EPA/600/2-80/011
States
ental Protection
Municipal Environmental Research EPA-600/2-80-011
Laboratory March 1980
Cincinnati OH 45268
and Development
_ jantity-Quality
Simulation (QQS): A
Detailed Continuous
Planning Model for
Urban Runoff Control
Volume 1. Model
Description, Testing,
and Applications
U.i
>"%&
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U S Environmental
Protection Agency have been grouped into nine series These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum inrerface in related fields
The nine series are
1 Environmental Health Effects Research
2 Environmental Protection Technology
3 Ecological Research
4 Environmental Monitoring
5 Socioeconomic Environmental Studies
6 Scientific and Technical Assessment Reports (STAR)
7 Interagency Energy-Environment Research and Development
8 "Special' Reports
9 Miscellaneous Reports
This report has been assigned to the ENVIRONMENTAL PROTECTION TECH-
NOLOGY series This series describes research performed to develop and dem-
onstrate instrumentation, equipment, and methodology to repair or prevent en-
vironmental degradation from point and non-point sources of pollution. This work
provides the new or improved technology required for the control and treatment
of pollution sources to meet environmental quality standards
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161
-------
EPA-600/2-80-011
March 1980
QUANTITY-QUALITY SIMULATION (QQS):
A DETAILED CONTINUOUS PLANNING MODEL
FOR URBAN RUNOFF CONTROL
Volume I
Model Description, Testing, and Applications
by
Wolfgang F. Geiger
Helmut R. Dorsch
DORSCH CONSULT LTD.
Toronto, Ontario, Canada M5H 1Z2
Grant No. R-805100
3
Project Officers
Richard Field
Douglas Ammon
Storm and Combined Sewer Section
Wastewater Research Division
Municipal Environmental Research Laboratory (Cincinnati)
Edison, New Jersey 08817
MUNICIPAL ENVIRONMENTAL RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OHIO 45268
-------
DISCLAIMER
This report has been reviewed and approved for publication by the Mu-
nicipal Environmental Research Laboratory, U.S. Environmental Protection
Agency, Cincinnati, Ohio. Approval does not signify that the contents nec-
essarily reflect the views and policies of the U.S. Environmental Protection
Agency, nor does mention of trade names or commercial products constitute
endorsement or recommendation for use.
11
-------
FOREWORD
The Environmental Protection Agency was created because of increasing
public and government concern about the dangers of pollution to the health
and welfare of the American people. Noxious air, foul water, and spoiled
land are tragic testimony to the deterioration of our natural environment.
The complexity of that environment and the interplay between its components
require a concentrated and integrated attack on the problem.
Research and development comprise that necessary first step in problem
solution and involve defining the problem, measuring its impact, and search-
ing for solutions. The Municipal Environmental Research Laboratory develops
new and improved technology and systems for the prevention, treatment, and
management of wastewater and solid and hazardous waste pollutant discharges
from municipal and community sources, for the preservation and treatment of
public drinking water supplies and for the minimization of adverse economic,
social, health, and aesthetic effects of pollution. This publication is one
of the products of such research; a most vital communications link between
the researcher and the user community.
This study describes the principles and use of the Quantity-Quality
Simulation (QQS) model. The QQS model defines the impact of urban runoff on
receiving waters and aids in the initial and detailed planning of abatement
alternatives for pollution from storm sewer discharges and combined sewer
overflows in urbanized areas.
Francis T. Mayo
Director
Municipal Environmental
Research Laboratory
ill
-------
PREFACE
Control of urban runoff pollution has become a focus of environmental pro-
tection activities. Satisfactory analysis of urban runoff abatement measures
requires the investigation of numerous alternatives and planning variables
to arrive at economical and efficient solutions. This has resulted in an in-
creasing interest in the modeling of urban runoff.
Urban runoff models have been developed to serve one or more of the ba-
sic functions of planning, system analysis or design, and operations. A num-
ber of models were developed with only one of these functions in mind. Selec-
tion of a suitable model becomes a difficult task when two or more engineer-
ing functions are involved simultaneously. This situation is further compli-
cated by the large variety of problems encountered in practice. A model
serving multiple functions ideally should not produce more information than
needed, yet should not yield such over-simplified results that its reliabi-
lity is in question.
The simulation model detailed in this report was developed to serve
multiple functions. It allows assessment of impacts of existing urban drain-
age systems on receiving waters as well as the evaluation of expected ef-
fects of structural and nonstructural mitigative alternatives. The model has
been applied in the detailed planning and initial design of main system com-
ponents and is considered to be applicable for the analysis of a wide vari-
ety of operations problems.
The theoretical background of this model and its data processing mode
were developed during the period from 1973 to 1975. The programming was par-
tially financed by the Federal Minister of Research and Technology of the
Federal Republic of Germany (Grant No. 0825027).
Because this simulation model may well suit certain American needs
within the framework of Section 201 and Section 208 studies, a research
grant was provided by the U.S. Environmental Protection Agency to make this
planning tool available to the American user. The model had already been
used in an American city.
This report i's in two volumes. Volume I describes the theoretical back-
ground, testing, and application of the model. Volume II is the user's
manual for the model program.
Persons interested in acquiring a copy of the model program and the
User's Manual (Volume II) should contact the EPA Project Officer through
the Wastewater Research Division, Municipal Environmental Research Labora-
tory, Cincinnati, Ohio, 45268.
iv
-------
ABSTRACT
A comprehensive mathematical model, the Quantity-Quality Simulation
(QQS) model, for calculation of urban stormwater and combined sewer overflow
pollution and the means for its control is presented. The model operates in
a continuous mode and accounts for the unsteady runoff and overflow behavior
of total drainage systems. Lumping techniques, that calculate the runoff
from drainage areas, are combined with detailed flow routing through main
and interceptor sewers as well as other structures, such as branches, over-
flows, basins, pump stations, control gates, and treatment facilities. The
computer program calculates the runoff in the storm or combined sewer system
and in the receiving waters. The program package, written in Standard For-
tran IV, comprises approximately 30,000 statements and can be used on any
BATCH processing system having Fortran IV compilers.
While the QQS model is designed to operate in the continuous mode,
single events may be used to calibrate and verify the model. A statistical
analysis routine yields total monthly or annual runoff and overflow figures
and related information such as the frequency and duration of receiving wa-
ter loadings. Continuous simulations may be employed to assess a stormwater
runoff, storm sewer outfall, or combined sewer overflow pollution problem
and to estimate improvements that would be achieved by structural and non-
structural control and corrective measures.
Applicability of the QQS model is demonstrated in a number of compari-
sons with measurements of runoff quantity and several water quality parame-
ters made in several catchments. Short descriptions of QQS model applica-
tions are given for overflow abatement studies made for: Rochester, N.Y.;
Vancouver, B.C., Canada; Toronto, Ont., Canada; Augsburg, Germany; and
Munich, Germany.
This is Volume I, subtitled "Model Description, Testing, and Ap-
plications" of the report, titled "Quantity-Quality Simulation (QQS), A De-
tailed Continuous Planning Model for Urban Runoff Control" submitted by
DORSCH CONSULT LTD., Toronto, Canada, in fulfillment of research Grant No.
R 805100 under the sponsorship of the U.S. Environmental Protection Agency.
Volume II is subtitled "User's Manual". Work was completed as of May 1979.
-------
CONTENTS
Page
Disclaimer ii
Foreword Hi
Preface •- iv
Abstract v
Figures ±x
Tables xii
List of abbreviations and symbols xiii
Acknowledgements xv
Section
1. INTRODUCTION 1
Consideration of urban runoff in planning 1
Continuous simulation concept 2
Unsteady-state simulation concept 3
Consideration of an entire system 4
The QQS concept 4
2. HYDROLOGICAL BACKGROUND 7
Runoff from drainage areas 7
Runoff quantity 7
Runoff quality ' 12
Sewer system routing 15
Flow routing in the sewer system 15
Pollutant routing in the sewer system 22
Receiving water system routing 24
Flow routing in the receiving water system 24
Pollutant routing in the receiving water system 25
Statistical analysis 25
Analysis possibilities 25
Rainfall statistics 26
Discharge and overflow statistics 28
Receiving water statistics 28
3. COMPUTER PROGRAM OVERVIEW 30
Overall computer program description 30
Computer program capacity 33
Computer program handling 35
vii
-------
Section Page
Computer requirements 35
Input data 35
Output data 36
4. MODEL TESTING 37
Hypothetical tests 37
Comparisons with measurements 37
Requirement for measurements 37
Comparisons of different catchments 39
Augsburg test area 39
Stuttgart test area 41
Munich test areas 41
Gothenburg test areas 54
Rochester test areas 64
Vancouver test areas 64
Toronto test areas 69
General results 69
Comparisons with a more detailed simulation model 69
5. APPLICATIONS 73
Augsburg, Germany 73
Munich, Germany 75
Rochester, N.Y 77
Vancouver, B.C., Canada 79
Toronto, Ontario, Canada 82
References 85
viii
-------
FIGURES
Number Page
1 Abstracted view of urban system used for QQS 5
2 Unit hydrograph principle for quantity of storm runoff 8
3 Determination of effective precipitation in the QQS model 10
4 Relation between t,t and d for wash-off calculation 14
5 Different types of system nodes considered in the QQS program 17
6 Difference equations as applied to a sewer element 19
7 Travel paths of pollutant loads in an element 23
8 Relationship between the frequency distribution and the
cumulative frequency or duration curve 27
9 Frequency surface obtained from two-dimensional statistics 27
10 Runstream of the individual QQS programs 31
11 Overall flow chart 32
12 Hypothetical test of flow routing for a 0.61 m (2 ft) diameter
pipe at a 0.05 % slope and with free inflow and outflow 38
13 Comparisons with measurements; Baerenkeller (74 ha), Augsburg,
Germany; storm events of July 17, August 27 and September 25,
1974 42
14 Comparisons with measurements; Buesnau (32 ha), Stuttgart,
Germany; storm events of September 6 and 19, 1967 43
15 Comparisons with measurements; Buesnau (32 ha), Stuttgart,
Germany; storm events of March 15, May 17 and December 3,
1968 44
16 Comparisons with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of September 3, 14, 15 and 17,
1976 46
ix
-------
Number Page
17 Comparison with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of September 28 and 30, and
December 2, 1976 47
18 Comparisons with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of November 3, 1976 and Septem-
ber 9, 1977 48
19 Comparisons with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of June 27 and 29 and July 13,
1977 49
20 Comparisons with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of July 19 and 20, 1977 50
21 Comparisons with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of July 25 and 31, 1977 51
22 Comparisons with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of August 1 and 9, 1977 52
23 Comparisons with measurements; Ingolstaedter Strasse (96 ha),
Munich, Germany; storm events of August 18 and September 4,
1977 53
24 Comparisons with measurements; Josef-Wirth-Weg (34 ha), Munich,
Germany; storm events of September 5, 14 and 17, November 3
and December 2, 1976 55
25 Comparisons with measurements; Josef-Wirth-Weg (34 ha), Munich,
Germany; storm events of September 15 and 28, 1976 and
July 13, 1977 56
26 Comparisons with measurements; Josef-Wirth-Weg (34 ha), Munich,
Germany; storm events of June 29 and August 18, 1977 57
27 Comparisons with measurements; Josef-Wirth-Weg (34 ha), Munich,
Germany; storm events of August 9 and September 1 and 9, 1977 ... 58
28 Comparisons with measurements; Bergsjoen (15 ha), Gothenburg,
29
30
31
Sweden* storm events of May 8 and June 27, 1973
Comparisons with measurements; Bergsjoen (15 ha), Gothenburg,
Sweden- storm events of July 10, 20 and 23, 1973
Comparisons with measurements; Bergsjoen (15 ha), Gothenburg,
Sweden' storm events of August 30 , 1973
Comparisons with measurements; Bergsjoen (15 ha), Gothenburg,
Sweden: storm events of Seotember 20 and October 29. 1973 ....
. . . 59
. . . 60
. . . 61
...62
-------
Number Page
32 Comparisons with measurements; Bergsjoen (15 ha), Gothenburg,
Sweden; storm events of November 11, 1973 63
33 Comparisons with measurements; Torslanda (105 ha), Gothenburg,
Sweden; storm events of September 9, 12 and 27 and November 28,
1976 65
34 Comparisons with measurements; Dewey Avenue (29 ha), Colvin
Street (78 ha), and North Street (34 ha), Rochester, N.Y.,
USA; storm events of August 29 and 30 and September 11, 18
and 20, 1975 66
35 Comparisons with measurements; Site 8 (390 ha), Rochester,
N.Y.; storm events of May 31 and June 19, 1975 67
36 Comparisons with measurements; Site 31 (878 ha), Rochester,
N.Y.; storm events of June 5 and 19, 1975 68
37 Comparisons of QQS catchment runoff simulations with
pertinent HVM simulations 70
38 Comparison of QQS overflow simulations with HVM simulations
for a once per year recurrence synthetic storm 71
39 Comparison of QQS overflow simulations with HVM simulations
for a 40 times per year recurrence synthetic storm 71
40 Main and trunk sewer and receiving water system of Augsburg,
Germany, as defined for QQS simulations 74
41 Total impact of BOD5 on the Lech River, Augsburg, Germany,
as predicated by QQS simulations 76
42 Main and trunk sewer and receiving water system of Munich,
Germany, as prepared for QQS simulations 78
43 Schematic of the main and trunk sewer system of Rochester, N.Y. ... 80
44 Schematic of the main and trunk sewer system of Vancouver,
B.C., Canada 81
45 Combined and separate sewer system of Toronto, Ontario,
Canada, as prepared for QQS simulations 83
XI
-------
TABLES
Number Page
1 Reference Values for Initial or Starting Losses
on Impervious Surfaces 11
2 Examples of Reference Values for Initial or Starting and
Soil Infiltration Losses for Pervious Surfaces 11
3 Reference Values for Maximum Pollution Build-Up 15
4 Tolerance Values for Iteration Control 22
5 Data Processing Capabilities of the Program System 34
6 Comparisons of QQS Simulations with Field Measurements
for Cases with Complete Data 40
xii
-------
LIST OF ABBREVIATIONS AND SYMBOLS
ABBREVIATIONS
BOD 5
cfs
cts
cts/5 min
COD
DEV
DWF
EPA
FC
ft
8
ha
hrs
HVM
kg
kg/5 min
km
1
1/s.ha
Ib
m
m3
m3/s
rag
mg/1
min
mm
mm/5 min
N
P
Pb
QQS
R & D
s
SS
STP
TSS
UWRR
yrs
29.8.75
biochemical oxygen demand (5-day)
cubic feet per second
counts
counts per 5 minutes
chemical oxygen demand
German standard methods for the exami-
nation of water, waste water and sludge
dry-weather flow
Environmental Protection Agency
fecal coliform
feet
gram
hectare
hours
Hydrograph Volume Method
kilogram
kilograms per 5 minutes
kilometer
liter
liter per second and hectare
pound
meter
cubic meter
cubic meters per second
milligram
milligrams per liter
minutes
millimeter
millimeters per 5 minutes
frequency
phosphorus
lead
Quantity-Quality Simulation
Research and Development
second
suspended solids
sewage treatment plant
total supended solids
Urban Water Resources Research
years
day.month.year, e. g. Aug. 29, 1975
xiii
-------
SYMBOLS
A -- flow area
a -- subscript indicating upper end of sewer segment
b -- subscript indicating lower end of sewer segment
cl» C2> C3 ~~ coefficients defining pollutant build-up
function
d -- time of the day
d.. -- differential of ..
3.. -- partial differentiation
DTI -- function of diurnal variation of pollu-
tant wash-off due to land-use
i"| -- fraction of pollutant load parcel
g -- acceleration of gravity
h -- water depth above weir crest
i -- sequence number
H -- flow depth
H — height of profile
I -- precipitation intensity
I, -- surface depression storage
I ,.,. -- effective precipitation
I -- evaporation loss
I . -- soil infiltration loss
C -i
I -- precipitation loss at beginning of rainfall
(I, and I )
T ^ de, we..
I -- total precipitation
I -- wetting loss
1 -- weir length
L — length of an element
p — overflow weir coefficient
n -- index in x-direction
P -- pollutant load
P -- pollutant build-up
P -- pollutant load of drainage area runoff
° and DWF
Q -- flow rate
Q — drainage area runoff including DWF
Q -- uniform flow
Q -- overflow rate
Q — full flow capacity
Rfil -- function for decrease of pollutant wash-off with
increasing rainfall duration
Sf -- friction slope
S -- invert slope
t, T -- time
u, $ -- unit impulse response
v -- flow velocity
x, 4 — coordinates
Y -- invert elevation
z -- number of time steps of 5 minutes
1 -- subscript indicating time 1
2 -- subscript indicating time 2
xiv
-------
ACKNOWLEDGMENTS
The cooperation and assistance of the municipalities and agencies con-
tacted, particularly those providing catchment data and rainfall-runoff
measurements, are gratefully acknowledged.
Especially acknowledged are the valuable guidance and assistance of
Richard Field, Chief of the Storm and Combined Sewer Section (Edison, New
Jersey), Municipal Environmental Research Laboratory, Cincinnati, Ohio.
The valuable comments provided by Harry Torno of the U.S. EPA Office of
Research and Development (Washington, D.C.) and by Douglas Ammon of the
Storm and Combined Sewer Section were gratefully accepted. M.B. McPherson,
Director of the ASCE UWRR Program, who reviewed and assisted in editing
the final report is thankfully acknowledged.
Particularly appreciated are the efforts of Simo Mrdja of Dorsch Con-
sult who proposed the programmable algorithms for the network routing meth-
ods. Appreciation is extended to Donatus Reich of Dorsch Consult who per-
formed most of the comparison tests presented. Computations were performed
at the Dorsch Consult Data Center in Munich, Germany. The principal author,
Wolfgang F. Geiger is now associated with the Technical University of Munich,
Department of Civil Engineering, Munich, Germany.
xv
-------
SECTION 1
INTRODUCTION
CONSIDERATION OF URBAN RUNOFF IN PLANNING
Control of urban runoff pollution has become a new focus for water pol-
lution control agencies. The basis for this control is to establish the im-
pact of stormwater runoff and combined sewer overflows on receiving waters.
To accomplish this objective by observations alone, runoff and overflows
would have to be monitored for a long sequence of events, with water quality
samples taken over short time intervals. Observations of all intermittent
loads from the immense number of storm sewer outfalls and combined sewer
overflows usually found in an urban drainage system would be technically and
financially impractical.
Even if such a data base were available, only existing conditions could
be assessed. More importantly, it would not be practicable to determine the
reasons for a combined sewer overflow occurrence, to select satisfactory
measures for the improvement of storm sewer discharge and combined sewer
overflow abatement, or to assess the impact of projected land use on receiv-
ing waters. Satisfactory analysis of urban runoff abatement measures re-
quires the investigation of numerous alternatives and planning variables to
arrive at economical and efficient solutions. Consideration of this factor
has resulted in an increasing interest in modeling urban runoff, including
its impact on receiving waters. The utility of urban runoff modeling was
debated at length at a national symposium (14).
Urban runoff models have been developed to serve one or more of the
basic engineering functions of planning, system analysis or design, and
operations. A number of models were developed with only one of these func-
tions in mind (1, 3 and 5). Features of major models have been compared by
several authors (1, 6, 8 and 12).
Selection of a suitable model becomes a difficult task when two or more
engineering functions are involved simultaneously. This situation is further
complicated by the large variety of problems encountered in practice. Ideal-
ly, a model serving multiple functions should neither produce more informa-
tion than needed, nor yield such- over-simplified results that its reliabi-
lity is in question.
This report is about the Quantity-Quality Simulation (QQS) model de-
veloped to cover the overlap in objectives and to yield the desired amount
of information. This model not only computes runoff hydrographs at any lo-
cation in a drainage or receiving water system, but also determines frequen-
cies, volumes and durations of storm sewer discharges and combined sewer
-------
overflows. Frequencies and durations of receiving water loadings are major
parameters required for determination of urban drainage impacts on receiv-
ing waters. The ability to assess such impacts for existing systems and to
predict the effects of structural and nonstructural alternatives for future
systems aids in the initial and detailed planning and design of abatement
facilities.
The reasoning behind the QQS concept may be explained as follows:
CONTINUOUS SIMULATION CONCEPT
Because rainfall is a stochastic process; the quantity and quality of
runoff resulting from rainfall also have stochastic properties. In addition
to the stochastic nature of precipitation, mainly two other variable factors
influence the quantity and quality of runoff: (1) surface depression capaci-
ties and sewer network detention capacities, which may be fully or only
partly restored after a prior rainfall; and (2) the amount of removable sur-
face pollutants and sewer sediments, which depend on the magnitudes and pat-
terns of preceding rainfalls and the durations of intervening dry-weather
periods.
Therefore, the resulting storm sewer discharge and combined sewer over-
flow quantities and pollution loads also have random behaviour patterns. Al-
so, because of these many influences and interdependences, the assigned fre-
quency of a given rainfall event and the frequencies of sewer network flows
and discharging and overflowing water quantities and/or overflowing pollu-
tant loads resulting from a given rainfall event may differ from one rain-
fall event to another and with respect to each other.
Therefore, it is not feasible to establish a representative one-event
design case (design storm) or a short series of events which by way of a
limited number of single event calculations would allow judgment on quanti-
ty, frequency and duration of storm water runoff and receiving water pollu-
tion emanating from an urban drainage system. Also, it does not suffice to
examine only those events which would cause overflow activity because: (1)
these events are unidentified a priori; and (2) the rainfalls and dry spells
which occur between overflow events change the initial conditions, which
must be known in order to proceed with sewer network flow simulations. These
conditions include: surface depression capacities, soil infiltration capaci-
ty, amount of removable pollution, retention potential in the sewer network,
etc.
A continuous simulation of all effective rainfall events together with
their associated runoff events is required to meet the main objective, i.e.,
the definition of reliable numerical conditions for storm sewer discharge
and combined sewer overflow simulations. Results from continuous simulations
are prerequisites for the statistical analysis of criteria which allow as-
sessment of the overflow activity of a sewer network, such criteria being:
frequency and duration of overflow quantities and loads of pollution con-
stituents discharged, such as BOD5, total suspended solids, settleable so-
lids, fecal coliform bacteria, etc. Furthermore, peak, mean, and total dis-
-------
charge and overflow amounts may be derived from continuous simulations.
These values may be extracted for any desired span of time, such as for a
total year, the summer months, or an individual month.
Thus, the availability of all discharge and overflow data makes it pos-
sible to define receiving water loading by determining loading frequencies
and loading durations as well as the coincidence of storm sewer discharges
and combined sewer overflows with the corresponding receiving water flows
and background pollution. Moreover, frequencies and durations of phases dur-
ing which regulations would be violated may be scrutinized (e. g., when wa-
ter quality standards are exceeded).
The objectives of a given project dictate the required scope and con-
tent for the statistical evaluation of data generated by continuous simula-
tions. For example, a considerable annual loading might be acceptable for
aquatic life in flowing waters provided the loading enters continuously and
at low concentrations, whereas intermittent loadings at high concentrations
might prove fatal to such life forms. On the other hand, consideration of
intermittent concentrated inputs would be less important than the determina-
tion of total annual load when a standing water body is involved.
As opposed to continuous simulation, single event or design storm sim-
ulations cannot provide loading frequencies or durations or total annual
loadings. Nevertheless, single event simulations applying suitable design
storms are adequate if the objective of a study is confined to the sizing
of sewers. In this case, the design event should be selected in such a way
that it represents patterns and characteristics of local rainfall and takes
into account specific properties, such as total precipitation, peak inten-
sity, time elapsed from start to peak intensity, precipitation volume before
and after the peak, etc. Detailed hydraulic sewer system calculation methods
and corresponding computer programs are available for the hydraulic design
of sewer networks (1, 3 and 10).
UNSTEADY-STATE SIMULATION CONCEPT
Changes in the storm runoff from urban areas and the resulting dis-
charges and overflows are phenomena which exhibit rapid variations, espe-
cially for periods of peak flow. Unsteady-state simulations of the rainfall-
runoff process using half-hour time steps or even longer calculation inter-
vals have been proposed; however, the inherent averaging of such methods
would conceal the peaks and such simulations, therefore, are not suited for
the definition of peak flows and peak loadings. Unsteady-state simulations
applying coarse time steps would tend to underrate total runoff and overflow
volumes and pollution loads. On the other hand, overrating may follow from
using simplified steady-state calculation formulas, which may result in
costly over-dimensioning of sewer facilities with the incorporation of a re-
sultant, but unknown, factor of safety.
Storm sewer discharge and combined sewer overflow behaviour is strongly
related to momentary flow processes and the momentarily available detention
capacity of the sewer network itself. Not only the ambient detention capabi-
-------
lity of the network and its structures must be accounted for in the simula-
tions, but also the causative backwater phenomena. Also, other components of
a network such as sewer branchings and connections, overflow chambers, de-
tention basins, pumping stations, and control gates, must be simulated rea-
listically. All of this can be accomplished in practice only by applying an
unsteady-state simulation method which discretizes the rainfall-runoff pro-
cess into short time steps, generally in the range of five to ten minutes.
Another factor contributing to unsteady and nonuniform runoff behaviour
is the typical movement of rain cells across an area with simultaneous tem-
poral and spatial variations in rainfall intensity. For example, in examin-
ing storms from a 20-year rainfall record for three rainfall stations in
Munich, Germany, it was found that in addition to a local variation of in-
tensities there was an average of 60-minute rainfall migration time from the
westerly station over the central to the easterly station. During the trav-
eling process the peak intensity shifted toward the end of the rainfall.
These results confirm that rainfall and runoff variations with respect to
location and time must be considered when dealing with study areas that en-
compass total cities or metropolitan areas.
CONSIDERATION OF AN ENTIRE SYSTEM
It is essential to study a sewer system in its entirety if mutual in-
terdependences of individual system components and interrelationships be-
tween network, runoff and overflow behaviour are to be taken into account.
Likewise, consideration of the entire system is required in order to select
an overall effective and economical program of improvements. At the same
time, such an overall consideration provides a basis for optimum system op-
eration.
THE QQS CONCEPT
The QQS model was designed to permit continuous simulation of various
rainfall series of several years duration together with discretization of
the simulation into 5-minute calculation intervals. However, in order to do
this at acceptable computer costs it is necessary to refrain from modeling
each sewer length in detail. Therefore, a study area is subdivided into 30
to 100 ha (75 to 250 acres) drainage areas. The division of a study area in-
to larger or smaller drainage areas depends mainly on the study objective,
but it also depends on the system geometry. If the objective is to estimate
the overall impact of runoff from an urban area on receiving waters, a coar-
ser division into larger drainage areas may suffice. A finer division is
necessary, however, for system component design or proof of compliance with
stream standards for individual storm discharges or combined sewer overflows
and also for looped and complicated sewer networks. The individual drainage
areas may be assigned one of four different area-types, e. g., residential,
commercial, industrial, or special.
Quantity of runoff from a drainage area is calculated according to the
Unit Hydrograph method, separately for the impervious and the pervious por-
-------
tions. Pollutant runoff from the drainage area is determined by modifying
the unit hydrograph to include pollution related parameters.
Hydrographs and pollutographs are added to their dry-weather flow equiva-
lents. The results are the quantity and quality inputs to the major sewer
system, which usually consists of the main sewers and their interceptors and
structures. Simulation of flow through the main sewer system is based on a
solution of the kinematic wave equations, whereas the pollution transport is
treated as a plug flow. This procedure accounts for the substantial effects
caused by detention and backwater.
Figure 1 illustrates the concept of linking the calculated lumped run-
off from individual drainage areas to the detailed flow simulation in main
sewers and structures. The extent of division into input drainage areas is,
however, subject to the study objectives and the system geometry as men-
tioned above.
.
iiliillM Mi
PRECIPITATION
DETENTION OF THE
SEWER NETWORK
SURFACE RUNOFF AND DWF
AND ITS POLLUTION RESULTING
FROM SMALL DRAINAGE AREAS
SPECFCKD N E G
COMMERCIAL AREAS
RESIDENTIAL AREAS
INDUSTRIAL AREAS
AREAS WITH SPECIAL
SEWER SYSTEM
COMBINED SEWER
OVERFLOW
RECEIVING
WATER
LOADING
STORM SEWER
OUTFALL
Figure 1. Abstracted view of urban system used for QQS.
Storm sewer discharges and combined sewer overflows and their respective
pollutant loads represent the inputs to the receiving water. They are super-
-------
imposed on the base flow and background pollution of the water body and then
followed through the receiving water system by applying the same principles
used in the major sewer system simulation.
According to experience, it is sufficient to simulate continuously the
rainfall-runoff of a number of carefully selected representative years rath-
er than a complete long series of 20 years or so, in order to determine
storm sewer discharge and combined sewer overflow behaviour of a drainage
system and the subsequent water quality characteristics of a receiving wa-
ter body. The data resulting from a continuous simulation are subjected to
a statistical analysis which produces, apart from totals, annual and monthly
frequencies and durations of flows and pollutant loads and their concentra-
tions, at locations such as detention basins, pumping stations, overflow
structures and selected sections of the receiving water body.
Single event simulations may be carried out in addition to continuous,
long-term simulations. They may cover either the total scope of quantity-
quality simulation or quantity simulation only. Single event simulations are
suited for comparisons between calculated and observed hydrographs and pol-
lutpgraphs and, if desired, also for comparisons with calculations made in a
more detailed way by means of strictly hydraulic models such as the Hydro-
graph Volume Method (HVM), a proprietary model of Dorsch Consult (10). An-
other objective of single event simulations may be to determine the behav-
iour of a trunk sewer system for a design storm, a critical rainfall, or
other single rainfall event. Although such single event simulations do not
produce sufficient data to assess receiving water pollution, they may be
utilized for the design and analysis of alternative improvements, for the
examination of intermediate phases of sewer system construction, and for
studying the effects of land use changes, as far as the major sewer system
and its structures are concerned.
A detailed description of the individual components of the QQS method
is presented in Section 2.
-------
SECTION 2
HYDROLOGICAL BACKGROUND
RUNOFF FROM DRAINAGE AREAS
Runoff Quantity
Quantity of Storm Runoff —
Storm runoff from drainage areas is calculated using the Unit Hydro-
graph method. The validity of this method is maintained because its applica-
tion in the QQS model is confined to small drainage areas where any distor-
tions of the runoff hydrographs, which might be caused by backwater effects
on the runoff surfaces or in the lateral sewers, are usually negligible.
Moreover, utilization of this "black box" method has proved to be suffi-
ciently accurate in small test areas of homogenous land use. If impervious
and pervious area portions are considered separately, a comprehensive,
mathematically linear runoff relation can reflect runoff behaviour from ur-
ban areas as well as from contiguous undeveloped areas.
The computer time and core requirements are substantially less than for
the application of a hydrodynamic mode of runoff calculation because the
transformation relation, i. e., the unit hydrograph, is of a relatively sim-
ple mathematical nature. In reality, this simplification makes it feasible
to apply the QQS model to continuous simulations over extended periods of
time .
A linear system is defined by its transformation relation. If the run-
off portion of the precipitation, i. e. , the effective precipitation,
and the transformation relation, u(t-l), are known, the precipita-
ff
tion-runoff relation may be represented by the convolution integral:
t
Q (t) = J u(t-T) Ie£f(T) dT (1)
o
where Q (t) is the drainage area runoff hydrograph ordinate value at time t
(Figure 2) . The transformation u(t-t) may either be derived from rainfall-
runoff observations or may be developed synthetically.
-------
ID
tu
-t-T
u(t-r)
t
Figure 2. Unit hydrograph principle for
quantity of storm runoff.
In addition to linearity, these two prerequisites must apply:
0 The properties of the drainage area in question do not vary with
respect to time
0 Precipitation is uniformly distributed over the drainage area,
a reasonable assumption for small sized watersheds.
The principle of linearity does not permit accounting for backwater ef-
fects which might be present on the surface and in the lateral pipes of the
drainage area. Still, the adoption of the unit hydrograph as part of the QQS
concept is regarded as a realistic decision, because the drainage areas as
defined are of small size and normally include only minor sewers.
-------
The effective precipitation, I (t) , is determined by subtracting the
losses from the gaged total precipitation, I t
where I = evaporation loss
I = wetting loss
Twe ,
I , = surface depression storage
I . = soil infiltration loss.
si
Apart from the evaporation, wetting, and depression losses which occur
on impervious areas, soil infiltration losses are effective on pervious
areas so that the determination of effective precipitation differs for im-
pervious and pervious areas. Evaporation is accounted for in the QQS model
only during dry spells, whereas it is neglected during a rainfall event be-
cause of its slight influence at that time. Decrease or increase in effec-
tive precipitation because of snow retention or snowmelt, respectively, is
not considered in the present model. Studies of Pfeiff (15) and Brunner (2)
carried out for Ludwigshafen and Munich, Germany, respectively, showed that
for the areas investigated the bulk of storm and overflow events occurred
during the summer months. Where this situation is true for a given study
area, simulation of rainfall-runoff events for the summer months might be
adequate.
Furthermore, it is presumed that wetting and depression losses are ef-
fective at the beginning phase of a rainfall. Therefore, these losses are
combined into an initial or starting abstraction, I .
s t
Soil infiltration is simplified by considering it as a constant loss
which does not vary with respect to time. The reason for this assumption is
computing costs and appears to be justified so that continuous long term
simulations may be performed economically.
For impervious areas equation (2) above is thereby reduced to
and for pervious areas to
Jeff(t) = W^ - Ist(t) -
Figure 3 is a schematic representation of the losses from precipitation
as they are applied in the QQS model.
-------
impervious surface
pervious surface
t
Figure 3.
Determination of effective pre-
cipitation in the QQS model.
The storm runoff calculation from drainage areas not only distinguishes
between impervious and pervious surfaces, but also among four different
land use types (e. g. , residential areas, commercial areas, industrial
areas, or areas with a special drainage system).
Initial or starting loss values for the impervious portions of the dif-
ferent land use types are derived from measured events by establishing vol-
ume balances between observed precipitation and observed runoff. Reference
figures for initial or starting losses are given for various land use types
in Table 1. These figures are based on experience and have been extracted
from a number of precipitation-runoff observation evaluations by the authors
(the number of such evaluations is given in Table 6) and are backed by val-
ues that have appeared in the literature.
10
-------
TABLE 1. REFERENCE VALUES FOR INITIAL OR STARTING
LOSSES ON IMPERVIOUS SURFACES
Initial or
Land Use Starting loss
(mm)
Commercial
Residential
Industrial
0.5 to 2.0
0.7 to 2.5
1.0 to 3.0
Pervious surfaces also have differing initial or starting losses de-
pending as well on the land use. The same is true for the soil infiltration
losses of pervious surfaces which, in addition, are a function of soil type.
Initial or starting loss values for pervious surfaces may be determined by
establishing volume balances, as for impervious surfaces, but in this in-
stance soil infiltration must also be taken into account. Table 2 lists ref-
erence values for both initial or starting losses and soil infiltration val-
ues for various types of pervious surface.
TABLE 2. EXAMPLES OF REFERENCE VALUES FOR INITIAL OR
STARTING AND SOIL INFILTRATION LOSSES FOR PERVIOUS SURFACES
Land use
Soil surface characteristics
(Typical examples)
Initial or
starting
loss
(mm)
Soil infiltration
(1/s ha)
Clay Clayey Sand,
sand loess,
gravel
Open space
Cultivated soil (corn, root crops,
viniculture, hop culture, etc.)
Landscaped strip, playground
Protected green areas and slopes
Garden or meadow
10
8
2
5
5
20
10
5
20
20
30
20
10
30
30
40
30
10
50
40
Precipitation data are the most substantial input to storm runoff cal-
culations. They may be extracted from tabulations of local, regional or na-
tional meteorological services.
11
-------
Initial Conditions--
The filling and emptying of surface depressions and the soil infiltra-
tion that affects pervious surfaces are tracked by balancing each calcula-
tion time step. Because the duration of the dry-weather period preceding the
rainfall in question is known, evaporation and infiltration which occur dur-
ing the dry spell may be determined so that the retention capacities of the
surface depressions present at the start of the rainfall are known. During
periods of rainfall the balance calculation is carried out neglecting eva-
poration losses. If precipitation begins when depressions are still partly
filled, the QQS program applies the degree of filling in order to calculate
the initial or starting losses.
Evaporation rates during dry-weather periods are computed as a function
of temperature and time. The temperature is interpolated according to the
day time from monthly maximum and minimum temperature values incorporated in
the input data. The temperature-evaporation relationship itself also is spe-
cified as associated values by input. This is a simplified way of consider-
ing evaporation but is adequate for long-term simulations.
Quantity of Dry-Weather Flow—
For combined sewer systems, the storm runoff must be added to the dry-
weather flow originating from domestic, commercial, and industrial sources,
and possibly from additional inputs such as ground water infiltration and
construction site water. The domestic dry-weather flow portion is calcu-
lated in accordance with the number of residents and the average per capita
water consumption rate. Defining the per capita consumption rate losses re-
sulting from sprinkling or leaks must be excluded. The calculated domestic
flow is enlarged by an allowance to account for smaller commercial water use
to which possible additional inputs are added. This total value may be addi-
tionally adjusted by a factor, called water consumption factor, for each
drainage area. The industrial sewage of major water users may be accounted
for also seperately for each drainage area.
Because the dry-weather flow rate is subject to considerable variation
during the course of a day, a variable diurnal dry-weather flow pattern may
be used. By applying an estimating factor, it is possible to change all dry-
weather flow input linearly by altering one single value, which can prove
helpful if various water consumption levels are to be quickly examined.
Runoff Quality
Quality of Storm Runoff—
During a rainfall, part or all of the pollution constituents present on
runoff surfaces are washed off. This process is thought to be controlled
mainly by rainfall intensity, rainfall duration and usage of the surfaces,
which varies with time of day (e. g. intensity of street traffic). These in-
fluences are nonlinear. In order to be able to formulate the relations math-
ematically, a new calculation concept was developed based on the Unit Hydro-
graph method. The convolution integral - Equation (1) - for the calculation
of storm runoff quantity was extended by nonlinear functions which represent
the effects of rainfall duration and diurnal variation in the process of
12
-------
pollution wash-off. If the pollutant unit impulse response, <$(t-T), is
known, then the pollution load P(t) washed off is as follows:
t
P(t) = J RDI(t) DTI(d) $(t-t) Ieff(T) dt (5)
o
where d = time of day
DTI(d) = function of diurnal variation of pol-
lutant wash-off due to land use
RDI(t) = function for decrease of pollutant wash-
off with increasing rainfall duration.
DTI(d) and RDI(t) are dimensionless functions. The relations between
the time variables t, t and d are illustrated in Figure 4. The diurnal vari-
ation factor usually varies between 0.7 and 1.0, but may be kept at 1.0 un-
til the modeler develops expertise to vary this parameter. Individual diur-
nal variation functions may be inserted to describe the wash-off of any of
the pollution parameters. The rainfall duration factor starts at 1.0 at the
beginning of a rainfall and is normally decreased to a minimum value of 0.2
within one hour. Individual rainfall duration functions, RDI(t), may be used
for any of the four land use types described above and for any of the pollu-
tion parameters to be described, thus characterizing the wash-off behaviour
of different land use types and pollutants.
The wash-off behaviour discussed above may be adapted for all conserva-
tive pollutants such as total suspended solids, settleable solids or metals.
If it can be assumed that the change of BODs, COD or fecal coliform values
with time can be neglected for the usually short periods of storm events,
these parameters may also be simulated with the method for conservative pol-
lutants .
Balance for Pollution Build-up and Wash-off —
During dry-weather periods, pollutants accumulate on the surfaces. The
effect of the dry-weather period on the magnitude of the pollution P (z)
which is present in the drainage area at the time of the beginning of a
rainfall is approximated by the following pollution build-up function:
puP(z) = ( l -
*
where z = number of time steps (of 5 minutes) and
P (z) = pollution accumulated after z time steps.
The form of the above semi-empirical function is based on the results
of various sampling programs (2 and 11). The magnitudes of the coefficients
cls c2 and c3 are in the range of 10 4 to 10 7 and depend on the pollution
parameter examined. In addition, these coefficients may be supplied sepa-
rately for different land use types. The maximum pollution build-up,
13
-------
Figure 4. Relation between t,T and d for wash-off calculation
P -\nn> varies with respect to the time of the year; and is therefore con-
trolled by an additional, dimensionless function, which may be different for
the individual pollution parameters but whose numerical values vary from 0.5
to 1.0. The maximum pollution values may be entered individually for each
pollution parameter and separately for different land use types.
14
-------
Total pollutant loads washed off during heavy storms that have been ob-
served in sampling programs give an indication of maximum pollution build-
up. Table 3 gives reference values for various pollution parameters based on
present experience. It should be understood that the values vary from case
to case and depend on land use type and other factors.
The mutual influence of pollution build-up and wash-off is mathemati-
cally formulated by a balance calculation. This balance includes the perio-
dic removal of pollutants by street cleaning. The street cleaning influence
is quantified by defining frequency and cleaning efficacy individually for
each land use type. Thus, the amount of pollution present at the beginning
of rainfall in all drainage areas is established.
TABLE 3. REFERENCE VALUES FOR MAXIMUM POLLUTION BUILD-UP
Pollution parameter P ,__
^ up,100
BOD5 15 to 40 kg/ha
KMn04 100 to 200 kg/ha
COD 100 to 200 kg/ha
Settleable solids 60 to 80 kg/ha
TSS 60 to 125 kg/ha
Lead up to approx. 0.03 kg/ha
Total phosphorus up to approx. 0.2 kg/ha
Fecal coliforms 5 x 1012 to 50 x 1012/ha
Quality of Dry-Weather Flow--
Analogous to the computation of combined flows in the case of combined
sewer systems, the pollution load of the storm runoff is superimposed on the
pollution originating from domestic, commercial, and industrial dry-weather
flow. The pollution values are entered as concentration values individually
for the parameters investigated for domestic (including commercial) sewage.
Because here quality may also be subject to diurnal variation, a variable
diurnal concentration pattern may be used for each of the parameters inves-
tigated. For industrial sewage individual constant concentrations may be
specified for each drainage area. This allows for special consideration of
wastewaters from particular industries.
SEWER SYSTEM ROUTING
Flow Routing in the Sewer System
The storm-runoff hydrographs determined for the drainage areas and the
drainage area dry-weather flows are, after superposition, transferred to the
nodes of the sewer system. However, nonlinear input-response behaviour must
be adopted for the major sewer system because of storage and backwater ef-
15
-------
fects and because of the controlling effects from diversions, overflows, and
detention structures. Thus, a hydrodynamic calculation method, based on the
continuity and energy equations, is applied in sewer network flow simula-
tion, as opposed to the hydrological method used for drainage area runoff
determination. It is this hydrodynamic approach that makes it possible to
model diversion structures, detention basins, overflow chambers, pumping
stations and operational control gates according to their actual hydraulic
performance.
The configuration of a sewer system prepared for calculation purposes
consists of a number of pipe elements (sewers) connecting the system nodes.
The logic of the system is based on these nodes. The different node types
considered in the QQS program are shown in Figure 5. The basic node (Figure
5-a) may have one to three inflows from sewers, one to three inflows from
drainage areas, and one or two outflows, thus allowing for simulation of
junctions and diversions. Overflow (Figure 5-b) and basins (Figure 5-c and
d) allow for one to three inflows from sewers and one to three inflows from
drainage areas as well, but for only one outgoing sewer. Gates for opera-
tional control (Figure 5-e) may be located in either one of the outflowing
branches of a diversion or in the outgoing sewer of an overflow. Because al-
most any sequence of sewers and nodes is allowed, the simulation of exist-
ing sewer systems can normally be accommodated.
The hydrodynamic calculations are based upon the two partial differen-
tial equations which describe an unsteady, non-uniform, frictional flow re-
gime. The energy equation is
o _ o 3H . 3 ( v2, ! 9v
bf - a - T (, — ) -i- (/)
3x 3x 2g g at
and the continuity equation is
— = 0 (8)
at
where A = flow area
g = acceleration of gravity
H = flow depth
Q = flow rate
Sf - friction slopes
S = invert slope
t = time
x = coordinate
The acceleration terms — 5— + ^— ( —) are omitted from Equation (7);
g dt dx ^g
however, the variation of the velocity head, ^— ( ^— ) is retained as a re-
16
-------
a. Basic Node
simplest case
IF 1 ->•
->OF1
1 outflow
IPS.
6 inflows 2 outflows
( 3 inflows from sewers.
backwater considered,
3 inflows from drainage areas.
no backwater effects)
e. Operational Control
easel: opening and closing of control gate is
dependent on water surface elevation of
node, where also gate action is induced
case 2. opening and closing of control gate is
dependent on time
! in both cases gate action may be induced by
water surface elevation of any node in the
system-)
located in branching
IF 1
IF 6/
or located behind overflow
IF1
IFi.
CG
IF6
IF6
c. Basin
case 1: detention basin
case 2: treatment basin
( for each case different cleaning effects as
dependent on time may be specified )
IF3
IF6/
d Pump Drained Detention Basin
IF1
Legend^
IF inflow B basin
OF outflow P pump
0V overflow structure CG control gate
inflow/outflow sewers
— --- inflow from drainage areas
Figure 5. Different types of system nodes considered in the QQS program.
17
-------
striction in Equation (12), below, by means of the unequal sign so that in
1 3v
fact only the acceleration component, - —, is neglected. These simplifica-
tions contribute to computer time savings, an important consideration in
long-term simulations.
With these mathematical simplifications and after transformation into a
difference form, Equations (7) and (8) may be expressed as follows (see Fig-
ure 6) :
H, - H = Y - Y. - S, Ax (9)
D a a b f
AH = (SQ - Sf) Ax (9a)
V * V V*V
(10)
Vl * Qb,2 .„ , A.,2 + Ab,2 .
—l z— At + —l J— Ax
Equation (10) is but one of several ways that Equation (8) can be written in
a difference form.
Equations (9) and (10) describe the flow behaviour in a sewer element. This
set of equations has four unknown variables, i. e., the flow depths H and
H, and the flow rates Q and Q, at the upper and lower ends of a flow ele-
ment. Depth and flow vaT.ues which have been determined for the preceding
time step are used as starting figures for the iteration calculation of the
time step in question. The sewer invert elevations Y, and thus the sewer
slope S , are known. The cross-sectional area A is defined by its correla-
tion with H and Q.
Equations (9) and (10) also provide the basis for the calculation of
the friction slope in sewer elements. The friction slope S for a partially
filled sewer element can be approximated by adopting a straight line water
surface slope in the element and by setting the ratio of the mean calculated
flow, 0.5 (Q + Q, ), to the uniform flow, Q , which corresponds to the mean
wetted cross-sectional area of the element:
Qa + Qb 2
S, = ( ) Sn (11)
18
-------
Sf.1 AX
Vb.i2/2g
Va.22'2g
i.22/2g
Ha.2
Figure 6. Difference equations as applied to a sewer element.
A further simplification is introduced by linearizing flow versus depth
relations, with the actual sewer profiles being approximated by rectangular
shapes of equivalent cross-sectional area and height. The friction slope
equation then becomes:
(Ha)
where
v
(H
full flow capacity of sewer element
full height of actual (and equivalent
rectangular) profile.
The full flow capacity of sewers, Q , is determined by means of the Manning
formula or the Prandtl-Colebrook formula. Additional energy losses, such as
those caused by manholes, bends, and roughness resulting from improper con-
struction, age or incrustation, must be accounted for in the friction fac-
tors of those formulas.
Flow under pressure is also described by Equation (Ha). In this case
the cross-sectional area is the full profile area, and the water depth fig-
19
-------
ures H and H are interpreted as pressure heights above the respective in-
verts. As opposed to the analysis of water hammer problems, compressibility
of water and elasticity of pipe walls do not have to be accounted for be-
cause the pressure heads which may occur in sewer systems are of small mag-
nitude .
Additional equations required for the determination of the four un-
knowns (H , H, , Q , and Q, ) are provided by node and boundary conditions.
The node conditions may be applied to as many as three inflowing elements
and as many as two outflowing elements at a basic node, as shown in Figure
5-a.
The energy equation applied to a system node in its simplified form is
as follows:
i=3 v2 . i=2 v2
^ ) Q. . £ I (Y.+H.+ -
2g b)1 i=l a>1 a>:L 2g
v .
I ( Y, . + H. . + -^ ) Q. . £ I (Y.+H.+ -^i ) Q . (12)
i=l b'1 b,i 2g
Production of negative water level differences at nodes is excluded by
the backwater condition:
min ( Y. . + H. . ) 5 max ( Y . + H . ) (13)
. + H. . ) 5
,i T),i
The continuity equation applied to a node is:
i=3 i=3 i=2
I Q . + IQ, .= I Q . (14)
where Q is the known inflow from a drainage area to a node during a given
calculation interval. The inflow value, Q , includes storm runoff and sewage
inflow from the drainage area. If the node is an overflow structure (Figure
5-b) , Q also includes the overflowing quantity Q , which is defined by the
Poleni formula:
2
QQ = f p 1 h0/" V 2g (15)
where p = overflow weir coefficient
1 = weir length
h = water depth above weir crest.
Where a node represents a pump, Q (downstream from the node) is the
maximum pumping rate possible for the prevailing pressure difference. How-
ever, if Q, (upstream of the node) is less than the maximum pumping rate
possible, Q equals Q, . Pumping capacity is entered as a function of pumping
head.
20
-------
Boundary conditions for sewers at the upstream ends of a study area are
either Q . = Q ., where Q . =0, or are (X . hydrographs contributed from
upstream areas. Imiform flow* conditions are 'assumed for the sewer sections
at the downstream boundaries of the study area, with the flow depths calcu-
lated as
H, = Q, . H /Q , where H< H .
T) xb v' xv' ID v
This system of equations takes the essential hydrodynamic features of
sewer network flow into account. Filling and emptying of the sewer network
is taken into account by the continuity equation - Equation (10) -, which is
applied to every element. Acceleration and deceleration of flow are repre-
sented in Equation (9). In the case of deceleration, AH becomes positive
(S < S ) and in the case of acceleration, AH becomes negative (S > S ).
The effects of backwater exerted on neighboring elements are accounted for
by Equation (13). For basins and overflow structures a horizontal invert and
a horizontal water level are adopted, for which Equation (9) reduces to
AH = 0.
The system of equations given above is solved successively for each
calculation interval, At. Because the system of equations has nonlinear ele-
ments, solution is accomplished by means of an iterative method applying
mathematical convergence aids.
The iteration method, applied to every time step, follows a specific
pattern. Unknown flows are determined in a downstream oriented calculation
by inserting the flow depths or pressure heads of the preceding iteration in
the equations. Then an upstream oriented calculation is carried out which
applies the flow values just determined and thus computes the corresponding
flow depths or pressure heads. However, in the case of overflows, only from
the outgoing element OF1 (Figure 5-b) backwater is considered. The down-
stream calculation applies both energy and continuity equations. The up-
stream calculation makes use of the energy equation for flow under pressure,
whereas for free-surface flow conditions only the continuity equation is
used. For the downstream calculation both equations are introduced to the
calculation process as a third degree polynomial which is solved using the
Newton-Raphson method.
The downstream-upstream procedure is repeated until certain programmed
accuracy requirements are fulfilled. At present the computer program con-
tains the tolerance values listed in Table 4, which control continuation or
interruption of the iteration procedure.
A limitation of the number of iterations per time step may be applied
as part of the data input. If the system is not convergent the tolerances
may not be reached prior to termination of the iteration procedure. Such a
condition may originate for instance at branching points, when the transi-
tion to backwater conditions is reached within the iteration procedure of a
time step. Note: The degree of precision of the calculation results depends
not only on the tolerance values and the maximum number of iterations per-
mitted, but also on the degree of skeletonization of the sewer system.
21
-------
TABLE 4. TOLERANCE VALUES FOR ITERATION CONTROL
Water level
Type of element tolerance
(mm)
Sewer elements 25
Sewer diversions 10-2
Overflows 10-2
Pumps 10-4
Pollutant Routing in the Sewer System
The pollutographs determined for the drainage areas by superimposing
storm runoff pollution and dry-weather flow pollution enter the main sewer
system at the system nodes. Sewer system flows and flow velocities deter-
mined by means of the hydrodynamic relations already described provide the
basis for the simulation of pollution transport in the network. The pollu-
tant routing part of the QQS model was conceived as a generally applicable
method which permits calculation of the transport of all conservative pol-
lutants involved.
In an analogy to Equation (14), a continuity equation is applied to balance
the pollutant loads at system nodes (refer to Figure 5):
i=3 i=3 i=2
I P . + I P, . = IP. (16)
•_-! C*1 •_-! Dj1 •_•] 3>1
where P = pollutant load in downstream element (upper end)
P, = pollutant load in upstream element (lower end)
P = pollutant load from drainage areas.
A uniform distribution of pollutants throughout the sewer cross-sec-
tions is assumed. At diversions and overflow structures, the pollutant loads
are divided according to the partition of the flows.
P is known from the drainage area runoff quality calculation. P, is
determined by moving the pollutant loads through the elements, using the
flow velocities at the upper and lower ends of the sewers which have pre-
viously been calculated for every time step as the transporting velocities.
The pollutant load which passes through the lower end of a sewer element or
a basin during a time step is thus computed by Equation (17) (see Figure 7):
Pb,l
22
-------
where r|
(17a)
in which
>n,l + A4n and W),2 ~ W),l
(17b)
time t!
P(n+3),l
P(n+2).l
Pn,l
t -
time t2=t1*At
a.i
P(n+3),2
P(n-t-3),l
p(n+2),2
-p(n+2), I
0-n)
b,i
5 n,l
Figure 7. Travel paths of pollutant loads in an element.
Assuming that the flow velocity variation dv/dt in a sewer element is
negligible, the flow path covered by the load during a calculation interval
is:
23
-------
At (v - v )
A b a
v Ax
A| = ( £ + — ) ( e ^ - 1 ). (18)
Vb ' Va
Settling and scouring of sediments in the sewers of the simulated major
system are not considered because the flow velocities in trunk sewers are
assumed to be sufficiently large to preclude sedimentation. Settling and
scouring occur mainly in lines at the extremities of systems and in lateral
sewers. These effects may be accounted for in the pollution build-up and
wash-off functions applied in calculations of drainage area runoff quality.
Pollution decay that occurs in detention structures is considered by
including decay functions in the appropriate equations. Two different types
of facilities may be accommodated, i. e., detention basins and storm or com-
bined sewage treatment basins.
For either type of facility, individual decay functions may be assigned
for any of the pollution parameters included. The decay function is entered
as a variable function of the settling time, which is determined by the pro-
gram using the flow velocity in the structure.
By means of these calculations, pollutographs are determined for every
node of the sewer system. Taking pollutographs and hydrographs as a basis,
associated concentration variations are computed for the pollution parame-
ters in question.
RECEIVING WATER SYSTEM ROUTING
Flow Routing in the Receiving Water System
In QQS terms, the receiving water system represents a system separate
from the sewer network. In the case of storm runoff or combined sewer over-
flow events, the receiving water system is computationally loaded by the
storm discharges or combined sewer overflow quantities previously determined
by the flow routing calculations for the sewer system. Nodal points of the
receiving water system are the entry points of the stormwater outfalls or
combined sewer overflow outlets, diversions, junctions, weirs, and invert
drops.
The base flow of receiving waters is entered as a table of mean diurnal
discharge values for the period simulated. These flows are superimposed on
the loading inputs from the sewer system. The resulting flows are routed
through the receiving water system in accordance with the same hydrodynamic
principles that are applied in sewer system calculations. In this way, super-
position of individual outfall or overflow quantities produces a loading
configuration for the receiving water system throughout the study area.
24
-------
Pollutant Routing in the Receiving Water System
Pollutant loads emanating from the sewer network are transferred to the
receiving water system at the loading nodes and superimposed on the receiv-
ing water background pollution. This background pollution may be presented
as functions of base flow rate and time of year. Changes in concentration
such as those caused by an oxygen increase or decay and chemical reactions
are not included as considerations in QQS. Because the flow path of receiv-
ing waters within a study area is normally of limited length and because the
flow time is short compared with biochemical reaction times, it is assumed,
within study area limits, that changes of water quality caused by biological
or chemical effects are of negligible magnitude. Therefore, transport of
pollutant loads through the receiving water system is calculated according
to the same principles used in the pollutant routing calculation of the sew-
er system.
Superposition of individual pollutant loads from the sewer system pro-
duces overall pollutant loading configurations for the receiving water sys-
tem for each pollution parameter investigated. These overall loadings com-
puted by the QQS model may be used as the inputs to a simulation model of
the entire receiving water to perform oxygen balance calculations.
STATISTICAL ANALYSIS
Analysis Possibilities
Continuous simulation deals with a multitude of events including rain-
fall as well as discharges and overflows from the sewer network that cause
loadings of the receiving water system. All of these events may be charac-
terized individually using various criteria such as durations, totals, mean
values, etc. However, it would be prohibitive to examine and evaluate each
and every individual event.
In order to obtain a clear picture of rainfall, runoff, overflow, and
receiving water loading activity and to establish interrelations that may
possibly exist, a statistical analysis of the fundamental variables must be
performed. The results of such a statistical analysis should be presented
first in the form of frequency distributions. A frequency distribution is
developed by allocating the individual values of a variable to classes pre-
viously established. A class is defined by an upper and a lower limit, and
only those values of the variable that remain within the range defined be-
long to that class. Each class contains a number of values which normally
differs from the numbers for other classes, i. e. , the individual classes
have different frequencies, that form the frequency distribution.
If the frequency distribution is then integrated, a cumulative frequen-
cy curve is obtained. If a statistical evaluation is made of the results,
which are related to the time span of the calculation interval (five min-
utes for the QQS program), the frequency can also be related to the dimen-
sion of time. In this case, the cumulative frequency curve is called a dura-
tion curve.
25
-------
The cumulative frequency curve or duration curve shows how often or for
how long a value is reached or exceeded, and for this reason is the most im-
portant and informative statistical curve that may be used to define dis-
charge and overflow behavior. The relationship between the frequency dis-
tribution and the cumulative frequency or duration curve is shown schemati-
cally in Figure 8.
Beside the simple one-dimensional frequency distribution discussed
above, the two-dimensional frequency distribution is also of practical im-
portance. Two variables are examined together, with the intention of estab-
lishing a statistical relation between the variables (e. g., overflowing
pollutant load in relation to receiving water flow). For example, critical
combinations of receiving water loadings and base flows may be determined,
in this way. For all individual events the two values of the variables con-
sidered are allocated to combined classes so that individual frequencies of
value pairs are obtained for every combination of two classes. Because these
classes have two dimensions, a frequency surface is obtained (Figure 9).
Because the numerical values of the variables examined may extend over
wide ranges, it was necessary to include the possibility of using geometric
or logarithmic values for classes. In addition, by input to the program the
user can specify any desired class limits. The broad choice of class ar-
rangements is also advantageous for handling the skewed frequency distribu-
tions that predominate in hydrological data because these distributions can
thereby be made similar to symmetrical distributions, thus easing evalua-
tion.
As a general principle, the class arrangements noted may be applied in-
dividually to every variable and, in the case of overflow and receiving wa-
ter statistics, to every overflow node and every receiving water node. How-
ever, the classes may also be specified collectively for all nodes analyzed.
Statistical analyses may be done on an annual basis, or for several
months (e. g. , May through September), or for individual months.
Rainfall Statistics
As mentioned earlier, it is sufficient to simulate continuously the
rainfall-runoff of a number of selected representative years rather than a
complete long series of 20 years or so, in order to determine storm sewer
discharge and combined sewer overflow behaviour of a drainage system and the
subsequent water quality characteristics of a receiving water body. There-
fore, to select a suitable time span for continuous runoff simulations,
rainfall statistics may be examined within the scope of the QQS concept.
Rainfall observations, which should extend over at least fifteen years, pro-
vide the basis for such statistical evaluation.
In most cases a sample of years can be found with basic rainfall cha-
racteristics that agree with those of the entire rainfall series available.
A satisfactory conformity exists if the cumulative frequency curves or dura-
tion curves of the more important variables are identical, such variables
for example, as rainfall duration, elapsed time to the interval with the ma-
26
-------
a
n
c
cumulative frequency
curve or duration curve respectively
frequency, cumulative frequency or duration *-
Figure 8. Relationship between the frequency distribution
and the cumulative frequency or duration curve.
frequency surface
variable 1 divided into classes
Figure 9. Frequency surface obtained from
two-dimensional statistics.
27
-------
ximum rainfall intensity, volume of rainfall, and mean intensity per rain-
fall.
In addition, the portions of storms occurring during the summer and
winter season may be studied by means of a statistical evaluation of rain-
fall duration, volume of rainfall, and mean intensity per rainfall. Know-
ledge of these values helps in deciding whether, with respect to overflow
behaviour, entire annual records have to be simulated or if simulation of
the summer seasons would be sufficient.
Appendix 5 of Volume II, User's Manual, defines the variables that can
be subjected to statistical analysis in the QQS program.
Discharge and Overflow Statistics
The stormwater outfall discharges and combined sewer overflow data for
several years of record, as determined by way of a continuous simulation of
a sewer system, form the basis for discharge and overflow statistics. Such
data are compiled especially for overflow chambers, overflow weirs of deten-
tion basins, stormwater outlets, and outlets of detention structures operat-
ing under pressure.
Those variables which characterize the overflow events with respect to quan-
tity as well as to quality and which can be subjected to statistical analys-
is are detailed in Volume II, User's Manual, Appendix 5. They total more
than 100 variables or pairs of variables. Among these, the most important
are overflow frequency and duration, storm discharges or combined sewer
overflows, and pollutant loads and concentrations. As a general principle,
all analysis possibilities described above may be applied to every variable
and to every discharge or overflow node. In most cases, however, it is suf-
ficient to examine only a small number of important characteristics in order
to arrive at an assessment of discharge and overflow activity.
Apart from the statistical results, a number of single values may be
derived for any time span desired (e. g. , a given year or season) from the
data used in the overflow statistics, such as
0 total volume of water discharged to receiving water
0 total pollutant loads discharged to receiving water
0 maximum and mean concentrations of pollutants dis-
charged to receiving water
0 duration of longest overflow event.
Single values of this kind help in assessing the portions of the total
annual pollution load which are conveyed respectively to the receiving wa-
ter body and to the treatment plant.
Receiving Water Statistics
A loading characterization determined through a continuous receiving
water simulation over several years provides the receiving water statistics.
Evaluations are first carried out for all loading nodes, viz., the outlet
28
-------
locations where stormwater discharges and combined sewer overflows enter the
receiving water. The characteristics which can be statistically examined are
listed in Appendix 5 of Volume II, User's Manual. These characteristics are
evaluated for sewer system outlets (i. e., inflows to the receiving water)
as well as for the receiving water cross-sections just upstream of loading
nodes.
Variables related to total hydrograph duration, such as mean intensity
of receiving water pollution or mean concentration, may also be statistical-
ly analyzed. These variables, however, have minor information value, because
pollutant load variations tend to be damped and to level out in the water
body.
All receiving water evaluations may be perfomed either taking into or
leaving out of consideration the base values of receiving water flow and
background pollution. The most important criteria for receiving water pol-
lution assessment are the loading concentrations with their pertinent load-
ing frequencies and durations.
29
-------
SECTION 3
COMPUTER PROGRAM OVERVIEW
OVERALL COMPUTER PROGRAM DESCRIPTION
The following explanation of the computer program package pertinent to
the QQS model concept is intended to provide condensed information on the
program structure to the extent that appears to be necessary to understand
Section 4 and 5 of this report. Detailed instructions about data input pre-
paration, job control language, program error messages, and program output
interpretation are given in Volume II, User's Manual.
The program package of the QQS model comprises of a number of indivi-
dual programs: a data edit program, a sequence of three programs necessary
to perform the actual runoff simulations, a program to print the runoff
hydrographs and pollutographs in the case of single events, and a statistics
program to evaluate continuous simulation results. In addition, there are
two optional programs included in the package. The one simulates rainfall-
runoff from individual catchments (not including major sewer routing), the
other analyses precipitation data statistically. Figure 10 shows the se-
quence in which the individual programs are applied. An overall flow chart
is provided in Figure 11.
The program DTCHCK checks the network quantity and quality data for
format and plausibility. In case of errorless data (according to input spe-
cifications) the internal input files for the simulation part of the pack-
age are created.
The program DWTFLC simulates one full day under dry-weather flow con-
ditions and establishes internal files (IF13 and IF21).
TXTFCE provides the headings for the output in English.
RCVRIN and MNTWKC perform the actual simulations of quantities and qua-
lities of runoff from the catchments and their transport throughout the sew-
er network.
MNTWSP prints single event simulation results which were written by
MNTWKC on an internal file (IF23).
For practical application of the QQS program package it has been found
beneficial to treat the sewer network and the receiving water as separate
systems, even though runoff and pollutant transport calculations in both
systems employ the same principles. This separation permits the performance
of data checks and simulations for verification or related purposes indivi-
30
-------
1 I/I
J£
o
+J
o>
s- z
£ s.
OI
C S
O O>
i- CO
-u
(O "O
r- 01
M
M- (_}
u-
O T3
C C
n re
a: _
Sto
(catchment runoff I QQSEGL
simulation, optional)
(print of single
event results)
(statistics of pre- |RAIN5C
cipitation data,
optional)
I MNTWSP
(datacheck)
(dry-weather flow
calculation)
(textfiles)
(storm runoff
simulation)
STATCS I (statistics)
\ Save Results
Delete
O O)
4-> S-
-------
03
J3
u
(0
u
O
00
•H
32
-------
dually for the sewer system or the receiving water system without handling
the other system component simultaneously. Therefore, for the sewer system
and receiving water system, individual program versions have been estab-
lished which differ mainly in their capacity. These individual program ver-
sions also contribute to more economical computer use.
The pertinent program names for receiving water system routing are
RCVRIN, MRBNTC and MRBNSP. Receiving water network routing can be started
only after a corresponding sewer network simulation has been completed.
The results of continuous long-term simulations are stormwater dis-
charges, combined sewer overflows, their pollution quantities, and receiv-
ing water loadings at specified nodes of interest. These data are organized
and statistically evaluated in STATCS. One-dimensional and two-dimensional
statistical analyses are performed for those properties and nodal points de-
signated in the initial input instructions.
QQSEGL is a program which facilitates calibration and verification of
drainage area calculations. This program simulates runoff and runo-ff quality
from individual drainage areas and is an excerpt of the pertinent subrou-
tines of MNTWKC. This program is included in the package but its use is op-
tional.
The program RAINSC provides a statistical analysis of precipitation da-
ta and serves to select a suitable time span from longer rainfall observa-
tions for continuous rainfall-runoff simulations. This program is included
in the package, but its use is optional.
The use of additional programs may prove beneficial, e. g., the HEC1
program for derivation of unit hydrographs (7), depending on individual pro-
ject needs.
COMPUTER PROGRAM CAPACITY
The capacity of the QQS program system as described in this report is
sufficient for normal applications of total service areas up to approximate-
ly 6,000 ha (15,000 acres) in size. At this approximate upper limit, with
100 drainage areas their average size would be 60 ha (150 acres). Consider-
able departures from these figures may be encountered in applications to
diverse service areas. Regardless of the size of the total service area,
three different rainfall records may be handled simultaneously in the simu-
lations. Each of these records is ascribed to a fixed area by input. To a
certain degree this will provide some consideration for nonuniformity in the
distribution of precipitation.
Single event simulations can be made for a period of up to 24 hours.
The time step is fixed at five minutes. Although it is desirable for the
scope of urban runoff simulation to derive the pertinent input from records
allowing for similar fine interpretation, it is still possible to interpo-
late the input data required in five-minute detail from less refined rec-
ords. Single event simulations required for more than one day of data may be
33
-------
performed by splitting the event and adjusting the initial conditions of
successive runs.
Continuous simulations of up to 20 years, the time step being fixed at
five minutes, are possible, but to save computing costs the selection of a
4 to 8 year period based on statistical analysis of precipitation criteria
has been found to suffice without sacrificing reliability.
From one to four pollutants can be considered simultaneously in one
simulation run. If necessary, additional pollutants may be handled in sub-
sequent runs.
Either metric or English units can be used. The major data processing
capabilities are summarized in Table 5. It should be noted that the program
dimensions given are not rigid but may be adjusted by the programmer to ac-
commodate larger areas and individual needs.
TABLE 5. DATA PROCESSING CAPABILITIES OF THE PROGRAM SYSTEM
Maxima
Length of data for a single event simulation 24 hrs
Length of data for a continuous simulation 20 yrs
Statistical variables, total cases 118
Precipitation stations 3
Conservative pollutants 4
Area types (considering pervious and impervious separately) 4(8)
Receiving waters arbitrarily connected 2
Sewer network - system elements 400
- nodes 370
Drainage areas 100
Special structures - overflows 35
- basins and pumps 15
- operational controls 20
External loadings - constant 18
- variable (only for single event simulation) 10
Receiving waters - system elements 150
- nodes 130
Specified nodes for statistical analyses - special structures 50
- arbitrary nodes 5
34
-------
COMPUTER PROGRAM HANDLING
Computer Requirements
The complete QQS program package is written in Fortran IV and consists
of approximately 30,000 statements. To date, the program package has been
operated on UNIVAC 1108, UNIVAC 1110, DEC 2050 and AMDAHL 470 computers.
However, the package can be used on all BATCH processing systems with For-
tran IV compilers. A maximum usable core storage of 122 K-words (UNIVAC),
which corresponds to 488 bytes, and a configuration with fast external mass
storage are required.
Input Data
Input data are divided into four groups: precipitation, network and
drainage area, quantity, and quality. All input may be provided in metric or
English units, except for programs QQSEGL and RAINSC which only allow for
metric input. Format requirements for all input data are provided in Section
3 of Volume II, User's Manual.
For continuous simulation, precipitation intensities must be supplied
at 5-minute time intervals and dry spells between individual events must be
specified. Gaps in precipitation records should be filled with suitable ap-
proximations. Control of measured precipitation records should be made by
checking them against totalizer registrations. For single event simulations,
precipitation data are entered with the network and drainage area data as
part of their input.
Network and drainage area data consist mostly of geometrical data de-
fining the sewer network or the receiving water system, logical connections
of the individual elements, drainage area characteristics, and assignment of
drainage areas to network nodes. Generally, these data may be derived from
sewer network plans and aerial photographs.
Quantity data consist mainly of the data necessary to determine runoff
quantities from drainage areas and dry-weather flows. Such data are unit
flow hydrographs and data necessary to evaluate the effective precipitation,
i. e., evaporation rates, initial or starting losses, and soil infiltration.
Diurnal variation of dry weather flow, boundary flow conditions and receiv-
ing water flows are also specified in this group. These data are obtained
from meteorological surveys, waterworks records, etc.
Quality data are analogous to the quantity data. They provide pollu-
tant load unit graphs and their modifying functions, dry-weather flow quali-
ty, pollution in external loadings, and receiving water background pollution
levels. Characteristic functions for pollutant accumulation and decay in the
catchments, detention and purification in detention basins and treatment fa-
cilities, and data specifying intervals and efficiency of street cleaning
are also contained in this group. Most of these data are derived from meas-
urements and engineering experience documented in the literature.
In order to facilitate application of the program package and to avoid
waste of computing time, the input data are checked by an extensive data-
35
-------
edit program for errors and for plausibility. Such a data-edit is critical
for an efficient and economical application of a comprehensive simulation
model.
Output Data
The major single event simulation results are hydrographs and polluto-
graphs, obtainable for any pipe element or any nodal point in the network.
The output values are water surface elevations, flows, pollution loads and
concentrations per time interval. Special outputs are given for structures
like outfalls, overflows, detention facilities, pumping stations, and treat-
ment facilities, providing information for all ingoing and outgoing con-
duits. For all outputs there is a choice of metric or English units. Output
headings are in English.
In addition to hydrographs, and pollutographs, total runoff and total
pollution loads washed from the study area are given for each simulation.
Critical backwater conditions and their locations are indicated by a special
output. Further, quantities and qualities of runoff from individual catch-
ments may be handled and printed out separately by the program QQSEGL, a
feature which has proved to save considerable time for catchment calibra-
tions and verifications.
In principle, continuous simulation results are similar to those for
single events, but the mass of information generated can be evaluated only
by means of statistical analyses.
The major outputs are annual and monthly frequencies, cumulative fre-
quencies, and durations of variables such as discharge and overflow volumes,
peaks, averages, intensities, and associated pollution values. In all, 118
different cases of variables may be considered. One-dimensional statistical
results are tabulated and graphed. Two-dimensional frequency distributions
can be formed for any combination of variables specified in the input. From
two-dimensional statistical results coincident frequencies and durations may
be obtained. One may also easily check ranges of concurrence of hazardous
conditions such as low receiving water flows and high overflow pollutant
loads.
Further, annual and monthly totals and averages of discharges and over-
flow quantities and pollutant loads are given for the individual outfalls
and overflows discharged to receiving waters, as well as their total impact
on the receiving water. As for single event results, total runoff and total
pollutant loads washed from the study area are provided over the time span
simulated.
Output examples for single event and continuous simulation results are
provided in Section 1 of Volume II, User's Manual.
36
-------
SECTION 4
MODEL TESTING
HYPOTHETICAL TESTS
Battelle Pacific Northwest Laboratories evaluated several mathematical
models that simulate time-varying runoff and water quality in storm and com-
bined sewer systems. Quantitative and qualitative routing was tested simu-
lating runoff through relatively simple hypothetical networks consisting of
segments with different slopes and diameters. Included in this evaluation
were some of the earliest tests made of the QQS program, the results of
which have been reported in Appendix E of the pertinent report (1).
Hypothetical networks were also employed in the early stages of testing
of the QQS program, to illustrate the plausibility of flow routing proce-
dures used. The pipe reach for which runoff was simulated had a very flat
slope of 0.05 percent and the entering hydrograph was for rapidly-varying
rates of flow, conditions which are usually particularly challenging in ma-
thematical modeling. Figure 12 shows the output flow hydrographs obtained at
points 600 m, 1,200 m and 1,800 m downstream from the inlet point. The hy-
drographs are quite smooth and exhibit the expected damping effect.
COMPARISONS WITH MEASUREMENTS
Requirement for Measurements
Comparisons of measurements with results from simulations should be
used to calibrate the model to a given study area and application. Some in-
put functions such as unit hydrographs may be directly derived from suitable
measurements made in the study area.
For derivation of input functions and model calibration, the following
information is normally necessary:
0 rainfall recordings of one or more stations;
0 corresponding runoff measurements or water
surface elevation recordings (and reliable ca-
libration curves), synchronized with rainfall
recordings;
0 samples taken at sufficiently short time in-
tervals during runoff for the pollutants con-
sidered;
37
-------
6 428
S 000 - -
3 r.70 - -
hydrograph at
600 in
1200 m
1800 m
downstream of inlet
a
2 14? - -
07U --002
0000
10
'.0
I
50
60 70 80
90
T i i i r r
100 110 120 130 HO 150 160
Time imin)
Figure 12. Hypothetical test of flow routing for a 0.61 m (2 ft) diameter
pipe at a 0, 05 % slope and with free inflow and outflow.
0 indications of dry-weather flow quantity and
quality prior to and after stormwater runoff;
and
0 time, duration and volume of preceding rain-
fall event(s).
A test catchment should be selected such that the boundary conditions
allow for an unequivocal evaluation of the measurement. The following should
be fulfilled:
0 there should be no runoff entering the area
under consideration from outside, unless it
has been monitored;
0 there should be no backwater effects within
the test area caused by downstream influences;
0 the rainfall station must be located within or
close to the test area; and
0 the characteristics of the test area should be
representative of suitable portions of the
study area.
In view of the intended comprehensive analysis of combined sewer over-
flow behavior, emphasis should be placed on the collection of data for storm
events that cause relatively large overflows.
38
-------
Comparisons of Different Catchments
A number of comparisons of QQS simulation results with field measure-
ments have been performed. Table 6 includes only those comparisons for which
complete sets of measurements were available. An extensive number of tests
using incomplete sets of field measurements have also been performed.
In general, the necessary simulations were made using the program por-
tion for drainage area runoff simulation only (QQSEGL). However, the simu-
lation results for sites 8 and 31 in Rochester, N.Y. , and for the seven mis-
cellaneous locations in the sewer network of Vancouver, B.C., Canada, were
taken from simulations using the programs DTCHCK, TXTFCE, MNTWKC, RCVRIN and
MNTWSP (drainage area runoff and flow routing simulations).
All test areas and locations are described briefly below. Comparisons
are shown in illustrations only if the calibration effort of the corres-
ponding project had been completed at the time of writing, and if permission
was obtained from the agencies, that made the measurements available, to
publish the complete sets of comparisions performed.
For the comparison results presented, the input functions such as unit
hydrographs and unit pollutographs are kept the same for all runoff simula-
tions of a particular test area or network location. For each test area, the
model was calibrated using all information available. Initial conditions are
derived for each event to the extent that information on dry spells and pre-
ceding rainfall events could be obtained. Time synchronization of precipita-
tion and runoff measurements frequently proved to be unreliable. Therefore,
individual calculated hydrographs are shifted by up to two intervals (10
minutes) on some of the figures shown.
Augsburg Test Area
The test area of Baerenkeller in Augsburg, Germany, is mainly residen-
tial and is drained by a combined sewer system. Approximately one-fourth of
the total area of 74 ha is impervious. The catchment houses 6,780 inhabit-
ants, 3,750 of which leave the area during working hours. Railway tracks
limit the test area to the north and south and to the west is another resi-
dential area. Flow measurements and samples that were specially taken for
the pertinent QQS project, were gathered in a sewer line located at a ground
surface drop representing the eastern boundary of the test area. All bounda-
ries are clearly definable. Because of an appreciable ground surface drop
and the resulting drop of the sewer line at that location, there are no
backwater effects from downstream influences.
A meteorological station of the German weather bureau is located 1 km
outside of the test area. All rainfall recordings used were obtained from
this station. All runoff measurements were taken in a 1.0 m diameter manhole
at the upstream end of the sewer line located below the ground surface drop.
For this purpose, a side street was closed to through traffic for six
months.
39
-------
CO
H
J2^
g
H
3
CO
^
W
s
9
W
PM
CO
Q
U
•H
4-1
10
•H
M
01
4-1
O
CO
S-l
CO
U
d
o
•H
4-1
CO
CJ
0
U
PM
CM
PQ
PM
co
CO
co
CO
Q
O
CJ
•vf
o
d
S
m
Q
O
o-
0)
o
I— 1
C/3
M
3
4-> 0
d -H
01 >
U i-l
S-l 0)
0) ft
CM E
1— 1
CO s~^
M J3
< ^
4->
d
0)
w J3
0) U
H 4->
CO
U
><
4-1
•H
H^
CO
ft
•H
CJ
•H
d
2
^
Jbj
4_)
d
3
O
U
IT)
> >
cO CO CO
LO F"*~ ^O
CM co co
o r-~ o
^J rH CO
r~-* co CM
s-i
0)
rH
pH
0)
d co u
0) d co
S-l W rH
0) 0) rH
co 3 3
00 M
S-l CO
3 oo .d
,a 4-> u
W 4-> -H
oo 3 d
34-13
u
rH -H
CO rH
H rQ
OJ 3
T3 ft
0) 0) 4-1
PM PS O
CO lO 00
i— 1 rH
^ ^o oo
i— t
rH O CO
rH IN rH
0)
00
CO
ti | | | t
W +J 4-J
0) (0 CO
> rH rH
CO M-l M-l
ON r^ ^o
CO CO rH
o o o
jsj \^ 00
•^ ^ co
in
S-j
0) 1
4-1 J3
00 T3 4-1
d 01 n
•H CO 'H
fH j_J QJ ^J
O W to 1
CO rH CO m
rH O CO 0)
S-t 00 S-l W 00
co d 4J o 01
•d Ji 43
u u o
•i-l -r-l -H
d d d
33 3
>>
d
CO
^
0)
o
vO
VO
LO rH
rH
in
rH ^3"
rH
01 01
00 00
CO CO
S-l S-l
0) 01
> >
CO CO
O *sO
*d" rH
^* CM
in m
rH O
>-l
d «
(U TJ
o d
•<-> CO
W rH
00 10
S-l S-l
0) O
pa H
00 00
3 3
f £>
d d
0) 0)
,d x!
4-1 4-1
0 O
0 0
d
0)
-o
OJ
?
to
CO
J>
CO
1
CO
rH
LH
rH
CO
vO
co
CO
4->
01
0)
J.J
4-1
CO
r"{
4-J
^_4
0
a
o>
4-1
ia
0)
(**}
CJ
O
PS
*^3
co
3
01 CM
> >
<0 CO
1 1
(0 CO
rH rH
Ifl 14-1
r^- rH
CO Sf
r-- m
oo r«-
CM ^^
4-1
0) 0)
3 0)
d S-l
0) 4-1
> CO
<
d
0) >
^ rH
0) O
Q U
D D
4-1 4-1
CO CO
01 01
f] f"j
U CJ
O 0
CN CM
CM CM
CM CM
CM CM
> >
CO co
I i
CO CO
rH T— 1
l4-| 14-1
rH in
vo m
0 O
O 00
CO 00
00 rH
CO
OJ 01
4-1 4-1
•H -i-l
co co
01 01
4-1 4-1
to w
0) 0)
pd *d
U CJ
0 0
1— 1
rH CO
co
1^, ,_,
rH
0) 0)
00 00
CO CO
Lj Lj
M W
OJ OJ
> >
CO CO
CN in
^- oj
00 SO
CO vO
r^. £\J
4-1 •
0) S-l
0) 4-1
S-l CO
4J
CO 0)
•H
CO N
4-i d
S-l 01
0) .id
rjQ CJ
rH CO
< a
0) 0)
> >
3 3
O 0
CJ U
d d
CO co
> >
CO
TJ
CO
{3
CO
CJ
^
>>'
CO
i i
4-1
«
1— 1
<4H
0
4-1
p**
i — I
O
4J
ft
H,
4-1
OJ
55
.
Lf-|
4-1
•H
Q
r--
0)
>
3
0
U
d
CO
>
vj" s^
U D
ft 00 00
0) CO co
rt 1 L_t LJ
UJ r-f M
4-1 0) 01
to > >
1 CO <0
m o o
o> in m
000
CM m r-~
r^» -^ CM
so
•
co 01
d ^
O 0) <
•H 3
CO 0) O
0 > 0
o < s
rH CU
CO rH
M 0) 4-1
S-l d CO
O O CO
* ^ CJ
o
4-1
d
o
o
H
40
-------
Water depths were recorded in the manhole and corresponding flows were
defined by rating curves. Samples were drawn manually at the manhole for
analysis of BOD5 and settleable solids in time intervals of 5 to 10 minutes
during storm events. This procedure was made possible by the fact that the
field crew was based in a treatment plant located approximately 7 km from
the sampling location. Once the crew anticipated a coming rainfall event, it
could usually arrive at the sampling site prior to the start of runoff in
the test area.
All samples taken were analyzed in the treatment plant laboratory for
BOD5 and settleable solids. Analyses were performed according to DEV guide-
lines (German standard methods for the examination of water, wastewater, and
sludge). During the test period in the summer of 1974, flow was recorded and
samples were taken and analyzed for approximately 20 rainfall events. The
data from only five of these events were sufficiently complete to be used
for model calibration purposes. Considering the difficult field conditions.
this low yield was quite good. Because two of these five rainfalls had very
low precipitation, emphasis was placed on the more intensive events of July 17,
1974, August 27, 1974, and September 25, 1974. Comparisons of the results
for flow, BODr, and settleable solids are shown in Figure 13 for these
three events.
Stuttgart Test Area
One of the first sets of comparisons between QQS simulations and field
measurements was for five storm events monitored in the Buesnau test area of
the Technical University of Stuttgart, Germany. This test area has a size of
32 ha, 38 percent of which is impervious, and a population of approximately
4,000 inhabitants. This catchment area is mainly residential with single
homes and is drained by a combined system. The slope of the catchment ranges
from flat to a slope of 6 percent. The steeper parts of the catchment, how-
ever, have no particular influence on surface runoff because individual
properties are mostly limited by walls that cause steps in the surface. More
details on this test area and on the data collecting system are contained in
a report issued by the Technical University of Stuttgart (11).
Figures 14 and 15 show comparisons of QQS simulations with measurements
for the storm events of September 19, 1967, March 15, 1968, May 17, 1968,
and December 3, 1968, for flow, BOD5, and total suspended solids. Figure 14
also shows the event of September 6, 1967, for which only the quantity of
runoff is compared.
Munich Test Areas
In Munich, Germany, rainfall-runoff measurements were made and samples
were collected in four test catchments of different sizes and characteris-
tics. The metering stations of the Pullach and Harlaching test areas were
installed and operated by the Technical University of Munich. The Ingol-
staedter Strasse and Josef-Wirth-Weg stations were operated by the engineer-
ing department of the City of Munich. All measurements available from the
four stations were used to calibrate QQS input functions for application
41
-------
5 o S S [ujtJU Q / mm] 'Q9Jd-*-
0)
in
CM
S! g S S S S 8 8 3 g
0^0 O
-g
[U!iug/6>|]
0)
01
o
60
V4
a
x>
M
oo
a <
<
/— N m
(0 CM
^ 2 [uiiUQ/uiui] -oeJd-*-
s a
OJ W
(-1
PQ
4->
•*• W
to a
4J 00
C 3
>.
[U|UJ9/6>|] SS
o o o o o o c
- [8/eUJ] 0
W 4-1
d d
O 0)
w >
•H D
t-l
0) S
E O
O 4J
U in
00
•H
SS
42
-------
•iH
-------
[tnwg/iuui] '
r3
O Q
2 2
0)
[s/eiuj
SSI
S-i
0)
ifl
60
+J 00
4-> v£>
3 CT\
Jj r-l
saoa
~ on
,~-S
ca ^
c^ S
CO 0)
v—' O
OJ
3 P
w d
W >>
•M (Q
c s
(D
e «
d d
O 0)
•H (U
00
CO
w
in
2 S g 3 3
o d o o o
•— [s/eiu]
-------
in the overflow abatement study presently under way on the entire urbanized
area of Munich.
The Pullach test catchment is 23 ha in size, of which 36 percent is im-
pervious, and has an average slope. This single-dwelling residential area is
drained by a separate storm sewer system. Catchment area and metering sta-
tion features, sampling procedures and the data that have been collected
have been reported and discussed (2). Complete information on flows, BOD5
and settleable solids measured were made available for comparisons with QQS
simulations.
Since the summer of 1976, precipitation, runoff, COD, some BOD5, set-
tleable solids, total suspended solids, and chemical data have been col-
lected from the Harlaching test catchment. This mainly residential area has
a size of 542 ha and is drained by a combined sewer system. Flow depths and
some quality parameters are monitored continuously and water samples are
drawn automatically. All collected data are checked and stored by data pro-
cessing equipment. Some of the early measurements were provided for the
calibration of the QQS model in connection with above-mentioned combined
sewer overflow abatement study. The data collection effort is being contin-
ued.
The Ingolstaedter Strasse test catchment with an area of 96 ha and an
imperviousness of 37 percent has a rather flat slope. The test area is
drained by a combined sewer system. Land use is mainly residential and rep-
resents a mixture of older single homes and newer high-rise apartment build-
ings. In order to obtain clear boundary definitions for this test catchment,
some sewers leading into the test area were cut off during the data collec-
tion period. A metering station was installed in an easily accessible area.
Two manholes were constructed to contain the metering equipment. In both
manholes, water depths were monitored whereby the measurements in the down-
stream manhole only served the purpose of checking for the occurrence of any
possible backwater effects that might penetrate into the test catchment. The
water quality samples were drawn in the upstream manhole. A calibrated me-
tering weir was installed between the two manholes. One water depth recorder
and the automatic sampler were stationed in a mobile shed that was parked
over the upstream manhole. The second water depth recorder was installed in
the downstream manhole. For the storm events of September 3, 14, 15 and 17,
1976, only runoff quantity measurements were available. In Figure 16, QQS
simulations of runoff quantity have been compared with corresponding meas-
urements. Figures 17 to 23 show comparisons for flow, BOD5, and total sus-
pended solids for the storm events of September 28 and 30, November 3 and.
December 2, 1976; and June 27 and 29, July 13, 19, 20, 25 and 31, August 1,
9 and 18, and September 4 and 9, 1977.
The fourth test area, monitored at Josef-Wirth-Weg, is a mostly single
dwelling residential area that is partially industrial. It is 33 ha in size,
has an imperviousness of 16 percent and a fairly flat slope. A peculiarity
of this sewer system is that while street runoff and domestic sewage is dis-
charged to sewers, roof and backyard runoff is infiltrated into the ground.
The metering station was installed over an existing manhole just upstream of
a pumping station and water depths were recorded and quality samples were
45
-------
[uwg/iuui] -
IN -J ID » O
O 000
10
[•/eiyj
(O
N
O)
CO
-S
ui
E
.1.
»5 n
« §18 •
iislfi
P: i!
[•/em]
(0
h.
0)
c
1
o>
.E
E
a
u
OJ
o
•H
a
2
(U
Id
QJ
(Q 'O
4J d
w n)
iH
O LO
60 I-H
d
CO
•M •>
d fo
0)
e n
0) tU
M 01
rt -M
D a,
J-) '-H
•rl O
£
M
M 4->
c a
O V
w >
•H OJ
u
(Q S
ft n
S O
O -P
O M
0)
(-1
fl
60
•H
PH
^— [8/cuJ]
»ounu
46
-------
[ujwg/uiuj]
jj
W CM
w
« l-i
S-l CJ
•U XI
W E
OJ
K O
0)
00
«N
I i ti^
•H O
CO
CO 4J
d d
O 4J
w >
•H (U
i-l
<0 £
E O
O ±J
u w
OJ
i-l
60
•H
[s/cui] O •*- ["!"J9/B>|] SQOB -^— [U!UJ9/B>(] SSI
47
-------
[uiiug/iuui] -oejd -»-
-3
o OOOOOOOQ
[s/euij o —-[UIUIS/6M] saO8
S 3 S S 2
iui 9/6>|] SSI
a a o' - - -
S 3
•S
J-3
[s/eiu]
SQO9
82332°
[uiais/6>(] SSI
a
a
M
0)
o
o
•1-1
C!
w
n)
+J
00 1-4
4J QJ
OJ ft
tfl 01
4J 00
w
i-l T>
o d
60 n)
d
W I-H
4J
01 00
01 1-1
1-4 01
w S
n3 0)
01 >
S O
•H O
S
w
W 4-1
d e!
o 01
w >
•H O)
S O
O 4-J
u w
00
D
1-1
00
•H
48
-------
[uiuig/uuiu] 'oejd -*•
[s/eoi] o -*-[u!Oig/6>f] 5aog
(fl
(-1
43
u
•rl
U
2
(S
SSI
CO
(U
w
w
-3
-8
a
•S E
-3
oo»*f^*in>^«r^^0
~'do d o o *3 <3 o o
§00000 ooooq
a *> * N 32§«a
[s/E"J] 0
Tuiuug/tuiul -
[uaug/6>i] SSi
r--
r-.
co
4-1 cn
T3 r-t
4-1 i-l
w 3
i-t *"3
O
00 T3
c d
W CM
4J
d -o
(U c
S «d
•r-4 01
1-1
S O
O 4J
U W
3
00
49
-------
N
N.'
K
N.
O)
>>
HI
C5
O
•H
a
0)
W
CO
S-l
OJ
QJ
OS
O
60
CO CN
4-1
C 13
•H 11
^
n) S
ft U
6 O
O *J
O to
O
CN
00
•H
essssse"
[s/eui] o -*-[ui
ssi
50
-------
[s/eui] o
saos
in
CM
M ^ V>
= C IS ?r c
o
r E E a> £
2 0> T W OJ
is^g 5
i£«sE
2 »O o a)
0.20^2
P: ii
I -8
S § 8 S S 3
§ § I
o o o o o
c
J,
0)
(U
o
X!
U
•H
0)
CO
w
CO
i-l
4J
C/3
S-l
CD
JJ
T)
CD
CO
o
60 c
C 1
d t?
CD d
S cd
01
s-i m
d CM
w
C8 >>
OJ r-l
S =>
!-)
4-> IH
•H O
S
w
w j-i
u d
O CD
W
OJ
« S
On M
S O
O -P
U W
CM
01
S-l
60
•H
[s/eiu] o
saog
SSI
51
-------
a a a a - -
t;
cq
0)
S § 8 S S 5 S
[s/ciu] D — [uiuig/6>(]
c
g 1
-S
-S
S-i
OJ
u
•rl
d
[u|iug/6>|] SSI
o\
4)
n)
IUJLU g/Lutu] '
•- w « •«
o ci O o d
CO
a c S
tfl
r-l r~~
O r~-
60 c^
a I-H
CO
4J
S
0) rH
3 4->
10 10
n) 3
0
^
4-> 4-1
•H O
CO
1=1 C
O 0)
CO >
•H 1)
ft n
S O
O 4J
(] SSI
52
-------
O)
SO O O O O O o O
tom^nr*-- iv
U
•H
d
2
01
to
en
«>
(-1
4-1
l-l
(U
-•— [s/Etu] O — [uit
saog
SSI
CO
00
OOOOO
f* & lf> -*1 T*l
Ift ;£ W
|!3°I
• § .§ s i
*s*s$
gS» 5 S
« eO-s 01
0.2OH2
P: i =
03 h
4-1 OJ
O 01
00 4-1
C ft
M 0)
W "O
4-> a
ci M-t
•H O
cn
U 4-1
a a
o 01
w >
•H 0)
(-1
« a
i|]
[u(UJ9/6>)] SSJ.
53
-------
drawn at this manhole. The water depth recorder and an automatic sampler
were located in a mobile shed stationed on top of the manhole. A calibrated
measuring weir was installed on the rim to the pump well for flow evalua-
tion. In Figure 24, flows are compared, for the storm events of September 5,
14 and 17, November 3 and December 2, 1976. In Figures 25, 26 and 27, com-
parisons are shown for flows, BOD5, and total suspended solids for the storm
events of September 15 and 28, 1976, and for June 29, July 13, August 9 and
18, and September 1 and 9, 1977.
For both the Ingolstaedter Strasse and Josef-Wirth-Weg metering sta-
tions, Manning dippers and Manning samplers were used. The equipment was ad-
justed so that sampling would start shortly after dry-weather flow runoff
rates were exceeded. The samplers each had a capacity of 24 half-liter sam-
ples. The time interval at which the samples were drawn was set initially at
15 minutes but was later shortened to 7.5 minutes. This change was found
necessary to record the first flush that occurred during the rising limb of
the runoff hydrograph.
The metering stations were inspected daily when the recording charts
for the dippers were exchanged and the equipment checked. From time to time,
deposits that accumulated upstream of the metering weirs had to be removed.
Samples taken were analyzed in the treatment plant laboratory for BOD5, COD,
and total suspended solids.
In the Pullach and Harlaching test catchments, mentioned earlier, pre-
cipitation was monitored within the catchment areas. Both the Ingolstaedter
Strasse and Josef-Wirth-Weg test catchment are situated close to the meteo-
rological station of Munich's Grosslappen treatment plant, and the rainfall
recordings of this station were used for the runoff simulations for these
two catchments.
Gothenburg Test Areas
The metering stations of the Bergsjoen and Torslanda test areas in Go-
thenburg, Sweden, were installed and operated by the Chalmers University of
Gothenburg, Sweden.
The Bergsjoen test area has a size of 15 ha with an imperviousness of
40 percent. The catchment has an average slope and is drained by a separate
storm sewer system. Land use is residential with high-rise apartment build-
ings. Details on the test catchment, the metering station and the data ac-
quisition system have been reported (13). Measurements of flow, KMn04 de-
mand, total suspended solids, lead, and phosphorus are compared with QQS
simulations in Figure 28 through 32 for the storm events of May 8, June 27,
July 10, 20 and 23, August 30, September 20, October 29 and November 11,
1973.
The Torslanda test catchment is a 105 ha residential area under de-
velopment; it includes a school. The imperviousness at the time of monitor-
ing was 16 percent. The area has an average slope. Details on the test
catchment, the metering stations and the data acquisition system have been
reported (17 and 18). Flow, COD, and total suspended solids measurements
54
-------
[UILUg/LULU]'00Jd
(S -4 tD •» O
o o b o **
d
«
S rH
)-l
i; **
O CM
U -fi
•H S
d
JS O
•P iz;
0) -O
w d
o w
to
4J ^
d m
01
e n
01 (U
W 0)
to jj
OJ A
e
•H (U
1-1
(Q E
ft H
E O
O J->
u to
eg
01
^
00
•H
0000000
[s/eiu]
[8/eiu] jjouny
55
-------
[uiujg/iuuj] -oajtj —•-
N
N
N
CO
[iHUjg/6>|] 9Qoa — [uiuug/6^] SSI
0)
O
u
•H
00
[ujiug/uuui] 'oajj ——
r* .j T5
a
• •N n)
in
J-> LO
d .-i
•H 0)
S-J
« s
& o
O 4J
u to
CM
(U
h
«
•H
[s/eai] O
SSI
56
-------
[uiuigyiuai] '
OJ
o
X!
u
•H
c
s
00
CD
•H
I 00
O W
^ 3
60
w <;
E «
0)
s-i IT\
a CM
w
n) 0)
'-H
•H O
CO 4-i
d c
O D
w >
•H 0)
S O
O 4-1
u w
CM
01
S-l
00
•H
[s/ciu] 0 •*-["!•«9/B»l] saO9 •*— [U|uug/B>i] SSI
57
-------
[UIUJQ/UJIU] -
o>
•
o>
£, cy
O
•H
I
[uiUJ9/6>|] SSI
[UILUS/UJO/] -oaid —••
-•I i
III'5!
P: 11
N
N
t
O>
—— fs/Euj] D — [u)ius/6>(] SOOB —— [Ujujg/6>(| SSI
CO
so e?i
-o
i a
X! «
4J
S-l 1-H
• H
S ^
I 0)
4-i ,n
CO CO
(0 3
0) 00
S 3
[ujuig/iuui]
N
•
CO
05
'SI °
u
CO 4->
a a
o
-------
n
N
cd
to
N
>
S 03
2
Jl
j-i VH
•rl O
ca
Ifl 4->
a a
o
•H CU
S O
O 4J
CJ W
oo
CM
01
V-i
00
•H
[s/Eiu] D
SSI
59
-------
[uiuig/ujuj] '09Jd —•-
tf» o «n °
o - - <*
e g o o
3 g S s
o o 3 °
[s/eUJ] ° -^•
h
h.'
c>
o b o o -
r» M ••
d
0)
•o
a;
60
(-1
d
,0
d
o
o
d co
aj r-~
o a\
•<-> I-H
to
oo •»
M ro
0) d
(0 (0
4-)
d o
.
CO
w +->
d d
o u
w >
•H
-------
[tnujg/ujuj] -oaad -»-
CO
CO
•
o
CO
-S
-s
s
« in .4 <•) M •»
o o a o o o
3 S
cJ
0)
•o
OJ
00
^
I
01
ja
4-1
o
o
«-2 „
olio I
£ E E o E
P: is
CO
c
•
O
CO
1 1 1 T
§33335
a
CJ
o
•<—>
in
00
a
OJ -
s o
•i-l OJ
3
00
•H
[8/eOl]
61
-------
[uiuug/uiui] •
8 2 8 ° S
co
N.
O)
CM
[uiiug/8]
[uiujg/6]d
in
C 71
«2 „
S S'l'o S
~ E E « £
fl»sl
5 2(0- «
CO
o>
6
CM
C
•a
C/3
00
I
1)
O
O
•> en
/-> r--
B3 CTi
ID
(0 4-1
00 U
H O
(U
CO t3
C3
[0
o
CM
S-i
w OJ
BJ 4J
CU OH
S OJ
4-) 4-1
•H O
CO
ca 4->
tf (3
O C)
M >
•H
O co
CO
CU
[8/BUl]
[ujtug/6]
[uiuug/6]d
62
-------
„! :
= c »- c
O o: 3 O ft)
= E E " E
P:IJ
c
o
O
efl
IT)
CO
CS
OJ
o
•>~>
CO
00
S-l
OJ •
pa co
p-
S •-!
0)
^4 U
d a)
CO ,Q
ID 6
II
O
XJ fe
+J
•H M-4
tfl •)->
d d
o oj
w >
•H OJ
M
RJ S
ft n
S O
O *J
U W
[u|UJS/6]d
CM
cu
S-l
60
•H
63
-------
available for comparison with QQS simulations are shown in Figure 33 for the
storm events of September 9, 12 and 27, and November 28, 1976.
Rochester Test Areas
Combined sewer overflow quantity and quality field data were compared
with simulations for three test areas and two other sites in the network
within a study by the Rochester, N.Y., Pure Water District.
Three of the test catchments are drained by combined sewers: Dewey Av-
enue (29 ha), Colvin Street (78 ha) and North Street (34 ha). Figure 34
shows comparisons of recorded and computed flows for the storm events of
August 29 and 30, and September 11, 18 and 20, 1975, for all three catch-
ments .
Site 8 in the Rochester, N.Y. , sewer network comprises an area of ap-
proximately 390 ha with an imperviousness of 61 percent; it is drained by
combined sewers and has mixed land uses. The slopes of the individual drain-
age areas vary from flat to average. In Figure 35, measurements of flow,
BOD5, total suspended solids, and fecal coliforms are compared with QQS sim-
ulation results for the storm events of May 31 and June 19, 1975.
Site 31 drains a combined sewer area of 878 ha with an imperviousness
of 55 percent. The individual drainage areas represent mixed land uses and
are of flat to average slope. The same parameters as for Site 8 are compared
for Site 31 in Figure 36 for the storm events of June 5 and 19, 1975.
More details on the measurements and test areas may be found in the
project report of the Rochester Pure Water District combined sewer overflow
pollution abatement study (9) .
Vancouver Test Areas
Vancouver, B.C., Canada, encounters mainly long-duration rainfalls but
also short, high intensity storms. Therefore, data from both short storms
and long rainfalls were used for catchment calibration of the QQS model in
connection with an overflow abatement study of parts of the City of Van-
couver.
The 74-ha area monitored at Alberta Street has an imperviousness of 42
percent and an average slope. The 27-ha area monitored at Mackenzie Street
has an average slope and an imperviousness of 25 percent. Both areas are
drained by combined sewers. For both sites, dippers and automatic samplers
were installed in manholes. The samples drawn were analyzed for BOD5, total
suspended solids, and fecal coliforms. Runoff measurements of storm of up to
20 hours duration were compared with simulations from the QQS model.
The runoff behavior in the sewer network of the Vancouver study area is
quite complex because of the interaction of several pump stations, control
gates, syphons and overflows; therefore extensive calibrations of the QQS
model throughout the study area were necessary.
64
-------
CO
CO
CM
o o o o a
0
[uiiu g/iuuil • 09jj _».
L ' J
S s c S
(O
N
0>
a>
e\J
[S/EUJ] o
0)
£
•O
(U
3
00 •
J-i vD
Xi CT*
d !""*
QJ
4-> 00
O CM
O
- 0)
01
O
d r-
co CM
rH
W T3
J-i C
O CO
H
CM
.». i—I
CO
4J ^
d CT>
OJ
E s^
D D
W Xi
W D
CO -P
OJ A
E
•H 4)
S-4
« e
ft H
s o
O 4J
C_) W
co
CO
00
•H
-------
#
\£)
C
« 2 01
ill:!
O a? - S <»
— ± C/5 c ±
35 • 2 «
u CD 52 oo
r-l 2
O <
u
m
•> o
nj W
^ 4J
01 S
d M
c o
01 4-1
> to
0) CO
S 1=
01
Q -
5 o'
O •- (M f>
m
h.
oj
^-1
T*
Q)
£
c/5
£
^
0
Z
J
/
/
x
!=>
^ <
,
J
^^
^ ^T
_y~^^\
D
^
i :
-8
s
•
-
is i
c
"E
L±^J
0)
: E
;S H
-
•
•
CO
4-> iz;
d
01 «
S t-t
0) 0)
>-l 4J
d to
CO 01
(Q 43
O! U
S 0
OH
43
4-> - uo
•H f-\ F~
S « (Ti
4S rH
W
d
-------
(0
o>
»
S Q)
E
eo°, cq eaoofiQQOoSfiOQOQQ
l***!^**^ PSS3RSS g S S S
§ §
•[8/eui] D —-
-*-[u!Ujg/B>f|sSi — -[u!iu9/s»unoooloi.] Od
u
o
o
ON
in
00
[uiuig/uiiu] -
.? ;
III?!
' 2 S
" a) re
0-5 »
Ot- 2
•H i— I
CO
OJ
M a
4-1 ^
rt
0) T3
e a
OJ «
w ro
«
(U >>
S a
si
4-) m
•H o
3
w
W 4->
c d
O 0)
CO >
S o
O 4J
LO
ro
3
00
•H
67
-------
[ujtug/ uiuj] -
IO
N
CO
O)
.8 I S ° 8 § § §
-*-[S/BUJ] D •*- [UHUS/B>|] sQOa —-[U!
-S
-g 0)
K
-S
•
-s
s-i
•H D
M
(8 S
O. J.J
£ o
o j->
cj to
60
•H
[s/cui] 0 —-[miug/BM]
68
-------
The calibration process involved comparisons of runoff quantity meas-
urements at seven additional locations in the sewer network with QQS simula-
tions. These locations comprise areas of up to 672 ha in size including
drainage areas of different land uses and slopes that are mostly drained by
combined sewers. Although eight storms were originally selected for use in
the calibration runs, a number of four storms was tolerable for the final
calibration of the overall system. Flow measurements were not synchrone-
ously at all locations for all calibration storms.
Toronto Test Areas
In connection with a study involving the QQS model for the City of To-
ronto, Ontario, Canada, the measured runoff from two residential areas was
compared with simulations. The one area is 49 ha in size and was measured at
Jones Avenue; the second catchment comprises an area of 27 ha and was moni-
tored at Castlewood Avenue. Both areas are of average slope and have an im-
perviousness of 50 percent. For calibration purposes four storm events have
been used. At the time of this writing the pertinent study report containing
more details was under preparation.
General Results
As a whole, it is concluded that the QQS methodology of lumped catch-
ment simulation combined with detailed flow routing in the main and trunk
sewers is able to reproduce field measurements closely for individual
storms. It is then assumed that the model is sufficiently calibrated to be
employed for continuous simulations for planning purposes. For any appli-
cation, emphasis must be placed on the calibration and verification effort.
This emphasis, of course, applies to all mathematical model applications.
COMPARISONS WITH A MORE DETAILED SIMULATION MODEL
Aside from the question of validating the QQS model with field measure-
ments, it proved beneficial to compare QQS simulations with simulation re-
sults obtained with a more detailed simulation model. Such comparisons were
made in connection with the Augsburg, Germany, overflow abatement study.
Runoff quantities calculated by QQS, for a number of drainage areas of dif-
ferent land uses and other characteristics and for different locations in
the main sewer system, were compared with corresponding simulation results
previously obtained by detailed simulations with the HVM model. (The HVM-
program package comprises approximately 10,000 Fortran statements and incor-
porates mainly a surface runoff and a sewer network flow simulation model.
The surface runoff part transforms the time-dependent rainfall into time-de-
pendent surface runoff solving the continuity and energy equations. These
equations are also used for sewer1 network flow simulation of all sewer seg-
ments including all laterals. Due to the detail employed this method allows
for single event simulation only.) Figure 37 contains comparisons for two
drainage areas, one commercial and the other industrial. For nine different
combined sewer overflow structures, the inflow and outflow hydrographs ob-
tained by QQS simulations were compared with corresponding results from the
more detailed HVM model in order to prove that the QQS model yields suffi-
69
-------
[U|UIS/lUUl] '
X
C
(0
£
o
•*
o
'c
*. =. S 8 s 5 g o
o o o o d <3 d
-[8/ciu]0
[u|ius/uiiu]
o* s>; * «
o O O O Q
[uiuifi/iuiu]'
« o «o p >" °. "*. P
OOvf-iN^nn
•^ o «»
-* irt «
(O
N
O)
in
4)
(A
(A
(0
**
(O
w
0)
*-
TJ
0)
n
*<
tfl
"o
O)
c
c w
« 3
|.§
iso
CO
[•/«"*]
CO
a
o
•H
e
•H
(a
o
i-l
e
4-> O
a -H
OJ 4-1
CO -H
U W
C/3 S
o* >
O" W
4-1 4-1
o a
CD
w a
fl -H
O 4-1
10 ^1
•H (U
-H
S
c
CD
E
w.
0)
O
r^
O)
0)
3
O [uioig/uiui]-o»Jd—•-
II
Z
ra
(0
3
•o
c
cu
5-1
3
60
•H
——[«/cUJ]
70
-------
AUGSBURG
Location R006
QQS-Simulation
HVM-Simulation
(Storm N«1)
1.0 60 BO
Time [min]
too
120
KO
Figure 38. Comparison of QQS overflow simulations with HVM simula-
tions for a once per year recurrence synthetic storm.
E 2.0-
i—i
1.0 -
0.5-
0
0
AUGSBURG
Location R006
QQS-Simulation
HVM-Simulation
(Storm N-40)
20
60 BO
Time [m\n]
100
120
11(1
0 •£•
-o.i E
in
•0.2 "P
•0.3 J,
-».<. i
h.
-0.5 O-
06 f
Figure 39. Comparison of QQS overflow simulations with HVM simula-
tions for a 40 times per year recurrence synthetic storm.
71
-------
ciently reliable results for planning of larger networks. Two examples of
these comparisons are provided in Figures 38 and 39. These comparisons were
made using storms with recurrences of approximately once per year (N = 1)
and 40 times per year (N = 40).
For Munich, Germany, similar comparisons were made for different drainage
areas and several locations in their main sewer system using synthetic de-
sign storms. One of the runoff comparisons for drainage areas (Ingolstaed-
ter Strasse, N = 0.5) is shown on Figure 37. In addition, both QQS and HVM
simulations have been compared with recorded rainfall-runoff events, such as
in the examples given in Figure 37 for the Ingolstaedter Strasse (September
5, 1976) and Josef-Wirth-Weg (September 28, 1976) test areas.
The lumped and detailed simulations compare so well that, if there is
sufficient trust in the detailed method, the unit hydrographs needed for the
lumped method of the QQS drainage area runoff calculations may be derived
using detailed simulation results such as from the HVM. The writer has em-
ployed this principle for deriving unit hydrographs, particularly for per-
vious areas.
Lastly, it should be noted that detailed simulation of all sewers in
smaller drainage areas cannot be avoided if, for instance, final design of
all sewers including lateral and collection sewers is the objective or if
local flooding problems have to be studied.
72
-------
SECTION 5
APPLICATIONS
AUGSBURG, GERMANY
The City of Augsburg, Germany, with a projected population of 470,000,
is situated at the junction of the Lech and Wertach rivers. A master plan to
span a planning period of 32 years was undertaken to determine the sequenc-
ing and timing of overflow abatement measures that should be taken to re-
lieve the partially overloaded combined sewer network taking into considera-
tion the future development of the city and its suburbs. At the same time,
the heavy pollution of the Lech and Wertach rivers and of city creeks caused
by combined sewer overflows had to be minimized and kept within set limits.
The master plan was developed on the basis of QQS and HVM calculations.
Figure 40 shows the main and trunk sewer and receiving water systems, as de-
fined for the QQS calculations. The study area consisted of 5,800 ha. First,
QQS simulation results were utilized to estimate roughly the effect of dif-
ferent abatement alternatives. These measures were finalized using detailed
HVM calculations. The final overflow abatement concept was investigated in
terms of its impact on receiving water quality using QQS simulations. For ex-
pected conditions in the year 2010, continuous simulations were made using
four years out of the last 15 years of rainfall records. Continuous simula-
tion results were statistically analyzed for such variables as overflow du-
ration, total, average and peak overflow volumes, BOD5, and settleable so-
lids.
The simulated overflows were then added to the corresponding background
flows of the receiving water system, which includes the Lech and Wertach
rivers and city creeks. Continuous simulation results of the runoff and pol-
lution loads were statistically analyzed mainly for the increase of BOD5 and
settleable solids concentrations and their durations. This analysis was done
for key points such as the junction of the Lech and Wertach rivers and down-
stream of the city area, where the greatest total impact of the city of
Augsburg occurs on receiving waters.
The overall intent in developing the overflow abatement measures was to
meet water quality classification "II" in the Lech river downstream from the
city of Augsburg, which among other criteria allows for BOD5 levels of 2 to
6 mg per 1. No regulations, however, were available that defined the number
and magnitude of violations acceptable per year. Rigid limits specified in
the guidelines could not be met at all times, as is the case for any hydro-
logical process. This uncertainty in criteria was overcome by analyzing the
statistical results of the continuous QQS simulations in such a way that
violations of the target classification "II" with its allowable BOD5 level
73
-------
CITY OF AUGSBURG
mom and trunk sewer system
™ • receiving water system
9 combined sewer overflow
O basin
p basm with overflow
O pumping station
O sewage treatment plant
Figure 40. Main and trunk sewer and receiving water systems of
Augsburg, Germany, as defined for QQS simulations.
74
-------
of 6 mg per 1 could be interpreted in terms of average receiving water back-
ground BOD5 levels. Figure 41 illustrates the total impact of the city's
overflows on receiving waters, both including and excluding treatment plant
bypasses and effluents. For example, one can read from the graphs in Figure
41 that for a background BOD5 level of 2 mg per 1, the target receiving wa-
ter classification is contravened 14 times per year when treatment plant by-
passes and effluents are taken into the consideration, but only approximate-
ly three times per year when treatment plant bypasses are excluded. The
treatment level was fixed in this case.
Another guideline required that an average of 90 percent of the annual
BOD5 and settleable solids loads running off during storm events should be
led to the treatment plant for processing. This provision was investigated
using the average annual totals calculated by the continuous simulation.
The overflow abatement schemes suggested in this study called for main-
ly inline storage. Storage would be in detention channels and basins, which
would be bypassed and emptied by pumping after major storm runoff occurren-
ces, and utilized via branchings and cutoffs. This use of storage would
guarantee a more equal usage of the overall storage volume of the sewer net-
work. The study results also showed that additional increases in storage
volumes within the sewer network would not yield further significant im-
provements in receiving water quality that would justify the additional
costs.
MUNICH, GERMANY
The City of Munich, Germany, is drained mainly by a combined sewer sys-
tem. Dry-weather flow and combined sewer runoff are mechanically and biolo-
gically treated at present in one treatment plant that discharges into a
channelized branch of the Isar river which discharges into fish ponds.
The receiving water system of the combined sewer overflows is the Isar
river with its connected creeks and channels. There are 33 overflow points
that discharge into this receiving water system, some of which have over-
flows shortly after dry-weather flow is exceeded. Some of the overflow
points are located at a part of the river that is dry or that carries only
minor flows resulting from ground water infiltrations and lateral inflows
for almost 300 days of the year.
The main purpose of the overflow abatement study for the City of Munich
was to guarantee, in the receiving waters downstream from the urbanized area
of Munich, the maintenance of water quality classification "II", mentioned
above in connection with the City of Augsburg. To satisfy this standard,
overflow abatement measures, that would reduce the frequency and duration of
overflows, had to be developed.
The project was divided into four phases: data collection and prepara-
tion, assessment of existing conditions, development of overflow abatement
measures, and presentation of evidence that the suggested measures would sa-
tisfy the water quality standards. The pollutant parameters investigated
75
-------
\
c
O
U -H
> <0
•H i— I
OS 3
u w
0)
hJ W
o-
Q «
O O
05 -H
4-1
O V]
(0 a)
t .
•H >>
O HI
H a
vr ro CN
[l/6ai]
ui
76
-------
were BOD5 and total suspended solids. The study area for existing condi-
tions has a size of approximately 13,600 ha and 1.56 million inhabitants are
connected to the present sewer network. Simulation for future conditions
were based on an area of 20,800 ha and 3.17 million inhabitants. This in-
crease is due to additional future connections of suburban areas to the
Munich sewer system. This future system also includes a second treatment
plant the location of which already was predetermined. The trunk sewer and
receiving water system used for the QQS simulations is shown in Figure 42.
The computer program had to be adjusted to accommodate for the study area
size. The QQS model was calibrated on a variety of flow and pollution mea-
surements from four test catchments and the treatment plant intake.
For simulation purposes, the rainfall record of three meteorological
stations located within the study area were used. Thus, the temporal and
spatial variations of rainfalls were considered to some extent. Continuous
simulations with the QQS program have been performed on the basis of four
representative years selected from the continuous records of 20 years from
each station.
From continuous simulation results of existing conditions the present
impact of combined sewer overflows on the receiving waters was defined. In-
vestigating the calculated differences in the individual overflow frequen-
cies and loadings, overflow abatement alternatives were developed, such as
adjustment of overflow weir heights, cutoffs, additional inline and offline
storage. The abatement scheme selected is a combination of above measures
and includes an additional offline storage volume on the order of 250,000 m3
and an extra inline storage of approximately 300,000 m3. For this scheme the
required criteria could be met.
Especially during the phase of development of overflow abatement meas-
ures QQS simulations were found to quickly indicate the effectiveness of in-
dividual measures with respect to receiving water loadings. The detailed
layout of the abatement scheme selected was finalized using the hydraulic
simulation package HVM.
ROCHESTER, N.Y.
The combined sewer system of the City of Rochester, N.Y., with a popu-
lation of 320,000 and an area of 5,300 ha, discharges pollutants into the
Genesee River. This pollution load causes a significant depression in dis-
solved oxygen levels and increases fecal coliform indicator organisms. With-
in the Rochester Pure Water District, combined sewer overflow pollution
abatement program, application of mathematical models has helped in the as-
sessment of present conditions and in the definition of an abatement scheme
which would comply with New York State Department of Environmental Conserva-
tion stream classification "B", which is aimed at the best usage for the re-
source, contact recreational use, and all other uses, except as a source of
potable water.
77
-------
E
Or
*« I
_ s s s
X I -5 s
o s s-
9 e
? o
u_
o
>
H;
O
r I
- = a
isg
t I j a D O
O
•H
Ci
S
4-1
O
+J
en
OJ c!
•P O
« -H
S 4J
BJ
00 r-l
C 3
•H S
> -H
•H W
>
« a
«
a s
•H h
03 CJ
S O
CN
(LI
i-l
&0
•H
78
-------
The computer program was adjusted to accommodate for the size of the
study area and was calibrated using local measurements. Continuous simula-
tions of the runoff and overflow behavior of the main sewer network of the
City of Rochester (Figure 43) were made for 8 years selected from a rainfall
record spanning 15 years. The overflow events simulated with the QQS model
were combined with recorded river flows and an average annual number of con-
traventions of the classification "B" was derived for each added storage and
treatment combination investigated.
The model output was also used to determine the total BOD5 loads pre-
sently overflowing into the Genesee River on an average annual basis, and
the reductions of annual BOD5 loads to be expected from storage facilities
of various sizes. Total suspended solids and fecal coliforms were also in-
vestigated.
A cost-effective system with a storage volume on the order of 300,000
m3 was found to comply with the stream standards. The results of this study
have been presented previously (5 and 16) and are contained in a study re-
port (9).
VANCOUVER, B.C., CANADA
The main objective of a study using QQS for an area of the City of Van-
couver, B.C., Canada, facing the shores of English Bay and False Creek, was
recreational use with full body contact while keeping bathing beaches open
for a maximum number of days in the season. There are two criteria which had
to be met. One, the Rawn Report, published in 1953, provides criteria for
design based on overflow events and is basically addressed to the mainte-
nance of the recreational waters in English Bay. It is advisory only and is
not legally binding. However, the Pollution Control Branch of the Government
of British Columbia, established in 1967, enforce guidelines for combined
sewer operation which apply to continuous, not seasonal discharges. To sa-
tisfy their requirements, data for the annual performance of the sewer im-
provements to be developed were required. In addition seasonal information
to justify any of the recommendations to the City's administration were de-
sired.
The study area has a size of approximately 3,000 ha and 110,000 inhab-
itants are connected to~ the storm and combined sewer system of the area un-
der investigation. The network of the study area contains almost all fea-
tures found in sewer systems, such as branches, outfalls, overflows, deten-
tion basins, pumping stations, time-dependent and water surface-dependent
operational control gates and syphons. Figure 44 shows a schematic of the
main sewer system of the study area. The complexity of the network, the sub-
stantial external loadings that enter the study area, and backwater effects
from reaches downstream of the study area, called for some adjustments of
the computer program and an extensive calibration effort for this QQS ap-
plication. Observations were available at seven locations within the network
and at two sewer system overflow points. Although eight storms were origin-
ally selected for use in the calibration runs, a number of four storms was
tolerable for the final verification of the overall system. Hereby flow
79
-------
CITY OF ROCHESTER
main storm and combined
sewer network
Genesee River
storm ourfall or combined
sewer overflow
sewage treatment plant
Figure 43. Schematic of the main and trunk sewer
system of Rochester, N.Y.
80
-------
T3
d
u
u
pq
s-i
OJ
O
u
d
o
s
-------
measurements were not simultaneously available at all locations for all
calibration storms.
After the successful calibration the impact of the existing system on
the receiving waters was assessed. This assessment was derived from contin-
uous rainfall-runoff simulations with the QQS program of five summer periods
(May to September) and of three winter periods (October to April). The peri-
ods were selected from 18 years of rainfall records. The parameters BOD5,
total suspended solids, and fecal coliforms were investigated.
On the basis of this problem identification a number of possible alter-
native measures to abate the combined sewer overflow problem was investi-
gated. These measures included upgrading of the inflow to the treatment
plant, changes in gate settings, additional storages and flow diversions.
The preferred alternative which represents a combination of these measures
was analyzed for its effectiveness again by continuous simulations. The con-
tinuous simulation for this future condition was done for the same five sum-
mer periods as used for the existing system assessment. However, in order to
check if above criteria were met it was acceptable to only investigate the
winter period of rainfalls which had effected the worst impact on receiving
waters under existing conditions.
TORONTO, ONTARIO, CANADA
A QQS study has been carried out for the City of Toronto, Ontario, Can-
ada, located on the shores of Lake Ontario. The storm and combined sewer
system studied serves an area of about 6,630 ha with 536,000 inhabitants
(Figure 45).
As a relief measure for two interceptor sewers constructed more than 60
years ago, construction of a new interceptor, the "Mid Toronto Interceptor",
and three interconnecting sewers linking the three interceptor sewers to-
gether was recently completed. The maximum conveyance capacity of the new
interceptor together with present inflows of all existing inflows to the
treatment plant during storms would exceed the capacity of the plant. Ac-
cordingly remote controlled, power operated gates in the interconnecting
sewers and conventional fixed gates for 24 inflow connection to the new in-
terceptor are used to limit the inflow to meet available capacities at the
treatment plant.
Objectives of the study were to assess the impact of the operational
behavior of the Interceptor and its effect on the City's sewer network and
the treatment plant. In addition annual peak and total quantity and quality
discharges to the Lake, the Harbour and the Don River originating from the
combined sewer system and the road storm sewer system connected to it were
to be computed for the chosen operational scheme. For the selected pumping
configuration maximum and annual total flows to the Treatment Plan were al-
so determined.
A family of four representative recorded rainfall events, was used to
examine the behavior of the Interceptor system under various gate settings.
82
-------
5S
II
•
<* 5
o <
H O
o. z
111 2
0 2
a: m
UJ
z S
_ -
2 uj UJ
O > >
z uj uj
a
r /
0
II- -«ll
^iC =
S
1
1
v>
A
p
bs
\
m
>«•
o
$
S
I
*
^
o
o
+J
fi
o
i-l
o •
H W
fi
4-1 O
O -H
4J
S rt
(U tH
4-> d
M S
>> -H
W W
(-1 M
-------
A continuous simulation for four selected years of rainfall records was car-
ried out for one selected configuration of gate settings.
For calibration of the QQS model measured runoff from two residential
catchment areas and several storm events were available. In addition de-
tailed HVM simulation results for model storms of one year and two year re-
turn frequency and several measured rainfall-runoff events were used for
comparison with QQS simulation results. The QQS program had been adjusted to
accommodate for the study area size.
In the future it is intended to use the overflowing loads of BOD5 and
settleable solids calculated by the QQS program as input for another purpose
made mathematical model which assesses and predicts the pollutant load dis-
tribution within the harbour area of Toronto.
84
-------
REFERENCES
1. Brandstetter, A. Assessment of Mathematical Models for Storm and Com-
bined Sewer Management. EPA-600/2-76-175a. U.S. Environmental Protec-
tion Agency. Cincinnati, Ohio, 1976. 509 pp.
2. Brunner, P. G. Die Verschmutzung des Regenwasserabflusses im Trennver-
fahren, Untersuchung unter besonderer Beruecksichtigung der Nieder-
schlagsverhaeltnisse im voralpinen Raum. (The Pollution of Storm Water
Runoff in Separate Systems: Studies with Special Reference to Precipi-
tation Conditions in the Lower Alpine Region.) Ph.D. Thesis. Technical
University of Munich, Munich, Germany, 1975. 200 pp.
3. Colyer, P. J. , and R. W. Pethick. Storm Drainage Design Methods. Re-
port No. IHT 154. Hydraulics Research Station, Wallingford, Oxon, Eng-
land, 1977. 85 pp.
4. Geiger, W. F. Urban Runoff Pollution Derived from Long-Time Simula-
tion. In: Proceedings of the National Symposium on Urban Hydrology and
Sediment Control, Lexington, Kentucky, 1975. pp. 259-270.
5. Geiger, W. F., S. A. LaBella and G. C. McDonald. Overflow Abatement
Alternatives Selected by Combining Continuous and Single Event Simula-
tions. In: Proceedings of the National Symposium on Urban Hydrology,
Hydraulics and Sediment Control, Lexington, Kentucky, 1976. pp. 71-79.
6. Haendel, H. Vergleichende Untersuchungen von elektronischen Kanalnetz-
berechnungen. (Comparative Evaluations of Computerized Sewer Calcula-
tion Methods .) Paul Parey Press, Hamburg, Germany, 1977. 134pp.
7. HEC1. Flood Hydrograph Package. Users Manual. The Hydrologic Engi-
neering Center, Davis, California, 1973. 25 pp.
8. Keeps, 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.
9. Joint Venture (Erdmann Anthony Associates, Lozier Engineers Inc. and
Steelye Stevenson Value & Knecht Inc.). Overflow Quantity-Quality Anal-
ysis. Volume III of Wastewater Facilities Plan, Combined Sewer Overflow
Abatement Program, Rochester Pure Waters District, Monroe County, New
York, 1976. 64 pp.
85
-------
10. Klym, H. , W. Koeniger, F. Meviu's, and G. Vogel. Urban Hydrological Pro-
cesses Computer Simulation. Paper presented at the Seminar on Computer
Methods in Hydraulics, Swiss Federal Institute of Technology, Zurich,
1972. 39 pp.
11. Krauth, K. Der Abfluss und die Verschmutzung des Abflusses in Misch-
wasserkanalisation bei Regen. (Wet-Weather Runoff and its Pollution in
Combined Sewer Systems.) Ph.D. Thesis, Technical University of Stutt-
gart, R. Oldenburg Press, Munich, 1970. 251 pp.
12. MacLaren, J. F. Review of Canadian Storm Sewer Design Practice and Com-
parison of Urban Hydrologic Models. Project No. 74 - 8 - 31. Environ-
ment Canada, Ottawa, Ontario, 1975. 212 pp.
13. Malmquist, P-A., and G. Svensson. Urban Stormwater Quality. Interim
Report. Urban Geohydrology Research Group, Chalmers University of Tech-
nology. Report No. 13, 1975.
14. McPherson, M. B. Utility of Urban Runoff Modeling. ASCE Urban Water
Resources Research Program. Technical Memorandum No. 31, 1976. 126 pp.
15. Pfeiff, S. Meteorologische, topographische und bautechnische Einflues-
se auf den Regenabfluss in Kanalisationsnetzen. (Meteorological, Topo-
graphical and Constructural Influences on Rainfall Runoff in Sewer Net-
works.) Ph.D. Thesis, Technical University of Darmstadt, Erich Schmidt
Verlag, 1971. 340 pp.
16. Shapiro, H. M. The Use of Modeling in Studying the Rochester Combined
Sewer System. SWMM Users Group Meeting, Toronto, 1976.
17. Svensson, G. , and P-A. Malmquist. Water Budget for a Housing Area in
Gothenburg. Proceedings of the Amsterdam Symposium, IAHS-AISH. Publi-
cation No. 123, October, 1977. pp. 101-108.
18. Svensson, G. , and K. 0ren. Planning Models for Sewer Networks. Urban
Geohydrology Research Group, Chalmers University of Technology. Report
No. 36, 1978.
86
------- |