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 //?/
-------
 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

-------