff. These losses are regulated by the
particular land-use characteristics that prevailed in each Northwood sub-
basin at the time of the given storm.
FACTORS CONTROLLING URBAH RUNOFF
The dominant hydrologic factors in the rainfall-runoff process are
rainfall, infiltration, depression storage, surface detention and gutter
detention, and storage in house drains, catch basins and major sewer elements.
Losses due to evaporation and to interception by vegetation are normally
considered small during a storm period and can be neglected. The influence
of infiltration, depression storage and surface detention, and the length of
the flow plane of overland flow from a pervious area is shown conceptually
in Figure 8. Figure 8 illustrates the manner in which rainfall excess
can be computed by deducting time-varying losses due to infiltration,
depression storage and surface detention. The overland flow hydrographs
shown in Figure 8' were developed using the Izzard relationship for steady-
state conditions:
0.324
0.0007 i + c 1,
3
in which D is the surface detention depth in inches, S is the ground slope,
L is the length of overland flow in feet, q is the overland flow supply in
inches per hour, i is the intensity of rainfall in inches per hour and c is
the coefficient of roughness. Overland flow leaving the watershed surface
would next be modified by storage and translation in the street gutter and in
the street lateral before it entered as an inflow to the main storm sewer.
INFLUENCE OF LAND USE ON URBAN RUNOFF
Land use exerts a profound influence on the quantity of urban
runoff through its effects on imperviousness and surface cover, which in
turn regulate surface depression, detention and infiltration. Representative
values of impervioutness are shown in Table 3 as a function of land use
classification. Representative values of depression and detention storage
associated with various land cover conditions are shown in Table 4 Both
Tables 3 and 4 contain assumed representative values for the Denver,
Colorado area.(4) While these values of surface depression and detention
are recommended only for use in connection with the Colorado Unit Hydrograph
procedure, they are in general agreement with commonly assumed values of
1/16-inch for impervious areas and 1/4-inch for pervious areas. The influ-
ence of land use on infiltration-capacity is shown in Figure 9. The
influence of land use on watershed outflow is further described in connection
with che Selby Street example, presented below.
187
-------
o.u
£5.0
o
cr
LU
Q.
CO
Lu
04.0
LU-
LL
o
ID
fr
i
cc 30
3
X
CC
LU
QL
(f\
V f
LU
X2.0
o
i=
„
_j
^
LL
Z
£ 1.0
o
— •
—
—
^^^
—
T
|
1
n.
uo
i^h
1
I
1
i
n
L_ „ r-i r-J LnJ
^uJl .iiJT-i^
1
1
••
H
•
n
j
J
'
i
/
f
M
>
/
•i
V i
N«
10 20
TIME FROM BEGINNING OF
FIGURE 7
••
1
1
f
1
m*
.
m
'
'
/
/
1
1
1
i
PAIMTAI 1
RUNOFF
MM
••M
—
^RUNOFF PEAK
i
••
/ 1
/
t
30
I
1
t
1
I
\
\
I
t -
\
1
\
\
\
N
\
\
40 50
RAN FALL, MINUTES
' HYETOGRAPH AND HYDROGRAPH FOR STORM
8-1-65 WITH RAINFALL STARTING AT 1639-
47.4 ACRE NORTHWOOD DRAINAGE AREA<2>
188
-------
•a
c
s
I
k
I
I
(a) Vi-in. depression storage
Overland (low and
depression storage supply
StllSSx^
(6) l/2'in- depression storage
Overland (low and
depression storage supply
45 50
FIGURE 8
60 G5 70 75 80 85 90
Time, in minutes, from beginning of design storm pattern
CONCEPT OF OVERLAND FLOW ON
PERVIOUS AREAS(3)
189
-------
TABLE 3. LAND USE VERSUS PER CENT OP
Land Use
Downtown Business Area, Airport
Terminal, Shopping Centers, etc.
Residential, Dense
Residential, Normal
Residential, Large Lots
Parks, Greenbelts, etc.
Per Cent
Pervious
0 -
40 -
55 -
60 -
90 -
5
55
65
80
100
Per Cent
Impervious
95 -
45 -
35 -
20 -
0 -
100
60
45
40
10
TABLE 4. TYPICAL DEPRESSION AND DETENTION FOR VARIOUS LAND COVERS^
Depression & Reconcnended,
Land Cover Detention. Inches Inches
Impervious:
Large Paved Areas 0.05-0.15 01
Roofs, Flat 0.1 - 0.3 Q*.l
Roofs, Sloped 0.05 - 0.1 o!o5
Pervious:
Lawn Grass 0.2 - 0.5 o 3
Wooded Area and Open Fields 0.2-0.6 0*4
190
-------
o:
x
LU
^
(T
o:
RESIDENTIAL AREAS
SANDY SOIL AREAS
INDUSTRIAL AND COMMERCIAL AREAS
zftfKMi
*n vi i.i ILI i. ...I'.1. .•-'.•..',
:t
0
0
RGURE
20
40
60 80 100
TIME, MINUTES
120
PERVIOUS SURFACE INFILTRATION-CAPACITY CURVES.
140
(3)
-------
CONCEPT OF A STORM DRAINAGE MODEL
Because of the complexities introduced by variable land use and
hydrological conditions, storm discharge from an urban watershed can best
be analyzed by dividing the total basin into smaller homogeneous units for
which their individual runoff contributions can be computed. Collection
of the individual subbasin outflows and their routing through main storm
sewers permits determination of the total watershed outflow at the basin
outlet. The concept of this approach is illustrated in Figures 10, 11
and 12,
APPLICATION OF STORMWATER MODEL TO LAND-USE PLANNING
The complete stonnwater model presented in Figure 3-8 has been
applied to the Selby Street watershed in San Francisco in order to determine
the influence on stormwater quality and quantity that might be expected with
a change in land use in the Selby Street combined sewer system. The
particular land use change investigated was the hypothetical conversion of
a park to a multiple-family residential area. The hydrological significance
of this hypothetical change would be to increase the imperviousness of the
park catchment from 11 per cent to 75 per cent and to generate a new dry-
weather wastewater flow from the area. The location of the Selby Street
drainage area and some of the results of the computer simulation are
presented in Figures 13, 14 and 15,
PONDING AND STORAGE AS A MEANS FOR CONTROLLING URBAN RUNOFF
Urban runoff management problems in the Selby Street watershed
center on the inability of the local treatment plant to handle combined
wastewater-stormwater flows, thereby contributing to pollution in San
Francisco Bay through combined sewer ovexflows. From the foregoing outline
of factors controlling urban runoff, it is clear that storage reservoirs and
detention ponds which increase runoff travel time in selected reaches of a
watershed can have important advantages in reducing peak flows during
storiuwater runoff. Ponding can be incorporated effectively in urban storm
drainage design through proper use of large flat-top roofs, plazas and
other natural on-site depressions. Swales, natural channels and grassed
waterways are becoming increasingly used as means for dispersing excess
stormwatar. In overflow areas normally reserved for recreational use,
urban ponds and lakes also are being designed with provisions for temporary
storage of storra runoff.
Other storage concepts include the use of in-line storage in
existing main sewers. As a means for vastly reducing the incidence of
combined sewer overflows, the City of San Francisco has developed a master
plan which includes provision for underground detention reservoirs in a
number of its watersheds„'"' Detention basin storage is coupled with a
cross-system storage and transmission tunnel, and in order to take full
advantages of all system capacities a total system automatic control
capability will be developed.
192
-------
iO
s,
FIGURE 10 CASCADE OF n PLANES DISCHARGING INTO THE J& CHANNEL SECTION
-------
INFLOW
Q,-f,(t)
RAINFALL INTENSITY
= f?(t)
I=I(t)
INFILTRATION
INFILTRATION:
.-k,t
Q
OUTFLOW
FLOW:
. -1-49
--
STORAGE:
M
At
FIGURE II ELEMENTARY FLOW CALCULATIONS
WATERSHED MODEL
194
-------
RAINFALL
*
J
PX
— .'., —
OVERLAND FLOW
ROUTING
QUAL^Y OF
OVERLAMD FLOW
SURFACE RUNOFF
MODULE
(OVERLAND FLOW)
I Inlet x I Inlet x
1A\* I pIZLt
FLOW ROUTING
THROUGH DRAINAGE
SYSTEM
Q
Outfall y
QUALITY ROUTING
THROUGH DRAINAGE
SYSTEM
P
Outfall y
\f
FLOW ROUTING
THROUGH
RECEIVING WATERS
QUALITY ROUTING
THROUGH
RECEIVING WATERS
Node z
TRANSPORT MODULE
(DRAINAGE SYSTEM
ROUTING)
t
RECEIVING WATER
MODULE
(RECEIVING WATER
RESPONSE)
FIGURE 12 . OVERVIEW OF THE STORM DRAINAGE MODEL
195
-------
SELBY
STREET
DRAINAGE
SYSTEM
v*
t»
a
t»
a
FIGURE 13 LOCATION OF SELBY STREET DRAINAGE
SYSTEM, SAN FRANCISCO(5)
196
-------
ourr«u.
SANITARY
f INTERCEPTOR
N[
LAUD USE LEGEND
FAMILY RESIDENCE
COMMERCIAL
INDUSTRIAL
REZQNE TO MULTIPLE FAMILY RESIDENTIAL
FIGURE 14
HYPOTHETICAL LAND USE CHANGE,
SELBY STREET DRAINAGE SYSTEM15*
197
-------
w:00
10,00
KXX)
ICCO 12:00
CLOCK TIME
FIGURE 15 EFFECT OF CHANGED LAND US=: ON
CHARACTERISTICS OF SUBCATCHWENT
RUNOFF, SELBY STREET^*
198
-------
REFERENCES
1. Lbwry, William P., "The Climate of Cities," Scientific American. Vol.
217, No. 2, August, 1967.
2. Tucker, L. S., "Northwood Gaging Installation, Baltimore — Instrumentation
and Data," ASCE Urban Water Resources Research Program, Technical
Memorandum No. 1, ASCE, New York, N.Y., August 1, 1968.
3. Design and Construction of Sanitary and Storm Sewers. ASCE Manuals of
Engineering Practice No. 37, First Edition, 1960.
4. Wright-McLaughlin Engineers, "Urban Storm Drainage Criteria Manual,"
Denver Regional Council of Governments, Denver, Colorado, 1969.
5. Roesner, Larry A., David F. Kibler and John R, Monser, "Use of Storm
Drainage Models in Urban Planning,11 pp. 400-405 in Proceedings of a
Symposium on Watersheds in Transition, AWRA Proceedings. Series No. 14,
1972.
6« Department of Public Works, "San Francisco Master Plan for Waste Water
Management, Supplement I," City and County of San Francisco, May 15,
1973.
199
-------
SIMPLIFIED METHODS OF COMPUTING THE QUANTITY OF URBAN RUNOFF
By
Robert P. Shubinski*
FLOW-FREQUENCY ANALYSIS
An example of the use of statistical methods in analyzing
urban runoff has been presented by Rantz(2) for the San Francisco Bay
region and is discussed in part here. Regional flood-frequency relations
were developed for 40 non-urban watersheds in the Bay area by means of
multiple regression. The resulting regression equations are of the form
QT = K Aa Pb, where QT is the peak discharge in cfs for a recurrence
interval of T years; A is drainage area in square miles; P is mean annual
precipitation in inches; K, a and b are regression coefficients. The
results of the regional multiple regression analysis for the San Francisco
Bay area are shown in Table 1. It is noted that other independent variables
related to climatology and topography were found to have no statistical
significance in the regression analysis and that drainage area and mean
annual precipitation were dominant factors.
In order to illustrate the use of the flood-frequency relations,
in Table 1, the following hypothetical problem is posed:
Given: Drainage area = 5.0 square miles
Mean annual precipitation = 40 inches
Degree of development = 70 percent of basin
urbanized and 70 percent of channels sewered
or lined.
Required: Peak discharge for 25-year flood = Q9
Principal Engineer, Water Resources Engineers, Springfield, Virginia.
200
-------
TABLE 1. RESULTS OF FLOOD-FREQUENCY ANALYSIS--
MULTIPLE REGRESSION EQUATIONS AND ASSOCIATED STATISTICS FOR
PEAK DISCHARGES AT SELECTED RECURRENCE INTERVALS^)
Recurrence
Interval
(years)
2
5
10
25
50
Coefficient
Multiple regression equation of multiple
correlation
Q2 « 0.069 A0'913 P1'965 0.964
Q5 - 2.00 A0'925?1'206 .976
Q • 7.38 A0'922 P°'928 .977
Q25 • 16.5 A * p°-'9' .950
Qw - 69.6 A°-847p0.511 ,902
Standard error of estimate
Loaarlthmlc Perccnt
Un1ts Plus Minus Mean
0.226 68.3 40.5 54.4
.175 49.6 33.2 41.4
.168 47.2 32.1 39.6
.178 50.7 33.6 42.2
.192 55.6 35.7 45.6
Solution Steps:
(1) Using the equations in Table 1 substitute drainage
area and precipitation to obtain for natural
watershed conditions:
Q2 * 422 cfs
Q5 =758 cfs
Q,Q = 998 cfs
Q25 - 1350 cfs
Q50 = 1800 cfs
(2) Plot the computed discharges on probability paper
and fit a smooth curve.
201
-------
a.—2-year recurrence
interval
u
a.
g
g
58
ea
u
g
la
B.
c.--10-year recurrence
interval
100
"0 20 40 60 80 100
e.--50-year recurrence
interval
I I I I
b.--5-year recurrence
interval
.—25-year recurrence
interval
0 20 40 60 80 100
f.—100-year recurrence
interval
PERCENTAGE OF CHANNELS SEWERED
FIGURE 1. RATIOS OF FLOOD-PEAK MAGNITUDE FOR URBANIZED BASINS TO
THAT FOR UNURBANIZED BASINS—FOR USE WITH FLOOD-FREQUENCY METHOET^
202
-------
(3) Select Q for natural conditions = 1390 cfs
b%)
(4) Adjust natural discharge for the effect of urbanization
by means of the coefficient 1.90 taken from Figure Id
based on 70 percent development and 70 percent of
channels sewered.
(5) Multiply the discharge of step (3) by 1.90 to obtain the
required discharge = 2640 cfs.
Thus the flood-frequency method is simple to apply. One limitation lies
in the fact that the regression equations of Table 1 were developed for
basins larger than five square miles under essentially non-urban conditions.
Increasing urbanization will tend to change the coefficients and exponents
of Table 1. The flood-frequency method also is restricted to the prediction
of peak flows and cannot be used in estimating the complete hydrograph.
RATIONAL METHOD
Another method of estimating urban peak flows of given frequency
is the so-called rational method which is commonly applied to areas less
than five square miles. Basis for the rational method Is the formula:
Q = C I A (1)
where Q is the maximum rate of runoff in acre-inches per hour or cfs; C
1s a runoff coefficient representing the integrated effects of infiltration,
surface detention, and other rainfall losses; I is the average intensity
of rainfall in inches per hour for a duration equal to the travel time or
time of concentration for the basin; and A is the drainage area in acres.
Travel time or time of concentration is the time required for flow from
the most remote part of the watershed to reach the design site.
203
-------
The two principal assumptions of the rational method are that:
(1) the maximum rate of runoff occurs when the entire area is contributing
flow; and (2) the maximum rate of rainfall occurs during the time of
concentration and the design rainfall depth during the time of concentration
can be converted to an average rainfall intensity. The use of the rational
method is best illustrated by means of an example which has been presented
previously by Rantz.
Given: Drainage area of 5.0 square miles
Mean annual precipitation = 40 inches
Degree of development is medium residential on
70 percent of the watershed and
natural on 30 percent.
Overland flow path is 500 feet long at 5 percent
slope in the undeveloped part of the
watershed.
Average channel characteristics below undeveloped
area are: length = 30,000 feet; slope =
4 percent; Manning n = 0.030; cross-
section area A = 200 fq. ft. (40 ft.
wide x 5 ft. deep) with wetted
perimeter • 50 feet.
Required: Peak discharge for a 25-year flood for storm sewer design.
Solution Steps:
(1) Compute overland and channel travel time to get time
of concentration for watershed and design rainfall
duration:
a. runoff coefficient for undeveloped area from
Table 2 is C = 0.30
b. from Figure 2 for travel path of 500 feet and
C * 0.30 overland travel time » 22 minutes
204
-------
TABLE 2. REGIONAL DESIGN VALUES OF PERCENTAGE OF IMPERVIOUS AREA
AND OF C IN THE RATIONAL METHOD 0)
Type of development
(1)
Density,
in units
per acre
(2)
Percent impervious
Santa Clara
County
(3)
San Francisco
Bay Region
(4)
C, In Rational Method
ASCE
(5)
San. Francisco
Bay Region
(6)
Residential:
Hill areas 0.5- -2
Low urbanization 3-6
Medium urbanization 7 -10
Heavy urbanization 11 -20
(apartments)
Industrial:
Nonmanufacturing
Manufacturing
Reserve
Commercial
Transporation
Public buildings
Public parks
Agricultural
Natural watersheds
6
10
20
32
50
40
20
50
70
40
12
4
2
8
15
25
40
60
50
25
60
75
50
12
4
2
— 0.11-0.30
0.25-0.
.30- .
.50- .
.60- .
.50- .
—
.50- .
.70- .
—
.10- .
—
—
40
50
70
90
80
95
95
25
.21-
.32-
.45-
.58-
.52-
.32-
.58-
.60-
.52-
.16-
.10-
.10-
.38
.52
.70
.88
.79
.52
.88
.90
.79
.32
.30
.30
c. substitute channel properties in Manning
equation to get a channel velocity of 25.2 fps
for the flood wave and a travel time a 20
minutes in the channel
d. total basin travel time = 22 + 20 = 42 minutes
(2) Compute basin-wide coefficient of runoff by weighting
both the residential and undeveloped parts of the basin;
a. according to Table 2 a medium residential area has
25 percent of its area impervious and C = 0.52
205
-------
1000
800
H4
•
u
600
400
I
200
co
ui
2
1
H
U.
o
H
O
cu
o
FIGURE 2. RELATION OF OVERLAND TIME OF TRAVEL TO OVERLAND TRAVEL DISTANCE,
AVERAGE OVERLAND SLOPE, AMD COEFFICIENT C--FOR USE IN RATIONAL METHOD
(Based on Figure 3-1, Wright-McLaughlin Engineers, Ref. 2,
Adapted and extended by S. E. Rantz)
206
-------
b. for the undeveloped portion C = 0.30
c. weighted basin-wide C = 0.70 x .52 + 0.30 x ,30
= 0.45
(3) Compute precipitation intensity, I
a. first get the rainfall depth corresponding to a
recurrence interval of 25 years and a duration of
42 minutes from depth-duration-frequency data for
San Francisco Bay region (TP 40 by USWB).
1. 25-year r.f. for 30 minutes = 0.88 inch and
for 60 minutes - 1.12 minutes
2. by interpolation 25 year r.f. for 42 minutes
= .98 inches
b. rainfall intensity = 60 x 0.98/42 = 1.40 iph
(4) Peak discharge $2$ ~ CIA = 0.45 x 1.40 x 3,200 acres
= 2020 cfs
The rational method has been used widely by drainage engineers for many
years and has been applied satisfactorily to small watersheds less than
five square miles in area. A basic disadvantage is the assumption of
average rainfall intensity throughout the time of concentration since it
is usually the intense, short-duration intensities which normally produce
major runoff events. The assumption of a gross"runoff coefficient is
also a short-coming, although it can be offset by dividing the total
watershed into smaller units and summing the individual contributions
from each to get total peak discharge. Again no routing is performed
and the total hydrograph is not constructed in the normal application
of this method.
207
-------
UNIT HYDROGRAPH METHOD
The unit hydrograph procedure is commonly employed in larger
drainage and flood control studies where the total hydrograph is required
in the design analysis. Briefly, the unto hydrograph (UH) is defined as
the hydrograph produced by one inch of rainfall excess occurring uniformly
over the basin area at a uniform rate during a specified duration. The
UH is based upon the principal assumption that the ordinates of a direct
runoff hydrograph are proportional to the ordinate of the UH times the
rainfall excess in a given time interval. The UH procedure thus represents
a linear, time-invariant process.
Use of the UH procedure involves two basic steps: (1) derivation
of the UH for specified duration; and (2) application to a given rainfall
event to determine the corresponding hydrograph of direct runoff. The
basic steps in deriving a UH from known rainfall-runoff data are:
(1) Determine rainfall excess available to direct
runoff process;
(2) Separate baseflow from corresponding hydrograph;
(3) Determine volume of direct runoff;
(4) Divide ordinates of hydrograph (without baseflow)
by volume of direct runoff.
Since the UH derived from a single storm may be in error it. is normally
desirable to average the UH from a number of runoff events and thereby
obtain an average UH of known duration for a given basin. Figures 3,
4, and 5 and Table 3 illustrate the basic steps in developing a six hour
UH from a rainfall which delivered 4.63 inches of rainfall excess over a
six hour period on a 40 square mile drainage basin.
20S
-------
FIGURE 3. RAINFALL HYETOGIWPH
FOR DERIVING UH
14 30
At
IW.MO
f
ISoo
FIGURE 4. OBSERVED RUNOFF
HYDROGilATH
C a
14
FIGURE 5. DERIVED SIX-HOUR Ul
209
-------
TABLE 3. STEPS IN COMPUTING SIX-HOUR UH FROM RAINFALL IN FIGURE 3
Day
1
2
Total
Hour
hr
6
8
10
12
14
16
18
20
22
24
2
4
6
8
10
12
14
16
Total
Flow
cfs
500
5,600
9,200
10,100
7,800
6,600
5,550
4,700
4,000
3,300
2,700
2,300
1,950
1,650
1,400
1,200
1,000
800
Base
Flow
cfs
500
450
400
400
450
450
500
550
600
600
600
650
650
700
700
750
750
800
Di rect
Runoff
cfs
0
5,150
8,800
9,700
7,350
6,150
5,050
4,150
3,400
2,700
2,100
1,650
1,300
950
700
450
250
0
59,850 =
Unit
Hydrogr,
Ordinates
cfs
0
1,120
1,915
2,110
1,600
1,340
1,100
900
740
590
460
360
280
210
150
100
50
0
Hours
After
Start
0
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
34
119,700 cfs-hours
From the data in Table 3 the area under the observed hydrograph
of Figure 4 represents 4.63 inches of runoff as follows:
Runoff'Volume =
119,700 cfs-hours x 3600 'p£- x 12 in.ches
hour A '*• feet
40 square miles x (5380
-
mi le
- 4.63 inches = depth of rainfall excess.
•
It is noted here that the six-hour UH of Figure 5 can easily be converted
to UH of either shorter or larger duration by techniques of superposition
or addition of UH lagged by appropriate time intervals.
210
-------
To demonstrate the use of the six-hour UH derived in Figure 5,
we now want to synthesize a hydrograph of streamflow resulting from a
pattern of rainfall excess shown in Figure 6. The computation of
rainfall oxcess is based upon known rainfall rates and estimates of
infiltration and other losses. The increments of rainfall excess for
successive six-hour periods are then multiplied by the UH ordinates of
Figure 5 and the resulting runoff rates ara summed as shown in Table A,
An arbitrary baseflow has been added to obtain a total basin hydrograph
as shown in Figure 7.
TABLE 4. DETERMINATION OF DIRECT RUNOFF HYDROGRAPH
FROM RAINFALL PATTERN IN FIGURE 6
Day
1
2
Hour
0000
0600
1200
1800
2400
0600
1200
1800
2400
0600
0.3 X
UH ord.
0
630
330
180
80
30
0
0.2 x
UH ord.
0
420
220
120
50
20
0
0.1 x
UH ord.
0
210
no
60
30
10
0
Base
Flow-cfs
100
150
150
200
150
100
100
100
SO
20
Total
Flow-cfs
100
780
900
600
560
290
180
130
60
20
A current example of the application of the UH is in San Diego
County where the UH is used to estimate the inlet hydrograph for a number
of developing watersheds in the San Diego region. In this application
the UH is part of a comprehensive watershed simulation program to evaluate
both the hydrologic effects of urbanization and the costs of required
drainage facilities;
211
-------
FIGURE 6. RAINFALL LOSS AltO
EXCESS IN UH EXAMPLE
0-0
Vf>o
>Mi«niv\A
FIGURE 7. DISCHARGE FROM RAINFALL
OF FIGURE 6 ft UN METHOD
*v»»»%> va<v« «"*
212
-------
REFERENCES
1. Rantz, S. E., "Suggested Criteria for Hydrologic Design of Storm-
Drainage Facilities in the San Francisco Bay Region, California,"
USGS Open File Report Prepared in Cooperation with the Department
of Housing and Urban Development, Menlo Park, California, 24 November
1971.
2. Wright-Mclaughlin Engineers, "Urban Storm Drainage Criteria Manual,"
Denver Regional Council of Governments, Denver, Colorado, 1969.
213
-------
THE WRE STORM MODEL
By
Robert P. Shublnski*
SECTION 1
INTRODUCTION
BACKGROUND
It has only been within the last decade that the real pollution
potential of urban runoff has come to be recognized. In a report published
in 1964 by the U.S. Public Health Service [l].* the nationwide signifi-
cance of pollution from urban'runoff was first identified. Since that time,
large amounts of effort and money have been devoted to the characteri-
zation of the quality of urban runoff and to the development of
methodologies and processes to control this source of pollution. Funding
for these studies has come from a number of municipalities, some states,
and from federal agencies-, notably the Environmental Protection Agency
(EPA) and recently the U.S. Army Corps of Engineers.
A review of these studies (see References 8 and 9) shows
that much work has been done in the following areas:
1. Development of stormwater treatment processes;
2. Sewer system control to maximize pipeline storage,
thereby reducing the amount and frequency of
overflows; and
3. Characterization of the quality of stormwater and
combined sewer discharges.
In addition, several sophisticated mathematical models have been
developed (some with funds from the private sector) that describe the
time-varying hydraulic response of an urban drainage system to rainfall.
A few of these models include descriptions of the quality of urban runoff.
The EPA Stormwater Management Model [2] is a typical example of the
detail and scope contained within these models..
The information and technological tools that are presently
available are deficient, however, in that they do not adequately address
some of the initial questions that must be answered in the preliminary
planning stage. One of the questions that must be answered prior to
developing a pollution control plan is: what is the present and expected
future magnitude of pollution loads carried by urban runoff from a given
watershed? Extensions of this question include such things as I) what
is the pollution load for an average event, 2) what is it for an extreme
event, and 3) how often does a given extreme event occur?
Prlntcipal Engineer, Water Resources Engineers, Springfield, Virginia.
214
-------
Given that we can answer these questions, it is then possible
to identify some constraints for the stormwater system that is ultimately
designed so that the receiving waters will be adequately protected (which
is the whole idea of "stormwater management" in the first place).
Therefore, it is necessary to identify those systems (i.e. -, combinations
of treatment rate and storage volume) that can meet the constraints. The
extremes of these combinations are obvious: All the runoff could be treated
as it arrives at the treatment plant, or all the runoff could be stored
for later treatment at a conventional treatment plant during off-peak
hours. Either of these two alternatives, however, will normally prove
to be highly uneconomical. In between these two extremes lie an extremely
large number of treatment rates and storage capacities that will satisfy
the environmental constraints placed on the system. The problem is
to identify the feasible combinations.
REPORT PURPOSE
The purpose of this document is to present an analytical method
that can be used in the preliminary planning stage to help answer the types
of questions posed above. The method has been coded into a computer
program called STORM (Storage, Treatment, Overflow, and Runoff
Model). This program represents a method of analysis to estimate the
quantity and quality of runoff from small, primarily urban, watersheds.
Nonurban areas may also be considered. Land surface erosion for urban
and nonurban areas is computed in addition to the basic water quality
parameters of suspended and settleable solids, biochemical oxygen demand
(BOD), total nitrogen (N), and orthophosphate (PO4). The purpose of the
analysis is to aid in the selection of storage and treatment facilities to
control the quantity and quality of urban stormwater runoff and land surface
erosion. The model considers the interaction of eight variables:
I, precipitation,
2. air temperature for snowpack accumulation
and snowmelt,
3. runoff, 'N
4. pollutant accumulation I related to
on the land surface / land use
5. land surface erosion, . . . }
6. treatment rates,
7. storage, and
8. overflows from the storage/treatment system.
Land uses accounted for in the model include: single family residential,
Multiple family residential, commercial, industrial, parks, and nonurban
or undeveloped areas. The program is designed for use with many years
of continuous hourly ' precipitation records. It is a continuous simulation
model but may be used for selected single events.
The City of San Francisco used this program in the preliminary
Planning phase of their Master Plan for Stormwater Management [3], The
Corps of Engineers is currently applying STORM in several of their Urban
215
-------
Studies, and East Bay Municipal Sewerage District No. 1 (which serves
seven cities on the east side of San Francisco Bay) has recently used the
model in an inflow/infiltration study of their sanitary sewer system.
The program has been documented by the Hydrologic
Engineering Center, Army Corps of Engineers in Davis, California, and
is available to the public [4],
HARDWARE AND SOFTWARE REQUIREMENTS
This program is available for the IBM 360/50, UNIVAC 1108,
and CDC 6600 or 7600 computer systems. It requires about 35.000 words
of core storage and a FORTRAN IV compiler that accepts multiple ENTRY
statements. Input is on the card reader and possibly a tape/disk. Output
is on a 132 position line printer. One to five additional tape/disk units
are required for temporary storage during the processing. The only
program differences among the three computer systems are due to
ENCODE/DECODE type statements and the way in which multiple output
files are handled. Up to three output files are generated on tape/disk
for printing at the end of the job.
216
-------
SECTION 2
COMPUTATION OF RUNOFF QUANTITY
CONCEPT OF "STORM"
The quantity of urban runoff has traditionally been estimated
by using a design storm through frequency-duration-intensity curves or
some other statistical means based on rainfall records. Such approaches
normally neglect the spacing between storms and the capacity of the
urban system to deal with.some types of storms better than others.
Often, through natural and artificial storage mechanisms,
intense short-duration storms maybe completely contained within storage
so that no untreated stormwater overflows to receiving waters.
Alternately, a series of closely spaced, moderately sized storms may
tax the system to the point that excess water must be released untreated.
Consider, for example, Figure 1 which shows the response of two different
systems to the same rainfall trace. System A, which has a relatively
high treatment rate and a small storage capacity, will overflow during
the high intensity, short duration storm. However, it will completely
contain the second storm of moderate intensity and longer duration. System
B, on the other hand, which has a low treatment rate and a large storage
capacity, completely contains the first storm. Notice that it would also
contain the second storm if the system were analyzed independently of
the antecedent storm. However, in this case the spacing of the storms
is such that the system analysis must include both rainstorms as a single
event to accurately describe the system's response to the rainfall trace
illustrated in the figure.
A storm cannot be defined by itself, but must be defined taking
into account tfee response characteristics of the urban stormwater system.
It is for this reason that an approach "was developed that would not only
recognize the properties of rainfall duration and intensity, but would
also consider storm spacing and the capacity of the urban stormwater
system.
Figure 2 shows, pictorially, the interrelationship of the eight
stormwater elements considered in this approach for estimating storm-
water runoff quality and quantity. In this approach, rainfall washes dust
and dirt and the associated pollutants off the watershed to the storage-
treatment facilities so that as much stormwater runoff as possible can be
treated prior to its release. Runoff exceeding the capacity of the treatment
plant is stored for treatment later, When the storage facilities become
inadequate to contain the runoff the untreated excess is wasted through
overflow directly into the receiving waters.
For a given precipitation record, the quantity, quality, and
number of overflows will vary as the treatment rate, storage capacity,
217
-------
t
CO
2
UJ
l~
z
A
RAINFALL TRACE
£ UNITS/HR
2 UNITS/HR
SYSTEM A: TREAT
STORAGE
TIME
4 UNITS/HOUR
4 UNITS
ro
»—"
00
UJ
e>
<
or
to
4 UNITS OF
STORAGE
UJ
CD
o:
o
SYSTEM B
24 UNITS
TREAT
STORAGE
TIME
2 UNITS/HOUR
24 UNITS
OVERFLOW
TIME
FIGURE 1
System Response Examples
-------
I, t t ' t
PRECIPITATION
i / • , / . rrtcwiri i«i iv
//' '//•/' (rainfall/snow)
1 i ' i •/
r i f i i ' / f
-T^V , , / ' '
i t i >^. ' i L~-**z—-^
STORAGE
POLLUTANT
ACCUMULATION
POLLUTANT
WASHOFF AND
SOIL EROSION
TREATMENT
FIGURE 2
Conceptualized View of Urban System Used in "STORM"
219
-------
and land use is changed. Land surface erosion is a function of land
use, soil types, ground slope, rainfall/ snowmelt energy and erosion
control practices, A typical method of investigation is to alter the
treatment, storage, and land use and note the resulting response of
the system. A group of alternatives can then be selected from among
those meeting the overflow quantity and quality objectives.
COMPUTATION OF THE QUANTITY OF RUNOFF
Runoff is calculated on an hourly basis as a function of rainfall
plus snowmelt using the following expression:
R = C(F - f) (1)
where
R = urban area runoff in inches per hour;
C = composite runoff coefficient dependent on urban
land use;
P r rainfall plus snowmelt in inches per hour over the
urban area; and
f - available urban depression storage in inches per hour.
For simplicity we will omit the snowmelt computation in our discussion
here. The interested reader is referred to the User's Manual [4] for
details of that computation.
runoff coefficient represents losses due to infiltration. It
is computed from land use data as follows:
C * Cp +(Cj - Cp) ^ X. F. (2)
where
C - runoff coefficient for pervious surfaces;
C. = runoff coefficient for impervious surfaces;
X. s area in land-use i as a fraction of total watershed area;
F. = fraction of land use i that is impervious; and
L = total number of urban land uses.
220
-------
Before the runoff coefficient is applied, depression storage
losses must be satisfied. Depression storage represents the capacity of
the watershed to retain water in ditches, depressions and on foliage.
The amount of depression storage at any particular time is a function of
past rainfall plus snowmelt and evapotranspiration rates. The function
is computed continuously using the following expression, where f is in
inches:
f = f +
o
for
f < D
(3)
where
N
D
k
= available depression storage, in inches, after
previous rainfall;
= number of dry days since previous rainfall;
= recession factor, in inches/day, representing the
recovery (evapotranspiration) of depression storage
and infiltration in inches; and
D = maximum available depression storage in inches.
Figures 3a and 3b show graphically the hourly precipitation
(P), depression storage (f), precipitation excess (P-f) and the resulting
runoff (R). Figures 3b and 3c show how the runoff is distributed between
treatment, storage and overflow for a system with a treatment rate of
0.02 inches/hour and a storage capacity of 0, 16 inches.
221
-------
IT
|.20-
Z
S •
UJ
L
< .10 •
oc
•J
_J
2
oe
o 4—i
i
-------
SECTION 3
COMPUTATION OF RUNOFF QUALITY
SOURCES OF POLLUTANTS
Basically, pollutant loads are introduced into urban runoff from
three sources:
1, The land surface itself, primarily impervious surfaces;
2. Catch basins; and
3. The sewers in combined systems.
Of these three sources, the land is the most important. Catch basins
can be a source of first-flush or shock pollution. An American Public
Works Association (APWA) study [5] in Chicago found that:
". . . the liquid remaining in a basin between runoff
events tends to become septic and that the solids trapped
in the basin take on the general characteristics of
septic or anaerobic sludge., The liquid in catch basins
is displaced by fresh runoff water in the ratio of one-
half the volume for every equal volume of added liquid.
During even minor rainfall or thaw this displacement
factor can release the major amount of the retained
liquid and some solids. The catch basin liquid was found
to have a BOD content of 60 ppm in a residential area.
For even minor storms, the BOD of the catch basin liquid
would be seven-and-one-half (7-1/2) times that of >the
runoff which had been in contact with street litter,
Improved design of catch basins, and better operational
and maintenance practices, could reduce this first-flush
pollutional effect. "
In combined sewer systems, wastewater is incorporated into
the storm runoff. In addition, the storm runoff, as it passes through
large sewers, scours sediment deposited by wastewater flows
during preceding dry-weather periods. Figures 4 and 5 illustrate the
effects of wastewater sewage and of catch basins and storm sewer scour
°n the quality of atormwater overflows [6],
As stated above, the most important contributor of pollutants
to urban runoff is the land surface itself, primarily the streets and gutters
and other impervious areas directly connected to streets or storm sewers.
Pollutants accumulate on these surfaces in a variety of ways. There is,
*or example: debris dropped or scattered by individuals; sidewalk
sweepings; debris and pollutants deposited on or washed into streets from
Yards and other indigenoxis open areas; wastes and dirt from building and
Demolition; fecal dropping from dogs, birds and other animals; remnants
223
-------
14:00
300
UJ
combined sanitary and
stormwater overflow
11=00 1200 isoo
CLOCK TIME
stormwater overflow only
w-oo
14=00
FIGURE 4
Comparison of Stormwater and Combined Overflows
(Selby Street, San Francisco)
224
-------
14-00
300i
contribution from catchbasins
and storm sewer scouring
combined overflow without
catchbasin or scour effects
14^00
I2:00 I3:CO
CLOCK TIME
14:00
FIGURE 5
First-Flush Pollutional Effects of Caixh Basin,
And Sower Scour on Combined Sewer Ovcrflov/
(Se!by Srreot. San Francisco;
225
-------
of household refuse dropped during collection or scattered by animals or
wind; dirt, oil, tire and exhaust residue contributed by automobiles; and
fallout of air pollution particles. The list could go on and on. Irrespective
of the way in which pollutants accumulate on the urban watershed, they
are generally associated with one of the following forms of street litter:
1, Rags,
2. Paper,
3, Dust and dirt,
4. Vegetation, or
5, Inorganics.
Table 1, which gives estimated street litter components for a residential
area in Chicago, provides a rough measure of the relative importance
of these components.
TABLE 1
Monthly Summary of Estimated Street Litter Components,
From a 10-acre (4 ha) Residential Area, Chicago**
Month
Street Refuir CoTponenti
(Tom/MontM
Root
Paper
OuH &
Dirt
Vepetalion
Inorganic
To'ol'
Jan.
Feb.
March
April
Moy
June
July
Aug.
Sept.
Oct.
Now.
Dec.
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.0015
.036
.034
.036
.036
.026
.036
.036
.036
.036
.036
.036
.036
.55
.55
.55
.55
.55
.55
.55
.53
.55
.55
.55
.55
.00
.00
.00
.08
.OP
.08
,013
.OR
.09
.83
.83
.00
.0"
.0?
.09
.0?
.09
.09
.09
.0?
.0"
.09
.W
.09
.6b
.76
.76
,7f-
,7f>
.76
.'6
.76
l.jfi
1..V.
.48
TOTAL'
.0180
.432
6.60
2,?2
1.03
10.43
'Some totals have been rounded off.
It is readily apparent that the most significant component is dust and
dirt except during the fall of the year when vegetation (primarily leaves)
becomes the dominant component.
<*This table is a reproduction of Table 4 in Reference 5.
226
-------
TYPES OF POLLUTANTS AND LOADING RATES
Nearly all of the pollutants found in urban runoff are associated
with the dust and dirt component of street litter. By type, COD, BOD,
and solids (suspended and settleables) are found in the greatest quantity,
Nitrogen and phosphorus are also found in significant quantities. In areas
where street deicing by salting is practiced, winter runoff contains very
high chloride concentrations. Other pollutants found in urban runoff
include pesticides, herbicides, fertilizers and other chemical additives,
heavy metals, and many other known and unknown pollutants,
Data on the rate at which pollutants accumulate on an urban
watershed is very scanty. In fact, it is almost non-existent. A lot
of data has been collected on the quality of combined sewer overflows
and stormwater discharges for various cities in the United States (Tulsa,
Oklahoma; Washington, D. C. ; Atlanta, Georgia; San Francisco,
California; Sacramento, California; and Roanoke, Virginia) as a result
of the U.S. Environmental Protection Agency's demonstration grants
program for abatement of stormwater pollution. The studies are reported
in EPA's Water Pollution Research Series; however, the dissimilar forms
in which the data is reported makes it difficult and in some cases
impossible to generalize. The problem is that data is often presented
as average concentrations or as pounds of pollutant runoff per inch of
rain, and the reported values may be for combined wastewater and storm
runoff rather than for storm runoff alone.
Even for a given watershed, there is no apt description of
"typical" stormwater runoff characteristics because of the variability
of rainfall-runoff patterns. Thus, reports of "mean concentration" or
pounds per inch of rainfall are meaningless as generalized variables and
they show poor correlation with any runoff parameters of interest,
Results from a demonstration project conducted in Tulsa,
Oklahoma [7], were summarized in terms of pounds of loading per day
Per mile of street for each of 15 areas sampled in the study. These
results, presented in Table 2, give an indication of the magnitude of
Pollutant buildup for different land uses. These findings must be viewed
vvith caution, however, because they were computed by taking the
"average" concentration of the pollutant for all events monitored, which
^hen multiplied by the total storm runoff gave "total annual mass
^missions" converted into a rate per day per mile, A much more useful
way to have developed this data, would have been to sum the products
°f discharge and concentrations over each of the observed events and then
fco sum all such total storm emissions over a year.
One of the best existing sources of information on the rate
pf accumulation of pollutants on urban watersheds is that data collected
ifi a field study in Chicago by APWA [S], This study determined the rate
°f buildup of dust and dirt in the streets on a number of different test
and then related the concentrations of various pollutants to the dust
dirt. A summary of their findin-gs is contained in Table 3,
227
-------
TABLE 2
Average Daily Loads Per Mile of Street*
(Tulsa, Oklahoma)
Total
Test Street
Area Miles
Residential
3 14,87
5 16.^52
7 6.84
8 6. 97
9 3. 11
11 49.05
13 5.58
15 2.06
Commerical
2 7.41
10 12.99
12 3.39
Industrial
1 11.46
4 28,40
6 12.24
Average Values
Residential
Commercial
Industrial
Average Load: Ibs /day /mile of
BOD
1.41
2.80
1.20
2.72
1. 12
1.60
2.58
2.47
2.54
2. 10
4,53
4.85
3,98
1.70
1.98
3.06
3.51
GOD
11.46
21.43
7.20
20.89
13.09
13,29
15. 16
8.67
15. 12
20.44
25.47
41. 10
29.29
12.73
13.9
20.3
27.7
Total
Solids
120
43
63
69
47
66
81
56
92
82
113
838
175
49
63. 1
95,7
354.
street
Organic
Kjeldahl Soluble
Nitrogen Orthophosphate
0. 26
0, 11
0, 12
0. 12
Of 07»
0,08
0, 25
0.07
0.32
0. 16
0.22
0.41
0."28
0.09
0. 14
0, 23
0, 26
0.34
0, 13
0, 10
0.21
0, 11
0, 15
0, HO
0.17
0.29
0, 13
0,30
1,30
0,30
0. 13
0. 18
0.24
0,57
"^Reproduced from Reference 7,
228
-------
TABLE 3
APWA Findings on Rate of Pollutant Buildup
On Urban Watersheds*
Amount of Ou»t and Dirt
Land the
Commercial
Induttriol
Multiple Family
Single family reiidence
Assumed weighted average
Amount of
Item
Water Soluble (mg/g)
Volatile Water Soluble (mg/g)
BOO (mg/g)
COO (mg/g)
PO4 (mg/g)
N(mg/g)
Total plate counrt/g (x 1000)
Confirmed coliform/g (« 1000)
Fecal enterococci/g
and Strength of 6OO
Ami . of 0/0
by land uie
by land U»e
BOO of D/D
lb/doy/100 ftof curb mg/g
3.3
4.6
2.3
_9JL
1.5
Pollutant by Type of
Single Family
6.0
3.8
5.0
40
.05
.48
10,900
1,300
645
7.7
3
3.6
5
5
land Uie
Multiple Family
5.6
3.4
3.6
40
.05
.61
18,000
2,700
518
Commercial
12.4
6.9
7.7
39
.07
.41
l'l,700
1,700
329
To convert the data contained in Table 3 into a form comparable
with that of Table 2, the Dust and Dirt is multiplied by 2 (gutters per
street) x 52.8 (100's of feet of gutter per mile) x (constituent concen-
tration/1000). E. g. , the rate of BOD accumulation on an urban area that
is single family residential is: 0.7 x (2 x 52.8) x 5.0/1000 = 0.36
Ibs/day/mile. 'Rates of pollutant buildup in pounds per day per mile are
given in Table 4.
TABLE 4
Average Daily Loads Per Mile of Street
(Chicago, Illinois)
Land Use
Single Family Residential
Multiple Family Residential
Commercial
Industrial
Average
BOD
0.36
0.87
2.70
1.45
Load: Ibs/day/mile
COD
?.95
9.70
13.6
N
0.03
0.15
0. 14
of street
PO4
0.004
0.01?
0.024
vSee Reference 5.
229
-------
Comparison of the values in Tables 2 and 4 indicates that
while the Chicago data is consistently lower than the Tulsa data, the
rates of buildup are similar in commercial areas except for PCM. Multiple
family residential values for the Chicago area also compare well with
the Tulsa data except for PO4. The principal reason for the lowe"r values
in the Chicago data is that the reported values are for the soluble portion
of the constituents and do not include that portion found in suspended and
settleable material, Also, the Tulsa data are approximations based
on over-simplified computations. Thus, of the two sets of reported data
the Chicago data is probably better, given that the amount of constituents
contained in £he solids can be determined.
One fact that is quite evident from both the Chicago and Tulsa
data is that the rate of buildup of pollutants on an urban watershed varies
significantly with land use. Intuition would tell us this is true. Both
sets of data indicate that industrial and commercial areas are much dirtier
than residential areas. This would be expected since there is higher
pedestrian and vehicular traffic densities in these areas. The data shows
that pollutant accumulation rates are approximately one and one-half to
five times as great in commercial and industrial areas as they are in
residential areas.
ENTRY OF POLLUTANTS INTO URBAN RUNOFF
The first raindrops that fall on an urban watershed simply wet
the land surface. As additional rain falls the impervious surface will
become wet enough that some of the water begins to form puddles, filling
the depression storage. This initial rain begins to dissolve the pollu-
tants in the gutters, streets, and on other impervious surfaces and
eventually, as this water actually begins to flow off the watershed it
carries the dissolved material in it.
As rainfall intensity increases, overland flow velocities
become sufficient to pick up solids. Suspended solids are, of course,
picked upat smaller velocities than settleable solids. The settleable solids
are carried off the watershed in two ways, If the velocity is sufficiently
high, the settleable solids may be suspended in the overland flow. At
lower velocities, particles may simply be rolled along the bottom surface
toward the stormwater inlet.
The rain that initially falls on pervious surfaces infiltrates
into the ground. If the rainfall is sufficiently intense, the infiltration
capacity may be exceeded and the excess rainfall begins to fill the
depression storage on the pervious surfaces. Finally, if the rainfall
is of sufficient intensity and duration, runoff will begin to flow off the
pervious areas, onto the impervious areas and thence into the stormwater
inlets. Present experience, however, indicates that the amount of runoff,
and hence the pollution loads, contributed from pervious surfaces in urban
areas are small compared to those coming from the impervious areas
and can be neglected in determining the quality of surface runoff. This is
230
-------
especially true of surfaces covered with vegetation such as lawns and
gardens, Figure 6 illustrates the differences in runoff and pollution
load from a watershed that would occur if it was converted from a park
(90% pervious) into a multiple residential area (20% pervious).
ESTIMATION OF THE RATE OF POLLUTANT BUILDUP
ON URBAN WATERSHEDS
Since dirt is the major component of street litter and is th,e
primary source of pollutants in urban runoff, the most basic approach
for estimating pollutant buildup rates would be to relate them to the dust.
and dirt accumulation rates.
Using APWA units (Table 3) for the rate of dust and dirt
accumulation, the rate of buildup DD_^ for a given land use L can be
expressed as:
where
DDL = ddLx (GjVlOO) x AL (4)
DD, = rate of dust and dirt accumulation on a watershed
of land use L in Ibs/day;
ddT = rate of dust and dirt accumulation on watershed L
in lbs/day/100 feet of gutter;
G. = feet of gutter per acre in watershed L; and
A T = area of watershed L in acres,
LJ
The rate factor dd. should be supplied by the user for his area, Default
i_»
values, which are those shown in Table 3 are incorporated into STORM
and can be used if no better data are avaiitble.
The initial quality of a pollutant p on watershed L at the
beginning of a storm can then be compxited as:
where
Pp = + Ppo
P = total pounds of pollutant p on the watershed L at the
P beginning of the storm;
F = pounds of pollutant p per pound of dust and dirt;
231
-------
ALTERNATIVE LAND USE
(Multiple f-'esidentiotj
EXISTING LAND US£
(Pork)
10:00
|i:CO 12:00 1300
CLOCK TIME
FIGURE 6
Effect of Chanced Land Use on Characteristics
Of Subcatchment Runoff, Selby Street
232
-------
N_j = number of dry days since the last storm, and
P = total pounds of pollutant remaining on watershed L
"° at the end of the last storm.
In practice, P is usually limited to the amount that would be
accumulated in a 90-day dry period. The reason for this is that the
efficacy of extrapolating daily buildup rates beyond this point (which
was arbitrarily selected) is uncertain. Moreover, if equation (5) is used
repetitively over long periods of time, positive errors could tend to
accumulate in P resulting in overly large values of P .
If street sweeping is practiced on the watershed, the number of
dry days since the last storm must be modified to account for the number
of street sweeping-s that occurred since the last rainfall. The correct
expression to use for P is then:
Pp = PpQ(l-E)n+ Ns xDDL xFp [(1-E) + . . ,(l-E)n]
XFP (6)
where
Nq = number of days between street sweepings;
n - number of times the street was swept since the
last storm; and
E = efficiency of street sweeping (0. 6 to 0. 95),
DETERMINATION OF URBAN RUNOFF POLLUTION LOADS
To compute the amount of pollutant washed off the watershed
during a storm, it is assumed that the amount of pollutant removed at
any time t is proportional to the amount remaining:
dP
£ - vr> t-7\
dt " ~*p (f'
We stated earlier that the runoff rate Q also affects rate of pollutant
removal, therefore K must be functionally dependent upon Q, However,
given two identical watersheds except for their area size, for the same
rainfall rate r on both watersheds a higher runoff rate would occur from
the larger watershed. This area effect can be eliminated by dividing the
runoff Q by the imagery ious area of the watershed. The impervious area
is used because only a negligible amount of the runoff comes from the
pervious area. Since cfs per acre are equivalent to inches per hour,
we can say that K is functionally dependent on the runoff rate R from
the impervious area, where R is in inches per hour. Finally, assuming
233
-------
that K is directly proportional to R and that a uniform rainfall of 1/2
inch per hour would wash away 90 percent of the pollutant in one hour
(a somewhat arbitrary assumption), we can say that K = 4,6R, Making
this substitution into equation (7) and integrating over a time interval
At (during which R is held constant) gives:
P (t + At) = P (t)e~4'6RAt (8)
Equation (8) is the basic form of the overland flow quality model developed
by Metcalf & Eddy, Inc. , as part of the EPA Stormwater Management
Model [2], Although it is simplistic and contains many assumptions,
it is the best overland flow water quality predictor or simulation model
that presently exists. Moreover, experience with that model (See
Reference 6 and Volume II of Reference 2) has shown it to give fairly
good results.
Some idea of how equation (8) behaves can be gained by exami-
nation of Figures 7 and 8. Figure 7 shows that for a constant runoff
rate R, the amount of pollutant remaining on the watershed decays
exponentially. Under a time varying R the picture is quite different, as
illustrated by the upper graph Figure 8. From the curve of P vs. t
it can be seen that the amount of pollutant removed during an interval
At is P(t)-P(H At). The _ra_te of removal of mass from the watershed M
is simply [P(t)-P(t+ A t)]/ At, which can be expressed as: P
Mp = P(t) x (l-e'4t6RAt)/At (9)
The variation of M with time for the associated hydrograph
is plotted in the lower graph of Figure 8. A plot of M versus t is termed
a pollutograph, one of the most informative methods for expressing the
pollutant load carried by urban runoff, To determine the concentration
of a pollutant in the runoff as a function of time, one simply divides the
pollutograph value M by Q (with appropriate conversion factors).
Equation (9) must be modified, however, because not all of
the dust and dirt on the watershed is available for inclusion in the runoff
at a given time t. Thus pollutants which are tied to the dust and dirt
are not all available either. The Storm Water Management Model study
[2] found that for suspended solids the available fraction at any time was:
A = 0.057 + 1.4R1' l (10)
sus l '
For settleable solids it has be en. assumed that the availability factfor is
A . = 0.028 + l.OR1'8 (11)
set * '
With regard to BOD, nitrogen and phosphate, recall that the APWA data
[5] described the dissolved fraction, which is independent of the amount of
234
-------
TIME, t
FIGURE 7
Basic Form of the Overland Flow Quality Model
235
-------
£
Q
LU
X
ex
TIME.t
AREA
VP>1
TIME,!
FIGURE 8
Development of Pollutograph (M vs, t)
From Time History of P
236
-------
solids available for runoff. In the Storm Water Management Model study
it was found that the BOD associated with the suspended solids was about
10 percent of the suspended solids load. We have further assumed that
the BOD tied to the settleable solids is 2 percent of the settleable solids,
For nitrogen and phosphate, we have assumed that BOD, N, and PC4
are associated with suspended and settleable solids in the same proportion
as they are in the dissolved state.
Thus, correcting equation (9) for available suspended and
settleable solids and adding the BOD, N and PO4 found in the solids,
we get the following set of equations which are used in STORM:
Suspended Solids
Msus(t> " Asus Psus <« * EXPT
(12)
1'
where
A = 0,057 -f 1.4R
sus
• EXPT = (1-e"4' 6RAt)/At, with At = 1 hour
Settleable Solids
"set'1' ' A
where
A
get
= 0.028+ R
set
1'8
Pset<'>xEXPT
BOD
Mbod(t) = Pbod
+0.02
Nitrogen
P04
= P
nit
EXPT + .045 M + ,01
gus
M
po4(t) = PpQ4 (t) x EXPT + ,
0045
+ .001 N
(15)
(16)
237
-------
SECTION 4
COMPUTATION OF TREATMENT, STORAGE AND OVERFLOW
PROCEDURE
Computations of treatment, storage and overflow proceed in
an hourly step-by-step method throughout a period of rainfall/snowmelt
record. For every hour in which runoff occurs the treatment facilities
are utilized to treat as much runoff as possible. When the runoff rate
exceeds the treatment rate, storage is utilized to contain the runoff.
When runoff is less than the treatment rate, the excess treatment rate
is utilized to diminish the storage level. If the storage capacity is
exceeded, all excess runoff overflows into the receiving waters and does
not pass through the storage facility. This overflow is lost from the
system and cannot be treated later. While the storm runoff is in storage
its age is increasing. Various methods of aging are used including
average, first-in: last-out, first-in: first-out, or others, depending on
the physical conditions encountered.
The computation of storage and the interplays among
rainfall/snowmelt, storage and treatment represent a simplistic approach
for dividing a rainfall record into unique events such that the event is
defined in terms of the urban system. For example, whether two "storms"
are considered as two isolated occurrences or as one large storm is
entirely dependent upon how the system will react to them. If the system
has not recovered from the first when the second arrives, the two definitely
will interact and hence must be considered together. "Events" are defined
as beginning when storage is required and continues until the storage
reservoir is emptied. All the rainfall occurring within this period is
regarded as part of the same event. If precipitation produces runoff that
does not exceed the treatment rate, the runoff will pass through the
treatment process but will not register as an event. From the standpoint
of the urban stormwater system, such precipitation is inconsequential and
hence is not part of an "event" even if it should occur immediately preceding
an obvious event.
The runoff coming into the storage/treatment system is given
by equation (1). The quantity of system overflows are computed using:
QQ = R -QT -Qs (17a)
Q_ = minimum of (R + Q , T) (17b)
1 8t-l
Qfl = minimum oi (R - QT , S) (17c)
238
-------
where
Q = watershed runoff overflow, in inches;
*>
Q = watershed runoff treated, in inches;
Q = watershed runoff stored, in inches;
s
Q = watershed storage remaining in previous hour, inches;
St-l
R = watershed runoff as calculated using equation (i), inches;
T = treatment rate in watershed inches/hour; and
S = storage capacity in basin, inches.
The quality of system overflows are computed as follows for each pollutant
for each hour:
Mpo = Mp'Qo/R) (18)
MpT/s ' Mp
where
M
M . = total pounds of pollutant overflowing from system;
M = total pounds of pollutant p coming into the system; and
T / = total pounds of pollutant p going to storage/treatment,
The program does not model the treatment process but it does
compute the quantity of water treated, It is assumed that the pollutants
will be reduced to an acceptable level before the storm water is released.
The age of pollutant in storage is computed as previously mentioned.
The format of STORM output data is given in Table 5..
239
-------
TABLE 5
"STORM" Output
I. STATISTICS BY EVENTS
A. RAINFALL
/ DURATION OF RAINFALL EVENT
2. HOURS OF RAIN
3.TOTAL RAINFALL
B. STORAGE
./. TIME SINCE LAST EVENT
2. DURATION OF STORAGE
J. TIME TO EMPTY
4. MAXIMUM STORAGE USED
C. OVERFLOW
/ TIME OVERFLOW STARTS
2. DURATION OF OVERFLOW
3. QUANTITY OF OVERFLOW
4. OVERFLOW IN FIRST THREE HOURS
D. TREATMENT
I DURATION OF TREATMENT
2 QUANTITY TREATED
E. QUALITY (susp. solids, sett, solids, BOD, nifrogen, phosphorous)
/ MASS £'MISSION IN RUNOFF
2. MASS EMISSION OF OVERFLOW
J. MASS EMISSION DURING
FIRST THREE HOURS OF OVERFLOW
I. AVERAGE STATISTICS (A-E ABOVE)
A. FOR ALL EVENTS
8. FOR ALL OVERFLOW EVENTS
C. EVENTS / YR
D. OVERFLOWS / YR
240
-------
REFERENCES
1. "Pollutional Effects of Stormwater and Overflows from Combined
Sewer Systems - A Preliminary Appraisal, " USPHS (November
1964).
2. "Stormwater Managment Model, Vol. I, Final Report. " 11024 DOC
07/71, Metcalf and Eddy, Inc., University of Florida, Gainesville
and Water Resources Engineers, Inc. , Report for EPA (July 1971).
3, "San Francisco Master Plan for Waste Water Management." Dept.
of Public'Works, City and County of San Francisco (Sept. 15, 1971).
4. "Urban Storm Water Runoff 'STORM'." Generalized Computer
Program, 723-S8-L2520, Hydrologic Engineering Center, U.S. Army
Corps of Engineers (May 1974).
5. "Water Pollution Aspects of Urban Runoff. " American Public Works
Association, Water Pollution Control Research Series, Federal
Water Pollution Control Administration, Report No. WP-20-15
(January 1969).
6. Roesner, L. A. , D.F. Kibler, and J.R. Monser, "Use of Storm
Drainage Models in Urban Planning." Presented at the AWRA
Symposium on Watersheds in Transition, Colorado State University,
Fort Collins, Colorado, June 1972, Proceedings, AWRA, Urbana,
Illinois, pp, 400-405 (1973).
7. "Storm Water Pollution from Urban Land Activity, " AVCO, Economic
Systems Corporation, Water Pollution Control Research Series,
Federal Water Quality Administration, Report No, 11034 FKL 07/70
(July 1970),
8. Field, R. and E.J. Struzeski, Jr., "Management and Control of
Combined Sewer Overflows." Journal Water Pollution Control Fed.
44, 1393 (.1972).
9- Lager, J. A. , and W. G. Smith, "Urban Stormwater Management and
Technology: An Assessment." Contract No. 68-03-0179, Metcaif &
Eddy, Inc. , Western Regional Office,, Report for USEPA (December
197.3) (Draft).
241
-------
THE EPA STORMWATER MANAGEMENT MODEL
By
Wayne C. Huber*
James P. Heaney*
INTRODUCTION
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 (1,2). The Environmental Protection
Agency (EPA) has recognized this problem and led and coordinated
efforts to develop and demonstrate pollution abatement procedures (1).
These procedures include not only improved treatment and storage
facilities, but also possibilities for upstream abatement alternatives
such as rooftops and parking lot retention, increased 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.
It was recognized that an invaluable tool to decision makers would be
a comprehensive mathematical computer simulation program that would
*Associate Professors, Environmental Engineering Sciences,
University of Florida, Gainesville, Florida.
242
-------
accurately model quantity (flow) and quality (concentrations) 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^) 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 encompassing rainfall, runoff, drainage, storage/
treatment, and receiving waters. The major components of the SWMM
are illustrated in Figure 1 in which program segments are seen to
correspond to physical components of the urban runoff process. The
SWMM also has preliminary decision-making capabilities by virtue of
cost estimation of abatement facilities, hydraulic pipe-sizing, and
adaptation to optimization techniques for overall urban water resources
planning.
Most of the theory behind the operation of the SWMM has remained
unchanged since its publication in Volume I, Final Report of the
243
-------
RUNOFF
(RUNOFF)
INFILTRATION
(INFIL)
TRANSPORT
(TRANS)
EXTERNAL
STORAGE
tSTORAO)
RECEIVING WATER
(RECEiV)
DRY WEATHER
FLOW
(FILTH)
INTERNAL
STORAGE
(TSTRDT)
COST
(TSTCST)
INPUT
> SOURCES
CENTRAL
> CORE
CORRECTIONAL
> DEVICES
EFFECT
Note; Subroutine names are shown in parentheses.
Figure 1 OVERVIEW OF MODEL STRUCTURE
244
-------
Storm Water Management MOdel (3). This document remains as the
primary background reference for the underlying principles of the
model, and material found in it is not repeated here.
However, many changes have been made in details of the program,
input/output requirements and data descriptions. As a result, an
official "Release 2" of the SWMM was issued in March 1975 (4).
The reader is referred to this publication which includes a revised
and improved User's Manual to accompany the "Release 2" program.
245
-------
REFERENCES
1. Field, R,, and E, J. Struzeski, Jr., "Management and Control of
Combined Sewer Overflows," J. Water Pollution Control Federation,
Volume 44, No. 7 (1972).
2. Gameson, A. L., and R. N. Davidson, "Storm Water Investigations
at Northampton," Institute of Sewage Purification, Conference
Paper No. 5, Annual Conference, Leandudno, England (1962).
3. Metcalf & Eddy, Inc., University of Florida and Water Resources
Engineers, Inc., "Storm Water Management Model," Water Pollution
Control Research Sertes. 'U.S. Environmental Protection Agency
Report No. 11024DOC07/71, Volume I, Final Report, July 1971.
4. Huber, W. C., et al. "Storm Water Management Model User's Manual
Version II." Water Pollution Control Research Series. U.S.
Environmental Protection Agency. Report No, 670/2-75-017,
March 1975.
246
-------
DECISION-MAKING FOR WATER QUANTITY AND QUALITY CONTROL
By
James P. Heaney*
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 et al. (1969)
and others have set up optimization models to determine tlie least cost
combination of sewage treatment and low flow augmentation.
More recently, Kirshen, Marks, and Schaake (1972) have devised a mathe-
matical model for analyzing stormwater quality control alternatives.
A linear programming model is used to determine the sizes and operating
policies for sewer pipes, storage reservoirs, and treatment plants. The
SWMM 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 stormwater 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
*
Associate Professor, Environmental Engineering Sciences, University of
Florida, Gainesville, Florida.
247
-------
for examining this more general problem has been developed. Addressing
the problem in this broader perspective, the determination of the over-
all 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.
If 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 wastewater focuses on design of
treatment units. There has been little systematic effort to devise
overall strategies for drainage and flood control. The studies in
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 partici-
p
pating in the layout of an 800-acre (3.24 Km ) planned urban development
which was 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 floodplain;
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
248
-------
that the development costs are competitive or lower than existing
practices.
The methodology used in addressing the problem can be 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;
F) specification of performance criteria;
6) statement of objectives; and
H) system optimization and examination of equity question.
Although detailed discussion of these components was presented in
this Short Course, this information is also contained in the following
technical report:
EPA-670/2-75-022
Urban Stormwater Management Modeling and Decision-Making
James P. Heaney
Wayne C. Huber
University of Florida
REFERENCES
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.
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.
Pyatt, E. E., J. P. Heaney, 6. R. Grantham, and B. J. Carter, "A Model
for Quantifying Flow Augmentation Benefits," Final Report to FWPCA.1969.
249
-------
Selected Case Studies Using Stormwater
Management Models - Quantity Aspects
By
Gerald T. Orlob*
INTRODUCTION
The problems encountered in putting the Storm Water Management Models
to practical use are best illustrated by "hands-on" experience—actual
cases. While no particular set of case studies will highlight the myraid
of unique problems that can be encountered in accord with Murphy's Law,
we have chosen a set that shows at least some of the major obstacles to
be overcome. These studies include investigations of:
1. Northwood Tract, Baltimore, Md.
2. Pegel Ti/T?, City of Hamburg
3. View Ridge; City of Seattle
4. Vine Street, Melbourne, Australia
5. Stability Problems—A Special Study
In addition, some special problems encountered with wiers and "surge
tanks" in storm drainage systems are discussed.
PROBLEMS ENCOUNTERED—POSSIBLE REMEDIES
Among the difficulties in adapting the models to practical use in
simulation of storm runoff the following seem to be the most common:
1. Observed and simulated peaks of runoff are not coincident in
time.
2. Watershed yields are not in agreement.
3. There is a lag between observed and simulated events.
4. Simulated runoff pattern, although similar in shape and area
to that observed, tends to be smoother, more attenuated.
5. Simulated runoff is. discontinuous.
6. Initial peaks are not well simulated although subsequent peaks
are, and vice versa.
A few brief comments on each of these problem areas may help us to
focus on some possible causes before we review the selected cases.
Senior Partner, G. T. Orlob & Associates, Orinda, California.
250
-------
COINCIDENCE OF PEAKS
This may result from an incorrect start time for the model, from
not offsetting the infiltration process in the initial stages of the
storm, or from an incorrect estimation of basin slope or imperviousness,
or both. Start time and basin characteristics may be adjusted*, the off-
set in infiltration simulation is conceptual and would have to be
corrected in the model.
WATERSHED YIELD
This discrepancy between model and prototype is a most common
problem. It may result from improper estimation of actual rainfall on
the basin, incorrect assessment of long-term infiltration, or incomplete
gaging of runoff. There is no substitute for good precipitation data
or careful hydrologic analysis. Infiltration estimates can be adjusted
by modifying Norton's equation, i.e., fc, the final constant value of
infiltration.
LAG
In addition to incorrect start time and poor estimation of basin
parameters, lag may be caused by over-application of other time dependent
processes, e.g., infiltration, depression filling, etc. Adjustment of
the infiltration decay rate or the amount of depression storage may
correct this difficulty.
ATTENUATION
Excessive attenuation of the runoff hydrograph may be attributed
to using too long a time step in integration, too coarse a structure for
representing basin detail, or over-application of storage routing.
Usually, better results are obtained by increasing detail, both spatially
and temporally. The inverse problem, i.e., excessive sharpening of peaks
and valleys, may be a result of failure to consider routing effects in
overland flow and conduits.
DISCONTINUOUS RUNOFF
Excessive long-term infiltration may cancel out precipitation
contribution to runoff. Failure to properly represent pervious areas
can result in a pattern where runoff ceases too abruptly. A remedy is
found in adjusting infiltration fates for pervious areas.
INITIAL PEAKS
Poor representation of initial peaks is most likely related to over-
correction for depression storage, interception, or initial infiltration.
Occasionally, for large basins, it can result from the non-uniformity of
Precipitation associated with a moving storm. In the latter case, a
remedy is to lag input precipitation to a more detailed basin representa-
tion.
251
-------
CASE STUDIES
Northwood Tract
Baltimore's Northwood Tract has served as an excellent study
example for calibration and verification of the EPA SWMM and WRE's
modifications of this model package. The Tract, illustrated in
Figure 1, encompasses approximately 47.4 acres of urban drainage with
an average imperviousness of 68 percent. The collected flow from the
area passes through a 12-foot Parshall Flume with a rated capacity of
about 400 cfs. Two continuous recording rain gages are available, one
of the weighing bucket type in the northeast quadrant of the area and
the other of the tipping bucket type near the southeast corner of the
area.
Figures 2, 3, and 4 illustrate the performance of the SWMM, as
modified by WRE, in simulating the runoff generated by three storms
over the area. Figure 2 may be considered the calibration run and
Figure 3 a verification check. Figure 4 illustrates the model's
capability to simulate a complex storm of highly variable temporal
pattern.
Two specific problems Were encountered in these studies. First,
there was a tendency for the model to produce accentuated peaks early
in the storm, while good checks were made in the balance of the simula-
tion. This was corrected in this example by increasing the allocation
to detention storage early in the simulation. Second, the model failed
to simulate the very low flows apparently observed at the Parshall Flume
midway through the storms of 1 August and 4 August (second storm). Upon
examining the data it was determined that this was primarily due to
inaccuracies in low flow measurement in the very wide-throated flume.
In general, however, the simulations are quite excellent and give
a good account of the model's capabilities.
Pegel Ti/T?, City of Hamburg
The Storm Water Management Models have been adapted for a unique
study of the urban drainage system for the City of Hamburg, Germany.
Figure 5, A and B, shows the configuration of the test basin and the
"testnetz", designated by the station at the lowest point, Pegel Ti/T2.
Results of simulation, following a calibration phase are summarized in
Figure 6. The computed runoff I"Berechnet") compares well with the
measured ("Gemessen") flow for a storm with two distinct precipitation
peaks.
Some special problems were encountered in calibration of the model*
however, and a few changes in the program were made to make the model
more realistic. One of these changes is illustrated in Figure 7, where
a "before" and "after" comparison is made. In the initial version of
the runoff model, referred to here as "Selby Street" after that used in
the San Francisco Study, runoff was delayed until the full detention
flow was developed over the drainage area. This resulted in a delay
in the rising limb of the hydrograph that was not realistic. Actually.
252
-------
Rain
RAIN GAGE
"Creefc
- WEIGHING
BUCKET TYPE
RAIN GAGE #2 - TIPPING BUCKET TYPE
•^— STORM CONDUIT
• INLET
• MANHOLE
— — — DRAINAGE AREA BOUNDARY
— SUB DRAINAGE AREA BOUNDARY
(1> SUB AREA NUMBER
Figure 1. PLAN OF NORTHWOOD DRAINAGE AREA
BALTIMORE, MARYLAND
253
-------
NORTHWQOD BflLTIMQRE 2ND STORM OF RUO 1 1965
Jjp.OO
0.10 0.20
0.30 0-40 0.50 0.60
TIME IN HOURS
0.70 0*00
Figure 2. RUNOFF SIMULATION, BALTIMORE NORTHWOOD TRACT
2nd STORM OF AUG. 1, 1965
.00
r\
N8RTHW8QD BRLTIMQRE 2ND STQRM 8F RUB 4 1965
0.20 0-40
0<60 0.80 I ,.00 T.20
TIME IN HOURS
1.40 1.60
Figure 3. RUNOFF SIMULATION, BALTIMORE NORTHWOOD TRACT
2nd STORM OF AUG. 4, 1965
254
-------
N8RTHW0QD BRLTIM0RE 1ST STORM
RUG 1 1965
0.60
t±Hf-
0.80 l-'.OO
IN HQURS
T.20
1.40
1.60
4. RUNOFF SIMULATION, BALTIMORE NORTHWOOD TRACT
1st STORM OF AUG. 1, 1965
-------
ro
en
CT*
Figure 5. TEST AREA AND "TESTNETZ," PEGEL T,/T9, CITY OF HAMBURG
(Ref. 2)
-------
•.III
I.Ml
I
IT.t II.« Il.l II.*
CLOCII Tint IMOIMI M|N «>»C
It.f
I I
U.I I«.M It.ll
illf UTO.I
Figure 6. SIMULATION OF STORMWATER RUNOFF
PEGEL TT, CITY OF HAMBURG (Ref. 2)
257
-------
Ntnvn)
Qll/s)
MO
Uhrurt t
IN
ro
en
oo
too
GangVnl*
,",
l\
Ganglini*
SMf S»MI Modri
Gcmnscnt
Ganglinit
Homborg
08
06
in 04
02
0.2 0.4 06 1X6 tO
VERHALTMS t/Tt
UhfMit I
Figure 7. COI^PARISON BETWEEN HAMBURGER AND
SELBX STREET MODELS (Ref. 2)
Figure 8. IMPERVIOUS RUNOFF FACTOR
(Ref. 2)
-------
runoff should become tributary to the gaging point shortly after rainfall
commences, i.e., after initial wetting and some surface detention is
satisfied, but not all. This was corrected for by allowing some fraction
of the flow from theTTmpervious area to be immediately contributing and
that fraction to increase steadily thereafter. Figure 8 shows the nature
of the function employed. It has the general form
At • Ao + " - V T, ek(1" T.}
where At is the fraction of calculated surface runoff from Impervious
areas that are fully tributary and t/Ti is the fraction of the time
required to reach full surface detention. An initial value, Artr can be
assumed, say 0.2 to 0.4,(depending on the nature of the drainage surface
and the shape of the curve of A vs t/Tj can be accommodated by the
coefficient k, usually taken in the Hamburg study as unity.
The result of making this correction .is apparent in Figure 7 where
the Hamburger Model is seen to agree closely with the measured value.
The Selby Street version is seen to produce a hydrograph displaced
generally to the right, i.e., delayed from the measured pattern.
View Ridge and Central Business District. Seattle
Figure 9 shows the location of a fully urbanized residential area,
View Ridge, in Seattle, Washington, that was used as a test case for
Jodel calibration in connection with an urban drainage study performed
by Water Resources Engineers and Kramer, Chin and Mayo, Inc. for the
Seattle District Corps of Engineers. The area is 605 acres and the
average imperviousness is 40 percent.
As one would expect, some difficulties were experienced in adjusting
[ainfall data from the raingage several miles away to the drainage basin.
Also, the basin terrain is generally hilly, whereas the gage is situated
Jn the somewhat flatter open area of the University of Washington campus.
Nevertheless, after several adjustments of the model an acceptable
calibration was obtained. This is shown in Figure 10 together with a
test for a small (25 acre), relatively impervious (99 percent)
iness area in downtown Seattle.
It is noted that the model gives a good account of the peak flow for
the View Ridge catchment, but is a bit too low in the trailing portion of
lye storm. For the Central District the reverse seems to be the case,
although we note that an adjustment of the peak upward, perhaps by modify-
surface detention, would probably result in a comparison similar to
for View Ridge. Apparently, in both cases, too much weight is given
sustained runoff from already wetted areas.
259
-------
f •
' :
jl^|^^§^]J£i|^
^;i^^^^Sj£^; ^
- — .'-_;.__ . 4u»A03?l • • *
-\ -'•:. t •••' ^ ;
^-fetpS
^T-^Q^Liiiz:
i^iSrif^FJiist
~3 - — :.; *-.^
\ Jjff^ SC/»L£ J".)
Figure 9. VIEW RIDGE URBAN DRAINAGE BASIN
VIEW RIDGE I
CENTRAL BUSINESS
DISTRICT
MOO 000 0200 0300 (MOO OMO
004-
q' • "'•• II.I '•
HfcroattAfH tr u.ar w
<-m. "'.
MrrnxHAf+i tr DCNHY mr
Figure 10. CALIBRATION OF RUNOFF MODEL, SEATTLE DRAINAGE STUDY
(Ref. 3)
-------
Vine Street, Melbourne, Australia
Heeps and Mein (4) carried out an excellent comparative study of
the EPA-SWMM and models developed by the Road Research Laboratory in
England (5) and University of Cincinnati (6). The three models were
tested against gaged runoff from two urban watersheds under a variety
of conditions. Among the model capabilities and properties tested were:
a. ability to simulate runoff from storms of diverse characteristics,
b. effect of degree of discretization,
c. effect of depression storage,
d. effect of infiltration,
e. treatment of pervious areas, and
f. sensitivity to input data.
An evaluation matrix was selected for purposes of comparisons; it included
percent of observed runoff, percent of observed peak, sum of squares of
errors, and processor time. It was generally concluded that the EPA-SWMM
was the superior in performance of the three model packages tested
although it was noted that processor time was invariably greater for
the SWMM.
To illustrate the effect of discretization each model was tested for
the same storm conditions on the Vine Street Basin at two or three levels
of detail. Figure 11 illustrates three levels of discretization employed.
In the upper sketch the 70 hectare catchment is divided into 12 subbasins
ganging in size from 0.1 to 18.6 ha. In the middle sketch the catchment
is divided into 51 subbasins and in the lower sketch, the maximum discreti-
zation there are 256 subbasins. The detail of inlets and drains was
increased roughly in proportion.
Figure 12 shows some results of the SWMM for coarse and medium
discretization. It is noted that the major effect of increased detail
"•s a reduction in peak flow, due primarily to increased conduit routing
and corresponding attenuation of the runoff hydrograph.
Figure 13 illustrates a comparison of the three models for a fairly
complex storm over the Vine Street Catchment. The SWMM is seen to be
superior model for this test. The RRL model tends toward excessively
peaks and the CURM, while giving a good account of the early peak,
to simulate the balance of the runoff pattern. This is attribute
ty Heeps and Mein to neglect of the runoff contribution from pervious
areas. The comparison statistics for this test of the three models are:
SWMM CURM RRL
% Observed Runoff 102 139 49
% Observed Peak 106 189 91
Sum of Squares of Errors 61 78 86
(m6/sec2)
Processor Time (min.) 34.23 31.05 1.87
261
-------
12 subbasins
13 Inlets
12 conduits
51 subbasins
55 Inlets
54 conduits
256 subbasins
all inlets
313 conduits
MEDIUM
Figure 11. VINE STREET CATCHMENT
MELBOURNE, AUSTRALIA
(after Heeps and Mein, Ref. 4)
262
-------
ro
en
u>
Intensity
(-/hr.)
STORM OF 12-02-1972
VINE SHEET
Observed
— • — SNMf-cnna
— > — SUM* -wdiun
50 75 100 125
Tiae fro* stmrt of storm (win.)
200
Figure 12. PERFORMANCE OF EPA-SWMM ON VINE STREET CATCHMENT
AT COARSE AND MEDIUM LEVELS OF DISCRETIZATION (Ref. 4)
-------
ro
CT>
( m /sec.)
1.50
1.25
1.00
.75
.50
.25
A
STORM OF 6-11-1971
VINE STREET
^•Cfcserved
—A— SKhM
—-o— RRL
— "— CURM
10
20
30
40
50
60
70
Rainfall
Intensity
( nm/hr.)
Time of start of storm (hours)
Figure 13. COMPARISON OF SWMM, CURM, AND RRL MODELS,
VINE STREET CATCHMENT (Ref. 4}
-------
Stability Problems—A Special Study
Simulation of the dynamics of storm flow through closed conduits
is fraught with possibilities for computational instability. Often,
however, it is difficult to distinguish numerically induced oscillations
in the solution from hydraulic behavior that may sometimes produce real
fluctuations, even reversals, in flow. The usual quick test for computa-
tional instability is to determine whether or not the oscillation
strengthens with time, i.e., diverges, damps, out, i.e., converges, or
remains steady in period and amplitude.
In the first instance computational instability is most likely.
One must devise methods for eliminating it. These include changes in
time step, averaging, or changes in computational techniques, among
the more prominent approaches.
In the second case, solution may be inefficient, but it may
ultimately converge. Unfortunately, in simulation of transient phenomena
these oscillations if damped slowly may musk the true nature of the
solution; resulting in a worthless simulation. In such cases, one must
seek means for accelerating convergence or removing oscillations from
the solution entirely.
In the third case, the mere fact of repetitive oscillation may
be indicative of a real phenomenon, for example, surge in a conduit
induced by tidal oscillations at an outfall location. One must be assured,
however, that the oscillations are pure, however, and not confused by
errors introduced by the numerics of the solution method.
Some simple examples of numerical instability in application of the
SWMM are illustrated in Figure 14 and 15. In the lower section of
Figure 14, water levels at a manhole vary widely and in more or less a
divergent manner when the time step of integration is 20 seconds. Reducing
the time step to 10 seconds allows a steady progression to a converged
solution (H = 1000.406 + 0.0002), as may be seen in the upper portion
of the figure. Similar, although more dramatic, is the experience in a
conduit connecting to the manhole. In the lower section of Figure 15
the flow is seen to vary from almost zero to more than 16"cfs with an
average of about 6.0 cfs. There is no distinctive pattern that suggests
convergence. Contrariwise, the fluctuations are rather random in ampli-
tude. However, when the time step is reduced to 10 seconds the flow
settles down rather quickly to a value of about 5.959 ± 0.004 cfs. Note
that while there are oscillations of about the same period as in the
lower plot, these tend to reduce in amplitude with time. Actually, they
are quite small; the horizontal line in the lower figure corresponds to
the trend line in the upper figure.
265
-------
•trtI ICMMCC* li«frftO, In*.
MiWlt CMIt, cutFU
MOM Milhtlt •CPU
I i«»dr fin- IIM.JI 'en
C«0««i Ulf* tlll.tl Hit
,~.«,
*tt* MM
im OB
ii
tfll.MI
IHI.7M
t!M.«4f
JMCTION
•tit tw
IIII.IH
|M«*4M
IW.4M
"""" """" /
! A. -b* o"
» / ^^
\ /
\ /
\ * •••»«•««•••»•«»•••.«•«•..•«•« ,
- i* **a»»^«»* lA4iJHKA4 I C?Qj 13T ^T"^F f
t / •••••
I / ••••
[ '.•«
t ••
t •
1 *•
I •
1 •
1 •
[
1 •
•
I t •
1 • • •. L , .«. _ ~ ~
; •. : At * IO e>cc,
! : /
i » •
i • •
i » •
i • *
i ti
i •
.1 ii.i 11.1 «.i ii.i ii.i ii.i t«.« n't nit n't
CtOCI ft«t (HOUM1
MUCTIO* NUHIU 411
INVCM VAt- lal«.l» fICT
CHOmi fit*- IIIU.lt FEET
MOUNO ILC«- IIII.lt FEE)
• • • I •
• •''• •• • * X. •
* / • • • •« •^
•' • « t » >
:V'f
*
«...'?
i'~ At«20*cc.
y
«. »*•
(*.• tl.« I*.* I*.I ll.l
ate* iim
It.l tl.t 11.1 11.1 II.J It.t
Cll*» 4MMM*
Figure 14. SIMULATED WATER LEVELS AT A
MANHOLE - TEST SYSTEM
266
-------
2 ••«***"* •••••••••«••
• •• • •• ••<• *•••••••••••••*•«•• i
•••*• • •• • •• •• »• •• •• •• •• •• • • #•** ••«
I M » M •* • • t • * • • • , . . • , ,_ t • t « t « .
I ** ** * * * * * * * •• • • .^^^fa^^K S^^f^f* •^^I"^T*^*^^^"I^^\
I •• ««•«••««• |. |J| \ i 1 «.».«»•««.
i ** M •! I* iZJ***'f^' ••••• •• ••
} Jj J** * *** * "
t '•'• •
1 M t
l.tM • M *
I» t * • • « • •
• */• » • •• •
> • • « •••• •• M •* M
'••• *• •• •• *• • •«
'••*«••••• *• • «•
I •»>«•«•• «* • •
• •/* * • •• • ** • •
I • f • • * •• •• «4 •
!•/>•<••• * «• •
!••*•••• • •
/•• • • *• ••
•.I
t
i
i
,i
n.*
• • « •* •• *
• • M ••
•«••••
** * *
A4 t
•* •
«• *
*• •
M
•
1 At«l
I • I f -JT • •
i*.i ti.» <•:• mi K.I u.i M,| ||;1 lt,l
ctoei itw i»euM)
0
— »:,
M
u.
H.i
t
{
I
t
f
c
r **^»***
•
••
**
»•
*•
••
* •
• • •
* * f t
• ** • *
* *• • •
• §• * •
• * ** * •
* • *• « •
•***•*
* *••*.•
* * • • • •
* *# »* »
* *
• *.
* •
• •
* *
* • *
• * * *
* •* *
* *
•
•
••
M
* •
• • 4
• *
• *
« *
• • * *
* •* •
*
* *
k« •• •
* «* *
• ** •
* * * *
• • • •
•*
•*
«4
^*
••
* *
* *
* «•
• *
• * *
* * *
*+ *
»* *
t ' •• • • t* •• •* • •
« • *• •• » • •
M H » «» • •
*« » M •• » •
• • • •• **
* • *» »»
• «« ••
• 20
11.*
aid TIM
-------
SUMMARY AND CONCLUSIONS
Application of Storm Water Management Models is somewhat more an
art than a science. There is considerable subjectivity involved and
sound hydrologic sense is essential to obtaining reasonable performance.
It is perhaps the general rule that there are no general cases, only
specific ones. Each requires particular attention.
Despite these broad reservations there are some fairly good models;
the EPA-SWMM seems to be one of the best. Examples we have chosen to
review here seem to indicate that with a little patience and this model,
one can obtain credible performance, that is; the model will simulate
most prototype events with reasonable reliability.
There seems to be no great advantage in carrying calibration and
verification to extremes, however, since the models have imbedded in
them many grand assumptions. If we are aware what these are and how
they may influence the model's performance, then we are in fair shape
to proceed with application.
As a final caveat let us remind ourselves—and others—that after
all, the EPA-SWMM is only a model; it is not the prototype. Let's not
expect it to do things it wasn't designed to do. The process of improving
its capabilities still goes on. After all, we modelers need to have some
new challenges ahead of us.
REFERENCES
1. Metcalf and Eddy, Inc., University of Florida, and Water Resources
Engineers, Inc., "Stormwater Management Model," Vol. 1, Final Report
to Environmental Protection Agency, July 1971.
2. Kibler, D.F.. et al, "Berechnung von stadtischen Kanalisationsnetzen,"
F. H. Kocks K.G. and Water Resources Engineers, Rep. presented at
Conf. on Sewer System Calculation Techniques, Dortmund Univ., Dortmund,
Germany, Dec. 1973.
3, Norton, W.R and S.^alrymple, "Documentation Report on Mathematical
Models-Seattle Drainage Study," Interim Rep. to Seattle District,
Corps of Engineers, July 1974.
4. Heeps, D.P. and R.G. Mein, "An Independent Evaluation of Three Urban
Stormwater Models," Rep. No. 4/1973, Civil Engineering, Monash
University, Melbourne, Australia, 1973.
5. Watkins, L.H., "The Design of Urban Sewers Systems," Dept. of Sci.
and Ind. Res., Road Research Laboratory, Tech. Paper No. 55, 1962.
6. Dept. of Civil Engineering, University of Cincinnati, "Urban Runoff
r Pollut1on Contro1 Research s*ries' EPA'
268
-------
Planning a Study Using Stormwater Management Models
By
Gerald T. Orlob*
INTRODUCTION
The great utility of the mathematical model as a tool for investiga-
tion lies in the facility with which one may expeditiously examine large
numbers of complex alternatives. It may serve for critique of designs,
sizing system elements, study of alternative routings, evaluating the
impact of off-line storage, testing control measures, and appraising
land and water use planning schemes. Given that the model can once be
made to represent prototype behavior with reasonable reliability, then
the user can apply it repetitively to a wide range of such alternative
plans or suggested solutions to specific Stormwater problems.
If the model package is well conceived and tested and the user follows
the user's manual (if it exists!) explicitly, he may anticipate a produc-
tive effort with only the "normal glitches" of computer usage to plague
him. However, things seldom run smoothly in the world of the computer and
the math model. If something can go wrong, it will; and we have sufficient
evidence to prove it.**
The purpose of this brief discussion on planning a study using
mathematical models is to highlight both the capabilities of models and
their fallibilities. We would like to begin by presenting a brief descrip-
tion of the model package, basically the Stormwater Management Model
developed under the auspices of EPA. In the course of this preliminary
discussion we'll try to point out where we may expect trouble and how to
avoid it, if possible. We shall see that in the face of certain problems
we may have to choose between the lesser of two evils, i.e., to make do
with a model of limited capability or to develop a new one. Either way
we must accept less than we would like to have. In the one case, we may
get a rough, cheap solution and in the other, a more elegant but costly
one. Which to choose?
We shall also look in some detail at the planning process itself,
that is, the technique for bringing the models up to speed and to bear on
the real problems to be solved. We'll pick this up somewhere along about
the middle of the model development process, at calibration, and carry
*Sen~ior Partner, G. T. Orlob & Associates, Orinda, California.
**
The first half of this statement is sometimes referred to as "Murphy's
Law. Murphy probably never saw a computer—the machine that surely
has contributed most to transforming his early maxim into "Law".
269
-------
forward to application. In this discussion we will see that the modeler's
work is never really done, it goes on forever! (But, that is the way we
like it—we need the challenge ahead!).
In the conclusion of this lecture we will point toward some actual
case studies. More to the point, perhaps we will set the spotlight on
some actual case difficulties, and what was done to solve them. These
will be discussed in more detail in the subsequent lecture,
A BRIEF DESCRIPTION OF STORMWATER MANAGEMENT MODELS
We have chosen to deal with stormwater management models in general
terms, rather than the specific of the EPA-SWMM. However, many of the
comments offered here apply equally well to the original model package
as to other versions, components* updates or so-called proprietary
packages. In all cases, the context will be that of urban stormwater,
although once again many comments are just as appropos of rural storm models,
The Urban Drainage System
The Urban Drainage System will be characterized by a high proportion
of impervious (or nearly impervious) surfaces. Moreover, it is a system
that will include man-made impervious pathways for guiding the flow of
water over the surface (curbs, gutters, lined channels, paved parking
areas, streets, etc.) and underground (storm, wastewater, and combined
sewers). The system includes all appurtenances that guide, control, or
otherwise modify either the quantity, rate of flow or quality of runoff
from urban drainage such as catch basins, storage basins, inlets, manholes,
sediment traps, wiers and outfall structures.
Figure 1 illustrates a somewhat simplified urban drainage system. In
this case, we view the system as an assemblage of subsystems dealing with
surface runoff, transport of flow and quality, and the receiving water.
Surface Runoff Subsystem
The Surface Runoff Subsystem is illustrated for our example in
Figure 2 which depicts the drainage area tributary to a sewer inlet as a
system of surface elements (the rectangles), gutters (the dotted line)
and drainage ditches (the dashed line). For the Surface Runoff Module of
the EPA-SWMM the following data are required for each catchment:
1. landuse (5 classes) 5. average watershed slope
2. total surface area, acres 6. average street (gutter) slope
3. total length of streets (gutters) 7. point in transport subsystem
100's feet to which the subarea drains
4. percent imperviousness
Input to the subsystem is comprised of rainfall that may be described
in terms of an intensity-time graph derived from direct measurements in
the watershed and applied to each subarea in accordance with the storm's
270
-------
SURFACE RUNOFF SUBSYSTEM
TRANSPORT
SUBSYSTEM
RECEIVING WATER
SUBSYSTEM .'/,
Figure 1. The Urban Drainage Subsystem
271
-------
precip.
intensity
time
OVERLAND
FLOW
SYSTEM
I
o
o-
INPUT
pollutant
load
0
b //?//
time
OUTPUT
time
time
Figure 2. Surface Runoff Subsystem
272
-------
pattern. The inset In the upper left of Figure 2 illustrates a typical
rainfall hyetograph. Values are given in inches per hour, usually for
every five minutes of the storm.
Within the subsystem a certain mass of a quality constituent
(pollutant) may exist at the outset of the storm and be delivered by the
flow at mass rates and concentrations that may depend on the nature of
the storm, the character of the surface, and the sources of the pollutant
(upper right insert, Figure 2). Data are provided on rates of pollutant
accumulation, pounds per 100 feet gutter per day (up to 17 types) and
pounds of pollutant per pound of dust and dirt delivered to the watershed.
The overland flow process modifies the rainfall hyetograph by infil-
tration, surface retention and transient storage, so that at the inlet one
observes a much modified inlet hydrograph, a temporal description of inlet
flow (lower left insert, Figure 2). In addition, the combined flow and
quality processes produce an inlet pollutograph, a time-concentration graph
of a particular pollutant as it leaves the Surface Runoff Subsystem and
enters the waste water conveyance system (lower right insert, Figure 2).
These two graphs, one of flow and the other of quality, comprise the output
of the Surface Runoff Subsystem and are input to the Transport Subsystem.
Transport Subsystem
The Transport Subsystem is comprised of the physical works for convey-
ing stormwaters and their associated pollutant loads from all of the inlets
in the system through a network of underground conduits to a point (or
points) of disposal. Enroute, flow and quality are both modified by accre-
tions to the system from other tributary areas and/or point sources of
pollution. In addition, flows and pollutant concentrations are attenuated
1n pass.ing through the system, the degree of modification depending on
such factors as system storage, "off channel" storage, phase relationships
of inflow hyetographs and pollutographs and certain hydraulic properties
of the system. The two lower inserts in Figure 3 illustrate a typical set
of outputs from the Transport Subsystem, a hydrograph and a pollutograph,
that in turn become inputs to the receiving water.
Receiving Water Subsystem
The Receiving Water Subsystem may be a stream, a lake, an estuary or
a coast. Discharge into an estuary will be used for illustration.
The impact of the discharge on the estjary will probably be assessed
in terms of the concentration of a particular quality constituent: its
distribution in space, its persistence in time, and its frequency of
exceedance of a certain critical level. For a given hydrological event,
the system may be observed synoptically (at the same instant in time) or
temporally (at the same point in space).* One gives the distribution in
space, the other the persistence in time. From the standpoint of quality
*These are sometimes referred to as the Lagrangran and Eulerfan viewpoints,
respectively.
273
-------
INPUT
time
time
time
OUTPUT
time
Figure 3. Transport Subsystem
274
-------
management both viewpoints are usually required for each hydrological event.
To obtain frequency of exceedance of a critical level the impact on the
receiving water must be observed a "statistically significant" number of
times. Figure 4 illustrates typical responses ("impacts") for an estuary.
TECHNICAL APPROACH TO MODEL APPLICATION
Application of models in real problem solving involves some necessary
preliminary steps that really begin with the conceptualization of the model
itself. For the purposes of our discussion here, however, we may assume
that we already have a tangible operating model in hand and that the
preliminaries of conceptualization, formulation, and computational represen-
tation as outlined schematically in Figure 5 have already been completed.
We will pick up the process at this point and describe briefly what is
entailed in each of the steps of calibration, verification, documentation,
and application.
Calibration
Calibration entails adjusting empirical relationships in the model to
ensure replication of in situ observation of the state variables under known
boundary conditions. This exercise is usually performed for a case consid-
ered representative of those most likely to be explored in model application.
The process is primarily subjective, involving adjustment by the
modeler of empirical coefficients during a succession of applications of
the model until the model response is in agreement with prototype data.
The degree of conformity between model and prototype that constitutes
"agreement" is generally judgmental, since many factors, some outside of
the modeler's control, influence results. Experience plays an important
role here.
A related part of this effort involves sensitivity testing of the
model, i.e., evaluating the model's response in terms of selected state
variable to changes in coefficients, boundary conditions, input data, and
sometimes even the structure of the model itself. A product of calibration-
sensitivity testing is a model suitable for application as a predictive or
management tool.
Verification
Verification is the modeler's demonstration of the model's predictive
capability—the first attempt to apply the model to a new set of conditions
and to generate responses that are representative of the prototype. Taking
the model as calibrated to known data, the modeler attempts to simulate an
unknown (to him) system response. The test of predictive capability is
whether the model can produce results that are within the range of accept-
ability used in calibration.
Sometimes in this step a recalibration is indicated; then, the processes
of calibration and verification are repeated until the results are accept-
able. A final product is a model ready for documentation.
275
-------
INPUT
time
RECEIVING WATER
SUBSYSTEM
time
ISO-CONCENTRATION
LINES
OUTPUT
critical
level
time
Figure 4. Receiving Water Subsystem
276
-------
GOALS & OBJECTIVES
prototype
descriptions
space and
time scales
solution
techniques,
hardware
test case data
coefficients (
boundary conditions
verification data
test case
results
prototype
data
FUNCTIONAL
REPRESENTATION
COMPUTATIONAL
REPRESENTATION
functional model
CALIBRATION
calibrated model
VERIFICATION
verified model
DOCUMENTATION
documented model
APPLICATION
A
MODEL
SENSITIVITY
USER
MANUAL
feedback to
improve model
Figure 5. The Processes of Mathematical Model Development
Conceptualization to Application
277
-------
Documentation
Documentation entails complete description of the model, its concep-
tual foundation, formulation, computational representation, data formats,
input-output requirements and options, and example problems with solutions.
Usually, the program is documented separately in detail, providing flow
diagrams, variable descriptions, characteristic dimensions and a program
listing and is accompanied by a Users' Manual that describes how to apply
it to a new case and gives a demonstration example. This has been accom-
plished for the EPA-SWMM under the original development contracts (1).
Under a continuing program of new development, improvements in the model
will also be documented and the Users' Manual revised from time to time.
Application
Application is a continuing process in model development that assures
gradual improvement through actual usage. No model is infallible; when
difficulties are encountered there is no assurance that the model, itself,
is infallible. After all, it is only a "model," not the prototype itself,
and should always be subject to reexamination and possible improvement.
Thus, the model development process is not closed, but rather it is
open-ended and requires that the user and. model developer continually re-
examine the model, the problems it is being asked to solve, and the data
provided. This should result in progressive improvement in capability to
simulate the prototype's behavior, commensurate with practical need and
available resources.
SOME COMMENTS ON DATA
Calibration, verification, and! application require that reliable
information be obtained from the prototype. Often data on actual storm
events are difficult to obtain or may not exist at all for the
drainage in question. Invariably, the raingage lies outside the actual
drainage basin, even miles away. We must somehow relate precipitation data
for this gage to gaged outflow from the basin. The usual hydrologic tech-
niques of inter-correlation between stations or construction of isohyetal
patterns over the larger area that includes our basin may be used to
develop the necessary rainfall-runoff relationship.
In this process, attention must be given to the temporal detail of
precipitation data. If the basin is small, say under 100 acres or so,
then it will not suffice to have hourly readings; we will need data at
5-minute spacings, surely not less than 10-minute intervals. If data are
too coarse we will lose definition of the storm and may not be able to
calibrate the model closely enough to allow description of runoff peaks.
278
-------
Measurement of the flow itself presents many difficulties. Of
course, what we measure will be the net contribution to surface runoff
after satisfying interception, depression storage, and infiltration.
Since we have no direct way of measuring these other quantities we must
distribute the aggregate difference between our uncertain input (precipi-
tation) and our measured output arbitrarily to each. Fortunately, the
functional relationships are distinctive; interception and depression
storage may be considered as constant amounts of precipitation while
infiltration varies throughout the storm in some unique fashion, e.g.
exponentially decreasing from some initial value. These assumptions
(that may not always hold) help us to make some reasonable judgments
about the amount of the flow that is not measured as runoff and how it
is distributed temporally. The adjustment process, that of reconciling
measured rainfall and gaged runoff, is basically subjective. There is
really no substitute here for a basic understanding of the hydrologic
phenomena and experience in their interpretation.
ROLE OF MODELS IN DECISION-MAKING
A model may be looked upon as an adjunct 'to decision-making, a tool
that when wisely applied in the assessment of alternative strategies can
enhance the decision makers' own judgment. It is really a representation
of the real world that permits the user to test his ideas, his preferences
or some conditions of prototype behavior before the fact of change. He
can make observations at low cost, quickly, and without permanent altera-
tion of the real world system. He can, with the aid of a properly conceived
model, explore interactions of such complexity that intuitive assessment is
impossible.
This is very much the case for the Storm Water Management Models.
Once the models are calibrated and verified, providing the user with
some appreciation of their reliabilities, and fallibilities, the models
can be employed best to discriminate between alternative choices. For
example, he may wish to test the effect of changing a land use pattern
on the runoff peak and time of occurrence. Intuitively, he would expect
that if he adds more impervious surfaces, the peak will rise; yet he may
not be able to say much about the magnitude of the peak and when it
^11 occur, except with the aid of the model or its equivalent. He can
Jlso test the effect of the area! distribution of the added increment of
impervious surface. Where should it be placed to produce the least impact
°n the basin's hydrologic response? Such questions can be answered in an
°^ganized program of model application.
If we may assume that we have in hand a reasonably good model, we can
now turn to a closer examination of some of the problems in applying it.
In the next lecture we will examine some actual case studies where such a
toodel has been employed in prediction of urban runoff events and where
certain common problems have been encountered and solved.
REFERENCES
!• Metcalf and Eddy, Inc., University of Florida, and Water Resources
Engineers, Inc., "Stormwater Management Model." Vol. 1, Final Report
to Environmental Protection Agency, July 1971.
279
-------
THREE CASE STUDIES ON THE
APPLICATION OF THE STORM WATER
MANAGEMENT MODEL
By
Jekabs P. Vittands*
CASE STUDY - 1 SEWER SYSTEM MODELING AND
IMPROVEMENTS IN THE EASTERLY DISTRICT, CLEVELAND,
OHIO
Introduction
The Easterly District of the City, shown on Figure 1,
totals approximately 16,000 acres and includes the downtown
area, with an additional 25,000 acres tributary from
surrounding areas outside the City.
The existing sewerage system within the Easterly
District is almost entirely combined. Tributary areas out-
side of the City use City sewers and drains for conveyance
of drainage to downstream water courses.
Flooding problems due to inadequate sewer capacity
were a major concern of the project. During the eight months
from January 1 to August 31 in 1972, over 560 complaints were
registered for cellar flooding, while another 168 complaints
were registered for pipe cave-ins. These pipe failures, or
cave-ins, are due to the existence of a large proportion
of old brick sewers, many of which are over seventy
years old. The locations of these flooding and cave-in
complaints are shown on Figure 2.
Because of the inadequate capacity and condition of
many of the sewers, a program was started for the systematic
evaluation of the existing system, and preparation of a
plan and program for the reconstruction of the inadequate
segments of the sewer system. This project was the first
step towards solving flooding problems in the Easterly
District of the City, with detailed design and construction
as the next steps.
The specific purposes of the project were to determine
the improvements needed to eliminate flooding problems via
computer modeling of the sewer system in the Easterly
*Project Manager, Metcalf & Eddy Engineers, Boston, Massachusetts.
280
-------
cx>
LEGEND
— LIMIT OF EASTERLY
SEWER DISTRICT
--- SURBURBAN DRAINAGE AREA
TRIBUTARY TO THE CITY
DRAINAGE AREA NUMBER
SCALE IN MILES
FIG. 1 DRAINAGE AREAS, EASTERLY DISTRICT, CLEVELAND, OHIO
-------
ro
CO
FIG. 2 LOCATION OF FLOODING AND CAVE-IN COMPLAINTS - CLEVELAND, OHIO
-------
District of the City, and to identify orders of priority for
necessary improvements along with estimated costs. Another
purpose of this project was to install this modeling
capability at the City by providing computer programs and
training City personnel in the use of the model.
Recognizing that in combined sewer systems the
elimination of flooding conditions is only one part of the
problem and that, in addition, remedies for overflow pollu-
tion will have to be determined in the future, consideration
was given to this need in the modeling of areas using the
Storm Water Management Model (SWMM). Only the parameters
necessary for the quantification of pollutants will have to
be added to the data in order to use the Storm Water
Management Model as an aid in determining alternative
remedies for combined sewer overflows.
The approach to the project was to use the appli-
cable computer programs of the SWMM to carry out studies
for necessary improvements. The two major blocks of the
programs used were the RUNOFF Block and the TRANSPORT Block
which were modified to permit their more effective use in
this project. The EXECUTIVE Block was used in addition to
provide its function of interprogram coordination.
The basic technique used was one of design by analy-
sis wherein computer simulation is used to determine the
conditions in a sewer system during a storm. An actual or
selected design rainfall is assigned to a drainage area
with a specific sewer configuration and the flow in sewers
and flooding conditions are determined by simulation.
In this project, first, the flooding conditions and
transport capabilities in the existing system were deter-
mined by computer simulation. Then, based on these results,
the necessary sewer sizes were estimated to eliminate
flooding under the design storm conditions. Computer simu-
lation was then used to test the adequacy of the selected
improvements. In some cases, a number of trials were
necessary in order to arrive at the solution.
All sewers 30 inches in diameter (or equivalent in
size) and larger were modeled explicitly. In addition, in
certain areas where the nature of the tributary area is
complex, smaller pipes were also included. These models
were referred to as large models and an example is shown on
Figure 3 for Drainage Area No. 3.
Pipes smaller than 30 inches in diameter were
analyzed on a tributary area basis. These were referred
to as fine models. In this case, selected typical areas
283
-------
ro
CO
< 'V
#* S V
FIG. 3 LARGE MODEL SUBDIVISION - CLEVELAND, OHIO
-------
were modeled In order to generate tables for flow quantifi-
cation which could be used to determine the required size
of sewers in the area on the basis of tributary area
characteristics such as size, slope, and percent impervious-
ness. This approach Is discussed further under the section
entitled Small Model Development.
Design Criteria
This section discusses only design criteria of the
Project pertinent to the use of the SWMM.
The criteria established by the City for the degree
°f protection to be provided was as follows:
1. Downtown areas - 25-year storm
2. Industrial and commercial areas - 10-year storm
3. Residential areas - 5-year storm
The design storms used In each of the drainage areas
a**e shown on Figure 4.
Hyetograph Development. In using simulation as a
technique for flow determination, time-varying parameters
are required. On this basis, rainfall data representative
°f actual conditions are necessary, rather than using time-
averaged rainfall values which are then multiplied by a
constant to obtain peak runoff as is done in the Rational
Method.
The procedure used In developing typical rainfalls
the above three design storms is as follows:
1. Select the desired range of frequencies and dura-
tions for the storms, i.e. 5-, 10-, and 25-year
storms and 1- and 6-hour durations. Durations
of 1 and 6 hours were Initially judged to
represent the range of storm durations appli-
cable to the drainage area sizes in the Easterly
District.
2. Using the US Weather Bureau Technical Paper
No. UO, select the appropriate isopluvial maps
corresponding to the particular storm frequencies
and durations. Locate the study area on the maps
and record the rainfall depths, interpolating
between isopluvial lines. This -rainfall depth
Information is presented in Table 1. Also shown
are the average rainfall intensities corresponding
to the depths and durations.
285
-------
LAKE
ERIE
L E GEND
5 YEAR FREQUENCY
10 YEAR FREQUENCY
25 YEAR FREQUENCY
— DRAINAGE AREA BOUNDARY
I DRAINAGE AREA NUMBER
FIG. 4 DESIGN STORM FREQUENCIES, EASTERLY DISTRICT, CLEVELAND, OHIO
-------
For rainfall durations less than 30 minutes,
isopluvial maps do not exist. However, Techni-
cal Paper No. 40 provides constants for con-
verting depths at 30-minute durations to depths
at 5-, 10-, and 15-minute durations.
3. Select the desired time increments for each storm
duration. A 5-minute increment was chosen for
the 1-hour storm and a 10-minute increment was
chosen for the 6-hour storm. This was judged to
best describe the storm characteristics and not
consume large amounts of computer time.
4. Select a storm pattern. Special consideration
must be given to locating the rainfall peak at a
point from beginning of storm that is representa-
tive of storms in the Cleveland area. Figure 5
was used as a basis for this. Figure 5 shows the
average distribution of rainfall and the time of
maximum rainfall intensity from beginning of
storm.
5. The design hyetographs for a storm of a given
frequency and duration were prepared as follows:
a. A bar graph was drawn with TIME in minutes
plotted on the x-axis and RAINFALL DEPTH in
inches plotted on the y-axis.
b. The point in time was located where the maxi-
mum rainfall intensity will occur. For a
1-hour storm duration, this point occurs
23 minutes after the start of the storm
according to Figure 5.
c. At the point of maximum intensity, the rain-
fall depth corresponding to the selected time
increment was plotted. For a 1-hour storm
with a 5-minute time increment, the rainfall
depth is 0.44 in., while the rainfall depth
for a 10-year frequency storm is 0.49 in.
287
-------
u.
U.
O
I-
tii
U
oc
I
UJ
O
100-
80 *"
on ,
POINT
RAINF-
^
DF MAXII\
\LL INTEf
/
IUM
MSITY-^
/
"t°
0
c
1
L
-«^^^
/
/
-i
0
0
y
1
i
/
G^j
f
$>
NOTE: D
01
ft
Jl
Er
^
X
i\TA FOR TH
JTAINED FR(
-EVE LA NO, C
?R POLLUTIC
INE 27, 1968
PERSON, COI
^
S DIAGRAM
3M FIG. 2-12,
)HIO. MASTE.
?N ABATEME
t>y HAVENS A
1SULTING Er
l
^*
WAS
PART 2,
RPLAN -
NJ,
NO
IGINEERS.
00 40- 50 60 70
DURATION AS PERCENT OF TOTAL RAINFALL
80
90
100
FIG. 5 CLEVELAND, OHIO, RAINFALL MASS DIAGRAM
288
-------
TABLE 1. RAINFALL DEPTH AND INTENSITY - DURATION -
FREQUENCY RELATIONSHIPS, CLEVELAND, OHIO
Rainfall depths in inches and intensities in inches
per hour for various frequencies
1 year 5 years 10 years 25 years
Duration in. in./hr in. in./hr in. in./hr in, In./hr
5 min.
10 min.
15 min.
30 min.
1 hr.
2 hr.
6 hr.
.28
.43
.54
.75
.95
1.13
1.50
3.
2.
2.
1.
0.
0.
0.
36
58
16
50
95
56
25
.44
.68
.86
1.19
1.50
1.65
2.3
5
4
3
2
1
0
0
.28
.07
.44
.38
.50
.83
.38
.49
.76
.96
1.34
1.70
2.10
2.65
5.
4.
3.
2.
1.
.1.
0.
88
55
84
68
70
05
44
.58
.90
1.14
1.58
2.0
2.2
2.95
6.96
5.40
4.56
3.16
2.0
1.1
0.49
d. Using the location of the hyetograph peak as
a starting point, rainfall depths were
plotted for each time increment corresponding
to the storm pattern indicated on Figure 5.
The rainfall depth for any given time period
should not exceed the maximum depth indicated
in Table 1 for a similar period.
e. The ordinate of the above graph should then
be converted to rainfall intensity in inches
per hour to produce the design hyetographs.
Figure 6 shows the design hyetographs for 5, 10, and
25-year storms of a 1-hour duration. It should be noted
that these hyetographs do not represent actual storm events,
but synthetic storms fulfilling the U. S. Weather Bureau
statistics on rainfall depth-duration-frequency and the
average rainfall distribution pattern in the Cleveland area.
Similar hyetographs were developed for storms of a
6-hour duration. Both the 1-hour and 6-hour duration storms
were tested against a selected large drainage area in the
Easterly District. Area No. 8 was used to determine Improve-
ments necessary under both storms. Area No. 8 was selected
on the basis of its size. From this it was found that there
were no marked differences in the results allowing selection
289
-------
8.0
7.0
6.0
I 5.0
>
H
z
4.0
< 3-°
DC
2.0
1.0
5MIN. INCREMENTS
fiflR
P
1.<
0.36
0.12j|
J ^^
!.28
I
TOTAL RAINFALL = 1.50"
H FREQUENCY- R YFARS
II DURATION = 1 HOUR
J PEAK = 5.28" PER HOUR
.88 ss? o
fOf^ \rntm
- . .
III
76
2.02
ll
§11.32
o.,l
3.:
30
I*K!I
[
24
WP»
ixW
>.8i
H
•
TOTAL RAINFALL -1.70"
FREQUENCY = 10 YEARS
DURATION = 1 HOUR
PEAK = 5.88" PER HOUR 3 p J
3.12
2.
—,
B4
1.80
1 1
llJj, O.J]
1
io-SiS
]
m
1
^
• v
3.
1
Li
i
TOTAL RAINFALL = 2.00"
FREQUENCY = 25 YEARS _
DURATION = 1 HOUR
PEAK = 6.96" PER HOUR
30
3.00
11 2.04
fl-1 1.08
:..... •.-,.•..!: IH>>1
J' 4 0.48
0 10 20 30 40 50 60
0 10 20 30 40 50 60
RAINFALL DURATION (MIN)
0 10 20 30 40 50 60
FIG. 6 5,10 and 25-YEAR 1-HOUR DESIGN STORMS, CLEVELAND, OHIO
-------
of the 1-hour duration storms as a basis for design.
Figure 6, therefore, represents the design storms used in
this project.
Land Use and Developments. For design of facilities
for urban runoff, it is necessary to measure the degree of
development in terms of the extent of roofs and pavements,
and other impervious surfaces in an area tributary to a
sewer. The data used in this report is shown on Figure 7.
Hydraulic Parameters. The following criteria were
used in the Storm Water Management Model:
Surface retention storage
1. impervious areas: 1/16 in. (inches)
2. pervious area: 1/4 in.
Ground infiltration as used in Norton's Equation
1. maximum rate: 3.0 in/hr. (inches per hour)
2. minimum rate: 0.52 in/hr.
3. decay rate: 0.00115 In/sec.
Flow friction factors as used in Manning's Equation
A. Overland flow
1. impervious surfaces: 0.13
2. pervious surface: 0.25
B. Existing Sewers
1. Cast or ductile iron: 0.015
2. Concrete and vitrified clay: 0.015
3. Brick or block: 0.016
4. Corrugated metal: 0.022
C. New Sewers (Concrete)
1. Less than 30 in. in diameter: 0.015
2. 30 in. to 84 in. in diameter: 0.013
291
-------
LAKE
ERIE
ro
DATA FOR THIS DIAGRAM WAS
OBTAINED FROM FIG. 2-14, PART 2,
CLEVELAND, OHIO, MASTER PLAN
FOR POLLUTION ABATEMENT.
JUNE 27, 1968 BY HAVENS AND
EMERSON, CONSULTING ENGINEERS
FIG. 7 IMPERVIOUS AREAS, EASTERLY DISTRICT
-------
3. Greater than 84 in. in diameter: 0.011
D. Culverts
1. Existing: 0.015
2. New: 0.011
Velocities
1. minimum: 3 fps (feet per second)
2. maximum: 15 fps.
Infiltration and Dry-Weather Flow. Sample calcula-
tions show that both infiltration and dry-weather flows
have negligible hydraulic significance on combined sewers
during design storm conditions and may be neglected if
pollution aspects are not being considered.
For example, in Area No. 8 inclusion of infiltration
would increase flows at the downstream end by less than
0.2 percent.
Similarly, selecting Subcatchment No. 1 in Area No. 8
inclusion of dry-weather flow would increase design flows by
under 2 percent.
Allowance for Inflow from Suburban Tributary Areas.
It was necessary to develop hydrographs of inflow in certain
locations where inflows from suburban communities were a
significant factor in the analysis of combined sewers.
Two approaches were used to determine the quantity of
dry-weather inflow from the Monticello Boulevard Sewer. The
first method was based on tributary area, population and
water use. The second method was based on the assumption
that the peak inflow is equal to the capacity of the inflow
Pipe. The greater of the two values was used as a constant
inflow in the determination of improvements.
In the modeling of those enclosed conduits which
transport drainage from suburban areas, it was necessary to
make allowances for such contributions. For each, a peak
inflow and time to peak was determined from previous studies.
An inflow hydrograph was developed using this peak discharge
and, the shape of hydrographs generated from areas with
similar basin characteristics that had been previously
modeled. These inflow hydrographs were incorporated in each
of the applicable drainage areas.
293
-------
Model Demonstration
For the purpose of demonstrating the application of
the SWMM the rainfall and flow during such rainfall
from Areas 8 and 13 was gaged and such conditions were
then simulated by computer for comparison.
Area No. 8 Data. The pertinent data for this area
is as follows:
Size: 692.4 acres.
Impervious area proportion: 55 percent.
General ground slope: 0.3 percent
Total length of major sewer: 12,000 feet.
Size of downstream sewer: 105 inch circular brick
sewer.
Capacity of downstream sewer: 660 cfs.
Area No. 13 Data. The pertinent data for this
area is as follows:
Size: 376.0 acres.
Impervious area proportion: 60 percent.
General ground slope: 0.35 percent.
Total length of major sewer: 7,200 feet.
Size of downstream sewer: 87 inches.
Capacity of downstream sewer:' 285 cfs.
Rainfall Data. The precipitation applicable to
flows measured in areas 8 and 13 was recorded by rainfall
gages located near the drainage areas. These gages recorded
the cumulative rainfall every ten minutes providing data
for the complete rainfall hyetographs shown on Figure 8
As shown, there is a large variability in rainfall between
locations for these recorded events. The input rainfall
for Area No. 8 is a weighted average of the rainfall records
recorded at both gages. The rainfall recorded at the East
105th Street Gage was used directly as input to Area No. 13.
The input hyetographs for each area are shown on Figures 9
and 10, respectively. &
Flow Measurements. Each gage and recording meter
was located in a selected manhole where it measured and
recorded the depth of flow at set time intervals. The chart
had a full scale reading equal to 60-inches permitting the
recording of flows from zero depth (pipe invert) to a depth
of 60-inches. This recorded depth was converted to a flow
rate by assuming that a condition of uniform flow exists,
therefore, using Manning's Equation for such conversion with
a pipe friction factor (n) of 0.016 conversion vn.™
294
-------
0.10
0.08
SJ 0.06
HI
Z
— — — EkST 105th STREET RAINFALL DATA
3090 BROADWAY RAINFALL DATA
0.04
ro
0.02
2010
2100
2200
2300
2400
0100
0200
0300
6-23-73-
•6-24-73
TIME
0400
FIG. 8 RAINFALL HYETOGRAPHS, TEST AREAS, CLEVELAND, OHIO
-------
LU
.04 -
<
cc
.08 ->
Q
1
CC
LU
O.
O
m
3
0400 0600 0800
10
TIME
FIG. 9 MEASURED AND SIMULATED FLOWS AREA NO. 8, CLEVELAND, OHIO
296
-------
z
3 .04 H
_i
LL
Z
.08 J
O
o
u
LLJ
CO
tc
LU
Q.
HI
111
LL
O
m
U
z
i
10
2010
0800
TIME
FIG. 10 MEASURED AND SIMULATED FLOWS EAST, AREA NO. 13, CLEVELAND, OHIO
297
-------
Model Comparison. The SWMM was then used to simu-
late conditions during the recorded rainfall events of
July 23rd and 24th, 1973. These results are also shown on
Figures 9 and 10 for Areas No. 8 and 13s respectively.
The agreement between computed and observed values
appears reasonable for Area No. 8, but is questionable for
Area No. 13. Some explanations for these variations may be
as follows:
Two operational problems that occur with the initial
flow measurements Involve: 1) the sudden recorded rise in
peak flow above the normal flow; and 2) the recording of the
correct time the rise in flow occurs. The first problem has
been attributed to the turbulence accompanying rapid rise in
the level of the flow in the pipe. An example occurs in
Area No. 8 during the evening storm of June 24th. The
sudden rise in the level of the flow from a Manning Chart
reading of 4 to a reading of 7 was attributed to such
turbulence and omitted in the analysis of the data. The
second problem, proper time,, can be caused by an inaccurate
starting time or inaccurate clock speed. An example of this
occurs at Area No. 13 during the evening storm of the 24th of
June. The recorded flow peaks almost an hour before the
rain gages record any precipitation. This was adjusted
also on Figure 10,
Precipitation variations can cause errors between
observed and computed data especially during small rainfalls,
such as those measured. This problem is more pronounced in
this case where the gages were not located within the
drainage area. The actual rainfall distribution, duration,
and intensity in Area No. 8 and Area No. 13 may not
have been represented by the rain gage data from the
stations at East 105th Street and 3090 Broadway as demon-
strated by the variability of the two storms shown on
Figure 8.
Another area requiring consideration when interpreting
the observed and computed data is the nature of the
individual storm events. The storms of June 23 and 24 were
small storms which can be severly affected by surface
storage, infiltration and other drainage basin characteris-
tics which would be dampened in a larger storm. In the
case of the storm on the evening of June 23, the maximum
intensity occurred during the first time increment (10
minutes) and, because of its small value, it is sensitive
to antecedent conditions and surface peculiarities.
298
-------
Fine Model Development
As mentioned earlier, evaluation of the sewers in the
Easterly District was divided into a two-part approach con-
sisting of the computer modeling of all pipes 30 inches and
greater, and the analysis by representative area of pipes
smaller than that.
The general approach for evaluation of the smaller
pipes was to generate a set of tables which could be used
to determine the peak-design flows tributary to a small
pipe. The parameters pertinent to this are: area size,
slope and shape; percent impervious area; and design storm
frequency.
Since all upstream sewers were made adequate in this
analysis, the condition of a limiting sewer capacity in a
specific reach did not exist. On this basis, sample areas
were selected to provide the basis for selecting typical
areas tributary to a small sewer system. Due to the
generally orderly nature of the streets in the City, the
shape was found generally rectangular with a uniform
distribution of sewers.
A study of available topographic data showed that
most of the Easterly District is relatively flat with a
ground slope under 2.0 percent averaging at about 0.5 percent
However, there were isolated cases within the City where
slopes exceed these values. These areas have slopes between
5 and 10 percent and were considered separately.
Design Table Development. Several areas were
selected and modeled to generate typical peak flows for
several areas under various design storms. A part of Area
No. 14 was finally used as the basis for generating the
design table data.
The model of this area was developed to a considerable
detail in order to represent actual conditions accounting
for flows on the ground, in the gutters and in pipes. A
City ordinance requires that all roof drains be connected
directly to the sewer system. On this basis, the runoff
from roofs was modeled to go to the sewers in that street,
With runoff from other areas going to the gutters in that
street and then via the gutters to the next downstream sewer
inlet. The fine model of this area was then developed to
cover the subareas for each reach. It was estimated that
30 percent of the area consists of roofs draining this
impervious area directly into the sewers (Due to computer
Program restrictions 95 percent imperviousness was used)
299
-------
with runoff from the remaining 70 percent entering gutters,
and proceeding to the downstream manhole of the reach.
Allowance of 30 percent of the area as representing
roofs was adhered to for all cases where the overall percent
of imperviousness was over 30 percent.
Various degrees of percent imperviousness (5, 30,
65, and 95 percent) were selected and run in the fine model
analysis of the sample area for the 5, 10, and 25-year
frequency storms. Straight line interpolation was used to
compute all intermediate points producing Tables 2 through
4.
A formula was used to separate roof areas from ground
areas while retaining various overall percent impervious-
ness. For example, if an area was determined to have an
overall percent imperviousness of 70 percent, exclusion of
30 percent of that area covered by roofs would make the
remainder of the area have an imperviousness of 60 percent.
It is noted that for 0.50 acres and less, between
30 percent and 95 percent imperviousness, the tables show
the runoff to be the same. Areas this small only represent
the roof runoff contributing to the upstream reach and are
always modeled at 95 percent imperviousness.
According to the assumption of no roof drains under
30 percent imperviousness this does not apply to less
developed areas. Here it was assumed that 30 percent of
its area at the actual percent imperviousness goes directly
into the upstream manhole and that the other 70 percent goes
into gutters and then into the downstream manhole.
Fine Model Verification. After completing the
tables for the fine model analysis, a section of Area No. 13
was subdivided and tributary flows were determined. These
flows were then checked against flows developed by actually
modeling the same area. The results were found to be in
general agreement as shown in the following table.
Representative area method Computer method
Expected Q (cfs) Expected Q (cfs)
3.69 3.57
9.32 8.68
13.00 H.5!
2.55 2.54
6.44 5.63
21.46 19.31
2.18 2.19
28.68 25*.87
300
-------
TABLE 2. FINE MODEL ANALYSIS - 5 YEAR DESIGN STORM, FLAT AREAS
AREA
(ACRES)
5.60
5.70
5.80
5.90
6*00
6*10
6.20
6.30
6*40
6.50
6*60
6.70
6.80
6.90
7.00
7.10
7.20
7.30
CO 7«*°
0 7.50
«-• 7.60
7.70
7.80
7.90
8.00
8.10
a. 20
8.30
8.40
8.50
8.60
8.70
8.80
8.90
9.00
9.10
9.20
9*30
9.40
9.50
9.60
9.70
9*80
9.90
10.00
5
1.68
1*72
1.76
1.79
1.83
1.87
1.90
1.94
1.98
2.02
2.05
2.08
2.11
2.14
2.17
2.21
2.24
2.27
2.30
2.33
2.37
2.40
2.43
2*46
2*50
2.55
2*59
2.64
2.69
2.74
2*79
2*84
2*89
2.94
3.00
3.05
3.09
3.14
3.19
3.24
3.29
3.34
3.39
3.44
3*50
10
2.99
3.05
3.10
3.15
3*20
3.25
3.30
3.35
3.40
3.45
3.50
3.55
3.60
3.65
3.70
3.7S
3.80
3.85
3.90
3.95
4.00
4.05
4.10
4.15
4.20
4.27
4.35
4.43
4.51
4.59
4.67
4.75
4.83
4.91
5.00
5.05
5.tl
5.17
5.23
5.29
5.35
5.41
5.47
9.53
5.60
n,;
15
4.30
4.37
4.43
4*50
4.56
4.63
4.69
4.76
4.ft2
4.89
4.95
5*02
5.09
5*16
5*22
5.29
5.36
5*42
5.49
5*56
5.63
5*69
5.76
5.83
5.90
6*00
6.11
6*22
6.33
6.44
6.55
6.66
6.77
6.88
7.00
7.06
7.13
7.20
7.27
7.34
7.41
7*48
7.55
7.62
7.69
SUM/
20
5.61
5.69
5.77
5.85
5.93
6.01
6.09
6.17
6.24
6.32
6.41
6.49
6.58
6.66
6.75
6.83
6.92
7.00
7.09
7.17
7.26
7.34
7.43
7.61
7.60
7.73
7.87
8.01
8.15
8.29
8.43
8.57
8.71
8.85
9.00
9.07
9.15
9.23
9.31
9.39
9.47
9.55
9*63
9.71
9.79
LTCiLI
25
6.92
7,02
7.11
7.20
7.30
7.39
7.48
7.57
7.67
7.76
7.86
7.96
8.07
8.17
8.27
8.37
8*48
8.58
8.68
8.78
8.89
8.99
9.09
9.19
9.30
9.46
9.63
9.80
9.97
10.14
10.31
10.48
10.65
10.82
11.00
11.08
11.17
11*26
11.35
11.44
11.53
11.62
11.71
11.80
11.69
f L.UW
30'
8.23
8.34
8.45
8.55
8.66
8.77
8.87
8.98
9.09
9.20
9.31
9.43
9.55
9.67
9.79
9.91
10*03
10.15
10.27
10.39
10.51
10.63
10.75
10.87
11.00
11.20
11.39
11.59
11.79
11.99
12.19
12.39
12.59
12.79
13.00
13.10
13.19
13.29
13.39
13.49
13.59
13.69
13.79
13.89
14.00
t<-*»J
35
9.32
9.45
9.58
9.71
9.84
9.97
10.09
10.22
10.35
10.48
10.62
10*76
10.90
11.04
11.18
11.31
11.45
11*59
11.73
11.87
12.01
12.15
12.29
12.43
12.57
12.77
12.98
13*19
13*39
13.60
13.81
14.02
14.22
14.43
14.64
14.77
14*89
15.02
15.15
15*28
15.41
15.54
15.67
15.79
15.92
iruti
40
10.41
10.56
10.71
10.86
11.01
11.16
11.31
11.47
11.62
11.77
11.92
12.08
12.24
12.40
12.56
12.71
12.87
13.03
13.19
13.35
13.51
13.66
13.82
13*98
14.14
14.35
14.57
14.78
14.99
15*21
15.42
15. '64
15.85
16.07
16.28
16.44
16.59
16.75
16.91
17.07
17.22
17.38
17.54
17.69
17.85
VAtU<
45
11.50
11.67
11.85
12.02
12.19
12.36
12.53.
12.71
12.88
13.05
13.23
13.41
13.58
13.76
13.94
14.11
14.29
14.47
14.65
14.82
15.00
15.18
15.35
15.53
15.71
15.93
16.15
16.37
16.59
16.82
17.04
17.26
17.48
17.70
17.92
18.11
18.29
18.48
18.67
18.85
19.04
19.22
19.41
19.69
19.78
uuor
50
12.59
12.78
12.98
13.17
13.37
13.56
13.75
13.95
14.14
14.34
14.53
14.73
14.93
15.12
15.32
15.51
15.71
15.91
16.10
16.30
16.50
16.69
16.89
17.08
17.28
17.51
17.74
17.97
18.19
18.42
18.65
18.88
19*11
19.34
19.57
19.78
19.99
20.21
20.42
20.64
20.85
21.07
21*28
21*49
21*71
ttu-j^n i irarcjitviuua AivrjAa
55 60 65 TO 75 80 85 90 95
13.68 14.77 15.85 16.87 17.89 18.90 19.92 20.94 21.95
13.89 15.00 16.11 17.16 18*20 19.25 20.29 21.34 22.38
14.11 15.24 16.37 17.45 18.52 19.59 20.66 21.74 22.81
14.33 15.48 16.63 17.73 18.83 19.93 21.03 22.13 23.23
14.54 15.72 16.89 18.02 19.15 20.28 21.41 22.53 23*66
14.76 15.96 17.15 18.31 19.47 20.62 21.78 22.93 24.09
14.97 16.19 17.41 18.60 19.78 20.96 22.15 23.33 24.51
15.19 16.43 17.67 18.89 20.10 21.31 22.52 23.73 24.94
15.41 16.67 17.93 19.17 20.41 21.65 22.89 24.13 25.37
15.62 16.91 18.20 19.46 20.73 22.00 23.26 24.53 25.80
15.84 17.14 18.45 19.72 20.99 22.26 23.53 24.80 26.08
16.05 17.38 18.70 19.98 21.25 22.53 23.80 2*. 08 26.36
16.27 17.61 18.95 20.23 21.51 22.79 24.07 2S.35 26.63
16.48 17.85 19.21 20.49 21.78 23.06 24.35 2S>.63 26.91
16.70 18.08 19.46 20.75 22.04 23.33 24.62 2*. 91 27.19
16.91 18.31 19.71 21.01 22.30 23.59 24.89 26.18 27.47
17.13 18.55 19.97 21.27 22.56 23.86 25.16 26.46 27.75
17.35 18.78 20.22 21.52 22.83 24.13 25.43 26.73 28.03
17.56 19.02 20.47 21.78 23.09 24.39 25.70 27.01 28.31
17.78 19.25 20.73 22.04 23.35 2\.66 2$. 97 27.28 28.59
17.99 19.49 20.98 22.30 23.61 24.93 26.24 27.56 28.87
18.21 19.72 21.23 22.55-23.87 25.19 26.51 27.83 29.15
18.42 19. 9S 21.49 22.81 24.14 25.46 26.79 28.11 29.43
18.64 20.19 21.74 23.07 24*40 25.73 27.06 28.39 29.71
18.85 20.42 22.00 23.33 24.66 26.00 27.33 28.66 30.00
19.09 20.67 22.25 23.60 24.96 26.32 27.68 29.04 30.40
19.32 20.91 22*50 23.88 25.26 26.64 28.03 29.41 30.79
19.56 21.15 22.75 24.15 25.56 26.97 28.38 29.79 31.19
19.79 21.39 23.00 24.43 25.86 27.29 28.73 30.16 31.59
20.03 21.64 23.25 24.70 26.16 27.62 29.08 30.54 31.99
20.27 21.88 23.50 24.98 26.46 27.94 29.43 30.91 32.39
20.50 22.12 23.75 25.25 26.76 28.27 29.78 31.29 32.79
20.74 22.37 24.00 25.53 27.06 28.59 30.13 31.66 33.19
20.97 22.61 24.25 25*80 27.36 28.92 30.48 32.04 33.59
21.21 22.85 24.50 26.08 27.66 29.25 30.83 32.41 34.00
21.45 23.12 24.80 26.40 28.00 29.59 31.19 32.79 34.40
21.69 23.39 25.10 26.71 28.33 29.94 31.56 33.18 34.79
21.94 23.67 25.40 27.03 2B.66 30.29 31.93 33.56 35.19
22.18 23.94 25.70 27.34 28.99 30.64 32.29 33.94 35.59
22.42 24.21 26.00 27.66 29.33 30.99 32.66 34.33 35.99
22.67 24.48 26.29 27.98 29.66 31.34 33.03 34.71 36.39
22.91 24.75 26.59 28.29 29.99 31.69 33.39 3!>.09 36.79
23*15 25.02 26.89 28.61 3C.33 32.04 33.76 3b.48 37.19
23.39 25.29 27.19 28.93 30.66 32.39 34.13 3b.86 37.59
23.64 25.57 27.50 29.25 31.00 32.75 34.50 36.25 38.00
-------
TABLE 3. FINE MODEL ANALYSIS - 10 YEAR DESIGN STORM, FLAT AREAS
AREA
< ACRES »
5*60
5.70
5*80
5*90
6.00
6.10
6.20
6.30
6.40
6.50
6.60
6.70
6*80
6.90
7*00
7.10
7.20
7.30
7.40
0 7.50
no 7.60
7.70
7*80
7.90
6.00
8*10
8.20
8.30
8.40
8.50
8.60
8.70
8.80
8*90
9.00
9.10
9.20
9.30
9.40
9.50
9.60
9.70
9.80
9.90
10.00
5
3.44
3.51
3*58
3.65
3.73
3.60
3.67
3.95
4.02
4*10
4.15
4.21
4.27
4.33
4.39
4.45
4.51
4.57
4.63
4*69
4*75
4.81
4.87
4.93
5.00
5.05
5.09
5.14
5.19
5.24
5.29
5.34
5.39
5.44
5.50
5.57
5.64
5.71
5.77
5.84
5.91
5.98
6.05
6.12
6.20
&
10 15
4.62 5.80
4.70 5.90
4.79 5.99
4.87 6*09
4.95 6.18
5.04 6*28
5.12 6.37
5.21 6.47
5.29 6*56
5.38 6.65
5.46 6.76
5.54 6.86
5.62 6.96
9.70 7.07
5.78 7.17
1.86 7.27
1.94 7.37
6.03 7.48
6.11 7.58
6.19 7*68
6.27 7.78
6.35 7.89
6*43 7.99
6.51 6.09
6.60 8.20
6.67 8*30
6*75 8.42
6.83 8.52
6.91 -8,63
6.99 8.74
7.07 8.85
7.15 8.96
7.23 9.07
7.31 9.18
7.40 9.30
7.48 9*40
7.57 9.50
7.65 9.60
7.74 9.70
7.82 9.80
7.91 9.91
8.00 10.01
8.08 10.11
8.17 10*21
8.26 10.31
ailfllAirjU .
20 25
6.99 8.17
7.09 8.29
7.20 8.40
7.30 8.52
7.41 8.63
7.51 8.75
7.62 8.87
7.72 8.98
7.83 9,10
7.93 9.21
8.06 9.36
8*18 9.51
8.31 9.65
8.43 9*80
8.55 9.94
8.68 10.09
8.80.10.23
8.93 10.38
9.05 10.52
9.17 10.67
9.30 10.81
9.42 10.96
9.55 11.10
9.67 11.25
9.79 11.39
9.93 11.56
10.07 11,73
10.21 11.90
10.35 12.07
10.49 12.24
10.63 12.41
10.77 12.58
10.91 12.75
11.05 12.92
11.20 13.10
11.31 13.23
11.43 13.36
11.55 13.50
11.67 13.63
11.78 13.76
11.90 13.90
12.02 14.03
12*14 14.17
12.26 14.30
12.37 14.43
DJL.UW \\^,rof
30 35
9.35 10.61
9.48 10.76
9.61 10.91
9.73 11.07
9.86 11.22
9.99 11.37
10.11 11.52
10.24 11.68
10.37 11.83
10.50 11*98
10.66 12.17
10.83 12.36
11.00 12*55
11.16 12.75
11.33 12.94
11*49 13*13
11.66 13*32
11.83 13.51
11.99 13.70
12.16 13*89
12.33 14.09
12.49 14.28
12.66 14*47
12.83 14.66
13.00 14.85
13.20 15.07
13.39 15.28
13.59 15*49
13.79 15.71
13.99 15.92
14.19 16.14
14.39 16.35
14.59 16.57
14.79 16.78
15.00 17.00
15.15 17.17
16.30 17.34
15.45 17.51
15.60 17.68
15.75 17.85
15.89 18.02
16*04 18.19
16.19 18.37
16.34 18.54
16*50 18.71
run. vAitu
40 45
11.86 13.12
12.04 13.32
12.22 13.53
12.40 13.73
12.58 13.93
12.75 14.14
12.93 14.34
13.11 14.54
13.29 14.75
13.47 14.95
13.68 15.19
13.90 15.43
14.11 15.67
14.33 15.92
14.55 16.16
14*76 16*40
14.98 16.64
15.20 16.88
15.41 17.12
15.63 17.36
15.84 17.60
16.06 17.84
16.26 18.08
16.49 18.33
16.71 18.57
16*94 18.81
17.17 19.05
17.39 19.29
17.62 19.54
17.85 19.78
18.08 20.02
18.31 20.27
18.54 20.51
18.77 20.75
19.00 21.00
19.19 21.21
19.38 21.42
19.57 21.64
19.77 21.85
19.96 22.07
20.15 22.28
20.34 22.49
20.54 22.71
20.73 22.92
20.92 23.14
7uo rdm-aM
50 55
14.37 15.63
14.60 15.88
14.83 16.14
15.06 16.39
15.29 16.65
15*52 16.90
15.75 17.16
15.98 17.41
16.21 17.67
16.44 17.92
16.70 18.21
16.97 18*50
17.23 18.79
17.50 19.09
17.77 19.38
18.03 19.67
18*30 19.96
18.56 20.25
16.83 20.54
19*09 20.83
19.36 21.12
19*63 21.41
19.89 21.70
20.16 21.99
20.42 22.28
20.68 22.55
20.94 22.82
21.19 23.09
21.45 23.37
21.71 23.64
21.97 23.91
22.22 24.18
22.48 24.45
22.74 24.72
23.00 25.00
23.23 25.25
23*47 25.51
23.70 25.77
23.94 26.02
24.17 26.28
24.41 26.54
24.64 26.79
24.88 27.05
25.12 27.31
25.35 27.57
ni imrciivvr
60 65
16.88 18.13
17.16 18.44
17.44 18.75
17.72 19.05
18.00 19.36
18.29 19*67
18.57 19.97
18.85 20.28
19*13 20.59
19.41 20.90
19.72 21.24
20.04 21.58
20.35 21.92
20.67 22.26
20.99 22.60
21*30 22.94
21.62 23.28
21.93 23.62
22.25 23.96
22.56 24.30
22.88 24.64
23.19 24.98
23.51 25*32
23.62 25.66
24.14 26.00
24.42 26*30
24.71 26.60
24.99 26.90
25.28 27.20
25.57 27.50
25.85 27.79
26.14 28.09
26*42 28.39
26.71 28.69
27.00 29.00
27.27 29.30
27.55 29.60
27.83 29.90
28.11 30.20
28.39 30.50
28.67 30.79
28.94 31.09
29.22 31.39
29.50 31.69
29.76 32.00
UUO -fl
70
19.23
19.56
19.88
20.21
20.54
20.87
21.19
21.52
21.85
22.18
22.52
22.86
23.21
23.55
23.89
24.24
24.58
24.92
25.27
25.61
25.95
26.30
26.64
26.98
27.33
27.65
27.98
26.30
28.63
28.95
29.28
29.60
29.93
30.25
30.58
30.90
31.23
31.55
31.88
32.20
32.53
32.85
33.18
33.50
33.83
LlVti/VO
75 80
20.32 21.41
20.67 21.78
21.02 22.15
21.37 22.52
21.72 22.89
22.07 23.26
22.41 23.63
22.76 24*00
23.11 24.37
23.46 24.75
23.81 25*09
24.15 25.44
24,£0 25*79
24.85 26.14
25.19 26.49
25.54 26*84
25.89 27.19
26.23 27.54
26.58 27.89
26.93 28.24
27.27 28.59
27.62 28.94
27,97 29.29
28.31 29.64
28.66 30.00
29.01 30.37
29.36 30.74
29.71 31.12
30.06 31.49
30.41 31.87
30.76 32*24
31.11 32.62
31.46 32*99
31.81 33.37
32.16 33.75
32.51 34.12
32.86 34.50
33.21 34.87
33.56 35.24
33.91 35.62
34.26 35.99
34.61 36.37
34.96 36.74
35.31 37.12
35.66 37.49
85
22.51
22.90
23.29
23.68
24.07
24.46
24.85
25.25
25.64
26.03
26.38
26.73
27.09
27.44
27.79
28.15
28.50
28.85
29*21
29.56
29.91
30.27
30.62
30.97
31.33
31.73
32.13
32.53
32.93
33.33
33.73
34.13
34. *3
34.93
35.33
35.73
36.13
36.53
36.93
37.33
37.73
38.13
38.53
38.93
39.33
90 95
23.60 24.69
24.01 25.13
24.43 25.56
24.84 25.99
2U.25 26.43
25.66 26.86
26.07 27.29
26.49 27.73
26.90 28.16
27.31 28.60
27.67 26.95
28*02 29*31
28.38 29.67
28.74 30.03
29.09 30.39
29.45 30.75
29.81 31.11
30.16 31.47
30.52 31.83
30.88 32.19
31.23 32.55
31.59 32.91
21*95 33.27
32.30 33.63
32.66 34.00
39.09 34.45
3?. 51 34.90
33.94 35.34
34.36 35.79
34.79 36*24
3t>.21 36.69
3fr.64 37.14
36.06' 37.59
36.49. 38.04
36.91 38.50
37.34 38.95
37.76 39.40
38.19 39.84
38.61 40.29
39.04 40.74
39.46 41.19
39.89 41.64
40.31 42.09
40.74 42.54
41.16 43.00
-------
TABLE 4. FINE MODEL ANALYSIS -.5 YEAR DESIGN STORM, STEEP AREAS
AREA
( ACRES J
5.70
5.80
5.90
6.00
6.10
6.20
6.30
6.40
6.50
6.60
6.70
6.80
6*90
7.00
7.10
7.20
7.30
7.40
., 7.50
0 7.60
«*> 7.70
7.80
7.90
8.00
8.10
8.20
8.30
8.40
8.50
8.60
8.70
8.80
8.90
9.00
9.10
9.20
9.30
9.40
9.50
9.60
9.70
9.80
9.90
10.00
5
3.81 5
3.90 5
4.00 5
4.07 5
4.14 5
4.21 5
4.29 5
4.36 5
4.43 5
4.51 5
4.54 5
4.58 5
4.61 c.
4.65 5
4.69
4.72
4.76
4.80
4.84
4.88
4*92
4.96
5.00
5.04
5.09
5.15
5.22
5.28
5.35
5.42
5.48 '
5.55 '
5.61 '
5.68 '
5.75 '
5.81
5.88 '
5.95 '
6.01
6.07 '
6.13
6.19
6.?5
6.31 •
10
.02
.11
.21
.28
.35
.42
.49
.57
.64
.71
.77
.83
.89
>.95
.01
.07
.13
.19
.25
.31
.37
.42
.48
.54
.60
.67
.73
.80
.87
.94
7.01
7.07
7.14
7*21
7.28
7.34
7.41
7.48
7.56
7.65
7.73
7.82
7.90
7.98
usTira
15 20
6.23 7.44
6.32 7.54
6.42 7*63
6.49 7.70
6.56 7.77
6.63 7.84
6*70 7.91
6.77 7.98
6.85 8.05
6.92 8*12
7.00 8.23
7.08 8.34
7.17 8.45
7.25 8.55
7.34 8.66
7.42 8.77
7.50 8.88
7.59 8.98
7.66 9.07
7.74 9.17
7.-81 9.26
7.89 9.35
7.96 9.44
8.04 9. S3
8.11 9.63
8.18 9.70
8.25 9.77
8.32 9.84
8.39 9.91
B»46 9.98
8.53 10.05
8.60 10.12
8.67 10.19
8*74 10.26
8.81 10.33
8*87 10.41
6.94 10.48
9.01 10.55
9.12 10.68
9.23 10.81
9.33 10.94
9.44 11.07
9.55 11.20
9,66 11.33
ATfiJLJ *IAM
25 30
8.65 9.87
8.75 9.96
8.84 10.06
8.91 10.12
8.98 10.19
9.05 10.26
9.12 10.33
9.19 10.40
9.26 10.47
9.33 10.54
9.46 10.69
9.59 10.84
9.72 11.00
9.85 11.15
9.99 11.31
10.12 11.46
10.25 11.62
10.38 11.78
10.49 11.90
10.60 12.03
10.71 12.15
10.81 12.28
10.92 12.40
11.03 12.53
11.14 12.66
11.21 12.73
11.29 12.80
11.36 12.88
11.43 12.95
11.50 13.02
11.57 13.10
11.65 13.17
11.72 13.25
11.79 13.32
11.86 13.39
11.94 13.47
12.01 13.54
12.08 13.62
12.23 13.79
12.39 13.97
12.54 14.14
12.70 14.32
12.85 14.50
13.00 14.67
V (UfO) fUK,
35 40
10.92 11.96
11.09 12.22
11.27 12.48
11*34 12.56
11.42 12.65
11.49 12.73
11.57 12.81
11*64 12.89
11.72 12.97
11.79 13,05
11.97 13.24
12.14 13.44
12.32 13.63
12*49 13.83
12*67 14.02
12.84 14.22
13.02 14.41
13.19 14.61
13.34 14.78
13.49 14.95
13.64 15.12
13.78 15.29
13.93 15.46
14*08 15.63
14.23 15.80
14.33 15.94
14.44 16.08
14.54 16.21
14*65 16.35
14*75 16.48
14.86 16.62
14*96 16.76
15.07 16.89
15.17 17.03
15.28 17.17
15.38 17.30
15.49 17.44
15*60 17*58
15.79 17.78-
15.98 17.99
16*17 18.20
16*37 18.41
16.56 18.62
16.75 18.83
VAK.1UU5 J:
45 50
13.01 14.06
13.36 14.49
13.70 14.91
13.78 15.01
13.87 15.10
13.96 15.19
14,05 15,28
14.13 15*38
14.22 15.47
14.31 15.56
14.52 15*80
14.74 16.03
14.95 16.27
15.16 16.50
15. 38 16.73
15.59 16.97
15.81 17..20
16.02 17*44
16.22 17*65
16.41 17.87
16.60 18.09
16.80 18*30
16.99 18.52
17.18 18.74
17.38 18*95
17.55 19.15
17.71 19.35
17.88 19.55
18.05 19.75
18.22 19.95
16.38 20.14
18.55 20.34
18.72 20.54
IB. 89 20.74
19.05 20.94
19.22 21.14
19.39 21.34
19.56 21.54
19.78 21.78
20.00 22.02
20.23 22.26
20.45 22.50
20.68 22.74
20.90 22.98
'EltU^IN T 1M
55 60
15.11 16.16
15.62 16.75
16*13 17.34
16.23 17,45
16*32 17.55
16.42 17.66
16.52 17.76
16.62 17.87
16.72 17.97
16.82 18.08
17.07 18.35
17.33 18.63
17.58 18.90
17.84 19.17
18.09 19.45
18.35 19.72
18.60 20.00
18.85 20.27
19.09 20.53
19.33 20.79
19.57 21.05
19.81 21*32
20.05 21.58
20.29 21.84
20.53 22.10
20.76 22.36
20.99 22.62
21.22 22.88
21.45 23.15
21.68 23.41
21.91 23.67
22.14 23.93
22.37 24.19
22.60 24.45
22.83 24.71
23.06 24.97
23.29 25.23
23.52 25.50
23.77 25.77
24.03 26.04
24.28 26.31
24.54 26*58
24.80 26.85
25.05 27.13
rutiv
65
17.22
17.88
18.56
18*67
18.78
18*89
19.00
19*11
19.22
19.34
19.63
19.92
20.22
20.51
20.80
21.10
21.39
21.69
21.97
22.25
22.54
22.82
23.11
23.39
23.68
23.97
24.26
24.55
24.84
25.14
25.43
25.72
26.01
26.31
26.60
26.89
27.18
27.48
27.76
28.05
28.34
28.63
28.91
29.20
iuua AK.&A
70 75
18.22 19.22
18.82 19.76
19.43 20*30
1.9.59 20.52
19.76 20.74
19.92 20.96
20.09 21.18
20.25 21.40
20.42 21.62
20.59 21.84
20.91 22.18
21.23 22.53
21.55 22.88
21.87 23.22
22.19 23.57
22.51 23.92
22.83 24.26
23.15 24.61
23.45 24.93
23.75 25.25
24.05 25.57
24.36 25.89
24.66 26.21
24.96 26.53
25,26 26.85
25.58 27.20
25.90 27.54
26.22 27,39
26.54 26*23
26.86 26*58
27.18 28.93
27.50 29,27
27.82 29.62
28.13 29.96
28.45 30,31
28.77 30.65
29.09 31.00
29.41 31.34
29.72 31,67
30.02 31.99
30.33 32*31
30.63 32.64
30.94 32.96
31.24 33.28
»
80 86
20.23 21.23
20.70 21.64
21.17 22.04
21*44 22.37
21.72 22.70
21.99 23.02
22.26 23.35
22*54 23.68
22.81 24.01
23.09 24.34
23.46 24.74
23.83 25.14
24.21 25.54
24.58 25.94
24.95 26.34
25.33 26.74
25.70 27.14
26*08, 27,54
26.41 27.89
26*75 26.25
27.09 28.61
27*43 28*96
27.76 29.32
28.10 29.67
28.44 30.03
28.81 30.43
29*18 30.83
29.56 31,23
29.93 31.62
30.30 32.02
30.67 32.42
31.05 32*82
31.42 33.22
31.79 33.62
32.16 34.02
32.54 34.42
32.91 34.62
33.28 35.21
33.62 35.57
33.96 35.93
34.30 36.29
34.64 36.65
34.98 37.01
35.32 37.36
90 95
22.24 23.25
22.58 23.52
22.91 23.79
23.29 24.22
23.68 24*66
24.06 25*09
24.44 25.53
24,82 25.96
26*20 26.40
25.59 26*84
26.01 27*29
26.44 27.74
26.67 28.20
27.29 28.65
27.72 29.10
26*15 29.56
28.57 30.01
29.00 30.47
29*38 30.86
29.75 31.25
30,12 31.64
30.50 32.03
30.87 32.42
31.24 32.81
31.62 33.21
32.04 33.66
32.47 34.11
32.89 34.56
33*32 35.01
33.74 35.47
34.17 35.92
34.60 36.37
35.02 36.82
35.45 37.28
35.87 37.73
36.30 38.18
36.72 38.63
37.15 39.09
37.53 39.48
37,90 39,67
38,28 40,27
38,65 40,66
39,03 41.05
39.40 41.45
-------
,LIMIT OF TRIBUTARY AREA
-EDGE OF ROAD-
o
ROOF DRAIN CONNECTIONS (TYPICAU-
I _
o
z
UJ
a.
B
AREA NO. 1
%IMPERVIOUSNESS'
SIZE=1 ACRE
75
-------
Fine Model Application. The set of tables were
then used for the determination of improvements needed for
the areas presently served by sewers smaller than 30 inches
in diameter. In using these tables, procedures must be used
that are in accordance with their development. Figure 11 is
included to demonstrate the procedure which is as follows.
The area tributary to Pipe No. 1 consists of the area
of the roofs of houses A, B, C, and D which discharge their
roof drains directly to the sewer. The above tables show
this as one part of the total area tributary to Pipe No. 1
(0.30 acres) at 95 percent imperviousness. The area
tributary to Pipe No. 2 consists of the entire Area No. 1
and the roofs of houses E, F, G, and H. This is shown in
the tables as 1.30 acres at 75 percent imperviousness.
CASE STUDY - 2
Introduction
COMBINED SEWER OVERFLOW REGULATION IN THE
CITY OP MIDDLETOWN, OHIO
The combined sewer area in the City of Middletown
consists of six drainage districts, three of which also
receive sanitary inflow from separate sewers tributary to
them. These drainage districts are as shown on Figure 12
and their related tributary areas are as follows:
Drainage
district
No.
1
2
4
5
6
Area tributary to
combined sewers,
acres
234
496
1,006
83
902
714
Area tributary to
separate sewers
upstream, acres
246
326
1,233
The districts drain in a westerly direction toward
the Great Miami River. A combined sewer interceptor,
located along the river, collects and transports dry-weather
flows from these areas to Middletown's water pollution
control plant. This plant is designed to provide secondary
treatment for flows up to the capacity of the combined sewer
interceptor.
305
-------
DESIGNATED MODEL IREAS
»RE< I . LJMVEriE tVENUE
*RE* 2 - I4TH *VCNJl
<*E* 3 - STH AVENUE
*Sf* 4 - 2ND IVEHUE
»RE» 5 - BULLS RUN
*RF* G - UKFSIOE 1VEUE
LFGfHP
LIIIIT OF COMBINED SEVER AREA'
— IIBIT OF OR»IN«GE DISTRICT
- SEPARATE SANITARY SEVERED
' AREAS TRIBUTARY TO THE
COttSINEO SEWER OYERFLOIS
'- EXISTING KAJOR COMBINED s
SFWEBS \
EXISTING INTERCEPTOR SE»ER
EXISTING REGULATOR LOCATIONS
CXI SI INC OVERFLO* LOCATIONS
FIG. 12 MAJOR DRAINAGE DISTRICTS, CITY OF M1DDLKTOWN, OHIO
306
-------
All combined sewer outfalls are connected to the
Interceptor by junction chambers designed to regulate flows
so that the major portion of flow during storm runoff passes
directly to the Great Miami River. All junction chambers
are equipped with float-type regulators except one, which
has a fixed orifice regulator.
In concept, the regulators should minimize the amount
of flow discharging to the river by diverting the maximum
possible flow to the treatment plant. Flow records at the
Water Pollution Control Plant sometimes show a reduction of
flow to the plant during periods of rainfall indicating that
the regulators, as presently operated, perform a negative
function for water pollution control.
Studies conducted during the period 1968-71,
established that dissolved oxygen is the most critical
problem on the Great Miami River when evaluated in terms of
established water quality standards.
The purpose of this investigation was to conduct a
comprehensive engineering evaluation with respect to
abatement of pollution from combined sewer overflows
to the Great Miami River within the City of Middletown.
Dry-Weather Flows
Dry-weather flow is comprised of domestic sewage from
residences, commercial, and institutional establishments and
industries; process flows from the connected industrial
establishments; and infiltration that enters the sewer
system.
Industrial Process Flows. The major industrial
process wastewaters in the combined sewer area that are
expected to contribute to the public system were obtained
from the industries in terms of flow, BOD and SS and their
variability.
Infiltration. Infiltration was assumed to be free
of pollutants.
Sewage Flows. Values selected as an allowance for
sewage flow are shown in Table 5 and were obtained from
analysis of water records and treatment plant data. Plant
records were analyzed in"order to determine an allowance for
BOD and SS contributions in dry-weather flow. These were
taken excluding data during storm conditions. In order to
arrive at a representative value of conforms, samples were
taken from the combined sewer interceptor during times of no
rainfall an.d were analyzed.
307
-------
TABLE 5. AVERAGE SEWAGE PLOW ALLOWANCES IN MIDDLETOWN,
OHIO
Item Per capita allowance
Flow 85 gcd
BOD5 244 rag/L
SS 150 mg/L
Conforms 6.1 x 10 most
probable number
Based on the previous criteria, the average dry-weather
flow discharged from each area of the combined sewer system
and its tributary upstream areas is as shown in Table 6.
TABLE 6. AVERAGE DRY-WEATHER PLOWS DISCHARGING
INTO COMBINED SEWERS IN MIDDLETOWN, OHIO
Area
No.
1
2
4
5
6
Total
Industrial process
flow, (l) mgd
1.0
0.95
1.95
Infiltration,
mgd
0.070
0.220
0.400
0.025
0.223
0.579
1.517
Sewage flow,
mgd
0.158
0.463
1.034
0.044
0.659
0.902
3.260
. ..-. .
process of being diverted directly to the combined
sewer interceptor.
Adjustment Factors. Adjustment factors are based
on 1972 records as shown in Table 7 for daily values of
flow, BOD and SS. For coliform, we have assumed no variation
from day to day.
308
-------
TABLE 7. DAILY ADJUSTMENT FACTORS
Day of week
Sunday
Monday
Tuesday
Wednesday
Thursday
Friday
Saturday
Flow
0.91
1.05
1.04
1.05
1.00
1.01
0.94
BOD
1.01
1.09
0.99
0.99
0.92
1.00
1.00
SS
0.83
1.11
1.01
1.09
0.96
1.00
1.00
By analysis of the Manning charts which are recorded
on a daily basis at the treatment plant, for dry-weather flow
conditions only, it was possible to develop hourly adjustment
factors for flow at the plant. These figures are presented
in Table 8.
TABLE 8. HOURLY ADJUSTMENT FACTORS FOR FLOW
Hour of day Adjustment factor
1
2
3
4
5
6
7
8
9
10
11
12 (noon)
13
14
15
16
17
18
19
1.01
0.97
0.90
0.85
0.82
0.80
0.80
0.81
0.83
0.95
1.03
1.11
1.14
1.14
1.14
1.14
1.15
1.15
1.10
Flow represented, mgd
9.71
9.32
8.69
8.19
7.89
7.69
7.71
7.79
7.96
9.12
9.89
10.64
10.92
11.01
10.92
11.00
11.10
11.10
10.57
309
-------
TABLE 8 (Continued). HOURLY ADJUSTMENT FACTORS FOR FLOW
Hour of day Adjustment factor Flow represented, mgd
20
21
22
23
24
(midnight)
1.
1.
1.
1.
1.
08
08
04
01
01
10
10
10
9
9
.41
.36
.04
.75
.70
Since BOD, SS, and conforms are not normally sampled
and tested on an hourly basis in Middletown, data from nearby
Cincinnati were used.
Wet-Weather Flows
In selecting the storm events to be used in this
study, hourly rainfall records were obtained from the United
States Department of Commerce, National Weather Records
Center, for Oxford and Dayton and daily averages were
obtained for Middletown.
These storm records were analyzed and compared in
terms of duration, frequency of occurrence, intensity, and
total and average monthly rainfall.
Design Storm Definition. A period of six hours
with no measured rain was selected to separate each event.
On this basis, storms start with the first measurable rain-
fall after six consecutive dry hours and end at the next six
consecutive hours of no measured rain. The period of
separation was judgmental and of no special hydrologic
significance. It was selected as a practical time needed
for system dewatering.
A separation time other than six hours would change
statistics associated with the data only in terms of the
number of storms and the total rainfall per storm. It
would not change the intensity of a storm or the total rain-
fall per year.
Rainfall Data Analysis. Total and summer rainfall
events were ranked by maximum hourly rainfall, total rain-
fall, and duration.
Summer storms (those storms occurring between June 1
and September 30) were treated separately to determine how
storm characteristics change during the critical water quality
season.
310
-------
The maximum expected hourly intensity occurring once
per year is about 1.1 inches per hour; the maximum total
expected rainfall event per year is about 2.5 inches; and
on the average, 46 storms events per year are expected to
last at least six hours.
Rainfall data compiled in Technical Paper No. 40 for
the Middletown area compares well with the data developed.
For this reason, intensity ratios derived from Technical
Paper No. 40 for durations of less than 60 minutes were used
for design hyetograph development. The procedure used is
similar to that described in Case Study No. 1. The one-year
design hyetograph is shown on Figure 13.
Surface Runoff Quality. Surface runoff, once con-
sidered as clean water, contains a relatively large composi-
tion of contaminants which cannot be ignored in water quality
control management. According to recent studies,* the magni-
tude of the three pollution components considered in this
report are characterized as follows:
BODc content of runoff equals about the strength of
domestic sewage after secondary treatment from the
same land use.
SS content of runoff is generally about three times
that of untreated sewage, but consisting mostly of
inorganic materials.
Goliform content of runoff is about two to four orders
of magnitude smaller than untreated sewage. However,
it is two to five orders of magnitude higher than is
considered safe for water contact recreation.
In the modeling of the combined sewer system,
allowances were made for. pollution from urban runoff in
addition to that contributed by dry-weather flows. Allow-
ances were based on general data recommended in the SWMM
using Middletown1s street-cleaning practices and design
standards along with land-use data as criteria defining the
amounts to be expected.
Durban Stormwater Management and Technology; An Assessment,
December 1973, Office of Research and Development, U.S.
Environmental Protection Agency by Metcalf & Eddy Engineers
(Draft)
311
-------
ro
i
cc
£
CO
HI
o
3.0
2.4-
1.8
1.2
0.6
0.0-
0
0.48
o 36 r
.24 r
0.12 PP1
or** i
i
- o
0.96
—
.60
.Ul
•i
1
,
*i
TOTAL R AINF ALL = 1.80"
MAX. 10MIN. = 0.51"
10MIN. INCREMENTS
0.96
10.60
l£^8
i 0.24
1 0-12
[ I I \ 1 -1
1 1 1
TIME (HOURS)
FIG, 13 1-YEAR - 6-HOUR DESIGN HYETOGRAPH FOR MIDDLETOWN, OHIO
-------
Evaluation of Existing System
The total capacity of the combined sewer system on
the basis of the capacity of the outfalls is 1,810 cfs. For
each drainage area, this is shown in Table 9.
TABLE 9. ONE-YEAR DESIGN STORM OUTFLOWS
Full flow Peak
Area capacity of flow^1'
No. outfalls, cfs cfs
1
2(15)
2(14)
3(9)
3(re-
lief)
4
5
6
65
85
198
14?
480
71
456
308
63
45
171
182
385
69
533
364
, BOD,-(2
Ib J
260
150
670
1,710
540
110
1,010
2,320
>, ss<2>,
Ib
3,590
1,190
7,200
12,720
8,130
1,180
10,430
9,850
Coliforms t-U,
most probable
number/100 ml
6.62 x 106
5.87 x 106
5.87 x 106
7.23 x 106
8.31 x 105
2.68 x 106
5.66 x 106
6.95 x 106
1. Maximum rate.
2. Total for entire storm.
In order to determine the amounts of flow and pollu-
tants being discharged from each area under storm conditions,
computer simulation was.used. The results for the one-year
design storm are shown in Table 9.
Combined Sewer Interceptor. In order that the
operation of overflow regulators could be evaluated
appropriately, a detailed hydraulic analysis of the combined
sewer interceptor was carried out under steady state
conditions.
Regulators. The float-operated, self-powered
regulators are housed in twin below-ground vaults located
along one side of a relief sewer. The downstream chamber
houses the float and is connected to the relief sewer by a
313
-------
telltale conduit. The water level in the relief sewer con-
trols the float movement which, in turn, controls the
regulator gate located in the upstream chamber. The float
is connected to the regulator sluice gate by a sprocket and
chain assembly. A low-profile invert dam located in the
relief sewer diverts dry-weather flow through the regulator
gate into the interceptor. As depth of flow in the relief
sewer increases, the float is raised and the regulator gate
is closed causing flow to be routed to the Great Miami
River.
Water Pollution Control Plant. Middletown's Water
Pollution Control Plant has been designed to provide
secondary treatment with an average of 26.6 cfs and a maxi-
mum of 56.3 cfs allocated to the area served by the combined
sewer interceptor.
Analysis of Existing System Operation. On the
average, there are about 90 overflows per year from the
existing system. This is based on the premise that once a
storm runoff occurs in excess of the regulator capacity,
regulators close bypassing all flows to the Great Miami River.
In studying the effects of 1/2-year and 1-year storms
it was found that such storms with a severity sufficient to
flush the street surfaces and to flush out a combined sewer
system will cause nearly equal pollution loads.
The City has been farsighted in its use of storage
basins in the upper reaches of the drainage area. These
facilities retain runoff and reduce peak flows in the
combined sewer system. However, little use has been made
of the storage capabilities in the large pipes of the sewerage
system itself to retain the pollution loads from the small,
but frequent, rainfall events for discharge to the water
pollution control facilities.
The regulator operation as designed does not permit
taking maximum advantage of the present treatment plant and
interceptor capacities.
Although it was not the purpose of this investigation to
evaluate drainage problems, the SWMM identified areas of
flooding problems supporting the City's flooding experiences.
Development of Alternatives
At the outset of the project, three objectives were
established in developing alternative solutions to the
abatement of pollution as a result of combined sewer over-
flows to the Great Miami River.
314
-------
The first objective was to initially take maximum
advantage of the presently existing facilities, such as the
storage available in the combined sewers, the transport
capacity of the interceptor, and the treatment capacity at
the water pollution control plant.
The second objective was to search out procedures for
additional abatement of pollution from overflows as may be
required by the Ohio Environmental Protection Agency.
The third objective was to look for remedial solutions
that do not require sophisticated equipment, nor specially
qualified personnel to operate, nor were based on limited
pilot-plant studies.
As a result of these objectives, a two-phase solution
for the abatement of overflow pollution was established.
Phase One consists of modifications to the existing system to
the extent of taking maximum advantage of existing facilities
with Phase Two consisting of providing further pollution
abatement.
Remedial Opportunities. The major categories of
available improvements are sewer separation; overflow con-
tainment; flow through overflow treatment without detention;
and overflow treatment with detention. Each of these methods
is discussed briefly as follows:
Sewer Separation. Sewer separation would cost on the
order of $38 million for Middletown. Separation does not
eliminate all the pollution because discharge resulting from
urban runoff includes pollutants stored on streets and park-
ing lots and in catchbasins and pipes. As mentioned
previously, the quality of surface runoff discharged is
generally equal to that of dry-weather flow after secondary
treatment in terms of BOD and equal to three times that of
dry-weather flow in terms of SS. In addition to these, a
considerable amount of floating visual pollution is dis-
charged. The inconvenience caused by the tearing up of
every street to provide for sewer separation cannot be
expressed in terms of cost.
Overflow Containment. Containment of overflows for
the purposes of ultimate treatment and discharge is an
effective method of pollution abatement. However, total
containment without overflows is very costly and requires
large land areas. The requirements for total containment of
the one-year and half-year design storms is as follows:
315
-------
One-year storm Half-year storm
Total off-line volume
required, cf 4,973,600 2,888,500
Area, acres* 7-6 4.4
~*The area required is based on storing 333,700 cf in the
combined sewer system, 2,167,500 cf in the outfall collec-
tor conduit, diverting 1,003,000 cf to the treatment plant,
and storing the remaining volume in a 15-foot deep reservoir,
Plow-Through Treatment. Treatment of overflows with-
out any storage are included in this category using devices
for screening out of pollutants or their separation from
runoff by concentration techniques. Although much research
is being conducted in this area, little experience exists at
this time on other than test-scale projects. Secondly, with-
out the provision for sufficient storage to provide for
disinfection contact time, the elimination of bacteria cannot
be achieved.
Overflow Detention and Treatment. The combination
of storage and treatment was considered the most effective
means for abatement of pollution for Middletown's combined
sewer overflows. This consists of capturing completely
the small frequent overflows for transmission to the
water pollution control plant once wet-weather flows
subside; and of providing for the detention, screening and
disinfection of overflows during the large and less frequent
storms; thereby achieving the elimination of bacteria, the
removal of visual pollution and the reduction of pollutants
discharged to the river.
In addition to being the most desirable in terms of
current technology, facilities associated with this process
are relatively .easy to operate avoiding problems associated
with highly sophisticated apparatus. In addition, the
systems developed in this process are expandable, should
requirements for additional pollution abatement become
necessary. For these reasons, alternatives were based on
this concept of treatment.
Various alternatives to regulator improvements,
in-system storage and external storage and treatment were
investigated leading to the recommended solution discussed
in the next section.
316
-------
Recommended Solution
The abatement of pollution from combined sewer over-
flows was recommended in two phases. In general, Phase One
consists of making maximum use of already existing facili-
ties for pollution abatement by diverting maximum design
flows to the water pollution control plant, by retaining
flows in the large combined sewers until their capacity is
reached,' and then by permitting excess flows to overflow
into the river.
Phase Two, on the other hand, consists of further
reduction of pollution by construction of facilities that
would carry out disinfection, and further reduce BOD and SS
discharge by capture of more flows and by treatment of all
overflows.
Phase One Facilities. Phase One recommendations
consist of modifying the existing regulators, adding a
regulator and installing gates for in-system storage in
outfall sewers serving three areas. The methods in which
these solutions are recommended to be accomplished are as
follows:
Regulator Modifications. The regulator system should
divert the maximum pollution loads to the combined sewer
interceptor in. accordance with maximum design capacity of
the water pollution control plant. The present regulator
controls do not best accomplish this objective since they
close during wet weather, thus diverting all flows to the
Great Miami River.
By changing the regulator controls, a constant maxi-
mum flow can be diverted to the interceptor.
With modifications, the regulator functions as a
governor and the float is actuated by the water level in
the gate chamber. In this way, the diverted flow can be
held within +10 percent of the design flow.
In-System Storage. The intention of in-system storage
•is to maximize use of the existing sewer system to store
combined sewage to the extent permitted by hydraulic condi-
tions and to transmit this flow to the treatment facilities
once space there becomes available after a storm. In-system
storage serves to capture all or part of small storms and
Diverts them directly to the interceptor. The flow resulting
fFom small storms, if not captured, tends to flush the
°ombined system and transport large pollution loads to the
Deceiving waters. Five locations appear to be available with
317
-------
storage totaling about 334,000 cubic feet. Of these, all
except one sewer are recommended, for Phase One providing a
storage of 294,000 cubic feet.
In order to create in-system storage, gates are
recommended to be placed in the combined system just down-
stream of the regulators. The gates are controlled by
sensing equipment which record the water surface elevation
upstream of the gate. The gate is held in the upright
position (for maximum storage) during dry weather and would
be lowered if it is necessary to permit large storm flows
to pass in emergencies.
Phase Two Facilities. On the basis of the analysis
of present discharge locations, available land, sewer system
configuration and water quality, use of two combined sewage
storage detention and treatment facilities are recommended,
as shown on Figures 14 and 15. These have been sized to
provide a minimum of 15-minutes detention time for the
peak flows during a one-year design storm, thereby providing
at a minimum for disinfection of overflows and for limited
treatment in terms of BOD and SS reduction.
Major components associated with these facilities
are shown in Table 10.
TABLE 10. MAJOR FACILITIES
Areas 4, 5, and 6,Areas 1, 2, and 3
Item Tank 1 • Tank 2
Tank
Size 160 ft by 340 ft 100 ft by 230 ft
Volume 6 8-ft
depth 473,640 cf 198,000 cf
Peak flow, one-
year storm 900 cfs 750 cfs
Outfall collector
conduit
Size 10 ft-10 in. by 10 ft-10 in. by
16 ft-3 in. 16 ft-3 in.
Length 2,500 If 3,600 If
Volume 337,500 cf 486,000 cf
318
-------
EMERGENCY RELIEF
OVERFLOW
ND LAKESIDE
TO PROPOSED
LL INTERCEPTOR
PROPOSED Iff-1
CORRUGATED AR
PROPOSED
RELOCATED
TAILRACE
PROPOSED RIVER
WATER SUPPLY LINE
PROPOSED
RETURN
LINE
PROPOSED CUTLET
FROM TANK AND
FLAP GAT
EXISTING
INTERCEPTOR
300
SCALE IN FEET
FIG. 14 LOCATION OF PROPOSED FACILITIES TANK ONE -
AREAS 4, 5, & 6, MIDDLETOWN, OHIO
319
-------
PROPOSED 101 10" x 16'-3
CORRUGATED ARCH
PROPOSED AREA OF
RECONSTRUCTION
TO LOWER ARMCO
PROCESS SEWER
EXISTING INTERCEPTOR
ROPOSED BfVER WATER
SUPPLY LlWfc TO PARALLEL
PROPOSE/) OUTLET F ROM "SANK
——s^-jj
/f/////'/7///j"/T'//. ft//////''.?/''' '
PROPOSED
RlP-flAP
CHANNEL
PROPOSED OUTLET FROM TANK TQ>
FOLLOW ALIGNMENT OF EXISTING I
LAFAYETTE AVENUE OUTFALL J
FIG. 15 LOCATION OF PROPOSED FACILITIES TANK TWO
AREAS 1, 2, &3, MIDDLETOWN, OHIO
320
-------
Reduction of Pollution Discharges. Since combined
sewer overflows are highly variable and stormwater dependent,
analysis of their effects on river pollution cannot follow
the traditional approach of steady state analysis of a con-
tinuously discharged uniform waste. Two types of analysis
are conducted in this report. First, an evaluation is made
on the reduction in annual pollution volumes as a result of
implementing Phases One and Two. Second, an evaluation is
made of the effects of each phase on water quality of the
Great Miami River.
Frequency of Overflows. The estimated number of over-
flows occurring during an average year under existing condi-
tions and after Phases One and Two are shown in Table 11.
The effect of the proposed facilities is to reduce the number
of overflows from 90 to 39 and 19, respectively, after
implementation of Phases One and Two.
TABLE 11. ESTIMATED ANNUAL FREQUENCY OF OVERFLOWS
(1)
Existing After Phase After Phase
Item conditions One Two
Number of storms causing
runoff 116 116 116
Number of storms causing
overflow 90 39 19
Percent of total rain-
fall volume
overflowing^1) 80 55 31
1^Includes allowance for a uniform distribution of rain-
fall, surface storage, surface infiltration, diversion
to the treatment plant, and in-line and off-line
storage.
Pollution Loads. As shown in Table 12, the reduction
in total pollution discharged annually is 93 percent BODc
and 89 percent SS. Table 13» on the other hand, shows the
reduction in pollution loads discharged during the one-year
design storm.
321
-------
TABLE 12. ESTIMATED TOTAL ANNUAL POLLUTION
LOADS DISCHARGED
ExistingAfter PhaseAfter Phase
Item conditions One Two
BOD5 in pounds 250,OQoU) 54,000 19,000
in percent of existing
conditions - 78 93
SS in pounds 985,00o(1) 310,000 110,000
in percent or existing
conditions - 68 89
T~.Pollution load without diversion to treatment plant.
Note: Total annual pollution load is based on grouping
storms which produce like pollution loads.
It should be noted that the overall reduction in
pollution is from 90 untreated overflows per year to 19
partially treated overflows including disinfection to
eliminate the chance of discharging pathogenic organisms.
TABLE 13. ESTIMATED POLLUTION LOADS DISCHARGED DURING
THE ONE-YEAR DESIGN STORM
Item
BODcj in pounds
In percent of
existing
conditions
SS In pounds
in percent of'
existing
conditions
Existing
conditions
6,800
-
54,300
••
After Phase
One
3,500
52
43,100
79
After Phase
Two
2,500
37
27,000
50
Receiving Water Quality
The method of approach in evaluating water quality
conditions in the Great Miami River resulting from combined
sewer overflow remedial actions was a combination of steady
state and dynamic analyses. This was done primarily to
322
-------
assist decision makers in evaluating how the conditions
achieved fit established criteria. To date, no meaningful
criteria have been set that are applicable to combined sewer
overflows.
A steady-state analysis was conducted to determine
the DO sag conditions in the river resulting from the
Middletown loads. In this analysis, loads from combined
sewer overflows were taken as the peak pollution discharges.
However, overflows are not steady continuous loads
but consist of short-term slugs that flow downstream growing
in extent and reducing in concentration as a result of
dispersion. For this reason, another steady-state analysis
was conducted in which the pollution discharge was represented
as a constant discharge of the average load discharged
during the peak hour of combined sewer overflow. This was
judged as a condition which dispersion around the peak would
cause.
Both analyses were conducted following the traditional
Streeter-Phelps approach of assuming all flows and water
quality conditions as not changing.
Recognizing the highly variable conditions of combined
sewer overflows, dynamic analyses were also conducted.
Initially, we followed the principle that storm conditions,
especially during a one-year storm, are associated with
increased river flows, thereby negating the use of low river
flows. However, upon further evaluation, it was felt that
low river flows should be used for a combined sewer overflow
analysis, thereby representing a very localized storm
condition.
Therefore, the dynamic analysis was conducted using
low river flow equivalent to the seven-day low flows exceeded
90 percent of the time, as presented earlier. Time varying
Pollution loads (pollutographs), however, were used to
Represent the discharge loads under existing, Phase One and
Phase Two conditions. The discharges from Middletown1s
water pollution control plant were also included as uniform
discharges.
The method is a numerical network model consisting of
interconnected channel reaches simulating the mass transport
Phenomena including advection, dispersion and biological
reaction processes following first-order kinetics. The
323
-------
technique was developed at the Massachusetts Institute of
Technology.*
The results of each analysis are shown on Figures 16
and 17 for existing conditions, Phase One and Phase Two.
This analysis includes the discharge of Middletown wastes
including those from its Water Pollution Control Plant, but
does not consider conditions above or below the Middletown
area. This analysis can be readily incorporated in a more
comprehensive river basin plan.
The effects on DO from the overflow pollution slugs,
as they move downstream in the dynamic analysis, are shown
on Figure 17. It should be noted here that these effects
are felt over only a two to three-hour period.
Superimposing the results of the dynamic analysis
(see Figure 16) shows that dispersion does have a marked
effect in reducing the peak DO sag as the slug moves down-
stream. Based on this analysis, it appears that the DO sag
of a combined sewer overflow would initially follow the
steady-state analysis of its peak but would recover con-
siderably sooner due to dispersion effects.
In evaluating the criteria associated with design
river flows which imply that criteria must be met 90 percent
of the time, a comparison to overflow conditions can be made.
As a result of Phase Two improvements, there will be about
19 overflows per year. Assuming that these will occur over
a six-hour period indicates that overflows will occur 114 hours
per year, or less than 10 percent of the time.
On the basis of annual pollution volumes discharged,
the reduction in the volume of BOD discharged as a result
of Phases One and Two is 78 and 93 percent, respectively.
*Numerlcal Model for the Prediction of Transient Water Quality
in Estuary Networks, October 1972. Jam^s TC. r>^iiey an
-------
ro
in
10-
REACH NUMBERS
Q 4
EXISTING SYSTEM
STEADY STATE "DISSOLVED OXYGEN SAG"
DUE TO PEAK BOD DISCHARGE RATE
STEADY STATE "DISSOLVED OXYGEN SAG"
DUE TO AVERAGE PEAK HOURLY BOD DISCHARGE RATE
LOWEST DISSOLVED OXYGEN CONCENTRATIONS FROM THE
DYNAMIC SOLUTION UNDER PHASE H
CONDITIONS AS POLLUTOGRAPHS MOVE DOWNSTREAM
8 10 12
RIVER MILE
14
16
18
20
22
24
FIG. 16 DISSOLVED OXYGEN SAG - STEADY STATE AND DYNAMIC ANALYSIS FOR ONE YEAR STORM
-------
00
ro
CT>
O
X
o
Q
UJ
12
10
8
6
EXISTING SYSTEM
PHASE I
PHASE H
15
17
19
21
23-
25
27
TIME (IN THOUSANDS OF SECONDS)
FIG. 17 DISSOLVED OXYGEN VS. TIME AT BEGINNING OF REACH NO. SIX - ONE YEAR STORM
-------
CASE STUDY - 3 COMBINED SEWER OVERFLOW
REGULATION IN THE METROPOLITAN BOSTON AREA
Introduction
Boston Harbor has three major tributary rivers, the
Mystic, Charles and. Neponset, which funnel through Boston
and empty their accumulated flows and impurities into the
Harbor. Prior to l889> the increased expansion of local
sewer systems discharging their wastes and street runoff
directly into these rivers gave rise to considerable public
concern. Accordingly, in 1889, the State Board of Health
completed an exhaustive investigation and recommended
passage of the act establishing the Metropolitan Sewer
District.
Today there are 43 cities and towns in the Metropoli-
tan Sewer District which now serves almost 2 million people
from an area greater than 400 square miles.
The Metropolitan Sewer District facilities include
approximately 225 miles of trunk sewers, serving nearly
5,000 miles of local sewers. The District has 11 pumping
stations, four headworks, and two large primary treatment
Plants at Deer Island and Nut Island. These plants have an
average treatment capacity of more than 450 million gallons
Per day, with a combined capability of handling maximum
flows at the rate of 1.2 billion gallons per day.
Completing the major components of the wastewater
transport and disposal system in the Metropolitan Sewer
District service area are 68 major combined sewer overflows
lh five member communities serving an area of 36 square miles
and 900,000 people. Figure 18 shows the general location of these
overflows.
After having the responsibility of providing sewer
for the past 80 years, the M.D.C. is now planning
for the next 80 years.
One component of this planning effort is a broad
ent of the combined sewer overflow problem in th<
Harbor area and Identification of remedial oppor-
s available for such pollution abatement.
327
-------
During a conference on the Boston Harbor*, the
biggest pollution problem confronting the Harbor was identi-
fied as the combined sewer discharge problem which has been
causing bacterial contamination and visual pollution of
beaches and shellfish areas.
There have been in the past, in addition to sewer
separation, two general approaches to combined sewer over-
flow regulation in the Boston area.
1. Diversion of overflows to chlorination-detention
facilities allowing for piecemeal solutions in
the immediate need areas.
2. Collection and ocean discharge of overflow
providing for a large-scale elimination from
the Boston Harbor area.
The Cottage Farm Chlorination-Detention facility
in Cambridge is an example of the first approach, and the
Deep Tunnel Plan, as proposed to the City of Boston,
represents the second approach.
Treatment Criteria
The suitability of the waters for bathing and other
recreational activities and for the harvesting of shellfish
is dependent on acceptably low pollution concentrations.
The expected future pollution control processes would
be biological treatment for dry-weather flow with pretreatment
of noncompatible industrial wastes. On this basis it is
expected that industrial pollutants such as toxics'and heavy
metals would be removed prior to entering the sewer system
and would not become subject to overflow.
P K ™e critical parameter for an area's acceptability
for bathing in the Boston Harbor area has been bacterial
contamination. On'this basis, the destruction of bacteria
is set as the primary requirement of any treatment process,
with reduction of solids pollution and elimination of visual
pollution as additional measures of accomplishment.
^Proceedings, Conference In the Matter of Pniin^^ of
the Navigable Waters of Boston Harbor a^ri -it-.* Tr1Hl4.^rl
Third Session, October *-f> j.m> Boston, Massachusetts,
U. S. Environmental Protection Agency UOC«,«,D,
328
-------
In order to meet the bacterial standard set by the
Public Health agencies, the facilities would be designed to
provide 15 minutes chlorine disinfection contact time for the
peak flow from a selected design storm. Other treatments
accomplished with this would be: Coarse screening of
influent, skimming of floatables, sedimentation, and fine
screening prior to discharge.
Application of the SWMM
The SWMM is being used to model approximately 24,000
acres of combined and separate sewered areas in Boston,
Brookline, Cambridge, Chelsea and Somerville. The modeling
process quantifies the combined sewer overflow pollution
problem in terms of quantity of discharge and amount of
pollutants as represented by BOD, SS and coliform bacteria.
The overall statistics on the modeling effort are
as follows:
Total population considered - 900,000
Total area modeled - 24,000 acres
No. of transport models used - 29
No. of elements modeled - 2,035
Total length of conduits modeled - 617,100 feet
No. of subareas modeled - 51?
Average size of subarea - downtown -25 to 30 acres
- parks -75 to 80 acres
Model Demonstration. Although the application of
the model has been demonstrated during its development and
in studies thereafter, it is pertinent to test its use
against any measured values in the locality of its applica-
tion. Measurements taken at an overflow in Cambridge* were
compared favorably with hydrographs generated by the SWMM
under similar conditions.
*Combined Sewer Overflows to the Charles Rivera August,
1972, Process Research, Inc., Cambridge, Mass., for the
Commonwealth of Massachusetts Water Resources Commission.
329
-------
Receiving Water Quality Analysis
The Massachusetts Division of Water Pollution Control
in conjunction with Hydroscience, Inc., are conducting water
quality modeling 'and simulation of quality conditions in
the Harbor to determine necessary steps in achieving remedial
objectives.
The model* was developed by Hydroscience, Inc., and
Figure 19vshows the discretization of Boston Harbor for
purposes df evaluating combined sewer overflow effects.
^Development of Hydrodynamic and Time Variable Water Quality
Models of Boston, Harbor, July 1973 Hydroscience, Inc.,
N. J., for the Commonwealth of Massachusetts Water Resources
Commission,
330
-------
4500 0 4500 9000
•J tei
SCALE IN FEET
FIG. 18 EXISTING MAJOR OVERFLOWS EMMA AREA
331
-------
NOTE DIAGRAM WAS OBTAINED FROM
FIGURE 2. DEVELOPMENT Of
HYDRODYNAMIC AND TIME
VARIABLE rtATER OUAilTV MODELS
OF BOSTON HARBOR, 1973. BY
HYOROSCIENCE INC
\>
FIG. 19 BOSTON HARBOR HYDRODYNAMIC MODEL
332
-------
COMPARATIVE ANALYSIS OF URBAN STORMWATER MODELS
By
Albin Brandstetter*
ABSTRACT
Eighteen mathematical models for the nonsteady simulation of runoff
in urban storm and combined sewerage systems were reviewed in a study
sponsored by the U.S. Environmental Protection Agency. The models were
evaluated on the basis of information published by the model builders and
model users. Seven models were also tested by computer runs using both
hypothetical and real catchment data. Most of the models evaluated include
the nonsteady simulation of the rainfall-runoff process and flow routing in
sewers; a few also Include the simulation of wastewater quality, options
for dimensioning sewerage system components, and features for realtime
control of overflows during rainstorms.
INTRODUCTION
In the past, the most common methods for analyzing and designing urban
storm and combined sewerage systems have been based on the rational formula
for rainfall-runoff computations and steady-state flow equations for sewer
network flow analyses. Recognizing the limitations of these methods for
highly complex systems under dynamic flow conditions, mathematical models
have been developed in recent years to provide better tools for the analy-
sis of existing and planned sewerage systems and system improvements.
In a study sponsored by the U.S. Environmental Protection Agency,
mathematical models simulating dynamic wastewater flow and quality condi-
tions for engineering assessment, control, planning and design of storm
and combined sewerage systems were evaluated. The most promising models
for practical application were tested using hypothetical and real urban
catchment data. The evaluations considered model accuracy, the cost of
model use, computer requirements, data requirements, input data prepara-
tion requirements and output options available to the user. The evalua-
tions were intended to aid the practicing engineer in deciding which of
the models will meet his requirements.
Research Associate, Battelle Pacific Northwest Laboratories, Richland,
Washington.
333
-------
Mathematical models of catchment hydrology are available which handle
all or any combination of the following phenomena:
• Catchment runoff
• Flow routing
• Water quality
• Realtime control
• Design
Some of the models, however, have been developed for nonurban catchments
and are not directly applicable to urban catchments without major
revisions to account for special urban hydrologic phenomena such as run-
off from pervious and impervious areas, flow routing in closed conduits,
and wastewater quality.
Eighteen models were identified which were specifically developed
or adapted for the dynamic simulation of urban catchments. All but one
of these models include the nonsteady simulation of the rainfall-runoff
process and all but one include flow routing in the sewers. Five models
include also the continuous simulation of wastewater quality. Six
models include options for dimensioning sewer pipes and three of them
use mathematical optimization schemes for least-cost design of new sewer
system components. Two models have provisions for the realtime control
of overflows during rainstorms.
A brief review of these models indicates a tremendous diversity in
scope and purpose, mathematical detail, system elements and hydrologic
phenomena being modeled, size of the system that can be handled, data
input requirements, and computer output. This diversity, of course,
is a result of the varying conditions and objectives which govern the
design and evaluation of individual sewerage systems.
The state of development of these models also varies significantly.
Some models have been developed and verified extensively, others have
been developed but not verified, while some have been developed con-
ceptually but not carried to the point of application. In addition,
because no standards exist for evaluating and comparing models, differ-
ing criteria have been used in the evaluations. The rigor with which
the models were tested varies greatly, from intuitive judgments to
graphical comparisons and more demanding statistical analyses.
The above considerations have hindered both the technology transfer
to interested municipalities and consulting engineers and the routine use
of these new methodologies. In fact, due to a lack of information on
data input requirements, model limitations, and program output options
combined with often overstated claims of capabilities, use of this
advanced methodology has received some criticism.
334
-------
The objective of this state-of-the-art survey is therefore to provide
the engineer and planner with a readily available reference containing
brief but precise descriptions and evaluations of all the models avail-
able, thus facilitating his selection of the appropriate model for a
particular application.
MODEL SELECTION
This study reviewed mathematical models which can be used for the
engineering assessment, planning, design, and control of storm and com-
bined sewerage systems. Engineering assessment involves the evaluation
of system performance under various hydrologic and wasteload conditions
to determine problem areas and needs for system improvement. This may
include the determination of undesirable surcharging, backwater and
flooding conditions, the extent of overflows during rainstorms and the
performance of storage facilities and treatment plants. For planning
purposes, the effects of both existing and potential hydrologic con-
ditions and land use decisions on existing and planned sewerage systems
and system components would be studied.
Design involves the computation of sizes of sewers, flow control
and diversion structures, and storage and treatment facilities which
will meet specified performance criteria. Design considerations include
the elimination of undesirable flooding, the reduction of untreated over-
flows, and better use of existing and planned facilities.
Control involves the regulation of flows to improve system effi-
ciency. It may include the regulation of controllable diversion struc-
tures and the operation of storage reservoirs and treatment plants during
rainstorms to maximize the utilization of available system capacity an«f
minimize untreated discharges to receiving waters. The need for potentiaV
system improvements and the implementation of realtime control schemes
for manual or fully automatic control of trttsttwater flows, storage and
treatment may also be considered in this phase.
To serve the purposes described above, mathematical models must
consider the spatial nonuniformity of rainfall; the time-varying runoff
resulting from rainstorms of different intensities and durations; spatial
and temporal variations in dry-weather flows; the effect on hydrograph
shapes of different flow travel times from various catchments; the
attenuation of flood peaks during overland, gutter, and sewer conduit
flow routing; and the operation of flow diversion structures and storage
facilities under dynamic wasteload conditions. The models should be
able to combine the runoff from several catchments and route the waste-
waters within the sewer networks.
Most of the 18 selected models meet the above requirements. The
few exceptions included are models with special provisions or advantages
for the evaluation of important aspects of urban catchments and sewerage
335
-------
systems, or with features which could improve existing models or provide
new modeling capability.
All 18 models were evaluated on the basis of published information
and communication with model developers. In addition, seven of the more
comprehensive models were selected for testing by computer runs using
hypothetical and real catchment data.
In selecting these models, a minimum requirement was the capability
to consider several raingauges, to compute runoff from several catchments,
and to route flows in a converging branch sewer network. Models which
rely heavily on mathematical formulations which cannot be derived readily
from catchment and sewer physical characteristics were not tested, nor
were models whose oversimplification restricts their use unnecessarily,
considering present computer capabilities and the state-of-the-art of
hydrologic modeling.
The 11 models which were evaluated solely on the basis of published
information in reports by model builders and model users are:
1) British Road Research Laboratory (BRRL) Model
2) Chicago Hydrograph Method
3) Colorado State University Urban Runoff Modeling
4) Corps of Engineers Hydrologic Engineering Center Storage, Treatment,
Overflow and Runoff Model
5) Hydrocomp Simulation Program
6) Minneapolis-St. Paul Urban Runoff Model
7) Seattle Computer Augmented Treatment and Disposal System
8) University of Cincinnati Urban Runoff Model
9) University of Illinois Storm Sewer System Simulation Model
10) University of Massachusetts Combined Sewer Control
Simulation Model
11) Wilsey & Ham Urban Watershed System
The seven models which were also tested by computer runs are:
1) Battelle Urban Wastewater Management Model
2) Dorsch Consult Hydrograph-Volume Method
3) Environmental Protection Agency Storm Water Management Model
(SWMM)
336
-------
4) Massachusetts Institute of Technology Urban Watershed Model
5) Metropolitan Sanitary District of Greater Chicago Flow
Simulation Program
6) SOGREAH Looped Sewer Model
7) Water Resources Engineers Storm Water Management Model
A review of all 18 models and preliminary results of the numerical
testing of four models is presented in the following sections.
MODEL REVIEWS
The two types of tables presented in the following pages provide a
quick overview of the features of each model. The first type, Table 1,
lists major model features and indicates whether a certain phenomenon is
being modeled or considered by a model. The table consequently indicates
the comprehensiveness of each model. The second type, Tables 2 through 5,
expands on Table 1 by listing additional categories, brief statements
indicating the mathematical formulations for particular phenomena, model
limitations, and other details which would be helpful in assessing the
applicability of a model for a particular purpose.
It is impossible to describe all the features of each model in a
state-of-the-art review and to discuss every model advantage and limitation
based solely on information provided by the model builders and selected
numerical tests. This would be possible perhaps for evaluations of
mathematical models of individual phenomena, but it becomes a tremendous
task when the interactions of all the submodels of the comprehensive
models in this study must be considered. Only extensive use and experi-
mentation with a particular model will reveal all its advantages and
shortcomings. Most of the models are being revised, expanded, and improved
continuously by the model developers. New versions of some models are
being developed by various model users.
It is hoped, however, that this review will provide sufficient infor-
mation to the potential user to aid him in selecting the appropriate model
for his needs. Once the user has decided that a model may be of interest
to him, careful study of the documents referenced for the model is recom-
mended before its computer implementation. This would familiarize him
with some model details which may be important to him but cannot be
described sufficiently in this review.
The model comparisons emphasize evaluation of the complete model,
rather than detailed critiques of submodels of individual phenomena.
Formulations of individual phenomena are discussed only in the context
of their use in the complete model, and no attempt is made to discuss
all available models for each phenomenon. As an example, a tremendous
variety of models exist for simulating the rainfall-runoff process alone.
No attempt was made to list all of them. Rather, only those used in the
337
-------
Table 1. COMPARISON OF MAJOR MODEL CATEGORIES
MISCELLNIIOUS
J
II
enetto rum
ommsrn
CO
CO
00
Vt. MOL
nmuin or
cacuujat
UUVKHZTT OP
•Utt MMOK38
-------
Table 2. COMPARISON OF HYDROLOGIC FEATURES OF MODELS
Battollo-
Horthweat
•xltiah Koad
•Maareh Labor-
atory
Chicago rlov
limulation
Chloago Hydro-
gnph Method
Colorado tut*
Unlvertlty
Carpi of
Cngineeri
Donah Conjult
environmental
Protection
Agency
•ydroaomp
•uaaahueette
Institute of
nahnalofy
liAneapOlie-
tt. Paul
Seattle
•ogreeh
Onivortlty of
linalnftetl
Oniveraity of
Xllinole
onlveriity of
Mataaahutettt
•MOT BMOIinet
Engineer!
Mlleey ud lam
•ample
catchment
inflow
xee
lea
let
Yet
Yet, but only
Into tingle pip*
•tring
For one catch-
ment only
tmm
y«.
Tn
Y»
T««
tioh oitohawit
ropr«**at«d in-
dcpndntly
!••
XM
In
IM, bat only
into *lngl« pip*
•trin«
Tot
TO!
Dir-DMthcc Uov
Bourly, d«ily,
pnttcnu*
Straight liM*
Bourly flow*
ttydroonph**
Uydr09r«ph>>
•o
Diunul pattern
eenputod fxoa
lend USA
lourly, daily,
and >«uoul
^attarna* or
oOBpot*d tnm
land am*
Bydraoraphi*
Kydrofrapha*
Diurnal pattern*
Diurnal patten*
Bydrofrapha*
Mo
•ydrooraphe*
COM taut flov*
Avg. flov*
tram termed into
tonrly and daily
ptttarni bated
on land uee
No
Snboatetaaant
preaipltation
W>i«htad •««.
of amril
ratngaoei
for each
aubaatchnent
Neighted a*«.
of aeveral
raingaaea
for aaeb
ittboatoheent
net Jeor« than
one par aub-
flatekaent
naiglited avg.
of aewaral
raingagaa
for eaoh
He
One rainaag*
for tingle
oatoheent
One rainoage
par eubeetobMat.
•p«ial ctetit-
tloal anelyaee
One rainqtge
per lubattchBent
Peroent of gage
preeip., one per
eabeatohnant
One ralngagi
per adwatahBent,
•torn can be
•e«ed eeroat
oatohment
Maighted avg. of
•everal raln-
gegei for eaoh
nifinitnlineiil
Six raingagea
lynthetio dealgn
•ton ooeputed
From aubeateh*
•tnt oharaotar-
ietioe
One raiagage
for entire
aewerage eyatea
lei**
One raingage
for entire
aewerage ayete««
generation of
•yn* rainfall
Up to three
raingtgee per
nbeatonnvnt
One raingage
foe entire
•operation
MO
•o
nonlinear
fuietioa of
aoil •oiatura
and daily avg.
air tenperatura
rized function.
Independent of
•eteorolegieal
variation!
Ho
Oonetant rate*
between
ninatona for
eaoh Booth
•0
•a
Baaed on
•matured poten-
tial evaporation
and available
•ell tan
Paaaan agu.
•0
He
Mo
•e
Ho
HO
He
Ho
Snow
aocomulatlon
and nelt
MO
MO
Bated on Bureau
of Meelamation
Engineering
Nomograph 19
method
Mo
No
Degree-day
method
Mo
Mo
Bated on Corpt
of Engineer!
method
Ho
HO
He
Me
He
Baaed en Cerpt
of Engineer!
method
»»
Ho
He
rroien ground
No
Mo
No
Ho
No
Mo
NO
No
Mat exchange
equ.
No
No
HO
No
Mo
He
Mo
Ho
Mo
Infiltration on
inparvloua areea
oay functioa(
dependent on
catchment moil-
tun
Ho
Ho
Ho.
No
Conetant frac-
tion of rainfall
minut depreeaion
itoraga
Mo
Mo
No
No
Ixponeriti*!
decay function
Bmpirieal
fiuwtiona**
ayntaatio
rainfall ueete
Me
•erton equ.
He
Conatant rate*
Noted Blank
•lank apace indloatea that relevant inlocmatlon it not available.
•Provided aa input data.
••Mathematical formulation and/or method of aolution net available.
339
-------
Table 2 (Continued)
otter rainfall
loeeea on iaper-
viou* arena
Initial loa«*
before runoff
beglne
LOM u function
Of tla»a
Ho
Initial loaa*
before runoff
beglaa, n-
atone
•a
Conatent ton
rate* aoaifled
by evaporation
•a depreaaion
etorage on roofe,
initial loaa OB
•treat* before
runoff begiae
initial loaa*
baton runoff
begine
Initial Ion*
•oalflad by
evaporation bo-
tan runoff
beglaa
Initial loee*
boforo runoff
begin.
No
Synthetic
rainfall exooaa
Dapreeelon loci
function of
rainfall
lea**
Initial leu*
bo fora runoff
haglna
Initial teat*
baton runoff
boflna
Tee**
f tor* runoff
from inporrioua
nroaa
Onit
hydrograph
Minfall eneea
routed Ming
tijw-area
oum
Conatant frac-
tion of ovarland
flow atorage
Hnrd*e aqu.
Rydrographa*
Conatant frac-
tion of rainfall
•Inua depnaeion
etoraga
Kinematic MO
aqu.
Kiiioaatio wan
aqu.. with
Manning aqu.
Unonatie van
aqu., with
Manning aqu.
Kinematic wan
equ. , with
Manning aqu.
Unit
hydrograph
Onit
bydrograph
Muakingun nMhod
Kinontic wnwo
aqu., with
Manning aqu.
Kinonatio vavo
aqu., with
Manning aqu.
Conatant f no-
tion of rainfall
•inua initial
loaa
Kinanatie van
aqu.. with
Manning aqu.
Xlnonatle van
aqa., with
Manning aqa.
Infiltration on
porvioua araaa
Modifiad >oltan
aqu., dapandant
on aoil noiaturo
no
rolynonial r»-
graMion ogu. ,
dopondant on
aoil noiatura
Modifiad
Horton'a oqu. ,
dopandant on
•oil nniatura
No
Conatant frac-
tion of rainfall
•inua dopraaaion
atoraqa
Option to u««
Borton or
•oltan aqu.
Borton aqu.,
indapandant of
aoil aeiatura
nvvirieal aqu.,
dapandant on
aoil aoiitura
opt. to nao nor-
ton aqu., Holtan
aqu., KB notnod,
runoff ooaffi-
eiant nathod
Modifiad Boltan
aqu., dapandant
on aoil noiaturo
mpirioal
funatioaa"
lynthotlo
rainfall axoaaa
norton aqu. ,
vith tin* offaat
baaad on actual
rainfall, indap.
of aoil aoiatun
norton aqa.
Mo
Borton aqu.
*••••
Othar rainfall
loaaaa on
parvloaa araaa
Initial loaa*
baton runoff
baa in*
Mo
Mo
function of
•oiatnra Condi*
tlona, racovary
botvaon atona
no
Cbnatant loaa
rata* nodifiad
by anporatioa
•oiatnra
condltlou
baton ztuwf f
bagina
Initial loaa*
•odiflad by
evaporation ba-
ton runoff
bagiM
Initial loaa*
baton rnnoff
bogina
Mo
•ynthotio
rainfall axeaaa
function of
rainfall and
infiltration
baton runoff
naqina
Xaa**
Iton rnnoff
from parvloaa
araaa
lyntnotic unit
hydregmph
Mo
llnaar atoraga
routing of onr-
laad flow ator-
aga
'aaard'a aqu.
•ydrographa*
Conatant frac-
tion of rainfall
minua dapraaaion
atoraoa
Klnanmtie van
aqu.
tin an tic van
aqu., with
Manning aqu.
flnaaaUo van
aqu.. vith
Manning aqu.
flnaaatia van
aqu., vith
Maonini aqu.
Onit
hydrograph
Onit
hydrograph
ttiiiMnfliwi nvtnod
ftoraga routing
vith Manning
aqu.
xinanatie van
aqu., vith
Manning aqu.
Mo
Kinaaatio van
aqu., vith
Manning oqu.
Klnaauttio van
aqu., vith
Manning aqu.
onundvatw
aiaulation
MO
Mo
BO-
MO
Mo
Mo
MO
MO
Monlinaar
funetlona for
atoraga and
aurfan flow
contribution
MO
BO
Mo
Mo
Mo
MO
Mo
Mo
Mo
Outtar flow
no
Avg. tranl tin*
fro. Manning
aqu.
•0
Kinaaatic vava
aqu., vith
Manning aqu.
Mo
Mo
Conputad aa
aurfaco runoff
or channel
routing
-• "'- van
aqu., vith
Manning aqu.
Mo
Kinanmtic vava
aqu., vith
Manning aqu.
Ho
Mo
Mo
Outflow aquala
inflow during
Mn* tUa
intanral' •
Unaaatlo van
aqu., vith
Manning aqu.
Ma
rinaaatio van
aqu., with
Manning aqu.
Unanatic van
aqu., vith
Manning aqu.
Mater balance
batoaan atone
MO
MO
nooountlag of
aoil noiatura,
overland and
onannal atoraga
Accounting of
•oil noiature,
deprea»ionj over-
land and
ohannel storage
Ho
Accounting of
dapraaaion and
reaervoir ator-
age
•aoovary of
depraaaion
atorage on
pervioua araaa
MO
Accounting of
•oil aolatun,
gronndwater and
aurtaoa vatar
atorago
Mo
Mo
•uaoff adjuitad
baaad on
prevlou atom**
Mo
MO
MO
naoovary of da-
preaaion atoraga
Mo
MO
340
-------
Table 3. COMPARISON OF HYDRAULIC FEATURES OF MODELS
•attella
Mortbweat
•ritlah lout
Reeeareh
Laboratory
Chicago flow
Simulation
Chicago Bydro-
greph Method
Colorado State
Oniverelty
Corps of
Bnglneere
Doraoh
Ceneult
Knvlronmantal
Protection
Hydroeomp
Maaaachuaetta
Inatltute of
Technology
Minneapolia-
St. leul
Seattle
Sogreah
Univeraity of
Cincinnati
Unlvaralty of
Illinoia
University of
Neaaachueette
Hater Kaeouroea
Bngineera
Vilaey end
•am
Boteat Blank apa
Open channel
network
Converging
branch network
Converging
branch network
Converging
branch network
Converging
branoh network
Single atrlng of
plpaa
Ho flow routing
Converging and
diverging branch
and loop network
without flow
reveraal
Converging
branch network
Converging
branch network
Converging
branch network
Converging
branch natvork
Converging
branch network
Converging and
diverging branch
and loop network
Converging
branch network
Converging
branch network
•ingle string of
plpaa
Converging and
diverging branch
and loop network
Converging
branch network
Free eurf ace
flew
Kinematic wave
equ., oharacter-
latic aolutlon.
Manning equ.
Storage routing.
Kenning equ.
Storage routing.
Manning equ.
Storage touting,
Manning equ.
Dynamic wave
equ., varioua
•olutlone, Man-
ning or oaroy-
Molebaoh equ.
Ho
Dynamic wave
equ., implicit
finite diff.
aol ., Manning or
Prandtl equ.
Kinematic wave
equ., explicit
finite diff.
aol.. Manning
equ.
Kinematic weve
equ., vith
diffuelon term**
Manning equ.
Kinematic weve
equ., explicit
or implicit fin-
ite diff, aol.,
Manning equ.
Frogreaalve avg.
lag method
Dynamic wave
equ.. explicit
finite diff. aol.,
Manning equ.
Dynamic wave .
equ., implicit
finite dltf.
aol.. Kenning
equ.
Translation of
hydrogzaph by
avg. flow time
of hydrograph.
Manning equ.
Dyn. wave equ.,
explicit finite
dltf. eel. of
char, aqu.'a,
Dejray-Neiabeoh
Dynamic wave
equ., Implicit
finite diff,
aol. Kenning
equ.
equ., explicit
finite dltf.
aol. .Manning
equ.
wave equ.,**
Manning equ.
Backwater
effect*
Mo
Me
Me
Mo
Dynamic wave
equ.
Ho
Dynamic wave
equ.
For known depth
va. •torage re-
letionehlp et
not more than
two looationa
Diffusion term
in kinematic
wave equ.
Ho
Ho
Dynamic wave
equ.
Dynamic wave
equ.
Ho
Dynamic wave
equ.
Yea
Dynamic weve
equ.
He
Plow rovereel
Mo
Ho
Mo
Ho
Dynamic wave
equ.
No
No
Ho
Ho
Ho
Ho
HO
Dynamic wave
equ.
No
Dynamic wave
equ.
Dynamic weve •
equ.
He
Surcharging and
preaeure flow
Considered in-
dependently in
eaoh pipe
Ho
No
No
Ho
No
Yea**
Conaidered In-
dependently in
eaoh pipe
No
NO
Ho
Conaidered only
for trunk
atorage upstream
of regulators
yes"
Ho
Ho
HO
Conaidered
independently
at eaoh
Junction**
Ho
Different pipe
cross-sec tiona
Circular pipe
Circular and
rectangular pipe,
trapaioidal
channel
Circular pipe,
trapetoidal
channel with
flood plain
Circular pipa,
trapaioidal
channel
Circular plpa
Ho
Varioua epeci-
fied ehapea plue
arbitrary ahape
Twelve specified
•hapea plua
three arbitrary
•hapea
Circular pipe,
trapaioidal
channel with
flood plain
Circular pipe,
triangular gut-
ter, triangular
and rectangular
channel
Circular pipa
Circular and
horeeahoe pipe
Circular and
egg-shaped pipe,
trapecoidal
channel plua ar-
bitrary ahapea
Circular pipe
and rectangular
channel
Circular pipe
Circular pipa
Clro., root.,
horaaehoe, bas-
ket handle and
egg-shaped pipe,
trap, channel
Different
diveraion
structures
Four common
typea of eri-
fico/volr com-
binations
Ho
Mo
No
Ho
Undefined ahape
Heir, diverging
pipe branches
Two types i en-
apecified ahape
or weir
Diversion
hydrographs*
Mo
Three typea of
common orifice/
veir comb..
three special
etructuree
Six typea of
gate control
atruoturee
Rectangular
weir and orifice
including time-
varying gate
settings
Ho
Hair or gate
flew control and
diversion at
outlet
Ho
Hairs, gates or
orifices, with
or without tide
gates
No
BO indicate-, that relevant intonation Is met available.
341
-------
Table 3 (Continued)
Divarelon
computation
Critic* and Mir
aqu. nagleetlng
downatraaB con~
dltiona
No
No
Do
No
Bxceaa ov*r
atoraga and
treatment eapae-
ity overflows
Hair «ju.,
oonaldarlng
upetraaB and
downatraaB
eoBditioiu
Conatant diver-
•ioa or wair
a«u. naglacting
downatraan con*
ditlona
Divereion hydro-
grapha*
Me
orliica and wair
equ.'e nealact-
ing dovnatreaB
oonditiona
orif tee and Mir
aqu.'a oonaidar-
ing upatraan and
downatraaB oon-
ditiona
Orifioa and walr
equ. 'a oonaider-
ing npetreaB and
dovnatzaa eand.
i flow revaraal
•o
•ale aqu. or any
diaeharge *a.
depth function
•0
orlfloe and Mix
equ.'a oonaider-
ing upatraeB and
downatreaB oond.
* flow reveraal
•0
Puaplng atatlona
No
Ho
Ho
HO
•0
HO
No
Conatant dia-
ohar«a 1C nat
vail axaaada
apaetflad
dapth
Moving
aohadula*
No
Tvo typaa i
proportional and
oonatant rata
»aa*«
roaplnf rata oan
ba outlat bound-
ary oondition
No
Dp to 3-«t*ja
poapina atatlona,
rnla ourva '
ooaritlon
No
Overflow atoraaa
For undaflnad
ahapa, only In-
flow and TOlua*
coBputad
Ho
HO
•o
Exeaaa ovar
traattaat
eapaoity la
•torad
Ona only« Bodala
(our typaa of
outlat
atrootoraa
Ta»««
NO
lalina atoraa*
ahapa « no dapth
ooBputationa
traatmant
capacity la
aterad
four typai of
ontlat
•tnoturaa
both apillway
and pooar tur-
bina flow
Itoraga In trunl
aavara upatraan
of ragulaton
•e
Storaga
eoBputation
CoBputaa inflow
and voluaa of
otrarflow atoraga
axoaading apaci-
fiad nax. out-
flow
No
Storad vatar ra-
tarDa to traat-
•ant plant
imetlon of
ahapa, inflow
and outflow
Junction
of ahapa. Inflow
and outflow
Hula eurva op-
aratlon or pra-
aat outflowa,
inolodin* povar
flow*
•0
Function of
ahapa, Inflow
and outflow
Vaa"
•0
ruootion
of ahapa. Inflow
and outflow
Ho
Function of
ahapa, inflow
and outflow
•o
Infiltration
into aawara or
opan channala
No
HO
HO
»0
ao
Conatant rata*
Conatant rata*
Nonlinaar
function
dapandant on
aoil nolatura
•o
No
Daciy function
of atom
Hydrographa*
Ho
No
No
Conatant rata*
NO •
:onputaa ataoa
hydrographa
For aach pipa
raach
No
For aaeh plpa
and channal
raaeh
No
At any point
•0
For aaeh aawar
ayatam alaawnt
Yaa, but not
priatad
For aaoh ohannal
raaeh and ator-
aga
For aaeh plpa
and ehannal
raaeh
Ho
For aaoh inflow
point and
Junction
For aach aawar
ayataB alaaant
i eoBputational
point
For aaeh
plpa taach, but
not printad
For aaeh aawar
ayataB alaawwt
and at any
daairad polnta
For aaeh plpa
raaon
For aaeh aawar
Junction
Yea
onputaa flow
valocltiaa
For aaeh pipa
aach
No
Ho
Yaa, but not
printad
At any point
No
For aach aawar
ayataB alanant,
but printa only
aalaotad valuaa
Vaa, but not
printad
•For aaoh pipa
and ohannal
raaoh and
etoraa*
For aaoh pipa
and ehannal
raaoh
Ho
For aaeh plpa
raaoh
For aaeh aawar
ayataB aleaant
I ooBputational
point
For aaeh pipa
raaeh , but pat
printad
For aaeh aawar
ayataB alaaant
and at any
daairad pointa
For aaeh pipa
raaoh
for aaeh pip*
and ehannal
raach
Yaa
342
-------
Table 4. COMPARISON OF WATER QUALITY FEATURES OF MODELS
Battalia
Mortnvaat
British load
Raaaarch Labor-
atory
Chicago flow
Simulation
Chicago Hydro-
graph Hathod
Colorado State
onlvaraity
Corp* of
Enginaara
Doraeh conault
BgwirOHaantal
Protection
Aganey
•ydroeoBB
laaaaflhnaat**
[natitUta Of
Taeanolooy
Minaaanolia-
St. falll
•aattla
•ograab.
gnlvoraity of
Cincinnati
Unlvaraity of
Illlnola
OUiraralty Of
Naaaaohuaatta
latav Bauourooa
Infllnaara
Mlaay and •••
MOtaai lltBk anal
•Proviawi 1
"MathaBBtli
Quality
conatituanta
7 arbitrary oon-
aarvatlva eon-
atltuanta
Mo
NO
Mo
No
Suapandad and
aattlaaola aol-
iaa, BOD, nitro-
aan, and phon-
pnato
(aourat* anoal
In praparatlon
Suapandad and
aattlaabla
•oltda, MD, COD,
oollform, M, POi
oil and oraaaa
17 oonatituanta
including watar
taiiBwcatiir*
Ho
Ho
•0
OonaarratiTa
oonatituanta
•o
Mo
•0
liupndad and
aattlaabla aol-
ida, MD, I, fO,
oil and graaaa,
17 ooni. oonat.
M
Sindioafaj tluit
Input data.
tal fennlntiea w
Dry-Mather
quality
Hourly, dally
and aaaaonal
pattarna*
Mo
Do
Mo
Mo
Mo
MO
Hourly, dally
and auaonal
pattonw* or
ooau>ut«d froB
land uaa
Xnf lev oonoaB-
trationa* or
Hfnl limr CUM*
tlon of inf lov
hyaroaraph
Mo
He
MO
Ho
Ho
Mo
•0
Krg. quality*
tranaromad into
bouly and duly
pattarna baaad
on land «aa
Mo
ralaraat UfaoMl
I/or Bathed of Mil
Stora runoff
Quality
Linear ragroa-
•lon aou. , with
a ton diacnaroa
and voluH
Mo
Mo
Mo
Mo
Moalinjau funct.
of oatoJaaant
char., poll.
aoeuBulation,
and runoff
Mo
•imltnoir fuaet.
of catefaMnt
okar., pall.
aoeuBulatlon,
and runoff
Honlinaar fnnct.
of catcoDant
obar., poll.
aoouMulatlon,
and runoff
Mo
Mo
Mo
Mo
Mo
Mo
MO
Moallitaar foaet.
of oalola»nt
•kar., poll.
ammunition,
and ranoff
on ii not arailabl
itian aat vrallabl
Quality
Intaraotlona on
catohBanta
Mo
Mo
Mo
Mo
Mo
•OO, nltrogan
and phoapnata
dap. on avapaad-
ad and aattla-
abla aolida
no
MD, nitnvm
and phoapnata
dap. on auapand-
ad and aattla-
abla aolida
Mo
Mo
Mo
Mo
Mo
MO
Mo
Me
All oooatltaanta
haialiiit OB
wttlaabla aol-
ida
MO
1.
a.
Quality routing
in ohannala
•ura advaotion,
•ixing batman
auocaiaiv* tin*
atapa
no
Mo
Mo
Mo
Mo
MO
fora advactlon,
Bixing batvaan
aaeoaaaiva tln»
atapa
mra advaotion,
vaiantad mUiing
batvaan aueeaa-
aiva tinia atapa
Mo
Mo
Mo
Mra advactlon
Mo
MO
Mo
fun adraction
Ho
Sadinwntatlon
and aoour in
ohannala
No
No
Ho
HO
HO
Ho, but land
aurfaoa aroaion
by univaraal
aoil loaa aou.
Do
Buapaadad aol-
ida, oonaidaring
partiela aita
diatribution
Ho
Ho
Mo
Ho
Mo
Mo
MO
Ho
Ho
Mo
Quality
raaetioaa in
enannala
Ho
Ho
Ho
•o
MO
Ho
HO
ririt-ordar da-
cay vlthout
iataraetioaa
tor MO
Varioua raae-
tioaa and intar-
actiooa
Ho
Ho
Ho
Mo
Mo
Ho
Mo
HO
Ho
343
-------
Table 4 (Continued)
Quality routing
throng* atorag*
M
•0
•0
•o
Ho
Mug flow
ring now or
iUtUtUMOW
pjlTlaW
tnm«» OM to
layara, outflott
AroB any Uy«
*>
•id**
•o
luw duHtlon
tor o»»<*h»>lno
duality
raaotiona in
•toraga
•o
•0
•o
•o
•o
Ho
nwtleu wd
intoTHtioao**
Ovctflov
tr*>tMnt
Tabular Junction
of flow and 000-
etatratian
No
Ho
•o
tafluoat (low
and pollutant*
only, aop.'a (or
nia* phraioal/
ehamicii pro-
uiioa tnataut
Tabular function
of flow and con-
centration
HO
No
Infliwnt
flov and pollu-
tanta
Ouality
iataraatiou
durlaf trcatavat
Mo
Colifom na»*ai
d*D«B4ant oa
Mipoadod aolida
Quality oalaao*
batwaan atoru
•o
BO
•O
•o
»0
Function ot pol-
lutant accumula-
tion and (traat
«M*elix
•o
rvMUon of pol-
lutant aeoujNla-
tioa and atroat
•vaaplng
•o
•o
•o
•o
•a
•a
•a
•o
•o •
Motiving vatar
flow aiaulation
•a
fai
vaa
Ya«
HO
•o" ~
KI
ror oa. a»d t«o-
diaaaaioaal
npraaaataUon
>••
Yaa
Do
DO
toa
Taa
•o
•o
Yaa
•o
RaMlTiiig vatar
quality
ilMulatloa
Bo
HO
•o
fS~ •
•o
No
taparat* •odal
in praparatiaD
ror on* and tav-
dlMialanal
Taa
•o
•o
No
•o
•o
•o
•o
r«
•o
344
-------
Table 5. COMPARISON OF MISCELLANEOUS FEATURES OF MODELS
•attelle-
Northweet
•rltiah Mood
Mary
Ckloegoriow
MaBletiOB
Chloego Hydro-
graph Natbod
Colorado f tat*
Oaivanity
•nglneen
Donah coBMlt
•nvii mBBWil «1
tretB-tlaa
•***•-
•ydrocoop
•eaaaohneetta
laatltuta of
teohnoloay
Hi nneepnll a~
ft. Faal
feattla
•o-raali
mivanity of
UBOiBBatl
univeralty of
XUiaeli
ualvanity of
Haaaaohaeette
•flneere
aieey and
CeatiKBOua
aiaulatioa
finale atora
period, no water
and quality bal-
aaoa between
•ten*
•ingla atora
balance between
•tone
f lagla aton
period, water
balanae between
•ton*
Cantinuoua flow
routing in
•ingla *trina of
Continnoua
for *everal
yeara
tiagl* aton
period, no water
balance between
•ton*
Magi* itora
period, no water
and quality bal-
atoraa
Coatinuoua for
COBtiauoua,
••rat I. for anow
f Ingle eton
period, no water
aelanoa between
finale atom
period, limited
water balanoe
tingle •ton
period, no water
>alanoa between
atom
fiagla a tor-
par lod, BO water
Mlanoa between
•torae
final* atorn
period onleaa
ootttiauou
hydrographa are
input
wjpt for enow
•ingle aton
period, BO water
md quality bal-
anoa between
•tona
•ingle •ton
Period, no water
wlanoa between
Ha* interval
Conatant Inter-
vale IB Binntea
(Bin. - 1. BU.
Constant later-
Conatant Inter-
vale in adautea
Conatant inter-
vale In Blnutaa
(1 Binute req'd
for rain, oaa be
aore for runoff)
oapaBdent on
auaarloal ata-
blllty condition
and eolected die-
tanoe interval
Nuet be 1 hour
Conatant Inter-
val la ainotea,
reatrloted by
nuaarioal ata-
bility ooadltloa
Conatant inter-
vala la alnutee
Mlautea,
dally, dally or
••Bi-ao.i ooapu.
uee aa. latarvla
Conatant inter-
val la alnutea
Conatant inter-
val In aeeonde
Conatant Inter-
val In nlnuta*
Conatant Inter-
val la alnutea.
oan be dlff . for
eaeh hour
Conatant inter-
val la alnutea
Conatant later-
val in aaoonde,
reatrloted by
nuaorioal ata-
bility ooadltloa
Huat be 1 hour
Conataat
intervel, re-
•trieted by
Buaarioal ata-
blllty ooadltloa
Mia. and au.
tin period for
iiaulatlon
Mia. 1 hr, BU.
f an or 5i tiaa
atapa
NO apparant
aia. , BU, 9M
tlna atepa
Mia. 24 noun,
BO apparent BU.
Mo apparent
ala., au. 90«
Blnutaa
MO apparent
aia. or au.
Nln. 24 hour a,
no apparent au.
No apparent
Bin., au. 900
tiaa eteoe
Ho apparent
BiB,, BU. ISO
tiaa atapa
NO apparent
_ i^^ ii
aln. or BU.
Mia. J boor a,
au. 14 houra
Mia. 1 hour, no
apparent au.
Mo apparant
aln., au. Uo
tiaa itep*
No apparent
ala, or au.
Mia. 1 hour,
au. 1 aonth
Me aia.
allow* Inpot
of initial
ooodltiono
cal nheint aoia-
turo oond.,
Init. oond. not
naann for
routing
ninh-aii aoli-
not for ohaanela
Mo
No
conetant dla-
ohar-a in entire
pipe •triaf
CatnUaant aola-
ture and quality
oond»not for
ohaanele end
•toroai
Catehaant aoU-
tnre oonditlou,
aot midia for
routing
catohaaat aola-
tnra and qaallty
oond., not for
ohennela and
atoraoe
roc all aawar
eatoBaaat Bole-
tore condition* ,
not for ohanaela
Catehaant aola~
tur* oonditlona,
not for onannela
No
ayataa eleaint
catohaant aoia-
tan oondltiou,
not for channel*
Conatant flow
tec all eewer
•man elaaeat*
Catohaent aola-
ttr* ooadltlou,
aot for -fc—— i-
DMlga
ocB-ntatiow
Leaat-soet alia*
of eewen, •tor-
ege and treat-
Beat faeilltlea
No
HO'
flw* pipe* for
peak flow, no
coat ooaputa-
tiona
No
No
NO
Coat* of
eeorage and
treatatat, *liaa
pipei to elialB-
ate auronarving
Mo
Leut-ooat alaee
of pipea, ator-
age and treat-
ant facilltio*
MO
No
Ho
Mo
flu* of pipe*
fox peak flew,
or au. depth
no ooet ooaputa-
tlOM
Mo
Mo
•iiee of pipea
for peak ilow
Lea*t-oo*t
do* ion
computation*
Modified gradi-
ent technique
conalderlng flow
and quality
MO
No
No
Mo
No
NO
MO
No
TJ near prograa*
flow only
Mo
No
Mo
Mo
Dyaaaia proanaa-
Ing with coat
and riak oouid-
eration in prep*
•ratio*
No
Ho
No
aealtlne control
optlaliatloa
Dynamic pro-
craBaiii— for
regulator opera-
tion
NO
Mo
No
Different
aohaae* ware
Investigated
MO
No
NO
flow (oreoaatiag
atioa without
optlalutlon
Mo
Iterative ran*
with trial regu-
lator aettinga
Autoa. regulator
operation, uiing
ml* curve*,
*«eh operated
indepeadwtly
No
Nc
No
Mo
No
No
, -.—<. e-ani Ia41c*tee tnec nuwan uzoaauoB la aot available.
••aMa.*B«-o*lfon»Utloa ant/or ntnod of •wlatlon aot avallabM.
345
-------
Table 5 (Continued)
Error maaaagaa
Umltad
Umltad
limited- ' "
Limited
Moderate
•xtenaive, pitta
eeparata data
oneakiwi pro-
„«, (350
maeaadOe)
Moderate
tttanalva
Moderate
Limited
Adequate for
real-time ovar-
flow control
Moderate
Nona
Moderate
Moderate
Principal
printed output
Iain, depth, vel..
dlaoh., itor,ou*l.
ooiti and aliea
of MHU*, (tor.
and treatment
Dlaoharge
lain. »ov,
depth, dle-
oharga, (tonga
Dlaoharge, lavar
dtamateri
Depth, ««aloa-
Ity, diaoharga
Iain, dlaeharge,
atoraga i
quality, land
aroaion
Daptb. dlaekarga,
(eaparataly for
aanltary and
•ton watar)
Rain, dlaoharge,
quality, coin
of Btor*aa and
traatjMnt, aawar
• Isaa
Rain, •nan, mail
aniatura, dapth,
diacharja, ator-
aoa, quality
Haln, dapth, dia-
olwraatooit* and
ai*« of a«w*ra,
atoraaa and
traatxant
Iain, diachart*.
"catoiw*nt aeia-
tura eonditioaa,
ragulator atatua
Rain, regulator
flow dapth and
control atatua
Dapth, nloeity.
diaeharga * prea-
aura, atraat
flooding dapth
Diaeharga
Dapth, velocity.
dleeharge, aawr
dlaMtari
Depth, dlioharge
Depth, veloolty,
dlaehazge,
quality
Dlaohacge, ae
-------
comprehensive urban hydrologic models were reviewed. This type of evalua
tion provides the engineer with the information he needs to select a
comprehensive model, but does not require him to be familiar with the
many available variations for modeling each urban hydrologic phenomenon.
Battelle Urban Wastewater Management Model
The Battelle Urban Wastewater Management Model is intended primarily
for the simulation of large urban catchments. It simulates the time-
varying runoff and water quality in combined sewerage systems consisting
of several catchments and a converging branch sewer network (Brandstetter,
et al.( 1973). Up to seven conservative water quality constituents can
be modeled. The model can be used for realtime control of overflows
during rainstorms and for least-cost design of sewerage system modifica-
tions. The model is limited to the simulation of single runoff events.
Rainfall for each catchment is computed as the weighted average of
rain at several raingauges. Losses are subtracted from rainfall for both
pervious and impervious areas. A modified Hoi tan's equation is used to
compute infiltration on pervious areas accounting for changes in soil
moisture conditions. Runoff from rainfall excess is computed using
synthetic unit hydrographs derived from physical catchment characteristics
by a method of the U.S. Soil Conservation Service.
Stormwater quality is computed from stormwater runoff using regression
equations including seasonal adjustment factors. Dry-weather flow and
quality is provided as input data in the form of average values for each
catchment and adjustment factors for diurnal, weekday and seasonal varia-
tions.
The combined flow is routed through circular sewers using the kine-
matic wave equations. Downstream flow control, backwater, flow reversal,
surcharging, and pressure flow are not considered. Diversions are com-
puted at regulators using the appropriate weir and orifice equations.
The filling of overflow storage facilities is computed, but not the
return of the stored wastewaters to the sewerage system.
The combined wastewater quality is routed through the sewers
using pure advection, neglecting scour and deposition, dispersion, decay,
reactions and interactions. Tables relating flow and concentration with
treatment efficiency are used to determine wastewater quality improve-
ment at treatment plants and overflow treatment facilities.
For realtime control of overflows during rainstorms, diversions at
controllable regulators are computed using dynamic programming to maxi-
mize the utilization of available sewer, storage and treatment capacities
and minimize pollutant overflows. For design studies, sizes of sewers,
overflow storage facilities, treatment plants, and overflow treatment
facilities are computed which will minimize costs for specified constraints
on the quality of overflows and treatment plant effluents.
347
-------
Considering the complexity of the model, it is comparatively easy to
use. The input data are arranged in logical groups, the program, however,
lacks input data diagnostics. Program improvements are necessary for the
efficient implementation of the realtime control optimization. The model
can be used in different modes, using various combinations of flow simu-
lation, quality simulation, overflow control and design optimization.
Program output Includes printed and plotted rainfall intensities,
water levels, discharges, velocities, and concentrations. Sizes and costs
of sewers, storage and treatment plants are printed for the design option.
The water flow and quality simulation is programmed in Fortran IV for a
DEC PDP-9 with 16K words of core memory and the realtime control and design
optimization in Fortran V for a UNIVAC 1108. The two segments can be run
independently or sequentially using batch mode operation. The program and
draft documentation are available. An IBM 360 computer version is operated
by the City of Cleveland (Pew, et al., 1972). The model has been verified
on very limited data.
British Road Research Laboratory Model
The British Road Research Laboratory Model simulates the time-varying
runoff in combined sewerage systems consisting of several catchments and
a converging branch sewer and open channel network (Watkins, 1962;
Terstriep and Stall, 1969). The model computes surface runoff only from
impervious areas and neglects the contribution of pervious areas. The
model is limited to the simulation of single runoff events. Water
quality, realtime control and design features are not included.
Rainfall for each catchment is computed as the weighted average of
rain at several raingauges. Losses are subtracted from the rainfall
according to functions of time provided as input data. The rainfall
excess is routed to inlets using time-area curves computed for each
catchment from Manning's equation. Dry-weather flow is provided as
input data in the form of a linear function of time.
The combined flow is routed through the sewers with a storage routing
technique using average travel times computed for each hydrograph from
Manning's equation. Circular and rectangular pipes and trapezoidal open
channels can be modeled. The model does not consider downstream flow
control, backwater, flow reversal, surcharging, and pressure flow. Flow
control, diversion, storage, and treatment facilities are not modeled.
The program is written in Fortran IV for an IBM 360 and is available.
Program output includes printed and plotted rainfall intensities and dis-
charges. Water levels and velocities are not computed.
The model has been tested extensively with.urban hydro!ogic data
(Stall and Terstriep, 1972). The model appears to produce satisfactory
results for drainage areas of less than 13 km' (5 sq mi) if the impervious
areas directly connected to the storm drainage system comprise more than
15 percent of the drainage area and if the frequency of the storm is less
than 20 years. The model requires a minimum of input data, is easy to
348
-------
use, and provides fairly accurate means of computing runoff from the paved
areas of urban catchments.
The Illinois State Water Survey is completing the development of a
new model based on the British Road Research method which considers also
the runoff contribution from pervious areas and includes a design option
to size circular sewers.
Chicago Flow Simulation Program
The Chicago Flow Simulation Program is intended primarily for the
simulation of large catchments consisting of both sewered and nonsewered
areas (Lanyon and Jackson, 1974). It simulates the time-varying runoff
in combined sewerage systems and nonurban drainage basins consisting of
several catchments and a converging sewer and natural channel network.
The flow routing formulation for natural channels includes provisions
for flow and storage in floodplains. The model can be used for
continuous simulation using hourly or smaller time steps. Water quality,
realtime control and design features are not included.
Storm runoff is computed from rainfall or snowmelt from records of
not more than one raingauge per subcatchment. Snowmelt is computed by
a method of the U.S. Bureau of Reclamation from precipitation, daily
average air temperature and daily average wind velocity. Different
formulations are used for the computation of runoff for sewered and
nonsewered areas.
Immediate runoff of the entire rain falling on impervious areas and
a constant fraction of the rain falling on pervious areas is assumed for
sewered areas. For nonsewered areas, it is assumed that the entire rain
falling on impervious areas becomes runoff, and that a constant fraction
of overland storage runs off during each time interval. For pervious
areas of nonsewered areas, losses are subtracted from rainfall using an
empirical function of rainfall, catchment shape, and soil moisture. An
empirical relationship is used for continuous accounting of soil moisture
considering deep percolation and evaporation. Dry-weather flow is com-
puted from subcatchment population and hourly values of per capita
contribution which are provided as input data.
The combined flow is routed in circular pipes and trapezoidal channels
by a storage routing technique using Manning's equation. The trapezoidal
channel formulation considers also the effects of floodplains of trapezoi-
dal cross section. Downstream flow control, backwater, flow reversal,
surcharging, and pressure flow are not considered. Storage reservoirs
which store all inflow exceeding a specified maximum outflow can be
simulated. No other flow control structures and diversion facilities
are modeled.
The model is very easy to use, but limited in its applicability by
the assumptions inherent in the precipitation-runoff computations which
349
-------
neglect catchment shape, slope and surface roughness for sewered areas and
impervious areas of nonsewered areas, and assume a constant surface rough-
ness for all pervious areas of nonsewered areas. Runoff and routing con-
stants are internal to the program and may have to be changed for applica-
tion of the model in different areas. Model testing on mostly nonurban
areas ranging in size from 12 to 264 km2 (4.7 to 102 sq mi) produced
generally satisfactory results.
The computer program is written in Fortran IV and is available in
an IBM and a CDC 6400 version. Program output includes precipitation
and both tables and line printer plots of stage, discharge and storage.
Flow velocities are not computed.
Chicago Hydrograph Method
The Chicago Hydrograph Method simulates the time-varying runoff in
combined sewerage systems consisting of several catchments and a converg-
ing sewer and open channel network (Tholin and Keifer, 1959; Keifer,
et al., 1970). The model computes diameters of circular pipes for peak
flows. Although it includes formulations for continuous catchment mois-
ture accounting it can be used only for the simulation of single runoff
events due to computer program limitations. Water quality and realtime
control are not included.
Rainfall for each catchment is computed as the weighted average of
several raingauge records or can be computed from internal design rainfall
equations for different frequencies. A modified Norton's equation
accounting for soil moisture changes is used to compute infiltration from
rainfall for pervious areas. Losses are subtracted for both pervious and
impervious areas to account for depression storage. Recovery of infiltra-
tion capacity and depression storage is simulated with internally fixed
equations.
The rainfall excess is routed separately from pervious and impervious
areas to the gutters by a modified storage routing method of Izzard.
Gutter routing of the combined flow is also accomplished by a storage
routing technique. Routing coefficients relating storage with discharge
have to be provided as input data. Equations to compute them from catch-
ment characteristics are provided in the user's manual. Dry-weather
flow can be provided in the form of hydrographs as input data at any
point in the channel or sewer system.
The combined flow is routed in circular pipes and trapezoidal open
channels using a linear storage routing technique. The routing coeffi-
cients relating storage and discharge are computed from the peak flow
of each hydrograph with Manning's equation. Downstream flow control,
backwater, flow reversal, surcharging, and pressure flow are not con-
sidered. Flow control, diversion, and storage facilities are not
modeled.
350
-------
The model includes considerable simplifications in the routing of
overland, gutter, sewer, and channel flow which would appear to limit
its accuracy and are not necessary considering present state-of-the-art
of flow routing techniques. The model is useful primarily for the design
of circular sewer pipes. No model testing using measured urban runoff
data is reported by the model developers.
The computer program is written in Fortran IV for an IBM 1130 and
is available. Program output includes precipitation and discharge and
sewer diameters for the design option. Water levels and flow velocities
except peak velocities are not computed.
Colorado State University Urban Runoff Modeling
The Colorado State University urban runoff modeling efforts are
included in this review because of special research activities which
are expected to provide new and improved techniques for the control of
overflows and the simulation of nonsteady flow in pipes. Two parallel
efforts are being conducted. One effort is concerned with the develop-
ment of concepts for the realtime control of overflows and the other with
physical and numerical modeling of nonsteady flow in circular pipes.
No computer programs have been developed under the realtime control
research effort which can be used directly for realtime applications.
Valuable concepts evolved under this program, however, related to the
future direction of urban wastewater management and the development of
automated realtime wastewater control systems (Smith, et al., 1972;
Grigg, et al., 1973).
Various explicit finite difference solutions of the dynamic wave
equations for gradually varied nonsteady open channel flow were investi-
gated (Yevjevich and Barnes, 1970). The diffusion scheme, Lax-Wendroff
scheme, and specified interval scheme of the method of characteristics
were computer programmed for a single string of circular pipes. The
programs consider input hydrographs at various points, constant initial
flow conditions, upstream and downstream hydraulic control, backwater,
and flow reversal. Surcharging and pressure flow are not considered.
The methods were tested with data obtained from a laboratory conduit.
The specified interval scheme of the method of characteristics was found
to be the most accurate of all methods investigated. Implicit finite
difference schemes were not investigated.
The programs of the three schemes were written in Fortran IV for a
CDC 6400 and 6600 and are available. The programs compute and print
water levels, velocities and discharges at any desired point along a
single pipe string. The programs are not set up for pipe networks and
diversion structures but merit consideration for incorporation into com-
prehensive urban runoff simulation models.
351
-------
Corps of Engineers STORM Model
The Storage, Treatment and Overflow Model (STORM) of the Corps of
Engineers Hydrologic Engineering Center is intended primarily for the
evaluation of stormwater storage and treatment capacity required to reduce
untreated overflows below specified values (Corps of Engineers, 1974).
The model can simulate hourly stormwater runoff and quality for a single
catchment for several years. Five water quality constituents are computed
for different land uses: suspended and setteable solids, biochemical
oxygen demand, nitrogen, and phosphorus.
Runoff is computed from hourly precipitation data of a single rain-
gauge. The rainfall excess is defined as the difference between the avail-
able rainfall and losses to depression storage. A constant recovery rate
for depression storage accounts for evapotranspiration. A weighted average
of runoff coefficients for the pervious and impervious areas defines
the fraction of the rainfall excess which becomes surface runoff during
periods of no precipitation. Runoff from snowmelt is computed by the
degree-day method.
Stormwater quality is computed from nonlinear functions considering
the daily rate of dust and dirt accumulation, the percent of each pollutant
contained in the dust and dirt, street sweeping practices, and days between
runoff events. The BOD, N, and PO/, runoff rates depend also on the rate of
runoff of suspended and settleable solids. Land erosion is computed with
the universal soil loss equation.
The model does not route the stormwater runoff and quality in a sewer
or channel network. Computations of treatment, storage and overflow proceed
on an hourly basis by simple runoff volume and pollutant mass balance for
the entire catchment. If the hourly runoff exceeds the treatment capacity,
the excess runoff is put into storage. If the storage capacity is also
exceeded, the excess runoff becomes untreated overflow. If the runoff
is less than the treatment capacity and water is in storage, then the
excess treatment capacity is utilized to diminish the storage volume.
Plug flow is assumed for the routing of pollutants through storage.
The water quality is not modified in storage. Treatment considers only
the hydraulic capacity of the treatment facilities and does not compute
water quality improvement.
The model depends on several empirical coefficients for both storm-
water runoff and quality computations. The runoff coefficients are pro-
vided as input data, the quality coefficients are internal to the program
and may have to be modified for different land uses and catchment charac-
teristics. Verification of the water quality equations has been limited,
and the verification of the complete model is still in progress. Its
accuracy consequently has not been determined.
352
-------
The model appears useful primarily for preliminary planning studies
to estimate approximate magnitudes of untreated stormwater overflows for
various combinations of storage and treatment capacities. The model does
not consider costs, however, and repeated runs with different capacities
are required to determine an optimal combination meeting constraints on
overflows. The model is limited in its application to stormwater drain-
age systems since it does not consider dry-weather flow and quality.
The program is written in Fortran IV for a UNIVAC 1108 and is avail-
able. Program output includes the input precipitation data and summaries
of overflow, storage, and treatment flow volumes and pollutant masses for
each runoff event. Line printer plots indicating the utilization of
storage are also available.
Dorsch Consult Hydrograph-Volume Method
The Dorsch Consult Hydrograph-Volume Method simulates the time-varying
runoff in combined sewerage systems consisting of several catchments and a
sewer and open channel network consisting of loops and converging and
diverging branches (Koniger, 1972; Klym, et al., 1972; Dorsch Consult,
1973). The flow routing is based on the dynamic wave equations and simu-
lates backwater, surcharging and pressure flow but not flow reversal. The
model also simulates retention basins and diversion structures. The model
is limited to the simulation of single runoff events. Water quality, real-
time control and design features are not included.
A diurnal pattern of dry-weather flow is computed representing base
flow, constant infiltration into the sewers from groundwater, and constant
sanitary flow computed from the number of residents and the per capita
contribution. Storm runoff is computed separately for pervious areas and
impervious areas with and without depression storage. One raingauge can
be used for each subcatchment. For areas with depression storage, runoff
does not begin until the smaller depressions are filled. Horton's or
Hoi tan's equation can be used to compute infiltration on pervious areas.
Evapotranspiration, snowmelt, and soil moisture accounting during periods
of no rain are not considered with the exception of depression storage
recovery on pervious areas. Overland and gutter flow routing is accom-
plished by a kinematic wave formulation.
Flow routing in sewers and open channels is accomplished with an
implicit finite difference solution of the dynamic wave equations. Several
cross sections are modeled. The formulation considers drop structures and
retention basins and is coupled with equations for overflow diversion
structures. An iterative solution technique is used which considers back-
water and downstream hydraulic control but neglects flow reversal. Special
formulations are included to handle surcharging and pressure flow.
This is one of the most complete models for the computation of runoff
from urban catchments and the routing of flows in sewer networks. Extreme
detail describing individual lots or at the most city blocks is required,
however, for the accurate computation of runoff from rainfall. The
353
-------
implicit solution of the dynamic wave equations provides an accurate means
of computing flow routing considering both upstream and downstream boundary
conditions and special hydraulic structures. The numerical solution is
very time consuming on the computer, however, and justified only if back-
water and surcharging are significant and the effect of downstream hydraulic
conditions on diversion and retention basin performance have to be con-
sidered.
This is a proprietary model of Dorsch Consult of Munich, Germany,
which operates a North American branch in Toronto, Canada. The computer
program is written in Fortran IV for a UNIVAC 1108, CDC 6600, and IBM 360/50.
Although the model has been applied extensively in Germany, reported model
testing with measured data has been limited to small catchments of less
than 20 ha (50 acres). Comparisons between measured and computed runoff
values have been good.
Program input and output can be either in U.S. Customary or metric units,
Program output includes water levels and discharges for all system ele-
ments, plus storage volume for retentioa basins and cumulative inflow and
outflow for both diversion structures and retention basins.
Environmental Protection Agency Storm Water Management Model
The Storm Water Management Model of the U.S. Environmental Protection
Agency is one of the most comprehensive mathematical models for the simula-
tion of storm and combined sewerage systems. It computes the combined
storm and sanitary runoff from several catchments and routes the flows
through a converging branch sewer network (Metcalf & Eddy, et al., 1971).
It can model two types of flow diversion structures, three storage basins,
and one overflow treatment plant.
Suspended and settleable solids, biochemical oxygen demand, carbona-
ceous oxygen demand, coliform bacteria, nitrogen, phosphorus, and oil and
grease are modeled, and the performance and cost of nine unit treatment
processes can be computed. A receiving water segment of the model com-
putes the flows and water quality impact of sewerage system effluents in
receiving waters. The model does not include realtime control. The model
is limited to the simulation of single runoff events. The model includes
an option to size sewers to eliminate surcharging.
Several raingauge records can be provided as input data but only one
raingauge can be assigned to each subcatchment. Runoff from both pervious
and impervious areas begins when the available depression storage is
filled. Infiltration on pervious areas is computed with Norton's equation
without accounting for changes in soil moisture conditions. Overland
and gutter flow routing is accomplished by a kinematic wave formulation.
Stormwater quality is computed from nonlinear functions considering
different land uses. The formulations consider the pollutant accumulation
between storms, street sweeping practices and the rate of stormwater
354
-------
runoff. Dry-weather flow and quality can be provided as input data or
computed from land use characteristics and account for weekday and
diurnal variations.
The combined flow is routed through the sewers using the kinematic
wave equations. Backwater can be approximated at up to two locations by
specifying a storage element and providing appropriate geometric input
data. If surcharging occurs, the model assumes that the flow in excess
of full pipe flow is stored in the next upstream manhole or the sewer
size can be increased to eliminate surcharging. Flow reversal is not
modeled. The geometries of 12 closed conduit shapes are programmed
internally and three additional aribtrary shapes can be specified by
input data. Two types of diversion structures can be simulated using
approximate equations for their performance. Two inline and one over-
flow storage facility can be modeled considering four types of outfall
structures. Pumping stations with constant pumping rates can also be
modeled.
The combined wastewater quality is routed through the sewers using
pure advection, with scour and deposition for suspended solids and decay
for biochemical oxygen demand. Either plug flow or complete mixing can
be selected for the water quality routing through storage. Quality
reactions and interactions are not simulated.
Nine unit treatment processes can be simulated in any series combina-
tion for a single overflow treatment facility. Functions are built into
the program to compute the annual cost of overflow storage and treatment.
The receiving water portion of the model solves the one-dimensional
dynamic wave equations for nonsteady gradually varied open channel flow
(De Saint-Venant equations) and advective transport of pollutants for
both one-dimensional and two-dimensional arbitrarily shaped grid systems.
The model does not consider chemical and biological reactions and inter-
actions between different water quality constituents.
Segments of the model have been tested on different sets of data
since adequate data of rainfall, runoff and water quality were not avail-
able for a single catchment to test the complete model. Testing of the
runoff and flow computations on catchments ranging in size from 5 to
2200 ha (13 to 5400 acres) showed good agreement with measured values.
The accuracy of the water quality computations, particularly the formula-
tions relating water quality with land use, has not been sufficiently
established to be used with confidence for prediction purposes.
The computer program is very complex and requires a major effort for
its implementation. The input data are arranged in logical groups, but
improvements in the user's manual and documentation of the model's theo-
retical bases is needed to understand the meaning and use of some data.
Program improvements, addition of new model features, and a new user's
manual are in progress at the University of Florida (Heaney, et al.t
1973).
355
-------
The program is written in Fortran IV and versions are available for
the IBM 360 and 370, UNIVAC 1108, and CDC 6400 computers. Program out-
put includes tables and line printer plots of rainfall intensities at
all raingauges, discharges and water quality for each subcatchment and
for up to ten sewer system elements. Summaries of treatment performance
and cost are also printed. Water levels and velocities in the sewerage
system are not computed.
Hydrocomp Simulation Program
The Hydrocomp Simulation Program (Hydrocomp, 1972) is an improved
version of the Stanford Watershed Model (Crawford and Linsley, 1966)
which was the first comprehensive mathematical model of catchment hydrol-
ogy based largely on physical concepts. Recently, a separate program
was developed for the simulation of water quality in river basins
(Hydrocomp, 1973) and interfaced with the hydrologic program. The
program is formulated for the continuous simulation of both water flow
and quality from several catchments and the routing in a converging
branch sewer and open channel network. Although originally developed
for nonurban areas, modifications give the program the capability to
simulate both sewered and nonsewered areas. The water quality model
simulates 17 water quality constituents, including their reactions and
interactions in natural water bodies. The model does not include waste-
water treatment, realtime control, design, and cost computations.
Precipitation for several raingauges is provided as input data to
compute runoff from rainfall or snowmelt. Other meteorologic data are
needed for the computation of evapotranspiration and snow accumulation
and melt. Empirical equations are used to compute interception, infiltra-
tion, percolation to the groundwater, groundwater contribution to surface
flow, and evapotranspiration from interception, depression storage,
groundwater storage and surface waters. A Corps of Engineers method is
used to compute snow accumulation and melt. The formulations require
several empirical coefficients whose values have to be determined by
calibration with measured data.
Both overland and channel flow routing is accomplished using kine-
matic wave formulations. A dispersion term in the kinematic wave equation
approximates backwater effects for the channel routing. The channel
routing is formulated for trapezoidal open channels with trapezoidal
floodplains and for circular closed conduits. Flow control and diversion
structures are not modeled by their hydraulic equations, but diversions
can be specified by providing diversion hydrographs as input data.
Reservoirs can be simulated in the channel network by defining their
geometry and rule curves for their discharge.
Dry-weather flow and quality is provided as input data at constant
time intervals. A power law is programmed as an option to compute dry-
weather quality from dry-weather flow. Stormwater quality is computed
with nonlinear functions of land use and stormwater runoff similar to
356
-------
the formulation of the EPA Storm Water Management Model. Special empirical
functions are built in to compute dirt and dust accumulation between run-
off events, and to account for street sweeping practices, natural decay,
wind effects, and seasonal variations.
The combined wastewater quality is routed in the channels using pure
advection, with a weighted mixing of concentrations between successive
time steps. Water quality reactions and interactions are formulated for
natural water bodies, but their application to sewers is not indicated
in the model documentation. Only one-dimensional flow and quality routing
is formulated. Mixing coefficients can be specified, however, for vertical
water quality interchange between up to nine horizontal layers in impound-
ments.
Although the model was originally formulated for nonurban catchments,
it was modified for application to urban catchments by adding formula-
tions for pervious and impervious areas and circular conduits. The
addition of other sewerage system phenomena and elements is needed,
however, to provide the capability to simulate comprehensive networks.
The hydrology formulation may be more complicated than needed for normal
application to urban catchments considering the lack of available data
for calibration purposes.
A major advantage of the model is the capability for continuous
simulation of both water flow and quality in complex networks since it
considers both catchment moisture and water quality accounting between
runoff events and allows time intervals from a few minutes to bi-monthly.
The hydrologic model has been tested and applied extensively in the .
United States and abroad in both nonurban and urban catchments. Testing
of the water quality model is in progress.
The computer program is written in PL/1 language for IBM 360 and 370
computers. This is a proprietary model of Hydrocomp International, Inc.
of Palo Alto, California. User's manuals are available and user's work-
shops are held periodically.
Program input or output can be in either metric or British units.
Program output includes precipitation; soil moisture status; water
stage, velocities, discharges and water quality concentrations for all
channels and storage; and volume of storage.
Massachusetts Institute of Technology Urban Watershed Model
The Massachusetts Institute of Technology (MIT) Urban Watershed Model
simulates the time-varying runoff of several catchments and a converging
sewer and open channel network (Harley, et al., 1970). The model appears
to be suitable for continuous simulation with the exception of snow
accumulation and melt. Water quality and realtime control features
are not included.
357
-------
The model computes the sizes and costs of sewers, storage and treat-
ment facilities which will result in the least-cost combination of
alternatives for the elimination of untreated overflows and the reduction
of flooding and surcharging (Kirshen, et al., 1972). The original model
was developed at MIT under various projects for the U.S. Office of Water
Resources Research but the model has been modified by Resource Analysis,
Inc., for routine applications (Schaake, et al., 1973). Some of these
modifications have not been fully documented.
Rainfall can be provided as input data for several raingauges or a
single rainstorm can be moved across the catchment by defining its speed
and direction of movement. Only one raingauge can be assigned to each
subcatchment.
Meteorological data for a single station are used to compute evapo-
transpiration with Penman's equation. Four options are available for
the computation of infiltration on pervious areas: Horton's equation,
Hoi tan's equation, a U.S. Soil Conservation Service method, and a runoff
coefficient method. The original MIT version computed infiltration based
on the advection-diffusion equation for nonsteady one-dimensional vertical
water flow in porous media.
A method based on filter theory can be used to compute the infiltra-
tion coefficients from measured rainfall and runoff (Leclerc and Schaake,
1973). Infiltration is computed from overland flow depth. The model
documentation does not describe the consideration of initial losses, such
as the filling of depression storage and differences in rainfall loss and
runoff formulations between pervious and impervious areas.
Flow routing is accomplished with kinematic wave formulations. The
equations are solved by a finite difference scheme for overland flow and
flow in triangular gutters and open channels, trapezoidal open channels,
and circular closed conduits. The formulations do not appear to include
provisions for downstream flow control, backwater, flow reversal, sur-
charging, and pressure flow.
Flow control and diversion structures are not modeled. The original
MIT version included an option to use a linearized solution of the dynamic
wave equations if backwater conditions were considered significant and a
formulation for shock waves.
The design option of the model computes the sizes and costs of circular
sewers, the volume of overflow storage basins, and the hydraulic capacity
of treatment plants needed to reduce undesirable flooding and surcharging
and to eliminate untreated overflows. Linear programming is used to deter-
mine the least-cost combination of these facilities for a design storm
event. The design does not consider water quality.
Model documentation is scattered through several project reports and
a draft user's manual was not released. It is therefore difficult to
evaluate the model since its present status is not fully documented.
358
-------
The model is based on physical principles and includes a minimum of
approximations. Of particular interest are its capability to compute
infiltration coefficients from measured rainfall and runoff data and the
design option which considers the dynamic response of the catchment and
sewerage system runoff to rainfall. Testing on the 9 ha (23 acres)
Gray Haven catchment near Baltimore, Maryland, produced good agreement
between measured and computed runoff values.
The computer program was written for an IBM 360/67 computer. The
design option requires the IBM linear programming package (MPSX 1971).
The user may need to contract with Resource Analysis, Inc. of Cambridge,
Massachusetts, for routine application of the complete and revised pro-
gram package.
Computer output includes tables of rainfall intensities, and overland,
catchment and channel depth and discharge. Flow velocities are not printed.
The output of the design option appears to include the volume and duration
of flooding for each sewer and the costs and sizes of sewers, overflow
storage facilities and treatment plants.
Minneapolis-Saint Paul Urban Runoff Model
The Minneapolis-Saint Paul Urban Runoff Model was developed for real-
time forecasting of flows in the major trunk and interceptor sewers of the
Minneapolis-Saint Paul combined sewer system. Its purpose was to compute
regulator settings which would reduce untreated overflows during rain-
storms to the Mississippi River (Minneapolis-Saint Paul Sanitary District,
1971). The model computes the runoff of several large catchments, diverts
the flows at controllable regulating structures, and routes them through a
converging branch sewerage network to the treatment plant. The model is
specifically designed for realtime control operation on a small computer
and cannot be used easily for the assessment of existing or the design
of new sewerage systems since it uses a highly simplified flow routing
procedure whose coefficients have to be calibrated with measured data or
derived from more sophisticated models.
Although the model was designed for continuous simulation by incor-
porating provisions for the accounting of catchment moisture between rain-
storms, the necessary formulations were never added and the program can
be used only for the simulation of single runoff events. The realtime
control operation of the model is not based on mathematical optimization
techniques but requires repeated trial and error runs with estimates of
regulator settings. The model does not include water quality, design and
cost computations.
Rainfall for each catchment is computed as the weighted average of
rain at several raingauges. Losses are subtracted from rainfall for both
pervious and impervious areas. An exponential decay function computes
losses on impervious areas, and a modified Holtan's equation computes
infiltration on pervious areas, both formulations accounting for changes
359
-------
in catchment moisture conditions. Unit hydrographs derived from catch-
ment characteristics by a method of the U.S. Soil Conservation Service
are used to compute runoff from the rainfall excess. Dry-weather flow
considering diurnal (hourly) variations is provided as input data.
The combined wastewater flow is routed through the sewers with the
progressive average lag method which requires two empirical routing
coefficients for each pipe. The formulation does not consider downstream
flow control, backwater, flow reversal, surcharging, and pressure flow.
The coefficients have to be derived from measurements of flow. Since
these were not available, a solution of the dynamic wave equations for
a single pipe was used to compute them for each pipe of the Minneapolis-
Saint Paul interceptor sewers.
Diversions are computed at six types of regulators by using the
appropriate combinations of weir and orifice or gate discharge equations.
The equations are formulated for controllable regulators, and gate settings
and inflatable weir heights can be read in as functions of time. Other
flow control facilities, such as pumping stations and storage facilities,
are not modeled.
The computer program was written in Fortran IV to be run on a DEC PDP-9
computer with 16K words of core memory at half hour intervals for realtime
predictions. Consequently, several simplifications were necessary in
the mathematical formulations to stay within these constraints. These
included the use of the unit hydrograph method for the rainfall-runoff
computations, the lumping of small catchments into larger catchments,
and the use of the progressive average lag method for the flow routing.
These methods, particularly the routing method, limit the model's accuracy
and its applicability to other areas without extensive calibration.
Of particular interest, however, is the interfacing of the model with
a realtime data acquisition system for rainfall, water level, and regulator
status data, and the remote control of the regulator gates and weirs.
The model is not used for realtime control anymore, however, since the
required trial and error procedure with estimates of regulator settings
is too time-consuming, and operator experience with the system appears to
be more efficient for the control of the regulators.
Seattle Computer Augmented Treatment and Disposal System
The Computer Augmented Treatment and Disposal System (CATAD) of the
Municipality of Metropolitan Seattle (Metro) is an operating system for
realtime control of untreated overflows from the main trunk and inter-
ceptor sewer regulators of the metropolitan Seattle, Washington, combined
sewerage system (Municipality of Metropolitan Seattle, 1971; Gibbs, et al.,
1972; Mallory and Leiser, 1973). The system does not include a compre-
hensive mathematical model for the simulation of runoff from several
catchments and the routing of flows in a sewerage network. A separate
model with these functions is being developed for interfacing with the
realtime control scheme. Present system operation is based on independent
360
-------
computer control of regulators and pumping stations using realtime data
acquisition during rainstorms and rule curve operation. Water quality
is not considered by the scheme.
The system provides automatic computer control of 16 regulators and
two pumping stations. The remote control is determined using realtime
data acquisition of rainfall from six raingauges and water levels upstream
and downstream of the controllable regulators and pumping stations. Rule
curves are used to determine the operation of the regulators and pumping
stations which will maximize the available sewer storage capacity and
minimize untreated overflows.
The rule curves represent sewer storage capacity as functions of
time for the sewer upstream of each controlled regulator and pumping
station. Each curve was developed from typical rainstorm patterns and
represents the desirable filling schedule for a sewer based on its storage
capacity. The rule curves are adjusted during realtime operation to reflect
actual rainfall patterns. Hydraulic equations have been programmed for each
regulator to determine the gate settings which will produce the sewer fill-
ing schedule and flow diversion computed by the rule curves. At present,
each regulator is controlled independently, and the effect of flow routing
on overall system performance is not considered. Manual override is possi-
ble for alarm conditions.
The computer program is written in Fortran for a Xerox Sigma 2 computer
with 49K words of core memory. The realtime data acquisition and remote
control operation (foreground operation) requires 26K while the remaining
13K are available for other purposes. The control scheme uses simulation
at half hour intervals to compute the required regulator and pumping
station operation for the following half hour.
Evaluation of the control scheme has shown significant reductions in
untreated overflows compared to uncontrolled conditions. The scheme's
effectiveness is impaired, however, since each regulator is considered
independently rather than as part of the whole sewerage system, and since
it neglects the effect of the quality of the overflowed wastewaters on
the receiving water quality.
SOGREAH Looped Sewer Model
The Looped Sewer Model of the French consulting firm Societe
Grenobloise d'Etudes et d'Application Hydrauliques (SOGREAH) simulates
the time-varying runoff of combined sewerage systems consisting of
several catchments and a sewer and open channel network including loops
and converging and diverging branches (SOGREAH, 1973 - 4 reports). The
model includes formulations for most hydraulic phenomena encountered in
closed conduit and open channel networks. The flow routing solves the
dynamic wave equations coupled with equations for special sewer system
facilities, such as diversion structures, pumping stations, inverted
siphons, and retention basins. The solution considers both upstream
361
-------
and downstream boundary conditions, backwater, flow reversal, surcharging,
and pressure flow. The model has been expanded recently to include the
advective transport of conservative pollutants. The model does not include
realtime control, design and cost computations.
The formulation for the catchment runoff from rainfall is limited to
the simulation of single runoff events, the flow routing portion of the
model, however, can be used for continuous simulation. This is a pro-
prietary model and not all details of the model formulations were available.
The model does not include a provision to read in rainfall data but
computes design rainfall excess hyetographs of specified frequency for
each catchment using formulas developed by Caquot. The formulas consider
catchment area, slope and percent imperyiousness and require ten empirical
coefficients which is a serious limitation for application of the model
to other than French hydrologic conditions. The design hyetographs are
routed to catchment inlets with the Muskingum flood routing method which
requires the estimation of two additional coefficients. One of the
coefficients is computed by a formula suggested by Schaake and Geyer
which considers catchment length, slope, and percent imperviousness and
uses four additional empirical coefficients. Dry-weather flow is provided
as input data in the form of hydrographs at constant time intervals.
Flow routing in sewers and open channels is accomplished with the
dynamic wave equations. An implicit finite difference solution is coupled
with hydraulic equations for special sewerage system facilities, such as
gates, orifices, weirs, inverted siphons, pumping stations, and retention
basins. The solution considers both upstream and downstream flow conditions
at all facilities and in the sewers, backwater, flow reversal, surcharging,
pressure flow, and flooding at inlets or manholes. The geometry of circu-
lar and egg-shaped pipes and of trapezoidal open channels is programmed
internally and any arbitrary shape can be defined as input data or by the
addition of appropriate subroutines. The flow routing is set up for
continuous simulation.
The model would be needed primarily for the simulation of systems
where the interactions of all modeled hydraulic phenomena are significant
and where loops and diverging branches need to be modeled in addition to
converging branches. A major model weakness is its formulation of the
rainfall-runoff process. Substitution by a different approach based on
more physical principles would be advisable. The flow routing computations,
due to their completeness, require much computer time, some of which can
be saved by specifying different lengths of time steps depending on the
expected rate of change of flow.
The model routing scheme is based on a river basin model developed
by SOGREAH earlier. The firm did not verify the sewer model with urban
hydrologic data since the river basin model verification produced satis-
factory results.
362
-------
The computer program consists of five main programs written in
Fortran IV for an IBM 360/65. This is a proprietary model of SOGREAH
of Grenoble, France. The firm does not have a North American repre-
sentative at present.
Program output includes tables of water depths (or pressures),
velocities and discharges, and plots of water depths and discharges
for each computational point. Program input and output is in metric
units.
University of Cincinnati Urban Runoff Model
The University of Cincinnati Urban Runoff Model simulates the time-
varying runoff of storm sewerage systems consisting of several catchments
and a converging branch sewer and open channel network (University of
Cincinnati, 1970; Papadakis and Preul, 1972). The model does not include
provisions for dry-weather flow, water quality, realtime control, and
design. It is limited to the simulation of single runoff events.
The University of Cincinnati Urban Runoff Model accepts input of only
one hyetograph for the entire drainage basin. Although this could be the
weighted average of several raingauge records, it neglects the areal non-
uniformity of rainfall. Losses are subtracted from rainfall separately
for pervious and impervious areas to account for depression storage.
Infiltration on pervious areas is subtracted from rainfall using Morton's
equation, with its time origin offset to balance cumulative rainfall with
potential infiltration.
Overland flow is computed with a storage routing technique and
Manning's equation. Gutter flow routing uses a steady-state approach
assuming that the gutter outflow equals the inflow during the same time
interval. Flow routing in circular sewers is accomplished by a simple
translation of the upstream hydrograph by an average travel time computed
with Manning's equation.
The model includes several simplifications which appear to reduce
the model's accuracy and applicability. This includes the assumption
of areal uniformity of rainfall, the neglect of dry-weather or base
flows, the computation of infiltration from rainfall rather than over-
land flow depth, the neglect of catchment moisture conditions in the
infiltration computations, the steady-state formulation for gutter flow
routing, and the neglect of dynamic effects in the sewer flow routing.
Although some encouraging results are reported by the model developers
(Papadakis and Preul, 1973) on a 5.2 ha (12.9 acre) and 964 ha (2380 acre)
catchment, the tests were restricted to fairly simple rainfall events and
required considerable calibration of the infiltration rate coefficients.
Testing by others on a 70 ha (173 acre) and 500 ha (1240 acre) catchment
in Australia using more complex storm patterns showed considerable differ-
ences between measured and computed runoff values (Heeps and Mein, 1974).
363
-------
The computer program is written in Fortran IV for an IBM 360 and
available from the University of Cincinnati. A draft user's manual
is also available. Program output includes primarily discharges at
selected points. Water depths and velocities are not printed.
University of Illinois Storm Sewer System Simulation Model
The University of Illinois Storm Sewer System Simulation Model
computes nonsteady flows in a converging sewerage network based on a
solution of the dynamic wave equations (Sevuk, 1973; Sevuk, et al.,
1973). The solution considers upstream and downstream flow controls,
backwater, and flow reversal in circular sewers and in-line storage
at sewer junctions. The model does not compute dry-weather flows from
land use or runoff from precipitation but requires inflow hydrographs
to the sewers as input data. Separate general hydrologic catchment
models have been developed, however (Chow and Kulandaiswaray, 1972;
Chow and Ben-Zvi, 1973), and research for the development of an urban
hydrologic model is in progress for potential interfacing with the flow
routing model. The routing model includes a feature for the sizing of
circular pipes to accommodate peak flows. Costs are not considered.
The model does not include water quality and realtime control features.
A separate model is being developed for sewer design using dynamic
programming with cost and risk considerations.
The flow routing is based on a first-order explicit finite differ-
ence solution of the characteristic equations of the dynamic wave equa-
tions. Surcharging and pressure flow are not computed. Pumping rates,
gate or weir control and diversions can be specified at the sewerage
system outlet, but not at other locations. The solution technique employs
an overlapping segment scheme which allows modeling large networks in
segments. The scheme requires two iterations for its solution, but some
of the resulting sacrifice in computer time can be saved by specifying
different time steps in different segments.
The design option of the routing model is based on hydraulic concepts
only and does not consider costs. An iterative approach is used starting
with the first upstream pipe of each branch, then sizing pipes in each
branch using a kinematic wave routing, and finally improving the sizes
by routing the flows with the full dynamic equations.
The model represents a comprehensive formulation of open channel
flow routing in circular sewers, but neglects special features such as
pressure flow, loops and diverging branches, and diversion structures
to be generally applicable. Long computer times have to be expected
for the explicit solution of the dynamic wave equations. Although the
model has continuous simulation capability, the cost of running it would
normally restrict its application to individual design events. The
increased accuracy in the flow routing may therefore be offset by
uncertainties with respect to defining satisfactory design inflow events.
364
-------
The model has not been verified with real sewerage network data or
applied to real systems. Testing of the routing scheme with experimental
pipe data indicates, however, that high accuracy can be expected.
The computer program is written in PL/1 and assembler language for
an IBM 360/75 system. The program and a user's manual are available
from the University of Illinois. The program is adaptable to IBM 360
and 370 systems but would require major reprogramming for implementation
on other makes of computers. Program output includes line printer plots
of inflow hydrographs, tables and Calcomp plots of depths and discharges,
and tables of velocities at selected points of the network, and sewer
diameters for the design option.
University of Massachusetts Combined Sewer Control Simulation Model
The University of Massachusetts Combined Sewer Control Simulation
Model simulates the time-varying runoff of several catchments and a
single string of circular sewers (Ray, 1972). The model computes runoff
from impervious areas only using hourly rainfall data. This restricts
the model to the simulation of runoff from large catchments with negli-
gible surface runoff contributions from pervious areas. A separate
model is available to compute synthetic hourly rainfall from recorded
rainfall using a Markov chain model.
The flow routing is accomplished with an implicit solution of the
dynamic wave equations which considers upstream and downstream flow
control, backwater, and flow reversal. Special sewerage system facili-
ties, such as diversion and flow control structures and storage facilities
are not modeled. The model is formulated for the continuous simulation
of runoff for periods of up to one month, but neglects catchment moisture
balance between runoff events and snow accumulation and melt. Water
quality, realtime control, and design features are not included.
Dry-weather flow is defined as a constant value for each subcatchment.
Runoff is computed from impervious areas only, and it is assumed that all
rainfall during an hour becomes runoff during the same hour after an
initial loss is satisfied. Synthetic hourly rainfall can be computed
from recorded rainfall data with a separate program. A first order Markov
chain computes hourly rainfall values during wet periods, and a sixth
order Markov chain computes the duration of wet and dry periods.
The combined wastewater flow is routed through a single string of
circular pipes with multiple inlets for the runoff from the modeled sub-
catchments. The flow routing is accomplished by an implicit finite
difference solution of the dynamic wave equations which consider upstream
and downstream flow control, backwater, and flow reversal. Special bound-
ary conditions, such as the simulation of flow control and diversion
structures, are not incorporated.
365
-------
The model has not been tested with real catchment data. Testing
with hypothetical data indicated that the combination of hourly runoff
computation from rainfall with the implicit flow routing scheme may be
a practical combination for the simulation of large catchments and trunk
and interceptor sewers where short duration peak discharges may be negli-
gible. Considerable additions would be needed to the program to make it
generally applicable, such as the computation of runoff from pervious
areas, the consideration of a time lag between rainfall and runoff, the
simulation of evapotranspiration and snow accumulation and melt, and
the expansion of the routing scheme to sewer networks and special
sewerage system facilities.
The computer programs are written for a CDC 3600 computer and are
available. A user's manual has not been written. Computer output con-
sists of rainfall intensities for each catchment, and flow depths and
discharges at selected points of the pipe string.
Water Resources Engineers Storm Mater Management Model
The Storm Water Management Model of Water Resources Engineers, Inc.
(WRE) is a modified version of the Storm Water Management Model of the
Environmental Protection Agency. The model simulates the time-varying
combined storm and dry-weather runoff and wastewater quality of several
catchments and a sewer and open channel network including loops and
converging and diverging branches (Shubinski and Roesner, 1973).
The flow routing is based on the dynamic wave equations and considers
both upstream and downstream control, backwater, and flow reversal. A
special formulation computes surcharging and pressure flow independently
for each junction. The solution is coupled with hydraulic equations
for flow control and diversion structures and pumping stations.
Both dry-weather and stormwater quality are computed for 23 constitu-
ents: suspended and setteable solids, biochemical oxygen demand, nitrogen,
phosphorus, oil and grease, and 17 arbitrary conservative constituents.
The pollutants are routed through the sewerage system, but treatment
processes are not modeled.
The model does not include realtime control, design, and cost computa-
tions. It is limited to the simulation of single runoff events. This is
a proprietary model and not all details of the model formulation and
computer program were available. A separate model for receiving water
flow and quality is available but it is not interfaced with the WRE Storm
Water Management Model (Chen and Orlob, 1972).
Records of several raingauges can be read in and up to three rain-
gauges can be assigned to a single subcatchment. Depression storage,
infiltration, and overland flow are computed separately for pervious and
impervious areas.
366
-------
Runoff begins when all depression storage is filled. A constant
loss rate can be specified for impervious areas. Infiltration on pervious
areas is computed with Norton's equation without accounting for changes in
soil moisture. Overland flow is computed by a kinematic wave formulation.
Special provisions allow the input of data for average urban blocks, the
number of blocks per subcatchment, and the length of each subcatchment to
allow larger spatial discretizations of the modeled catchments without
significant loss in the accuracy of the runoff computations.
Stormwater quality is computed from nonlinear functions for different
land uses. The formulations consider the pollutant accumulation between
storms, street cleaning practices and the rate of stormwater runoff.
Average values of dry-weather flow and quality is provided as input data
and varied by internal functions to account for time and land use
differences.
The combined flow is routed through the sewers using the dynamic
wave equations. A special explicit finite difference formulation solves
the continuity equation at nodes (sewer junctions and connections) and
the momentum equation along links (sewer and channel reaches). The
formulation computes backwater and flow reversal and is coupled with
hydraulic equations for weirs, gates, orifices, and pumping stations.
The geometries of five closed conduit shapes and a trapezoidal open
channel are modeled. Surcharging and pressure flow conditions in the
sewer network are considered independently at each junction.
The combined wastewater quality is routed through the sewers using
pure advection. Dispersion, sedimentation and scour, decay, reactions
and interactions are not simulated.
The model represents a very comprehensive formulation of sewerage
system flow phenomena coupled with water quality computations. The
explicit solution of the dynamic wave equations requires very small time
steps and is time-consuming on the computer. The model would be
needed where the simulation of most modeled hydraulic phenomena is
important and water quality routing is desired. The computation of water
quality from land use characteristics has not been sufficiently verified.
Model testing showed good agreement between measured and computed
runoff values for a 19 ha (47 acre) and 1540 ha (3800 acre) catchment.
Water quality predictions were of the right order of magnitude.
The computer program consists of three main programs and is written
in UNIVAC 1108, IBM 360/65, and CDC 6600 versions. This is a proprietary
model of Water Resources Engineers of Walnut Creek, California. Program
output includes tables and Calcomp plots of depths, velocities, discharges
and water quality (concentrations and mass rates) for the subcatchments,
the sewer system outlet, and selected internal subcatchment and sewerage
system points. Calcomp plots of depths and discharges can also be obtained.
367
-------
Wilsey and Ham Urban Watershed System
The Wilsey and Ham Urban Watershed System computes the time-varying
stormwater runoff of several catchments and a converging branch sewerage
network (Amorocho, 1972). A design option sizes circular pipes for peak
flows. The model does not include provisions for dry-weather flow, water
quality, and realtime control. The model is limited to the simulation
of single runoff events. This is a proprietary model and not all details
of the model formulations are available.
Stormwater runoff is computed separately for pervious and impervious
areas from rainfall intensities of a single raingauge for the entire
modeled catchment. Losses are subtracted from rainfall, but details of
the loss computations are not available. Kinematic wave formulations
are used for the routing of overland flow to the gutters, and for flow
routing in standard shapes of gutters and in circular sewers. A special
feature routes catchbasin inflow exceeding the free flow capacity of the
sewer in the gutter to the next downstream catchbasin. Downstream flow
control, backwater, flow reversal, surcharging, and pressure flow are
not modeled. The model does not simulate flow control and diversion
structures.
A special design option sizes circular sewer pipes to accommodate
the upstream inflow under free flow conditions. Costs are not considered
in the design.
The model includes special provisions which considerably reduce the
input data requirements and computer running time. Data need to be
defined only for typical urban subcatchment elements, rather than all
elements, and the appropriate hydrologic computations are performed only
for these typical elements. The flow routing, however, considers the
actual location of the elements.
Although sufficient details are not available on the mathematical
formulations of the model, it appears to be an efficient model for the
evaluation and design of small storm sewerage systems for which the
limitations of the flow routing scheme are acceptable. The program has
been applied extensively by Wilsey and Ham, but testing with real catch-
ment data to evaluate model accuracy has not been reported.
Details describing the computer program, including estimates of
running times, are not available. The program is written in two versions,
to run on a CDC 6600 computer and on Tymshare's XDS 940 system. This is
a proprietary model of Wilsey and Ham, Inc. of Foster City, California.
Computer output includes discharges at selected points of the storm
sewerage network and pipe diameters for the design option.
368
-------
MODEL TESTING
Most reviewed models were tested and verified by the model developers
and subsequent model users on very limited data as a result of the unavaila-
bility of sufficiently complete and reliable urban catchment, precipita-
tion, runoff, and water quality data. Quite commonly, only portions of a
model were tested on measured data and not the complete model. Sometimes
different portions of the same model were tested with data from different
catchments since comprehensive data were not available from a single
catchment, especially when both flow and quality data were required. Only
in rare instances were two or more models tested on the same data.
Numerical testing of the seven more comprehensive models was therefore
conducted in this study using both hypothetical and real catchment data to
compare model performance. The purpose of the testing was to show that
different models with different mathematical formulations for the same
phenomena will produce differing results given the same input data. The
testing consequently provides useful information to the model users by
indicating ranges of applicability of each model. Testing with the hypo-
thetical data shows the sensitivity of each model to model parameter
variations, while testing with real catchment data also provides infor-
mation on model accuracy.
The reviewed models require widely varying detail with respect to
the spatial discretization of data describing the catchment and sewerage
system. Some models require extremely detailed information such as sizes
and slopes of individual roofs, driveways, lawns, gutters, etc., while
others can lump areas of several hundred acres into single subcatchments.
Great differences exist also between the models with respect to the
required time discretization for precipitation, runoff and water quality
data. Some models require time steps of less than one minute to satisfy
numerical stability conditions, while others can be run with hourly, daily,
and up to semi-monthly data.
The model testing with real catchment data required the collection of
available data on urban catchment and sewer system characteristics, precipi-
tation, runoff, and wastewater quality. A large number of data sources were
investigated and data for several U.S. cities were collected. Selected
data were digitized in the computer formats required by the models being
tested. It became rather difficult, however, to find real catchment data
meeting the requirements of all models. Additional difficulties which were
encountered included missing information on physical characteristics of the
subcatchments and sewerage systems, uncertainties with respect to watershed
infiltration characteristics and moisture conditions, insufficient docu-
mentation with respect to the instrumentation and data analysis techniques
used during the measurement periods, and uncertainties with respect to the
accuracy of the measured rainfall, runoff, and water quality data.
Great reluctance was expressed by model developers, therefore, to run
real catchment data whose accuracy was uncertain and over whose collection
they had no control. The use of hypothetical data removed these uncertainties.
369
-------
Consequently, the model testing concentrated on hypothetical data. All
seven models were tested using the same set of hypothetical data, but
only the Battelle, Chicago Flow Simulation, EPA Storm Water Management
and Dorsch models were also tested on the same set of real catchment
data.
The numerical testing concentrated only on the most common features
of the selected models, that is the rainfall-runoff simulations and the
flow and quality routing. Special features such as snowmelt, the simula-
tion of treatment processes, realtime control and design aspects were
not tested. The hypothetical data testing was applied separately to
the rainfall-runoff formulations and the routing procedures to test the
response of the formulations of each phenomenon to physical parameter
variations. Testing with the real catchment data served the objective
of testing the interactions of the runoff computations for several catch-
ments and the routing of flows and quality through a sewer network.
Hypothetical Data Tests
The hypothetical data runs were not intended to portray accuracy,
but to determine the sensitivity of the models to parameter variations.
The data were also selected to determine limits of applicability of the
models, such as the ability to handle surcharging and backwater conditions.
In addition, the hypothetical data runs provided information with regard
to the cost of running each model.
Two independent sets of hypothetical data were selected. One set
tests the rainfall-runoff relationship for various catchment conditions.
The second set tests the sewer flow routing (and water quality routing,
if applicable) for various combinations of pipe parameters.
Examples of hypothetical data runs are presented only for the Battelle
Urban Wastewater Management Model (BUWMM), EPA Storm Water Management Model
(SWMM) and Chicago Flow Simulation Program (FSP) since results from the
proprietary model runs were not available in time for incorporation in
this report.
Hypothetical Catchment Tests
Hypothetical catchment data tests were performed for two different
catchment sizes, two orientations of each catchment with respect to the
direction of surface flow, ten combinations of physical catchment
characteristics, two different initial catchment moisture conditions,
and four different rainstorms, representing a total of 320 data combina-
tions. Five-minute time steps were selected for the simulation runs.
The effect of different time discretizations on computed runoff was not
tested.
Five examples are described here to illustrate model differences.
They are for a 2-hour triangular storm with a peak rainfall intensity
370
-------
of 50.8 nm/hr (2.0 in/hr) and a 4.65 ha (11.5 acre) rectangular catchment.
The catchment width in the direction of surface flow is 152 m (500 ft)
and the catchment length perpendicular to the direction of surface flow
is 305 m (1000 ft). A Manning's roughness coefficient of 0.25 was assumed
for the pervious area and 0.025 for the impervious area of the catchment.
Initially dry catchment moisture conditions were assumed. For pervious
areas, the maximum infiltration rate is 50.8 mm/hr (2.0 in/hr), the minimum
infiltration rate 12.7 mm/hr (0.5 in/hr), and the infiltration decay rate
0.001 sec"1 as defined by Norton's infiltration equation. The retention
storage capacity for pervious areas is 5.1 mm (0.20 in.) and for impervious
areas 1.3 mm (0.05 in.). The selected examples are for different combina-
tions of catchment slope (0.1 and 10 percent) and impervious areas (0, 50
and 100 percent).
Considerable differences exist between the models in the computed
runoff for the catchments with 0 percent imperviousness as a result of
differences in the mathematical formulations for infiltration (Figures 1
and 4). The differences are particularly great for the 0.1 percent slope
(Figure 1). For both slopes, runoff computed by the BUWMM is consid-
erably higher than for the SWMM. The reason appears to be that Norton's
equation as used in the SWMM neglects catchment moisture conditions in
the computation of potential infiltration. When the ground becomes
rapidly saturated, Morton's equation overestimates infiltration and under-
estimates surface runoff. For steeper slopes where runoff occurs more
rapidly, or low intensity rain, this effect is less pronounced since less
water is available for infiltration.
Hoi tan's equation as used by the BUWMM accounts for changes in
soil moisture, but on the other hand introduces a different approximation
by computing potential infiltration from rainfall rather than the depth
of water on the catchment. It therefore underestimates infiltration for
flat slopes where runoff is sufficiently slow and rainfall occurring
during one t.ime step contributes to infiltration not only during the
same but also succeeding time steps.
The FSP assumes that a constant fraction of the rain falling on the
pervious areas of sewered catchments runs off during the same time step
and neglects catchment slope, shape, and roughness. For the FSP simula-
tions presented here, nonsewered catchment runoff was specified, however,
which computes losses from rainfall with an empirical equation accounting
for soil moisture changes and evaporation and considers catchment shape
and slope. For the flat slope (Figure 1) the FSP computes an extremely
low and slow response, and the runoff hydrograph does not recede during
the 6-hour simulation period.
Differences in the computed results are introduced also by the over-
land flow routing procedures. The BUWMM uses a unit hydrograph
approach, the SWMM a kinematic wave formulation, and the FSP a linear
storage routing which considers catchment shape and slope but neglects
surface roughness.
371
-------
CO
^J
ro
8.
CM
C_) in.
CD
GC
Ole
Q
CD
.in
CM
LU
CO
00
9.00
1 SWMM
2 BUWMM
3 FSP
en
10.190
11.00
12.00 13.00
TIME. HOURS
T1
15.00
8
8
.to
m
o
n>
—I
I—I
CD
CO
"»
8
8
16.00
Figure 1. Hypothetical catchment simulation - 0.1% slope and 0% impervious
-------
CO
-*4
CO
9.1
11.
T
12.00 13.00
TIME, HOURS
14.00
15.1
16.00
Figure 2. Hypothetical catchment simulation - 0.1% slope and 50% impervious
-------
CO
1 SWMM
2 BUWMM
3 FSP
10.00
11.00
12.00 13.00
TIME, HOURS
14.00
15.1
16.00
Figure 3. Hypothetical catchment simulation - 0.1% slope and 100% impervious
-------
U)
"O
en
1 SWMM
2 BUWMM
3 FSP
12.0(9 13.00
TIME, HOURS
16.00
Figure 4. Hypothetical catchment simulation - 10% slope and 0% impervious
-------
CO
i sum
2 BUWMM
3
9.1
10.00
11,190
12.00 13.
TIME, HOURS
T*
15.00
16.00
Figure 5. Hypothetical catchment simulation - 10% slope and 100% impervious
-------
The effect of these differences can be seen by comparing the results
of the runoff computations for the 100 percent impervious areas, since
for these runs rainfall losses were considered negligible and the com-
puted differences are primarily the result of differences in the over-
land flow formulations. As Figures 3 and 5 show, the results of the BUWMM
model and SWMM computations are quite comparable for the steep slope
(Figure 5) and show only small differences for the flat slope (Figure 3).
The FSP assumes that a constant fraction of the overland flow storage on
nonsewered areas runs off during each time step. The resulting runoff
hydrograph is much flatter and longer than the BUWMM and SWMM hydrographs.
Since the FSP formulation neglects catchment shape, slope, and surface
roughness for impervious areas, the computed runoff hydrographs for the
flat and steep slopes are the same.
The combined effects of the formulations for infiltration and overland
flow for pervious and impervious areas is illustrated in Figure 2 for the
0.1 percent slope and a 50 percent imperviousness. The BUWMM produces
the highest runoff, both with respect to peak flow and volume. The SWMM
and FSP appear to produce the same volume of runoff, but the FSP has a
much lower and earlier peak discharge and a longer recession curve.
Effects of these differences for real catchment simulations are shown
in a later section.
Hypothetical Pipe Data Tests
Hypothetical pipe data tests were performed for two different pipe
diameters and three invert slopes. In addition, two types of upstream
and downstream boundary conditions were specified. The first assumes
free inflow into the upstream end of the pipes and free outfall at the
downstream end. The second assumes a storage tank at the upstream end
of the pipe and a diversion structure at the downstream end. Four inflow
hydrographs each with three inflow quality constituents were specified for
the flow and water quality routing. These data represent a total of 48
data combinations.
One example is described here to illustrate model differences. It
is for a 0.61 m (2 ft) diameter circular pipe, 3048 m (10,000 ft) long,
having a slope of 0.05 percent, and a Manning's roughness coefficient of
0.01. Free inflow and outflow is assumed. A 2-hr triangular inflow
hydrograph is used with a peak flow of 0.168 or/sec (5.94 cfs) and a
constant flow of 0.017 m3/sec (0.59 cfs) before and after the triangular
inflow. The peak flow represents 90 percent of full pipe flow as computed
by Manning's equation. Three conservative inflow quality parameters are
specified: a constant concentration of 100 mg/1, a 2-hr triangular shape
with a peak concentration of 100 mg/1 and a constant concentration of
0 mg/1 before and after the triangular shape, and an inverted triangular
shape with a minimum concentration of 0 mg/1 and a constant concentration
of 100 mg/1 before and after the triangular shape.
377
-------
The inflow hydrographs and concentrations and the computed outflow
hydrographs and concentrations are shown in Figures 6 to 9. The routing
was performed with 5-minute time steps. The routed hydrographs of the
BUWMM and SWMM agree quite closely during the rising and falling limb
(Figure 6). The SWMM, however, appears to assume zero initial conditions
although the inflow hydrograph has a finite flow value, and the peak is
considerably more rounded than for the BUWMM. Both models use a kinematic
wave formulation and Manning's equation for the flow routing, but the
BUWMM solves the equations along a forward characteristic which introduces
less numerical dispersion than the finite difference scheme of the SWMM.
The FSP also assumes initially zero flow conditions, and produces a much
flatter routed hydrograph with a much lower and later peak than the other
two models. The model uses a storage routing technique with Manning's
equation, but assumes linear relationships between depth, flow area, and
discharge. The result is a considerable smoothing of the hydrograph.
Water quality computations are shown only for the BUWMM and the SWMM
since the FSP does not include water quality simulation. The constant
inflow concentration is routed without modification by the BUWMM
(Figure 7). The SWMM, on the other hand, assumes zero initial conditions,
and computes a rising limb which reaches the constant concentration value
at the time of the peak of the hydrograph, then drops below the constant
value again, until it finally approaches the constant value 6 hours after
the beginning of the simulation period.
Similarly shaped routed concentrations are computed by the BUWMM and
the SWMM for the triangular inflow concentrations. The BUWMM, however,
computes a higher and earlier peak since its mass routing uses the
kinematic wave celerity rather than water flow velocity to facilitate
interfacing with the realtime control scheme of the model.
The inverted triangular inflow concentration is slightly modified
by the BUWMM quality routing (Figure 9). The SWMM assumes again initially
zero concentrations which results in a rising limb until the routed con-
centration eventually changes into a shape resembling the inflow concen-
trations. As before, the smoothing is greater and the minimum routed
inflow concentration occurs later for the SWMM than for the BUWMM.
It is apparent from the above observations that the BUWMM in general
may produce earlier arrival and less smoothing of the routed concentrations,
It also appears that the SWMM model has to be run for a period representing
the actual flow time through the modeled system prior to the desired
simulation period in order to reduce the effects of its inability to
read in and consider nonzero initial pipe flow and concentration condi-
tions. Additional comments concerning this deficiency are given for the
real catchment data tests.
378
-------
CO
9.00
10.00
11,00 12.00 13.00
TIME, HOURS
14.00
15.00
16.00
Figure 6. Hypothetical pipe simulation ^ outflow hydrographs
-------
CO
00
o
in_
CC
QCe
t—Q
UJin
O
O
O
If).
CM
©
• .
8
9. 00
- INFLOW
1 SWMM
2 BUWMM
-i-
10.00
11.00 12.00 13.00
TIME, HOURS
15.00
16.00
Figure 7. Hypothetical pipe simulation ^ outflow concentrations for
constant inflow concentrations
-------
U)
00
- INFLOW
1 SWMM
2 BUWMM
9.00
10.00
11.00
12.00 13.00
TIME, HOURS
15.00
16.00
Figure 8. Hypothetical pipe simulation «- outflow concentrations for
triangular inflow concentrations
-------
CO
CO
CD®
in.
O
0=
DC©
f—®
CJ
z
O
O
in.
CM
9.1
10.00
11.00 12.00 13.1
TIME, HOURS
- INFLOW
1 SWMM
2 BUWMM
14.00
15.1
16.00
Figure 9. Hypothetical pipe simulation
-------
Real Catchment Data Tests
An attempt was made to select three real catchments with drainage
areas covering three orders of magnitude which have reliable concurrent
rainfall, runoff, and water quality data. Since no small catchment was
found which had also quality data, the Oakdale Avenue catchment in
Chicago, Illinois, with a drainage area of 5.22 ha (12.9 acres) was
selected because it has reliable rainfall and runoff data. An intermediate
size catchment was selected which had rainfall, runoff, and water quality
data, but it was subsequently discovered that the data were unreliable.
It was then too late in the study to find substitute data. The Bloody
Run catchment in Cincinnati, Ohio, with a drainage area of 964 ha (2380
acres) was selected to represent a large catchment. It has rainfall,
runoff, and water quality data. Difficulties were encountered, however,
in ascertaining the reliability of the data, and the data had to be
scaled from plotted hyetographs, hydrographs, and water quality graphs
since only small portions of the measured data were available in tabular
form. Only Oakdale catchment simulations are presented here since the
Bloody Run simulations were not completed in time for incorporation
in this report.
Oakdale Avenue Catchment Simulations
The Oakdale Avenue catchment is located in an urban area about 6 miles
northwest of downtown Chicago, Illinois (Tucker, 1968). The Oakdale
Avenue catchment is 5.22 ha (12.9 acres) in size (approximately 2-1/2
blocks long by 1 block wide) and consists entirely of residential lots
and adjoining streets.
The backbone of the drainage system consists of a 76-cm (30-in.)
diameter reinforced concrete combined sewer that drains east along Oakdale
Avenue for about two blocks. It drains into a 3.20 m x 3.20 m (10.5 ft x
10.5 ft) concrete combined trunk sewer that drains north toward the North
Branch of the Chicago River. The diameters of the lateral sewer pipes
vary from 25 cm (10 in.) to 76 cm (30 in.). The pipe slopes vary from
0.30 to 4.20 percent. The Manning coefficients of these pipes were
assumed to be 0.012.
As shown in Figure 10, the catchment was divided into 13 subcatchments
for the runoff simulations. Each of the 13 subcatchments has its individual
inlet manhole. The subcatchments vary in size from 0.33 ha (0.82 acres)
to 0.65 ha (K60 acres). The imperviousness of the subcatchments varies
from 39.5 to 56.5 percent. Ground slopes vary from 0.37 to 0.90 percent.
Manning's roughness coefficient was assumed to be 0.012 or 0.013 for the
impervious areas and 0.350 for the pervious areas.
The maximum infiltration, the minimum infiltration, and the decay
rate of infiltration were assumed to be 63.5 mm/hr (2.50 in/hr), 11.4
mm/hr (0.45 in/hr), and 0.00115 sec"', respectively. The overall
volumes of retention storage on the pervious and the impervious areas
were assumed to be 5.08 mm (0.20 in.) and 2.03 mm (0.08 in.),
respectively.
383
-------
I—I I"
_l LJ
•
L. .
»
OAKDALEAVE j
I I
LM
i
! j ALLEY I
J I /. , _ ,
)Um / T-IL.T} ^ r——j
RAINGAGE
I U-1551-*
ry.Jl.|TLL_j~[
CO
00
1 i r
I i J
n r~ — ^
J ! j ©
•
)""*
1
I
i
f
3
j
i
2
>
If "
i
1
VI
J
L_
__j_
•
in i m
i
{ _\
LJ~L __
rT
-JL. 1
\J
r
10
I
>
o — i —
1
L-I.
o L
l_
\
1
1
1
1
1
1
J
FLC
o
LECLAIRE
LEVERGNE
FLOW MEASUREMENT
VAULT
3) SUBCATCHMENT
J] INLET MANHOLE
i SEWER PIPE
Z= STREETS AND ALLEYS
0
80 160
FEET
•4-
0 25 50
METERS
Figure 10. Map of Oakdale Avenue catchment showing subcatchments and
sewer system elements
-------
Rainfall and runoff In the Oakdale Avenue catchment have been
periodically measured and recorded since 1957 by the Chicago Department
of Public Works, Bureau of Engineering. Detailed rainfall and runoff
data for the monitored storms have been reported by Tucker (1968).
Runoff has been measured with a Simplex 76-cm (30-in.) type "S" para-
bolic flume housed in a vault located on the corner of Lamon and Oakdale
Avenues. Rainfall measurements have been conducted using a tipping
bucket raingauge which is located about one block north of the drainage
area on top of Falconer Elementary School. No runoff water quality
measurements were conducted.
In the present study, the rainstorms on the following four dates
were selected to test several methods of computing urban runoff quantity:
May 19, 1959; July 2, 1960; July 26, 1960; and August 2, 1963. These
rainstorms were selected since they represent a typical range of possible
combinations of rainfall intensities and durations.
The storm of May 19, 1959, had high intensity rainfall with a short
duration followed by low intensity rainfall with a long duration
(Figures 11 to 14). This storm generated runoff with one good peak.
The storm of July 2, 1960, had lighter long-duration rainfall followed
by heavier short-duration rainfall (Figure 15 to 18). Runoff produced
by this storm had three peaks and the second peak was the largest.
Since the flume was flooded at the flow measuring location during the
last high-intensity rainfall, the recession curve of the last peak was
not recorded. The storm of July 26, 1960, had relatively light long-
duration rainfall which resulted in successive small peaks of runoff
(Figures 19 to 22). The storm of August 2, 1963, had two high inten-
sity, short-duration rainfalls which produced two successive medium
peaks of runoff (Figures 23 to 26).
Figures 11 to 26 show comparisons between the measured runoff and
the runoff computed by the Battelle Urban Wastewater Management Model
(BUWMM), EPA Storm Water Management Model (SWMM), Chicago Flow
Simulation System (FSP), and Dorsch Consult Hydrograph Volume Method
(Dorsch). Each computed hydrograph was plotted in a separate figure
to better show differences with the measured runoff. Plotting of all
computed runoff hydrographs for a given storm in the same figure may
provide better comparisons between models, but preliminary plots showed
that the lines would be too close together to adequately distinguish
between the different model hydrographs. Runoff periods lasting longer
than 2 hours and 20 minutes are plotted on more than one figure with a
20-minute overlap between successive figures.
The measured rainfall and runoff data were reported at 1-minute
time intervals and the same interval was chosen for the SWMM and Dorsch
model simulations. Five-minute time intervals were used for the BUWMM
and FSP due to model limitations. The BUWMM is restricted by a nflnimum
time step of 2 minutes and a maximum of 56 time steps. The 5-minute
time interval was selected to reduce the need for dividing a long
385
-------
runoff period into too many segments for the simulation. The FSP requires
input data and simulation for a minimum of one day and is restricted by
a minimum time step of 5 minutes.
Figures 11 to 26 indicate that the BUWMM, SWMM and Dorsch model can be
expected to produce satisfactory runoff hydrographs in most cases for
small catchments similar to the Oakdale Avenue catchment. The FSP,
however, did not produce satisfactory results which confirms the model
developers' original intent to use the model only -for the continuous
simulation of runoff from large, mostly unsewered catchments using
1-hour time steps. FSP formulations are not adequate to accurately
simulate runoff from small sewered areas requiring short time steps.
The model's assumption of immediate runoff from sewered areas produces
peaks which are too high and occur too early. The assumption that all
rainfall falling on impervious areas and a constant fraction falling on
pervious areas becomes runoff generally produces also a higher volume
of runoff than measured.
The use of 5-minute time steps for the BUWMM resulted in smoothing
of the peaks and valleys compared to the measured values. Exceptions
are the second peaks of the July 2, 1960 (Figure 16b) and August 2, 1963
(Figure 24) storms which show computed peaks and runoff volumes that are
higher than the measured values. This appears to be the result of the
model's lack of formulations for catchment moisture balance between
storms. The model does not compute the recovery of depression storage
and infiltration capacity during dry periods, and consequently computed
insufficient rainfall losses for the second runoff period of successive
storms.
Similar overestimates of the second peaks and runoff volumes are
observed for the SWMM for the August 2, 1963 storm (Figure 23) for the
same reasons. The second peak and runoff volume of the storm of July 2,
1960 (Figure 15b) were underestimated, however. This storm had to be
split into two simulation periods since it was too long for a single
simulation run. The underestimate is caused by insufficient knowledge
of initial catchment moisture conditions at the beginning of the second
simulation period.
The simulations with the BUWMM, SWMM and FSP were performed by
Battelle-Northwest, but the Dorsch model results were furnished by
Dorsch Consult, Dorsch simulated only the major runoff periods of
each storm and split longer runoff periods into separate simulation
periods. This generally produced satisfactory results, with the excep-
tion of overestimating the second peak and runoff volume of the storm
of August 2, 1963 (Tigure 26). This again is caused by uncertainties
with respect to initial moisture conditions at the beginning of the
second runoff period. The model does not include provisions for comput-
ing catchment moisture balance during dry periods, with the exception
of the recovery of depression storage on pervious areas.
386
-------
CO
CD
•vj
- MEASURED RUNOFF
1 SWMM
T
.99 1.33
TIME, HOURS
1.66
1.99
2.33
Figure 11. Oakdale storm of May 19, 1959 - EPA Storm Water Management Model
-------
oo
00
Q
•
8.
od
.OJ
0.00
.33
- MEASURED RUNOFF
E BUWMM
.66
\ \
.99 1.33
TIME, HOURS
1.66
1.99
cn
p-
CO
ro
• •
(3
ID
rn
o
l!
ID
-P
©
2.33
Figure 12. Oakdale storm of May 19, 1959 - Battelle Urban Wastewater Management
Model
-------
u>
00
- MEASURED RUNOFF
3 FSP
I T
.99 1.33
TIME, HOURS
1.66
1.99
2.33
Figure 13. Oakdale storm of May 19, 1959 - Chicago Flow Simulation Program
-------
CO
U3
o
- MEASURED RUNOFF
4 DORSCH
0.00
.99 1.33
TIME, HOURS
1.66
1.99
2.33
Figure 14. Oakdale storm of May 19f 1959 - Dorsch Hydrograph Volume Method
-------
CO
.
CM
LU
CD
CC
CCs
Q
nr
u 11111111 u
LLJ
CO
0.00
- MEASURED RUNOFF
1 SWMM
CD
CO
©
ID
en
ro
©
* •
©
fO
DO
rn
o
ID
—!
i—i
O
.OT
,66
.99 1.33
TIME, HOURS
1.66
1.99
2.33
Figure 15a. Oakdale storm of July 2, 1960, part 1 - EPA Storm Water Management
Model
-------
GO
VO
tM
- MEASURED RUNOFF
1 SWMM
2.00
2.33
2.66
I f
2.99 3.33
TIME, HOURS
3.66
3.99
4.33
Figure 15b. Oakdale storm of July 2r 1960, part 2 - EPA Storm Water Management
Model
-------
CO
VO
CO
s
Q
•
8.
CM
LU
CD
OL
or©
.
10
rV
m
CM
•o
LU
00 e\
.c\iC_>
S> I S
0.120
.33
,66
I
nnm
- MEASURED RUNOFF
2 BUWMM
I \
.99 1.33
TIME, HOURS
CJI
-
00
Q
or
M
rn
o
ID
—i
l—i
O
ID
1.66
1.99
2.33
Figure 16a. Oakdale storm of July 2, 1960, part 1 - Battelle Urban Wastewater
Management Model
-------
CO
- MEASURED RUNOFF
2 BUWMM
2.33
2.66
T T
2.99 3.33
TIME, HOURS
3.66
3.99
4.33
Figure 16b. Oakdale storm of July 2, 1960, part 2 - Battelle Urban Wastewater
Management Model
-------
cn
CM
UJ
CJE>
QC
nr
NUULlll II
CM
.s*
_
UJ
cn
00
66
T
- MEASURED RUNOFF
3 FSP
,99 1.33
TIME, HOURS
1.66
or
ro
1.99
ro
"0
ID
rn
o
ZD
-H
i—i
O
0)
2.33
Figure 17a, Oakdale storm of July 2, 1960, part 1 - Chicago Flow Simulation
Program
-------
00
vo
fl UU ULJL
- MEASURED RUNOFF
3 FSP
2.99 3.33
TIME, HOURS
Figure 17b. Oakdale storm of July 2, 1960, part 2 - Chicago Flow Simulation
Program
-------
CO
CM
£^
CJin.
LjJ
C£>
CC
s
ID
F?
uuuuuuu u
.3
LU
en
B.flfc
33
66
I
I
- MEASURED RUNOFF
4 DORSCH
G &
o>
O>
^
.99 1.33
TIME, HOURS
1.66
ro
ID
rn
o
3>
—i
I—H
O
" •
»•—•
CD
• •
S
1.99
2.33
Figure 18a. Oakdale storm of July 2, I960, part 1 - Dorsch Hydrograph Volume
Method
-------
GO
10
CO
UUUll U ML
- MEASURED RUNOFF
4 DORSCH
1 I
2.99 3.33
TIME, HOURS
Figure 18b. Oakdale storm of July 2, 1960, part 2 - Dorsch Hydrograph Volume
Method
-------
vo
UJ
CD
OC
CC
Q
1O
CO
OJ
.OJ
'CJ
UJ
CO
0.0(9
.33
,66
- MEASURED RUNOFF
1 SWMM
•4 1-
~~\ 1
.99 1.33
TIME, HOURS
T
Ul
®'
•
CO
en
ZE
IE
or
to
ro
s
CO
1.66
1.99
ro
rn
r>
-l
ID
•33
CO
2.33
Figure 19a. Oakdale storm of July 26, I960, part 1 - EPA Storm Water Management
Model
-------
o
o
to,
.in
'lO
CD
QC
CC
^.
CD to
i—i
O
in
«
OJ
T
CO
.(M
UJ
cn
2.00
2.33
B U LJLJ HILJ U BII0 DDUU
- MEASURED RUNOFF
1 SWUM
2.66
2.99 3.33
TIME, HOURS
S Q
Ul
s
•
00
0>
©
ur
3.66
.TO
5
rn
o
i-H
ID
—I
i—t
O
IE
ID
CO
3.99
4.33
Figure 19b. Oakdale storm of July 26, I960, part 2 - EPA Storm Water
Management Model
-------
8
UJ
CD
cc
en
0s.
tO »n
u uuuuu u u u u u u u u uo n r
*CJ
UJ
(n
4.00
4.33
- MEASURED RUNOFF
1 SWMM
4.66
I I
4.99 5.33
TIME, HOURS
e 8
«
00
§
INJ
M
8
8
ID
rn
o
s—i
ID
O)
IE
ID
5.66
5.99
6.33
Figure 19c. Oakdale storm of July 26, I960, part 3 - EPA Storm Water Management
Model
-------
-p>
O
LU
CD
oc
(-J.
CO 10
O
00
(_>
UJ
.66
- MEASURED RUNOFF
2 BUWMM
t I
.99 1.33
TIME, HOURS
I
•
1.66
1.99
03
©
N)
00
ro
s
• *
Q
fO
~0
ID
m
o
s—I
ID
O
,p>-
®»—i
ID
.00
s
2.33
Figure 20a. Oakdale storm of July 26, 196O/ part 1 - Battelle Urban Wastewater
Management Model
-------
s
•
Q.
COQ
U_m
CJ •
LU
CD
QC
CC
0.
COm
i — i
Q
in
© ,s
oo
.CM
.CM
UJ
CO
:S
2.00
2.33
UUUUIIUUULJUIUUU UUU
- MEASURED RUNOFF
2 BUWMM
2.66
I I
2.99 3.33
TIME, HOURS
s
s
Ul
S'
00
3.66
&'
*-»
•
CO
ro
13
ID
rn
o
ID
3.99
4.33
Figure 20b. Oakdale storm of July 26, 1960, part 2 - Battelle Urban Wastewater
Management Model
-------
LU
CD
OC
0=
CO JO
I — I
Q
OIIUUl] UU U 1 U U U DUD D—IT
.CM
.C\J
LU
CO
4.33
- MEASURED RUNOFF
2 BUWMM
4.66
4.99 5.33
TIME, HOURS
5.66
5.99
Q
Ul
©•
*
CO
IE
ZD
ur
NJ
.ro
s
"D
DD
rn
ID
—I
i—i
O
en
IE
ID
.00
8
Q
6.33
Figure 20c. Oakdale storm of July 26f 1960, part 3 - Battelle Urban Wastewater
Management Model
-------
o
in
33
- MEASURED RUNOFF
3 FSP
=8=
,66
I I
.99 1.33
TIME, HOURS
1.66
-9-
1.99
3-
»
00
Q
i—«
•
0>
or
ro
IV)
s-
O3
ro
-1°
©
m
o
ID
-H
i—i
O
OJ
CO
' •
0
2.33
Figure 21a. Oakdale storm of July 26, 1960, part 1 - Chicago Flow Simulation
Program
-------
O
(ft
U-K
r^
LJ
CD
CC
CH
COin'
in
OJ
oo
.CM
.CM
LLJ
cn
© Q
2.00
1ULJI1 UUULJUUULJIIH DI B II
2.33
- MEASURED RUNOFF
3 FSP
UI
©
•
CD
O)
©
ID
2.66
2.99 3.33
TIME, HOURS
©
©
©
DD
rn
o
ID
3.66
3.99
4.33
Figure 21b. Oakdale storm of July 26, I960, part 2 - Chicago Flow Simulation
Program
-------
IIUUULJU U U II D U
©
*
Q.
O
UJ
CD
OC
OZ
in
OJ"
CO
CM
CM
UJ
CO
.20
. 00
I
- MEASURED RUNOFF
3 FSP
8 Q
©
Ul
©'
•
CD
4.66
4.99 5.33
TIME, HOURS
eu-
ro
ro
rn
D
ID
-H
i—i
O
21
.0)'
®»-«
o
5.66
5.99
6.33
Figure 21c. Oakdale storm of July 26, I960, part 3 - Chicago Flow Simulation
Program
-------
.£>
O
CO
co(
o!
LU
CD
DC
CO io-
I—1
Q
in
T
00
.Oi
.CM
*(_>
QJ
CO
2.00
u u nuiu IILJ u 11«in
- MEASURED RUNOFF
4 DORSCH
s
G
G
00
©
2.33
2.66
2.99 3.33
TIME, HOURS
3.66
Q-
v-*
.
O)
fO
' *
5
©
rn
o
ID
—i
t—i
O
3.99
t. 33
Figure 22a, Oakdale storm of July 26, 1960, part 2 - Dorsch Hydrograph Volume
Method (part 1 of storm not simulated by Dorsch)
-------
©
©.
Li- in
UJ
CD
QC
en
Q
in
II1ULJU U U U U II II IMlin0"
00
.CM
4.00
4.33
4.66
- MEASURED RUNOFF
4 DORSCH
^r I
4.99 5.33
TIME, HOURS
Ul
CD
OJ
§
ro
ro
ISJ
* •
8
rn
o
ID
-H
i—i
O
ID
QO
5.66
5.99
6.33
Figure 22b. Oakdale storm of July 26, 1960, part 3 - Dorsch Hydrograph Volume
Method
-------
UJ
CD
OI
cn
03
.CM
0.00
nr
.33
I
.66
- MEASURED RUNOFF
1 SWMM
I I
.99 1.33
TIME, HOURS
01
1.66
1.99
8
' •
8
ro
©
~D
ID
rn
o
2.33
Figure 23. Oakdale storm of August 2, 1963 - EPA Storm Water Management Model
-------
LU
O
O=
cn
0s.
Q
1O
OJ
8 ;L
CO
-CM
I
0.0®
I
.33
T
- MEASURED RUNOFF
2 BUWMM
,66
I \
.99 1.33
TIME, HOURS
8
©
" »
8
8
Ol
'
CO
8
ro
ID
rn
o
•P=~D
1.66
1.99
2.33
Figure 24. Oakdale storm of August 2, 1963 - Battelle Urban Wastewater
Management Model
-------
ro
MEASURED RUNOFF
0.00
66
\ I
.99 1.33
TIME, HOURS
1.66
1.99
2.33
Figure 25. Oakdale storm of August 2, 1963 - Chicago Flow Simulation Program
-------
CO
MEASURED RUNOFF
4 DORSCH
0.013
1 T
.99 1.33
TIME, HOURS
Figure 26. Oakdale storm of August 2, 1963 - Dorsch Hydrograph Volume Method
-------
The numerical model testing indicates that equally satisfactory
results can be obtained for the BUWMM, SWMM and Dorsch models for small
catchments and sewerage systems where backwater and surcharging condi-
tions do not exist or are negligible. Difficulties arise for these
three models as a result of approximations in the computation of catch-
ment moisture conditions and uncertainties with respect to initial
catchment moisture conditions, particularly for long intermittent storm
periods. The Dorsch model would have to be selected among these four
models if backwater, downstream flow control, surcharging and pressure
flow have to be simulated, but water quality does not have to be simu-
lated. The SWMM becomes cumbersome to use and requires unnecessary
computer time if only flow simulations are needed since the water quality
computations cannot be suppressed and have to be performed whether needed
or not.
The BUWMM has considerable flexibility in performing only portions
of the modeled phenomena, and can suppress flow routing, water quality
computations, realtime control optimization, and design optimization as
desired by the user. The BUWMM, however, does not simulate water quality
from land use, unit treatment processes, and receiving water flow and
quality.
The Chicago FSP is suited primarily for the continuous simulation of
large nonsewered catchments, but its general applicability may be limited
due to considerable simplifications for the catchment hydrology and
channel flow routing and due to the need for modifications of model
coefficients internal to the computer program.
CONCLUSIONS
All of the 18 reviewed mathematical models are suitable for the
simulation of storm and combined sewerage systems or for incorporation
in comprehensive simulation models. Considerable differences exist
however, in the types of phenomena that are modeled and in the mathe-
matical formulations for each phenomenon. The model reviews in a
previous section summarize the objectives, advantages and limitations
of each model. For some applications, models are available with con-
siderable simplifications in their mathematical detail to reduce input
data requirements, computer storage requirements, and computer running
time. Some models include unnecessary approximations considering present
state-of-the-art of hydrologic modeling and computer capability. Some
of the simplifications, however, are based on the need for realtime
control of overflows using a small process computer with slow execution
times but the requirement of repeated simulations within fixed time
constraints.
414
-------
Tables 1 to 5 can be used to select a model based on features needed
for simulating specific storm and combined sewerage system conditions.
If only a few selected physical phenomena need to be modeled, then the
simplest model simulating these phenomena with adequately accurate
mathematical formulations should be selected. Generally, input data
requirements and computer running times decrease with decreasing
complexity of the model. Some models include options to suppress por-
tions of the simulation if only selected phenomena are of interest.
Although this feature is not listed in the tables, it should be con-
sidered in the model selection.
ACKNOWLEDGMENTS
The work performed for this study was conducted under contract
No. 68-03-0251 of the U.S. Environmental Protection Agency with valuable
guidance from Messrs. Chi-Yuan Fan (project officer), Richard Field,
and Harry C. Torno. Battelle staff members contributing significantly
include Mr. Larry V. Kimmel, who performed much of the model and data
collection, and Mr. Stacy E. Wise and Ms. Annette S. Myhres, who per-
formed the computer program conversions and data analyses. The
cooperation of model developers and users is greatly appreciated.
Dr. Paul E. Wisner of MacLaren, Ltd., Toronto, Canada, furnished the
data for the Oakdale catchment in the SWMM formats.
REFERENCES
1. Amorocho, J. The WH-1 Urban Watershed System for Hydrologic
Simulation and Sewer Design. Wilsey & Ham, Consulting Engineers,
Foster City, CA, Unpublished Manuscript, 1972.
2. Anderson, J. J. Real-time Computer Control of Urban Runoff.
Journal of the Hydraulics Division, American Society of Civil
Engineers, 96(HY1):153-164, Proc. Paper 7028, January 1970.
3. Brandstetter, A., R. L. Engel, and D. B. Cearlock. A Mathematical
Model for Optimum Design and Control of Metropolitan Wastewater
Management Systems. Water Resources Bulletin, 9(6):1188-1200,
December 1973.
4. Chen, C. L. and V. T, Chow. Formulation of Mathematical Watershed-
Flow Model. Journal of the Engineering Mechanics Division, American
Society of Civil Engineers, 97(EM3):809-828, Proc. Paper 8199,
June 1971.
5. Chen, C. W. and R. P. Shubinski. Computer Simulation of Urban Storm
Water Runoff. Journal of the Hydraulics Division, American Society
of Civil Engineers, 97(HY2):289-301, Proc. Paper 7924, February 1971,
6. Chow, V. T. and V. C. Kulandaiswamy. General Hydrologic System
Model. Journal of the Hydraulics Division, American Society of
Civil Engineers, 97(HY6):791-804, Proc. Paper 8]76, June 1971.
415
-------
7. Chow, V. T. and A. Ben-Zvi. The Illinois Hydrodynamic Watershed
Model III (KHW Model III). University of Illinois at Urbana-
Champaign, Department of Civil Engineering, Hydraulic Engineering
Series No. 26, September 1973.
8. Crawford, N. H. and R. K. Linsley. Digital Simulation in Hydrology:
Stanford Watershed Model IV. Stanford University, Palo Alto, CA,
Department of Civil Engineering, Technical Report No. 39, July 1966.
9. Crawford, N. H. Studies in the Application of Digital Simulation
to Urban Hydrology. Hydrocomp International, Inc., Palo Alto, CA,
for Office of Water Resources Research, September 1971.
10. Dorsch-Consult. Pollution of Receiving Water Bodies From Combined
and Storm Sewer Systems. Dorsch-Consult, Ltd., Toronto, Canada,
June 1973.
11. Gibbs, C. V., S. M. Alexander, and C. P. Leiser. A System for
Regulation of Combined Sewage Flows. Journal of the Sanitary
Engineering Division, American Society of Civil Engineers, 98(SA6):
951-972, Proc. Paper 9402, December 1972.
12. Harley, B. M., F. E. Perkins, and P. S. Eagleson. A Modular
Distribution of Catchment Dynamics. Massachusetts Institute of
Technology, Cambridge, Ralph M. Parsons Laboratory for Water
Resources and Hydrodynamics, Report No. 133, December 1970.
13. Heaney, J. P., W. C. Huber, H. Sheikh, J. R. Doyle, J. E. Darling,
and W. A. Peltz. Storm Water Management Model: Refinements, Testing
and Decision-Making. University of Florida, Gainesville, Department
of Environmental Engineering Sciences, June 1973.
14. Heeps, D. P. and R. G. Mein. Independent Comparison of Three Urban
Runoff Models. Journal of the Hydraulics Division, American Society
of Civil Engineers, 100(HY7):995-1009, Proc. Paper 10685, July 1974.
15. Hydrocomp International, Inc. Simulation of Continuous Discharge
and Stage Hydrographs in the North Branch of the Chicago River.
Palo Alto, CA, undated (1970?).
16. Hydrocomp International, Inc. Hydrocomp Simulation Programming—
Operations Manual. Palo Alto, CA, February 1972.
17. Hydrocomp International, Inc. Hydrocomp Simulation Programming--
Mathematical Model of Water Quality Indices in Rivers and Impound-
ments. Palo Alto, CA, undated (1973?).
18. Kareliotis, S. J. and V. T. Chow. Computer Solution of a Hydrodynamic
Watershed Model (IHW Model II). University of Illinois at Urbana-
Champaign, Department of Civil Engineering, Hydraulic Engineering
Series No. 25, March 1971.
416
-------
19. Keifer, C. J., J. P. Harrison, and T. 0. Hixson. Chicago Hydrograph
Method, Network Analysis of Runoff Computations. City of Chicago
Department of Public Works, Preliminary Report, July 1970.
20. Kirshen, P. H., D. H. Marks, and J. C. Schaake, Jr. Mathematical
Model for Screening Storm Water Control Alternatives. Massachusetts
Institute of Technology, Cambridge, Ralph M. Parsons Laboratory for
Water Resources and Hydrodynamics, Report No. 157, October 1972.
21. Klym, H., W. Koniger, F. Mevius, and G. Vogel. Urban Hydrological
Processes Computer Simulation. Paper presented at the Seminar on
Computer Methods in Hydraulics, Swiss Federal Institute of
Technology, Zurich, February 1972.
22. Koniger, W. Comparison of Measured Runoff and Runoff Computed
by HVM. Dorsch Consult, Ltd., Munich, Germany, and International
Engineering Corp., Inc., San Francisco, CA, Promotional Brochure,
1972.
23. Lager, J. A., R. P. Shubinski, and L. W. Russel. Development of a
Simulation Model for Stormwater Management. Journal of the Water
Pollution Control Federation, 43(12):2424-2435, December 1971.
24. Lanyon, R. F. and J. P. Jackson. Metropolitan Sanitary District
of Greater Chicago Flow Simulation System. Paper Presented at
the American Society of Civil Engineers Hydraulics Conference,
Ithaca, NY, August 1972.
25. Lanyon, R. F. and J. P. Jackson. A Streamflow Model for Metropolitan
Planning and Design. American Society of Civil Engineers, Urban
Water Resources Research Program, Technical Memorandum No. 20,
January 1974.
26. Leclerc, G. and J. C. Schaake, Jr. Methodology for Assessing the
Potential Impact of Urban Development on Urban Runoff and the Relative
Efficiency of Runoff Control Alternatives. Massachusetts Institute
of Technology, Cambridge, Ralph M. Parsons Laboratory for Water
Resources and Hydrodynamics, Report No. 167, March 1973.
27. Linsley, R. K. A Critical Review of Currently Available Hydrologic
Models for Analysis of Urban Stormwater Runoff. Hydrocomp Inter-
national, Inc., Palo Alto, CA, for Office of Water Resources
Research, August 1971.
28. Mai lory, T. W. and C. P. Leiser. Control of Combined Sewer Overflow
Events. Paper Presented at the 1973 International Public Works
Congress and Equipment Show, Denver, CO, September 1973.
29. Metcalf & Eddy, Inc., University of Florida, and Water Resources
Engineers, Inc. Storm Water Management Model. U.S. Environmental
Protection Agency Report 11024 DOC 07/71, 4 Volumes, October 1971.
417
-------
30. Mevius, F. Analysis of Urban Sewer Systems by Hydrograph-Volume
Method. Paper Presented at the National Conference on Urban
Engineering Terrain Problems, Montreal, Canada, May 1973.
31. Minneapolis-Saint Paul Sanitary District. Dispatching System for
Control of Combined Sewer Losses. U.S. Environmental Protection
Agency Report 11020 FAQ 03/71, March 1971.
32. Municipality of Metropolitan Seattle. Maximizing Storage in Combined
Sewer Systems. U.S. Environmental Protection Agency Report 11022
ELK 12/71, December 1971.
33. Normand, D. Etude Experimentale du Ruissellement Urbain. Societe
Hydrotechnique de France, XImes Journees de 1'Hydraulique, Paris,
France, Question II, Rapport 1, 1970.
34. Papadakis, C. and H. C. Preul. University of Cincinnati Urban
Runoff Model. Journal of the Hydraulics Division, American
Society of Civil Engineers, 98(HY10):1789-1804, Proc. Paper
9298, October 1972.
35. Papadakis, C. N. and H. C. Preul. Testing of Methods for Deter-
mination of Urban Runoff. Journal of the Hydraulics Division,
American Society of Civil Engineers, 99(HY9):1319-1335, Proc.
Paper 9987, September 1973.
36. Pew, K. A., R. L. Callery, A. Brandstetter, and J. J. Anderson.
The Design and Operation of a Real-Time Data Acquisition System
and Combined Sewer Controls in the City of Cleveland, Ohio.
Paper Presented at the Annual Conference, Water Pollution Control
Federation, Altanta, GA, October 1972.
37. Preul, H. C. and C. N. Papadakis. University of Cincinnati Urban
Runoff (UCUR) Model—User's Manual. University of Cincinnati, OH,
Department of Civil Engineering, October 1973.
38. Ray, D. L. Simulation of Control Alternatives for Combined Sewer
Overflows. University of Massachusetts, Amherst, Department of
Civil Engineering, Report No. EVE 33-73-4, April 1973.
39. Schaake, J. C., Jr., G. Leclerc, and B. M. Harley. Evaluation and
Control of Urban Runoff. American Society of Civil Engineers
Annual and National Environmental Engineering Meeting, Preprint
2103, New York, NY, October/November 1973.
40. Sevuk, A. S. Unsteady Flow in Sewer Networks. University of
Illinois, Urbana-Champaign, Department of Civil Engineering, Ph.D.
Thesis, 1973.
418
-------
41. Sevuk, A. S., B. G. Yen, and G. E. Peterson. Illinois Storm Sewer
System Simulation Model: User's Manual. University of Illinois,
Urbana-Champaign, Water Resources Center, Research Report No. 73,
October 1973.
42. Sharon, J. D. and J. A. Gutzwiller. Verification and Testing of
the EPA Storm Water Management Model. University of Cincinnati,
Department of Civil Engineering, M.S. Research Report, 1972.
43. Shubinski, R. P. and L. A. Roesner. Linked Process Routing Models.
Paper Presented at American Geophysical Union Annual Spring Meeting,
Washington, DC, April 1973.
44. Smith, G. L., N. S. Grigg, L. S. Tucker, and D. W. Hill. Metro-
politan Water Intelligence System. Colorado State University,
Fort Collins, Department of Civil Engineering, Completion Report--
Phase 1, for Office of Water Resources Research, June 1972.,
45. SOGREAH. Mathematical Model of Flow Simulation in Urban Sewerage
Systems, CAREDAS Program. Societe Grenobloise d1Etudes et
d1Applications Hydrauliques, Grenoble, France, undated (1973?).
46. SOGREAH. Modele Mathematicjue de.Simulation des Ecoulements dans
les Reseaux d'Assainissement Urbains, Programme Caredas. Societe
Grenobloise d1Etudes et d1Applications Hydrauliques, Grenoble,
France, Partial Draft Report, April 1973.
47. SOGREAH. Modele Mathematique de Simulation des Ecoulements dans
les Reseaux d'Assainissement Urbains, Programme Hydrologiques.
Societe Grenobloise d1Etudes et d1Applications Hydrauliques,
Grenoble, France, October 1973.
48. SOGREAH. Mathematical Flow Simulation Model for Urban Sewerage
Systems, Caredas Program. Societe Grenobloise d'Etudes et
d'Applications Hvdrauliques, Grenoble, France, Partial Draft Report,
April 1973, Translated from French by David Vetrano, December 1973.
49. Stall, J. B. and M. L. Terstriep. Storm Sewer Design—An Evaluation
of the RRL Method. U.S. Environmental Protection Agency Report
EPA-R2-72-068, October 1972.
50. Terstriep, M. L. and J. B. Stall. Urban Runoff by Road Research
Laboratory Method. Journal of the Hydraulics Division, American
Society of Civil Engineers, 95(HY6):1809-1834, Proc. Paper 6878,
November 1969. Discussions: 96(HY4):1100-1102, April 1970;
96(HY7):1625-1631, July 1970; 96(HY9):1879-1880, September 1970.
Closure: 97(HY4):574-579, April 1971.
419
-------
51. Tholin, A. L. and C. J. Kelfer. The Hydrology of Urban Runoff.
Journal of the Sanitary Engineering Division, American Society of
Civil Engineers, 85(SA2):47-106, Proc. Paper 1984, March 1959.
Discussions: 85(HY8):119, August 1959; 85(SA5):37-51, September
1959. Closure: 86(SA2):112, March 1960.
52. Tucker, L. S. Oakdale Gaging Installation, Chicago--Instrumental on
and Data. American Society of Civil Engineers, Urban Water
Resources Research Program, Technical Memorandum No. 2, August 1968.
53. University of Cincinnati. Urban Runoff Characteristics. U.S.
Environmental Protection Agency Report 11024 DQU 10/70, October
1970.
54. University of Cincinnati. Urban Runoff Characteristics: Volume I--
Analytical Studies, Volume II—Field Data. U.S. Environmental
Protection Agency Draft Report 11024 DQU 10/72, October 1972.
55. U.S. Corps of Engineers. Urban Runoff: Storage, Treatment and
Overflow Model "STORM." U.S. Army, Davis, CA, Hydrologic Engineer-
ing Center Computer Program 723-S8-L2520, May 1974.
56. Watkins, L. H. The Design of Urban Sewer Systems. Department
of Scientific and Industrial Research, London, England, Road
Research Technical Paper 55, 1962.
57. Wood, E. F., B. M. Harley, and F. E. Perkins. Operational
Characteristics of a Numerical Solution for the Simulation of
Open Channel Flow. Massachusetts Institute of Technology,
Cambridge, Ralph M. Parsons Laboratory for Water Resources and
Hydrodynamics, Report No. 150, June 1972.
58. Yen, B. C. Methodologies for Flow Prediction in Urban Storm Drain-
age Systems. University of Illinois, Urbana-Champaign, Water
Resources Center, Research Report No. 72, September 1973.
59. Yevjevich, V. and A. H. Barnes. Flood Routing through Storm Drains.
Colorado State University, Fort Collins, Hydrology Papers 43, 44,
45, and 46, November 1970.
60. Grigg, N. S., J. W. Labadie, G. L. Smith, D. W, Hill and B. W.
Bradford. Metropolitan Water Intelligence Systems. Colorado State
University, Department of Civil Engineering, Completion Report--
Phase II, for Office of Water Resources Research, June 1973.
420
-------
SHORT COURSE
APPLICATIONS OF STORMWATER MANAGEMENT MODELS
LIST OF PARTICIPANTS
Donald D. Adrian
Dept. of Civil Engineering
Univ. of Mass.
Amherst, MA 01002
Patrick A. Ahern
Dept. of Civil Engineering
University of Toronto
Toronto, Ontario Canada
Mark E. Alpert
MD Environmental Service
Tawes State Office Building
Annapolis, MD 21401
Charles App
EPA Region III
716 Raynham Rd.
Collegeville, PA 19426
K. E. Armstrong
Alberta Environment
305 Southridge
45 Ave & 106 St.
Edmonton, Alberta Canada
David Aronson
US Dept. of Int. Geol.Survey
19 Laurel Place
Bethpage, NY 11714
Laurence E. Benander
Gannett Fleming Corddry Fleming, Inc.
PO Box 1963
Harrisburg, PA 17105
Scott P. Berdine
EPA
1421 Peachtree St. NE
Atlanta, GA 30309
Bernard B. Berger
Director of Water Resources Research
Center
Univ. of Mass.
Amherst, MA 01002
Ken Black
Dept. of Public Works
Greenfield, MA 01301
Dennis D. Blair
City of Phi la. Water Dept.
Room 1270, Municipal Service Bldg.
15th & JFK Blvd.
Philadelphia, PA 19107
Harry Bostian, Staff Engineer
Advanced Waste Treatment Res. Lab
National Environmental Research Center
Cincinnati, Ohio 45268
Albin Brandstetter, Manager
Environmental Resources Planning Sec.
Battelle Pacific Northwest Labs.
Richland, WA
Alvan Bruch
Independent Consultant
113 Harvard St.
Cambridge, MA 02139
Ted S. Buczek
Cleveland Reg. Sewer District
3090 Broadway Ave.
Cleveland, Ohio 44115
Theodore B. Burger
Nassau County Health Dept.
26 Coolidge Ave.
Glen Head, NY 11545
421
-------
Donald Carr
R.V. Anderson Assoc.
Toronto, Ontario Canada
Carlos Carranza
Springfield College
Springfield, MA
Steven C. Chapra
GLERL/NOAA
2300 Washtenow Ave.
Ann Arbor, MI 48104
Max S. Clark
Camp Dresser & McKee, Inc.
1 Center Plaza
Boston, MA 02108
Raul Cuellar
Center for Research in Water Res.
Rt. 4, Box 189
Austin, TX 78757
K. C. Das
VA State Water Control Board
4010 W. Broad St.
Richmond, VA 23230
Francis A. DiGiano
Asst.Prof. of Civil Engineering
Univ. of Mass.
Amherst, MA
J. Obiukwu Duru
Howard Univ., Washington, D.C.
4004 Hanson Oaks Dr.
Landover Hills, MD 20784
H. Lawrence Dyer
Argonne National Laboratory
Argonne, 111. 60439
John B. Erdmann
Mass. Div. of Water Poll. Control
PO Box 545
Westhorough, MA 01581
Chi-Yuan Fan, Staff Engineer
Storm & Combined Sewer Section
Advanced Waste Treat. Res. Lab.
National Environmental Res. Center
Woodbridge Ave.
Edison, NJ 08817
Dr. T. H. Feng
Dept. of Civil Engineering
Univ. of Mass.
Amherst, MA 01002
Richard Field, Staff Engineer
Storm & Combined Sewer Section
Advanced Waste Treatment Res. Lab.
National Environmental Research Center
Woodbridge Ave.
Edison, NJ 08817
Gerald W. Foess
Curran Associates, Inc.
182 Main St.
Northampton, MA 01060
David Gaboury, Research Assistant
Dept. of Civil Engineering, Univ. of Mass,
175 Summer St., Apt. 15
Amherst, MA 01002
James A. Hagarman
Univ- City Science Center
3508 Science Center, Suite 100
Philadelphia, PA 19104
James P. Heaney
Assoc. Prof, of Environmental
Engineering Sciences
University of Florida
Howard C. Hoi tan
234 East 23rd Ave.
Anchorage, Alaska 99503
Wayne C. Huber
Assoc. Prof, of Environmental
Engineering Sciences
University of Florida
422
-------
Thomas K. Jewell, Research Assistant
Dept. of Civil Engineering, Univ. of Mass,
RR2, 35 Harkness Rd
Amherst, MA 01002
Milton B. Kendrick
Polytech, Inc., Consulting Engineer
1836 Euclid Ave.
Cleveland, OH 44115
Alan W. Klimek
210 East Peace St.
Raleigh, NC 27611
Kenneth Klos
Brevard Engineering Co.
5750 Major Blvd.
Orlando, Fla 32805
James Kutzman
EPA
1421 Peachtree St.
Atlanta, GA 30309
E. E. Lindsey
Chemical Engineering
Univ. of Mass.
Amherst, MA 01002
Kuang-Mei Lo
Connell Assoc., Inc.
1320 S. Dixie Highway
Coral Gables, Fla. 33193
David J. Lystrom
US Geological Survey
Water Research Division
PO Box 3202
Portland, OR 97208
Dwight A. MacArthur
O'Brien & Gere Engineers, Inc.
1304 Buckley Rd.
Syracuse, NY 13201
John R. MacLennan
E. A. Hickok & Assoc.
545 Indian Mound
Wayzata, MN 55391
Peter A. Mangarella
Asst. Prof, of Civil Engineering
Univ. of Mass.
Amherst, MA 01002
Mario Marques
Federal Highway Administration
110 Lynbrook Ct.
Greenbelt, MD 20770
Richard N. Marshall
Oceanographic & Environmental Services
Raytheon Co.
P.O. Box 360
Portsmouth, RI
Jerald F. McCain
US Geological Survey
7417 S. Lamar St.
Littleton, Colo.
Robert F. McGhee
EPA, Region III
603 Allen Lane
Media, PA 19063
Murray B. McPherson, Director
ASCE Urban Water Resources Program
Walter A. Mechler
Gr. Vancouver Sewerage & Drainage Dist.
2294 W. 10th Ave.
Vancouver, BC Canada V6K 2H9
Anne N. Miller
EPA Region II
26 Federal Plaza
New York, NY 10007
423
-------
Edward A. Moe
Tighe & Bond, Inc.
268 Howard St.
Ludlow, MA 01056
Michael Moss
Dept. of Natural Resources
Environmental Protection Div.
47 Trinity Ave., SW
Atlanta, GA 30334
Joanne Murphy
The Franklin Institute
20th & Cherry Sts.
Philadelphia, PA 19103
Thomas J. Murphy
Fay, Spofford & Thorndike, Inc.
11 Beacon St.
Boston, MA 02108
Rogert Nemergut
Cahn Engineers
52 Trumbull St.
New Haven, CT
George A. Nossa
US EPA Region II
26 Federal Plaza
New York, NY 10007
Glenn A. Ofcarcik
Pharmer Engineering Corp.
582 Amherst Rd.
S. Hadley, MA 01075
Gerald T. Orlob, Partner
G. T. Orlob § Associates
Orinda, CA
August B. Palmitessa
EPA Region II
26 Federal Plaza, Rm. 837
New York, NY 10007
William Paraskevas
Chem. & Biochem. Engin. Dept.
Rutgers Univ., Engineering College
New Brunswick, NJ 08903
Donald B. Partridge III
Chester Engineers
229 Goldsmith Rd
Pittsburgh, PA 15237
Richard Perna
EPA, Municipal Permits
26 Federal Plaza, Rm. 837
New York, NY 10007
Stephen E. Poole
Nashua River Program
76 Summer St., #117
Fitchbrug, MA 01420
Thomas E. Radigan
O'Brien & Gere Engineers, Inc.
1304 Buckley Rd.
Syracuse, NY 13201
Richard S. Reed
Metcalf & Eddy Engineers
Bo1 ton Rd.
Harvard, MA 01451
J. Steve Reel
Central & Southern Florida Flood Control
District
PO Box V
West Palm Beach, Fla 33402
Larry A. Roesner, Principal Engineer
Water Resources Engineers, Inc.
Walnut Creek, CA
Victor Scottron
Director, Inst. of Water Resources, CT
U-37, Univ. of Conn.
Storrs, CT 06268
424
-------
Kent Scully
Elson T. Killam Assoc., Inc.
48 Essex St.
Mill burn, NJ
Joe Shane
210 East Peace St.
Raleigh, NC
Robert P. Shubinski, Principal Engineer
Water Resources Engineers, Inc.
Springfield, VA
Cheryl Signs
Leonard Rice Con, Water Engin., Inc.
2785 North Speer Blvd
Denver, Colo. 80211
Joseph Skupien
Elson T. Killam Assoc., Inc.
43 Essex St.
Millburn, NJ
Richard P. Slutzah
Polycomp Sys., Inc.
188 Montague St.
Brooklyn, NY 11201
Richard D. Stalker
Area Planning Board of Palm Beach Cty.
PO Box 3643
West Palm Beach, Fla 33402
Tony Tafuri, Staff Engineer
Storm & Combined Sewer Sect.
Advanced Waste Treatment Res. Lab
National Environmental Research Center
Woodbridge Ave.
Edison, NJ 08817
John S. Tapp
EPA
1421 Peachtree St. NE
Atlanta, GA 30309
Harry Torno, Staff Engineer
Munic. Pollution Control Div. (RD67A)
Transport & Trt. Sys. Branch
EPA
Waterside Mall
Washington, DC 2046-
William P. Tully
State University of New York
College of Environm. Science & Forestry
Syracuse, NY
Frank G. Underwood
Wright, Pierce, Barnes & Wyman
25 Vaughan St
Portsmouth, NH
Sandor Vamosi
Proj. Eng. of Assoc. Eng. Services
15007 56 Ave.
Edmonton, Alberta Canada
Guillermo J. Vicens
Resource Analysis, Inc.
1033 Mass. Ave.
Cambridge, MA 02138
Jekabs P. Vittands, Project Manager
Metcalf & Eddy Consulting Engineers
Boston, MA
Vladimir Vlahovich (10CC)
United Engineers & Constructors
1401 Arch St.
Philadelphia, PA 19105
James Walsh
Springfield College
Springfield, MA
Andrew Warren
EPA
26 Federal Plaza
New York, NY 10007
425
-------
Alan Wengell
Cahn Engineers
52 Trumbull St.
New Haven, CT
David Wins low
Espey-Huston & Assoc., Inc.
Austin, TX
Chin-Lien Yen
Howard University, Washington, D.C.
Ill Rosemere Ave.
Silver Spring, MD 20904
Nabil Zaghoul
James F. MacLaren, Inc.
435 McNicoll Ave.
Willowdale, Ontario M2H 2R8
426
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-67Q/2-75-065
3. RECIPIENT'S ACCESSIOI*NO.
4. TITLE AND SUBTITLE
Short Course Proceedings:
APPLICATIONS OF STORMWATER MANAGEMENT MODELS
5. REPORT DATE
June 1975 (Issuing Date)
6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S>
Editors: Francis A. DiGiano and Peter A. Mangarella
8. PERFORMING ORGANIZATION REPORT NO.
9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Civil Engineering
University of Massachusetts
Amherst, Massachusetts 01002
10. PROGRAM ELEMENT NO.
1BB034 (ROAP 21-ATA, Task 026
11. CONTRACT/GRANT NO.
R-803069
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
1974 Short Course Proceedings
14. SPONSORING AGENCY CODE
15. SUPPLEMENTARY NOTES
16. ABSTRACT
This Short Course was sponsored by the U.S. Environmental Protection Agency. The
specific objectives were to encourage the consulting profession to implement storm-
water models in solving the problem of storm and combined sewer overflows and to
make state pollution control agencies aware of this tool in their pollution abatement
efforts. Emphasis was placed on presentations of various types of models, their data
requirements and case studies of their use. The EPA Stormwater Management Model
(SWMM) was highlighted. It is hoped that this compilation of instructional papers,
prepared by the Short Course faculty, will enable practicing engineers to broaden
their use of stormwater management models. The Short Course was held at the
University of Massachusetts August 19-23, 1974. Registration totaled 81 with
representation by consultants; Federal, State and Municipal engineers, including
the Canadian government; and University researchers. This report submitted in
partial fulfillment of Project Number R-803069-01-1 by the Department of Civil
Engineering at the University of Massachusetts, under the sponsorship of the
Environmental Protection Agency.
7.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b.lDENTIFIERS/OPEN ENDED TERMS C. COS AT I Field/Group
Computer programs, *Mathematical models,
*Storm sewers, *Combined sewers,
*Simulation, *Rainfall intensity, *Runoff,
Pollution, *Stream pollution, *Waste
treatment, Cost analysis, Cost effective-
ness, Optimization, *Water quality
infiltration, *Peak
storm flow, Combined
sewer overflows, Urban
runoff, Water quality
control
13B
18. DISTRIBUTION STATEMENT
RELEASE TO PUBLIC
19. SECURITY CLASS (ThisReport}
=IITY CLASS (ThisRt
UNCLASSIFIED
21. NO. OF PAGES
435
20. SECURITY CLASS (Thispage)
UNCLASSIFIED
22. PRICE
EPA Form 2220-1 (9-73)
427
• MI
------- |