EPA-R2-72-068
October 1972 Environmental Protection Technology Series
Storm Sewer Design
An Evaluation of the RRL Method
o
Office of Research and Monitoring
U.S. Environmental Protection Agency
Washington, D.C. 20460
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and
Monitoring, Environmental Protection Agency, have
been grouped into five series. These five broad
categories were established to facilitate further
development and application of environmental
technology. Elimination of traditional grouping
was consciously planned to foster technology
transfer and a maximum interface in related
fields. The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
U. Environmental l^onitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ENVIRONMENTAL
PROTECTION TECHNOLOGY series. This series
describes research performed to develop and
demonstrate instrumentation, equipment and
methodology to repair or prevent environmental
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.
-------
EPA-R2-72-068
October 1972
STORM SEWER DESIGN - AN EVALUATION OF THE RRL METHOD
By
John B. Stall and Michael L. Terstriep
Project 11030 FLN
Project Officer
Harry C. Torno
Office of Research and Monitoring
Environmental Protection Agency
Washington, D.C. 20^-60
Prepared for
OFFICE OF RESEARCH AND MONITORING
U.S. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DoCo 20k60
For sale by the Superintendent of Documents, U.S. Government Printing Office, Washington, D.C. 20402 - Price $1.25
-------
EPA Rev few Notice
This report has been reviewed by the Environ-
mental Protection Agency and approved for
publication. Approval does not signify that
the contents necessarily reflect the views and
policies of the Environmental Protection
Agency, nor does mention of trade names or
commercial products constitute endorsement
or recommendation for use.
-------
ABSTRACT
Storm rainfall and runoff data were assembled from 10 urban basins in the U.S.A. ranging
in size from 14 acres to 8 sq mi. The British RRL method of storm drainage design was
applied to the 10 basins. The RRL method considers the urban basin to be comprised of
the paved area of the basin which is directly connected to the artificial storm drainage
system. In 3 of the 10 basins the RRL procedure was deemed to be appropriate and suit-
able for the design of a storm drainage system within the normal range of frequency of
design rainfall events, from 2 to 20-year events. For greater storms and for certain cases
within this frequency range, the RRL method breaks down because the runoff coming
from the grassed area of the basin is significant. If the basin is highly steep or if the paved
area comprises less than 15% of the total basin, this breakdown occurs. An example is
given of the use of the RRL method in the re-design of an existing storm drainage system,
as is current practice in Great Britain.
This report is submitted in fulfillment of Project 11030 FLN under the partial sponsorship
of the U.S. Environmental Protection Agency.
in
-------
CONTENTS
Section^ Page
I Conclusions 1
II Recommendations 3
III Introduction 5
Need 5
Objective 6
IV Assembly of Data 7
Criteria 7
Methods 8
V The Road Research Laboratory Method 11
Origin 11
The Procedure 11
Computer Program 15
Application of the RRL Method 15
Layout for Presentation of Results 18
VI Description and Results by Basins 19
Woodoak Drive Basin, Westbury, Long Island, New York 19
Ross-Ade (Upper) Basin, West Lafayette, Indiana 23
Sewer District No. 8 Basin, Bucyrus, Ohio 27
Echo Park Avenue Basin, Los Angeles, California 31
Crane Creek Basin, Jackson, Mississippi 35
Tripps Run Tributary Basin, Falls Church, Virginia 39
Tar Branch Basin, Winston-Salem, North Carolina 43
Third Fork Basin, Durham, North Carolina 47
Dry Creek Basin, Wichita, Kansas 51
Wingohocking Basin, Philadelphia, Pennsylvania 55
VII Overall Results 59
General 59
Rainfall Intensity in Britain and America 60
Other Urban Hydrologic Models 61
VIII Re-Design of a Storm Drainage System 63
IX Acknowledgments 69
X References 71
XI Glossary of Terms 73
-------
FIGURES
Figure Page
1 Two-hour storm rainfall in inches expected at a 5-year recurrence
interval for mainland United States 7
2 Elements in the development of the hydrograph 13
3 Elements in the storage routing technique 14
4 Flow diagram for the computer program of the RRL method 16
5 Input format for RRL method for all basin parameters, storm
rainfall, and storm runoff 17
6 Outlet of the single storm drain in the Woodoak Drive basin 20
7 RRL results for Woodoak Drive basin, Long Island, New York 21
8 Aerial photo of Woodoak Drive basin 22
9 Street view in Ross-Ade (Upper) basin 24
10 RRL results for Ross-Ade (Upper) basin, West Lafayette, Indiana 25
11 Aerial photo of Ross-Ade (Upper) basin 26
12 Flat terrain and indeterminate drainage pattern typical of much of
the upstream basin of Sewer District No. 8 28
13 RRL results for Sewer District No. 8 basin, Bucyrus, Ohio 29
14 Aerial photo of Sewer District No. 8 30
15 Street view in Echo Park Avenue basin 32
16 RRL results for Echo Park Avenue basin, Los Angeles, California 3 3
17 Aerial photo of Echo Park Avenue basin 34
18 Ditch along street in Crane Creek basin 36
19 RRL results for Crane Creek basin, Jackson, Mississippi 37
20 Aerial photo of Crane Creek basin 38
21 Typical large lawns in the downstream portion of Tripps Run
Tributary basin 40
22 RRL results for Tripps Run Tributary basin, Falls Church, Virginia 41
23 Aerial photo of Tripps Run Tributary basin 42
24 Typical houses and street in the Tar Branch basin 44
25 RRL results for Tar Branch basin, Winston-Salem, North Carolina 45
26 Aerial photo of Tar Branch basin 46
27 Open channel which is representative of most of the drainage in
Third Fork basin 48
28 RRL results for Third Fork basin, Durham, North Carolina 49
29 Aerial photo of Third Fork basin 50
30 View of street typical of those which carry most surface runoff
in the Dry Creek basin 52
31 RRL results for Dry Creek basin, Wichita, Kansas 53
32 Aerial photo of Dry Creek basin 54
33 Typical row-houses in the Wingohocking basin 56
VI
-------
Figure Page
34 RRL results for Wingohocking basin, Philadelphia, Pennsylvania 57
35 Street map of the Wingohocking basin 58
36 Frequency curves of 2-hour point rainfall in Great Britain and
for four locations in the United States 60
37 Time distribution of storm rainfall, median curve for point rainfall 63
38 Reach identifications for a portion of the Tar Branch basin,
Winston-Salem 66
vn
-------
TABLES
Table Page
1 Storm data and results for Woodoak Drive basin 20
2 Storm data and results for Ross-Ade (Upper) basin 24
3 Storm data and results for Sewer District No. 8 basin 28
4 Storm data and results for Echo Park Avenue basin 32
5 Storm data and results for Crane Creek basin 36
6 Storm data and results for Tripps Run Tributary basin 40
7 Storm data and results for Tar Branch basin 44
8 Storm data and results for Third Fork basin 48
9 Storm data and results for Dry Creek basin 52
10 Storm data and results for Wingohocking basin 56
11 General results of RRL method for all basins 59
12 The 5-year 2-hour design storm for Woodoak Drive basin,
Long Island, and the Tar Branch basin, Winston-Salem 64
13 Results of the RRL method re-design of the storm drainage
system for Tar Branch basin 65
vin
-------
SECTION I
CONCLUSIONS
1. The RRL method provides an accurate means of computing runoff from the paved area
portion of an urban basin.
2. The RRL method adequately represents the runoff from actual urban basins under the
following conditions:
a. the basin area is less than 5 sq mi,
b. the directly connected paved area is equal to at least 15 percent of the basin area, and
c. the frequency of the storm event being considered is not greater than 20 years.
3. The RRL method cannot be used for all urban basins in the United States; the method
breaks down when significant grassed-area runoff occurs, which happens if one or more of the
following conditions exist:
a. the directly connected paved area is less than 15 percent of the basin area,
b. the frequency of the event being considered is greater than 20 years,
c. the grassed area of the basin has steep slopes and tight soils, regardless of the anteced-
ent moisture condition,
d. the grassed area of the basin has steep slopes, moderately tight soils, and an anteced-
ent moisture condition of 3 or 4, and
e. the grassed area of the basin has moderate slopes, moderately tight soils, and an ante-
cedent moisture condition of 4.
4. The principal strength of the RRL method is that, by confining runoff calculations to
the paved area of a basin, it utilizes hydraulic functions which are largely determinate such as
gravity flow from plain sloping concrete surfaces, gutters, pipes, and open channels. Physical
understanding of these flow phenomena is much greater than the present understanding of
the many complex phenomena governing runoff from rural areas such as antecedent moisture
conditions, infiltration, soil moisture movement, transpiration, evaporation, etc.
5. A modification of the RRL method that would provide a function for grassed-area con-
tributions to runoff could be developed into a valuable design tool for urban drainage. This
is believed to be possible in spite of the many complexities involved. Further flexibility
could be offered by the additional provision for routing surface runoff through surface stor-
age.
6. The input data requirements for use of the RRL method on an urban basin are reasonable
for the engineering evaluation of a basin for storm drainage design. The necessary data are no
more complex nor elaborate than the data usually compiled for a traditional storm drainage
design.
7. It appears that rainfall occurs in greater amounts in the United States than in Great Brit-
ain. This may account for the fact that the RRL method is successful and widely used in
Great Britain and yet suffers the above-described breakdowns for some of the basins studied
in the United States.
8. Better urban rainfall and runoff data are required for proper testing of all mathematical
models. Research basins that do not have hydraulic problems, such as undersized drains or
inadequate inlets, should be selected and instrumented.
-------
SECTION II
RECOMMENDATIONS
1. It is recommended that research and development be carried out to incorporate within
the RRL method a series of parameters and functions which would accommodate the runoff
from grassed areas within an urban basin and to provide other minor improvements in the RRL
method to provide greater flexibility in its application.
2. It is recommended that, until the above improvements are achieved, the method not be
promoted as having general applicability for widespread-use. This is because the method
breaks down for basins where the grassed-area contribution is significant, as shown in this re-
search. If the method can be suitably altered to accommodate grassed-area runoff, then it
would be appropriate to carry out further developmental work on the method in order to
streamline it as a design procedure.
3. Immediate efforts should be made to establish stations and instrumentation for the col-
lection of higher quality data on storm rainfall and runoff from urban basins. This is neces-
sary for the adequate testing and calibration of many of the modern existing methods for
accommodating urban storm rainfall and runoff. This instrumentation should probably also
include water quality parameters.
4. It is recommended that further use and applications be made of total-basin-accounting
models published and used elsewhere, such as the hydrologic simulation programming, HSP.
The use of such a model in the urban setting offers promise of contributing understanding of
all the many complex and critical hydrologic processes in an urban basin.
-------
SECTION III
INTRODUCTION
Need
$2.5 billion per year is the construction cost for storm drainage systems in urban areas of the
United States estimated by the American Public Works Association (1966). This monumental
expenditure represents the amount that city dwellers pay in order that storm runoff water can
be adequately collected and removed from the rooftops and streets of an urban area and emp-
tied into a convenient natural stream outside of the city limits.
When storm rainfall occurs in a rural area, much of it soaks into the earth; the remainder runs
off to the nearest stream. The excess surface runoff may cause some temporary flooding on
the land surface along ditches, drainageways, and small stream channels. When a city is con-
structed, much of the natural landscape is covered with rooftops, paved streets, and other
paved areas. The remaining natural earth is usually covered with grass lawns. Several research-
ers have shown the effects of urbanization on the storm runoff of a region. Stall and others
(1970) showed that the complete transformation of a 3.5 sq mi rural basin in east-central
Illinois to an intensely urbanized basin would increase the flood peak by about 4 times for the
50-year recurrence interval. It would increase the mean annual flood by about 8 times.
An artificial storm drainage system for an urban area usually includes a collection network of
storm drains consisting of underground conduits. Engineering design practice in 1972 utilizes
almost exclusively the rational method for determining the size and hydraulic capacity of
these storm drainage systems. Design practice in 32 cities has been summarized by Ardis and
others (1969). The rational method is described in most hydrologic text books and is given
by Chow (1964) as being Q = CIA where Q is the peak discharge in cfs, C is a runoff coeffi-
cient depending on the characteristics to the drainage basin, I is the rainfall intensity in inches
per hour, and A is the drainage area in acres. The term rational is used because the units of
the quantities are numerically consistent. The method has widespread acceptance but its use
still relies heavily on engineering judgment.
Practicing engineers have recognized the need for an improved method for understanding the
storm rainfall-runoff process in urban areas. The American Society of Civil Engineers (1969)
gave an extremely high priority to the need for better knowledge of the rainfall-runoff-quality
process in urban drainage systems. Under this impetus a number of different models have been
developed in recent years for accommodating the storm rainfall-runoff process for an urban
region. A critical review of about 12 of these models has been provided by Linsley (1971),
who states as one conclusion:
"The present limited amount of urban hydrologic data is a serious deterrent to develop-
ment and testing of storm runoff models. It seems unlikely that any significant improve-
ment in current models is possible until more data and better quality data are available."
The Illinois State Water Survey in recent years carried out a major evaluation of the method
of storm drainage design developed at the British Road Research Laboratory (called the RRL
method) which is described in a later section of this report. Terstriep and Stall (1969) used
the RRL method to carry out an analysis of 39 storms on three urban watersheds in the
United States and illustrated the general applicability of the method for conditions in this
country. It was considered desirable to have additional cases for testing this method. Explo-
ration revealed that some data on storm rainfall and runoff existed for urban areas of the
United States. A grant was provided by the federal Environmental Protection Agency to
-------
finance part of the cost of the collection of the necessary physical information and rainfall-
runoff data and the analysis of these data by the RRL method for about 10 basins in the
United States. This activity is described in this report.
All of the basin data, storm drain data, maps, and storm data assembled for this project have
been provided to the EPA. Persons interested in obtaining copies of these data for research
purposes should contact Harry C. Torno, Staff Engineer, Municipal Pollution Control Section,
U.S. Environmental Protection Agency, Washington, D.C. 20460.
Objective
The object of this project has been to provide a catalog of actual applications on about 10
urban basins in the United States of the British Road Research Laboratory method for the
design or re-design of storm drainage systems. This included an evaluation of the applicability
of this method for United States conditions. Because the RRL method considers runoff from
paved areas only, a secondary objective was to evaluate the influence of grassed-area runoff
from typical urban watersheds.
-------
SECTION IV
ASSEMBLY OF DATA
Criteria
Relatively few data exist for storm rainfall and storm runoff measurement from urban drain-
age basins. The American Society of Civil Engineers recently carried out a study to locate data
of this type. As reported by Tucker (1969) there do exist in some cities and other govern-
mental offices throughout the United States some data which seem suitable for hydrologic
analysis. The U.S. Geological Survey has emphasized in recent years the measurement of run-
off from urban basins. A number of USGS offices were contacted regarding data for this proj-
ect. Of the 10 basins ultimately studied and reported herein, data for 6 were obtained from
the USGS, usually collected in cooperation with some other agency. In the usual case these
urban runoff data had been collected and used for some particular project, but had not been
published. A major study by the USGS on the effects of urbanization on floods has been pub-
lished by Anderson (1970).
In considering basins for this project, priority was given to the following items: 1) basins less
than 5 sq mi in size, 2) basins that were intensely urbanized, 3) basins that had extensive
storm drainage systems, 4) basins with a high amount of paved area, 5) long records of rain-
fall and runoff, 6) the degree of quality of the data on storm rainfall and runoff, 7) the degree
of information available on the storm drainage system, and 8) data which had not already been
published in one form or another.
WOODOAK
WINGOHOCKING
2.5
TAR BRANCHT-1".1^.^
Figure 1. Two-hour storm rainfall in inches expected at a 5-year recurrence interval
for mainland United States (U.S. Weather Bureau, 1961)
-------
Another major item in the selection of basins was the effort to provide coverage of a variety
of hydrologic regimes in the United States. In order to do this, consideration was given to the
variability of storm rainfall within the United States. Figure 1 is a map of the 2-hour storm
rainfall in inches expected at a 5-year recurrence interval for mainland United States, as shown
by the U.S. Weather Bureau (1961). Although some other combination of duration and fre-
quency could be considered as critical for storm drainage design, it is felt that the 2-hour 5-
year storm is one which is certainly in the critical zone as far as storm drainage design is con-
cerned.
The following 10 basins were selected and used: Woodoak Drive basin, Westbury, Long Island,
New York; Ross-Ade (Upper) basin, West Lafayette, Indiana; Sewer District No. 8, Bucyrus,
Ohio; Echo Park Avenue basin, Los Angeles, California; Crane Creek basin, Jackson, Missis-
sippi; Tripps Run Tributary basin, Falls Church, Virginia; Tar Branch basin, Winston-Salem,
North Carolina; Third Fork basin, Durham, North Carolina; Dry Creek basin, Wichita, Kansas;
and Wingohocking basin, Philadelphia, Pennsylvania.
Methods
In order to apply the RRL method to a basin it was necessary to obtain street maps, aerial
photographs, and the location, size, and slope of the existing storm drainage system. Normally
this information was obtained from city officials. In scouting around to locate data suitable
for this project, a great number of data collection agencies and cities were contacted. Many
leads were obtained from the publication by Tucker (1969). For the more promising basins
a letter was written to the data collection agency and to the city outlining the needs of the
project, and asking for a description of the exact nature of the data.
The second step was to examine samples of the raingage and streamgage charts to determine
the general quality of the data and to determine the minimum time intervals of rainfall and
runoff which could be interpreted from the charts. If the data were suitable for the project,
arrangements were made to borrow or copy either the charts or any rainfall and runoff data
that had already been reduced for the shortest feasible time interval. A detailed inspection
trip was made through the basin, and notes were made of the exact drainage pattern. Obser-
vations were made of the practice in the disposition of runoff from rooftops and notes were
made as to the number or percent of rooftops which emptied onto the driveway or were
directly connected to the storm drainage system or to the street by tile drain. Further inspec-
tion and notes were made of existing open channels or drainage ditches, noting the general
conditions, shapes, and configurations. An inspection was usually made of the stream gaging
installation and the raingages, to note their general condition, exposure, and setting. Photo-
graphs were taken to aid in further interpretation of the field notes at the office.
In several of the basins a considerable amount of work was entailed in the reading of the rain-
gage charts and recorder charts of flow. This was accomplished as a part of the present project
and considerable use was made of the automatic graphical digitizing machine, the model 3400
auto-trol, which was available at the Illinois State Water Survey.
An aerial photograph of each basin was obtained. The principal source of these photos was
the U.S. Department of Agriculture, Agricultural Stabilization and Conservation Service
(ASCS). These aerial photos were used in determining the total amount of the paved area in
the basin.
Information has been assembled for every basin to characterize the predominant surface soils.
This information was obtained from the federal Soil Conservation Service and followed their
-------
designations as hydrologic soil groups:
A. Have low runoff potential, high infiltration rates, consist of sand and gravel,
B. Have moderate infiltration rates and are moderately well-drained,
C. Have slow infiltration rates; have a layer which impedes the downward movement
of water, and
D. Have high runoff potential, very slow infiltration rates, clays with a permanent high
water table and a high swelling potential.
The antecedent moisture condition used later in this report was determined from Weather
Bureau (National Weather Service) Climatological Data. The gage selected for each basin was
the nearest Weather Bureau recording gage to the basin. Antecedent moisture conditions were
classed in four divisions based on the 5-day antecedent rainfall, as follows: 1, bone dry, 0 rain-
fall; 2, rather dry, 0 to 0.5 inch; 3, rather wet, 0.5 to 1.0 inch; and 4, wet, over 1.0 inch.
-------
SECTION V
THE ROAD RESEARCH LABORATORY METHOD
Origin
The conception and development of the British RRL method for the accommodation of storm
rainfall and runoff for the design and re-design of storm drainage systems has been the result
of an elaborate, varied, and sustained research program. The Road Research Laboratory is an
agency of the Ministry of Transport of the British government. In the era after World War II
the research personnel of this laboratory recognized the need for better information for the
design of storm drainage systems. The design method being used at that time was the Lloyd
Davies method which is the British counterpart of the rational method. Recognizing the need
for observed data on storm rainfall and runoff, suitable measuring equipment was developed.
Ultimately, a series of 12 urban basins were selected and instrumented. Storm rainfall and run-
off data were collected for 286 storm events on these watersheds during the period 1950-59.
These results were analyzed by the rational method and the unit hydrograph method. After
recognizing the limitations of these two methods, the new British RRL Hydrograph Method
was devised. This research and development has been described by Watkins (1962). Some of
the conclusions of this study were:
1. The rational equation is satisfactory as a design method for small areas in which there
are no drainage pipes larger than 24 inches in diameter. Here the errors introduced by
the rational method are tolerable.
2. The RRL hydrograph method is accurate and reliable for calculating hydrographs for
all urban areas.
3. The unit hydrograph method is unsuitable for use in designing urban drainage systems
due to the difficulty of determining the shape of the unit hydrograph.
Having completed this research and having devised the RRL method, the Road Research Lab-
oratory carried out further work to reduce this hydrograph method to an actual design method.
Computer programs were written and made available to design engineers at computing centers.
The Road Research Laboratory (1963) later provided a guide for engineers to use in providing
input into the computer program in order to obtain design output. This greatly simplified the
use of this design procedure for practicing engineers. It is reported that in Great Britain in
1972 about 80 percent of the design or re-design of storm drainage systems is carried out by
this RRL method.
The Procedure
The dominant feature of the RRL method is that it accommodates runoff only from the
paved areas of the basin which are directly connected to the storm drainage system. Grassed
areas are excluded from consideration as are paved areas which are not directly connected.
The principal elements of the procedure are as follows. Equal time increments of rainfall are
applied to the directly connected paved area in a small sub-basin of the total urban basin.
Next a computation is made of the travel time required for each increment of runoff to reach
the inlets at the downstream end of the sub-basin. In this way a surface hydrograph is pro-
vided for each sub-basin. These surface hydrographs from each sub-basin are accumulated in
a downstream order through the basin. This cumulation of inflow hydrographs is routed
through each section of pipe to accommodate the temporary storage within each pipe section.
The result is a computed outflow hydrograph from each section of pipe and this is ultimately
11
-------
provided at the outlet of the total basin. The procedure was described by Watkins (1962) and
by Terstriep and Stall (1969). In both of these studies storage for the basin was lumped to-
gether and the routing was accomplished in one step for the entire basin. A later improvement
to the method devised by the British and used in this study provides for multi-step routing of
the hydrograph through successive sections of pipe.
The RRL method is applied by first dividing the basin to be studied into sub-basins. A sub-
basin is normally a part of the basin contributing to a single inlet or set of inlets into one
storm drain pipe. Two physical factors must be evaluated for each sub-basin. First, the paved
area directly connected to the storm drainage system must be determined; second, the travel
time must be calculated for flows on paved areas and in gutters.
The various elements and steps used in developing a runoff hydrograph for an urban sub-basin
are illustrated in figure 2. Extending down the middle of the small sub-basin map (figure 2a)
is a city street with a pair of inlets at the lower end which allow water to enter a storm drain
pipe. Shown also are rooftops along this city street. The area shaded has been determined by
a field survey to be directly connected to the street. In each case about half of the driveway
has been considered to be contributing. The flow from roof No. 1 is not connected to the
street, but the flow from roof No. 2 reaches the drainage system either by way of the driveway
which flows into the street or by a direct underground connection.
After the directly connected paved area has been determined, calculations are made to deter-
mine the time-of-travel for the runoff water from various parts of the paved area to the inlets
at the downstream end of the sub-basin. In earlier studies the velocity and travel times for
overland flow were based on an equation developed by Hicks (1944) as described by Jens and
McPherson (1964). In the present project, travel times were computed in two steps. First, a
design flow of 0.5 cfs per acre of paved area was assumed: in a few cases this was increased to
1.0 cfs per acre. This design flow was considered to be flowing down various reaches of street
gutters.
The second step was to apply Manning's equation to compute the velocity of flow in the gutter.
By this means the travel times were computed for various reaches of the paved area in each sub-
basin. These travel times were plotted at various locations on the paved area, and by connect-
ing points of equal travel time a series of isochrons were drawn on the paved area (figure 2a).
The directly connected paved area between these isochrons was measured and designated area
Al, A2, A3, A4, and A5 as shown in figure 2a. These various areas are accumulated and plot-
ted against travel time to the inlet as shown in figure 2b. This time-area curve shows the
amount of paved area within the sub-basin which is contributing water at the inlet at any
time after the beginning of runoff. In the computer program described later, the time-area
curve was assumed to be a straight line connecting the origin and the end-point of the curve.
The end-point of the curve, as illustrated in figure 2b, represents the travel time of runoff
water from the furthest point of the directly connected paved area, and the total amount of
the directly connected paved area.
In constructing the hydrograph for each sub-basin the input is the rainfall pattern as a series
of intensities of equal duration (figure 2c). The rainfall input can be an actual event or a
design storm. The time increment used should be the same as the time interval between the
isochrons; this time interval, At, is used throughout the computations. In general it should be
as short as the quality of rainfall data will allow, except that for very large basins or very long
storms it may be more convenient to use a longer At.
Shown in figure 2d are the losses for the same time intervals as given for rainfall. For applica-
12
-------
a. SUB-BASIN MAP
(DIRECTLY CONNECTED PAVED AREA SHADED)
c. RAINFALL
1-minute
isochrons
Roof #1 o
not connected ^
i.
!3
<4
Ti
Roof #2
directly
connected
Inlets to
storm drain
d. LOSSES
j I
i 1
SR„
•=t.
Q
b. TIME-AREA CURVE
A5-
A
SR3
SR4
e. SUPPLY RATE
jR-i
^
012345
TRAVEL TIME TO INLET, minutes
01 234 56789 10
TIME, minutes
f. HYDROGRAPH
Q i
11
i.
o
-------
tion to a paved area the losses consist of initial wetting and depression storage. These losses
are combined and treated as an initial loss to be subtracted from the beginning of the rainfall
pattern. In figure 2d the entire initial loss LI occurs during the first minute or first time
interval.
After subtracting these losses from the rainfall pattern, the remainder of the rainfall will ap-
pear as runoff from the paved area. This runoff is shown in figure 2e and is referred to as the
paved area supply rate; it is also plotted at 1-minute intervals.
The ordinates of the surface hydrograph are developed by applying the supply rate to the time-
area curve, which is done by the series shown in figure 2f. The hydrograph developed in figure
2f occurs at the sub-basin inlets illustrated in figure 2a. Such a hydrograph is developed for
each sub-basin and becomes an input to the system at a particular point. If the sub-basin in
question happens to be at the upper-most end of a series of pipe or open channel reaches, the
hydrograph is entered into the system by routing it downstream to the next input point. If
the sub-basin occurs somewhere below the upper end, its hydrograph is combined with the
upstream hydrograph and the resulting combined hydrograph is routed downstream to the
next input point. If the sub-basin is located at the confluence of two or more pipes, its hydro-
graph is combined with the converging hydrographs before routing downstream to the next
input point.
2out
2t
TIME, FROM BEGINNING OF RUNOFF
Figure 3. Elements in the storage routing technique
14
-------
A simple storage routing technique is used to pass the hydrograph from one input point to the
next. In order to use this technique a determinate relationship must exist between discharge
and storage for the reach of channel or pipe between the input points. Such a relationship is
developed by first using Manning's equation to compute a stage-discharge curve for the cross
section in question. Since the length and geometry of the reach are known, the required dis-
charge-storage relationship may be computed by assuming uniform flow in the particular reach.
Errors incurred by this assumption are minimized by keeping the time increment and reach
length as short as practical.
Figure 3 shows two curves OQun Q2m which is a section of the inflow hydrograph at the upper
end of the reach and OQloutQ2out which is a section of the outflow hydrograph at the lower
end of the reach. Let SI and 52 be the total storage at times t and 2t respectively. From fig-
ure 3, area Qac = areas Oae + Qec or
1/2 (Qlin t) = S1 + 1/2 (Qlout t) (1)
Since Qlin and t are known, the right side of equation 1 may be evaluated. Because S^ is
known in terms of Q\out from the discharge-storage relationship described earlier, the equation
can be solved for Qlout. For the period t to 2t, area abdc = areas efdc + abfe, or
t/2(Qlin+Q2in) = t/2(Qlout + Q2out) + (S2-S1) (2)
t/2 (Qlin + Q-Lin ~ Qlout) + Sj. = t/2 Q2out + S2 (3)
Since the left side of equation 3 is known, the right side may again be solved for Q2out using
the discharge-storage relationship. By this step-by-step procedure all ordinates of the down-
stream or routed hydrograph may be determined.
Computer Program
A Fortran IV program has been developed to perform the operations just described. The pro-
gram also provides the summary tables and plotted hydrograph presented in Section VI of
this report. The program was written for and run exclusively on the University of Illinois
IBM 360/75 and Calcomp Plotter System. The major functions of the program are indicated
in the flow diagram of figure 4. Interested persons may contact the Environmental Protection
Agency to obtain a listing or program deck for this program.
Application of the RRL Method
Input Format. Figure 5 is a reproduction of the computerized input format for all the data
on one basin. Shown is only a skeleton amount of data in order to illustrate the layout of the
total input format. The data are for part of the Wingohocking Basin in Philadelphia. Items 1
and 2 and 3 are computer signals, and item 4 contains the basin name. Item 5 shows the many
individual facts regarding the basin parameters which were derived from the various aerial
photographs, street maps, and physical inspection of the basin. The data in item 5 comprise
the basin data. Item 6 is a computer signal and item 7 contains physical information on basin
parameters in a particular sub-basin.
Item 8 signals the beginning of storm data and item 9 is the storm identification. Item 10
carries the storm information on base flow, raingage weights, and scale parameters for provid-
ing a plotted output of the storm hydrograph. Item 11 gives the antecedent moisture condi-
tion which is used as only a general indication of conditions and is sometimes valuable in the
15
-------
READ INPUT DATA
LOOK AT SUB-BASIN
COMPUTE
TIME-AREA CURVE
COMPUTE SURFACE
HYDROGRAPH
COMBINE SURFACE
AND CHANNEL
HYDROGRAPHS
NO-
IS THIS AN
INITIAL REACH
-YES-
• NO-
/IS THIS A
"I CONFLUENCE,
-YES'
COMBINE
HYDROGRAPHS
[SELECT CROSS SECTION FOR DISCHARGE STORAGE RELATIONSHIP}
CIRCULAR
PIPE
L
RECTANGULAR
PIPE
OPEN
CHANNEL
J
ROUTE
HYDROGRAPH
TO NEXT
INPUT POINT
n
PRINT AND PLOT
OUTPUT
IS THERE
ANOTHER REACH
-YES-
Figure 4. Flow diagram for the computer program of the RRL method
16
-------
ITEM
2 02000 Signals the beginning of a new basin
2 03000 Signals the beginning of a new etorm
3 04000 Basin data will follow
4 HINGOHOCKINC, BASIN PHILADELPHIA Basin name as it will
appear on printout
a se < S
5 5326 3246 5 .015
6 05000 Sub-basin data will follow
15
£M
12 300
1 0 12 11.0 4030 1800 2.74 .015
711 8 25.7 R250 945 4.00 .015
20 8 10.2 2650 2250 2.00 .015
2 1
8 06000 Storm data will follow
36 1
243
3 4 0.5
9
Storm date for plotted output
WINGOHOCKING BASIN PHILADELPHIA TIME (HOURS) STORM OF NOV 24 1964
to is 4C tx _o a c '*3
Baseflow °f Thiessen X t«, ^ 6 '!, ,3 6 °,i *&§ ^ g £
Coordinates ,§ c Heights | | go- '| g «• | "g „ £ ^ g g
10
11
12
13
14
IS
16
1530
3
07000
Date
1*6
£<§ £
112564
112564
112564
112564
112564
015000
08000
Date
c * &
o a
-------
interpretation of the storm results. Item 12 signals that rainfall data will begin; item 13 gives
the actual rainfall data from the various gages. Item 15 signals that flow data will begin, and
item 16 provides the actual data on time and discharge amount in cfs. Item 17 is the end of
the input data.
Output. The computer program accomplishes the complete processing of all of the sub-basins
by the RRL method, and provides a printed output as well as plotted hydrographs of each
storm. As illustrated by table 1 for Woodoak Drive basin presented in the next section, the
output includes first (columns 1-7) the identification and observed storm data for each storm
on a basin. Column 8 is the observed runoff ratio if runoff had occurred only on the directly
connected paved area of the basin. In other words, this method assumes that runoff from the
directly connected paved area is virtually 100 percent and that no runoff occurs from any
other areas. On the basis of this assumption, the numbers in column 8 should all be 1.00;
consequently, the actual values in column 8 give some sense for the precision of the results.
Computed runoff and peak flow results are then presented (columns 9-12). Given along the
bottom of table 1 are the number of cases and the average error in the computations of the
peaks and of the runoff volume.
The complete graphical results for the application of the RRL method are exemplified by
figure 7 for the Woodoak Basin. Included are plots of the observed peaks versus the computed
peaks (figure 7a) and the observed runoff volume versus the computed runoff volume (figure
7b) for all of the storms listed in table 1.
Figure 7 illustrates the plotted computer output showing the goodness-of-fit of the total storm
runoff hydrograph as observed and as computed by the RRL method for one storm (October
19, 1966). Such plotted outputs were obtained for all of the storms, but only a sample plot
for one storm is presented in this report for each of the 10 basins.
Layout for Presentation of Basin Results
For each of the 10 basins included in this project, the basin results are presented as an output
table similar to table 1 and as a series of graphs similar to that shown in figure 7 for the Wood-
oak Drive basin. The basins are presented in the following section of the report in increasing
order of basin size.
18
-------
SECTION VI
DESCRIPTION AND RESULTS BY BASINS
Woodoak Drive Basin, West bury. Long Island, New York
Woodoak Drive basin is a 14.7-acre residential area all of which drains to one set of inlets. Be-
cause of its small size and the fact that only one length of pipe exists, it was not necessary to
divide the basin into smaller sub-basins. Street slopes are less than 1 percent and yard slopes
were estimated to be less than 3 percent. The dominant soil in the area is Haven Loam which
is classified in hydrologic group B. The existence of highly permeable gravelly sand at a depth
of 2 or 3 feet accounts for the success of recharge basins which are common in this area of
Long Island.
Most driveways in the area are paved and are either full-width drives or narrow strips of con-
crete that accommodate car tires. Roof drains appear to flow onto full-width drives where
such drives exist and onto the grass in other cases. The directly connected paved area there-
fore consists of the streets, all driveway aprons, all full-width driveways, and the front half of
roofs located adjacent to full-width drives.
Data. Flow measurements are made at 5-minute intervals by a digital stage recorder located
behind a V-notch weir in a 24-inch concrete pipe. The instrumentation, including a water
table measuring system, are described in detail by Seaburn (1970). Flow data were provided
for this project in printed form with discharge in cfs at 5-minute intervals.
Rainfall is recorded on a weighing-bucket gage located about 300 yards southeast of the basin.
Copies of the original weekly charts were provided for this study. The charts were replotted
using a larger time scale and read at 10-minute intervals.
Results. There appears to be no significant grassed area runoff from this basin. The permeable
soils and flat slopes combine to provide infiltration rates which can accommodate not only the
rain falling directly on the grassed area, but also runoff onto grass from the unconnected paved
areas. The RRL method works well on this basin.
19
-------
Table 1. Storm data and results for Woodoak Drive basin
Total Basin Area
14.7 acres
Total Paved Area
4.9 acres
33.9 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
10
Mean
Date
(2)
91466
92166
101966
102066
42765
50667
50767
82567
82667
52968
values
AMC
(3)
2
1
2
4
3
2
4
2
4
3
Rain
inches
(4)
1.42
3.48
2.38
0.76
1.49
0.65
1.13
1.74
0.72
3.60
Peak
flow
cfs
(5)
0.76
5.93
1.98
1.84
0.59
0.53
1.07
2.30
2.10
4.00
Data
Runoff
volume
inches
(6)
0.24
0.75
0.45
0.14
0.20
0.09
0.29
0.21
0.10
0.75
Directly Connected Paved Area
2.85 acres
19.4 percent
Computed Results
Runoff
ratio
(7)
0.17
0.22
0.19
0.18
0.13
0.15
0.26
0.12
0.14
0.21
0.18
Virtual
runoff
ratio
(8)
0.88
1.14
1.00
0.97
0.71
0.77
1.35
0.64
0.74
1.10
Runoff
volume
inches
(9)
0.25
0.64
0.43
0.12
0.26
0.10
0.19
0.31
0.12
0.66
Error
percent
(10)
4.7
-15.6
-4.9
-11.4
30.4
9.7
-32.8
45.1
16.1
-12.2
18.3
Peak
flow
cfs
(11)
0.85
2.53
1.35
1.27
0.68
0.69
0.86
3.02
2.73
3.12
Error
percent
(12)
12.3
-57.3
-31.7
-30.9
16.0
29.4
-19.9
31.3
30.2
-22.9
28.2
Computed peaks were high in 5 cases, average + error = 23.9 percent
Computed peaks were low in 5 cases, average - error = 32.4 percent
Computed runoff volumes were high in 5 cases, average + error = 21.2 percent
Computed runoff volumes were low in 5 cases, average — error = 15.4 percent
Figure 6. Outlet of the single storm drain in the Woodoak Drive basin
20
-------
6
"„ 4
CO
LU
D-
Q
LU
|—
^> o
Q- 2
§
O
0
1 1 1 1 1 /
A. PEAKS /
/
'
8 / 1.°
— • X
/ 2
9 /
/X
/y/--
5/ 7
/ 1 1 1 1 1
24
OBSERVED PEAKS, cfs
1.2
V>
OJ
Jo. 8
U_
u_
1
o;
o
£0.4
Q_
O
O1
1 1 1 1 1 /
B. VOLUMES /
/
— / —
x.10
— 2 _
/
/3
8'1 /X
~ vx 7'
XII II
0.4 0.8
OBSERVED RUNOFF, inches
1.2
2.0
1.5
in
o
LU
gi.o
<:
1C
o
0.5
I I I
C. HYDROGRAPHS
I I I I I
STORM OF OCTOBER 19, 1966
OBSERVED
Lg o
5 6
TIME, hours
10 11
Figure 7. RRL results for Woodoak Drive basin. Long Island, New York
21
-------
WOODOAK DRIVE BASIN
WESTBURY, LONG ISLAND, N.Y.
A STREAM GAGE
• RAIN GAGE
SCALE OF FEET
0 500
(Photo by Lockuood, Kessler, & Bartlett)
Figure 8. Aerial photo of Woodoak Drive basin
22
-------
Ross-Ade (Upper) Basin, West Lafayette, Indiana
The Ross-Ade upper basin is one of four research basins established by Purdue University. This
basin is entirely residential and relatively uniform in character. The basin has a definite valley-
type configuration. Woodlawn Avenue which runs down the center of the valley has slopes of
1 to 3 percent, but some of the side streets are steeper. Yard slopes vary from nearly flat in the
upper part of the basin to about 25 percent near the center of the basin. The basin was repre-
sented in the model by 12 sub-basins ranging in size from 0.2 to 1.6 acres.
Storm drainage is provided by an interceptor that runs down Woodlawn Avenue. The water-
shed boundary is difficult to determine. Inlets exist along the watershed boundary that are
part of a combined sewer system which does not pass through the gage. If these inlets become
inoperative or are overtopped during large storms, runoff can flow into the basin that would
otherwise be intercepted by this combined sewer system. Soils in the basin vary from Crosby
silt loam of hydrologic group C in the flood plain to Miami silt loam of hydrologic group B on
the steeper portions of the basin and Eel silt loam of hydrologic group C on the uplands.
Impervious areas were measured on large-scale aerial photographs. A field survey indicated
that about one-half of the roof drains had underground connections. In the lower portion of
the basin it was assumed that these were connected to the storm drain, but in the upper reaches
it was assumed that the connections were made to the combined system. The directly connect-
ed paved area then consisted of all streets, all driveways, and connected roofs in the lower part
of the basin.
Data. A Columbus-type deep-notch weir with a 6-foot crest length provides accurate flow
measurement at the gaging site. Stage hydrographs are recorded on a Stevens A-25 recorder
with a 20-inch chart. The chart speed of 6 inches per hour allows 1 minute stages to be read.
Rainfall is collected by a 16-inch diameter receiver located 8 feet above the ground. The re-
ceiver is connected to an 8-inch float chamber which records rainfall magnified 10 times on
the same chart as stage. Temperature is also recorded on these charts. Although the gages are
in operation at the time of this writing, data for the past few years have not been reduced.
Results. Data are limited to the two small storms shown in table 2. For these, the RRL meth-
od underestimates peaks and volumes, because of the grassed area contribution caused by
steep slopes.
23
-------
Table 2. Storm data and results for Ross-Ade (Upper) basin
Total Basin Area
54.0 acres
Total Paved Area
13.3 acres
24.7 percent
Observed Storm
Storm
(1)
1
2
Mean
Date
(2)
80267
82667
values
AMC
(3)
2
1
Rain
inches
(4)
0.37
0.14
Peak
flow
c/s
(5)
13.6
6.6
Data
Runoff
volume
inches
(6)
0.07
0.03
Directly Connected Paved Area
7.8 acres
14.4 percent
Computed Results
Runoff
ratio
(7)
0.19
0.20
0.19
Virtual
runoff
ratio
(8)
1.30
1.36
Runoff
volume
inches
(9)
0.04
0.01
Error
percent
(10)
-45.3
-80.4
62.8
Peak
flow
eft
(11)
8.8
2.3
Error
percent
(12)
-35.4
-65.6
50.5
Computed peaks were high in 0 cases
Computed peaks were low in 2 cases, average — error = 50.5 percent
Computed runoff volumes were high in 0 cases
Computed runoff volumes were low in 2 cases, average — error = 62.8 percent
Figure 9. Street view in Ross-Ade (Upper) basin
24
-------
15
M-
0 10
to
•a:
0.
O
LU
1 5
0
O
0
A. PEAKS
/ 1^ ~
.
.
/
/ '
/ i 1
5 10
OBSERVED PEAKS, cfs
0. 06
o
c
SO.04
o;
o
O.02
15
i 1 1 I 1
~ B. VOLUMES
/
/
/
// 2
/
1 >:
f
i
•
i i
0.02 0.04 0.06
OBSERVED RUNOFF, inches
15
I
C. HYDROGRAPHS
OBSERVED
00
M-
o
o
oo
10
COMPUTED
T
STORM OF AUGUST 2, 1967
TIME, hours
Figure 10. RRL results for Ross-Ade (Upper) basin. West Lafayette, Indiana
25
-------
>**•
ROSS-ADE DRAIN (UPPER) BASIN
WEST LAFAYETTE, INDIANA
A STREAM GAGE
• RAIN GAGE
#\
SCALE OF FEET
500
u
I 1
(USDA photo of June, 1971)
Figure 11. Aerial photo of Ross-Ade (Upper) basin
1000
J
26
-------
Sewer District No. 8 Basin, Bucyrus, Ohio
The 206-acre No. 8 Sewer District basin lies within the older section of Bucyrus, Ohio. Land
use varies from residential through commercial and heavy industrial. The entire basin is served
by a combined sewer system, but there does not appear to be an adequate number of inlets
to drain the basin properly. Dry-weather flow is intercepted above the gage. The combined
sewer system is quite extensive and was represented in the model by dividing the basin into 42
sub-basins. With the exception of a few short roadside swales there is no open channel drain-
age. Street slopes are generally less than 1 percent and yard slopes less than 3 percent. Princi-
pal soils in the area are Bennington Silt loam and Luray silty clay loam. These soils are both
classified in hydrologic group C-D.
Determination of the directly connected paved area was complicated by the lack of curb and
gutter on many streets in the northern part of the basin. Runoff from many of these streets
(figure 12) would apparently find its way into adjacent low-lying yards and vacant lots or be
ponded on the street. A sample inspection of roofs during field investigations indicated that
about 60 percent of the residential roofs are directly connected to the combined sewer system.
The directly connected paved area thus consisted of all streets with curb and gutter, a 10-foot
strip for streets without curb and gutter, driveways on curb and guttered streets, major build-
ings, and 60 percent of the residential roofs.
Data. Instrumentation, as described by Burgess and Niple (1969), consisted of a Stevens
Type-F stage recorder located behind an 8-foot rectangular weir. The data were provided in
the form of plotted discharge hydrographs. Since the distance between the 42-inch outfall
pipe and the weir was necessarily small, approach velocities could have had an effect on the
measurement of high flows.
Rainfall data were collected on a Bendix weighing-bucket gage with a 24-hour chart. The data
received for this study had been digitized and was read at 10-minute intervals.
Results. The runoff ratios of 10 to 15 percent for storms of 1 to 1.5 inches seem reasonable
for a basin with 18 percent directly connected paved area. Runoff that might be expected to
occur from the relatively tight soils in the basin is delayed by the lack of inlets and the result-
ant ponding. Infiltration is also enhanced by the ponded conditions.
The storm shown in figure 13 illustrates the problem of the RRL method on this basin. Com-
puted peaks are too quick and too high and flow recedes too quickly. A shortcoming of the
RRL method under these conditions is its inability to provide routing through surface storage
such as the ponding that occurs on this basin.
27
-------
Table 3. Storm data and results for Sewer District No. 8 basin
Total Basin Area
206 acres
Total Paved Area
43 acres
21 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
10
Mean
Date
(2)
32469
40569
51769
61369
71169
71769
72769
80969
81669
90669
values
AMC
(3)
2
2
3
2
4
4
1
3
1
2
Rain
inches
(4)
0.31
1.47
1.37
1.20
1.55
1.01
0.40
0.51
0.70
0.23
Peak
flow
eft
(5)
4.4
22.8
32.4
29.5
50.8
25.8
20.9
22.7
23.1
15.9
Data
Runoff
volume
inches
(6)
0.03
0.16
0.16
0.18
0.20
0.15
0.05
0.07
0.10
0.02
Directly Connected Paved Area
37.5 acres
18.2 percent
Computed
Runoff
ratio
(7)
0.10
0.11
0.11
0.15
0.13
0.14
0.12
0.14
0.14
0.09
0.12
Virtual
runoff
ratio
(8)
0.58
0.58
0.63
0.83
0.70
0.79
0.67
0.78
0.80
0.52
0.69
Runoff
volume
inches
(9)
0.04
0.25
0.23
0.20
0.26
0.17
0.05
0.08
0.11
0.02
Error
percent
(10)
18.4
59.1
45.7
10.6
32.0
17.1
10.4
1.8
7.6
11.0
21.4
Results
Peak
flow
cfs
(11)
6.1
43.6
58.8
58.3
129.5
49.9
27.5
32.4
29.9
12.8
Error
percent
(12)
37.7
91.5
81.4
97.7
155.0
93.4
31.7
42.7
29.6
-19.7
68.0
Computed peaks were high in 9 cases, average + error = 73.4 percent
Computed peaks were low in 1 case, average - error = -19.7 percent
Computed runoff volumes were high in 10 cases, average + error = 21.4 percent
Computed runoff volumes were low in 0 cases
Figure 12. Flat terrain and indeterminate drainage pattern typical of much
of the upstream basin of Sewer District No. 8
28
-------
120
80
Lul
O.
40
"I 1 1 T
A. PEAKS .
7 /
•9 /
'•12
'10
/•I 3
i i
,
40 80 120
OBSERVED PEAKS, cfs
»0.3
O)
u
c
£0.2
•z.
=>
0
h-
—I
^ 0.1
O
Q
III/
B. VOLUMES /
/ -
•- '9 ,/X/
''•X
x/9
/
li/
1-!/
i.,''io
x'3 '3
/5'2| | | | |
0 0.1 0.2 0.3
OBSERVED RUNOFF, inches
50
40
30
« 20
Q
10
-I I T-1
C. HYDROGRAPHS l|
'I
STORM OF JULY 17, 1969
Figure
TIME, hours
13. RRL results for Sewer District No. 8 basin, Bucyrus, Ohio
29
-------
Figure 14. Aerial photo of Sewer District No. 8
SEWER DISTRICT No. 8 BASIN
BUCYRUS, OHIO
STREAM GAGE
• RAIN GAGE
SCALE OF FEET
0 500 1000
:=Z3
ft
(USDA photos of 1970)
30
-------
Echo Park Avenue Basin, Los Angeles, California
The Echo Park Avenue basin is primarily a residential area with commercial strips along the
main streets. The basin has a deep valley configuration. Runoff flows down very steep side
streets to an interceptor flowing north-to-south along the center of the valley. Minimum
slopes in the basin occur down the center of the valley where they vary from 2 to 4 percent.
On the side streets slopes approach 20 percent and on landscaped areas slopes of 30 percent
are not uncommon. The dominant soil in the basin according to a 1916 survey is Altamont
loam. Under natural, undisturbed conditions, this soil would be in hydrologic group B or C
depending on the depth to bedrock and the degree to which the rock is weathered.
Surveys by the City of Los Angeles fixed the total paved area at 136 acres. These surveys
showed that 54 percent of the total paved area was in streets and parking, and that the other
46 percent was in roofs. An additional roof survey indicated that 40 percent of the roofs are
connected to the streets and 60 percent to the lawns. The directly connected paved area thus
consisted of 73 acres of streets and parking (136 X 0.54), and 25 acres of connected roofs
[(136-73) X 0.40], for a total of 98 acres.
Data. Stage hydrographs are recorded in a 51-inch concrete storm sewer by the Bureau of
Engineering at the City of Los Angeles. The original charts along with a rating table based on
Manning's equation assuming uniform flow and a 0.013 "n" value were furnished by the
bureau. Crawford (1971) has recently commented on the Echo Park data: "the flow data
could be in error by more than 20 percent due to uncertainty in the roughness and the super-
critical flow velocities in the sewer."
Rainfall is recorded on a weighing-bucket type gage on a standard 24-hour chart. These charts
were provided by the Bureau of Engineering and were digitized, as a regular part of this project,
by the Water Survey Model 3400 auto-trol. A 4-minute time interval was used for rainfall re-
duction. Because of the short entry times and quick response of this basin, an even shorter
time interval would have been desirable.
Results. The RRL method clearly does not apply to this basin. Grassed-area contribution has
a significant effect on both the peak and the volume of runoff.
31
-------
Table 4. Storm data and results for Echo Park Avenue basin
Total Basin Area
252 acres
Total Paved Area
136 acres
53.8 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
Mean
Date
(2)
20358
20458
21958
21262
21962
20963
12164
12264
40865
40965
111967
112067
12669
20669
21569
22569
30470
122170
values
AMC
(3)
2
4
1
4
4
1
2
4
4
4
1
4
4
2
2
4
4
4
Rain
inches
(4)
0.66
1.10
3.43
0.68
1.54
2.38
1.06
0.54
1.11
1.30
0.88
0.49
0.85
1.01
1.00
1.33
1.35
1.35
Peak
flow
c/s
(5)
275
260
295
234
204
170
178
187
182
199
260
284
187
240
196
146
147
116
Data
Runoff
volume
inches
(6)
0.29
0.56
1.43
0.42
0.68
0.73
0.41
0.25
0.44
0.74
0.49
0.31
0.56
0.60
0.45
0.84
0.55
0.24
Directly Connected Paved Area
97.7 acres
38.8 percent
Computed Results
Runoff
ratio
(7)
0.45
0.51
0.42
0.61
0.44
0.31
0.38
0.47
0.39
0.57
0.55
0.62
0.66
0.59
0.45
0.63
0.41
0.18
0.48
Virtual
runoff
ratio
(8)
1.15
1.31
1.08
1.59
1.14
0.79
0.99
1.22
1.01
1.47
1.42
1.61
1.69
1.53
1.16
1.63
1.06
0.46
Runoff
volume
inches
(9)
0.22
0.38
1.28
0.22
0.55
0.87
0.37
0.17
0.39
0.46
0.30
0.15
0.29
0.35
0.35
0.47
0.48
0.48
Error
percent
(10)
-26.9
-31.1
-10.7
-46.6
-18.3
19.3
-9.7
-33.5
-11.0
-37.9
-38.3
-51.0
-48.4
-41.9
-23.4
-43.9
-13.3
100.9
30.3
Peak
flow
cfs
(11)
161
137
207
110
104
85
93
74
101
89
125
155
85
104
105
65
90
91
Error
percent
(12)
-41.6
-47.4
-30.0
-53.4
-49.3
-49.9
-47.9
-60.6
-44.2
-55.6
-51.9
-45.5
-54.5
-56.4
-46.4
-55.2
-38.6
-21.3
47.2
Computed peaks were high in 0 cases
Computed peaks were low in 18 cases, average — error = —47.2 percent
Computed runoff volumes were high in 2 cases, average + error = 60.1 percent
Computed runoff volumes were low in 16 cases, average — error = —30.4 percent
Figure 15. Street view in Echo Park Avenue basin
32
-------
300
20°
LU
Q-
100
O
O
1
A. PEAKS
~1 T
/
/
/
9 15 •
^ _ _c -
/ '" * ;*;"°
QV. L
1.5
T—[—I—I—I—T
- B. VOLUMES
0.5
100 200
OBSERVED PEAKS, cfs
300
0
V
X
0 0.5 1.0 1.5
OBSERVED RUNOFF, inches
200
150
Si 00
<£.
^
O
oo
I
C. HYDROGRAPHS
50
I I I
STORM OF APRIL 8, 1965
OBSERVED
2 3
TIME, hours
Figure 16. RRL results for Echo Park Avenue basin, Los Angeles, California
33
-------
SCALE OF FEET I'C,
0 500
ECHO PARK AVENUE BASIN
LOS ANGELES, CALIFORNIA
STREAM GAGE
RAIN GAGE
(Photo by Geotronios, Long Beach, California, 1968)
Figure 17. Aerial photo of Echo Park Avenue basin
-------
Crane Creek Basin, Jackson, Mississippi
The Crane Creek basin is a 273-acre residential area. Two large schools, a church, and an
apartment complex have a significant effect on the paved area runoff. There are large open
areas around the schools and in the flood plain in the lower part of the basin. Street slopes
range from 1 to 30 percent and yard slopes vary from 2 to 6 percent. The drainage system as
represented in the model has 11 open channel reaches with a total length of 5700 feet and 15
closed conduits with a total length of 6800 feet. The primary soil in the basin is a Loring silt
loam which is classified in hydrologic group C. In the flood plain area, Falaya series soils of
hydrologic group D should be expected.
The absence of curb and gutter on many streets complicate the determination of directly con-
nected paved area. All such streets have well-maintained roadside ditches and conceivably the
contributing roadway could include everything between the centerline of the ditches. For this
study, however a 20-foot strip of contributing area was used for streets without curb and gut-
ter. In addition to the streets, all major buildings, parking lots, and an approximation of resi-
dential driveways were included in the directly connected paved area. Residential roofs gen-
erally drain onto grass.
Data. The instrumentation for this basin is typical for a USGS installation. One digital record-
er provides stage hydrographs at a rated culvert on Crane Creek. Another digital recorder pro-
vides rainfall at the same site. The recorders operate from the same clock at 5-minute intervals.
Rainfall and discharge data were provided in both tabular and plotted form for this study.
Urban runoff effects for several basins in Jackson have been published elsewhere by Wilson
(1968).
Results. The RRL method seriously underestimates peak flow and runoff volumes on this
basin. A substantial amount of grassed-area runoff occurs both from yards and from the large
open areas. This is substantiated by the fact that the 8 storms which fall lowest on the two
plots (figures 19a and b) occurred when the soil was saturated, an antecedent moisture condi-
tion of 4.
35
-------
Table 5. Storm data and results for Crane Creek basin
Total Basin Area
273 acres
Total Paved Area
65.5 acres
23.9 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Mean
Date
(2)
51565
62465
62565
72465
81265
82065
91065
91165
92265
100665
10466
20166
22666
30366
42066
42666
52366
values
AMC
(3)
2
2
4
2
4
2
1
4
3
2
4
4
3
4
2
4
4
Rain
inches
(4)
0.38
1.44
0.78
2.00
1.81
0.64
1.74
1.27
0.58
1.14
1.79
0.45
0.65
0.57
3.23
1.09
1.16
Peak
flow
eft
(5)
39
106
65
161
149
30
78
253
20
70
60
23
22
137
154
116
248
Data
Runoff
volume
inches
(6)
0.07
0.24
0.15
0.42
0.47
0.08
0.25
0.54
0.04
0.16
0.46
0.08
0.13
0.29
0.70
0.38
0.56
Directly Connected Paved Area
39.7 acres
14.5 percent
Computed Results
Runoff
ratio
(7)
0.18
0.17
0.20
0.21
0.26
0.12
0.15
0.42
0.07
0.14
0.26
0.17
0.20
0.50
0.22
0.35
0.26
0.23
Virtual
runoff
ratio
(8)
1.34
1.24
1.42
1.51
1.88
0.89
1.06
3.06
0.47
1.04
1.87
1.22
1.46
3.65
1.58
2.55
3.49
Runoff
volume
inches
(9)
0.04
0.18
0.09
0.25
0.23
0.07
0.21
0.16
0.06
0.13
0.22
0.05
0.07
0.06
0.41
0.13
0.15
Error
percent
(10)
-42.9
-27.0
-43.1
-39.6
-50.5
-13.0
-16.0
-70.6
59.9
-17.6
-53.3
-40.0
-48.2
-78.3
-41.7
-66.2
-73.2
45.9
Peak
flow
cfs
(11)
38
91
35
127
157
46
53
108
54
81
32
26
15
64
112
53
87
Error
percent
(12)
-2.1
-14.3
-46.1
-20.8
5.5
55.2
-31.7
-57.3
166.4
15.2
-47.5
11.1
-33.1
-53.4
-27.2
-54.1
-64.9
41.5
Computed peaks were high in 5 cases, average + error = 50.7 percent
Computed peaks were low in 12 cases, average — error = 37.7 percent
Computed runoff volumes were high in 1 case, average + error = 59.9 percent
Computed runoff volumes were low in 16 cases, average — error = 45.1 percent
Figure 18. Ditch along street in Crane Creek basin
36
-------
300
if_
o
. 200
CO
^
UJ
O-
o
LU
1—
1 100
o
O
0
1 1 1 1 — • •• /
A. PEAKS /
X
X
/
.
- .' -
'
/
5 /
,
/ 4
10 /' • .
*/ 17
9 / . ,*
~ 6« / * 16 —
-/* r,'3
/ *3 i i i i i
0.6
t/1
B. VOLUMES /
/
/
/
/
. 10.71— »
/ 15
/
/
/
/ : s
"' . -
/ 2 i
• 17
/ • •
.. 10 16
~ / 0 —
9 _>6 . 3
/ •* ' 2 ' 3 14
/ ' 1 1 II
TOO 200
OBSERVED PEAKS, cfs
300 0 0.2 0.4 0.6
OBSERVED RUNOFF, inches
200
150
100
5
•—i
Q
50
I I
C. HYDROGRAPHS
OBSERVED
I I 1
STORM OF JULY 24, 1965
012345
TIME, hours
Figure 19. RRL results for Crane Creek basin, Jackson, Mississippi
-------
#•
In
CRANE CREEK BASIN
JACKSON, MISSISSIPPI
A STREAM GAGE
• RAIN GAGE
SCALE OF FEET
0 500 1000
20.
p/?ofo o^ Crane Creek basin
38
-------
Tripps Run Tributary Basin, Falls Church, Virginia
Tripps Run is primarily a residential basin, but there is a significant amount of commercial
development adjacent to U.S. Route 50 which crosses the basin in an east-west direction.
North of Route 50 the residential area is relatively dense compared with the large lots and open
areas to the south. The streets south of Route 50 are asphalt strips laid on existing grade with-
out curb and gutter or roadside ditches, as illustrated in figure 21. Of the 15 reaches used to
represent the drainage system, 5 were open channels with a combined length of 2370 feet and
10 were closed conduits with a combined length of 8325 feet. Storm drain information was
difficult to obtain. In several locations missing data had to be filled with what seemed appro-
priate. Street slopes in the basin vary from 1 to 6 percent and yard slopes vary from 3 to 10
percent. Dominant soils in the general area of the basin are Appling and Louisburg in hydro-
logic group B and Colfax in hydrologic group C.
The directly connected paved area includes all of the streets, all of the commercial area, and
driveways in the residential areas.
Data. The USGS provided the data for this study in the form of original charts from a Stevens
graphical stage-recorder located on a rated culvert. The recorder was equipped with a second
pen that recorded blips from a tipping bucket raingage on the same chart. The time scale of
0.2 inches per hour was adequate to define the stage hydrographs but not for accurate timing
of bucket tips. The 0.1 inch tipping bucket was a further limitation on this data. Since it was
recognized that a different interpretation of the rainfall between bucket tips was possible, the
most intense storms from the available data were read at 10 minute intervals.
Results. The RRL method does not work on this basin. There is a significant amount of
grassed-area contribution to runoff. The fact that the prediction of peak values is slightly
better than the prediction of runoff volumes shows the influence of the concentrated paved
areas in the commercial zone.
39
-------
Table 6. Storm data and results for Tripps Run Tributary basin
Total Basin Area
322 acres
Total Paved Area
100 acres
31.0 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
Mean
Date
(2)
62963
81963
82063
60265
81865
82665
100765
82467
102567
values
AMC
(3)
1
2
4
3
2
3
1
4
1
Rain
inches
(4)
2.75
2.55
2.45
0.85
0.85
1.35
3.10
2.55
0.90
Peak
flow
c/s
(5)
225
219
285
47
131
203
221
312
62
Data
Runoff
volume
inches
(6)
0.78
0.58
1.23
0.08
0.17
0.30
1.00
2.27
0.17
Directly Connected Paved Area
56.9 acres
17.7 percent
Computed Results
Runoff
ratio
(7)
0.28
0.23
0.50
0.09
0.20
0.22
0.54
0.44
0.14
0.29
Virtual
runoff
ratio
(8)
1.61
1.28
2.85
0.52
1.15
1.24
1.83
5.03
1.10
Runoff
volume
inches
(9)
0.47
0.44
0.42
0.14
0.13
0.22
0.54
0.44
0.14
Error
percent
(10)
-39.4
-23.6
-66.0
76.4
-22.5
-24.2
-46.2
-80.7
-18.1
44.1
Peak
flow
cf,
(11)
101
155
174
68
125
210
90
92
84
Error
percent
(12)
-55.1
-29.2
-38.9
44.3
-4.9
3.3
-59.1
-70.7
36.0
37.9
Computed peaks were high in 3 cases, average + error = 27.9 percent
Computed peaks were low in 6 cases, average — error = 43.0 percent
Computed runoff volumes were high in 1 case, average + error = 76.4 percent
Computed runoff volumes were low in 8 cases, average — error = 40.1 percent
Figure 21. Typical large lawns in the downstream portion of
Tripps Run Tributary basin
40
-------
300
to
(J
-
< 200
LU
O_
Q
LU
ID
o 100
o
0
1 III'
A. PEAKS
/
/
/
/
'3
~ ,' 2 —
/
/ 1.
•9// '7 '8
•yx
/I 1 1
100 200 300
OBSERVED PEAKS, cfs
.3
01
JC
o
u_
|2
oi
C3
LU
O
O
0
c
1 1 /'
B. VOLUMES /
/
/
/
/
- /C - "
/^l 9 1 1 1 1
) 1 2 3
OBSERVED RUNOFF, inches
240
180
60
0
I r^
C. HYDROGRAPHS
I I I
STORM OF AUGUST 19-20, 1963
123456
TIME, hours
Figure 22. RRL results for Tripps Run Tributary basin, Falls Church, Virginia
41
-------
TRIPPS RUN TRIBUTARY BASIN
FALLS CHURCH, VIRGINIA
A STREAM GAGE
• RAIN GAGE
(USDA photo of May 25, 1962)
J
Figure 23. Aerial photo of Tripps Run Tributary basin
42
-------
Tar Branch Basin, Winston-Salem, North Carolina
A large part of downtown Winston-Salem and major industrial areas lie within the boundaries
of the Tar Branch basin and account for the high percentage of paved area. The remainder of
the basin is light commercial or residential. In order to represent the extensive storm drainage
system in reasonable detail, 103 sub-basins were used. Of the 103 reaches, 15 were open chan-
nels with a combined length of 7200 feet. Pipes in the system ranged from 10 inches up to 72
inches in diameter. Information on the drainage system was not complete, and storm drain
slopes were assumed to be the same as street slopes in many cases. Street slopes are highly
variable. In the downtown area they are gentle; but range up to 10 percent in other parts of
the basin. Yard slopes are also variable ranging from 3 to 10 percent. The dominant soil in the
basin is a Pacolet fine sandy loam which in the undisturbed state is in hydrologic group B.
The directly connected paved area consists of all of the downtown commercial area, all other
streets, and other major buildings and parking lots. Residential roofs are not generally connect-
ed to the drainage system and private driveways are usually not paved.
Data. The instrumentation on this basin was the USGS standard installation for urban basins.
Two digital recorders punch the rainfall and stage synchronously at 5-minute intervals. In this
case the instruments are located on an open channel above a rated culvert. For this study the
rainfall and the discharge were both provided in tabular form at 5-minute intervals.
Results. There is no significant grassed-area runoff contribution from the Tar Branch basin.
The RRL method works well because of the large percentage of paved area and because of the
permeable soil.
43
-------
Table 7. Storm data and results for Tar Branch basin
Total Basin Area
384 acres
Total Paved Area
227 acres
59 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
Mean
Date
(2)
62668
71268
101868
61169
61569
61869
62169
72869
61370
121867
31668
42368
41569
52769
60969
80369
121069
values
AMC
(3)
2
2
4
4
4
4
2
2
3
1
4
2
2
3
2
3
4
Rain
inches
(4)
0.67
1.42
2.88
3.03
2.22
0.38
0.82
1.38
2.05
1.00
1.14
0.52
0.88
1.22
0.71
0.43
1.23
Peak
flow
cfs
(5)
265
397
175
945
171
210
316
290
857
97
132
73
71
105
134
80
82
Data
Runoff
volume
inches
(6)
0.34
0.69
0.71
1.97
0.46
0.12
0.33
0.53
1.80
0.43
0.32
0.10
0.30
0.41
0.13
0.07
0.31
Directly Connected Paved Area
195 acres
51 percent
Computed Results
Runoff
ratio
(7)
0.51
0.48
0.25
0.65
0.21
0.31
0.40
0.39
0.88
0.43
0.28
0.20
0.34
0.33
0.18
0.17
0.25
0.37
Virtual
runoff
ratio
(8)
1.00
0.95
0.49
1.28
0.41
0.61
0.78
0.76
1.73
0.85
0.54
0.38
0.67
0.66
0.35
0.35
0.50
Runoff
volume
inches
(9)
0.29
0.67
1.41
1.48
1.06
0.15
0.37
0.59
0.98
0.45
0.53
0.21
0.39
0.56
0.31
0.17
0.57
Error
percent
(10)
-15.0
-3.1
97.8
-24.9
127.8
29.9
12.0
11.1
-45.6
5.0
66.7
110.4
32.2
38.7
147.6
123.3
83.4
57.3
Peak
flow
cfs
(11)
302
540
299
713
335
260
456
297
686
60
129
79
91
165
219
96
66
Error
percent
(12)
13.9
36.0
70.9
-24.5
96.2
23.8
44.4
2.5
-20.0
-38.4
-2.2
7.7
28.7
56.9
63.2
19.6
-20.1
33.4
Computed peaks were high in 12 cases, average + error = 38.6 percent
Computed peaks were low in 5 cases, average — error = —21.0 percent
Computed runoff volumes were high in 13 cases, average + error = 68.1 percent
Computed runoff volumes were low in 4 cases, average — error = —22.2 percent
Figure 24. Typical houses and street in the Tar Branch basin
44
-------
1200
800 -
A. PEAKS
<=c
UJ
D.
O
UJ
400
o
o
/
/
ol^
II
400 800
OBSERVED PEAKS, cfs
1 7
/
1200
o
c
o
UJ
o
o
I r~
B. VOLUMES
?.• v
/
_L
II
"I /I
/
J L
_L
1 2
OBSERVED RUNOFF, inches
1 1
C. HYDROGRAPHS
COMPUTED
OL
5 min
RAINFALL
1 1
STORM OF JULY 28, 1969
OBSERVED
FT ' s s s sim-A PI tyx ni
3 4
TIME, hours
Figure 25. RRL results for Tar Branch basin, Winston-Salem, North Carolina
45
-------
TAR BRANCH
WINSTON SALEM, N.C.
A STREAM GAGE
• RAIN GAGE
SCALE OF FEET
0 500 1000
(USDA photo of Oat. 1966)
Figure 26. Aerial photo of Tar Branch basin
46
-------
Third Fork Basin, Durham, North Carolina
The Third Fork basin contains a variety of land uses. There is a high-density commercial area
and a significant industrial area along the northern watershed boundary. The residential area,
which makes up most of the basin, is itself highly variable ranging from simple frame homes on
dirt streets to homes on large lots. Surrounding the channel in the southern part of the basin
are over 100 acres of open park area. Soils in the flood plain are primarily Cangaree loams.
Although these are classified in hydrologic group B, the high water table in this area could add
significantly to the runoff potential. Upland soils consist of White Store soils and are classified
in hydrologic group D. With the exception of a few pipes in the upper reaches of the basin, all
drainage is by open channel. Of the 39 reaches used to describe the storm drainage system,
only 8 are closed conduits. Street and channel slopes are moderate, ranging from less than 1
to about 5 percent. Yard slopes range from 5 to 10 percent.
The total paved area of the basin was determined by zoning out the 100-percent paved areas
and the park areas, and measuring sample blocks in the remaining area. The residential area
was divided into 3 zones; low income, middle income, and high income. It was assumed that
zero, 10 percent, and 12.5 percent, respectively, of these roof areas were connected to the
storm drainage system. In the areas where paved streets did not exist a 15-foot strip was
assumed to be connected to the system. The directly connected paved area thus consisted of
147 acres of commercial area, 126 acres of streets, and 20 acres of residential roofs and drive-
ways.
Data. This is a standard USGS installation. Two digital recorders operating from the same
clock punch the stage hydrograph and the rainfall at 5-minute intervals. The stage hydrograph
is recorded at a rated culvert section in an open channel. For this study both rainfall and dis-
charge were provided in tabular form at 5-minute intervals.
Results. The RRL method works on small storms but underestimates large storms. Much of
the directly connected paved area is in the upper reaches of the basin. Flood runoff from this
area flows downstream through an earth channel. For small storms the flood peak reaches the
gage later than it would if the paved area were evenly distributed or located in the lower por-
tions of the basin. The RRL method correctly reproduces this delayed peak for small storms.
For big storms a large grassed-area contribution is generated in the lower part of the basin and
arrives at the gage at about the same time as the delayed paved-area contribution. The RRL
method which ignores the grassed-area contribution thus seriously underestimates both the
peak and the volume of runoff.
47
-------
Table 8. Storm data and results for Third Fork basin
Total Basin Area
1075 acres
Total Paved Area
397 acres
37 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
Mean
Date
(2)
60969
61569
72869
80169
80369
80469
81369
81569
90269
91769
92469
121069
122169
122569
21670
values
AMC
(3)
2
4
3
3
4
4
3
3
1
1
4
3
1
3
1
Rain
inches
(4)
0.64
1.80
0.97
0.72
0.83
0.50
0.53
1.96
0.73
1.36
0.60
1.05
0.83
0.73
2.11
Peak
flow
eft
(5)
77
500
485
137
593
199
120
1700
593
732
217
205
105
116
245
Data
Runoff
volume
inches
(6)
0.15
0.79
0.33
0.19
0.48
0.19
0.17
1.39
0.32
0.37
0.31
0.36
0.27
0.32
1.01
Directly Connected Paved Area
293 acres
27 percent
Computed
Runoff
ratio
(7)
0.23
0.44
0.35
0.26
0.58
0.38
0.31
0.71
0.43
0.27
0.52
0.34
0.32
0.44
0.48
0.40
Virtual
runoff
ratio
(8)
0.83
1.61
1.27
0.97
2.14
1.40
1.15
2.60
1.59
1.00
1.90
1.26
1.18
1.62
1.75
Runoff
volume
inches
<9)
0.14
0.46
0.23
0.17
0.20
0.11
0.11
0.50
0.17
0.34
0.13
0.25
0.19
0.17
0.53
Error
percent
(10)
-1.7
-42.3
-31.2
-12.4
-59.5
-43.7
-31.2
-63.9
-46.4
-8.9
-57.2
-29.3
-27.2
-48.0
-46.9
36.7
Results
Peak
flow
cfs
(11)
86
324
202
151
279
131
128
1021
346
646
81
183
87
75
159
Error
percent
(12)
12.0
-35.2
-58.3
9.9
-52.3
-34.0
6.4
-39.9
-41.7
-11.8
-62.5
-10.9
-17.3
-35.1
-35.3
31.2
Computed peaks were high in 3 cases, average + error = 9.4 percent
Computed peaks were low in 12 cases, average — error = —36.2 percent
Computed runoff volumes were high in 0 cases
Computed runoff volumes were low in 15 cases, average - error = -36.7 percent
Figure 27. Open channel which is representative of most of
the drainage in Third Fork basin
48
-------
1200'
800'
"i i 1 1 r
A. PEAKS
o
LU
i 400
o
o
X
I/'.,.
7* *6 '
/
1 1 1
1.2
Ol
o
•^0.8
LU
O
^^
0
£ 0.4
a.
o
o
0
1 1 1 1 1 '
B. VOLUMES //
/
/
/
/ 15 (1.39)-- •
/ *
/i
/ 3.12
- ,//•;, \
/ XV6 *
/ i i i i i
400 800
OBSERVED PEAKS, cfs
1200
0 0.4 0.8 1.2
OBSERVED RUNOFF, inches
500
400
£ 300
200
100
1 1 1 r
C. HYDROGRAPHS
OBSERVED
COMPUTED
5 min
RAINFALL
1 1 1 1 1 1"
STORM OF JULY 28, 1969
i i
0 1
5 6 7 8 9 10 11 12 13 14
TIME, hours
Figure 28. /?/?£ results for Third Fork basin, Durham, North Carolina
49
-------
THIRD FORK BASIN
DURHAM, N.C.
A STREAM GAGE
• RAIN GAGE
SCALE OF FEET
0 1000 2000
(VSDA photo of Oat. 1966)
Figure 29. Aerial photo of Third Fork basin
50
-------
Dry Creek Basin, Wichita, Kansas
This is primarily a residential basin with a few strips of commercial area. There is no under-
ground storm drainage in the basin. Runoff is transported via streets to either the East or
West Branches of Dry Creek. The East Branch has had some improvement but is essentially
a natural stream. The West Branch flows for much of its length through specially modified
street cross sections which are in effect a concrete canal. This is illustrated in the photo in
figure 30. As a result, flow down the West Branch is much faster than flow down the East
Branch. Street slopes are quite flat, averaging less than 0.5 percent. Yard slopes vary from
2 to 8 percent. Dominant soils in the area are Dale silt loam and Farnum loam, both classified
in hydrologic group B. There is a small area of Bethany silt loam in the upper reaches of the
basin which is in hydrologic group C.
Twenty-three sample blocks were used to determine the paved area of the basin. The directly
connected paved area includes all of the streets, major buildings and parking lots, and 25 per-
cent of the remaining paved area.
Data. Both rainfall and stage data on this basin are collected by digital punch-type recorders
located on a rated bridge-opening on Dry Creek. A graphical stage recorder originally installed
was found impractical because of the rapid changes in stage. The data provided for this project
were for 1964 through 1969, but only the 1964-1965 data were used because, after a dry peri-
od during 1966-1967, there appeared to be a shift in the rating curve. This shift has been
fairly well documented by the USGS personnel, but there is still some question about the
1968-1969 data.
Results. There is no significant grassed-area runoff from this basin. Although the results are
erratic the RRL method does seem to apply. The flat slopes and relatively permeable soil are
factors which favor the RRL method.
51
-------
Table 9. Storm data and results for Dry Creek basin
Total Basin Area
1882 acres
Total Paved Area
583 acres
31 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
Mean
Date
(2)
50564
80964
82764
91964
22865
51365
52465
70965
values
AMC
(3)
2
4
4
4
2
2
4
3
Rain
inches
(4)
0.72
2.20
1.88
0.56
0.60
2.38
0.74
1.68
Peak
flow
eft
(5)
250
365
580
226
212
608
148
505
Data
Runoff
volume
inches
(6)
0.15
0.34
0.39
0.11
0.15
0.55
0.08
0.56
Directly Connected Paved Area
365 acres
19 percent
Computed Results
Runoff
ratio
(7)
0.21
0.15
0.21
0.19
0.25
0.23
0.11
0.33
0.21
Virtual
runoff
ratio
(8)
1.07
0.80
1.07
1.00
1.32
1.18
0.58
1.71
Runoff
volume
inches
(9)
0.12
0.40
0.34
0.09
0.09
0.43
0.12
0.30
Error
percent
(10)
-21.0
17.4
-13.1
-19.0
-39.1
-20.9
45.7
-46.0
27.8
Peak
flow
cfs
(11)
219
355
645
175
155
522
210
351
Error
percent
(12)
-12.6
-2.6
11.3
-22.7
-26.7
-14.1
41.9
-30.4
20.3
Computed peaks were high in 2 cases, average + error = 26.6 percent
Computed peaks were low in 6 cases, average — error = 18.2 percent
Computed runoff volumes were high in 2 cases, average + error = 31.6 percent
Computed runoff volumes were low in 6 cases, average — error = 26.5 percent
Figure 30. View of street typical of those which carry most
surface runoff in the Dry Creek basin
52
-------
600-
400-
LU
Q.
200-
1 1 1 1 1 71
A. PEAKS .3 /
-
/
/ * ^
/
/ -
/* 2 'Q
_ /
/
/
'V^'1
" r
,-'
/
/ 1 1 ! 1 ] 1
0.6
I/)
OJ
_C
O
t 0.4
o
— i
cc
Q
LU
| —
£ 0.2
Q
O
O1
i i 1 1 1 1 7
B. VOLUMES /
/
/'
/ -
/
/ '6
"z / ~"
/
/ • 3
,'
/ '8 ~
/
/
-
/
'7//'1
/
/
t 1 1 1 1
200 400 600
OBSERVED PEAKS, cfs
0 0.2 0.4 0.6
OBSERVED RUNOFF, inches
STORM OF OCTOBER 9-10, 1964
01 23456789
TIME, hours
Figure 31. RRL results for Dry Creek basin, Wichita, Kansas
10 11
53
-------
Ifflrf
**»• WvW i
fr8R3l ,
I
ft1
*&
!!« b
,
(£/5&4 ptoto of J^ne,
Figure 32. Aerial photo of Dry Creek basin
54
-------
Wingohocking Basin, Philadelphia, Pennsylvania
Wingohocking is the largest and the most highly urbanized basin in this report. There are a
few areas of separate single-family residences in the basin, but row-houses are by far the most
common. Extensive commercial and industrial areas also exist in the basin. There are no open
channels. An extensive combined sewer system with arch-shaped pipes up to 21 by 24 feet
provides storm drainage. A sanitary interceptor sewer picks up dry weather flow just above
the gage. The basin is represented in the model by 128 separate sub-basins ranging in size
from 1.2 to 117 acres. Street slopes generally range from 0.5 to 2 percent and yard slopes
from 3 to 10 percent. Soils in the area are either in the Chester Complex group which is
classified 85 percent hydrologic group B and 15 percent C-D, or Ho well Complex which is
classified 75 percent hydrologic group B and 25 percent C-D.
Paved areas were based on studies previously made by the city of Philadelphia and confirmed
during this study by measuring sample blocks on aerial photographs. All of the paved areas in
the basin, including residential rooftops, are directly connected to the drainage system.
Data. The flow measurement program, as described by Tucker (1969), was established by
the U.S. Public Health Service in 1963. A graphical stage recorder was installed 450 feet up-
stream from a low broad crested weir. The weir, which itself is 87 feet above the outfall, was
rated by a physical model built and tested at Swarthmore College in 1964. The Research and
Development unit of the Philadelphia Water Department took over the gage in July of 1967 and
again built a model of the weir at the city's Northeast Water Pollution Control Plant. For use
in this project discharge hydrographs were provided in digital form.
The city also operates a network of recording raingages. Four of these gages were used for the
Wingohocking basin. These are shown on the accompanying map as 1, Roosevelt; 2, Heintz;
3, Queen Lane; and 4, Harrowgate. All of the raingages were of the weighing-bucket type. As
a part of this project the original raingage charts were digitized for 5-minute intervals with the
Water Survey's auto-trol model 3400 X-Y digitizer.
Results. Both peaks and volumes are overestimated by the RRL method for storms occurring
between 1964 and 1967. The runoff ratios for these storms seem to be quite low for a basin
which is 61 percent paved. The method works satisfactorily on the 1968 storms which have
an average runoff ratio of 0.72.
55
-------
Table 10. Storm data and results for Wingohocking basin
Total Basin Area
5326 acres
Total Paved Area
3246 acres
61 percent
Observed Storm
Storm
(1)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
Mean
Date
(2)
42764
42964
92864
112564
71165
80465
80965
82165
92465
100765
70267
72967
80967
61268
72468
80168
values
AMC
(3)
2
3
4
2
2
2
4
3
1
1
3
3
3
4
3
2
Rain
inches
(4)
1.02
1.00
1.26
1.41
2.52
1.02
1.97
1.16
1.22
1.11
1.38
1.20
1.34
3.23
1.68
1.31
Peak
flow
eft
(5)
470
860
1145
1960
1860
789
1960
800
1570
880
2325
1587
2640
5248
3402
3402
Data
Runoff
volume
inches
(6)
0.27
0.37
0.38
0.36
0.89
0.31
0.50
0.31
0.39
0.26
0.70
0.50
0.35
2.72
1.20
0.80
Directly Connected Paved Area
3246 acres
61 percent
Computed
Runoff
ratio
(7)
0.26
0.37
0.31
0.26
0.35
0.30
0.25
0.27
0.32
0.23
0.51
0.41
0.26
0.84
0.71
0.61
0.39
Virtual
runoff
ratio
(8)
0.43
0.61
0.50
0.43
0.58
0.50
0.42
0.43
0.52
0.38
0.83
0.67
0.43
1.38
1.17
1.00
Runoff
volume
inches
(9)
0.55
0.54
0.70
0.79
1.46
0.58
1.13
0.64
0.67
0.61
0.77
0.66
0.74
1.88
0.95
0.72
Error
percent
(10)
107.2
45.9
81.5
116.5
64.2
86.6
124.6
106.9
72.3
137.1
10.6
33.7
112.7
-30.8
-20.5
-9.2
72.5
Results
Peak
flow
cfs
(11)
849
1400
1690
2225
2583
1124
3301
1308
3101
1667
2967
2692
3928
3680
2806
3417
Error
percent
(12)
80.6
62.8
47.6
13.6
38.9
42.5
68.4
63.5
97.5
89.5
27.6
69.6
48.8
-29.9
-17.5
0.5
49.9
Computed peaks were high in 14 cases, average + error = 53.7 percent
Computed peaks were low in 2 cases, average - error = 23.7 percent
Computed runoff volumes were high in 13 cases, average + error = 84.6 percent
Computed runoff volumes were low in 3 cases, average — error = 20.2 percent
Figure 33. Typical row-houses in the Wingohocking basin
56
-------
6000
to
it-
" 4000
U1
s£
«t
UJ
Q_
O
I —
^>
^ 2000
o
o
0
i i i i I X
A. PEAKS /
X
/
X
/
. / -
X
i j
,•
x
r " ' / .
• • / "
i; '
f /
,0 /
• !/
»•••' /
- . •/
/
/ i i i i i
1.2
t/i
O)
i:
o
•- 0.8
^
U-
LL-
o
•z.
ZD
tX
0
LjJ
^ 0.4
O-
?r
O
O
0
1 I 1 It-' i^"i-^;
B. VOLUMES ,,!„, j
( I ,m]
' /
/
/
— •* X ~
u. "X .
• J y 1 f i
, . . /
• t i;' /
lo» / —
r .i, ,
• . /
j /
X
X
X
X
X
- X
X
X
X 1 1 1 1 1
2000 4000
OBSERVED PEAKS, cfs
6000
0.4 0.8 1.2
OBSERVED RUNOFF, inches
5000
4000
3000
S2000
1000
C. HYDROGRAPHS
OBSERVED
I I I I
STORM OF JUNE 12, 1968 —|
0 1
5 6 7
TIME, hours
10 11 12 13 14
Figure 34. RRL results for Wingohocking basin, Philadelphia, Pennsylvania
57
-------
i
... it jn . fj1
'J''-!--
-^5^V^'>"
' ^ 'Av^r ^^ -^
•i r *z%&%-*2^.
* .'- *2*.',~ ^4-^\^^ Y-. -
•Y .^^^f ''• --'\\., •' a!**
v •?
^ WINGOHOCKING BASIN
W$ PHILADELPHIA
«N
PV/.r^^VW-Vfi*-
Figure 35. Street map of the Wingohocking basin
58
-------
SECTION VII
OVERALL RESULTS
General
The results and basic parameters for the 10 basins used in this study are presented in table 11.
Columns 9 and 10 show the number of storms for which the computed peak was higher and
lower, respectively, than the observed peak. Column 11 gives the difference between the ob-
served and computed peaks expressed as a percent of the observed peak and averaged without
regard to sign for all storms in the basin. The last three columns give a similar summary for
the computed runoff.
The 10 basins fall into three classifications with regard to the RRL method. The first includes
the 3 basins on which no significant grassed-area runoff was present and on which the RRL
method does apply. These basins are Woodoak Drive, Tar Branch, and Dry Creek. These
basins have in common flat to moderate slopes and soils in hydrologic group B. The weighted
average errors for all of the storms in this group are 28.9 percent for the peaks and 39.4 per-
cent for the runoff volumes. These percentage values are influenced greatly by large percentage
errors in relatively small peaks and runoff volumes.
In the second class, computed peaks and runoff volumes are much lower than the corresponding
observed values. The RRL method does not apply because significant grassed-area runoff oc-
curs frcm these 5 basins. Included in this group are Ross-Ade, Echo Park, Crane Creek, Tripps
Run, and Third Fork. These basins have in common moderate to steep slopes and soils in
hydrologic groups C or D. The weighted average errors are 40.4 percent for peaks and 39.4
percent for the runoff volumes for the storms on these 5 basins having significant grassed-area
runoff.
The third classification includes 2 basins: Sewer District No. 8 and Wingohocking. For these
basins the RRL overestimates peaks and runoff volumes. Surface ponding is known to be a
problem on both of these basins and would appear to be the reason for the overestimates by
the RRL method. There is no provision in the RRL method for attenuating the surface hydro-
graph because of ponded conditions before the runoff reaches the inlet. Although this is a
shortcoming of the method, it should be pointed out that even if such a procedure were avail-
able the evaluation of these flooded conditions would be subjective at best.
Table 11. General results for the RRL method on all basins
Computed peaks
Basin
(1)
Woodoak Drive
Ross-Ade (Upper)
Sewer Dist No. 8
Echo Park Avenue
Crane Creek
Tripps Run Trib.
Tar Branch
Third Fork
Dry Creek
Wingohocking
Basin
area
acres
(2)
14.7
54.0
206
252
273
322
384
1075
1882
5326
Total paved
area
acres percent
(3) (4)
4.9 33.9
13.3 24.7
43 21.0
136 53.8
65.5 23.9
100 31.0
227 59.0
397 37.0
583 31.0
3246 61.0
Dir. connected
paved area
acres percent
(5) (6)
2.8 19.4
7.8 14.4
37.5 18.2
97.7 38.8
39.7 14.5
56.9 17.7
195 51.0
293 27.0
365 19.0
3246 61.0
Hydro-
logic soil
group
(7)
B
B-C
C-D
B-C
C-D
B-C
B
B-D
B-C
B-D
Basin
slope
(8)
flat
steep
flat
steep
mod
mod
mod
mod
flat
mod
No.
high
(9)
5
0
9
0
5
3
12
3
2
14
No.
low
(10)
5
2
1
18
12
6
5
12
6
2
Mean
absolute
error
percent
(11)
28.2
50.5
68.0
47.2
41.5
37.9
33.4
31.2
20.3
49.9
Computed
No.
high
(12)
5
0
10
2
1
1
13
0
2
13
No.
low
(13)
5
2
0
16
16
8
4
15
6
3
runoff
Mean
absolute
error
percent
(14)
18.3
62.8
21.4
30.3
45.9
44.1
57.3
36.7
27.8
72.5
59
-------
During the evaluation of these results considerable emphasis was placed on the plotted results
and the general shape and fit of the computed and observed hydrographs. It is difficult to
describe objectively the overall fit of the computed hydrographs for all of the storms on a
basin. In a number of basins excellent reproductions of the entire observed hydrograph tend
to be obscured by the average percentage errors reported in table 11.
Overall, it is the authors'judgment that the RRL method is satisfactory for describing runoff
characteristics of an urban basin that is smaller than about 5 square miles, and has a paved
area directly connected to the storm drainage system of at least 15 percent of the basin area.
This applicability is limited to the normal design storms for storm drainage systems, storms
having frequencies from 2 to 20 years.
Rainfall Intensity in Britain and America
The British RRL method was devised in Great Britain and is extensively used today in all of
Britain for the routine design or re-design of storm drainage system. The research described in
this report, however, indicates that for some basins in the United States the storm runoff can-
not be predicted merely by computing the runoff from the paved area of a basin that is direct-
ly connected to the storm drainage system. This difference suggests that the rainfall regime in
the United States may be more severe.
0.6
50
RECURRENCE INTERVAL, years
Figure 36. Frequency curves of 2-hour point rainfall in Great Britain
and for four locations in the United States
60
-------
Figure 36 shows a series of point rainfall frequency curves for 2-hour storms which are impor-
tant for the urban runoff application. The curves show the comparison of British and American
conditions. The curves for the United States were derived from the maps of Technical Publica-
tion 40 of the U.S. Weather Bureau (1961). The Great Britain curve was published by Watkins
(1962) and represents the Bilham formula for interpretation of a wide range of heavy rainfalls
recorded throughout Great Britain.
It is readily noted that the Great Britain curve is lower than all of the curves for the United
States. This may be a significant indication that rainfalls in the United States which are criti-
cal for storm drainage design are greater in amount than those that generally occur in Great
Britain.
Other Urban Hydrologic Models
A number of existing models for the urban rainfall-runoff process are noted here. These accom-
plish a variety of purposes; the RRL method is not a duplication of any of these. The effect of
urban development and drainage improvement on runoff has been shown by James (1965) for
a 43.8 square mile basin in California using the Stanford Watershed Model.
A major storm water management model has been published by Metcalf and Eddy Inc. and
others (1971). This model was devised for the EPA by a triumvirate of Metcalf and Eddy Inc.,
the University of Florida, and Water Resources Engineers, Inc. The 4-volume report describes
the model which accommodates rainfall, runoff, and water quality.
An hydrologic model to accommodate the rainfall-runoff process for an urban basin has been
devised by the St. Paul Metropolitan Sewer Board (1971). The Chicago method of storm drain-
age design has been described by Tholin and Keifer (1960). The University of Cincinnati (1970)
has derived an urban rainfall-runoff-quality model. Brater (1968) provided a better understand-
ing of urban runoff processes.
The value of the total-basin-water-accounting procedure called Hydrocomp Simulation Program-
ming, HSP, has been shown by Crawford (1971) in its application to urban hydrology.
61
-------
SECTION VIM
RE-DESIGN OF A STORM DRAINAGE SYSTEM
One object of this project has been to illustrate how the RRL method could be utilized to re-
design a storm drain system. All of the basins included in this project were inspected as to the
suitability of the RRL method for the re-design of the existing storm drainage system. For
only 3 basins was the RRL method valid, as has been described in the preceding section. Of
these, the Dry Creek basin was omitted because there are no existing storm drainage pipes in
the basin. The remaining 2 basins, Woodoak Drive and Tar Branch, were selected for this
analysis. A relatively good fit was obtained by the RRL method to the observed storms on
both of these basins. Also good information on the existing storm drainage system is available.
Tar Branch basin contains considerable existing artificial storm drainage.
100
80
60
7
O
40
O
20
20 40 60 80
CUMULATIVE PERCENT OF STORM TIME
Figure 37. Time distribution of storm rainfall, median
curve for point rainfall (Huff, 1967)
100
63
-------
In order to re-design a storm drain system, an applicable design storm is required. In this case
use has been made of the map on storm rainfall frequency shown in figure 1. This map shows
that the 5-year 2-hour storm rainfall amount would be 2.25 inches for Woodoak Drive and
2.40 inches for Tar Branch.
Illinois studies have provided considerable information on the time distribution of rainfall
(Huff, 1967). The use of this time distribution in urban studies has been demonstrated by
Stall and others (1970). Further, a complete discussion of the occurrence of storm rainfall and
its many dimensions has been provided by Stall and Huff (1971). Primarily on the basis of
these published papers and the descriptions of the storm rainfall phenomena, a time distribu-
tion for storm rainfall was selected and used in this study. This time distribution (Huff, 1967)
is shown in figure 37. This graph is the median curve for the distribution of storm rainfall with-
in thunderstorms for which the majority of the rainfall occurred during the first quartile of the
storm time. The first quartile storm was shown by Huff to be the most common type in Illi-
nois. Also the curve in figure 37 represents the distribution of storm rainfall at a point.
This time distribution was used to provide a cumulative mass curve for the 2-hour 5-year storm
rainfall amounts for Woodoak Drive and Tar Branch (table 12). Column 1 is the cumulative
storm time at 5-minute intervals from 0 to 120 minutes. These are the incremental values
needed for the distribution of the design rainfall for both basins and they are converted into
percent of storm time in column 2. The values from the curve in figure 37 are read and given
in column 3 as cumulative amount of storm rainfall. These cumulative amounts are then ap-
plied to the rainfall totals for the two basins. The last value in columns 4 and 5 (2.25 and
Table 12. The 5-year 2-hour design storms for Woodoak Drive basin,
Long Island, and Tar Branch basin, Winston-Salem
Cumulative
storm
time
minutes percent
(1) (2)
0
5
10
15
20
25
30
35
40
45
50
55
60
65
70
75
80
85
90
95
100
105
110
115
120
4.2
8.3
12.5
16.7
20.8
25.0
29.2
33.3
37.5
41.7
45.8
50.0
54.2
58.3
62.5
66.7
70.8
75.0
79.2
83.3
87.5
91.7
95.8
100
Cumulative
percent of
storm
rainfall
(3)
10
21
33
44
52
59
64
68
72
75
78
80
82
84
85
87
88
90
92
94
95
97
98
100
Design rainfall
Woodoak Tar Branch
inches inches
(4) (5)
0
0.22
0.47
0.74
0.99
1.17
1.33
1.44
1.53
1.62
1.69
1.76
1.80
1.84
1.89
1.91
1.96
1.98
2.02
2.07
2.11
2.14
2.18
2.21
2.25
0
0.24
0.50
0.79
1.06
1.25
1.42
1.54
1.63
1.73
1.80
1.87
1.92
1.97
2.02
2.04
2.04
2.11
2.16
2.21
2.26
2.28
2.33
2.35
2.40
64
-------
2.40 inches) is the design amount for the basin.
In order to re-design the storm drainage system for these two basins, the two design storms
(table 12) were applied to the basins with the existing pipes and storm drainage system. A de-
sign version of the RRL program was used. In cases where a pipe section was not large enough
to accommodate the design flow, the computer automatically specified one size larger pipe and
recomputed that section until an adequate pipe size was obtained.
This process was carried out for the Woodoak Drive basin, which is a small basin with only one
section of storm drain pipe. This one pipe was found to be adequate to carry the flow from the
5-year design storm, so no re-design was needed or accomplished.
For the Tar Branch basin, the re-design process did provide considerable results. Figure 38
shows a street map of a part of the Tar Branch basin identifying the various reaches (upstream
reach 1-0 to downstream reach 1-8) of existing storm drains. Table 13 gives the results of the
re-design of these reaches. If the design peak flow in column 6 exceeds the capacity of the ex-
isting pipe in column 5, the ratio of the two flows, called the surcharge ratio, is given in col-
umn 7.
Table 13 shows that the design flow exceeded that of the existing pipe for 8 of the 26 reaches
of pipe in the Tar Branch basin. For these 8 pipe sections the design program was written to
select the next available pipe size larger than the existing pipe. This pipe was tested for its
capacity to carry the design flow. If not large enough, a still larger pipe was selected until one
adequate to carry the design flow was found. The re-designed pipe sections and their capac-
Table 13. Results of the RRL method re-design of the storm
drainage system for Tar Branch basin
Physical data
Existing pipe
5-year design
Required pipe
Reach
(1)
1-0
1-1
1-2
2-0
3-0
3-1
1-3
4-0
4-1
1-4
5-0
1-5
1-6
6-0
6-1
6-2
6-3
6-4
1-7
7-0
7-1
8-0
8-1
8-2
1-8
Length
feet
(2)
330
200
250
259
260
410
238
240
300
430
300
350
400
350
300
390
325
300
150
410
360
240
350
260
500
Slope
percent
(3)
2.4
2.5
4.8
4.7
4.7
4.7
3.5
4.5
2.0
5.0
9.5
5.0
5.0
5.0
9.5
3.5
7.0
7.0
0.9
6.5
8.5
7.5
2.5
5.0
2.5
Diam
inches
(4)
18
18
18
10
18
24
24
18
18
open
15
36
36
18
18
24
24
open
open
18
18
18
18
18
60
Capacity
eft
(5)
14.1
14.4
19.9
4.1
19.7
42.5
36.6
19.3
12.9
channel
17.2
129.2
129.2
20.3
28.0
36.6
51.8
channel
channel
23.2
26.5
24.9
14.4
20.3
356.7
Design
peak
eft
(6)
15.0
21.1
26.2
3.5
9.9
14.3
57.2
8.6
15.7
86.9
6.3
99.8
103.6
15.2
32.7
47.9
63.0
64.7
167.0
13.4
15.1
9.7
11.2
13.4
202.6
Surcharge
ratio
(7)
1.06
1.47
1.32
1.56
1.22
1.17
1.31
1.22
Diam
inches
(8)
21
21
21
30
21
33
21
27
27
27
57
Capacity
eft
(9)
21.3
21.7
30.1
66.4
19.4
102.0
42.3
50.2
71.0
71.0
187.0
65
-------
X
Figure 38. Reach identifications for a portion of Tar Branch Basin, Winston-Sal em
66
-------
ities for the 8 pipes are given in columns 8 and 9 of table 13.
The results in table 13 can be further interpreted as to the degree that the pipe system is in-
adequate. For example, for reach 1-0 the surcharge ratio is 1.06 meaning the design flow is
only 6 percent greater than pipe capacity. For reach 1-3, however, the surcharge ratio is 1.56
which depicts a relatively important deficiency. The re-design suggests that a 30-inch pipe is
needed to replace the 24-inch pipe in the existing system. The open channels shown on three
reaches have been replaced in the re-design by appropriate circular pipes.
67
-------
SECTION IX
ACKNOWLEDGMENTS
Michael L. Terstriep, Associate Engineer of the Illinois State Water Survey, has been the prin-
cipal professional person carrying out this project, and John B. Stall, Engineer, has been direc-
tor. The work has been carried out as a part of their regular work in the Water Survey Hydro-
logy Section, H. F. Smith, Head. The entire work has been under the general supervision of
Dr. William C. Ackermann, Chief. Robert A. Sinclair, Systems Analyst, provided the program-
ming expertise for this project. The drafting has been carried out by John W. Brother, Jr., and
William Motherway, Jr., of the Research Support Group. Mrs. J. Loreena Ivens, Technical Ed-
itor, reviewed the final report and contributed much to its value. John Lichter worked on the
project as an Engineering Assistant part-time while a student in Sanitary Engineering at the
University of Illinois. The total effort of the Water Survey comprised 70 percent of the project
budget.
An amount equivalent to 30 percent of the budget of this project was provided to the Illinois
State Water Survey as a Research and Development Grant from the federal Environmental
Protection Agency, designated as Project 11030 FLN. During most of the project, George A.
Kirkpatrick was Project Officer of the EPA, a role later assigned to Harry C. Torno. The aid
and counsel of both these Project Officers is gratefully acknowledged.
Help and advice to the project personnel were provided by L. Scott Tucker of the American
Society of Civil Engineers, who had earlier compiled a report (Tucker, 1969) listing various
sources of rainfall-runoff data on urban basins.
A considerable amount of the data used in this project was obtained from various offices of
the U.S. Geological Survey. The data used here were usually of the type collected by the
USGS for their own analysis for other projects.
The data for the Woodoak basin near Westbury, New York, were provided by the USGS. In a
paper by Seaburn (1970) the instrumentation was described, as well as the results of urbaniza-
tion on runoff. Gerald E. Seaburn was helpful in providing the data and taking photos of the
basin.
Data for the Ross-Ade basin were provided by Professor Jacques W. Delleur, Professor of Hy-
draulic Engineering in the Civil Engineering Department at Purdue University, and Dr. Rama-
chandra A. Rao, Associate Professor.
For the basin of Sewer District No. 8 in Bucyrus, Ohio, the primary information was obtained
from Richard Noland of the consulting engineering firm of Burgess and Niple, Ltd. of Colum-
bus, Ohio. The firm had earlier provided a special report on this basin for the federal EPA;
see Burgess and Niple, Ltd. (1969).
Data for the Echo Park Avenue basin in Los Angeles were provided by the Department of Pub-
lic Works, Mr. Lyall A. Pardee, City Engineer. Special help and advice were provided by Irving
R. Cole, Head of the Division of Storm Drainage Design in the Bureau of Engineering, and by
Walter R. Naydo of the same Division.
Data for the Crane Creek basin were provided by personnel of the USGS District Office in
Jackson. The special help of Kenneth V. Wilson and James Hudson is especially acknowledged.
Data for the Tripps Run basin in Falls Church were provided by the Fairfax Sub-District Office
of the USGS with the special help of Pat L. Soule.
69
-------
Data for the Tar Branch basin in Winston-Salem and the Third Fork basin in Durham were pro-
vided by personnel of the Durham District Office of the USGS with the special help of Arthur
L. Putnam. For the Third Fork basin, information on existing storm drainage was provided
by Larry S. Kerr, Street Engineer, City of Durham.
The data for Dry Creek basin in Wichita were provided by David Richards of the USGS Dis-
trict Office in Lincoln, Nebraska, and the personnel of the Sub-District Office in Wichita.
Data for the Wingohocking basin in Philadelphia were provided by personnel of the Research
Division of the Water Department, City of Philadelphia, Joseph V. Radziul, Chief; William
Green, of the Planning Division also provided help.
70
-------
SECTION X
REFERENCES
American Public Works Association, Urban Drainage Practices, Procedures, and Needs, Special
Report 13, Chicago (1966).
American Society of Civil Engineers, Water and Metropolitan Man, New York (1969).
Anderson, Daniel G., Effects of Urban Development on Floods — Water in the Urban Environ-
ment, U. S. Geological Survey Water Supply Paper 2001-C (1970).
Ardis, Colby V., Kenneth J. Dueker, and Arno T. Lenz, "Storm Drainage Practices of Thirty-
two Cities," ASCE Journal of the Hydraulics Division, 95, No. HY1, (January 1969).
Brater, E. F., "Steps Toward a Better Understanding of Urban Runoff Processes," Water Re-
sources Research, 4, No. 2, pp 335-347 (April 1968).
Burgess and Niple, Ltd., Stream Pollution and Abatement from Combined Sewer Overflows,
Bucyrus, Ohio, Federal Environmental Protection Agency Water Pollution Control Series
11024 FKN (1969).
Chow, Ven Te, Handbook of Applied Hydrology, McGraw-Hill Book Co., Inc., New York,
pp 14-17 (1964).
Crawford, N. H., Studies in the Application of Digital Simulation to Urban Hydrology, U. S.
Office of Water Resources Research Contract Report 14-31-0001-3375, Washington, D.C.,
100 pp (1971).
Hicks, W. I., "A Method of Computing Urban Runoff," American Society of Civil Engineers
Transactions, 109, pp 1217-1253 (1944).
Huff, F. A., "Time Distribution of Rainfall in Heavy Storms," Water Resources Research, 3,
No. 4, pp 1007-1019 (1967).
James, L. Douglas, "Using a Digital Computer to Estimate the Effects of Urban Development
on Flood Peaks," Water Resources Research, 1, No. 2, (1965).
Jens, Stifel W., and W. B. McPherson, "Hydrology of Urban Areas," Section 20 in Handbook
of Applied Hydrology, edited by Ven Te Chow, McGraw-Hill Book Co., Inc., New York (1964).
Linsley, Ray K., A Critical Review of Currently Available Hydrologic Models for Analysis of
Urban Stormwater Runoff, Office of Water Resources Research, Washington, D.C. (1971).
Metcalf and Eddy, Inc. University of Florida, and Water Resources Engineers, Inc., Storm
Water Management Model, 4 volumes, Federal Environmental Protection Agency Water Pollu-
tion Control Research Series 11024DOC (1971).
71
-------
Road Research Laboratory, A Guide for Engineers to the Design of Storm Sewer Systems,
Road Note 35, London (1963).
Seaburn, G. E., Preliminary Results ofHydrologic Studies at Two Recharge Basins on Long
Island, New York, U. S. Geological Survey Professional Paper 627-C, 17 pp (1970).
St. Paul (Minnesota) Metropolitan Sewer Board, Dispatching System for Control of Combined
Sewer Losses, Part II, Mathematical Model, Federal Environmental Protection Agency Water
Pollution Control Research Series 11020FAQ, p 121 (1971).
Stall, John B., and Floyd A. Huff, The Structure of Thunderstorm Rainfall, Illinois State Water
Survey Reprint 165, Urbana, 30 pp (1971).
Stall, John B., Michael L. Terstriep, and F. A. Huff, Some Effects of Urbanization on Floods,
Illinois State Water Survey Reprint 133, Urbana (1970).
Terstriep, Michael L., and John B. Stall, "Urban Runoff by the Road Research Laboratory
Method," ASCE Journal of the Hydraulics Division, 95, No. HY6, pp 1809-1834 (November
1969).
Tholin, A. L., and C. J. Keifer, "The Hydrology of Urban Runoff," American Society of Civil
Engineers, Transactions, 125, p 1308 (1960).
Tucker, L. Scott, Availability of Rainfall-Runoff Data for Sewered Catchments, American
Society of Civil Engineers Urban Water Resources Research Program, Technical Memorandum
No. 8, New York, 43 pp (1969).
U. S. Weather Bureau, Rainfall Frequency Atlas of the United States, Technical Paper No. 40,
Washington, D.C. (1961).
University of Cincinnati, Urban Runoff Characteristics, Federal Environmental Protection
Agency Water Pollution Control Research Series 11024DQU (1970).
Watkins, L. H., The Design of Urban Sewer Systems, Road Research Technical Paper No. 55,
Dept. of Scientific and Industrial Research, London, Her Majesty's Stationery Office (1962).
Wilson, K. V., A Preliminary Study of the Effect of Urbanization on Floods in Jackson, Mis-
sissippi, U. S. Geological Survey Professional Paper 575D, pp 259-261, Washington, D.C.
(1968).
72
-------
SECTION XI
GLOSSARY OF TERMS
Design Area — The paved portion of a basin from which runoff water can reach the gage with-
out first passing over grassed area. This term is synonymous with the directly connected paved
area.
Design Storm — A rainfall pattern of specified amount, intensity, duration, and frequency.
Directly Connected Paved Area — The paved portion of a basin from which runoff water can
reach the gage without first passing over grassed area.
Entry Time — The time in minutes for runoff water to flow from the most remote point on
the directly connected paved area to a specified inlet.
Reach — The smallest subdivision of the drainage system consisting of a uniform length of
open channel or underground conduit.
Sub-Basin — A physical division of a larger basin which is associated with one reach of the
storm drainage system.
73
-------
SELECTED WATER
RESOURCES ABSTRACTS
INPUT TRANSACTION FORM
1. Report No.
2.
3. Accession No.
w
4- Title 5. Report Date
STORM SEWER DESIGN — AN EVALUATION OP THE RRL «•
METHOD
7. Author(s)
John B. Stall and Michael L. Terstriep
9. Organization
Illinois University, Urbana, Illinois 61801,
Illinois State Water Survey
8. Performing Organization
Report No.
10. Project No.
EPA 11030 FLN
12. Sponsoring Organization
IS. Supplementary Notes
11. Contract/ Grant No.
13. Type of Report and
Period Covered
Environmental Protection Agency report
number EPA-R2-72-068, October 1972.
16. Abstract
Storm rainfall and runoff data were assembled from 10 urban basins
in the U.S.A. ranging in size from 14 acres to 8 sq mi. The British
RRL method of storm drainage design was applied to the 10 basins.
The RRL method considers the urban basin to be comprised of the paved
area of the basin which is directly connected to the artificial storm
drainage system. In 3 of the 10 basins the RRL procedure was deemed
to be appropriate and suitable for the design of a storm drainage
system within the normal range of frequency of design rainfall events
from 2 to 20-year events. For greater storms and for certain cases
within this frequency range, the RRL method breaks down because the
runoff coming from the grassed area of the basin is significant. If
the basin is highly steep or if the paved area comprises less than
15% of the total basin, this breakdown occurs. An example is given
of the use of the RRL method in the re-design of an existing storm
drainage system, as is current practice in Great Britain.
17a. Descriptors
*Storm drains, *storm runoff, rainfall-runoff relationships, closed
conduit flow, sewers, hydrographs, surface drainage, peak discharge,
surface runoff, urbanization, design flow.
17b. Identifiers
ITc.CO WRR Field & Group Q 2 E
18. Availability
19. Security Class.
(Report)
20. Security Class.
(Page)
21. No. of
Pages
22. Price
Institution
Send To:
WATER RESOURCES SCIENTIFIC INFORMATION CENTER
U.S. DEPARTMENT OF THE INTERIOR
WASHINGTON, D. C. 20240
GPO 913.2gl
------- |