United States EPA/600/R-09/077
Environmental Protection Agency July 2009
STORM WATER MANAGEMENT MODEL
APPLICATIONS MANUAL
By
Jorge Gironas
Larry A. Roesner
Jennifer Davis
Department of Civil and Environmental Engineering
Colorado State University
Fort Collins, CO 80523-1372
Project Officer
Lewis A. Rossman
Water Supply and Water Resources Division
National Risk Management Research Laboratory
Cincinnati, OH 45268
NATIONAL RISK MANAGEMENT RESEARCH LABORATORY
OFFICE OF RESEARCH AND DEVELOPMENT
U.S. ENVIRONMENTAL PROTECTION AGENCY
CINCINNATI, OH 45268
-------
DISCLAIMER
The information in this document has been funded wholly or in part by the U.S.
Environmental Protection Agency (EPA). It has been subjected to the Agency's peer
and administrative review, and has been approved for publication as an EPA document.
Mention of trade names or commercial products does not constitute endorsement or
recommendation for use.
-------
FOREWORD
The U.S. Environmental Protection Agency is charged by Congress with protecting the
Nation's land, air, and water resources. Under a mandate of national environmental
laws, the Agency strives to formulate and implement actions leading to a compatible
balance between human activities and the ability of natural systems to support and
nurture life. To meet this mandate, EPA's research program is providing data and
technical support for solving environmental problems today and building a science
knowledge base necessary to manage our ecological resources wisely, understand how
pollutants affect our health, and prevent or reduce environmental risks in the future.
The National Risk Management Research Laboratory is the Agency's center for
investigation of technological and management approaches for reducing risks from
threats to human health and the environment. The focus of the Laboratory's research
program is on methods for the prevention and control of pollution to the air, land,
water, and subsurface resources; protection of water quality in public water systems;
remediation of contaminated sites and ground water; and prevention and control of
indoor air pollution. The goal of this research effort is to catalyze development and
implementation of innovative, cost-effective environmental technologies; develop
scientific and engineering information needed by EPA to support regulatory and policy
decisions; and provide technical support and information transfer to ensure effective
implementation of environmental regulations and strategies.
Water quality impairment due to runoff from urban and developing areas continues to
be a major threat to the ecological health of our nation's waterways. The EPA
Stormwater Management Model is a computer program that can assess the impacts of
such runoff and evaluate the effectiveness of mitigation strategies. This manual
presents a number of worked-out examples that shows new users how to set up and
apply SWMM to the most common types of Stormwater management and design
problems encountered in practice.
Sally C. Gutierrez, Director
National Risk Management Research Laboratory
MI
-------
ABSTRACT
The EPA Storm Water Management Model (SWMM) is a dynamic rainfall-runoff
simulation model that computes runoff quantity and quality from primarily urban areas.
This manual is a practical application guide for new SWMM users who have already
had some previous training in hydrology and hydraulics. It contains nine worked-out
examples that illustrate how SWMM can be used to model some of the most common
types of stormwater management and design problems encountered in practice. These
include: computing runoff for both pre- and post development conditions; analyzing the
hydraulics of simple collection systems; designing a multi-purpose detention pond;
modeling distributed low impact runoff controls; simulating the buildup, washoff,
transport and treatment of stormwater pollutants; analyzing both dual drainage and
combined sewer systems; and running long-term continuous simulations. Each example
is accompanied by a complete SWMM input data file.
IV
-------
Contents
Foreword iii
Abstract iv
Figures vi
Tables ix
Acknowledgment xi
Introduction 12
Example 1. Post-Development Runoff 15
Example 2. Surface Drainage Hydraulics 35
Example 3. Detention Pond Design 50
Example 4. Low Impact Development 69
Example 5. Runoff Water Quality 85
Example 6. Runoff Treatment 103
Example 7. Dual Drainage Systems 117
Example 8. Combined Sewer Systems 132
Example 9. Continuous Simulation 154
References 177
-------
Figures
Figure 1-1. Undeveloped site 16
Figure 1-2. Developed site 16
Figure 1-3. Idealized representation of a subcatchment 17
Figure 1 -4. SWMM representation of the undeveloped study area 21
Figure 1 -5. Computation of width of the undeveloped subcatchment 22
Figure 1-6. Design storm hyetographs 23
Figure 1-7. Runoff hydrographs (flow Q versus time) for the undeveloped site 27
Figure 1-8. Discretization of the developed site into subcatchments 28
Figure 1-9. Definition of overland flow length and slope for subcatchment S2 29
Figure 1-10. Width and slope computation for subcatchments S3 (a) and S4 (b) 30
Figure 1-11. Runoff hydrographs (flow Q versus time) for the developed site 32
Figure 2-1. Pre-development site 36
Figure 2-2. Post-development site without conveyance system 37
Figure 2-3. Links and Nodes 37
Figure 2-4. Post-development site with runoff conveyance added 40
Figure 2-5. SWMM representation of the post-development conveyance system 41
Figure 2-6. Post-development outflow hydrographs for 2-yr storm 46
Figure 2-7. Post-development outflow hydrographs for 10-yr storm 47
Figure 2-8. Post-development outflow hydrographs for 100-yr storm 47
Figure 3-1. Detention pond location for the post-development site 51
Figure 3-2. Schematic representation of a detention pond 55
Figure 3-3. Water quality capture volume (UDFCD, 2001) 55
Figure 3-4. Average depth (inches) of runoff producing storms in the US
(Driscoll, et al., 1989) 57
Figure 3-5. Geometry of the pond's WQCV 57
Figure 3-6. Study area map with storage unit SU1 59
Figure 3-7. Properties of storage unit SU1 59
Figure 3-8. WQCV drainage times for the iterations shown in Table 3-3 61
Figure 3-9. Final design of the detention pond and outlet structure 65
Figure 3-10. Detail of the pond outlet structure 65
Figure 3-11. Outlet hydrographs for the 2-yr storm 67
Figure 3-12. Outlet hydrographs for the 10-yr storm 67
Figure 3-13. Outlet hydrographs for the 100-yr storm 68
Figure 4-1. Post-development site with LIDs in place 70
Figure 4-2. Schematic representation of a filter strip 71
Figure 4-3. Schematic representation of an infiltration trench 72
Figure 4-4. Re-discretization of subcatchments S3 and S4 75
Figure 4-5. Representation of subcatchments S_FS_1, S3.1, and S3.2 76
Figure 4-6. Representation of subcatchments S_FS_3, S_FS_4, S4.1, S4.2, and S4.3 . 76
Figure 4-7. Re-discretization of subcatchments SI and S2 for infiltration trenches 78
VI
-------
Figure 4-8. Influent and effluent hydrographs for filter strip S_FS_J 81
Figure 4-9. Influent and effluent hydrographs for infiltration trench S IT 1 81
Figure 4-10. Comparison of outlet discharges with and without LID controls 83
Figure 4-11. Percent reduction in outlet peak flows and runoff volumes with LIDs .... 83
Figure 5-1. Post-development site with no runoff controls 86
Figure 5-2. TSS buildup curve 92
Figure 5-3. TSS concentrations for the 0.1 in. storm with EMC washoff 96
Figure 5-4. TSS concentrations for the 0.23 in. storm with EMC washoff 96
Figure 5-5. TSS concentration at site outlet with EMC washoff 97
Figure 5-6. Runoff flow and TSS load at site outlet for the 0.1 in. storm with EMC
washoff 98
Figure 5-7. TSS concentrations for the 0.1 in. storm with Exponential washoff 99
Figure 5-8. TSS concentrations for the 0.23 storm with Exponential washoff. 99
Figure 5-9. TSS concentrations for the 2-yr (1 in.) storm with Exponential washoff. 100
Figure 5-10. TSS concentration at the site outlet for Exponential washoff 100
Figure 6-1. Developed site with LIDs and detention pond 104
Figure 6-2. Schematic of the re-designed pond outlet structure 105
Figure 6-3. TSS and runoff reduction through filter strip S_FS_J for the 0.1 in.
storm 110
Figure 6-4. TSS and runoff reduction through infiltration trench S IT 4 for the 2-yr
storm Ill
Figure 6-5. TSS concentrations in pond SU2 with and without treatment
(k = 0.01 ft/hi) 112
Figure 6-6. TSS mass load released by pond SU2 for the 0.23 in storm
(ฃ = 0.01 ft/hi) 113
Figure 6-7. TSS concentrations in pond SU2 with and without treatment
(ฃ = 0.3 ft/hi) 114
Figure 6-8. TSS mass load released by pond SU2 for the 0.23 in. storm
(ฃ = 0.3 ft/hi) 114
Figure 6-9. Total TSS load discharged at site outlet under different treatment
scenarios 116
Figure 7-1. Post-development site with simple drainage system 118
Figure 7-2. Parallel pipe and gutter conveyance 118
Figure 7-3. Elements of streets defined in the drainage criteria 120
Figure 7-4. Three-dimensional layout of the site's dual drainage system 122
Figure 7-5. Full Street cross section. The Half Street section is half of this section. 122
Figure 7-6. Post-development site with dual drainage system 125
Figure 7-7. Examples of surcharge and flooding 127
Figure 7-8. Flows in pipes P5 and P6 during the 100-yr storm 130
Figure 7-9. Surcharge behavior along the eastern boundary of the site 131
VII
-------
Figure 8-1. Combined sewer system study area 133
Figure 8-2. Conceptual representation of overflows in a combined system
(Field and Tafuri, 1973) 134
Figure 8-3. Transverse weir flow regulator 136
Figure 8-4. Alternative ways to represent trnaverse weir flow regulators in SWMM 136
Figure 8-5. (a) Pump Property Editor and (b) Pump Curve Editor 137
Figure 8-6. Schematic representation of the combined sewer system 138
Figure 8-7. Layout of the combined sewer system before adding regulators and
pump 141
Figure 8-8. Force main line 143
Figure 8-9. Final combined sewer system model layout 146
Figure 8-10. Flow (Q) along sections of the stream for the 0.23 in. storm 147
Figure 8-11. Flow (Q) along sections of the interceptor for the 0.23 in. storm 148
Figure 8-12. Flow (Q) diverted from each regulator to the interceptor for the
0.23 in. storm 148
Figure 8-13. Flow (Q) along sections of the stream for the 2-yr storm 149
Figure 8-14. Flow (Q) along the sections of the interceptor for the 2-yr storm 149
Figure 8-15. Flows (Q) in regulators Rl and R4 for the 2-yr storm 150
Figure 8-16. Pump behavior for the 0.23 in. storm: (a) wet well water depth,
(b)pump flow 151
Figure 9-1. Drainage system and detention pond (SU1) designed in Example 3 155
Figure 9-2. Ten year rainfall record for Fort Collins, CO (Source: National
Climatic Data Center) 160
Figure 9-3. Rainfall event producing the largest flow rates to the detention pond 162
Figure 9-4. Water depth in the detention pond between days 3061 and 3069 163
Figure 9-5. Inflow and outflow (Q) for the detention pond between days 3061
and 3069 164
Figure 9-6. Losses for subcatchment 57 between days 3061 and 3069 165
Figure 9-7. Statistics Selection dialog for analyzing peak pond outflows 166
Figure 9-8. Summary statistics for peak inflow to node J out (same as peak
pond outflow) 167
Figure 9-9. Event listing of peak inflow to node J out (same as peak pond outflow) 167
Figure 9-10. Histogram of peak inflow to node J_out (same as peak pond outflow).. 168
Figure 9-11. Cumulative frequency of peak inflow to node J out (same as peak
pond outflow) 168
Figure 9-12. Statistical query for daily peak water depth in the detention pond 170
Figure 9-13. Frequency plot of daily peak water depth in the detention pond 170
Figure 9-14. Selection of anMTTfor analyzing a rainfall record 172
Figure 9-15. Frequency plots for event duration and depth 173
VIM
-------
Tables
Table 1-1. Properties of the undeveloped subcatchment 22
Table 1-2. Flow lengths and slopes of the undeveloped subcatchment 23
Table 1-3. Analysis options 25
Table 1-4. Summary of results for the undeveloped site 27
Table 1-5. Geometric properties of the subcatchments in the developed site 29
Table 1-6. Land use categories in the developed site 31
Table 1-7. Land use coveage (ac) and imperviousness for subcatchments in the
developed site 31
Table 1-8. Summary of results for post-development conditions 33
Table 1-9. Comparison of runoff for pre- and post-development conditions 33
Table 2-1. Invert elevations of junctions 41
Table 2-2. Subcatchment outlets 42
Table 2-3. Characteristics of the conduits used in the model 43
Table 2-4. Conduit properties 43
Table 2-5. Available culvert sizes 44
Table 2-6. Maximum depths and flows for conduits during the 100-yr event 45
Table 2-7. Comparison of runoff for post-development conditions with and without
routing 48
Table 3-1. Pre- and post-development peak discharges 50
Table 3-2. Post-development subcatchment data 56
Table 3 -3. Design of the WQCV outlet (Orl} 60
Table 3-4. Characteristics of the pond's outlet structure 66
Table 4-1. Subcatchments containing filter strips 72
Table 4-2. Properties of the new junctions 74
Table 4-3. Properties of the new conduits 74
Table 4-4. Properties of the subcatchments derived from S3 and S4 75
Table 4-5. Widths of the filter strip subcatchments 76
Table 4-6. Properties of the filter strip subcatchments 77
Table 4-7. Properties of the subcatchments derived from SI and S2 78
Table 4-8. Subcatchments containing infiltration trenches 79
Table 4-9. Properties of the infiltration trench subcatchments 79
Table 4-10. Properties of the new junctions 80
Table 4-11. Properties of the new conduits 80
Table 4-12. Runoff coefficients for filter strips 82
Table 4-13. Runoff coefficients for infiltration trenches 82
Table 5-1. Rainfall time series for the 0.1 and 0.23 inch events 86
Table 5-2. Typical dust and dirt buildup rates after Manning et al. (1977) 91
Table 5-3. Parameters for TSS buildup 92
Table 5-4. Curb length and land uses for each subcatchment 92
Table 5-5. Washoff characteristics for each land use 94
IX
-------
Table 5-6. Average TSS concentration for the 0.23 in. event 101
Table 6-1. Storage curve for the re-designed pond 104
Table 6-2. Properties of the pond's re-designed outlet structure 105
Table 6-3. Curb lengths and land uses for LID subcatchments 108
Table 6-4. Detention pond TSS treatment performance summary 115
Table 7-1. General drainage system criteria for Fort Collins (City of Fort Collins,
1984 and 1997) 121
Table 7-2. Cross section data for street transects 124
Table 7-3. Junction invert elevations for dual drainage system 126
Table 7-4. Conduit shapes and offsets for dual drainage system 126
Table 7-5. Calculated gutter grades 128
Table 7-6. Available drainage pipe sizes 129
Table 7-7. Iterations for 2-yr storm pipe sizing 129
Table 8-1. Properties of the combined sewer system nodes1 139
Table 8-2. Properties of the combined sewer system conduits1 140
Table 8-3. Subcatchment outlets 140
Table 8-4. Properties of the regulator structures 143
Table 8-5. Properties of the force main line 144
Table 8-6. Pump curve data 145
Table 8-7. Summary of dry weather flows 146
Table 8-8. Summary of system performance for different storm events 152
Table 9-1. Monthly evaporation for Fort Collins, CO (Source: Western Regional
Climate Center) 159
Table 9-2. Summary statistics for various rainfall event properties 173
Table 9-3. Ten most severe events based on duration, depth and intensity 174
Table 9-4. Correspondence among the most severe events 175
-------
ACKNOWLEDGEMENT
The authors would like to acknowledge the active participation of EPA's Project
Officer, Lewis Rossman in the development of this Manual. Dr. Rossman participated
in the development of all examples contained in the manual, reviewed and commented
on the model results and participated in the write-up of the examples. His active
participation in this project has contributed to the high quality of the resulting Manual.
XI
-------
Introduction
The EPA Storm Water Management Model (SWMM) is a dynamic rainfall-
runoff simulation model that computes runoff quantity and quality from primarily
urban areas. The runoff component of SWMM operates on a collection of
subcatchment areas that receive precipitation and generate runoff and pollutant loads.
The routing portion of SWMM transports this runoff through a system of pipes,
channels, storage/treatment devices, pumps and regulators. SWMM tracks the quantity
and quality of runoff generated within each subcatchment and the flow rate, flow depth,
and quality of water in each pipe and channel during a simulation period comprised of
multiple time steps.
SWMM was first developed in 1971 and since then has undergone several
major upgrades. It continues to be widely used throughout the world for planning,
analysis, and design related to stormwater runoff, combined sewers, sanitary sewers,
and other drainage systems in urban areas and has also been used for modeling non-
urban areas. The most current implementation of the model is version 5.0 which was
released in 2005. It has modernized both the model's structure and its user interface,
making SWMM easier to use and more accessible to a new generation of hydrologists,
engineers, and water resources management specialists.
OSWMM 5 - Examplel post.inp
I-INฎ
File Edit
D a^
Report Tools Window Help
Data Map
;'-. Study Area Map
Title/Notes
Options
Climatology
+: Hydrology
E6 Hydraulics
E Quality
it; Curves
Time Series
Time Patterns
k.j -.-1 -.u-i
H & 9 H O V O t=J
00ฎ
Dates
Time Steps
Dynamic Wave
Interface Files
Auto-Length: Off * Flow Units: CFS " Offsets: Depth ~
Zoom Level: 100?i X,Y: 620.042,1447.757
12
-------
The objective of this manual is to serve as a practical application guide for new
SWMM users who have already had some previous training in hydrology and
hydraulics. It contains nine worked-out examples that illustrate how SWMM can be
used to model some of the most common types of stormwater management and design
problems encountered in practice. The following applications are discussed in the
manual:
1. Post-Development Runoff. Surface runoff from a 29 acre residential site is
computed for several design-storm events under both pre- and post-
development conditions.
2. Surface Drainage Hydraulics. A surface runoff conveyance network is added to
the post-development catchment area of Example 1 and is analyzed with
SWMM's various hydraulic routing options.
3. Detention Pond Design. A detention pond and outlet structure are designed for
the post-development condition of Example 2 that meets both water quality
control and peak flow reduction criteria.
4. Low Impact Development. Two typical Low Impact Development (LID)
controls, filter strips and infiltration trenches, are placed within the post-
development catchment modeled in Example 2.
5. Runoff Water Quality. The buildup, washoff and routing of total suspended
solids (TSS) is simulated within the post-development catchment modeled in
Example 2.
6. Runoff Treatment. The removal of TSS in the LID controls and detention pond
of Examples 3 and 4 is modeled.
7. Dual Drainage Systems. The surface drainage system of Example 2 is converted
into a parallel system of below ground storm sewers and above ground streets
and gutters that is subjected to both surcharged flow and street flooding.
8. Combined Sewer Systems. The sewer system of Example 7 is converted into a
combined system that carries both dry weather sanitary flow and wet weather
runoff and includes various flow diversion structures and a pumped force main.
9. Continuous Simulation. SWMM's statistical tools are used to analyze the
performance of the detention pond designed in Example 3 over a continuous ten
year period of historical rainfall.
All the examples are developed for the same catchment area and each one
builds in some degree on the results of a previous example. Therefore, it is
recommended that the reader begin with Example 1 and work sequentially through
Example 9, while hopefully building the required input data files and running them
with SWMM for each example. They can then compare the input files they build with
the files created by the authors. These files, as well as the backdrop image files that are
needed to complete the examples, are available in a compressed file named
epaswmm5_apps_manual.zip. It can be downloaded from the EPA SWMM web site
at: http://www.epa.gov/endnrmrl/models/swmm.
13
-------
This manual assumes that readers have a basic knowledge of how to run EPA
SWMM 5 and perform such functions as opening a new project, setting project
defaults, adding drainage system objects to a project, editing the properties of these
objects, and viewing simulation results. These topics and more are covered in the
SWMM Users Manual which is also available from the EPA SWMM web site. That
manual includes a tutorial example that leads new users through each of these steps.
14
-------
Example 1. Post-Development Runoff
This first example demonstrates how to construct a hydrologic model of an urban
catchment and use it to compare stormwater runoff under both pre- and post-development
conditions. It illustrates the process of spatially dividing a catchment into smaller
computational units, called sub catchments, and discusses the characteristics of these
subcatchments that SWMM uses to transform rainfall into runoff. This example considers
runoff only. Flow routing of runoff through the drainage pipes and channels contained
within the catchment is addressed in Example 2.
Models of this type are very common in practice. Many local stormwater
ordinances and agencies require that new developments limit peak runoff flows relative
to those under pre-development conditions. To meet environmental sustainability
objectives, similar criteria are being applied to total runoff volume as well.
1.1 Problem Statement
Figure 1-1 is a contour map of a 29 acre natural catchment area where a new
residential development is planned. This undeveloped area is primarily pasture land that
has a silt loam soil type. Figure 1-2 shows the proposed development for this site. With
the exception of the depressions located in the parkland area, no major changes in
topography are expected. This implies that future streets will, in general, follow the
natural slope. However, the residential lots will be graded toward the street at a slope of
2% so they can drain easily. The developed site will drain to a stream through a culvert
under the street located on the southeast side of the site, which is considered to be the
outlet point of the catchment.
The objective is to estimate the stormwater discharges at the catchment's outlet
and compare them to the ones generated prior to urbanization. The approach typically
employed in stormwater drainage manuals will be used, which is to compute the
hydrologic response of the catchment to a series of synthetic design storms associated
with different return periods. The design storms used here will be for a 2-hour event with
return periods of 2, 10, and 100 years. Most of the parameter values used in this example
were taken from tables published in the SWMM User's Manual (Rossman, 2008). These
were supplemented with design guidelines published by the Denver Urban Drainage and
Flood Control District (UDFCD) (UDFCD, 2001).
Two models will be built: one that represents the catchment in its current
undeveloped condition and one that represents the catchment after it is fully developed.
Because this is an initial estimation of the discharges at the outlet of the catchment under
its current and future conditions, no channelized flows will be defined and only runoff as
overland flow will be simulated. Example 2 in this manual will add a conveyance system
of swales, channels, and culverts to this model.
15
-------
4975
4975
Figure 1-1. Undeveloped site
4975 .. ' :>
4975
4%8
4%7
4973 ^^ Swak-, iMlurjl
i mmmm Culvert*
4970
4968
ป\ 4967
Figure 1-2. Developed site
16
-------
1.2 System Representation
SWMM is a distributed model, which means that a study area can be subdivided
into any number of irregular subcatchments to best capture the effect that spatial
variability in topography, drainage pathways, land cover, and soil characteristics have on
runoff generation. An idealized subcatchment is conceptualized as a rectangular surface
that has a uniform slope and a width W that drains to a single outlet channel as shown in
Figure 1-3. Each subcatchment can be further divided into three subareas: an impervious
area with depression (detention) storage, an impervious area without depression storage
and a pervious area with depression storage. Only the latter area allows for rainfall losses
due to infiltration into the soil.
W
Figure 1-3. Idealized representation of a subcatchment
The hydrologic characteristics of a study area's subcatchments are defined by the
following set of input parameters in SWMM:
Area
This is the area bounded by the subcatchment boundary. Its value is determined
directly from maps or field surveys of the site or by using SWMM's Auto-Length
tool when the subcatchment is drawn to scale on SWMM's study area map.
Width
The width can be defined as the subcatchment's area divided by the length of the
longest overland flow path that water can travel. If there are several such paths
then one would use an average of their lengths to compute a width.
In applying this approach one must be careful not to include channelized flow as
part of the flow path. In natural areas, true overland flow can only occur for
distances of about 500 feet before it begins to consolidate into rivulet flow. In
urbanized catchments, true overland flow can be very short before it is collected
into open channels or pipes. A maximum overland flow length of 500 feet is
appropriate for non-urban catchments while the typical overland flow length is the
17
-------
length from the back of a representative lot to the center of the street for urban
catchments. If the overland flow length varies greatly within the subcatchment,
then an area-weighted average should be used.
Because it is not always easy to accurately identify all of the overland flow paths
within a subcatchment, the width parameter is often regarded as a calibration
parameter whose value can be adjusted to produce a good match between
observed and computed runoff hydrographs.
This is the slope of the land surface over which runoff flows and is the same for
both the pervious and impervious surfaces. It is the slope of what one considers to
be the overland flow path or its area-weighted average if there are several such
paths in the subcatchment.
Imperviousness
This is the percentage of the subcatchment area that is covered by impervious
surfaces, such as roofs and roadways, through which rainfall cannot infiltrate.
Imperviousness tends to be the most sensitive parameter in the hydrologic
characterization of a catchment and can range anywhere from 5% for
undeveloped areas up to 95% for high-density commercial areas.
Roughness Coefficient
The roughness coefficient reflects the amount of resistance that overland flow
encounters as it runs off of the subcatchment surface. Since SWMM uses the
Manning equation to compute the overland flow rate, this coefficient is the same
as Manning's roughness coefficient n. Separate values are required for the
impervious and pervious fractions of a subcatchment since the pervious n is
generally an order of magnitude higher than the impervious n (e.g., 0.8 for dense
wooded areas versus 0.012 for smooth asphalt).
Depression Storage
Depression storage corresponds to a volume that must be filled prior to the
occurrence of any runoff. Different values can be used for the pervious and
impervious areas of a subcatchment. It represents initial abstractions such as
surface ponding, interception by flat roofs and vegetation, and surface wetting.
Typical values range between 0.05 inches for impervious surfaces to 0.3 inches
for forested areas.
Percent of Impervious Area Without Depression Storage
This parameter accounts for immediate runoff that occurs at the beginning of
rainfall before depression storage is satisfied. It represents pavement close to the
gutters that has no surface storage, pitched rooftops that drain directly to street
gutters, new pavement that may not have surface ponding, etc. By default the
value of this variable is 25%, but it can be changed in each subcatchment. Unless
special circumstances are known to exist, a percent imperviousness area without
depression storage of 25% is recommended.
18
-------
Infiltration Model
Three different methods for computing infiltration loss on the pervious areas of a
subcatchment are available in SWMM. They are the Horton, Green-Ampt and
Curve Number models. There is no general agreement on which model is best.
The Horton model has a long history of use in dynamic simulations, the Green-
Ampt model is more physically-based, and the Curve Number model is derived
from (but not the same as) the well-known SCS Curve Number method used in
simplified runoff models.
The Horton model will be used in the current example. The parameters for this
model include:
Maximum infiltration rate: This is the initial infiltration rate at the start of
a storm. It is difficult to estimate since it depends on the antecedent soil
moisture conditions. Typical values for dry soils range from 1 in/h for
clays to 5 in/h for sands.
Minimum infiltration rate: This is the limiting infiltration rate that the soil
attains when fully saturated. It is usually set equal to the soil's saturated
hydraulic conductivity. The latter has a wide range of values depending
on soil type (e.g., from 0.01 in/hr for clays up to 4.7 in/hr for sand).
Decay coefficient: This parameter determines how quickly the infiltration
rate "decays" from the initial maximum value down to the minimum
value. Typical values range between 2 to 7 hr"1.
Precipitation Input
Precipitation is the principal driving variable in rainfall-runoff-quantity
simulation. The volume and rate of stormwater runoff depends directly on the
precipitation magnitude and its spatial and temporal distribution over the
catchment. Each subcatchment in SWMM is linked to a Rain Gage object that
describes the format and source of the rainfall input for the subcatchment.
1.3 Model Setup - Undeveloped Site
The SWMM model for the undeveloped site is depicted in Figure 1-4. It consists
of a rain gage Rl that provides precipitation input to a single subcatchment SI whose
runoff drains to outfall node Ol. Note that the undeveloped site contour map has been
used as a backdrop image on which the subcatchment outline has been drawn. The
SWMM input file for this model is named Examplel-Pre.inp.
19
-------
Utilizing a Backdrop Image in SWMM
To help facilitate the placement of drainage-system objects, SWMM can
utilize an image as a backdrop behind a project's study area map. This image is
typically a site map of some kind with known dimensions. Any bitmap image file
(BMP extension), Windows metafile (WMF or EMF extension) or JPEG image
file (JPG or JPEG extension) can be used as a backdrop. These images would
typically come from a CAD or GIS drawing of the site or perhaps from an
electronically published or scanned topographic or street map.
Before adding the backdrop image, the actual horizontal and vertical
dimensions represented in the image must be known in order to scale the map
correctly. To add a properly scaled backdrop image to a SWMM project do the
following:
1. Select View | Backdrop | Load from the main menu.
2. Enter the name of the backdrop image file to be loaded in the Backdrop Image
Selector dialog that appears. After closing the dialog, the backdrop will appear
on the Study Area Map with a default set of dimensions.
3. To properly scale both the backdrop and the study area map, select View |
Backdrop | Resize from the main menu. In the Backdrop Dimensions dialog
select the "Scale Map to Backdrop Image" checkbox. This will automatically
adjust the dimensions of the map to be the same as the backdrop. Then enter 0,0
for the lower left coordinates and the backdrop's width and height for the upper
right coordinates.
Lower Left
Backdrop Map
X-eoordinate: 0 000 '0 000
Y-coordinate:
Upper Righl:
X-coordinate:
Y-coordinate:
0000
Backdrop
'1423000
1475 000
jOOOO
Map
]1423.000
1475.000
O Resize Backdrop Image Only
O Scale Backdrop Image to Map
CfjlScaie Map to Backdrop |maฃ^
Help
4.
Finally, select View | Dimensions from the main menu and select the
appropriate units for the map's dimensions (typically feet or meters).
At times one will want to lighten the backdrop image so that the added drainage
system objects will stand out more clearly on the map. This can be done by toggling
the View | Backdrop | Watermark option on the main menu.
20
-------
Subcatchment Properties
According to the site's contour map, its topography is fairly homogenous and no
well-defined channels exist within the basin which means that mainly overland flow takes
place. There are no roads or other local impervious areas and the type of soil is similar
throughout the watershed (Sharpsburg silt loam). Therefore, no disaggregation is required
based on the spatial distribution of catchment properties. The single subcatchment SI
drains to the free outfall node Ol whose elevation is 4967 ft.
O Study Area Map
I. llnllxl
4975
4975
.4972
4971
4970
496S
Figure 1-4. SWMM representation of the undeveloped study area
The area shown in Figure 1-4 is not the entire pre-development natural catchment.
It has been bounded by the post-development roadways to come so that comparisons
between the two conditions (developed and undeveloped) can be made.
The properties assigned to the single subcatchment SI are summarized in Table 1-
1. Their values were developed on the basis of the undeveloped site being primarily
pasture land containing a silt loam soil. Parameter values for this soil type can be found
in the SWMM User's Manual and the UDFCD Guidance Manual.
21
-------
Table 1-1. Properties of the undeveloped subcatchment
Property
Area
Width
Slope
Imperviousness
Roughness coefficient,
impervious areas
Roughness coefficient,
pervious areas
Value
28.94 ac
2521 ft
0.5 %
5%
0.015
0.24
Property
Depression storage,
pervious areas
Depression storage,
impervious areas
% of impervious area without
depression storage
Maximum infiltration rate
Minimum infiltration rate
Infiltration decay coefficient
Value
0.3 in.
0.06 in.
25%
4.5 ia/hr
0.2 ia/hr
6.5 hr1
The subcatchment's area was determined using SWMM's Auto-Length tool. The
subcatchment width was arrived at by first assuming a maximum overland-flow length of
500 ft, as recommended for undeveloped areas. By the time runoff has travelled this
distance it has consolidated into rivulets and therefore no longer behaves as overland
flow over a uniform plane. Based on this assumption, the subcatchment was divided into
subareas with flow-path lengths of 500 feet or less. Figure 1-5 shows that there were
three such areas whose flow-path lengths are each 500 ft. Thus, the average flow length is
also 500 ft. When the total subcatchment area of 28.94 acres (1,260,626 ft2) is divided by
the average flow length the resulting subcatchment width is 2,521 ft. The average slope
of the subcatchment was derived from the area-weighted average of the slopes of the
three sub-areas that comprise the overland flow paths as shown in Figure 1-5 and Table
1-2. Its value is 0.5.
1 + 2+3
S = 0.5% -
Figure 1-5. Computation of width of the undeveloped subcatchment
22
-------
Table 1-2. Flow lengths and slopes of the undeveloped subcatchment
Sub-Area
1
2
3
Average
Flow Path
Length
(ft)
500
500
500
500
Associated
Area (Aj)
(ac)
11.13
14.41
3.4
Upstream
Elevation
(ft)
4974.8
4973
4970
Downstream
Elevation
(ft)
4973
4970
4967
Elevation
Difference
(ft)
1.8
3
3
Slope
(Si)
(%)
0.4%
0.6%
0.6%
0.51
, '=
Weighted average corresponding to V
'=3 V . A
The Horton method was selected as the infiltration model for this analysis. The
values assigned to its parameters are typical of those for a silt loam soil as found in this
watershed and are listed in Table 1-1. It is strongly recommended to use any available
site-specific data from the study area BEFORE relying on values from the literature.
Rain Gage Properties
The properties of rain gage Rl describe the source and format of the precipitation
data that are applied to the study area. In this example, the rainfall data consist of three
synthetic design events that represent the 2-, 10- and 100-year storms of 2-hour duration.
Each storm is represented by a separate time series object in the SWMM model that
consists of rainfall intensities recorded at a 5 minute time-interval. The time series are
named 2-yr, 10-yr and 100-yr, respectively and are plotted in Figure 1-6. The total depth
of each storm is 1.0, 1.7, and 3.7 inches, respectively. These design storms were selected
by the City of Fort Collins, CO to be used with SWMM (City of Fort Collins, 1999).
1 2-year 10-year D 100-year
30
60
Time (minutes)
90
120
Figure 1-6. Design storm hyetographs
23
-------
Measuring Tools Available in SWMM
SWMM's graphical user interface has several tools that can assist in
measuring distances and areas on a project's study area map. One such tool is the
Auto-Length option. If Auto-Length is turned on, any subcatchment's polygon
outline that is drawn or edited will have its area computed automatically and stored
in its Area property. The same holds true for conduit lengths. The current status of
the Auto-Length option is shown in the status bar at the bottom of SWMM's main
window. It can be switched on/off by clicking the drop-down arrow next to the
Auto-Length display
Another feature of Auto-Length is its ability to re-compute the areas of all
subcatchments and the lengths of all conduits in a project at once. This capability
proves useful when a user changes the map scaling, switches between using US and
metric units, or has done a lot of map editing with Auto-Length turned off. To
implement this feature, make sure Auto-Length is turned on and select View |
Dimensions from the main menu. In the Map Dimensions dialog that appears,
check on the option to re-compute all lengths and areas. There is no need to change
any of the other settings in the dialog unless one so desires. Once OK is clicked, all
of the subcatchments and conduits on the map will have their areas and lengths
updated.
A second measurement tool is the Ruler tool. It is used to measure the
distance along a polyline as well as compute an area if the polyline is closed to form
a polygon. To activate the Ruler tool select the S button on the Map Toolbar. Then
click the mouse on the first point to begin measuring from, and continue to click at
intermediate points along the path being measured. To complete the path and
determine its length, right click (or press Enter) at the terminating point. To
measure the perimeter and area of a polygon, make the terminating point the same
as the initial point of the path.
24
-------
1.4 Model Results - Undeveloped Site
Analysis Options
Table 1-3 shows the analysis options used to run the model. Three runs of the
model were made, one for each design storm event. To analyze a particular storm event
one only has to change the Series Name property of the rain gage to the rainfall time
series of interest. The total discharges to the outfall for each storm were then plotted on
one graph for comparison.
^Option
Value
Flow units
Routing Method
Start analysis time and date
Start reporting time and date
End analysis time and date
Reporting time step
Runoff dry-weather time step
Runoff wet-weather time step
Routing time step
CFS
Kinematic wave
01/01/07-00:00
01/01/07-00:00
01/01/07 - 12:00
1 minute
1 hour
1 minute
1 minute
U.S customary units used throughout
Must specify a routing method, but it is not used
in the overland flow computations in this example
Not important for single event simulation
Start reporting results immediately
12 h of simulation (storm duration is 2 h)
Good level of detail in results for a short
simulation
Not important for single event simulation
Should be less than the rainfall interval
Should not exceed the reporting time step
Simulation Results
Figure 1-7 shows the outlet hydrographs obtained for each of the design storms. It
was created by following the procedure outlined in the sidebar titled Exporting Data from
SWMM. Note the significant increase in the peak discharge as the return period increases
and how sensitive it is to the rainfall intensity. (The 2-yr storm hyetograph is plotted for
comparison.) The rate at which the discharge volume increases is much greater than the
rate at which the rainfall volume changes for different return periods. This is because the
soil becomes more saturated during larger storms resulting in more of the rainfall
becoming runoff.
Table 1-4 compares the peak rainfall intensity, total rainfall, total runoff volume,
runoff coefficient, peak runoff discharge and total infiltrated volume for each design
storm. Additionally, the last column provides the percent of the rainfall that is infiltrated
in each case. These values came directly from the Subcatchment Runoff Summary table
that appears in the Status Report of a SWMM run.
25
-------
Exporting Data from SWMM
While data can be plotted in the SWMM interface for one run's results, it is
not possible to plot the results of an additional run with those of an older run. To do
this the results from each run must be exported to a spreadsheet or other plotting
software. Data from both plots and tables can be easily exported in SWMM. The
steps below explain how this is done for the runoff seen at the watershed outlet for
the 2- and 10-yr storms.
Run the 2-yr simulation. Click on the watershed outlet Ol and then select
the Table icon (H) on the Standard Toolbar. For this example, choose "by
Object..."
In the Table by Object dialog box select Date/Time for the Time Format
and for the variable select Total Inflow. Click Ok.
1.
2.
3. A table of runoff flow rates and their times will appear. Select Edit | Select
All and then Edit | Copy to... from the main menu. In the Copy Table
dialog box you have the choice of copying the data to the Clipboard and
pasting it from there directly into the spreadsheet or saving it to a text file.
Choose Clipboard for this example. Open a spreadsheet document, and
paste the Clipboard contents into it.
4. Return to SWMM and run the 10-yr storm and repeat steps 1 to 3. Paste the
data into the same spreadsheet, next to the data from the 2-yr storm.
5. Now use the Scatter plot and the formatting tools of your spreadsheet to plot
the Total Inflow data for both runs on the same graph.
26
-------
45
0:00
1:00
2:00
3:00 4:00
time (hr:min)
5:00
6:00
Figure 1-7. Runoff hydrographs (flow Q versus time) for the undeveloped site
Table 1-4. Summary of results for the undeveloped site
Design
Storm
2-yr
10-yr
100-yr
Peak
Rainfall
(in./h)
2.85
4.87
9.95
Total
Rainfall
(in.)
0.978
1.711
3.669
Runoff
Volume
(in.)
0.047
0.22
1.87
Runoff
Coeff.
(%)
4.8
13.1
50.8
Peak
Runoff
(cfs)
4.14
7.34
31.6
Total
Infiltration
(in.)
0.93
1.48
1.80
%of
Rainfall
Infiltrated
95.1
86.5
49.1
1.5 Model Setup - Developed Site
The increase in impervious surface and reduction of overland flow length are the
main factors affecting the hydrologic response of a catchment when it becomes
urbanized. The reduction in infiltrative surface creates additional surface runoff as well as
higher and faster peak discharges. In this section, the runoff hydrology of our example
site will be modeled in its post-development condition. The SWMM input data for this
model is named Examplel-Postinp. Again, the focus will be on just the rainfall-runoff
transformation and overland flow processes. Routing through channelized elements will
be covered in Example 2.
Catchment Discretization
In the urbanized catchment there are channelized elements (gutters and swales)
that conduct runoff to the site's outlet. The partitioning of the study area into individual
27
-------
subcatchments depends not only on the spatial variability in land features but also on the
location of the channelized elements. Inspection of the developed site plan for the
example study area (see Figure 1-2) suggests that a total of seven subcatchments would
be sufficient to represent both the spatial differences in planned land uses and the location
of channelized elements within the site. The subcatchment boundaries were determined
by aggregating together sub-areas whose potential overland flow paths share a common
direction and drain to the same collection channel. The resulting discretization is shown
in Figure 1-8.
497J
Figure 1-8. Discretization of the developed site into subcatchments
Figure 1-8 shows all of the subcatchments discharging their overland flow
directly into the watershed's outlet node, Ol. In reality, the discharge outlet of each
subcatchment should be the point where its runoff enters the channelized drainage
system. However, since this example does not consider routing through any channelized
elements in the watershed (Example 2 covers this issue) it is acceptable to use the study
area's outlet node (Ol) as a common outlet for all of the subcatchments. The elevation of
this point is 4962 ft, which corresponds to the bottom elevation of a planned culvert
under the street.
28
-------
Geometric Parameters
Table 1-5 lists the area, flow path length, width, slope and imperviousness of each
subcatchment. The areas were computed using SWMM's Auto-Length tool as the outline
of each subcatchment was traced on the scaled backdrop image. (See the sidebar
"Measuring Tools Available in SWMM" for more information.)
Table 1-5. Geometric properties of the subcatchments in the developed site
, Area Flow Width
Subcatchment , , T ,, ,,,. ,,,.
(ac) Length (ft) (ft)
1
2
3
4
5
6
7
4.55
4.74
3.74
6.79
4.79
1.98
2.33
125
125
112
127
125
125
112
1587
1653
1456
2331
1670
690
907
Slope
(%)
2.0
2.0
3.1
3.1
2.0
2.0
3.1
Percent
Impervious
56.8
63.0
39.5
49.9
87.7
95.0
0.0
Figure 1-9 illustrates how the overland flow path length was estimated for
subcatchment S2 which consists entirely of residential lots. This subcatchment can be
represented as a rectangular area with an overland flow length equal to the distance from
the back of a typical lot to the middle of the street (125 ft in this case). SWMM's width
parameter can then be computed as the area (4.74 ac = 206,474.4 ft2) divided by the
overland flow length, which results in a value of 1650 ft.
Figure 1-9. Definition of overland flow length and slope for subcatchment S2
29
-------
In contrast to S2, subcatchments S3 and S4 contain both residential lots and grass-
covered areas. Their overland flow lengths are computed as an area-weighted average of
the flow lengths across each type of area as shown in Figure 1-10. Their widths are then
found by dividing their areas by their overland flow lengths.
Total Area = 3.74 ac
Total Area = 6.79 ac
Figure 1-10. Width and slope computation for subcatchments S3 (a) and S4 (b)
Slopes characterizing overland flow in mostly urbanized subcatchments will be
the lot slope, which is usually about 2%. By way of illustration, Figure 1-10 shows how
the slopes of subcatchments S3 and S4 correspond to the area-weighted average of the
slope of the overland flow paths over both the residential lots and the grass-covered
areas.
Imperviousness
The imperviousness parameter in SWMM is the effective or directly connected
impervious area, which is typically less than the total imperviousness. The effective
impervious area is the impervious area that drains directly to the stormwater conveyance
system, e.g. a gutter, pipe or swale. Ideally, the imperviousness should be measured
directly in the field or from orthophotographs by determining the percent of land area
devoted to roofs, streets, parking lots, driveways, etc. When these observations are not
available, it is necessary to use other methods. A conservative approximation that tends
to overestimate runoff discharges is to use runoff coefficients as the imperviousness
value. A runoff coefficient is an empirical-constant value that represents the percentage
of rainfall that becomes runoff. Using the runoff coefficient to represent the percent
imperviousness of a subcatchment results in a higher estimate of impervious area because
the value is calculated from the runoff of both the impervious and pervious areas of the
subcatchment. For purely illustrative purposes, runoff coefficients will be used in this
example to estimate the imperviousness for each subcatchment within the developed
watershed. The steps involved are as follows:
1. Identify all of the major land uses that exist within the subcatchment.
2. Compute the area Aj devoted to each land usey in the subcatchment.
30
-------
3. Assign a runoff coefficient Q to each land use category j. Typical values are available
in drainage criteria and basic literature (see for example UDFCD, 2001; Akan, 2003).
Pervious areas are assumed to have a runoff coefficient of 0.
4. Compute the imperviousness / as the area weighted average of the runoff coefficients
for all of the land uses in the subcatchment, / = (SCjA^/A, where A is the total area of
the subcatchment.
When this approach is applied to the current example the results listed in Tables
1-6 and 1-7 are obtained. Table 1-6 displays the various land-use categories that appear in
the developed site along with their runoff coefficients. The latter were obtained from the
City of Fort Collins Storm Drainage Design Criteria and Construction Standards (City of
Fort Collins, (city of Fort Collins, 1984 and 1997). Table 1-7 lists the amount of area
devoted to each land use within the site's subcatchments. These areas were used to
compute a weighted-average runoff coefficient that is used as a surrogate for the
imperviousness of the given subcatchment.
Table 1-6. Land use categories in the developed site
Id
M
L
DL
M2
S
RT
T
P
Land Use Runoff coefficient (C)
Medium density
Low density
Duplex
Medium density
Apartment, high
density
Commercial
Commercial
Natural (park)
0.65
0.45
0.70
0.65
0.70
0.95
0.95
0
Table 1-7. Land use coveage (ac) and imperviousness for subcatchments in the developed site
Sub- Total Area Area Area Area Area Area Area Area Imperviousness
catchment Area(ac) M L DL M2 S RT T P (%)
SI
S2
S3
S4
S5
S6
S7
4.55
4.74
3.74
6.79
4.79
1.98
2.33
2.68
0
0
0.61
0
0
0
1.87
1.32
0
0
0
0
0
0
3.42
0
0
0
0
0
0
0
1
0
0
0
0
0
0
1.18
2.05
0
0
0
0
0
0
1.64
0.7
0
0
0
0
0
0
3.72
1.98
0
0
0
1.56
2.49
0.37
0
2.33
56.8
63
39.5
49.9
87.7
95
0
31
-------
Remaining Parameters
The remaining subcatchment properties for the developed site (roughness
coefficients, depression storages, and infiltration parameters) are kept the same as they
were for the undeveloped condition. Likewise, the same analysis options were used to run
the simulations. Refer to Tables 1-1 and 1-3 for a listing of the parameter values used in
the undeveloped condition.
1.6 Model Results - Developed Site
Outlet Hydrographs
Figure 1-11 shows the outlet hydrographs (the Total Inflow to node Ol) obtained
for each of the design storms under post-development conditions in the study site. As
with the pre-development hydrographs, the peak runoff flow occurs close to when the
peak rainfall does and there is a significant increase in peak discharge as the return period
increases. Unlike the pre-development case, the post-development's hydrographs show a
more rapid decline once the rainfall ceases. This behavior can be attributed to the much
larger amount of imperviousness under the post-development condition (57%) as
compared to pre-development (5%). Table 1-8 summarizes the results obtained for each
design storm in the same fashion that Table 1-4 did for the pre-development condition.
300
270
240
210
180
iso
120
90
I 2-yr storm
-2-yr
10-yr
100-yr
0:00
1:00
4:00
- 0
1
2
3
4
5
6
7
2:00 3:00
time (hr:min)
Figure 1-11. Runoff hydrographs (flow Q versus time) for the developed site
1 10
5:00
32
-------
Table 1-8. Summary of results for jjost-devdonment conditions
Design
Storm
2-yr
10-yr
100-yr
Peak
Rainfall
(in./h)
2.85
4.87
9.95
Total
Rainfall
(in.)
0.978
1.711
3.669
Runoff
Volume
(in.)
0.53
1.11
3.04
Runoff
Coeff.
(%)
54.5
64.7
82.7
Peak
Runoff
(cfs)
46.7
82.6
241
Total
Infiltration
(in.)
0.42
0.58
0.61
%of
Rainfall
Infiltrated
42.9
33.8
16.6
Pre- and Post-Development Comparison
Table 1-9 compares total runoff volumes, runoff coefficients, and peak discharges
computed for both the pre- and post-development conditions. For larger storm events,
where infiltration plays a minor role in the runoff generation, the responses become more
similar between the two cases. Total runoff volume under post-development conditions is
approximately 10, 5, and 2 times greater than under pre-development conditions for the
2-yr, 10-yr, and 100-yr storms, respectively. Peak flows are about 10 times greater for
both the 2-yr and 10- yr storms but only 7 times greater for the 100-yr event.
Table 1-9. Comparison of runoff for gre- and gost-develojjment conditions
Runoff Volume (in.) Runoff Coeff. (%)
Design
Storm
2-yr
10-yr
100-yr
Total
Rainfall
(in.)
0.978
1.711
3.669
Pre
__
0.24
1.87
Post
__
1.11
3.04
Pre
__
13.1
50.8
Post
__
64.70
82.70
Peak Runoff (cfs)
Pre
__
7.34
31.6
Post
__
82.64
240.95
1.7 Summary
This example used SWMM to estimate the runoff response to different rain events
for a 29 ac development that will be built in a natural area. Comparisons were made
between the runoff peaks and total volume for each event for both pre- and post-
developed conditions. The key points illustrated in this example were:
1. Building a SWMM model for computing runoff requires that a study area be properly
partitioned into a collection of smaller subcatchment areas. These can be determined
by examining the potential pathways that runoff can travel as overland flow and the
location of the collection channels, both natural and constructed, that serve to
intercept this runoff.
2. Initial estimates of most subcatchment parameters can be based on published values
that are tabulated for various soil types and land uses. The primary exception to this is
the width parameter which should be based on the length of the overland-flow path
that the runoff travels.
33
-------
3. Path lengths for true overland flow should be limited to about 500 ft or to the distance
at which a collection channel/pipe is reached if it is less than 500 ft.
4. Urban development can create large increases in the imperviousness, peak-runoff
flow rate, and total-runoff volume.
The next case study, Example 2, will further refine the model built in this example
by adding a stormwater collection system to it and routing the runoff flows through this
system.
34
-------
Example 2. Surface Drainage Hydraulics
Example 1 showed how to construct a hydrologic model of an urban catchment
that compared stormwater runoff under both pre- and post-development conditions.
Hydraulic routing was not considered. This example will demonstrate how SWMM's
hydraulic elements and flow routing methods are used to model a surface drainage
system. A conveyance network will be added to the post-development model built in
Example 1 and be sized to pass the 2 hour synthetic storm events with return periods of
2-, 10-, and 100-years. For simplicity, open channels (e.g. swales or gutters) will be used
to convey flow. The simple routing network developed in this example will be built upon
further in Example 7 where additional open channels and below-ground pipes will be
added that are designed according to typical drainage design criteria.
2.1 Problem Statement
Figures 2-1 and 2-2 show the SWMM model layouts of the pre- and post-
development study area developed in Example 1. In Figure 2-1 the pre-development site
was represented by a single subcatchment whose width parameter was determined by
assuming a maximum overland flow length of 500 ft, as recommended for undeveloped
areas. For the developed case (Figure 2-2), the post-development site was discretized into
seven subcatchments, the subcatchment widths were computed using an area-weighted
average of the flow lengths across each type of area, and all subcatchments discharged
their overland flow directly to the site's outlet, node Ol (see Section 1.5 of Example 1).
The objective of this example is to add a simple surface drainage system to the
post-development site. A system of gutters, grass swales, and culverts will be designed
and sized to convey the 100-yr storm. Runoff from three design storms (the 2-, 10- and
100-yr storms) will be routed through this system using the three alternative hydraulic
routing methods available in SWMM. The resulting outflow hydrographs at the site's
outlet will be compared to those generated in Example 1 where no hydraulic routing was
used.
2.2 System Representation
SWMM models a conveyance network as a series of nodes connected by links
(Figure 2-3). Links control the rate of flow from one node to the next and are typically
conduits (e.g. open channels or pipes) but can also be orifices, weirs or pumps. Nodes
define the elevation of the drainage system and the time-varying hydraulic head applied
at the end of each link it connects. The flow conveyed through the links and nodes of the
model is ultimately discharged to a final node called the outfall. Outfalls can be subjected
to alternative hydraulic boundary conditions (e.g. free discharge, fixed water surface,
time varying water surface, etc.) when modeled with Dynamic Wave. The properties of
these drainage system elements are explained in detail in the sidebar "Hydraulic Elements
in SWMM".
35
-------
4975
4975
4972
4971
Figure 2-1. Pre-development site
Hydraulic routing is the process of combining all inflows that enter the upstream
end of each conduit in a conveyance network and transporting these flows to the
downstream end over each instance of time. The resulting flows are affected by such
factors as conduit storage, backwater, and pipe surcharging. SWMM can perform
hydraulic routing by three different methods: Steady Flow, Kinematic Wave and Dynamic
Wave. These three methods are summarized below.
Steady Flow
Steady Flow routing is an instantaneous translation of a hydrograph from the
upstream end of a conduit to the downstream end with no time delay or change in
shape due to conduit storage. Steady Flow routing will simply sum the surface runoff
from all subcatchments upstream of the selected node through time.
36
-------
4'.'?.,
4'l-s
4T2
Figure 2-2. Post-development site without conveyance system
Network Drainage System
SWMM Representation
-Nodes
Links
Figure 2-3. Links and Nodes
37
-------
Hydraulic Elements in SWMM
All hydraulic elements modeled in SWMM are classified as either nodes or links.
The hierarchy of these elements is shown below along with both the required and optional
properties that characterize each element's hydraulic behavior.
Type of Hydraulic Category of Examples or Required Optional
Element Element Types Properties Properties
Junction -Manholes
-Points of change
Oin conduit slope or
cross section
Divider -Cutoff
-Tabular
O-Weir
-Overflow
T3
ฎ Storage -Reservoirs
f TT . -Peak shaving
Unit detention ponds
-BMPs
Outfall -Free
-Normal
\ / -Fixed
V -Tidal
-Time series
Conduit -Natural channels
-Closed conduits
-Open Channels
Pump -Off-line
-In-line
f^~3 incremental
^-^ -Variable speed
in-line
a3 Orifice -Circular orifice
" -Rectangular
S3 . . orifice
3 ^
Weir -Transverse
-Side-flow
0-V-notch
-Trapezoidal
Outlet -Used to model
special head-
j^ji discharge
TXXf relationships
-Invert elevation
-Maximum depth
-Initial depth
-Invert elevation
-Diversion link
-Type
-Maximum depth
-Initial depth
-Invert elevation
-Storage curve
-Maximum depth
-Initial depth
-Invert elevation
-Type
-Inlet node
-Outlet node
-Shape and section
geometry
-Length
-Roughness
-Inlet offset
-Outlet offset
-Inlet node
-Outlet node
-Pump curve
-Inlet node
-Outlet node
-Type
-Shape and geometry
-Inlet offset
-Discharge coefficient
-Inlet node
-Outlet node
-Type
-Geometry
-Inlet offset
-Discharge coefficient
-Inlet node
-Outlet node
-Inlet offset
-Rating curve
-Surcharge depth
-Treatment
-Inflows
-Ponded area
-Ponded area
-Surcharge depth
-Treatment
-Inflows
-Treatment
-Inflows
-Evaporation factor
-Inflows
-Treatment
-Tide gate
-Initial flow
-Maximum flow
-Entry loss coefficient
-Exit loss coefficient
-Average loss
coefficient
-Initial status
-Startup depth
-Shutoff depth
-Flap gate
-Time to open/close
-Flap gate
-End contractions
-End coefficient
-Flap gate
38
-------
Kinematic Wave
Kinematic Wave uses the normal flow assumption for routing flows through the
conveyance system. In Kinematic Wave routing, the slope of the hydraulic grade line
is as the same as the channel slope. Kinematic Wave routing is most applicable to the
upstream, dendritic portions of drainage systems where there are no flow restrictions
that might cause significant backwater or surcharging. It can be used to approximate
flows in non-dendritic systems (i.e., those that have more than one outflow conduit
connected to a node) only if "flow divider" nodes are used.
Dynamic Wave
Dynamic Wave routing is the most powerful of the flow routing methods because it
solves the complete one-dimensional Saint Venant equations of flow for the entire
conveyance network. This method can simulate all gradually-varied flow conditions
observed in urban drainage systems such as backwater, surcharged flow and flooding.
Dynamic Wave can simulate looped conduit systems and junctions with more than
one link connected downstream (bifurcated systems). The ability to simulate
bifurcated systems allows one to model pipes and gutters in parallel; this more
advanced level of modeling will be described in Example 7.
2.3 Model Setup
System Layout
Figure 2-4 shows the layout of the runoff conveyance system that will be added to
the developed site. It consists of 7 grass swales, 3 culverts, and one street gutter. The
objective in this example is to estimate the discharges at the outlet of the site and not
design all of the elements within the entire drainage system. For this reason, only the
main surface conduits that route runoff to the outlet in the aggregated model will be
considered. These will purposely be oversized to ensure that all the generated runoff is
conveyed to the outlet and that no flooding occurs within the site (see Example 7 for an
analysis of surcharged pipes and flooded junctions). The starting point for building this
model is the input file Examplel-Post.inp that was created in Example 1.
In Example 1, the subcatchment widths were set to properly represent the
overland flow process. All the subcatchments were directly connected to the outlet of the
study area; flows through channels were not modeled. In this example, the subcatchment
properties will remain the same as defined previously, but conduits representing the
channelized flow throughout the site will be added into the model.
The definition of the conveyance system begins by specifying the location of its
nodes (or junctions). A node is required wherever runoff enters the conveyance system,
whenever two or more channels connect and where the channel slope or cross section
changes significantly. They are also required at locations with weirs, orifices, pumps,
storage, etc. (see Example 3 where orifices and weirs are used as outlets to a storage
unit). The locations of the nodes for this example are shown in Figure 2-5. They are
labeled Jl through Jll. The invert elevation of each node (i.e., the elevation at the
bottom of the lowest connecting channel) is shown in Table 2-1.
39
-------
4975
4973
Swale, natural
channel
Culverts
4975
V 4967
Figure 2-4. Post-development site with runoff conveyance added
40
-------
RainGags
Culvert
Swale
Gutter
Figure 2-5. SWMM representation of the post-development conveyance system
Table 2-1. Invert elevations of junctions
Junction ID
Jl
J2
J3
J4
J5
J6
J7
J8
J9
J10
Jll
Invert Elevation (ft)
4973.0
4969.0
4973.0
4971.0
4969.8
4969.0
4971.5
4966.5
4964.8
4963.8
4963.0
41
-------
The definition of the conveyance system continues by adding the feeder channels
Cl, C2 and C6 that convey runoff into the main drainage way that runs through the
undeveloped park area of the site. Conduit Cl is a grass swale that drains subcatchment
57's runoff to the watershed's main drainage way; Conduit C2 is a gutter that carries
subcatchment S2's runoff to the upstream end of the culvert (C77) that discharges to the
site's outlet (07); conduit C6 carries runoff from subcatchment S4 into culvert C7. At
this point, the elevations of the bottom bed of these channels correspond to the invert
elevations of their respective upstream and downstream junctions. Their lengths are
automatically determined by drawing them with Auto-Length turned on. SWMM uses
this information to compute the slope of each channel.
Finally, the remaining conduits C3 through C77 that comprise the main drainage
way through the park to the outlet need to be defined. As before, they connect to their end
nodes with no vertical offset and the Auto-Length tool is used to compute their lengths.
Subcatchments S3 through S7 drain to different locations of the main drainage channel.
S3 drains to the culvert (C3) at the beginning of the main drainage channel, S4 drains to
the swale C6 that connects to the second culvert (C7) on the main drainage channel, S7
and ฃ5 drain to the main drainage channel at J10, and S6 drains directly to the last culvert
(C77) on the main drainage channel. Table 2-2 summarizes the outlet junction and
conduit associated with each subcatchment.
Table_2^2._Subcatchmratoutlets
Subcatchment Outlet Junction Outlet Conduit
SI Jl Cl(Swale)
S2 J2 C2 (Gutter)
S3 J3 C3 (Culvert)
S4 J7 C6 (Swale)
S5 J10 CIO (Swale)
S6 Jll Cll (Culvert)
S7 J10 CIO (Swale)
Note that the conveyance system modeled in this example (Figure 2-5) ignores the
storage and transport provided by the street gutters within each subcatchment. In some
applications, however, these conveyance elements can play a significant role and should
be represented, perhaps by adding a channel within each subcatchment that represents the
aggregated effects of routing through all of its street segments. To keep the example
simple, this level of detail is not included and only the major drainage channels within the
site are considered.
System Properties
Properties can now be assigned to the conduits and junctions that have been
defined. Table 2-3 shows the cross-sectional shapes of the three conduit types used in this
example. The swale side slopes (Zl and Z2 (horizontal:vertical)), roughness coefficient
(n), bottom width (b) and height (h) of the swale section are as recommended by the
42
-------
UDFCD Manual (2001). The gutter cross-slopes (Zl and Z2), roughness coefficient (n),
bottom width (b) and height (h) are based on typical design practice. The culvert
diameters will be sized as described in the next section to convey runoff from the 100-
year storm.
Table 2-3. Characteristics of the conduits used in the model
Type of Shape Cross_Section Representing DJft) Zt ^ bj|t) h (f t) n
Trapezoidal \ / " Swales -5553 0.05
h
Trapezoidal ,.--''""""" ' Gutters - 0.00017 25 0 1 0.016
U-'-' ,
Circular ( n } Culverts 4.752 .... 0.016
1 SWMM cannot accept a slope of zero
2 This is the initial diameter of the culverts, not the final value
Table 2-4 shows the SWMM properties assigned to each conduit. Conduit lengths
were computed using the Auto-Length option as described in the previous section. The
inlet and outlet offset of all the conduits, with one exception, were set to zero which
means that the bottom elevation of each conduit coincides with the elevation of its inlet
and outlet junctions. The exception is conduit C2 (the gutter), whose outlet offset of 4 ft
represents the difference in elevation between the bottom of the gutter and the channel
bed in the park. The diameters of the three circular culverts will be determined in the next
section.
Conduit
ID
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CIO
Cll
Type of
Conduit
Swale
Gutter
Culvert
Swale
Swale
Swale
Culvert
Swale
Swale
Swale
Culvert
Inlet
Node
Jl
J2
J3
J4
J5
J7
J6
J8
J9
J10
Jll
Outlet
Node
J5
Jll
J4
J5
J6
J6
J8
J9
no
Jll
Ol
Length
(ft)
185
526
109
133
207
140
95
166
320
145
89
h (ft) or
D(ft)
3
1
TBD
o
J
3
o
J
TBD
3
o
J
3
TBD
b(ft)
5
0
-
5
5
5
-
5
5
5
-
Zi
5
0.0001
-
5
5
5
-
5
5
5
-
Z2
5
25
-
5
5
5
-
5
5
5
-
TBD = To Be Determined
43
-------
Because the conduits are all surface channels and not buried pipes, it is sufficient
to set the maximum depth of all the junctions as zero. This will cause SWMM to
automatically set the depth of each junction as the distance from the junction's invert to
the top of the highest conduit connected to it. Thus, junction flooding (the only flooding
allowed by SWMM) will occur when the channel capacity is exceeded. Finally, the
outfall node Ol is defined as a free outfall (see the sidebar "Hydraulic Elements in
SWMM'} with an elevation of 4962 ft. The resulting SWMM input file has been named
Example2-Post.inp
2.4 Model Results
Culvert Sizing
Before SWMM's alternative routing methods are compared, the diameters of the
three culverts in the conveyance system must be determined. This is done by finding the
smallest available size for each culvert from those listed in Table 2-5 that will convey the
runoff from the 100-year, 2-hr design storm without any flooding. This process involves
the following steps:
1. Start with each culvert at the largest available diameter.
2. Make a series of SWMM runs, reducing the size of conduit C3 until flooding occurs.
Set the size of C3 to the next larger diameter.
3. Repeat this process for conduit C7 and then for Cll.
Note that one proceeds systematically from upstream to downstream, making sure
that each culvert in turn is just big enough to handle the flow generated upstream of it.
This procedure is appropriate because there is no flooding under the baseline condition
(with the culverts at their maximum possible size). The approach should not necessarily
be applied when pipe diameters have to be enlarged (or when there is flooding under the
baseline condition). It is very common to find design situations in which changes
downstream have significant effects upstream, so that a minor change in the diameter of a
pipe located downstream may solve flooding problems upstream.
Table 2-5. Available culvert sizes
Diameter Diameter
(ft) (m)
1 12
1.25 15
1.5 18
1.75 21
Diameter Diameter
(ft) (in)
2 24
2.25 27
2.5 30
2.75 33
Diameter Diameter
ffi) (in)
3 36
3.25 39
3.5 42
3.75 45
Diameter Diameter
ffi) fisL_
4 48
4.25 51
4.5 54
4.75 57
These culvert-sizing runs are made using the Example2-Post.inp file with the
routing method set to KW (Kinematic Wave), the Rain Gage's time series set to 100-yr,
and the following set of time steps: 1 minute for reporting, 1 h for dry weather, 1 minute
for wet weather and 15 s for routing. Note that the routing time step is somewhat
stringent for the drainage system being modeled and the routing method used (KW). It is
44
-------
used here, however, because Dynamic Wave routing will be used later in this example
and it typically requires smaller time steps than Kinematic Wave to produce stable results.
If just KW routing were used for this model, the routing time step could probably be
safely set to 1 minute (see sidebar "A Note About Time Steps"}. The presence or absence
of flooding is determined by examining the Node Flooding Summary section of a run's
Status Report.
A Note About Time Steps
SWMM requires that four time steps be specified: runoff time steps for both
wet weather and dry weather, a flow routing time step, and a reporting time step.
The most common error new users make is to use time steps that are too long. The
runoff wet weather time step should not exceed the precipitation recording interval.
The flow routing time step should never be larger than the wet weather time step,
and in most cases should be 1 to 5 minutes (or less) for Kinematic Wave routing and
30 s (or less) for Dynamic Wave routing. Dynamic Wave routing can also employ a
Variable Time Step option that automatically lowers the time step during periods
when flows change rapidly. High continuity errors typically result when runoff or
routing time steps are too large. If the reporting time step is set too high important
details in the output results might be missed. Setting the reporting time step equal to
the routing time step helps prevent this, but can generate very large output files.
Starting with smaller time steps, users can experiment with larger time steps to find
one that produces acceptably accurate results most efficiently.
After making these sizing runs with Example2-Post.inp, the final diameters
selected for the culverts are 2.25 ft for C3, 3.5 ft for C7, and 4.75 ft for Cll. Table 2-6
lists the fraction that each conduit is full and the fraction of its full Manning flow that is
reached under peak flow conditions for the 100-year event. These values are available
from the Link Flow Summary table of SWMM's Status Report.
Table 2-6. Maximum depths and flows for conduits during the 100-yr event
Conduit Maximum Depth Maximum Flow
ID
Cl
C2
C3
C4
C5
C6
C7
C8
C9
CIO
Cll
Full Depth
0.37
0.96
0.70
0.38
0.67
0.44
0.71
0.70
0.88
0.88
0.78
Full Flow
0.11
0.94
0.83
0.12
0.41
0.16
0.85
0.44
0.76
0.74
0.95
45
-------
Comparison of Routing Methods
Having adequately sized the culverts, the model is next run using all three routing
methods (Steady Flow, Kinematic Wave, and Dynamic Wave) to obtain the outlet
discharges which are then graphed by design storm along with the discharges found from
Example 1. Figures 2-6, 2-7 and 2-8 show the outflow hydrographs (Total Inflow to node
Ol) generated for each design storm using all three hydraulic routing methods. As in
Example 1, these graphs were created by first exporting the pertinent SWMM results for
each run to a spreadsheet and then using the spreadsheet's graphing tools. For Steady
Flow routing the outlet flows are identical to the flows generated in Example 1 (shown as
a dotted line) except for the case of the 100-yr design storm which produced flooding
within the system, This is because the discharge that appears at the outlet of each
subcatchment in Steady Flow routing instantaneously appears at the site outlet where it is
added to the discharges of the other subcatchments of the watershed. Thus, Steady Flow
routing produces results at the site outlet that would have been generated had no channels
been simulated in the model.
Regarding flooding, the Steady Flow method indicates potential for flooding in
the system by computing the flow depth in the conduits using Mannings equation; if this
depth exceeds the channel capacity, the flow is truncated to the full-flow capacity of the
conduit and flooding is reported. Figure 2-8 shows this difference between the outlet
discharges generated by the Steady Flow routing method and the simulation without
routing when flooding does occur.
100
90
80
70
60
,", 50
a
40 -
30
20
10
0
2-yr storm
2-yr, steady
2-yr, K.W
2-yr, D.W.
Example 1
10
0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00
time (hr:min)
Figure 2-6. Post-development outflow hydrographs for 2-yr storm
46
-------
150
135
f
DOIT
I 1 10-yr storm
10-yr, steady
10-yr, K.W
10-yr, D.W.
Example 1
9
10
0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00
time (hr:min)
Figure 2-7. Post-development outflow hydrographs for 10-yr storm
300
100-yr storm
100-yr, stead
100-yr, K.W
- 100-yr, D.W.
Example 1
0:00 0:30 1:00 1:30 2:00 2:30 3:00 3:30 4:00 4:30 5:00
time (hr:min)
Figure 2-8. Post-development outflow hydrographs for 100-yr storm
47
-------
The other two routing methods, Kinematic and Dynamic Wave., both produce a
time-lag and a reduction in the peak flow, spreading the volume of the outlet hydrograph
out over time. These effects are more pronounced under Dynamic Wave routing because
it accounts for backwater that can increase even further the storage utilized within the
conveyance system.
Table 2-7 compares total runoff volumes, runoff coefficients, and peak discharges
at the outlet computed for the post-development model without routing (from Example 1)
and with Dynamic Wave (DW) routing from this example. These values come directly
from SWMM's Status Report. In terms of runoff volumes and coefficients, the results
obtained with routing are identical to those found in Example 1 where no hydraulic
routing was considered. The effects of routing are observed in the comparison of peak
flows, which decrease when routing is considered. In the case of this example, peaks
produced with Dynamic Wave routing are 28.7% smaller for the 2-yr storm, 24.8%
smaller for the 10-yr storm and 32.4% smaller for the 100-yr storm in comparison to the
peaks produced when no routing was considered.
TabJe_2:7.JI!ojnj)an^^
Design
Storm
2-yr
10-yr
100-yr
Total
Rainfall
(in.)
0.98
1.71
3.67
Runoff Volume (in.)
No Routing1
0.53
1.11
3.04
DW
0.53
1.11
3.04
with and
Runoff Coeff. (%)
No Routing
54.5
64.7
82.7
DW
54.5
64.7
82.7
withoutj'outiiig
Peak Runoff (cfs)
No Routing
46.7
82.6
241
DW
33.4
62.2
164.1
'Results from Examplel-Postinp
2.5 Summary
This example introduced the use of hydraulic routing in SWMM. It demonstrated
how a surface runoff collection system is laid out, how to size the elements of this
system, and the effect that routing of runoff flows through this system has on the
catchment's outlet hydrograph. Comparisons were made between the runoff peaks and
total volume for different design storm events using each of SWMM's three routing
options as well as with no routing. The key points illustrated in this example are:
1. A runoff collection system can be represented as a network of links and nodes, where
the links are conduits (such as grass swales, street gutters, and circular culverts) and
the nodes are the points where the conduits join to one another.
2. An iterative process that proceeds from upstream to downstream can be used to
determine the minimum conduit size needed to prevent flooding under a particular
extreme design event.
3. Steady Flow hydraulic routing produces outlet discharges identical to those produced
without routing unless there is flooding in the drainage system. The method
instantaneously translates hydrographs from the upstream end of a conduit to the
downstream end, with no delay or change in shape.
48
-------
4. Dynamic Wave and Kinematic Wave routing produce smaller peak runoff discharges
than models without routing (Example 1) due to storage and possible backwater
effects within the channels. Routing with Dynamic Wave resulted in a decrease of
32.4% for the 100-yr peak outlet flow.
5. Except for flooding, the choice of routing method (Steady Flow, Kinematic Wave or
Dynamic Wave) does not affect the total volume of runoff that leaves the study area
through the outlet.
In the next case study, Example 3, a storage unit will be added to the post-
development drainage system developed in this example to mitigate the effects that
urbanization has on receiving streams.
49
-------
Example 3. Detention Pond Design
This example illustrates how to define, design, and evaluate a detention pond
using SWMM. Storage units, orifices and weirs will be used to model a multi-purpose
detention pond built to detain a water quality capture volume (WQCV) and control peak
post-development release rates to their pre-development levels. The urban catchment
studied in Examples 1 and 2 will also be used in this example.
Storage is widely used in urban runoff quantity and quality control, providing
both peak flow reduction and suspended solids removal. The design criteria for storage
structures have changed over time due to an improved understanding of the effects that
urban runoff has on the environment. Facilities must control not only the extreme runoff
events to prevent flooding, but also the more common smaller events that produce a "first
flush" pollution phenomenon and thereby impact the quality of receiving water bodies.
3.1 Problem Statement
In Example 1 a model was built to estimate the pre-development runoff from a 29
acre site. Additional models were constructed to estimate the site's post-development
runoff both without flow routing (Example 1) and with routing through a surface
collection system (Example 2). Total site runoff for the 2-, 10- and 100-yr design storms
was computed for both the pre- and post-development conditions. Based on the results of
these models, it is now required to design a detention pond immediately downstream of
the planned urban development to both prevent flooding and protect water quality in a
receiving stream. It is required that the pond reduce the peak discharges of the 2-, 10- and
100-yr storms to those of the undeveloped site, and that extended detention be provided
for a specific water quality capture volume.
Table 3-1 shows the discharges to be controlled by the pond. The pre-
development peaks were determined in Example 1 (Table 1-9 in Section 1.5) and post-
development peaks were determined in Example 2 (Table 2-7 in Section 2.4). The
SWMM input files that produced these results are Examplel-Pre.inp and Example2-
Post.inp, respectively. The 2-, 10- and 100-yr rainfall hyetographs are included in both
these input files.
TjJbleJ^JJ^re-.
Return
Period (yr)
2
10
100
and post-developmei
Rainfall Depth
(in.)
0.98
1.71
3.67
it peak discharges
Pre-development Peak
Discharge (cfs)1
4.14
7.34
31.6
Post-development Peak
Discharge (cfs)2
33.5
62.3
163.8
1 From Examplel-Pre.inp
2 From Example2-Post.inp
50
-------
In addition to controlling the discharges listed, it is required that a Water Quality
Capture Volume (WQCV) be controlled. The WQCV is defined as a suitable volume
expressed in units of watershed depth that is detained for a long enough period of time to
achieve a targeted level of pollutant removal. The required volume and drawdown time
vary under different stormwater control policies (Akan and Houghtalen, 2003). In this
example, the WQCV must be released over 40 hours, during which a significant portion
of particulate pollutants found in urban stormwater are removed. Finally, for safety
reasons, a maximum storage depth of 6 ft is considered for the final design. In this
example the minor storms (the WQCV and 2-yr storm) and the major storms (the 10- and
100-yr storms) runoff will be detained in separate sections of the detention pond. Both
sections will have the shape of a trapezoidal prism. The location of the pond within the
developed study area is shown in Figure 3-1.
4975
4973
Swale, ii.iini.il
channel
4975
Figure 3-1. Detention pond location for the post-development site
51
-------
3.2 System Representation
The main elements used to design detention ponds in SWMM are storage units
with orifice and weir outlets. These three elements are described below:
1. Storage units
Storage units in SWMM are modeled as nodes. They are similar to the conveyance-
system junction nodes introduced in Example 2 but have some fundamental
differences: storage volume is described by a Storage Curve, an Evaporation Factor
can be specified, and a Maximum Depth of storage must be defined.
Storage Curve: This curve defines the shape of a storage unit by describing
how the surface area within the unit varies with water depth. This curve is
integrated by SWMM to compute stored volume as a function of depth. It can
be specified to the model as either a functional equation or as a tabular curve
(area-depth pair data).
Evaporation Factor: To allow evaporation from the surface of a storage unit
one sets its Evaporation Factor to 1 and provides evaporation data to the
model using SWMM's Climatology Editor. The default value for this
parameter is 0, implying that evaporation is ignored.
Maximum Depth: The maximum depth of a storage unit must be defined and
should not be left at the default zero value. If the storage unit's depth is not
defined, the model will assume the storage unit has a zero depth even if a
storage curve has been assigned or a conduit is connected to the storage unit.
If the largest depth on the storage curve is below the maximum depth, the last
area value on the curve will be extended outward.
2. Orifices
SWMM's orifice-type link can be used to represent the opening along the side or
bottom of the storage unit that serves as an outlet. The upstream node of the orifice is
the storage unit while its downstream node would be a junction that connects it to a
downstream conduit. Orifice properties that need to be defined include its height
above the bottom of the storage unit (invert offset), its type (side or bottom orifice),
its geometry (rectangular or circular shape and the respective dimensions) and its
hydraulic properties (discharge coefficient and the presence/absence of a flap gate to
prevent backflows).
3. Weirs
SWMM's weir-type link can be used to represent the opening at the top of the storage
unit that serves as an overflow structure. As with the orifice, the upstream node of the
weir is the storage unit while its downstream node connects it to a downstream
conduit. Weir properties that need to be defined include the weir's height above the
bottom of the storage unit, its type (transverse, V-notch or trapezoidal), its geometry
and its hydraulic properties (discharge coefficients, end contractions and the
presence/absence of a flap gate to prevent backflows).
52
-------
Converting Node and Link Elements
Nodes need not be deleted and links redrawn to replace a node with a
storage unit or to define orifices and weirs in SWMM. Instead, they can be
converted. For example, a node can be converted to a storage unit by following the
procedure outlined below.
Convert a Node to a Storage Unit
1. Right-click on the node to be converted and choose "Convert to..." from the
pop-up menu as show in the figure below.
2. Select "Storage Unit" from the sub-menu that appears.
3. Open the Property Editor for the new storage unit and define its new name
(e.g. SUJ). The new unit is given the invert elevation and maximum depth of
the node from which it was converted.
4. Enter any additional properties needed to define the behavior of the storage
unit (such as its Storage Curve).
Convert a Conduit to an Orifice
Like nodes, links can be converted to other types of links. Follow the steps
below to convert the conduit connecting a storage unit to an outfall into an orifice.
1. Right-click on the conduit downstream of the storage unit and choose
"Convert to..." from the pop-up menu.
2. Select "Orifice" from the sub-menu that appears..
3. Open the Property Editor of the orifice and define its dimensions, invert
offset and discharge coefficient.
Convert an Outfall to a Node
4. If the storage unit (SU1) is to have more than one orifice, then the orifices
cannot be connected directly to an outfall (Ol). The outfall must be converted
into a node (Ol) in this case and a new outfall created (O2). Convert the
outfall using the same procedure described above.
53
-------
3.3 Model Setup
SWMM can be used to model storage facilities that capture runoff from different
design storms and release it to a receiving channel at a controlled rate. This example
demonstrates how the design of a storage pond is an iterative process in which the
dimensions of the pond and its outlets are changed to satisfy the design criteria and
constraints for the design storms considered. The three main steps used to design the
storage pond are:
1. Estimate the water quality capture volume (WQCV).
2. Size the storage volume and the outlet to control the release rate of the WQCV.
3. Size the storage volume and the outlet to control the peak runoff rates from the 2-, 10-
and 100-yr design storms.
The final design will be a storage unit with a shape specific to its location, rainfall
and climate conditions; a defined relationship between its surface area and storage depth;
and a multi-outlet structure designed to control different runoff events. Figure 3-2 shows
the schematic of a detention pond and its outlets designed to control a WQCV and the
peak discharges for three design storms. The stacked trapezoidal prism shape shown in
this figure will be used in this example; the upper prism will control the major storms
(10- and 100-yr) while the lower prism will control the minor storms (WQCV and 2-yr).
Note that the discharge for different storms is controlled by a combination of
orifices and weirs rather than a single unique outlet. Orifice 1 in Figure 3-2 controls the
release of the WQCV; orifices 1 and 2 control the release of the 2-yr storm; orifices 1, 2
and 3 control the release of the 10-yr storm and all the orifices combined together with
the weir (4) control the release of the 100-yr storm.
Estimation of the Water Quality Capture Volume
The WQCV is the critical runoff volume to be used in the design of stormwater
quality enhancement facilities. Detailed investigation based on calibrated long-term
runoff simulations is the preferred method to determine this volume for a given site (Guo
and Urbonas, 1996). However, several methodologies or "rules of thumb" have been
proposed to estimate the WQCV that are simpler to use but still reliable when long-term
records are not available (see for instance Guo and Urbonas, 1995, 1996 and 2002; Water
Environment Federation, 1998). This example will use the methodology proposed by the
UDFCD (2001). Figure 3-3 shows the curves defined in this methodology to estimate the
WQCV as a function of the tributary catchment's total imperviousness and the drain time
of the capture volume. The following steps are used to estimate the WQCV for the
detention basin being designed in this example:
54
-------
100-yr
10-yr
2-yr
WQCV ,
t ^
,
; ;
=
3
2
r 1
1
c
Figure 3-2. Schematic representation of a detention pond
0.50
0,45
0.40
s 0.30
I
0.25
0.20
0.15
0.10
0,05
0.00
40-hour Dram Tim
^
24-hour Drain Tim
.WQCV=a'(0.91i--1.1
6-hr drain time a = 0.7
12-hr drain time a =0.8
~~ 24-hr drain time a = 0.9
40-hr drain time a = 1.0
e|^
12-hour Drain Time
6-hour Drain Time
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9
Total Imperviousness Ratio (i = I,ซ,I100)
Figure 3-3. Water quality capture volume (UDFCD, 2001)
55
-------
1. First, determine the developed site's average Directly Connected Impervious Area
(DCIA). DCIA is the impervious area that is directly connected to the stormwater
drainage system; it does not include rooftops, patios, etc. that drain to lawns or other
pervious areas, and is smaller than the gross or total impervious area that is typically
estimated through aerial photography. These areas were previously estimated for each
of the seven subcatchments defined for the post-development site condition in
Example 1 and are presented in Table 3-2.
iTable^2._Post^eveloฃnien|_subcatchment data
Jnibcatchment SI S2 S3 S4 S5 S6 S7
Area(ac) 4.55 4.74 3.74 6.79 4.79 1.98 2.33
Imperviousness (%) 56.8 63 39.5 49.9 87.7 95 0
2. Next calculate the site's average percent imperviousness by weighting the
imperviousness of each subcatchment by its area and dividing by the total area
(28.94-ac) of the study area. The average percent imperviousness of the site
determined by this method is 57.1% ~ 57%.
3. The next step is to determine the WQCV in watershed inches. Assume that the
example site is located in Colorado's high plains near the foothills and that the
storage unit is to have a 40 hr drain time. From Figure 3-3 the corresponding WQCV
in watershed inches is 0.23 in. Thus the total water quality control volume is 28.94-
ac-0.23 in/12 = 0.555 acre-ft or 24,162 ft3.
4. If the design location were not in Colorado's high plains near the foothills, one would
need to adjust the WQCV determined from Figure 3-3. The curves shown in Figure 3-
3 are defined to control the 80th percentile runoff event and are appropriate for use in
Colorado's high plains near the foothills. For other locations, the WQCV from Figure
3-3 can be adjusted to obtain an appropriate volume, WQCVo, using Equation 3-1. In
this equation, d6 is the average precipitation depth of the runoff-producing storms.
Storm events for Equation 3-1 were defined for a 6-hour inter-event period and have a
minimum depth of 0.1 in. Figure 3-4 shows estimates of d$ for the contiguous United
States (UDFCD, 2001).
WOCV
WQCV0 = d6 y^^- (3_1)
6
Pond Geometry and Dimensions
The shape of the storage unit will depend on the regulations in the location where
the structure will be constructed. Generally, it is recommended that the distance between
the inlet and outlet of the facility be maximized; a length to width ratio of 2:1 to 3:1 is
adequate. This example will use a length to width ratio of 2:1, a WQCV depth (hi) of 1.5
ft, and a side slope of 4:1 (H:V). Figure 3-5 shows the geometry of the WQCV and
equations developed based on the length to width ratio (2:1) and the storage unit side
slope (4:1) that describe the unit's geometry. The steps used to determine the dimensions
of the WQCV are:
56
-------
(L50
0.40
0.50
3.SO
0.60
D.M
0,63
0,70
0.30
Figure 3-4. Average depth (inches) of runoff producing storms in the US (Driscoll, et al., 1989)
\
J
= Ij + 2 4&J = 2L3 + 2
WQCV
1 3 2 4
2L-
Figure 3-5. Geometry of the pond's WQCV
1 . Solve for L3 using the WQCV found in the previous section (24, 162 ft3) and hi equal
to the WQCV depth (1.5 ft). Rearranging the fifth equation listed in Figure 3-5 yields
the following quadratic equation
4L23 + 24^4 + (64hf - 1VWQCV I \ ) = 0
Solving for L3 gives L3 = 85.15 ft ~ 86 ft.
57
-------
2. Next solve for the other dimensions of the WQCV using LS and hi. From the first
equation from the top in Figure 3-5, L} = 170.3 ft ~ 171 ft, from the second equation
L2 = 184 ft, and from the third equation L4 = 98 ft.
3. Then define the storage curve for the WQCV portion of the storage unit. At 0 depth
the area is LrL3 = 14,706 ft2 while at the full depth of 1.5 ft the area is L2-L4 = 18,032
ft2. These pairs will be entered into the model in the following section together with
new points in the surface area-depth curve representing the shape shown in Figure 3-2
to control larger volumes.
Adding a Storage Unit to the Model
The Example2-Post.inp file will be used as a starting point to add a storage unit into the
model that represents the detention pond. The following steps are taken to define the
storage unit.
1. A new Storage Curve object named SU1 is created to represent the shape of the
storage unit.
2. The two previously determined depth-area points are entered into the Curve Editor
dialog for curve SU1. These two points are di = 0, AI = 14706 ft2 and d2 = 1.5 ft, A2 =
18032 ft2.
3. A new storage unit node, also named SU1, is placed onto the study area map as
shown in Figure 3-6, and is left disconnected from the drainage system. The
following properties are assigned to SU1', Storage Curve = Tabular; Curve Name =
SU1; Invert Elevation = 4956 ft (six feet lower than the outfall node elevation defined
in the previous examples); Maximum Depth and Initial Depth = 1.5 ft (the maximum
allowable depth defined to control the WQCV).
4. An additional node (J out), conduit (C out) and outfall node (O2) are added that will
connect the orifices and weirs draining the storage unit (SU1) to the outfall node
(O2). This must be done because it is not possible to connect more than one hydraulic
link to an outfall node in SWMM. The invert elevations of J out and O2 are set to
4954 ft to avoid backwater effects (again, details in the geometric definition of the
storage unit will depend on the local conditions and it is not the intention of this
example to cover these details) and C out is given a length of 100 ft and a roughness
of 0.01. Figure 3-7 shows the independent storage unit system, the tabular storage
curve SU1 for the WQCV and the storage unit's Properties table.
Initially, the storage unit and its WQCV orifice are modeled independent of the
watershed to size the WQCV orifice to drain in 40 hours. Although the storage unit and
the watershed are shown in the same input file in Figure 3-6, they will run as independent
systems in the model because they are not hydraulically connected. The location of the
pond in Figure 3-6 will be its final location in the model. The pond could have been
placed in the park area since there is significant open space for it, but for clarity it was
placed at the downstream end of the park.
58
-------
RainGage
Figure 3-6. Study area map with storage unit SU1
SU1
Or1
Cjjut
Storage Curve SU1
Tag
Inflow*
inert:
Et
Mat OffKh
Ponded fuea
Shape Cuv*
Uซfna*MBปrt rant of ttaqe i
Figure 3-7. Properties of storage unit SU1
4356
15
15
0
1000
0
0
59
-------
Sizing the WQCV Orifice
The next step is to design the pond outlet so that the entire WQCV is released
within 40 hours. The outlet will be an orifice connecting the storage unit to the
downstream outfall O2. This orifice could be located at the bottom or side of the storage
unit and be either circular or rectangular in shape. The following steps are used to size the
orifice so the WQCV drains in 40 hours.
1. A side orifice (Orl) is added between the storage unit (SU1) and the node (J out)
leading to the outfall node. It is given a rectangular shape and assigned an inlet offset
of zero so that its invert is the same as the storage unit. Its discharge coefficient is
assumed to be the default value of 0.65.
2. The simulation time step options are set as follows: reporting, wet-weather and
routing time steps to 15 seconds and the dry-weather time steps to 1 hr. The
simulation duration must be longer than 40 hours so that the performance of the
orifice can be properly evaluated; this example uses 72 hours.
3. The final dimensions of orifice Orl are determined by running SWMM several times
using Dynamic Wave flow routing while iteratively changing the orifice dimensions
until a size is found that drains the WQCV in approximately 40 hours. For each run,
the dimensions of the orifice are varied while keeping the initial water depth in the
storage unit at the depth of the WQCV, 1.5 ft. One can assume that the basin is
essentially empty once the water depth is 0.05 ft. Note that the runoff discharge
generated by the rainfall falling on the subcatchments does not affect the storage unit
during this part of the example because it is not connected to the drainage system.
Figure 3-8 shows the drainage time for three iterations as well as that for the final
design. Table 3-3 shows the dimensions assigned to the orifice by iteration. The final
orifice design has a height of 0.3 ft and a width of 0.25 ft. This small size is typical of a
WQCV orifice. That is why the orifice must be protected by a screen to prevent plugging
during the storm and maintenance must be done regularly to ensure the screen remains
free of debris.
Table 3-3. Design of the WQCV outlet (Orl)
Iteration
Height (ft)
Width (ft)
Inlet Offset (ft)
Discharge Coefficient
Drainage time (hrmin)
1
0.166
0.25
0
0.65
53:58
2
0.25
0.25
0
0.65
43:21
3
0.25
0.4
0
0.65
27:07
Final
0.3
0.25
0
0.65
40:12
60
-------
Iteration 1
Iteration 2
Iteration 3
Final
1.5
1.2
o 0.9
ฃ-
JS
a.
S 0.6
0.3
0
0 8 16 24 32 40 48 56 64 72
time (hr)
Figure 3-8. WQCV drainage times for the iterations shown in Table 3-3
Sizing the 2-yr Design Storm Orifice
The runoff volume generated by the 2-yr storm will be larger than the WQCV
volume designed for in the previous section. The volume of the storage unit must now be
enlarged and a new outlet must be defined. This new outlet, which will be placed at a
height of 1.5 ft above the basin floor, will begin to discharge when the runoff volume
from any storm just exceeds the WQCV. This outlet will control not only the peak runoff
rate of the 2-yr storm but also partially control the runoff rate from storms greater than
the 2-yr storm. The required increase in storage volume will be achieved by extending the
sides of the storage unit above the WQCV depth while keeping a lateral slope of 4:1
(H:V) as shown in the basin schematic, Figure 3-2. The following steps outline how the
storage unit is sized for the 2-yr design storm orifice.
1. The storage unit is first connected to the rest of the drainage system. This can be done
by changing culvert CJTs outlet to SU1 and deleting the original outfall node Ol.
Culvert Cll is given a downstream offset of 1ft so that for minor storms it has no
backwater but still has its crown below the top of the storage pond.
2. Next the size of the pond is enlarged for flood control by expanding its height while
keeping a constant slope (refer to Figure 3-2 for an illustration). This is done by
entering a new pair of surface area-depth pairs to the storage curve SU1. The values
for this new pair are: ds = 6 ft, A3 = 29583 ft2. The initial depth of the storage unit is
set to zero and its maximum depth to 6 ft to account for the new volume.
3. The model is then run for the 2-yr storm using only the WQCV orifice to determine
the maximum depth in the storage unit, the peak discharge of the WQCV orifice
(Orl), and the time it takes the storage unit to empty. The results show a maximum
61
-------
storage unit depth of 2.82 ft, a maximum orifice discharge of 0.64 cfs and an
emptying time of 56:23 (hrmin).
Based on the results in step 3, the peak outflow for the 2-yr storm can be increased
because the pre-development 2-yr peak runoff (4.14 cfs from Table 3-1) is larger than
the discharge through the WQCV orifice (0.64 cfs). Increasing this discharge is
advantageous because it will reduce the final volume of the storage required and save
in costs. To increase the 2-yr storm pond outflow, a second orifice (Or2) is added
directly above the WQCV depth (inlet offset =1.5 ft) as illustrated in Figure 3-2. This
orifice is assigned a rectangular shape with an inlet offset of 1.5 ft and a discharge
coefficient of 0.65. It should be drawn on the map with at least one intermediate
vertex so that it can be distinguished from the existing orifice Orl (see the sidebar
"Drawing Links as Polylines"}.
Drawing Links as Polylines
SWMM allows links to be drawn as
polylines containing any number of
straight-line segments that define the
alignment or curvature of the link. Once
a link has been drawn its interior points
can be added, deleted, and moved. This
feature is especially useful when two or
more links share the same set of end
nodes and would otherwise appear
directly on top of one another on the
map. The figure on the right shows how
polylines were used to draw the various
orifices and weirs that comprise the
outlet structure of a storage unit so that
they could be distinguished from one
another.
5. An initial estimate of 0r2's area A is made using the orifice equation:
(3-2)
with C = 0.65, Q = (4.14 - 0.64) cfs = 3.5 cfs and h = (2.84 - 1.5) ft = 1.34 ft. This
produces an orifice area of 0.58 ft2. Assume Or2 has an initial height of 0.58 ft and a
width of 1 ft.
Running the model with these dimensions for Or2 produces a discharge of 2.84 cfs.
This value is less than the target discharge (4.14 cfs). Therefore, iterations must again
be used to size orifice Or2 as was done for Orl until the combined peak discharge of
the two orifices is equal to or a little less than the 2-yr pre-development peak
discharge (4.14 cfs).
62
-------
7. To simplify the iterations for sizing Or2 its height is fixed at 0.5 ft and its width is
varied in 0.05 ft increments until the combined discharge of both orifices is close to
4.14 cfs. A size of 0.5 by 2 ft produces a peak discharge of 4.11 cfs and a maximum
storage unit depth of 2.21 ft. These dimensions will thus be used for the 2-yr orifice.
Sizing the 10-yr Design Storm Orifice
Up to this point the storage unit has been represented as a single trapezoidal
prism. This shape was determined for the WQCV back in the "Pond Geometry and
Dimensions" section and increased in size (keeping the 4:1 side slope) to contain the 2-yr
storm runoff. In this section the storage unit shape is redefined by adding an additional
trapezoidal prism over the minor storm prism to contain the 10- and 100-yr storm
volumes (See Figure 3-2). The steps below show how this is done to size the 10-yr storm
orifice:
1. The storage curve SU1 is modified by replacing the surface area-depth pairs d2,A2
and d3,A3 with the following three surface area-depth pairs: d2 = 2.22 ft, A2 = 19659
ft2; d3 = 2.3 ft, A3 = 39317 ft2 and d4 = 6 ft, A4 = 52644 ft2. Note that the new d2,A2
point is the highest point of the original shape (the maximum height of the 2-yr
storm) . The area of point 3 (A3 = 39317 ft2) doubles the area of point 2 (A2 = 19659
ft2). The depth, d3 is set 0.1 feet above d2 so that the change in cross sectional area is
not too abrupt. In practice this transitional area would have a slope of 2% so that
water in storage will drain into the smaller basin after the storm. The area of point 4 is
computed by extending the sides of the storage unit above the point 3 while keeping a
lateral slope of 4:1 (H:V).
2. The model is run with the 10-yr storm and the existing orifices to determine if a 10-yr
storm orifice is needed. The resulting maximum water depth is 3.20 ft and the
combined peak discharge from both existing outlets is 6.96 cfs. The pre-development
peak discharge for the 10-yr storm is 7.34 cfs which means that the storage unit
volume can again be decreased by adding another orifice.
3. A new 10-yr storm orifice (Or3} is added directly above the depth of the volume
designed to control the 2-yr storm runoff (inlet offset = 2.22 ft). As with Or2, Or3 is
drawn with intermediate vertices so that it can be seen easily on the system map. The
orifice equation (3-2) is used to estimate its required area. For C = 0.65, Q = (7.34 -
6.96) cfs = 0.38 cfs and h = (3.20 - 2.22) ft = 0.98 ft the resulting orifice area is 0.073
ft2. A height of 0.25 ft and a width of 0.25 ft are used as an initial estimate of the
orifice's size.
4. When the model is run with the 10-yr storm for this size of Or3, the combined
discharge is 7.22 cfs. Because this discharge is less than the pre-development
discharge (7.34 cfs), the orifice's width is increased to 0.35 ft and the model is re-run.
The new combined discharge is 7.32 cfs and the maximum depth in the storage unit is
3.17 ft. This is sufficiently close to the target discharge to accept this orifice size
(height = 0.25 ft, width = 0.35 ft).
63
-------
Designing the 100-yr Weir
The model can now be run with the 100-yr storm using the combined WQCV, 2-
yr and 10-yr orifices to determine if the 100-yr weir is needed. The peak discharge of
these combined orifices for the 100-yr storm is 12.57 cfs, which is not enough to pass the
100-yr storm's runoff (31.6 cfs), and the storage unit floods. A weir will be designed to
control this extreme event so that the pre-development discharge is matched and the total
water depth in the pond does not exceed the 6 ft depth that was given as the maximum
depth for safety reasons. This is accomplished as follows:
1. A new weir link Wl, drawn with intermediate vertices, is added between the storage
unit and the node (J out) leading to the final outfall. It is specified as a transverse
weir whose inlet offset is 3.17 ft above the storage unit bottom (the maximum depth
reached by the volume controlling the 10-yr storm runoff), and whose discharge
coefficient is 3.3. The height of the weir opening is set at 2.83 ft which is the distance
between the volume controlling the 10-yr storm runoff and the maximum depth of the
storage unit.
2. The weir equation (3-3) is used to determine an initial width L for the weir:
Q = CLh312 (3-3)
Using Q = (31.6 - 12.57) cfs = 19.03 cfs, C = 3.3 and h = 2.83 ft produces a width of
3.43 ft-3.45 ft.
3. With weir dimensions of height = 2.83 ft, width = 3.45 ft and invert offset = 3.17 ft
the model is run for the 100-yr storm. The resulting peak total discharge from the
storage unit is 42.4 cfs which exceeds the target flow of 31.6 cfs.
4. Step 3 is repeated with successively smaller weir widths until a combined discharge
close to 31.6 cfs is obtained. A width of 1.75 ft produces a combined 100-yr
discharge of 31.2 cfs and a maximum depth of 5.42 ft in the storage unit.
5. The final step is to insure that adequate freeboard is maintained in the storage unit.
The current design provides 6.0 - 5.43 = 0.53 ft. The required amount will depend on
local design guidelines. For example, the UDFCD (2001) requires a freeboard of 1
foot above the maximum water surface elevation when the weir is conveying the
maximum discharge.
3.4 Model Results
The final SWMM model for the post-development site with the detention pond is
shown in Figure 3-9 and its input file is named ExampleS.inp. Table 3-4 summarizes the
characteristics of the different discharge elements included in the pond's outlet, which are
illustrated in Figure 3-10.
64
-------
497J
RainGage
Figure 3-9. Final design of the detention pond and outlet structure
Orifice & Weir
Dimensions
2.83 ft
0.25 ft
_*_
0.5ft
i 0.3 ft
Hmax=5.43ft
Offsets
6ft
-K 1 3.17ft
T 2 22 ft I
T 1 1
Figure 3-10. Detail of the pond outlet structure
65
-------
Table 3-4. Characteristics of the jjond's outlet structure
Type of Event , Height Width Invert Discharge
Element Controlled pe h (ft) b (ft) Offset Z (ft) Coefficient
ID
Orl
Or2
Or3
Wl
Orifice
Orifice
Orifice
Weir
WQCV
2-yr
10-yr
100-yr
Side
Rectangular
Side
Rectangular
Side
Rectangular
Rectangular
0.3
0.5
0.25
2.83
0.25
2
0.35
1.75
0
1.5
2.22
3.17
0.65
0.65
0.65
3.3
The outflow hydrographs produced by this final design for the detention pond
were compared against those resulting with no control as well as against the pre-
development discharge targets. Those target discharges were 4.14, 7.34 and 31.6 cfs for
the 2-, 10- and 100-yr storms, respectively. The post-development controlled
hydrographs were generated using the input file for the final design of this example
(ExampleS.inp), while the post-development uncontrolled discharges were generated
using the final design developed in Example 2 (Example2-Post.inp). The pre-
development hydrographs were generated using the final design developed in Example 1
for the pre-development watershed (Examplel-Pre.inp). The models were run using a 15
second time step for reporting, wet-weather runoff and flow routing and a 1 hr dry
weather runoff time step.
The resulting sets of hydrographs are shown in Figures 3-11 through 3-13. Once
again a spreadsheet program was used to combine results from different SWMM runs
onto one graph. They verify that the detention pond was able to control post-development
peak discharges from the site to their pre-development levels. Note, however, that the
storage unit had no effect on reducing the total volume of post-development runoff that
resulted from the large increase in impervious area.
3.5 Summary
This example showed how SWMM could be used to design a detention pond and
its outlet structure to provide both a water quality capture volume (WQCV) and peak
runoff control. The WQCV was designed to provide a 40 hour drawdown time to satisfy
water quality treatment requirements while the peak runoff goal was to limit the
maximum post-development discharges for the 2-, 10- and 100-yr storms to their pre-
development values. The key points illustrated in this example were:
66
-------
2-yr, uncontrolled
2-yr, controlled
2-yr, pre-developed
time (hr:min)
Figure 3-11. Outlet hydrographs for the 2-yr storm
10-yr, uncontrolled
10-yr, controlled
10-yr, pre-developed
time (hr:min)
Figure 3-12. Outlet hydrographs for the 10-yr storm
67
-------
100-yr, uncontrolled
100-yr, controlled
100-yr, pre-developed
1
1
1
o
o
time (hr:min)
Figure 3-13. Outlet hydrographs for the 100-yr storm
1. The WQCV's outlet structure can be designed by using a full storage unit (volume =
WQCV) disconnected from the drainage system and an orifice whose dimensions are
varied until the storage is drained in a time equal to that defined by local regulation
(40 hr in this example).
2. The dimensions of the other component of the outlet structures (e.g. orifices and
weirs) used to control peak flows can be designed sequentially. The maximum water
depth reached using one design storm is the location of the invert offset of the orifice
or weir used to control the next larger design storm.
3. The orifice and weir equations are useful for making initial estimates of an outlet's
dimensions.
4. Although detention storage is effective in controlling peak runoff rates it has no effect
on reducing runoff volume.
This example will be extended to include water quality treatment associated with
the storage pond in Example 6 and continuous simulation in Example 9.
68
-------
Example 4. Low Impact Development
This example demonstrates how to model two Low Impact Development (LID)
control alternatives, filter strips and infiltration trenches. The detention pond modeled in
Example 3 is an example of a subdivision-scale drainage control structure. The LIDs in
this example are hydrologic source controls that operate on a smaller scale and rely
heavily on infiltration and distributed small-scale storage to reduce the overall runoff
volume from a watershed and control water quality.
SWMM is better suited to simulating some types of hydrologic source control
techniques than others. Filter strips and infiltration trenches are two such source controls.
This example will illustrate how they can be represented in SWMM by applying them to
the same catchment area studied in the previous Examples 1, 2 and 3.
4.1 Problem Statement
In Examples 1 and 2, runoff estimates for the 29 acre residential development
shown in Figure 4-1 were made without any source controls in place. In this example, the
effects that infiltration trenches (IT) and filter strips (FS), two commonly used LIDs, have
on the site's runoff will be examined. As illustrated in Figure 4-1, four infiltration
trenches will be placed at each side of the east-west street in the upper part of the study
area. Additionally, filter strips will be used to control the runoff from lots S, M and M2,
located in the southwest section of the site. These strips will be built along the sidewalks
so they control runoff from the lots before it reaches the gutters.
In Example 3 the quantitative objective for the design of a detention pond was to
reduce the post-development watershed discharge to pre-development levels. The LIDs
modeled here will not be designed to accomplish a specific quantitative objective but
rather to achieve a general reduction in runoff volume to help meet sustainability goals
and place a lower burden on stormwater controls further downstream in the basin. The
performance of the LIDs for the 2-, 10- and 100-yr storms will be analyzed.
4.2 System Representation
The two LIDs modeled here are filter strips and infiltration trenches. Guidance in
the representation of these and other LIDs can be found in Huber at al. (2006).
69
-------
497;
4975
Natural Channels
or Sw ales
Culverts
Filler Strip
111 lilir.ni<>n Trench
Figure 4-1. Post-development site with LIDs in place
Filter Strips
Filter strips are grassed or vegetative areas through which runoff passes as sheet
flow. They do not effectively reduce peak discharges but are effective in removing
particulate pollutants for small storms (< 1 year storm) (Akan and Houghtalen, 2003).
Flat slopes (<5%) and low to fair permeability (0.15 to 4.3 mm/h or 0.006 to 0.17 in/h) of
natural subsoil are required for their effective operation (Sansalone and Hird, 2003).
SWMM does not have a unique visual object to represent filter strips but they can be
represented as a pervious subcatchment that receives runoff from an upstream
subcatchment as illustrated in Figure 4-2. The two most important processes that must be
simulated with filter strips are infiltration and storage. A filter strip can be simulated as a
100% pervious subcatchment whose geometry (area, width and slope) is obtained directly
70
-------
from the field. This subcatchment receives water from an upstream contributing area
(impervious or semi-impervious) and drains to a conduit representing the gutter or street.
The infiltration for a filter strip can be simulated using any of SWMM's infiltration
options.
Contributing
area
Filter Strip
\7
Subcatchment 2
Figure 4-2. Schematic representation of a filter strip
Infiltration Trenches
Infiltration trenches are excavations backfilled with stone aggregate used to
capture runoff and infiltrate it to the ground (Guo, 2001). Subsoil with a minimum
permeability of 13 mm/h (0.5 in/h) is required for a good performance (U.S. EPA, 1999).
The most important processes that must be simulated for an infiltration trench are
infiltration, storage and the water flow along the trench. Figure 4-3 shows a workable
SWMM representation of an infiltration trench. It consists of a rectangular, fully pervious
subcatchment whose depression storage depth equals the equivalent depth of the pore
space available within the trench.
71
-------
w
Flow depth, ' *^-^m ffr
created by \ /-I-h ' DePซss'ฐซ "o'^-
runoff ' Jj "ฐ
Subcatcliment 2
h= depression storage
Figure 4-3. Schematic representation of an infiltration trench
4.3 Model Setup - Filter Strips
Figure 4-1 shows the locations of the filter strips (FS) to be modeled in this
example. The initial SWMM file to which the LIDs will be added is Example2-Post.inp,
which includes both subcatchments and a runoff conveyance system in it. Figure 4-1 will
be used as the backdrop of the new model to help facilitate the placement of the filter
strips. The image file for this backdrop is named Site-Post-LID.jpg. Table 4-1 lists the
subcatchment to which each filter strip belongs and the length of each strip.
Table 4-1. Subcatchments containing filter strips
Subcatchment controlled
S3
S3
S3
S4
S4
S4
S3 and S4
Filter Strip
FS 1
FS2
FS3
FS4
FS5
FS6
FS7
Filter Strip Length (ft)
410
105
250
359
190
345
375
72
-------
Sub-Area Routing in SWMM
An alternative method to modeling LIDs, other than creating additional
subcatchments as used in this example, is to use Sub-Area Routing. Each
subcatchment modeled in SWMM is composed of two sub-areas, an impervious
sub-area and a pervious sub-area, and options called Sub-Area Routing and Percent
Routed appear in their Properties tables.
Sub-Area Routing has three options: Outlet, Impervious and Pervious. The
Outlet option ((a) in the figure below) routes the runoff from both sub-areas directly
to the subcatchment's outlet. The Pervious option ((b) in the figure below) routes
the runoff from the impervious sub-area across the pervious sub-area and then to
the outlet, and the Impervious option ((c) in the figure below) routes the runoff
from the pervious sub-area across the impervious sub-area and then to the outlet.
When the runoff from the impervious surface is routed across the pervious
surface ((b) in the figure below), some of the runoff is lost to infiltration and
depression storage in the pervious sub-area. The graph below shows the typical
effect on runoff these three sub-area routing methods have when applied to a single
subcatchment. Note that because the runoff produced by the pervious area in this
case is negligible, the "100% routed to Outlet" and the "100% routed to
Impervious" cases are practically the same.
3.5 -
3 -
2.5 -
stS
& 2 -
0-
1.5 -
1 -
0.5 -
^ ]0o% routed to
I) Outlet, 2-yr
100% routed to
Pervious, 2-yr
1| 100% routed to
TA
TI\
-* 1 \
' / ^v-
-/ / _S: --
0:00:00 1:12:00 2:24:00 3:36:00 4:48:00
Time (hr)
(b) pervious Area (c) Impervious Area
Routing Routing
The Sub-Area Routing - Pervious option can be used to model LIDs. This is
done by representing the LID as the pervious sub-area, setting the subcatchment's
pervious values to those of the LID, routing the runoff from the impervious sub-
area of the subcatchment to the pervious subarea, and defining the Percent Routed
to represent the percentage of impervious surface connected to the LID.
Modeling LIDs with this method implies that the entire subcatchment's
pervious surface represents the LID. This approach is not as flexible as the one
presented in this example because the slope and width of the LID must be those of
the subcatchment, which is not always the case.
73
-------
The filter strips to be constructed are identified in Table 4-1 with letters FS and a
number. In the model being built some of the strips will be aggregated together as they
are represented as new subcatchments. These filter-strip subcatchments will be identified
as "S_FS_number".
To better estimate the runoff treated by the filter strips, subcatchments S3 and S4
will be discretized further. S3 will be divided into three subcatchments S3.1, S3.2 and
S3.3 and S4 into four subcatchments S4.1, S4.2, S4.3 and S4.4 as shown in Figure 4-4. It
will be necessary to re-estimate the average overland runoff length and new
subcatchment widths, slopes and percents imperviousness to place in the model. Refer to
Example 1 for help in determining the new subcatchments' hydrologic properties. The
slopes of the new subcatchments representing lots will be 2%. The same Horton
infiltration rates already defined in the previous examples will be used here: 4.5 in/hr for
the maximum, 0.2 in/hr for the minimum and 6.5 hr"1 for the decay constant. With the
addition of these new subcatchments and the re-discretization of some of the original
ones, it is necessary to define new channels and junctions to connect them to the drainage
system. These new elements are illustrated in Figure 4-4 (pipes in red and junctions in
blue). Table 4-2 lists the properties of the new junctions and Table 4-3 does the same for
the new conduits.
Table 4-4 summarizes the properties of the new subcatchments. The outlets of the
new subcatchments (S3.2, S3.3, S4.2, and S4.3) are not junctions but filter strips
represented as subcatchments. This cascade layout used in the SWMM model is shown in
Figures 4-5 for subcatchments S3.1, S3.2 and Figure 4-6 for subcatchments S4.1, S4.2,
S4.3andS4.4.
The length of the filter strips (i.e., the overland flow length existing between lots
and streets) will be 4 feet and the slope of each strip along this length is the same as the
typical lot slope of 2%. The widths of the strips (perpendicular to the overland flow
direction) is computed directly from the map with the Ruler tool as explained in Example
1 and are given in the last column of Table 4-5.
Table 4-2. Properties of the new junctions
New Junction Invert Elevation (ft)
J15
J16
J17
4974.5
4973.5
4973.5
New Inlet Outlet Length Type of
Conduit Node Node (ft) Section1
C15 J15 J3 444.75 Swale
C16 J17 J5 200.16 Swale
C17 J16 J7 300.42 Gutter
Manning Maximum Bottom Left Right
Coefficient Depth (ft) Width (ft) Slope Slope
0.05 3 555
0.05 3 555
0.016 1.5 0 0.0001 25
1 Type of section based on those defined in Example 2
74
-------
497S
4973
4972
4969
S FS 1
J
S_FS_2
S_FS_4
Original boundaries
of S3 and S4
Filter strips
Open-space sub.
Urbanized sub.
Mixed use
New junctions
^^ New conduits
Figure 4-4. Re-discretization of subcatchments S3 and S4
496tt
Table 4-4. Properties of the subcatchments derived from S3 and S4
New ,, ,
, , , Outlet
Subcatchment
S3.1
S3.2
S3.3
S4.1
S4.2
S4.3
S4.4
J3
S_FS_1
S_FS_2
J6
S_FS_3
S_FS_4
J9
Area (ac)
1.29
1.02
1.38
1.65
0.79
1.91
2.40
Width (ft)
614
349
489
580
268
657
839
Slope (%) Imperviousness (%)
4.7
2
2
5
2
2
2
0
65
58
0
70
65
69
75
-------
Table 4-5. Widths of the filter strip subcatchments
Filter
Strip
Subcatchment
Controlled
Composition
Partial Widths (ft)
Total Widths
(ft)
S_FS_1
S_FS_2
S_FS_3
S FS 4
S3.2
S3.3
S4.2
S4.3
FS 1
FS2 + FS3+partFS7
PartFS?
FS4 + FS5+FS6
410
105+250+108
267
359+190 + 345
410
463
267
894
S FS 1
J15
SKI
Figure 4-5. Representation of subcatchments S_FS_1, S3.1, and S3.2
S FS 4
S_FS_4
Figure 4-6. Representation of subcatchments S_FS_3, S_FS_4, S4.1, S4.2, and S4.3
Table 4-6 summarizes the characteristics of the subcatchments used to model the
filter strips in this example. The areas of the filter strips are calculated from the lengths of
the strips measured in the model and the flow path length (4 ft) used in this example. This
76
-------
means the areas calculated by Auto-Length for these subcatchments will be replaced with
those calculated in Table 4-6.
Table4-6.Propertiesoft^
Subcatchment
S
S
S
S
_FS
_FS
_FS
_FS
_1
_2
_3
_4
Upstream ,, ,
c , . . . Outlet
Subcatchment
S3
S3
S4
S4
.2
o
.5
.2
.3
J15
J17
J16
J16
Area1
(ft2)
1640
1852
1068
3576
Area
(ac)
0.038
0.043
0.025
0.082
Width2
(ft)
410
463
267
894
Slope
(%)
2
2
2
2
Depression3
Storage (in)
0.3
0.3
0.3
0.3
Once the filter strip properties shown in Tables 4-5 and 4-6 are added to the
model, all filter strips are assigned a zero percent imperviousness, an impervious
roughness of 0.015 and an impervious depression storage of 0.06 in (although neither of
the latter properties are used due to 0% imperviousness); a pervious roughness of 0.24,
and a pervious depression storage of 0.3 in. These last two values are the same as those
for the rest of the pervious areas within the watershed. Finally, both the maximum and
minimum infiltration rates for Horton infiltration are set to the same value, 0.2 in/hr,
which is the minimum infiltration rate of the soil in the study area. This is a conservative
approach that takes into account possible losses to infiltration capacity as well as the fact
that soil may be saturated at the beginning of a new storm.
All subcatchments must have a rain gage assigned to them in order for the
SWMM model to run. However, no rain should fall on the subcatchments representing
the filter strips because their areas are considered part of the subcatchments within which
they were placed. Thus, a new Time Series called "M//7" is defined with a single time and
rain value equal to zero. A new rain gage also called "Nuir is created and linked to the
time series "TVw//". This gage is defined to be the rain gage of all the filter strip
subcatchments, while the subcatchments that generate runoff are connected to the original
rain gage "RainGage", as in the previous examples.
4.4 Model Setup - Infiltration Trenches
The infiltration trenches (IT) included in this example are located in the north part
of the study area (see Figure 4-1). To model these devices subcatchments SI and S2 are
first divided into six smaller subcatchments as shown in Figure 4-7. Their areas are
determined automatically by using the SWMM's Auto-Length tool. Their widths are
based on an assumed urban runoff length of 125 ft (measured from the back of the lot to
the street), as discussed in Example 1. The imperviousness of each newly sub-divided
Subcatchment is based on the type of land use (M, L or DL) as listed in Table 1-6 of
Example 1. Table 4-7 summarizes the properties of the newly sub-divided
subcatchments. The remaining properties not shown in the table (Horton infiltration,
pervious depression storage, etc.) are the same as those given to the original
subcatchments.
77
-------
4973
S FS 1
S FS 2
S_FS_4
Original boundaries
Filter strips '
Open-space sub.
Urbanized sub.
Mixed use
Figure 4-7. Re-discretization of subcatchments SI and S2 for infiltration trenches
Table 4-7. Properties of the subcatchments derived from SI and S2
New ,, , Area
, , , Outlet , ,
Subcatchment (ac)
Sl.l
S1.2
S1.3
S2.1
S2.2
S2.3
J12
S_IT_1
S_IT_2
J14
S_IT_3
S_IT_4
.21
.46
.88
.30
.50
.88
Width
(ft)
422
509
655
453
523
655
Slope
(%)
2
2
2
2
2
2
Imperviousness
(%)
65
65
45
45
70
70
The infiltration trenches added to the model are identified in Table 4-8. Their
names begin with the letters IT followed by a number. The location of these infiltration
trenches is shown in Figure 4-7. The four infiltration trenches are added into the model as
78
-------
sub catchments and have an S_ prefix added to their name. They serve as the outlets of
subcatchments SI.2, SI.3, S2.2 and S2.3 as shown in Table 4-9. The outlets of the other
new subcatchments, Sl.l and S2.7, are junctions .772 and J7ฅ, respectively.
When sizing the infiltration trenches, note that the definition of length and width
is opposite of the manner in which length and width were defined for filter strips. The
length of each infiltration trench is listed in Table 4-8. The trench area can be computed
using this length and assuming a width of 3 ft. The widths and areas of each infiltration
trench are listed in Table 4-9. Each trench is modeled as having a depression storage
depth of 1.2 ft (below this depth water is infiltrated, while above this depth there is flow
along the trench). The actual storage depth of the infiltration trenches is 3 ft, but when the
40% porosity of the 1-1/2 in. diameter gravel filling the trenches is considered, the depth
of the trenches available to store the water to be infiltrated is reduced to 40% of the actual
trench depth. Table 4-9 summarizes the properties assigned to the infiltration trench
subcatchments.
Table 4-8. Subcatchments containing infiltration trenches
Subcatchment Controlled Infiltration Trench
SI IT1
SI IT 2
S2 ITS
S2 IT 4
Trench Length (ft)
450
474
450
470
Table 4-9. Properties of the infiltration trench
Subcatchment
S_IT_1
S_IT_2
S_IT_3
S IT 4
Upstream
Subcatchment
S1.2
S1.3
S2.2
S2.3
Outlet
Jl
Jl
J13
J13
subcatchments
Area1
(ft2)
1350
1422
1350
1410
Area1
(ac)
0.031
0.033
0.031
0.032
Width
(ft)
3
3
3
3
Slope2
(%)
0.422
0.422
0.444
0.468
Depression3
Storage (in)
14.4
14.4
14.4
14.4
1 Area = length of the trench x 3 ft width perpendicular to the flow direction.
2 Slope as computed from the site map.
3 This corresponds to the effective depth of the trench, 0.4 x 36 in.
The slopes of the infiltration trenches are calculated directly from the site map,
their imperviousness is 0%, their Manning's roughness coefficient is 0.24, and their
storage depths are accounted for by setting their depression storage to the effective
storage depth of 14.4 in (1.2 ft). As was done for the filter strips, a constant (minimum)
soil infiltration capacity of 0.2 in/hr will be used for the infiltration trenches. This ignores
any horizontal infiltration that might occur through the sides of the trench. Also the rain
gage assigned to each trench Subcatchment is the newly created "M///" rain gage so that
no rainfall occurs directly over the trench's area.
79
-------
This trench design assumes there is no barrier affecting the flow of water into the
trench. A layer of grass and soil above the 1-1/2 in. diameter gravel, for example, might
slow the rate of flow into trench; causing a smaller quantity of flow to be treated by the
trench. This will be touched on again in the results section of this example.
To complete the model it is necessary to define the additional channels and
junctions that connect the newly added subcatchments to the drainage system. These new
elements are illustrated in Figure 4-7 (conduits in red and junctions in blue). Table 4-10
lists the properties of the new junctions and Table 4-11 does the same for the new
conduits.
New Junction Invert Elevation (ft)
J12
J13
J14
4973.8
4970.7
4972.9
New Inlet Outlet Length Type of
Conduit Node Node (ft) Section1
C12 J12 Jl 281.7 Swale
C13 J14 J13 275.5 Gutter
C14 J13 J2 157.48 Gutter
Manning Maximum Bottom Left Right
Coefficient Depth (ft) Width (ft) Slope Slope
0.05 3 555
0.016 1.5 0 0.0001 25
0.016 1.5 0 0.0001 25
1 Type of section based on sections defined in Example 2
4.5 Model Results
The final model with all LIDs included can be found in the file Example4.inp. It
was run under Dynamic Wave flow routing using a wet runoff time step of 1 minute, a
reporting time step of 1 minute, and a routing time step of 15 s for each of the three
design storms. Figure 4-8 compares the resulting influent and effluent runoff hydrographs
for filter strip S FS1 for each of the design storms. Figure 4-9 does the same for
infiltration trench S IT' 1. Results for the other LIDs look similar to these. Tables 4-12
and 4-13 list runoff coefficients for each filter strip and infiltration trench, respectively.
As used here, the runoff coefficient is the ratio of the effluent-runoff volume flowing out
of the LID to the influent-runoff volume flowing into the LID. The fractional reduction in
runoff volume provided by the LID is simply 1 minus the runoff coefficient.
80
-------
2-yr: Inflow
10-yr: Inflow
100-yr: Inflow
2-yr: Outflow
10-yr: Outflow
100-yr: Outflow
0:00 0:15 0:30 0:45 1:00 1:15 1:30 1:45 2:00 2:15 2:30 2:45 3:00
Time (hr:min)
Figure 4-8. Influent and effluent hydrographs for filter strip S_FS_1
2-yr. Inflow
10-yr: Inflow
100-yr: Inflow
2-yr: Outflow
ฐ 10-yr: Outflow
100-yr: Outflow
o
0:00 0:15 0:30 0:45 1:00 1:15 1:30 1:45 2:00 2:15 2:30 2:45 3:00
Time (hr:min)
Figure 4-9. Influent and effluent hydrographs for infiltration trench S_IT_1
81
-------
Table4-12.
Runoff Coefficient
Filter Strip _____
2-year Storm 10-year Storm 100-year Storm
S_FS_1 0.95 0.98 0.99
S_FS_2 0.95 0.98 0.99
S_FS_3 0.96 0.98 0.99
S FS 4 0.94 0.97 0.99
Table 4-13. Runoff coefficients for infiltration trenches
1nEltntion^^^_^SS!L^S^S^L
Trench 2-year Storm 10-year Storm 100-year Storm
S_IT_1 0.45 0.73 0.89
S_IT_2 0.35 0.71 0.90
S_IT_3 0.51 0.75 0.90
S IT 4 0.59 0.79 0.92
It is apparent that filter strips provide negligible runoff control, with the outflow
rate very nearly equaling the inflow rate for all storm magnitudes. This confirms what
was mentioned in section 4.2 regarding the utility of filter strips, i.e. they are primarily
pollutant removal devices and provide no benefit in controlling runoff flow rates or
volumes. In contrast, all of the infiltration trenches provide significant reductions in
runoff volume, particularly for the smaller rainfall events. It should be remembered,
however, that the trenches in this example do not have grasses planted above the gravel
backfill. Adding such a vegetative layer to the trench may reduce its effectiveness,
depending on how it is designed.
Figure 4-10 compares the discharges simulated at the outlet of the study area for
each design storm (2-, 10- and 100-yr return period) both with and without LIDs. For
each design storm LIDs reduce both runoff volumes and peak discharges. As the storm
event becomes larger, LIDs become less effective and the attenuation of their volumes
and peak-discharges is reduced. These percent reductions in outlet volumes and peaks are
compared in Figure 4-11. It shows how the benefit of LID controls decrease with
increasing size of storm.
82
-------
2-yr: no LIDs
10-yr: No LIDs
100-yr: No LIDs
2-yr: LIDs
10-yr: LIDs
100-vr: LIDs
000 0:15 0:30 0:45 100 1:15 1:30 1:45 2:00 2:15 2:30 2:45 3:00
Time (hrrmin)
Figure 4-10. Comparison of outlet discharges with and without LID controls
40
35
1 Peak Volume
30
O
o
u
20
0*
10
Q
2-yr Storm
10-yr Storm
100-yr Storm
Figure 4-11. Percent reduction in outlet peak flows and runoff volumes with LIDs
83
-------
4.6 Summary
This example illustrated how SWMM can be used to evaluate two types of Low
Impact Development (LID) alternatives, filter strips and infiltration trenches. The key
points illustrated in this example were:
1. A filter strip can be modeled as a rectangular, 100% pervious subcatchment with a
constant infiltration rate.
2. An infiltration trench can be modeled as a rectangular, 100% pervious subcatchment
with a constant infiltration rate whose depression storage is the effective pore volume
depth of the trench.
3. Modeling these types of LIDs can require a finer level of subcatchment discretization
to properly account for their localized placement.
4. Infiltration trenches (without a top soil layer) are more effective than filter strips in
reducing runoff volumes and peaks.
5. The effectiveness of LIDs at reducing runoff volumes and peaks decreases with
increasing size of storm event.
Although this example used a series of design storms to evaluate LID
performance, a more accurate estimate of their stormwater control capabilities would
require that a continuous long-term simulation be run for the site. Using several years'
worth of actual rainfall inputs would allow the model to properly account for the
variation in antecedent soil conditions between storm events. This factor becomes a
critical concern when infiltration-based controls are being considered. Example 9 in this
manual illustrates how to perform such a continuous simulation.
84
-------
Example 5. Runoff Water Quality
This example demonstrates how to simulate pollutant buildup and washoff in an
urban catchment. The influence of different land uses on pollutant buildup is considered
and both Event Mean Concentrations (EMCs) and exponential functions are used to
represent the washoff process.
Surface runoff quality is an extremely important, but very complex, issue in the
study of wet-weather flows and their environmental impacts. It is difficult to accurately
represent water quality within watershed simulation models because of a lack of
understanding of the fundamental processes involved as well as a lack of sufficient data
needed for model calibration and validation. SWMM has the ability to empirically
simulate nonpoint source runoff quality as well as water quality treatment (an example of
which will be shown in Example 6). It provides a flexible set of mathematical functions
that can be calibrated to estimate both the accumulation of pollutants on the land surface
during dry weather periods and their release into runoff during storm events. The same
study area used in Examples 1 through 4 will be used to illustrate how these functions can
be applied to a typical urban catchment.
5.1 Problem Statement
The 29 acre urban catchment and drainage system presented in Example 2 will be
extended to include water quality modeling. Pollutant buildup, washoff and routing will
be simulated in order to estimate the quality of the water released at the catchment outlet
under post-development conditions with no runoff controls applied (meaning no BMPs or
flow detention in the system). The study area site is shown in Figure 5-1 and the input file
that will be modified to include water quality is named Example2-Post.inp.
Examination of long precipitation records reveals that most storms are quite
small. For instance, in Example 3 the water quality capture volume (WQCV) of a
detention pond located in the Colorado high-plains near the foothills was estimated to be
only 0.23 inches. (See Example 3 for a methodology to calculate the WQCV in other
areas in the country.) This volume corresponds to a depth that is exceeded by only 1 in 4
storms and is only 25% of the 2-yr design storm that was used in the previous examples
(1.0 in). Therefore, small-sized, frequently occurring storms account for the predominant
number of recorded events. It is these storms that result in significant portions of
stormwater runoff and pollutant loads from urban catchments (UDFCD, 2001).
To explore the effect of storm volume on pollutant loading, this example will
compute runoff loads produced by two smaller-sized 2-hour storms with volumes of 0.1
in. and 0.23 in. respectively. These loadings will be compared against those generated
from the 2-year design event storm used in the previous examples whose volume is 1.0
in. The time series of intensities at five minute intervals for each of the two smaller
storms are shown in Table 5-1.
85
-------
4973
ainGage
Figure 5-1. Post-development site with no runoff controls
Table 5-1. Rainfall time series for the 0.1 and 0.23 inch events
Time
(min)
0:00
0:05
0:10
0:15
0:20
0:25
0:30
0:35
0:40
0:45
0:50
0:55
0.1 in Storm
(in./h)
0.030
0.034
0.039
0.065
0.083
0.160
0.291
0.121
0.073
0.043
0.036
0.031
0.23 in Storm1 Time
(in./h) (min)
0.068 1:00
0.078
0.089
0.150
0.190
0.369
0.670
0.277
0.167
0.099
0.082
0.071
:05
:10
:15
:20
:25
:30
:35
:40
:45
:50
:55
0.1 in Storm
(in./h)
0.020
0.019
0.018
0.017
0.017
0.016
0.015
0.015
0.014
0.014
0.013
0.013
0.23 in Storm1
(in./h)
0.047
0.045
0.042
0.040
0.040
0.038
0.035
0.035
0.033
0.033
0.031
0.031
1 0.23 in. corresponds to the WQCV.
86
-------
These new hyetographs are defined in SWMM using its Time Series Editor. The
names of the new rainfall series will be 0.1-in and 0.23-in, respectively. They will be
used by the model's single rain gage in addition to the 2-yr, 10-yr and 100-yr storms used
in previous examples. This example will only employ the 2-yr storm along with the 0.1
in. and 0.23 in. events.
5.2 System Representation
SWMM employs several specialized objects and methods to represent water
quality in urban runoff. These tools are very flexible and can model a variety of buildup
and washoff processes, but they must be supported by calibration data to generate
realistic results. The following is a brief description of the objects and methods used by
SWMM to model water quality.
Pollutants
Pollutants are user-defined contaminants that build up on the catchment surface and
are washed off and transported downstream during runoff events. SWMM can
simulate the generation, washoff and transport of any number of user-defined
pollutants. Each defined pollutant is identified by its name and concentration units.
Pollutant concentrations in externally applied water sources can be added directly to
the model (e.g. concentrations in rain, groundwater and inflow/infiltration sources).
Concentrations generated by runoff are computed internally by SWMM. It is also
possible to define a dependency between concentrations of two pollutants using the
co-pollutant and co-fraction options (e.g., lead can be a constant fraction of the
suspended solids concentration).
Land Uses
Land uses characterize the activities (e.g. residential, commercial, industrial, etc.)
within a subcatchment that affect pollutant generation differently. They are used to
represent the spatial variation in pollutant buildup/washoff rates as well as the effect
of street cleaning (if used) within a subcatchment. A subcatchment can be divided
into one or more land uses. This division is done independently of that used for
pervious and impervious sub-areas, and all land uses in the subcatchment are assumed
to contain the same split of pervious and impervious area. The percentages of named
land uses assigned to a subcatchment do not necessarily have to add up to 100. Any
remaining area not assigned a land use is assumed to not contribute to the pollutant
load.
Buildup
The buildup function for a given land use specifies the rate at which a pollutant is
added onto the land surface during dry weather periods which will become available
for washoff during a runoff event. Total buildup within a subcatchment is expressed
as either mass per unit of area (e.g., Ib/acre) or as mass per unit of curb length (e.g.,
Ib/mile). Separate buildup rates can be defined for each pollutant and land use. Three
options are provided in SWMM to simulate buildup: the power function, the
87
-------
exponential function and the saturation function. The mathematical representation of
each function is described in the SWMM 5 Users Manual (Rossman, 2008). These
formulations can be adapted, by using the proper parameters, to achieve various kinds
of buildup behavior, such as a linear rate buildup or a declining rate buildup.
Defining an initial pollutant loading over the subcatchment is an alternative to using a
buildup function for single event simulations. Initial loading is the amount of a
pollutant over the subcatchment, in units of mass per unit area, at the beginning of a
simulation. This alternative is more easily adapted to single-event simulations and
overrides any initial buildup computed during the antecedent dry days.
Washoff
Washoff is the process of erosion, mobilization, and/or dissolution of pollutants from
a subcatchment surface during wet-weather events. Three choices are available in
SWMM to represent the washoff process for each pollutant and land use: event mean
concentrations (EMCs), rating curves and exponential functions (see the SWMM 5
Users Manual for mathematical representations). The main differences between these
three functions are summarized below.
EMC assumes each pollutant has a constant runoff concentration throughout
the simulation.
Rating curves produce washoff loads that are functions of the runoff rate only,
which means that they simulate the same washoff under the same discharge,
regardless of the time in the storm that the discharge occurs.
Exponential curves differ from rating curves in that the washoff load is a
function not only of the runoff rate but also the amount of pollutant remaining
on the watershed.
Buildup functions are not required when EMCs or rating curves are used to
represent the pollutant concentrations. If buildup functions are used,
regardless of washoff function, buildup is continuously depleted as washoff
proceeds, and washoff ceases when there is no more buildup remaining.
Because rating curves do not use the amount of buildup remaining as a
limiting factor, they tend to produce higher pollutant loads at the end of a
storm event than do exponential curves which do take into account the
amount of buildup remaining on the surface. This difference can be
particularly important for large storm events where much of the buildup may
be washed off in early stages.
After pollutants are washed off the subcatchment surface, they enter the conveyance
system and are transported through the conduits as determined by the flow routing
results. Here they may experience first-order decay or be subjected to reduction at
specific nodes where treatment functions have been defined.
Pollutant Reduction from Land Surfaces
Two procedures for reducing surface pollutant loads within subcatchments are
available in SWMM. They are:
88
-------
BMP Treatment: This mechanism assumes that some type of BMP has been
utilized in the subcatchment that reduces its normal washoff load by a
constant removal fraction. BMP treatment will not be used in this example but
will instead be illustrated in Example 6.
Street sweeping: Street sweeping can be defined for each land use and is
simulated in parallel with buildup prior to the beginning of the first storm
event and in-between the next events. Street sweeping is defined by four
parameters used to compute the pollutant load remaining on the surface at the
start of a storm: (1) days between street sweeping, (2) fraction of the buildup
that is available for removal by sweeping, (3) number of days since last
sweeping at the start of the simulation and (4) street sweeping removal
efficiency (in percent). These parameters are defined for each land use while
the fourth one is defined for each pollutant as well.
5.3 Model Setup
Total Suspended Solids (TSS) will be the lone water quality constituent
considered in this example. TSS is one of the most common pollutants in urban
stormwater and its concentration is typically high. The U.S. EPA (1983) reported TSS
EMCs in the range of 180 - 548 mg/L while the UDFCD (2001) reports values between
225 mg/L and 400 mg/L depending on the land use. Some of the receiving water impacts
associated with this pollutant are habitat change, stream turbidity, and loss of recreation
and aesthetics. The solids associated with TSS can also contain toxic compounds, such as
heavy metals and adsorbed organics. The following paragraphs discuss how to modify
the model built in Example 2 (file Example2-Post.inp) to consider the buildup, washoff,
and transport of TSS within the post-development site.
Define the Pollutant
The first step is to define TSS as a new pollutant under the Quality category in
SWMM's Data Browser. Its concentration units will be mg/L, and a small amount (10
mg/L) is assumed to be present in rainwater. Concentrations in groundwater as well as a
first order decay are not considered in this example, nor will any co-pollutant be defined
for TSS.
Define Land Uses
Three different land uses will be considered in this example: Residential 1,
Residential 2 and Commercial. The Residential' 1 land use will be used in residential
areas with low and medium densities (lot types "L", "M" and "M2") while the
Residential 2 land use will be used with high density apartments and duplexes (lot types
"DL" and "S"). The Commercial land use will be used with lot types "T" and "RT".
89
-------
Land uses are defined in SWMM under the Quality category in the Data Browser.
Street sweeping is not considered in this example so sweeping parameters are not
defined. A mixture of land uses will be assigned to each subcatchment area. This is done
by opening the Property Editor for a given subcatchment, selecting the Land Use
property and clicking the ellipsis button. A Land Use Assignment dialog will appear
where one enters the percentage of surface area that is assigned to each land use.
Percentages are estimated visually from the study area map. Table 5-4, shown later in this
example, summarizes the assignment of land uses in each of the subcatchments.
Defining Pollutants and Land Uses
Pollutants. Pollutants are defined in the Pollutant Editor
of SWMM under the Quality category of the Data
Browser. The minimum amount of data needed to define
a new pollutant is a name and concentration units. Other
characteristics include the pollutant's concentration in
various external (non-buildup) sources (rainwater,
groundwater, and RDII), its first-order decay coefficient
(day"1) and name of a co-pollutant that its buildup is
dependant upon.
Property
Name
Units
Rain Concert.
GW Concen.
l&l Concen.
Decay Coeff.
Snow Only
Co-Pollutant
Co-Fraction
JTSS
MG/L
10
0.0
00
0.0
NO
User-assigned name of the pollutant.
Land Uses: Different land uses will generate pollutants at different rates. Land uses
are defined in SWMM under the Quality category of the Data Browser. Their
properties are edited using the Land Use Editor which is divided into three
categories: General, Buildup and Washoff. The General tab contains the land use
name and details on street sweeping for that particular land use. The Buildup tab is
used to select a buildup function, and its parameters, for each pollutant generated by
the land use. The choice of normalizer variable (total curb length or area) is also
defined here. Finally, the Washoff tab is used to define the washoff function and its
parameters, for each pollutant generated by the land use, as well as removal
efficiencies for street cleaning and BMPs.
Land Use Name
Description
User assigned n
Max. Buildup
Rate Constant
Power^S at. Constant
Ncrmalizer
Buildup function: ROW
SAT = saturation.
0.11
0.5
0.0
CURB
= power, EXP = exponential.
General Buildup Washolt
Pollutant
Property
Function
Coefficient
Exponent
Cleaning Eftic.
BMP Effic.
Washolf (unction: EXP
curve, EMC = event m
|TSS jฃ[
Value
!EMC
160
0.0
DO
0.0
= exponential RC = rating
an concentration.
90
-------
Specify a Buildup Function
One of SWMM's buildup equations will be selected to characterize the
accumulation of TSS during dry weather periods. Unfortunately, the choice of the best
functional form is never obvious, even if data are available. Even though most buildup
data in the literature imply a linear buildup with time, it has been observed that this linear
assumption is not always true (Sartor and Boyd, 1972), and that the buildup rate tends to
decrease with time. Thus, this example will use an exponential curve with parameters Ci
(maximum buildup possible) and 2 (buildup rate constant) to represent the buildup rate
B as a function of time t:
B = Cl(\-e~c*) (5-1)
Buildup data for TSS reveal that commercial and residential areas tend to generate
similar amounts of the dust and dirt that comprise the TSS (again, there is a large
variation for different cases). Similarly, high-density residential areas tend to produce
more of this pollutant than low-density residential areas. Typical values of dust-and-dirt
buildup rates based on a nationwide study by Manning et al. (1977) are shown in Table 5-
2.
_Land_Use _ MฃELlJ!?f!^^
Commercial 116 3-365
Multiple family residential 113 8-770
Single family residential 62 3 - 950
Table 5-3 shows the parameters Ci and 2 used in equation 5-1 for each land use
defined earlier. A graphical representation of the exponential buildup model with these
parameters is shown in Figure 5-2. In SWMM the buildup function and its parameters are
defined for each land use on the "Buildup" page of the Land Use Editor. The Buildup
Function used here is Exp, the constant Ci is entered in the field Max. Buildup and
constant 2 is entered in the field Rate Constant. The field Power/Sat. Constant is not
defined when the Exponential model is used.
The values of the parameters used in this TSS buildup function were obtained
from the literature. No other justification supports their use and it is strongly
recommended that modelers define them based on site data specific to their project.
Buildup in all the subcatchments will be normalized in this example by the curb
length (typically there is more literature data per unit length of street/gutter than per unit
area). This choice is specified for each land use in the Land Use Editor. Curb lengths can
be estimated by using SWMM's Ruler tool to trace over the streets within the study area
map (see the sidebar "Measuring Tools Available in SWMM" in Example 1). They should
be similar to those listed in Table 5-4. These values are assigned to each subcatchment by
using the Property Editor. The curb length units (e.g. feet or meters) must be consistent
with those used for the buildup rate (e.g. Ibs/curb-ft or kg/curb-m) in the Land Use
Editor; do not mix units between the two systems.
91
-------
Table 5-3. Parameters for TSS buildup
Land Use
Ct (Ib/curb-ft) C2 (I/day)
Residential_l
Residential_2
Commercial
0.11
0.13
0.15
0.5
0.5
0.2
0.14
0.12
(9 o.os
-3
* O.Qi
I
TJ 0.04
0.02
-------
EMCs
An estimation of the EMCs can be obtained from the Nationwide Urban Runoff
Program (NURP) conducted by EPA (U.S. EPA, 1983). According to this study, the
median TSS EMC observed in urban sites is 100 mg/L. Based on the general observation
that residential and commercial areas produce similar pollutant loads, and taking into
account the differences among land uses, this example uses the EMCs shown in Table 5-
5 at the end of this section. These EMCs are entered into the model using the Land Use
Editor's Washoff page for each defined land use. The entry for the Function field is
EMC, the concentration from Table 5-5 is entered in the Coefficient field and the
remaining fields can be set to 0. The resulting SWMM input file is saved as ExampleS-
EMC.inp.
Exponential Washoff
The exponential washoff function used in SWMM is:
W = Cl-qC2-B (5-2)
where:
W = rate of pollutant load washed off at time t in Ibs/hr
/-I I
Ci = washoff coefficient in units of (in/hr)" 2(hr)"
C2 = washoff exponent
q = runoff rate per unit area at time t, in/hr
B = pollutant buildup remaining on the surface at time t, Ibs.
According to sediment transport theory, values of the exponent C2 should range
between 1.1 and 2.6, with most values near 2 (Vanoni, 1975). One can assume that
commercial and high-density residential areas (land uses Commercial and Residential_2\
because of their higher imperviousness, tend to release pollutants faster than areas with
individual lots (Residential 1). Thus a value of 2.2 is used for 2 in the Residential 2 and
Commercial land uses and 1.8 is used for the Residential 1 land use.
Values of the washoff coefficient (C/) are much more difficult to infer because
they can vary in nature by 3 or 4 orders of magnitude. This variation may be less extreme
in urban areas, but is still significant. Monitoring data should be used to help estimate a
value for this constant. The current example assumes a Ci equal to 40 for Residential 2
and Commercial and a Ci equal to 20 for the land use Residential 1.
Table 5-5 summarizes the Ci and 2 coefficients used for each land use under
exponential washoff. These are entered into the model using the Land Use Editor
Washoff page for each defined land use. The entry for the Function field is EXP, the Ci
value from Table 5.5 is entered in the Coefficient field, and the C2 value from the table is
entered into the Exponent field. The remaining fields can be set to 0. The resulting
SWMM input file is saved as ExampleS-EXP.inp.
93
-------
iTable 5-5. Washoff ch^actetiฃฃifฃฃฃachjand_use
Land Use EMC (mg/L) Ct [(in/hr)C2 sec *]
Residential 160 20 1.8
Residential_2 200 40 2.2
Commercial 180 40 2.2
5.4 Model Results
Both the EMC washoff model (ExampleS-EMC.inp) and the exponential
washoff model (ExampleS-EXP.inp) were run for the 0.1 in., 0.23 in. and 2-yr rainfall
events under the following set of analysis options:
Simulation Period: 12 hours
Antecedent Dry Days: 5
Routing Me thod: Dynami c Wave
Routing Time Step: 15 seconds
Wet-weather Time Step: 1 minute
Dry-weather Time Step: 1 hour
Reporting Time Step: 1 minute
A discussion of the results obtained from each model is next presented.
EMC Washoff Results
Figures 5-3 and 5-4 show the runoff concentrations simulated at different
subcatchments both for the 0.1 in. (Figure 5-3) and the 0.23 in. (Figure 5-4) storms. The
concentrations are constant and correspond to the summation of the constant
concentration in the rain (10 mg/L) and the EMCs assigned to the land uses within each
subcatchment. Once the surface runoff ceases, the TSS concentration goes to zero. That is
why no concentration is displayed for subcatchment 57, since it generates no runoff (all
rainfall is infiltrated). Note that with EMC washoff, the size of the storm has no effect on
a subcatchment's runoff concentration.
Figure 5-5 shows the TSS concentration over time (pollutograph) simulated at the
study area outlet for each of the three storm events (0.1, 0.23, and 1.0 in.). The outlet
concentration reflects the combined effect of the TSS washoff produced from each
subcatchment and routing through the conveyance network. The peak-concentrations and
shapes for the pollutographs are very similar. Compared with the washoff concentrations
generated by the individual subcatchments (Figures 5-3 and 5-4), the outlet
concentrations are not constant but attenuate over time. This attenuation is caused
primarily by the longer time it takes runoff from the lower EMC subcatchments (such as
S3 and S4) to reach the outlet. Some of it is also a result of the numerical dispersion in the
model resulting from the assumption of complete mixing within each conveyance conduit
during the pollutant routing process.
94
-------
How to Read the SWMM Status Report in Terms of Water Quality
The simulation of water quality generates added information in SWMM's Status
Report. This information can be broken into four general sections (A, B, C and D). These
additions to the status report are discussed below using ExampleS-EMC.inp with the 0.1
inch precipitation event as an example.
Part A shows the runoff-quality
continuity balance over the entire study
area. The "input" loads include (a) Initial
Buildup before the start of the simulation,
(b) Surface Buildup during all dry
weather periods, and (c) Wet Deposition
(from pollutant in the rainfall). The
"output" loads include (1) Sweeping
Removal (not simulated), (2) Infiltration
Loss for any direct rainfall or runon from
other subcatchments (simulated
automatically), (3) removal associated
with BMP Removal (not simulated in this
example), and (4) pollutant load in the
Surface Runoff (which includes the
portion of buildup that is washed off as
well as any loads produced by direct
deposition and runon). Finally, the
continuity report indicates the Remaining
Buildup.
Part B shows the quality-routing
continuity balance. In this example, only
runoff loads are routed through the
conveyance system. No dry weather,
groundwater, RDII, or user-supplied
external inflows, nor is any treatment or
decay considered. Therefore, the only
three variables represented in this
summary are the Wet Weather Inflow, the
External Outflow and the Final Stored
Mass. Note that the Wet Weather Inflow
in Part B is equal to the Surface Runoff in
Part A.
Part C provides a summary of the load of
pollutant washed off from each
subcatchment. Subcatchments S7 and S3
EPA STORM WATER MANAGEMENT MODEL - VERSION 5.0 (Build 5.0.014- beta]
Runoff Qua!i ty conti nui ty
mi ti al Bui 1 dup
surface Bui 1dup
wet Deposition
sweepi ng Removal
Infil trati on Loss
BMP Removal
surface Runoff
Remaining Buildup
Continuity Error (
Quality Routing continuity
Dry weather Inflow
wet weather Inflow
Groundwater Inflow
Inflow
-nal Infl ow
rnal Floodi ng
rnal outf1ow
Reacted
i al stored Mass
1 stored Mass
i nui ty Error (%)
B
Subcatchment Washoff Summary
Subcatchment
TSS
ibs
. 327
.213
. 099
. 339
. 494
.333
. 000
System
Outfall Loading Summary
D
outfall Node
Flow
Freg.
Pent.
Avg.
Fl ow
CFS
Max. Total
Fl ow vol ume
CFS Mgal
System
8 Li. 63
80.S3
0. 09
0.09
0.70 0.024
0.70 0.024
33.307
33.307
generate the lowest loads of TSS. S7 does not
generate any load because it does not produce
any runoff while S3 produces a smaller load
due to a large amount of pervious surface in the
subcatchment.
Part D shows the total loads leaving the
system through its outfalls.
95
-------
Z4U
210
180
150
60
30
n
i
i
i
i
i
i
i
S3
S7
,
S2
S4
i
,
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00
time (hr:min)
Figure 5-3. TSS concentrations for the 0.1 in. storm with EMC washoff
240
210
180
90
0
S3
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00
time (hnmin)
Figure 5-4. TSS concentrations for the 0.23 in. storm with EMC washoff
96
-------
|
220
200
180
160
140
120
100
80
60
40
20
0
0.1 in. storm
0.23 in. storm
2-yr storm
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00
time (hr:min)
Figure 5-5. TSS concentration at site outlet with EMC washoff
Figure 5-5 also shows that TSS concentrations continue to appear at the outlet for
an extended period of time after the end of the storm event. This is an artifact of the flow
routing procedure wherein the conduits continue carry a very small volume of water
whose concentration still reflects the high EMC levels. Thus although the concentrations
appear high, the mass loads carried by these small discharges are negligible. This is
evident when the outlet hydrograph is plotted alongside the outlet loadograph for a given
storm. A loadograph is a plot of concentration times flow rate versus time. An example
for the 0.1 in. event is shown in Figure 5-6. This plot was generated by exporting the time
series table for Total Inflow and TSS concentration at the outfall node Ol into a
spreadsheet, using the spreadsheet to multiply flow and concentration together (and
converting the result to Ibs/hr), and then plotting both flow and load versus time. Note
how the TSS load discharged from the catchment declines in the same manner as does the
total runoff discharge.
97
-------
mass rate
discharge
0.8
0.7
0.6
0.5
0.4
0.1
0
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:0011:0012:00
time (hr:min)
Figure 5-6. Runoff flow and TSS load at site outlet for the 0.1 in. storm with EMC washoff
Exponential Washoff Results
Figure 5-7 shows the simulated TSS concentration in the runoff from different
subcatchments using the 0.1 in. storm and the Exponential washoff equation. Unlike the
EMC results, these concentrations vary throughout the runoff event and depend on both
the runoff rate and the pollutant mass remaining on the subcatchment surface. Figure 5-8
shows the same plots but for the 0.23 in storm. Note two significant differences with
respect to the results obtained for the 0.1 in storm. The maximum TSS concentrations are
much larger (around 10 times) and the generation of TSS is much faster, as seen by the
sharper-peaked pollutographs in Figure 5-8. Finally, Figure 5-9 shows the same graphs
for the larger 1-in., 2-yr storm. The TSS concentrations are slightly larger than those for
the 0.23 in. storm but the difference is much smaller than the difference between the 0.1
in and 0.23 in storms. Similar results hold for the pollutographs generated for the
watershed's outlet as seen in Figure 5-10.
98
-------
300
250
200
150
two
100
50
0
0:00
1:00
3:00
2:00
time (hr:min)
Figure 5-7. TSS concentrations for the 0.1 in. storm with Exponential washoff
4:00
0:00
1:00
3:00
2:00
time (hr:min)
Figure 5-8. TSS concentrations for the 0.23 storm with Exponential washoff
4:00
99
-------
wo
2000
1800
1600
1400
1200
1000
800
600
400
200
0
SI
S3
S7
0:00 1:00 2:00 3:00
time (hr:min)
Figure 5-9. TSS concentrations for the 2-yr (1 in.) storm with Exponential washoff
4:00
2000
1800
1600
1400
j 1200
1000
800
600
400
200
0
M
0.1 in. storm
0.23 in. storm
2-yr storm
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00
time (hr:min)
Figure 5-10. TSS concentration at the site outlet for Exponential washoff
100
-------
Even though the EMC and Exponential washoff models utilize different
coefficients that are not directly comparable, it is interesting to compute what the average
event concentration in the runoff from each subcatchment was under the two models. The
resulting averages are shown in Table 5-6 for the case of the 0.23 in. storm. The point
being made here is that even though the pollutographs produced by the two models can
look very different, with the proper choice of coefficients it is possible to get event
average concentrations that look similar. Although the results of the Exponential model
are more pleasing to one's sense of how pollutants are washed off the watershed, in the
absence of field measurements one cannot claim that they are necessarily more accurate.
Most SWMM modelers tend to use the EMC method unless data are available to estimate
and calibrate the coefficients required of a more sophisticated buildup and washoff
model.
Subcatchment
SI
S2
S3
S4
S7
EMC Model
^^(SS^)^^
170
199.2
117.2
131.2
0
Exponential Model
^^sis/y^^
180.4
163.6
67.7
91.4
0
5.5 Summary
This example illustrated how SWMM is used to model the quality of stormwater
runoff within an urban catchment without any source or regional BMP controls. One
pollutant, TSS, was simulated with one buildup method (exponential) and two different
washoff methods (EMC and exponential). The key points illustrated in this example
were:
1. SWMM models runoff water quality through the definition of pollutants, land uses,
pollutant buildup, and pollutant washoff. Any number of user-defined pollutants and
land uses can be modeled. Pollutant buildup and washoff parameters are defined for
each land use and more than one land use can be assigned to each subcatchment.
2. There are several options available to simulate both pollutant buildup and washoff.
Buildup expressions are defined by a buildup rate and a maximum buildup possible
per unit of area or curb length. Pollutant washoff can be defined through an event
mean concentration (EMC), a rating curve, or an exponential function. The
exponential method is the only one that directly depends on the amount of buildup
remaining on the surface. Rating-curve calculations are dependant only on the runoff
across the subcatchment, while EMCs have constant concentrations throughout the
simulation.
101
-------
3. Exponential washoff produces a runoff pollutograph with rising and falling limbs,
similar to that of a runoff hydrograph. The EMC pollutograph is flat throughout the
duration of the event.
4. Small storms can have a high impact on receiving waters because they are more
frequent and can still generate significant washoff concentrations.
There are many uncertainties associated with both the process representation and
the data required to properly estimate, calibrate and validate a runoff water quality model.
It is strongly recommended that modelers use site specific data whenever possible when
building a runoff water quality model with SWMM.
102
-------
Example 6. Runoff Treatment
This example illustrates how to model water quality treatment in the BMPs that
were used in two earlier examples to control runoff from a new residential development
on a 29 acre site. Treatment of total suspended solids (TSS) is applied at both the
detention pond introduced in Example 3 and the filter strips and infiltration trenches
added in Example 4. The pond from Example 3 is re-designed to a smaller volume for
this example to account for the runoff reduction associated with the upstream infiltration
trenches. TSS removal in the detention pond is modeled as an exponential function of
time and water depth. TSS removal in the filter strips and infiltration trenches is a fixed
percent reduction in loading.
6.1 Problem Statement
In Example 3 a regional detention pond was designed for a 29 acre residential site
to detain a water quality capture volume (WQCV) for a specific period of time and to
reduce peak runoff flows to their pre-development levels. Example 4 added two different
types of distributed Low Impact Development (LID) source controls throughout the site
to help reduce the volume of runoff generated. Example 5 illustrated how to model total
suspended solids buildup and washoff within the site without considering any TSS
removal that might occur within the LIDs or the pond. These previous models will be
extended to explicitly account for the removal of TSS that occurs in both the LIDs and
the pond. A comparison will be made of the TSS concentrations and loads in the runoff
produced by the site both with and without considering treatment.
Figure 6-1 shows the example study area with the LIDs and detention pond
included. The SWMM storage unit that represents the detention pond as designed in
Example 3 was renamed to SU2 and resized for this example to incorporate a smaller
Water Quality Capture Volume (WQCV) due to the runoff reduction produced by the
LIDs. It was concluded from Example 4 that of the two types of LIDs modeled,
infiltration trenches have the greatest impact on reducing the volume of water that needs
to be treated in the pond's WQCV. The total volume of water that can be captured by the
infiltration trenches was determined to be 6,638 ft3. Because the WQCV is a requirement
for the watershed as a whole, the volume captured in the infiltration trenches can be
subtracted from the volume required in the regional pond, reducing it from the 24,162 ft3
required in Example 3 to 17,524 ft3 (24,162 - 6,638).
103
-------
Figure 6-1. Developed site with LIDs and detention pond
The storage unit's shape and outlet structures were redesigned to control this new
WQCV and meet the general design criteria introduced in Example 3 (e.g. 40 hr
drawdown time for the WQCV and peak shaving of the 2-, 10- and 100-yr storms). Table
6-1 compares the storage curve of the pond designed in Example 3 without LIDs and the
pond designed for this example with LIDs. Table 6-2 does the same for the dimensions
and inverts of the orifices and weirs that comprise the pond's outlet structure. Figure 6-2
shows the general placement of the various outlets.
Table 6-1. Storage curve for the re-designed pond
This Example
Depth (ft) 0 2.2
Area (ft2) 10368 14512
2.3 6
32000 50000
Example 3
Depth (ft)
Area (ft2)
0
14706
2.22
19659
2.3
39317
6
52644
104
-------
Table 6-2. Properties of the pond's re-designed outlet structure
Type of Event
Element Controlled
, Height Width, Invert Discharge Orifice or Weir
pe h(ft) b(ft) Offset z (ft) Coefficient Area(ft2)
Example 3
Orl Orifice WQCV
Or2 Orifice 2-yr
Or3 Orifice 10-yr
Wl Weir 100-yr
^lde , 0.3 0.25 0 0.65 0.08
Rectangular
^lde , 0.5 2 1.5 0.65 1
Rectangular
^lde , 0.25 0.35 2.22 0.65 0.09
Rectangular
Rectangular 2.83 1.75 3.17 3.3 4.95
This Example
Orl Orifice WQCV
^ ~ ^ -f 1- and 10-
Or2 Onfice
yr
Wl Weir 100-yr
6ft
Z3
Z2
,,
^ldet , 0.16 0.25 0 0.65 0.04
Rectangular
^lde , 0.5 2.25 1.5 0.65 1.13
Rectangular
Rectangular 2.72 1.6 3.28 3.3 4.35
. wi
b3
I',
K. ~~
Or 2
-Orl
Figure 6-2. Schematic of the re-designed pond outlet structure
6.2 System Representation
Water Quality Treatment in LIDs
As defined in Example 4, the filter strip and infiltration trench LIDs are modeled
as subcatchments in order to represent the combined effects of infiltration and storage on
storm runoff. This example will also consider their ability to reduce pollutant loads in the
surface runoff they handle. There are no widely accepted mechanistic models of pollutant
removal through these types of LIDs. The best one can do is to apply average removal
efficiencies for specific contaminants based on field observations reported in the
literature.
105
-------
SWMM can apply a constant BMP Removal Efficiency for any pollutant in the
washoff generated from a particular land use. At each time step the pollutant load
generated by a given land use is reduced by this user-supplied value. This reduction also
applies to any upstream runoff that runs onto the subcatchment. The LIDs added in
Example 4 receive runoff from upstream subcatchments and do not generate any
pollutant load themselves. Thus it will be convenient to define a new land use, named
"LID", that is used exclusively for the LID subcatchments and has a specific BMP
Removal Efficiency for TSS associated with it.
Water Quality Treatment in Detention Ponds
Detention ponds are modeled as storage unit nodes within SWMM. By adding
Treatment Functions to the storage node's properties, SWMM can reduce the pollutant
concentrations in the pond's outflow. This example uses an empirical exponential decay
function to model solids removal through gravity settling within a pond. For controlling
the WQCV event, the pond fills relatively quickly over a period of 2 hours and then
drains slowly over an extended period of 40 hours during which solids removal occurs.
At some interval At during this drain time, and assuming homogenous concentration, the
fraction of particles with a settling velocity ut that are removed would be utAt/d where d
is the water depth. Summing over all particle settling velocities leads to the following
expression for the change in TSS concentration AC during a time step At:
(6-1)
where Ct is the total concentration of TSS particles at time t and/ is the fraction of
particles with settling velocity ut. Because ^_ifiui is generally not known, it can be
replaced with a fitting parameter k and in the limit Equation (6-1) becomes:
(6-2,
\ /
,
ซ. / I
dt d
Note that k has units of velocity (length/time) and can be thought of as a representative
settling velocity for the particles that make up the total suspended solids in solution.
Integrating Equation (6-2) between times t and t+At, and assuming there is some
residual amount of suspended solids C* that is non-settleable leads to the following
treatment function for TSS in the pond:
C,+A, = C * +(C, - C*)e-(kid]^ (6-3)
Equation (6-3) is applied at each time step of the simulation to update the pond's TSS
concentration based on the current concentration and water depth.
106
-------
Simulating Treatment within a Conveyance Network
SWMM can apply water quality treatment at any node of a drainage
system's conveyance network. Treatment for a node is defined by opening its
Property Editor and clicking the ellipsis button next to the Treatment property. This
brings up a Treatment Expression dialog box in which the user can define a
treatment function for each pollutant that passes through the node.
The treatment function for a given pollutant can have one of the following forms:
R = f(P,R_P,V)
C = f(P,R_P,V)
where ,R is the fractional removal, C is the outlet concentration, P is one or more
concentrations given by the pollutant names (e.g., TSS), R P is one or more
pollutant removals (e.g., R TSS), and V is one or more of the following process
variables: FLOW (Row rate into the node), DEPTH (water depth above node invert),
HRT (hydraulic residence time), DT (routing time step) and AREA (node surface
area). Some examples of treatment expressions are:
C = BOD * exp(-0.05*HRT)
R = 1 - (1 + (0.001/(2*FLOW/AREA))^(-2)
With a fractional removal expression, the new concentration at the node, C,
is defined as Cin(l-R) where Cin is the inflow concentration to the node. Also, when
a concentration P appears in an expression applied to a non-storage node, it is the
same as Cin for the node whereas for a storage node it is the current concentration C
in the storage unit.
107
-------
6.3 Model Setup
The starting point for adding treatment to the runoff controls placed on the study
site is the file Example6-Initial.inp. This input file already contains the local LIDs
defined in Example 4 and the redesigned storage unit and outlet structures described in
section 6.1. The land uses in this file were re-calculated and re-assigned to each
subcatchment based on the discretization employed in Example 4. The washoff function
for TSS is the EMC washoff function used in Example 5. The curb lengths and land uses
assigned to each subcatchment are listed in Table 6-3. The subcatchments treated by
infiltration trenches or filter strips are marked in grey in this table.
Table 6-3.
Curb lengths and land uses for LID subcatchments
Subcatchment Area (ac)
Sl.l
S1.2
SI. 3
S2.1
S2.2
S2.3
S3.1
S3.2
S3. 3
S4.1
S4.2
S4.3
S4.4
S5
S6
S7
1.21
1.46
1.88
1.3
1.5
.88
.29
.02
.38
.65
0.79
1.91
2.4
4.79
1.98
2.33
Curb Length (ft)
450
600
630
450
600
630
0
430
500
0
400
1150
700
2480
1100
565
Residential_l (%)
100
100
100
100
0
0
0
100
0
0
0
36
0
0
0
0
Residential_2 (ฐ/
0
0
0
0
100
100
0
0
85
0
100
64
0
0
0
0
'o) Commercial (%)
0
0
0
0
0
0
0
0
0
0
0
0
71
98
100
0
LID Treatment
It is assumed that each filter strip and infiltration trench can provide 70% TSS
removal for the runoff that passes over it. This is a typical removal observed for
infiltration-based LIDs (Sansalone and Hird, 2003). A new land use, named "Z/Z>" is
created with no TSS buildup function, an EMC TSS washoff function with 0 mg/L of
TSS, and a TSS "BMP efficiency" of 70%. The land use assignment for each of the LID
subcatchments (S_FS_J through S_FS_4 and S IT 1 through S IT 4) is set to 100% LID.
As a result, all of the runoff generated from upstream subcatchments that flow over these
LID subcatchments will receive 70% TSS removal.
108
-------
Detention Pond Treatment
The removal of TSS in the detention pond is simulated using the exponential
model given by Equation 6-3. One can roughly estimate what the removal constant k in
this expression must be so that a targeted level of pollutant removal is achieved within a
40 hour detention time for the 0.23 in. WQCV design storm. If Equation 6-3 were applied
over a 40 hour period to achieve a target TSS reduction of 95%, then an estimate of k
would be:
k = -dxln(0.05)/40 (6-4)
where d is some representative value of the pond depth during the 40 hour release period.
As can be verified later on, the average depth in the pond for the 0.23 in design storm
over a 40 hour duration is 0.15 ft. Using this value in the expression for k yields an
estimate of 0.01 ft/hr. This value is of the same order as the 0.03 ft/hr figure quoted in US
EPA (1986) that represents the 20-th percentile of settling velocity distributions measured
from 50 different runoff samples from seven urban sites in EPA's Nationwide Urban
Runoff Program (NURP).
With this value of k and assuming a minimum residual TSS concentration C* of
20 mg/L, the following expression is entered into SWMM's Treatment Editor for the
storage unit SU2: C = 20 + (TSS - 20) * EXP(-0.01 / 3600 /DEPTH * DT)
Note the meaning of the individual terms in this expression with respect to those
in Equation (6-3): 20 is the value assumed for C*, TSS is the identifier given to the TSS
concentration C for this model, 0.01/3600 is the value of k expressed in units of ft/s,
DEPTH is the reserved word that SWMM uses for the water depth d in feet, and DT is
the reserved word that SWMM uses for the routing time step At in seconds. When
SWMM sees reserved words like DEPTH and DT within a treatment expression it knows
to automatically insert their current values into the expression at each time step.
The following analysis options should be used for all of the simulations made
with this treatment-augmented input file:
Flow Routing Method: Dynamic Wave
Wet Weather Time Step: 1 minute
Flow Routing Time Step: 15 sec
Reporting Time Step: 1 minute
Total Duration: 2 days (48 hr)
The 48 hour duration was chosen so that the full effect of the drawdown of the
WQCV in the detention pond could be observed. Finally, it is suggested to increase the
number of significant figures (from 2, the default value, to 4) for the subcatchment
parameter "Runoff and the node parameters "Quality" and "Total Inflow" by selecting
Tools | Program Preferences | Number Formats from SWMM's main menu bar. This
will help when tabulated results are copied from SWMM to a spreadsheet program to
compare results between different runs in cases where variations are small and would not
be visible if only two significant figures were used. The resulting input file is named
Example6-Final.inp.
109
-------
6.4 Model Results
Results for several sets of comparison runs will be discussed. First, the effect of
TSS treatment at the LIDs is considered. Figure 6-3 compares the TSS concentration in
the treated runoff from filter strip S_FS_1 to that of the upstream runoff from
subcatchment S3.2 for the 0.1 in. storm. Also shown for reference are the runoff flows
from each area. The reduction from 170 mg/L to 51 mg/L through the filter strip matches
the 70 % TSS removal efficiency specified for the LID. Similar results are obtained for
the remaining filter strips under all design storms.
160
100
80
/:n
jn
i iiu *,...ป Tee
/
7
0:00 1
T~
^CI^;
L:00 2:00
s
^^ FllterS trip TS S
Ups tream Runoff
3:00 4:00 5:00 6:
Time (hr:min)
0.08
0.06
0.04
0.02
0
90
Figure 6-3. TSS and runoff reduction through filter strip S_FS_1 for the 0.1 in. storm
Figure 6-4 presents a similar comparison for the infiltration trench S IT 4 that
treats runoff from subcatchment S2.3. These results are for the 2-yr (1 in.) event since for
the smaller storms all rainfall infiltrates through the trenches. Once again note how the
70% removal causes a drop in TSS concentration from 210 mg/L down to 63 mg/L. As
was the case with the filter strips, the remaining infiltration trenches show a similar
behavior to that of S IT 4.
110
-------
Upstream TSS
Trench TSS
Upstream Runoff
Trench Runoff
3
2 is
o
i:00
1:00
2:00
3:00
lime (hr:min)
4:00
5:00
6:00
Figure 6-4. TSS and runoff reduction through infiltration trench S_IT_4 for the 2-yr storm
The next comparison is between the levels of treatment provided by the detention
pond for the various design storms. Comparing the time series of pond influent TSS
concentrations with the treated effluent concentrations is not particularly useful since the
flow rates of these two streams are so different. Instead the pond effluent concentration,
both with and without treatment will be compared for each design storm. The result is
shown in Figure 6-5. These plots were made by running the model for each design storm
(0.1 in., 0.23 in., and 2-year) both with the treatment function for the storage unit node
SU2 and without it. After each run a time series table of the TSS concentration at node
SU2 was generated and exported to a spreadsheet program from which Figure 6-5 was
generated.
The following observations can be drawn from Figure 6-5:
With the lvalue of 0.01 ft/hr, the pond behaves as designed for the WQCV storm
(0.23 in) by removing essentially all of the settleable solids over a period of 40
hours.
The larger the 2-hour design storm, the less effective is treatment in reducing TSS
levels over time because of the deeper water depths experienced in the pond.
For all size storms, it takes a considerable amount of time for any significant
removal of TSS to occur; a 50% reduction in settleable TSS requires 10, 19, and
35 hours for the 0.1 in, 0.23 in, and 2-year storms, respectively.
111
-------
k = 0.01 ft/hr
w/o Treatment
0.1 in
0.23 in
2-year
w/ Treatment
10 15 20 25 30 35 40 45 50
Figure 6-5. TSS concentrations in pond SU2 with and without treatment (k = 0.01 ft/hr)
Another way to evaluate treatment in the pond is to compare the TSS mass
loading that it releases both with and without treatment. This is done in Figure 6-6 for the
0.23 WQCV storm. Plots for the other design storms look similar to this one. The effect
of treatment on reducing the mass of TSS released is not as significant as it was for
concentration. In fact, the Status Report for the run with treatment shows that of 72.5 Ibs
of TSS washed off for this storm event, only 15.7 Ibs were removed in the pond. This
yields an overall mass removal of only 21.7 %. The percent mass removals for the other
design storms were 44.4 % for the 0.1 in. storm and 6 % for the 2-year (1 in.) storm.
These low to moderate mass removals are a consequence of the extended time required
for solids to settle in the pond during which it still is releasing an outflow.
112
-------
k = 0.01 ft/hr
Untreated Effluent
Treated Effluent
0 4 8 12 16 20 24 28 32 36 40 44 48
Figure 6-6. TSS mass load released by pond SU2 for the 0.23 in storm (k = 0.01 ft/hr)
These rather modest levels of detention pond performance were computed using a
removal constant k that represents particles with a very low settling velocity, below that
of the lowest 20% determined from a nationwide survey. Suppose the size distribution of
particles comprising the TSS washoff were larger, as reflected in a lvalue of 0.3 ft/hr.
This represents the 40-th percentile of the settling velocities found from the NURP study
(US EPA, 1986). Figure 6-7 shows the resulting TSS concentrations in the pond
discharge with this higher lvalue. Figure 6-8 does the same for the TSS discharge
loading for the 0.23 in. event. Table 6-4 summarizes the pond's treatment performance
for the two different ^-values. These results show that uncertainty in the removal constant
will significantly impact predictions of TSS removal within the detention pond.
Unfortunately, as noted in US EPA (1986), there can be high variability in solids settling
velocity distributions from site to site and from storm to storm within a given site. This
variability makes it very difficult to make reliable estimates of detention pond treatment
effectiveness.
113
-------
10 15
Time (hr)
20
25
Figure 6-7. TSS concentrations in pond SU2 with and without treatment (k = 0.3 ft/hr)
k = 0.3 ft/hr
Untre ate d Efflue nt
Treated Effluent
o
ซ
o
0 4 8 12 16 20 24 28 32 36 40 44 48
Figure 6-8. TSS mass load released by pond SU2 for the 0.23 in. storm (k = 0.3 ft/hr)
114
-------
Time to achieve 50% reduction, hr
Time to achieve full reduction, hr
Overall mass removal, %
0.1 in. Storm
A = 0.01 A = 0.3
10 1
30 7
44.4 81.8
0.23 in. Storm
A = 0.01 A = 0.3
19 3
40 10
21.7 75.0
1.0 in. Storm
A = 0.01 A = 0.3
35 6
>48 20
6.0 35.7
Finally, Figure 6-9 compares the total pounds of TSS discharged from the study
area site for each design storm with no treatment, with just LIDs, and with both LIDs and
the detention pond. These loadings can be read from the Status Reports generated by
running an analysis of each storm with (a) both types of treatment, (b) with the treatment
function for the pond removed (LID treatment only), and (c) with the input file developed
for EMC washoff in Example 5 (no treatment). The k value used in this comparison is the
0.01 ft/hr value. Note the consistent pattern of load reduction as more treatment is
applied. Also note that the pond provides a lower increment of overall load reduction
than do the LIDs even though the pond is a regional BMP that treats all of the
catchment's runoff while the LIDs are local BMPs that receive runoff from only 41 % of
the catchment's area. This is a result of the conservative k value used in the pond's
treatment expression. Using a higher value, which would reflect a larger particle size
distribution in the runoff, would result in lower loadings from the pond.
6.5 Summary
This example showed how water quality treatment could be modeled within
SWMM. Total Suspended Solids (TSS) removal was considered in both local LID source
controls as well as in a regional detention basin. The key points illustrated in this example
were:
1. LID controls can be modeled as distinct subcatchments with a single landuse that has
a constant removal efficiency assigned to it. SWMM applies this removal efficiency
to the runoff that the control receives from upstream subcatchments.
2. Treatment within a detention pond is modeled with a user-supplied Treatment
Function that expresses either the fractional removal or outlet concentration of a
pollutant as a function of inlet concentration and such operational variables as flow
rate, depth, and surface area.
3. An exponential treatment function was used to predict TSS removal within this
example's detention pond as a function of a removal constant and the pond's water
depth, where the removal constant reflects the settling velocity of the particles to be
removed.
4. SWMM's use of constant removal efficiencies for LID controls makes its LID
treatment performance insensitive to size of storm.
5. Treatment performance for detention ponds decreases with increasing size of storm
due to an increase in pond depth. It also decreases with decreasing size distribution of
the sediments that constitute the TSS in the runoff.
115
-------
6. For the treatment function used in this example, the pond provided less incremental
TSS load reduction than did the LIDs. This result, however, is completely dependent
on the value of the removal constant used within the pond's treatment function.
7. The large variability reported for particle settling velocities in urban runoff makes it
extremely difficult to estimate a removal constant for a detention pond's treatment
function that can consistently provide reliable estimates of the pond's treatment
performance.
0.1 in Storm
0.23 in Storm
j
$ 10
H
n r-
I I
No LIDs Only LIDs +
Treatment Pond
TSS Load (Ibs)
4- 90
s o o
No LIDs Only LIDs +
Treatment Pond
2-yr Storm
^ 500
X
C, 400
1 300
x 200
H 100
0
No LIDs Only LIDs +
Treatment Pond
Figure 6-9. Total TSS load discharged at site outlet under different treatment scenarios
116
-------
Example 7. Dual Drainage Systems
The post-development model in Example 2 simulated simple hydraulic routing
within a surface drainage system that employed open channels in the form of gutters and
swales. The three hydraulic routing methods (Steady Flow, Kinematic Wave and Dynamic
Wave) were introduced and their effects on the drainage system behavior shown.
Example 7 will convert some of the open channels in Example 2 to parallel pipe and
gutter systems. A series of storm sewer pipes placed below the existing swales will also
be added to help drain the downstream section of the site's park area. Both the 2-yr and
100-yr design storms will be used to size and analyze the performance of this expanded
dual drainage system. Particular attention will be paid to the interaction between the
below-ground storm sewer flows and the above-ground street flows that occurs during
high rainfall events.
7.1 Problem Statement
The objective of this example is to simulate the interaction between the minor and
major drainage systems through the interconnection of part of their underground and
surface sections. For frequent events the minor or "initial" system operates (Grigg, 1996);
overland flows are conveyed by gutters and enter into the pipe system. For large events
these pipes surcharge and flood, and the major system handles the flows (Grigg, 1996). In
particular, the entire street (not only the gutters) becomes a conveyance element.
Figure 7-1 shows the post-development layout of the site analyzed in Example 2.
Example 2 modeled the drainage system with open channels and culverts whose invert
elevations were the same as the ground surface elevations found on the site contour map.
A series of below ground pipes will be added to the site that share inlets with the open
(surface) channel running through the park. Gutters will also be added to the upstream
section of the study area. The cross sections of these gutters will be that of a typical
street, representing the surface "channel" through which water would flow if the pipe
system surcharged and flooded the street. Thus, in this example the pipes and flow in
gutters represent the minor system, and the channel in the park and the flow in streets
represent the major system. The pipe system will be sized for the 2-yr event and its
behavior observed during the major storm (100-yr event).
7.2 System Representation
Example 2 introduced junction nodes and conduit links as the basic elements of a
drainage network. An example of the parallel pipe and gutter conveyance arrangement
that will be used in this example is shown in Figure 7-2. It consists of a below-grade
circular pipe connected to manhole junctions on either end, plus an above grade street
and gutter channel also connected to the same two manhole junctions. The details of the
gutter inlet and drop structures that make the actual connection with the manholes are not
117
-------
important for our purpose. The parameters needed to characterize this type of node-link
arrangement are as follows:
RainGage
4070
Figure 7-1. Post-development site with simple drainage system
Lengthj Manning roughness
coefficient; Cross section
shape and geometry
Inlet no tie
Outlet node
Inlet offset
"ฃ"
Maximum depth
Invert elevation
Outlet offset
Figure 7-2. Parallel pipe and gutter conveyance
118
-------
Manhole Invert Elevation
The invert elevation of the manhole junction is the elevation of the bottom of the
manhole relative to the model's datum (such as mean sea level). The invert elevation
establishes the junction's vertical placement in the SWMM model.
Manhole Maximum Depth
The maximum depth of the manhole junction is the distance from its invert to the
ground surface elevation where street flooding would begin to occur. If this
maximum depth is left set at zero, SWMM automatically uses the distance from the
junction's invert to the top of the highest connecting link, which would be the top of
the gutter/street channel in this particular example.
Surcharge Depth
The surcharge depth of a junction is the additional depth of water beyond the
maximum depth that is allowed before the junction floods. SWMM uses this depth to
simulate pressurized conditions at bolted manhole covers or force main connections
but it will not be used in this example. Different types of surcharge will be discussed
later in this example.
Conduit Offsets
The inlet offset for a conduit is the distance that its inlet end lies above the invert of
the junction that it connects to. A similar definition applies to the offset for the outlet
end of a conduit. In the parallel pipe and gutter system, the elevations of the gutters
are set above the pipes using their inlet and outlet offsets (Figure 7-2).
Note that because this representation of a dual drainage system has more than one
conduit exiting a junction node, Dynamic Wave flow routing must be used to analyze it
hydraulic behavior. An alternative way to represent these systems is to use the Overflow
variety of Flow Divider nodes to connect pairs of pipes and gutter channels together. This
scheme can be analyzed using the simpler Kinematic Wave method, where any flow in
excess of the sewer pipe capacity would be automatically diverted to the gutter channel.
The disadvantage of this approach is that cannot model a two-way connection between
the pipes and gutters/streets, nor can it represent the pressurized flow, reverse flow and
backwater conditions that can exist in these systems during major storm events.
Drainage System Criteria
The general drainage system criteria that will be used in this example are listed
below. These criteria are based on those defined for the city of Fort Collins (City of Fort
Collins, 1984 and 1997). Figure 7-3 shows the different elements of the street considered
in these criteria. Two storms will be used to design the drainage system: a minor or initial
storm (2-yr) and a major storm (100-yr). The initial storm is one that occurs at fairly
regular intervals while the major storm is an infrequent event. In this example, the streets
are classified as "collectors". Note that the general drainage criteria presented here apply
only to this example and will change depending on the location of the system being
designed. The criteria are:
119
-------
The minimum gutter grade (Si in Figure 7-3) shall be 0.4%, and the maximum shall
be such that the average flow velocity does not exceed 10 ft/see.
The cross-slope (Sx in Figure 7-3) of all streets will be between 2% and 4%.
The encroachment of gutter flows onto the streets for the initial storm runoff will not
exceed the specification presented in the second column from the left in Table 7-1.
The encroachment of gutter flows on the streets for the major storm runoff will not
exceed the specification presented in the third column from the left in Table 7-1.
The pipe system should carry the 2-yr storm and work as an open channel system.
Street
SL Gutter grade Sx Cross slope
Figure 7-3. Elements of streets defined in the drainage criteria
120
-------
Table 7-1. General drainage system criteria for Fort Collins (City of Fort Collins, 1984 and 1997)
Street
classification
Initial storm
Major storm
Local (includes ., ,. , , ... , ,. , ... , ,
, ,, ,T , , . , Residential dwelling and other dwelling cannot be
places, alleys No curb-topping. Flow may . Jtjtt. j i- r, , ., ^
j . , , t r-tt inundated at the ground line. The depth of water i
and marginal spread to crown of street ., . , , . ,
, ฐ the crown cannot exceed 6 inches.
access)
over
Collector
Major Arterial
No curb-topping. Flow spread
must leave at least one lane
width free of water
Residential dwelling and other dwelling cannot be
inundated at the ground line. The depth of water over
the crown cannot exceed 6 inches. The depth of water
over the gutter flowline cannot exceed 18 inches. (The
most restrictive of the last two conditions governs)
,T , . , , Residential dwelling and other dwelling cannot be
No curb-topping. Flow spread . ,,,, .1- r * *&
^ , ฐ-. i i r. inundated at the ground line. The street flow cannot
must leave at least one-half ^ . , ^ ,, ^
,- _, ..i r. ,- overtop the crown. The depth of water over the gutter
of roadway width free of ~ ,. , , . . ,. ,
J fir,T,,iin^ cannot exceed 18 inches. (The most
water in each direction
flowline cannot exceed 18 inches.
restrictive of the last two conditions governs)
7.3 Model Setup
Figure 7-4 shows the layout of the dual drainage system with the pipes, streets
and swales that will be included in the model. Note that runoff from subcatchments <57
and S2 is introduced first to the street system (gutters) through nodes Auxl and Aux2, and
then enters the storm sewers through inlet grates represented by nodes Jl and J2a. For the
rest of the system the runoff is assumed to enter directly into the pipe system. The
following steps are used to build the complete dual drainage model starting from the
layout defined in the input file Example2_Post.inp.
Surface Elements
1. The first step is to add the additional junction nodes Auxl, Aux2 and J2a, shown in
Figure 7-4. The invert elevations of these nodes for now are simply their surface
elevations (Auxl = 4975 ft, Aux2 = 4971.8 ft, and J2a = 4970.7 ft).
2. Next the additional gutters C_Auxl, C_Auxlto2, C_Aux2, and C2a are added into
the model and their lengths are determined using the Auto-Length option. These
gutters have a roughness coefficient of 0.016.
3. Instead of the trapezoidal cross section used in Example 2, irregular shapes will be
used to represent the cross sections of the gutter conduits. The transects that define
these shapes are created using SWMM's Transect Editor (see the sidebar "Defining
Channel Cross Sections as Irregular Channel Transects"}. The transect Full Street
shown in Figure 7-5 is used to represent the entire section of all streets within the
study area. A cross-slope Sx = 4% is used to define this section. The transect
Half Street depicts only half of the Full Street section, from the sidewalk to the
crown of the street. It is used to represent the north-south street that runs down the
121
-------
east side of the development because its crown is considered to be the boundary of
the catchment. The station and elevation data for both cross sections are listed in
Table 7-2.
4. The two types of transect cross-sections just created are then assigned to the
corresponding street conduits shown in Figure 7-4. Conduits C Auxl, C Auxlto2,
and C_Aux2 represent streets that have the Full Street section. C2a and C2 represent
streets that have the Half Street section.
S3
C2a
C2
Inflow from subcatchment
Figure 7-4. Three-dimensional layout of the site's dual drainage system
40ft
20ft
Figure 7-5. Full_Street cross section. The Half_Street section is half of this section.
122
-------
Defining Channel Cross Sections as Irregular Channel Transects
Channel cross sections can be assigned to each channel using the Cross
Section Editor as was shown in Example 2 or by defining a general cross section
Transect object. The main advantages of using a Transect are: 1) the shape of a
channel cross section can be entered into the model once and then assigned to
multiple channels that share the same shape, 2) changes to channels with the same
cross section (e.g. in Mannings roughness and depth) can be made more efficiently
through the Transect Editor and 3) any channel cross section shape can be entered.
Transects are generally used to model natural irregular channel shapes in SWMM
but they can also be used to model gutter cross sections as well. The process used to
add a new Transect object to a project is:
1. Open the Transect Editor by selecting Hydraulics
Data Browser.
2. Enter a name for the new Transect.
Transects I "+" in the
3. Define the shape of the Transect's cross section by entering Stations
(distance from the left edge of the Transect) and their corresponding
Elevations into the editor's data grid.
n
I
Transect GutteMvlinor
| o Ovefhank o Channel \
4. The location of the left bank, the right bank and the channel section is
defined by entering the corresponding station locations in the Bank Stations
entry fields. In this example, there is no left or right bank so the Left and
Right bank stations are set to zero.
5. The Mannings roughness values are defined for the cross section's left bank,
right bank and channel sections. No values are required for the bank
roughness if they are not modeled.
To assign a specific Transect object to a conduit's cross-section:
1. Open the Property Editor for the conduit.
2. Click the ellipsis button next to the Shape property to bring up the Cross-
Section Editor.
3. Select Irregular for the choice of cross-sectional shape and then select the
Transect of choice from the Transect Name combo box.
123
-------
Table 7-2. Cross section data for street transects
Full
Street
Half-
Street
Station
Elevation
Station
Elevation
-40
1.3
-40
1.3
-20
0.5
-20
0.5
-20
0
-20
0
0
0.8
0
0.8
20
0
0
1.3
20
0.5
-
40
1.3
-
Below-Ground Elements
The changes made so far have been to surface elements of the watershed. Now the
pipes below the streets will be added to the drainage system and the invert elevations of
the surface conduits will be re-defined.
1. Conduit Cl is replaced with a pipe (PI) and two pipes are added between nodes J2a
and J2 (pipe P2) and J2 and Jll (pipe P3) The roughness of these three pipes is
0.016 and their diameter is 1 ft. As usual, the lengths of the pipes are calculated
automatically with the Auto-Length tool.
2. At this point there are pairs of parallel links connected to the same input and output
nodes (for example C2a and P2\ C2 and PS). Right now, they are located at the same
elevation (the surface elevation). These elevations must be adjusted so that the sub-
surface pipes lie below the surface gutters. This is done by decreasing the invert
elevation of Jl and J5 by 4 ft (Jl = 4969 ft and J5 = 4965.8 ft), and raising the outlet
offset of C Auxl to 4 ft and the inlet offset of C Auxlto2 to 4 ft. Then, the invert
elevations ofJ2a and J2 are decreased by 4 ft (J2a = 4966.7 ft and J2 = 4965 ft), an
outlet offset of 4 ft is given to C_Aux2, an inlet and outlet offset of 4 ft is given to
C2a, and an inlet offset of 4 ft and outlet offset of 6 ft to C2.
3. The properties of the different elements of the model should be those listed in Tables
7-3 and 7-4. Table 7-3 summarizes the junction inverts; Table 7-4 shows the conduit
shapes and their inlet and outlet offsets.
Park Area Elements
The next phase is to add the series of pipes that drain the park area containing the
natural channel swales, shown in Figure 7-4. This process consists of the following steps.
1. A new node, Aux3, is added to the model with an invert elevation of 4974.5 ft. A new
swale, C_Aux3, is defined connecting nodes Aux3 and J3. Its roughness is 0.05 and
its cross section is the same as the cross section used in Example 2 for swales (for
instance, the cross section of conduit C4).
2. Next the series of pipes that drain the park area is created as shown in Figure 7-4.
These pipes (P4, P5, P6, P7 and P8) connect with the surface system at junctions
Aux3, J4, J5, J7, J10, and Jll. The latter represent inlet grate structures in the surface
swales. The roughness coefficient of these pipes is 0.016.
3. The pipes underneath the park have an initial diameter of 3 ft and lie 6 ft below the
invert elevation of the natural channel through the park. These pipes are 2 ft deeper
124
-------
than the pipes buried below the streets to avoid giving P5 an adverse slope. The
junctions where the swales and the pipe system connect together are assigned invert
elevations that are 6 ft below the ground surface. The invert elevations of these
junctions (in feet) become: Aux3 = 4968.5, J4 = 4965, J7 = 4963.5, J10 = 4957.8, Jll
= 4957 and Ol = 4956.
4. As currently defined, one pipe in the model, P6, has an adverse slope. This negative
slope is corrected by lowering the invert elevation ofJ7 from 6 ft below the surface to
8 ft (elevation J7 = 4963.5 ft).
5. The next step is to set the inlet and outlet offsets of the surface conduits where they
share nodes with the series of pipes to maintain their surface elevations. The inlet and
outlet offset elevations of C Aux3, C4 and CIO are 6 ft and 0 ft, 6 ft and 4 ft, and 6 ft
and 6 ft, respectively. Additionally, the inlet offset of C5 and C6 are increased to 4 ft
and 8 ft respectively, and the outlet offset of C3 and C9 are increased both to 6 ft.
These new offsets are defined to keep the swale at the surface. Note that the depths of
the gutters are assumed negligible in this example. The depths of the swales (3 ft) are
not considered negligible but were accounted for in Example 2 when their invert
elevations were modeled.
6.
The invert elevation of junctions J3, J6, J8 and J9 defined in Example 2 remain the
same for this example. The final layout should look like the one shown in Figure 7-6.
The final invert elevations of the junctions as well as the size, inlet and outlet offsets
of the conduits are summarized in Tables 7-3 and 7-4.
Full Street
Parallel System
(C2a&P2) and (C2&P3)
Half Street
^ "I
- i
i
i
i
i
i
Parallel System
(C10&P8)
01
Shared junctions
Park Pipe System
Parallel pipes
Swales
Figure 7-6. Post-development site with dual drainage system
125
-------
Junction Surface Elev. Change for Final Invert
ID Elevation (ft) Parallel Pipe (ft) Elevation (ft)
Jl
J2a
J2
J3
J4
J5
J6
J7
J8
J9
J10
Jll
Auxl
Aux2
Aux3
Ol
Table 7-4. Co
Conduit
ID
C2a
C2
C3
C4
C5
C6
C7
C8
C9
CIO
Cll
C_Auxl
C_Aux2
C_Auxlto2
C_Aux3
PI
P2
P3
P4
P5
P6
P7
P8
4973.0
4970.7
4969.0
4973.0
4971.0
4969.8
4969.0
4971.5
4966.5
4964.8
4963.8
4963.0
4975.0
4971.8
4974.5
4962.0
Type
Half_Street
Half_Street
Culvert
Swale
Swale
Swale
Culvert
Swale
Swale
Swale
Culvert
Full_Street
Full_Street
Full_Street
Swale
Pipe
Parallel Pipe
Parallel Pipe
Pipe
Pipe
Pipe
Pipe
Parallel Pipe
4
4
4
0
6
4
0
8
0
0
6
6
0
0
6
6
Initial Depth or
Diameter (ft)
1.3
1.3
2.25
3
3
3
3.5
3
3
3
4.75
1.3
1.3
1.3
3
1
1
1
3
3
3
3
3
4969.0
4966.7
4965.0
4973.0
4965.0
4965.8
4969.0
4963.5
4966.5
4964.8
4957.8
4957.0
4975.0
4971.8
4968.5
4956.0
jdjrainagje_sy^tein
Length
(ft)
157.48
526.00
109.00
133.00
207.00
140.00
95.00
166.00
320.00
145.00
89.00
377.31
239.41
286.06
444.75
185.39
157.48
529.22
567.19
125.98
360.39
507.76
144.50
Inlet Offset
(ft)
4
4
0
6
4
8
0
0
0
6
0
0
0
4
6
0
0
0
0
0
0
0
0
Outlet Offset
(ft)
4
6
6
4
0
0
0
0
6
6
0
4
4
0
0
0
0
0
0
0
0
0
0
126
-------
7.4 Model Results
General System Behavior
The first step in analyzing this dual drainage system is to run the model in its
preliminary sized state for the 2-yr storm with Dynamic Wave routing to check its general
performance. The hyetographs for both the 2-yr and 100-yr storms are the same as those
used in Example 1. For this run the reporting and wet-weather time steps were set to 1
minute, the routing time step to 15 seconds and the dry-weather runoff time step to 1
hour. In addition, the "Report Input Summary" check box on the General page of the
Simulation Options dialog was checked so that the slopes computed for all of the
system's conduits would be displayed in the run's Status Report. The input data file for
this initial system design is named ExampleT-Initial.inp.
After running this model, the resulting Status Report shows that no nodes are
either surcharged or flooded (see the Node Surcharge Summary and Node Flooding
Summary sections of the report). However, conduits PI, P2, and P3 were all surcharged
(see the Conduit Surcharge Summary section) indicating that the system was undersized.
The differences in these surcharge reports illustrate that node surcharge and
conduit surcharge are two different behaviors. Figure 7-7 shows several examples of
surcharge and flooding. Figure 7-7a shows conduit surcharge; Figure 7-7b shows conduit
surcharge, node surcharge and node flooding; Figure 7-7c shows conduit surcharge and
node surcharge; and Figure 7-7d shows conduit surcharge, node surcharge and node
flooding. Note that in the case of the open channel (Figures 7-7a and 7-7b) node
surcharge occurs simultaneously with node flooding.
Ground surface
Conduit
Surcharge
Invert Elevation
(a)
Conduit surcharge,
no node surcharge
Node surcharge
= node flooding
(c)
Node
surcharge
(d)
Node surcharge
& node flooding
Figure 7-7. Examples of surcharge and flooding
127
-------
The drainage criteria in Section 7.3 also require a minimum gutter grade of 0.4%
and maximum average flow velocity of 10 ft/sec. Therefore, it is required to check that
the minimum slope of the conduits representing gutters is larger than 0.4%. Table 7-5,
taken from the Link Summary table of the Status Report, shows the slopes of the conduits
representing the gutters. All the gutters have a longitudinal grade of 0.4% or greater and
thus meet the design standards considered for this example. The velocities in the gutters
and all the conduits will be checked once the pipes of the system are properly sized.
Table 7-5. Calculated gutter grades
Conduit ID
C2a
C2
C Auxl
C Aux2
C_Auxlto2
Length (ft)
157
526
377
239
286
Slope (%)
1.1
1.1
0.5
0.5
0.4
Design for the 2-yr Storm
The next step of the analysis is to size the pipes so that they can carry the 2-yr
storm without surcharging. In addition, the design criteria listed in Section 7.3 require the
collector streets to be sized so there is no curb topping and at least one lane remains free
of water. This phase of the analysis is carried out as follows:
1. The sizes of the pipes are iteratively modified, and the model run, until the Status
Report's Conduit Surcharge Summary shows that no conduits are surcharged and that
the values of the Max/Full Depth ratio in the Link Flow Summary are close to 0.85 for
all the pipes (a safety factor of 15% is considered to reduce the risk of surcharging).
Table 7-6 lists the standard pipe sizes that were used in this process. Table 7-7 shows
the results from three iterations, including the final one. Note how increasing the size
of the upstream pipes allows some of the downstream pipes to be reduced in size (but
made no smaller than the upstream ones).
2. Using the final pipe sizing from the previous step (Trial 3 in Table 7-6), the peak
velocities in the conduits representing the streets are checked in the Link Flow
Summary section of the Status Report. The peak velocity for all these conduits is less
than 10 ft/sec, the maximum allowed by the drainage criteria.
3. According to the drainage criteria, no curb-topping should occur for the minor storm
in any street. Figure 7.5 shows that curb-topping implies a Max/Full Depth ratio of
0.5/1.3 = 0.38. The Link Flow Summary in the Status Report for the final pipe sizing
shows that the highest value of the Max/Full Depth ratio for the street conduits is 0.29
for conduit C_Aux2. Thus, the final design of the system is appropriate in terms of the
flows in the streets.
The input file with the final pipe sizes for this example is named Example7-Final.inp.
128
-------
Table 7-6. Available (drainage nine sizes
in.
6
12
16
ft
0.5
1
1.33
in.
18
20
22
ft
1.5
1.67
1.83
in.
24
28
36
ft
2
2.33
3
in.
38
42
48
ft
3.17
3.5
4
Table 7-7. Iterations for 2-jr storm nine sizing
Pipe
PI
P2
P3
P4
P5
P6
P7
Size
(ft)
' I"""
1
1
o
J
3
3
3
Trial 1
Surcharged
yes
yes
yes
no
no
no
no
Max/Full
Depth
0.85
1
1 J
0.27
0.29
0.36
0.48
Size
(ft)
7.33
1.33
1.33
2
2
2
2
Trial 2
Surcharged
no
yes
no
no
no
no
no
Max/Full
Depth
0'.69
0.90
0.90
0.52
0.58
0.71
0.81
Size
...(**)_
1.33
7.5
7.5
7.67
7. S3
2
2
Trial 3, final
Surcharged
no
no
no
no
no
no
no
flax/Full
Depth
__
0.67
0.80
0.64
0.65
0.72
0.81
P8
no
0.53
.17
no
0.51
3.77
no
0.51
Major Storm Performance
Finally, the model is run for the 100-yr storm by changing the rainfall time series
used by its rain gage. The Status Report's Node Surcharge Summary and Node Flooding
Summary tables show that no nodes are surcharged or flooded during the 100-yr storm.
Additionally, the Conduit Surcharge Summary shows that, as expected, all of the pipes
are surcharged. According to the Link Flow Summary table, the highest velocities occur
in conduits C7, Cll and P8 (11.25, 11.93 and 10.76 ft/sec respectively), all of them over
10 ft/sec. Standard drainage criteria define a maximum allowable velocity in pipes and
culverts of 15 to 18 ft/sec (CCRFCD, 1999; Douglas County, 2008). Thus, the maximum
velocities predicted for pipe P8 and culverts C7 and Cll are acceptable, and no re-sizing
of the conveyance system is required. Finally, the Link Flow Summary table shows that
the Max/Full Depth value of all street conduits is less than one. Therefore the water depth
in the streets never reaches the crown or the highest point of the sidewalks, and the
requirements for the 100-yr storm are successfully satisfied.
To illustrate the complex flow conditions that can occur in these systems, Figure
7-8 shows the hydrographs in pipes P5 and P6 under the 100-yr storm. Negative flows
caused by backwater effects occur in both pipes after 30 minutes of simulation. These
special flow conditions can be only simulated when Dynamic Wave routing is used,
which demonstrates the capabilities of this method. These negative discharges do not
occur for the 2-yr storm. This shows that the behavior of a system under a minor and a
major event can be quite different. In this case, the runoff generated by subcatchment S4
into junction J7 for the 100-yr storm is large enough to cause a backwater effect that
generates these negative flows.
129
-------
15.0
10.0
OT
LL
O
-5.0
-10.0
| Link P6 Link P5~~|
0.5
1 1.5 2
Elapsed Time (hours)
2.5
Figure 7-8. Flows in pipes P5 and P6 during the 100-yr storm
Another way to visualize the behavior of the dual drainage system is with Profile
Plots. Figure 7-9 contains two such plots stacked on top of one another. They depict the
surcharge condition that occurs between nodes J2a, J2, and Jll at 34 minutes into the
100-yr storm. The lower plot applies to the below-ground sewer pipes P2 and P3 and
shows that both pipes are surcharged. The upper plot is for the streets C2a and C2 that lie
above P2 and P3. The water levels at junctions J2a and J2 are high enough to cause
water to flow out of the sewer pipe P2 and onto the street, but not high enough to flood
the street. On the other hand, the invert elevation of junction Jll is low enough so that
the surcharged pipe P3 does not create street flooding and instead, street flow re-enters
the sewer system there. (The plot makes it appear that the flow at the downstream end of
street C2 is zero, but this is just an artifact of the way that SWMM draws the water
surface profile within a conduit (by connecting the water elevations between its end
nodes without allowing the profile to cross either its bottom or top surface)).
7.5 Summary
This example built on the simple drainage system modeled in Example 2. It
converted some of the open surface channels used in that example to parallel pipe and
gutter systems and added a new series of pipes to drain the downstream section of the
park. The system was sized for a minor (2-yr) storm event and its behavior was also
analyzed under a major (100-yr) storm event. The key points illustrated in this example
were:
130
-------
-
4,972-
4,970-
4,968-
g
^ 4,966:
ฃ 4,964:
nj
4,962:
4,960^
4,958:
4,972^
4,970^
n 4,968^
'B 4,966:
is
ฃ 4,964^
LLl
4,962:
4,960:
4,958:
Street
Sewer
ion
200
300 400
Distance (ft)
500 600
01/01/200700:34:00
Figure 7-9. Surcharge behavior along the eastern boundary of the site
1. The three-dimensional structure of dual drainage systems can be modeled by using
manhole junctions set below ground that connect parallel pairs of sewer pipes and
gutter/street channels, where the latter are offset to ground elevation.
2. These systems are designed so that the below-ground sewer system will flow only
partly full during the more frequent minor storm events and will surcharge into the
street channels during the larger, less frequent major events without causing any
overtopping of the street.
3. An iterative process can be used to properly size the sewer pipe elements to meet both
the small storm and large storm design criteria.
4. Most of the results needed to evaluate the performance of a dual drainage system can
be found in the various tables produced by SWMM's Status Report.
As this example shows, dual drainage system models require a significant amount
of additional effort to set up. The need to represent both the minor and major portions of
a drainage system will depend on the objectives set forth for the analysis and on the level
of detail required from the model.
131
-------
Example 8. Combined Sewer Systems
This example demonstrates how to model systems that convey both sanitary
wastewater and stormwater through the same pipes. Systems like these are known as
combined sewer systems and are still quite common in older communities and cities.
During periods of moderate to heavy rainfall the capacity of these systems to convey and
properly treat the combined flow can be exceeded, resulting in what are known as
Combined Sewer Overflows (CSOs). CSO discharges can cause serious pollution
problems in receiving waters. Contaminants from these discharges can include
conventional pollutants, pathogens, toxic chemicals and debris.
This example will use SWMM to analyze the occurrence of overflows in a
combined sewer system. Particular attention is paid to properly representing the flow
regulators that divert flow between collection sewers, treatment plant interceptors and
CSO outfalls. In addition, the example shows how to model a pump station that conveys
the intercepted flow through a force main pipeline to the headworks of a treatment
facility. Although the focus of this example is on combined systems, many of the same
modeling elements it employs (wastewater inflows, pump stations, and force mains) can
also be used to model separate sanitary sewer systems.
8.1 Problem Statement
Rather than being a new development, the 29 acre urban catchment studied in
Example 2 is now assumed to be an older area that is served by an existing combined
sewer system. The hydraulic behavior of this combined system, including the magnitude
of any overflows, will be analyzed for several different size storms. These include the
0.23 in. water quality storm defined in Example 3 as well as the 1.0 in., 2-yr and 1.7 in.,
10-yr storms used throughout the previous examples.
Combined sewer pipes conveying both wastewater and stormwater flows
generated within different sewersheds (i.e., areas that contribute wastewater flows to a
single point) will be added to the model. Constant wastewater flows (also known as dry-
weather flows) will be based on average generation rates per capita. An interceptor pipe
will be sized to convey both the base dry weather flow and a portion of the combined
stormwater flow to a pump station that pumps to the headworks of a wastewater
treatment plant (WWTP) through a force main. Various combinations of orifices, weirs
and pipes will be used to represent different types of flow diversion structures located
within the interceptor. The CSOs that cannot be diverted by these devices will discharge
directly into the stream running through the site's park area.
The schematic representation of the combined sewer system modeled in this
example is shown in Figure 8-1. It includes the combined sewer pipes (in green) that
drain the subcatchments (or sewersheds) SJ, S2, S3, S4 and ฃ5, the stream (in blue), the
interceptor (in brown), the flow regulators (red boxes), and the pump station.
132
-------
To WWTP
Stream
Culvert in the stream
Interceptor
Combined sewer pipes
Flow regulators
Figure 8-1. Combined sewer system study area
8.2 System Representation
Combined sewer systems are systems that convey both sanitary sewerage and
stormwater through the same pipes. Interceptors are pipes designed to capture 100% of
the sanitary flows during dry weather periods and convey them to a WWTP. During
periods of moderate or heavy rainfall, however, the wastewater volume in the combined
sewer system can exceed the capacity of the interceptor or the WWTP. For this reason,
combined sewer systems are designed to discharge the excess wastewater directly to a
nearby stream or water body through diversion regulators. Figure 8-2 shows a schematic
representation of a combined sewer system and CSO occurring in the system. The figure
shows how for wet-weather flows the interceptor at the bottom is able to convey only part
of the flow into the WWTP and CSOs occur.
133
-------
Figure 8-2. Conceptual representation of overflows in a combined system (Field and Tafuri, 1973)
Dry Weather Flows
To create a combined sewer system in SWMM one adds dry weather wastewater
flows into the appropriate nodes of a previously created stormwater conveyance system.
These nodes typically represent locations where collector sewers discharge into trunk
sewers. Their number and location will depend on the level of aggregation used to
combine individual wastewater sources (homes, businesses, etc.) together. The sidebar
below explains how to use a node's Inflow Editor to specify the time series of dry
weather flow entering the node.
134
-------
Adding Dry Weather Flows into SWMM
This example requires that dry weather flows representing the wastewater
discharges be added into the combined sewer system model. The Dry Weather page
of the Inflow Editor is used to specify a continuous source of dry weather flow or
any pollutant entering a specific node of the drainage system. The Inflow Editor is
accessed through the Inflows property of the node for which the flow is being
defined. Dry weather flows in SWMM are characterized by an average (or baseline)
value and up to four optional time patterns (TP) that can represent monthly, daily
and hourly (for both weekday and weekend) variations. Dry weather flows are then
computed as shown below:
Dry weather flow at time t = (average value)*(TP lt)*(TP 2t)*...
where TPlt is the multiplier for time pattern 1 at time t, TP2t is the multiplier for
time pattern 2 and so on.
If no time patterns are defined then the dry weather flow entering the node is
simply the average value. The Inflow Editor for the dry weather flow at a particular
node is shown below.
ซ|RDM |
Avenge Value
(CFS]
~34|X|
^1*1 X|
NOT! L2.=,ซe-..-.v2rE;g2 value-he-l1:,l=r-,l t,:, rerr,,:,",? .=,",;. Jr.'
weaiher inflow (or .3 given i:on:hhi|,?nt ai this node
Flow Regulator Structures
Flow regulators (or diversion structures) are used to control the flow between
collection sewers and the interceptor. These regulators allow the conveyance of
wastewater to treatment facilities during dry weather conditions. During wet weather
conditions the regulators divert flows away from the interceptor and discharge directly
into a water course to avoid surcharge and flooding of the combined sewer system. Flow
regulator devices include side weirs, leaping weirs, transverse weirs, orifices and relief
siphons. Metcalf & Eddy, Inc. (1991) presents a detailed description of these different
devices. This particular example will use the transverse weir with orifice type of
regulator illustrated in Figure 8-3. In this regulator there is a weir or a small plate placed
directly across the sewer perpendicular to the line of flow. Low flows are diverted to the
interceptor through an orifice located upstream of the weir. During periods of high flow,
the weir is overtopped and some flow is discharged through the overflow outlet,
eventually reaching a CSO outfall.
135
-------
Overflow
outlet
Transverse weir
Pipe or orifice
diversion to
interceptor
Combined
sewer flow
To the
interceptor
Figure 8-3. Transverse weir flow regulator
A transverse flow regulator can be represented in SWMM by using weir and
orifice elements. Because these elements correspond to hydraulic links, additional
junction nodes must be added into the model. A schematic representation of three
possible definitions of a transverse flow regulator in SWMM is shown in Figure 8-4.
Each of these three configurations will be used in this example. Configurations (a) and
(b) both contain the weir shown in Figure 8-3, but use different elements to divert to the
interceptor: (a) uses a bottom orifice while (b) uses a pipe. Finally, the third configuration
(Figure 8-4c) uses neither a weir nor an orifice. Instead, it simply diverts flow by using
different inlet offsets for the pipes that convey flows to the interceptor and to the stream.
The first pipe has an inlet offset of zero while the pipe linked to the stream has a larger
invert elevation.
(a)
w
0
Transverse Weir
Bottom Orifice
Junction
Combined sewer
Interceptor
Stream
Regular conduit
Figure 8-4. Alternative ways to represent tranverse weir flow regulators in SWMM
136
-------
This example uses each of these regulator configurations for purely illustrative
purposes. The choice of a particular configuration to use in a real application will depend
on the specific conditions encountered in the field and on how numerically stable the
resulting model will be. Some unwanted hydraulic phenomena that can be artificially
introduced into the model by these different representations include surcharged weirs,
instabilities caused by short pipes and excessive storage associated with large pipes.
Pump Stations
Pumps are devices used to lift water to higher elevations. They are defined in the
model as a link between two nodes and can be in-line or off-line. The principal input
parameters for a pump include the identification of the inlet and outlet nodes, its pump
curve, initial on/off status and startup and shutoff depths. A pump's operation is defined
through its characteristic curve that relates the flow rate pumped to either the water depth
or volume at its inlet node or to the lift (i.e., hydraulic head) provided. Its on/off status
can be controlled dynamically by defining startup and shutoff water depths at the inlet
node or through user-defined control rules. A pump is defined in the model in the same
fashion as any other link, while the pump curve is created using the Pump Curve Editor
and is linked to the pump by the latter's Pump Curve property as shown in Figure 8-5.
(a)
(b)
Name
Inlet Node
Outlet Node
Description
Tag
Pump Curve
Initial Status
Startup Depth
Shutoff Depth
15
JI7
02
ON
0
0
Name of pump curve (or * for ideal pump). After specifing
a curve, you can double-click to edit it]
Pump Type
|TYPE1
Description
1
2
3
4
5
6
7
8
9
Volume
(113)
Fk>ป
(CFS) _
2.
Figure 8-5. (a) Pump Property Editor and (b) Pump Curve Editor
8.3 Model Setup
Preliminaries
Figure 8-6 shows the system to be modeled in this example. Example 7
(ExampleT-Final.inp) is the starting point for the model setup, although major changes
are required. Because combined sewer systems are no longer used for new developments,
this example assumes that the combined sewer system being modeled has been in place
for many years.
137
-------
IR1
I Diversion Junctions W Transverse Weir
O Bottom Orifice
f J J Junction
Combined sewer
I Flow
Regulator |
Stream ' _ _ J
Figure 8-6. Schematic representation of the combined sewer system
In modifying the model of Example 7, the gutter elements will be removed as will
the pipes along the stream running through the park. A new interceptor sewer will be
placed along the north side of the stream, as illustrated in Figure 8-1. This interceptor will
convey wastewater flows to a pump station comprising a storage unit, which represents a
wet well, and a pump. The pump discharges through a force main line to a constant head
outfall (02) representing the inlet to a hypothetical WWTP. New pipes representing the
combined sewer system need to be added as well as several weirs and orifices that define
the flow regulators. The streambed for this example is lowered by 5 ft compared to the
original stream bottom elevations used in Example 2 so that backwater from the stream
will not flood the regulators in the combined sewer system.
These modifications are made by removing junctions Auxl and Aux2 as well as
conduits C2a, C2, C Auxl, C Aux2, C Auxlto2, P5, P6, P7 and P8. Then, the elevations
of the nodes and the inlet/outlet offsets of the links that comprise the stream in the park
must be changed as well. The new invert elevations of the nodes in the stream (Aux3, J3,
J4, J5, J6, J8, J9, J10, Jll and outlet Ol which are colored blue in Figures 8-1 and 8-6)
are listed in Table 8-1. Their maximum depths are set to zero so that the program will
automatically adjust their depths to match the top of the highest connecting stream
conduit. The remaining junctions (Jl, J2, J2a, and JT) have their depths set equal to the
ground elevation minus their invert elevation. The offsets of the swales and culverts that
run through the park (links C3 through Cll) are set back to their original values of zero.
Combined Sewer Pipes
The next step is to create the combined sewer pipes, shown in green on Figure 8-6
and identified with the letter P. All of them have a roughness coefficient of 0.016. Pipes
PI, P2, P3 and P4 were defined already in Example 7 and two more combined sewer
pipes are added; P5 will convey flows from subcatchment S4 and P6 from subcatchment
138
-------
S5. The upstream nodes (or inlet nodes) for these pipes, J13 and J12, are added to the
model as well. The properties of these junctions are shown in Table 8-1. The properties
of the combined sewer pipes are shown in Table 8-2.
Table 8-1. Properties of the combined sewer system nodes1
Node ID
Jl
J2a
J2
J3
J4
J5
J6
J7
J8
J9
J10
111
Invert
Elevation (ft)
4969.0
4966.7
4965.0
4968.0
4966.0
4964.8
4964.0
4960.0
4961.5
4959.8
4958.8
4958.0
Maximum
Depth (ft)
4.2
4.0
4.0
0.0
0.0
0.0
0.0
12.0
0.0
0.0
0.0
0.0
Node ID
J13
Aux3
JI1
JI2
JI3
JI4
JI5
JI6
JI7
JI8
JI9
Ol
Well
Invert
Elevation (ft)
4968.0
4968.5
4958.0
4957.0
4955.0
4952.0
4950.0
4967.0
4967.0
4962.0
4960.0
4957.0
4945.0
Maximum
Depth (ft)
4.8
0.0
16.0
15.8
16.0
14.0
16.0
7.2
6.0
6.2
5.2
-
14.0
1 Colors indicate whether the nodes belong to the stream (blue), the sewer pipes (green), the
interceptor (brown) or they are the diversion junction (grey)
Subcatchment Outlets
With the sewer pipes entered in the model, it is necessary now to redefine the
outlet nodes of the different subcatchment. These nodes will receive the stormwater
runoff generated by the subcatchments as well as the flows corresponding to the
wastewater flows, defined later in this section. In other words, the subcatchments used for
drainage are also defined as sewersheds in this example. No other changes to the
properties of the subcatchments are required. Table 8-3 shows the new outlets for the
subcatchments.
Interceptor Pipe
An interceptor line is now added to the model that runs along the north side of the
stream and conveys all wastewater flows to the pump station located on the east side of
the study area. Its pipes are identified with the letter I and brown color, as shown in
Figure 8-6. Conduits II, 12,13,14 and 19 are the main pipes of the interceptor. The new
nodes belonging to the interceptor are identified with the letters JI with the last node
being the pump's wet well which is a storage unit named Well. Properties of the nodes
and pipes of the interceptor are summarized in Tables 8-1 and 8-2. Figure 8-7 displays
how the combined sewer system looks before the interceptor is connected to the rest of
the system with flow regulators and the pump station is defined.
139
-------
Table 8-2. Properties of the combined sewer system conduits1
Pipe ID
C3
C4
C5
C6
C7
C8
C9
CIO
Cll
C_Aux3
PI
P2
P3
P4
P5
P6
11
12
13
14
15
16
17
18
19
Shape
Circular
Trapezoidal
Trapezoidal
Trapezoidal
Circular
Trapezoidal
Trapezoidal
Trapezoidal
Circular
Trapezoidal
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Circular
Inlet
Node
J3
J4
J5
J7
J6
J8
J9
J10
Jll
Aux3
Jl
J2a
J2
Aux3
J13
J12
Jll
JI2
JI3
JI4
JI7
J7
JI8
JI9
JI5
Outlet
Node
J4
J5
J6
J6
J8
J9
J10
Jll
01
J3
JI7
J2
JI9
JI6
J7
JI8
JI2
JI3
JI4
JI5
JI2
JI3
JI4
JI5
Well
Length
(ft)
109.00
133.00
207.00
140.00
95.00
166.00
320.00
145.00
89.00
444.75
185.39
157.48
529.22
567.19
377.76
498.42
150.36
230.38
578.27
124.45
10.65
153.02
32.88
47.72
100
h or d
(ft)2
2.25
3
o
j
3
3.5
3
o
j
3
4.75
3
1.33
1.5
1.5
1.67
1.67
1.67
1
1
1.5
1.5
0.33
0.66
0.5
0.5
2
Rough.
Coeff.
0.016
0.05
0.05
0.05
0.016
0.05
0.05
0.05
0.016
0.05
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
0.016
b
(ft)3
0
5
5
5
0
5
5
5
0
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Zl4
0
5
5
5
0
5
5
5
0
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
ซ s Inlet
Offset (ft)
0
5
5
5
0
5
5
5
0
5
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
5
0
0
0
0
0
6
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Outlet
Offset (ft)
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
4
1 Colors indicate whether the pipes belong to the stream (blue), the sewer pipes (green) or the
interceptor (brown)
2 h or d corresponds to the depth (Trapezoidal shape) or diameter (Circular shape)
3 b corresponds to the bottom width (Trapezoidal shape)
4'5 Zl and Z2 correspond to the left and right slope (Trapezoidal shape)
Table 8-3. Subcatchment outlets
Subcatchment Outlet Node
SI Jl
S2 J2a
S3 Aux3
S4 J13
S5 J12
S6 Jll
S7 J10
140
-------
J2a
Well
Stream
Sewer
Interceptor
S6"
Figure 8-7. Layout of the combined sewer system before adding regulators and pump
Flow Regulators
The flow regulator structures are the next elements to be represented in the model.
Five regulators identified with the letter R will be used to control the flows from the five
combined sewers (PI, P3, P4, P5 and P6) into the interceptor. These identifiers
(RJ,...,R5) are given only for reference purposes in the text; in the actual model
regulators are not defined as elements directly, but rather through a combination of
orifices, weirs and pipes (i.e., in the model there is no element named Rl), as seen in
Figure 8-6. The steps needed to create each of these regulators are as follows: (Note: it is
recommended that SWMM's Auto-Length feature be turned off so that conduit
lengths are not altered as connections are made to the flow regulators.)
Regulator Rl:
1. Place a new junction named JI6 somewhere between J4 and JI1.
2. Change the downstream node of pipe P4 to JI6 (using the pipe's Property Editor).
3. Add a weir link Wl connecting JI6 to J4.
4. Add a bottom orifice link Ol connecting JI6 to JI1.
141
-------
Regulator R2:
1. Place a new junction JI7 somewhere between J5 and JI2.
2. Change the downstream node of pipe PI to JI7.
3. Add weir link W2 connecting JI7 to J5.
4. Add a new pipe 75 connecting JI7 to JI2.
Regulator R3:
1. Add a new pipe 16 between J7 and JI3.
Regulator R4:
1. Place a new junction JI8 somewhere between J10 and JI4.
2. Change the downstream end of pipe P6 to JI8.
3. Add a weir link W3 that connects JI8 to J10.
4. Add a pipe 17 connecting JI8 to JI4.
Regulator 5:
1. Place a new junction JI9 somewhere between Jll and JI5.
2. Change the downstream node of pipe P3 to JI9.
3. Add a weir link W4 that connects JI9 to Jll.
4. Add a pipe 18 that connects .7/9 to JI5.
The resulting layout of the combined system with the regulators added is shown
in Figure 8-6. Note that adding the regulator junctions JI6, JI7, JI8, and JI9 required
changing the discharge nodes of combined sewer pipes P4, PI, P6, and P3, respectively,
from their original stream junctions to their respective regulator junctions.
To complete the process of adding in the regulators, the properties of the new
junctions (JI6, JI7, JI8, and JI9) must be set using the data from Table 8-1. The same
holds true for the new interceptor pipes (75, 16, 17, and IS) whose properties are listed in
Table 8-2. The dimensions and offsets for the newly added weirs and orifices (WJ, W2,
W3, W4, and Of) are listed in Table 8-4. Note that for each weir, the sum of its inlet
offset and its opening height is smaller than the depth of the corresponding inlet node,
defined in Table 8-1. The discharge coefficient for each of the weirs is 3.3 and is 0.65 for
the orifice. In addition to these newly added elements, it is necessary to set the inlet offset
for stream channel C6 to 5 ft higher than the invert of junction J7 so that it can serve as
an overflow diversion while the new pipe 16, with no offset, connects J7 to the
interceptor (see Table 8-2)
142
-------
Flow
regulator
Rl
R2
R3
R4
R5
Sewer
pipe
controlled
P4
PI
P5
P6
P3
Inlet
node
JI6
JI7
J7
JI8
JI9
Discharge into stream
ID
Wl
W2
C6
W3
W4
Height
(ft)
6.5
5.5
5
4.5
Length
(ft)
o
J
5
Pipe, see
4
4
Inlet
offset
0.4
0.33
Table 8-3
0.4
0.35
Outlet
node
J4
J5
J10
Jll
Discharge
Type
I Orifice
Pipe
Pipe
Pipe
\ Pipe
ID
Orl
15
16
17
18
into interceptor
Diameter Outlet
(ft) node
0.5
See Table
See Table
See Table
See Table
Jll
8-2
8-2
8-2
8-2
Pump Station and Force Main
To complete the combined system model the pump station and force main must
be added at the downstream end of the interceptor. As described earlier, node Well will
serve as the wet well for the pump station. It is represented by a storage node with an
invert elevation of 4945 ft, a maximum depth of 14 ft, an initial depth of 3 ft, and has a
surface area of 300 ft2 that remains constant over its entire height. To specify the latter,
the Storage Curve entry in the node's Property Editor is set to Functional, the
Coefficient and Exponent entries are zero and the entry for the Constant field is 300.
To make room for the force main on the study area map, the map's dimensions
must be expanded. Select View | Dimensions from the main menu bar. In the Map
Dimensions dialog that appears enter -235.8, -70.2 for the Lower Left coordinate and
1960.5, 1514.2 for the Upper Right coordinate. Do not select the "Re-compute all lengths
and areas" option which will appear if the Auto-Length option is currently on.
There should now be room to add a series of four junctions downstream of the
wet well node that defines the path of the force main (JI10, Jill, Jll2, and Jll3) along
with the final outfall node (02} that represents the WWTP (see Figure 8-8). The invert
elevations of these nodes are 4947.0, 4954.8, 4962.6, 4970.4, and 4968.0 ft, respectively.
As explained in the sidebar "Defining a Force Main Pipeline" all of the junction nodes in
the main are assigned zero maximum depth and a surcharge depth of 265 ft so that they
can pressurize (to at least 115 psi) without flooding. The outfall O2 is of type Fixed, with
a fixed stage elevation of 4970.0 ft.
O2, 4968 ft
-= Fixed stage,
112, 4962.6 ft
4970 ft
Jll 1,4954.8 ft
Jll 0,4947ft
Well
Figure 8-8. Force main line
143
-------
Defining a Force Main Pipeline
The junctions used to connect sections of a closed
force main together are typically welded or bolted fittings
that do not allow water to escape the junction when it
becomes surcharged. This type of junction can be
modeled in SWMM by using a junction node whose
maximum depth is zero and whose surcharge depth is set
to an arbitrarily high number (such as the burst pressure
of the main). Using a maximum depth of zero means that
the actual depth will equal the distance from the node's
invert to the top of the highest connecting conduit. A high
surcharge depth is required to allow the node to remain in
a pressurized state without causing any flooding.
D
Property Value
Name
X-Coordinate
Y-Coordinate
Description
Tag
nflows
Treatment
nveit El.
Maw. Depth
Initial Depth
Surcharge Depth
Ponded Area
JI10
1573.004
618.390
NO
NO
4947
0
0
50 !
0
Depth in excess of maximum depth before flooding
occurs [ft]
Any of SWMM's closed conduit shapes can be used for the cross-section of
a force main, although circular pipes are most commonly used. SWMM also
provides a special circular shape named Force Main which uses either the Hazen-
Williams or the Darcy-Weisbach formulas to compute friction losses during
pressurized flow instead of the Manning equation as would otherwise be the case.
Some engineers prefer to use one of the former two head loss formulas instead of
the Manning equation for pipes that are known to have pressurized flow almost all
of the time. In the current example, the force main consists of simple circular pipes
and the Manning equation is used throughout to compute losses.
After these nodes are created, a set offeree main pipes 110, 111, 112 and 113 are
added between nodes JI10, Jill, JI12, JI13, and 02, respectively (refer to Figure 8-8).
The properties of these pipes are listed in Table 8-5.
Table 8-5. Properties of the force main line
_. T_ , T , , -T , i ,, , , i Length Diameter Rough. Inlet Outlet
Pipe ID Shape Inlet Node Outlet Node * ,,,. 5,. ,. , ,,,. ,. , ,,,.
1 l (ft) (ft) Coeff. Offset (ft) Offset (ft)
110
111
112
113
Circular
Circular
Circular
Circular
JI10 (4947 ft)
Jill
JI12
JI13
(4954.8
(4962.6
(4970.4
ft)
ft)
ft)
Jill
JI12
JI13
O2 (4968)
500
500
500
500
2
2
2
4
0
0
0
0.
.016
.016
.016
.016
0
0
0
0
0
0
0
0
1 Number in parenthesis indicates the invert elevation of the node.
The final step is to add a pump link named Pumpl between the nodes Well and
JI10. The pump curve (yet to be defined) associated with this pump is also named
Pumpl, the initial status is OFF, and the startup depth is 5 ft while the shutoff depth is 2
ft. This means that the pump turns on when the water depth in the wet well reaches 5 ft
and it shuts down when the depth drops to 2 ft.
144
-------
Each pump added to a SWMM model is required to have a curve that defines how
the pump's discharge flow rate depends on the volume, depth or head added at its inlet
(suction) side. This example will use a Type 3 pump curve where discharge varies
inversely with the head (or lift) added by the pump between its inlet and outlet nodes.
This type of curve provides the most realistic representation of how pumps actually
behave and is usually available from the pump's manufacturer. The head-discharge data
that define the pump curve used here are listed in Table 8-6. To add this curve to the
model:
1. In the Data Browser, select the Pump Curve sub-category.
2. Click the + button to bring up the Pump Curve Editor dialog.
3. Enter Pumpl as the name of the curve and select TypeS as the type.
4. Enter the Head-Discharge data from Table 8-6 into the grid on the form.
TableJM>.J>umj)_curve data
Head (ft) 11 12 13 14 15 16 17 18 19 20 21 22 23 24
Discharge (cfs) 7.2 6.9 6.5 6.1 5.7 5.2 4.7 4.1 3.6 2.9 2.3 1.5 0.8 0
Dry Weather Flows
Wastewater flow from the sewersheds is added into the model using SWMM's
Dry Weather Inflow tool (see the sidebar "Adding Dry Weather Flows into SWMM'}.
This example assumes that the individual subcatchments in the model also represent the
individual sewersheds. Only average daily dry weather inflows will be used to keep
things simple. In reality, peak daily flows can be two to four times greater than the
average. Therefore, if a more accurate analysis is required, a diurnal time pattern should
be used to capture the full range of dry weather inflows. This, together with a continuous
rainfall record, would allow one to simulate the dynamic performance of the system
under all types of events.
Typical per capita domestic wastewater generation rates vary between 40 gpd
(gallons per day) for apartments and 150 gpd for luxury residences and estates (Nicklow
et al., 2004). ASCE (1992) defines a range of average per capita domestic loading rates
between 50 gpd and 265 gpd. Based on these ranges and an estimate of 3 to 5 inhabitants
per lot, an average domestic loading rate per lot of 300 gpd is assumed. In addition to the
domestic rates, the discharge rates from the commercial areas in subcatchments S5 and S6
also need to be included. These are estimated to be 2850 gpd in subcatchment S5 and
8100 gpd in subcatchment S5. The dry-weather flows for the subcatchments are then
computed as the sum of the domestic loading rate and the commercial rate. Table 8-7
summarizes this computation, showing the number of residential lots in each
subcatchment (sewershed), the corresponding dry weather flows (in gpd and cfs) and the
nodes that receive these inflows. Note that no dry weather flows are used in
subcatchments S6 and S7 because they contain no residential or commercial lots.
145
-------
Table 8-7. Summary of dry weather flows
c. i. .. , .. TWT j Number of
Subcatchment Node . , ,. , , ,
residential lots
SI
S2
S3
S4
S5
S6
S7
Jl
J2a
Aux3
J13
J12
Jll
J10
17
22
10
17
-
-
-
Commercial dry Dry weather Inflow
weather inflow gpd
5100
6600
3000
2850 5100+2850
8100 8100
0
0
cfs
0.008
0.010
0.004
0.0123
0.0125
0
0
Precipitation Data and Simulation Options
The 0.23 inches storm defined previously in Table 5-1, Example 5, and both the
2-year and 10-year storms will be used to evaluate the performance of the combined
sewer system. The 0.23 in. storm is added to the model by creating a new time series
called 0.23-in. with the corresponding values of time and intensity. Dynamic Wave flow
routing with a time step of 15 s, a 1 minute wet weather runoff time step, a 1 hr dry
weather runoff time step, a 1 minute reporting time step and a total duration of 12 hours
will be used in all the simulations. All the information for the model has been
summarized in Tables 8-1 to 8-7. The complete model input data can be found in the file
named ExampleS.inp and the study area map should resemble that shown in Figure 8-9.
Figure 8-9. Final combined sewer system model layout
146
-------
8.4 Model Results
0.23 in. Storm
The model is first run for the 0.23 in. storm event. Viewing the Link Flow
Summary of the resulting Status Report shows that there is no flow in any of the links that
divert water from the combined sewers into the stream (Wl, W2, W3, W4 and C6). Thus
no CSOs occur for this size storm. Figure 8-10 shows the flow through each channel of
the stream. The only channel with any flow is Cll which receives only stormwater runoff
(and no wastewater flow) from subcatchment 57. Flows in any other of the stream's
conduits would imply that CSOs are occurring somewhere in the system.
C_Aux3
C5
CIO
C4
C7
Cll
0:00 1:00 2:00 3:00 4:00 5:00 6:00 7:00 8:00 9:00 10:00 11:00 12:00
time (hhrmm)
Figure 8-10. Flow (Q) along sections of the stream for the 0.23 in. storm
Figure 8-11 shows the flows through each section of the interceptor for the 0.23
in. storm. Note how the peak discharges and volumes increase in the downstream
direction (from II to 19) as the different combined sewer pipes discharge into the
interceptor through the flow regulators, as illustrated in Figure 8-12. The latter figure
shows that all the orifices and pipes that connect the combined sewers to the interceptor
are contributing flows to the interceptor. In this figure, the first ID identifies the link of
the regulator that carries the diverted flow; the second ID corresponds to the flow
regulator to which it belongs.
2-yr Storm
Results obtained for the 2-yr storm (with a volume of 1.0 in.) are now presented.
Figures 8-13 and 8-14 show the flows through the various stream and interceptor
sections, respectively. Note that for this larger storm, flow occurs in all sections of the
stream. The Link Flow Summary of the Status Report shows that CSOs occur and all the
147
-------
flow regulators start discharging flow into the stream once the capacity of diversion into
the interceptor has been reached. The irregular fluctuations in flow through the
interceptor seen in Figure 8-14 are caused by fluctuations in the flow pumped into the
force main at the pump station.
7.5
7
6.5
6
5.5
5
4.5
4
3.5
3
2.5
2
1.5
1
0.5
0
II
12
13
14
19
0:00 1:00 2:00 3:00 4:00
5:00 6:00 7:00
time (hhrmm)
8:00 9:00 10:00 11:00 12:00
Figure 8-11. Flow (Q) along sections of the interceptor for the 0.23 in. storm
a
2.5
2.25
2
1.75
1.5
1.25
i
0.75
0.5
0.25
0
Orl, Rl
16, R3
18, R5
I5,R2
17, R4
0:00 1:00 2:00 3:00 4:00
time (hhrmm)
5:00
6:00
Figure 8-12. Flow (Q) diverted from each regulator to the interceptor for the 0.23 in. storm
148
-------
0:00
1:00
2:00 3:00
time (hhrmm)
Figure 8-13. Flow (Q) along sections of the stream for the 2-yr storm
4:00
10
4 -
0:00
1:00 2:00
time (hhrmm)
4:00
Figure 8-14. Flow (Q) along the sections of the interceptor for the 2-yr storm
Figure 8-15 compares the flows that are diverted into the interceptor and to the
stream CSO for flow regulators Rl and R4. The interceptor diversions are from orifice
Orl and conduit 77, while the CSO discharges are from weirs Wl and W3. The regulators
are able to convey the discharges into the interceptor for a while, but at a certain point the
maximum capacity is reached and a CSO occurs. Note that the durations of CSOs are
149
-------
smaller than those of the discharges into the interceptor; however, the peak discharge is
larger.
10
9
8
7
6
<& c
y 5
x'
a 4
3
2
1
0
0:00
Rl, flow to interceptor
Rl, flow to stream
R4, flow to interceptor
R4, flow to stream
1:00
2:00 3:00 4:00
time (hhrmm)
Figure 8-15. Flows (Q) in regulators Rl and R4 for the 2-yr storm
5:00
6:00
Pump Station Behavior
Figure 8-16 shows how the pump in this combined system behaves under the 0.23
in. event. Part (a) of the figure plots the water depth in the pump's wet well and part (b)
shows the pump's discharge flow rate. The startup and shutoff depths are shown with
dashed lines. At the start of the simulation the water depth is 3 ft and the pump is initially
off (Point 1). The pump turns on once the startup depth of 5 ft is reached (Point 2). Inflow
to the pump station is large enough so that the pump continues working and the wet well
water depth stays above the shutoff level. The water depth eventually reaches a maximum
(Point 3) and then starts decreasing until reaching the 2 ft shutoff depth at which time the
pump stops operating (Point 4). After the runoff flow ceases some 2+ hours into the
simulation, only wastewater flows are received at the wet well, and the water depth
increases slowly again until reaching the startup depth (Point 5); the pump turns on but
rapidly stops when the shutoff depth is reached (Point 6). From hereafter pumped
discharges fluctuate between the startup and shutoff limits
150
-------
8.0
u
x
I
7.0
6.0
5.0
4.0
3.0
20
1.0
0.0
468
Elapsed Time (hours)
10
12
(b)
\
HapsedTime (hours)
0 2 4 6 8 10
II I
12 4
Figure 8-16. Pump behavior for the 0.23 in. storm: (a) wet well water depth, (b) pump flow
12
151
-------
Node ID
Max. wet well water depth (ft)
Max. interceptor pump flow (cfs)
Max. flow at stream outfall Ol (cfs)
Max. overflow at regulator Rl (cfs)
Max. overflow at regulator R2 (cfs)
Max. overflow at regulator R3 (cfs)
Max. overflow at regulator R4 (cfs)
Max. overflow at regulator R5 (cfs)
0.23 in.
5.35
7.2
0.88
0
0
0
0
0
2-yr (1.0
13.49
7.2
20.02
3.11
5.88
7.13
9.35
6.60
Overall Performance
Table 8-8 compares the main results for the 0.23 in., 2-yr, and 10-yr storms. No
CSOs occur for the 0.23 in. storm and all the wastewater flow is diverted into the
interceptor. For the 2- and 10-yr storms, all the regulators are releasing discharges into
the stream. Note how the occurrence of CSOs is reflected in the large increase in peak
discharge at the receiving stream outfall Ol. The peak discharge changes from 0.88 cfs
for the 0.23 in. storm to 20.02 cfs for the 2-yr storm and 45.65 cfs for the 10-yr storm.
Table 8-8 also shows that the maximum discharge conveyed by the interceptor barely
changes with the magnitude of the storm once all the regulators are discharging CSOs to
the stream. The maximum water depth in the wet well is practically the same for the 2-
and 10- year storm. This result clearly shows that flow regulators work in a way such that
all the flows above the diversion capacity are directly discharged into the water body.
n.) 10-yr (1.7 in.)
13.86
7.2
45.65
5.92
8.76
13.44
13.44
8.72
8.5 Summary
This example showed how to model a combined sewer system, composed of
combined sewer pipes, flow regulators, a pump station and a force main line, within
SWMM. The resulting model was used to determine the occurrence of Combined Sewer
Overflows (CSOs) under different size storm events. The key points illustrated in this
example were:
1. The main components of a combined sewer system model are pipes that carry both
dry weather sanitary and wet weather runoff flows, flow regulators that divide flow
between an interceptor pipe and a CSO outfall, and, if required, pump stations that
carry interceptor flows to a treatment facility through a force main.
2. Continuous wastewater flows, which can vary periodically by time of day, day of
week, and month of the year, are directly added into nodes associated with collector
sewers that service individual sewersheds.
3. Flow regulators can be represented using a combination of pipes, orifices and weirs.
The best way to model these regulators will depend on local conditions in the
combined sewer system under analysis.
4. A pump station can be modeled using a storage unit node to represent the wet well
that connects a pump link to the inlet node of a force main. The operation of the pump
152
-------
is defined through a pump curve and a set of wet well water levels that determine
when the pump starts up and shuts down.
A force main line can be defined by using a set of pipes between junction nodes that
are assigned a high surcharge depth so that flooding does not occur when the main
pressurizes.
By examining a number of design storms (or by running a continuous simulation as
described in the next example) one can determine the frequency at which combined
sewer overflows will occur. In this particular example, the 0.23 in. storm produced no
overflows while the 2-yr and 10-yr storms did. As shown from an analysis of the
rainfall record for this site made in Example 9, roughly 1 in 4 storms is larger than
0.23 in. and would thus have the capability to result in a CSO.
153
-------
Example 9. Continuous Simulation
This example shows how to run a continuous simulation with SWMM using a
long-term rainfall record. It will analyze the performance of the drainage system and
BMP detention pond for the 29 acre residential site designed in Example 3. The multi-
purpose pond was designed to detain a water quality capture volume (WQCV) and
control peak post-development release rates to their pre-development levels for the 2-yr,
10-yr and 100-yr design storms. This model will be re-run using a set of monthly average
evaporation rates and a continuous precipitation record, so that its behavior over a 10-
year period can be studied. The use of SWMM's Statistics tool for analyzing the results
of a continuous simulation will also be demonstrated. Because it only takes a few steps to
set up Example 3 for a continuous simulation, the bulk of this chapter will focus on
analyzing the results produced by the simulation.
Continuous simulation is important because it allows actual historic data to be
used to analyze the performance of drainage systems and their components. Drainage
systems are typically designed using synthetic design storms. These single event design
storms do not take into account varying patterns of rainfall duration and intensity,
variation of time between storms, changing antecedent soil and storage conditions within
the watershed, and the effect of evaporation. Continuous simulation considers all of these
factors and allows for a more accurate and robust comparison of the long-term water
balance and hydrologic performance of alternative stormwater management scenarios.
9.1 Problem Statement
Figure 9-1 shows the drainage system model developed in Example 3 for a new
residential development on a 29 acre site. The system includes a BMP detention pond
(SU1) that was designed to provide a water quality capture volume for the more
frequently occurring small storms and also provide peak runoff control for the 2-, 10- and
100-yr design storms of 2-hr duration. In this example the site will be analyzed using
continuous rainfall records from the city of Fort Collins, CO that were downloaded from
the National Climatic Data Center (NCDC) along with a set of monthly average
evaporation rates (in/day). The performance of the conveyance system and the detention
pond will be studied over the ten year period of 1968 to 1978. The results of the
simulation will be analyzed with regard to the following questions:
Was the flow target used to design the detention pond adequate?
How significant is evaporation within the overall system water balance?
How significant are antecedent conditions in affecting system behavior?
How effective is the detention pond in reducing peak discharges?
How effective is the pond in capturing the majority of runoff events within its
water quality control volume?
154
-------
What statistical properties of the rainfall record set it apart from the design storms
used previously?
RainGage
Figure 9-1. Drainage system and detention pond (SU1) designed in Example 3
9.2 System Representation
Continuous or long term simulation involves the simulation of multiple events
over a continuous period of months to years. In single-event simulation the rainfall record
is first analyzed separately to create an Intensity-Duration-Frequency (IDF) curve from
which design storms can be chosen; these design storms have a specific duration and
return period (e.g. the 2-hr 10-yr design storm) and are used within SWMM to generate a
single runoff hydrograph for design. In comparison, continuous simulation applies a long
term record of rainfall directly within SWMM to generate a long-term record of
simulated runoff; statistical analysis is then run on this generated record to characterize
the long-term performance and achieve a final design. It is advantageous to use
continuous simulation because it accounts for antecedent soil conditions and other initial
values of variables, such as the initial water level in storage units, which affect the
response of the drainage system to individual storm events. It also allows the
155
-------
representation of actual storm events of varying magnitudes, durations, and occurrence
intervals. The main disadvantages of continuous simulation are the additional
computation time required and the lack of high quality, long-term rainfall records for
many locations.
Rainfall
SWMM can utilize long-term rainfall data stored in external files. The program
recognizes several different file formats for these data: (1) the National Climatic Data
Center (NCDC) DSI-3240 format for hourly rainfall and the DSI-3260 format for 15 min
rainfall (both available from www.ncdc.noaa.gov/oa/ncdc.html), (2) HLY03 and HLY21
formats for hourly rainfall and the FIF21 format for 15 min rainfall at Canadian stations,
available from Environment Canada (EC) at www.climate.weatheroffice.ec.gc.ca, and (3)
a standard user-prepared format as described in the SWMM Users Manual. The quality
and quantity of the record will vary from station to station, but it is unusual to find long
precipitation records with no missing or incorrect data. It is very important to check the
quality of records before using them.
Evaporation
Single event simulations are usually insensitive to the evaporation rate. Thus,
evaporation is typically neglected when a single rainfall event or a synthetic storm is
simulated. However, this process is more significant when a continuous simulation is
performed because it is through evaporation that depression storage is recovered and
water levels in extended detention and wet ponds are reduced; thus it becomes an
important component of the overall water budget. Several options are available for
representing evaporation data in SWMM, including: (1) a single constant value, (2)
historical daily average values stored in an external file, (3) a time series when high
temporal resolution is available, and (4) monthly averages. Evaporation data are supplied
on the Evaporation page of SWMM's Climatology Editor.
Although conceptually evaporation should also affect the recovery of infiltration
capacity within the pervious areas of the watershed, SWMM's infiltration models do not
explicitly take it into account. Instead, they employ simple empirical functions for this
purpose. The Horton infiltration model, used in the examples throughout this manual,
employs an exponential function to restore infiltration capacity during dry periods. The
rate coefficient in this function is inversely proportional to the soil's drying time, i.e., the
number of days it takes a saturated soil to drain completely. A drying time of 7 days is
used for the site being analyzed, which is a typical value for the silt loam soil that is
assumed to cover the site (see Example 1).
156
-------
How to Load a Continuous Rainfall File into SWMM
A continuous precipitation record can be supplied to SWMM by using an
external file as the data source. What follows are the steps used to identify such a
file to SWMM.
1. Open the Property Editor of the Rain Gage that will use the rainfall data and
select FILE as the Data Source.
2. Click the ellipsis button in the File Name field. Navigate through your files and
select the file containing the continuous rainfall data.
3. If the file is in the User-Prepared format then the Rain Format and the Rain
Interval of the data in the file must be specified in the Property Editor. For data
from NCDC or EC files, SWMM automatically determines the appropriate rain
format and rain interval to use, and will override any values specified for these
properties in the editor.
4. Provide an entry for the Station ID field in the Property Editor. For NCDC and
EC files any identifier can be used - it need not be the station ID used in the
file. However for User-Prepared files it must correspond to an ID that appears
in the file since these files can contain data for more than one station.
5. If a User-Prepared rainfall file is being used then the correct Rain Units for the
rain depths in the file (inches or millimeters) must be specified. For NCDC and
EC data files this property is determined automatically.
6. Finally, specify the desired start and end analysis dates in the Simulation
Options Editor based on the rainfall record. These dates must fall within the
period of record. Time periods that fall outside those contained in the rainfall
file will have no rainfall associated with them.
The Rain Gage Property Editor for Example 9 is shown in the figure below.
Note that the entry in the File Name property will be different depending on where
you installed the rainfall file for this example on your computer's file system.
Property
Name
X-Coordinate
Y-Coordinate
Description
Tag
Rain Format
Rain Interval
Snow Catch Factor
Data Source
value |
RamGage
-148.485
1207.602
VOLUME
1 00
1.0
FILE
- Series Name
- File Name
- Station ID
Rain Units
T \pn:i|ects\HHSPro|ects\Eซamples SWMM5. EPA'Jenniler\4 - ,
053005
ilN
Units of rainfall data
157
-------
Statistical Tools
The large amount of output produced by a continuous simulation requires
statistical tools to analyze it in a concise and meaningful manner. SWMM provides an
interactive statistical query tool that can be applied to any output variable associated with
any specific object (subcatchment, node, or link) or the system as a whole. The tool
performs the following steps when analyzing the statistics for a specific output variable:
It first segregates the simulation period into a sequence of non-overlapping events
either by day, month, year or cluster of consecutive reporting periods. The occurrence
of these events is defined by the following minimum threshold values:
Analysis Variable Threshold - the minimum value of the variable under
analysis that must be exceeded for a time period to be included in an event.
Event Volume Threshold - the minimum flow volume (or rainfall volume) that
must be exceeded for a result to be counted as part of an event.
Separation Time - the minimum amount of time that must occur between the
end of one event and the start of the next event. Events with fewer hours are
combined together. This value applies only to events formed from consecutive
reporting periods, not to daily, monthly or annual event periods.
It then computes a user-specified event statistic for the analysis variable over all
reporting time periods that fall within each event period. This could be the event
mean value, the peak value, the total volume, etc. Thus each event is characterized by
a single value for whatever variable is being analyzed.
Summary statistics for the event values over the entire set of events are then
computed. These statistics include the maximum, minimum, mean, standard deviation
and skewness.
Finally a frequency analysis for the collection of event values is performed. The
events are rank ordered by value in a table that lists the event's date, its duration, its
event value, its cumulative exceedance frequency, and estimated return period. A
histogram of event values and a cumulative frequency plot of these values are also
produced.
As an example, a typical query might ask SWMM to segregate the output record
of flows discharged from a particular outfall into periods where the flow is above 0.05 cfs
and there are at least 6 hours between events, and compute the summary statistics and
frequency distribution of the peak flows within these events.
158
-------
9.3 Model Setup
Only three steps are required to convert the SWMM input file for Example 3
(ExampleS.inp) into one that can be used for continuous simulation. These are:
1. Add evaporation data:
In this example evaporation is supplied as monthly averages in units of in/day.
Table 9-1 shows the monthly averages for the city of Fort Collins that are used in this
example; the values were obtained from the Western Regional Climate Center
(www.wrcc.dri.edu/htmlfiles/westevap.final.html#colorado). The first row shows the
total monthly pan evaporation in inches. These amounts are adjusted in the second row
by multiplying them by 0.70 to more closely estimate the evaporation from naturally
existing surfaces, as proposed by Bedient and Huber (2002). Finally, the third row shows
the daily rates obtained by dividing the monthly totals by the number of days in each
month. Note that total values from December to February are 0.00. Typically this means
that the station does not measure pan evaporation during these months. The values in the
last row are entered on the Evaporation page of SWMM's Climatology Editor.
Evaporation from the surface of the site's detention pond should also be
accounted for. To do this, open up the Property Editor for the storage unit node SU1 and
enter a value of 1.0 for the Evap. Factor property.
Table 9-1.
i___f_^^
Monthly average pan Q QQ Q QQ ^ ^ 5M ^ ^ g QJ ^ ^ L4g Q QQ
evaporation (in)
Monthly average QQQ QQQ L?5 3 16 3 ?9 4 42 4 ง4 4 25 3 32 2 15 LQ4 QQQ
corrected (in)
Monthly average rate QQQ QQQ QQ6 QU QU Q 15 Q 16 Q u QU QQ1 Q Q3 QQQ
(in/day)
2. Specify a rainfall file:
The 30-year hourly rainfall record provided for Fort Collins is linked into the
model as explained in the sidebar "How to Load a Continuous Rainfall File into SWMM".
The name of this file is Record.dat. Because this file is from the NCDC and adheres to
the DSI-3240 format there is no need to alter the Rain Format, Rain Interval, or Rain
Units properties of the model's single rain gage. The Station ID property can be set to the
ID associated with the data's recording station which is 053005. The period of record for
these rain data extends from 1949 to 1979, but only the 10 year period from 1968 to 1978
will be simulated. Figure 9-2 shows the precipitation record for this period of time. There
are many factors to consider when selecting a sub-period of a long-term record, such as
the quality of the data it contains and how representative it is of both the overall record
and of current rainfall patterns.
159
-------
1968 1969 1970 1971 1972 1973 1974 1975 1976 1977
Year
Figure 9-2. Ten year rainfall record for Fort Collins, CO (Source: National Climatic Data Center)
3. Revise the simulation options:
Finally it is necessary to change some of the original simulation options so that
the correct period of rainfall record is simulated and a manageable amount of output is
generated. Using the Dates page of the Simulation Options dialog, both the simulation
start date and reporting start date are set to 01/01/1968. The simulation ending date is set
to 01/01/1978. Next the following changes are made on the Time Steps page of the
dialog: reporting step = 15 minutes, wet weather runoff step = 15 minutes, dry weather
runoff step = 6 hours, and routing step = 60 sec. Note that these time steps are longer than
those used in most of the previous examples. This is done to cut down on the amount of
time needed to run the simulation and on the amount of output results that are produced.
After making these changes the modified input file should be named
Example9.inp. Running this model will take about three to five minutes on a 2.41 GHz
computer. Most of this time is consumed by the flow routing computations. In general,
the run time will depend on the complexity of the watershed being modeled, the routing
method employed and the size of the routing time step used. The larger the time steps, the
faster the simulation but the less detailed the results. If accurate simulation of peak flows
on small watersheds is desired, then smaller time steps must be used. One way to
determine the proper time step is to simulate a single event at different time steps, then
choose the longest time step that produces the desired resolution in the hydrograph but
still produces a stable solution with acceptable continuity error.
160
-------
The run time for this example could be cut in half if Kinematic Wave routing were
used instead of Dynamic Wave routing. This is a viable alternative for long-term
simulations if one is willing to ignore possible backwater effects and pressurized flows,
and if the drainage network is not bifurcated. SWMM also has another feature that can
possibly reduce run times. This is the Skip Steady Periods option that appears on the
General page of the Simulation Options dialog. When this option is used the simulation
skips over periods of time when there is no runoff and no changes in flow. The criteria
used to determine when such periods exist are quite stringent, and for this example it
resulted in a negligible savings in run time.
9.4 Model Results
General Results
SWMM's Status Report summarizes overall results for the 10-yr simulation. The
runoff continuity error is -0.412 %, the flow routing continuity error is -1.557 %, and no
flooding occurred within the conveyance network (listed as Internal Outflow in the Flow
Routing Continuity section of the report). This is expected because the conveyance
system in Example 3 was designed for the 100-yr storm. The Runoff Quantity Continuity
table shows the significance of evaporation in the total water budget. Almost 20 inches of
water were evaporated from the depression storage surfaces of the catchment with
another 1.2 inches evaporated from the pond; this corresponds to 19.2 % of the total rain
over the 10 years (110.35 inches).
Detention Pond
The Status Report section "Node Inflow Summary" shows that the maximum flow
rate into the detention pond (SU1) was 31.7 cfs. The outflow from the pond never
exceeded 8.75 cfs (as determined by the maximum inflow for junction J out in the same
table). This control occurs over a period of 37 minutes, given that the maximum inflow to
the pond was at 12:02 AM of day 2781, while the maximum discharge was at 12395 AM
of the same day. Note that the maximum flow entering the pond is much lower than the
initial estimate for the 10-yr storm computed in Example 3, 62.1 cfs. Interestingly,
however, the maximum discharge released by the pond in the continuous simulation (8.48
cfs) is somewhat larger than 7.34 cfs, the 10-yr peak discharge target used in Example 3.
In other words, even though the peak flow entering the pond during the 10 year record is
lower than the design value, the peak outflow is actually larger than its original design
value.
To explain this result, consider Figure 9-3 which shows the rainfall hyetograph
for the storm event that produced the maximum inflow and outflow from the pond. The
volume of rain that falls in the first two hours is 1.83 in. This is larger than the 1.71 in
associated with the 10-yr 2-hr design storm used in Example 1, yet the inflow rate into
the pond is smaller (31.7 cfs versus 62.1 cfs). The main reason for this discrepancy is the
different rainfall interval used for the two storm events. The 10-yr design storm used a 5
161
-------
minute interval and its maximum intensity was 4.87 in/hr (see Figure 1-6 in Example 1);
in contrast, the rain interval for the storm from the continuous record is only 1 hr and its
maximum intensity was 1.47 in, 30 % of the maximum intensity of the design storm.
1.6
1.4 -
1.2 -
1 -
O.f
0.6 -
0.4 -
0.2 -
1.47
0.36
0.24
0.03
0.02
0.01
C
-------
Antecedent Conditions
One of the benefits of continuous simulation is that the model accounts for the
initial state of the catchment and its conveyance network at the beginning of each new
storm event. For instance, with continuous simulation the model simulates the initial
water depth in the detention pond at the beginning of a new rainfall event. A high water
depth will reduce the volume available for controlling the next storm, so that the
discharges released will be larger than for the case where the pond is empty at the start of
the storm.
Figure 9-4 shows the water depth in the pond over a period of 8 days, from day
3061 to day 3069 (from 5/19/1976 to 5/26/1976). Figure 9-5 shows the inflow and
outflow rates for the pond over the same period of time. Rainfall is also included in both
plots. Two storm events separated by 4 dry hours are shown: Event 1 has a maximum
intensity of 0.24 in/h and a total volume of 0.41 in.; Event 2 has a maximum intensity of
0.08 in/h and a total volume of 0.33 in. Despite the smaller volume and intensity of the
second storm, the water depth in the pond reaches 1.33 ft, which is larger than the water
depth reached with the first event, 1.17 ft. Because of the short dry time in between the
two events, the water level in the storage unit has only descended to a depth of 0.80 ft
when the next event stars. The storage available to control the second event is such that
the smaller peak inflow produced by the second event, 1.25 cfs is controlled to a peak
outflow of 0.43 cfs, while the larger peak inflow produced by the first event, 3.91 cfs, is
controlled to an outflow of only 0.39 cfs.
2
1.8
1.6
1.4
1.2
!
0.8
0.6
0.4
0.2
Rainfall
Water depth
ooooooooooooooooo
ooooooooooooooooo
o cs' o cs' o cs o cs o cs o cs o cs o cs o
Time (hh:mm)
Figure 9-4. Water depth in the detention pond between days 3061 and 3069
163
-------
t*2
u
3
4.5
4
3.5
3
2.5
2
1.5
1
0.5
ft
r T
-
-
-
_
-
-
i
1
i
-f
-
-
_
-
Rainfall
Discharge entering into Storage Unit
Discharge released by Storage Unit
^*\_
u
0.05
0.1
0.15 w
0.2 5
5'
0.25 g
0.3 ^
0.35^
0.4
0.45
ft e
o o
o o
ooooooooooooooo
ooooooooooooooo
Time (hh:mm)
Figure 9-5. Inflow and outflow (Q) for the detention pond between days 3061 and 3069
Evaporation
Finally, another advantage of continuous simulation is the inclusion of
evaporation in the overall system water balance. Evaporation from both the pervious and
impervious depression storage areas takes place between rain events. Thus, the amount of
depression storage available to capture the initial portion of the next storm depends on the
interval between storms. As an example, Figure 9-6 shows the rainfall and losses
(infiltration plus evaporation) simulated in subcatchment 57 over the period of time from
5/19/1976 to 5/26/1976. It is seen that after large events, and once infiltration has
stopped, the losses stabilize at 0.0021 in/h. These losses are caused by evaporation acting
over the water stored on the impervious area. To confirm this, consider that the average
evaporation rate in May is 0.12 in/day or 0.005 in/h; the percent imperviousness of
subcatchment SI is 56.8%, and 25% of that area does not have depression storage.
Therefore, evaporation acts over 0.568 (1-0.25) = 42.6% of the total area, and so the
loss rate over the entire subcatchment SI equals 42.6% 0.12 in/day = 0.0021 in/hr, the
same as the loss rate shown in Figure 9-6.
164
-------
0.25
0.225
0.2
0.175
0.15
0.125
0.1
0.075
0.05
0.025
0
Losses
Rainfall
0.0021 in/h
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
o
Time (hh:mm)
Figure 9-6. Losses for subcatchment SI between days 3061 and 3069
Detention Pond Outflows
To illustrate the use of SWMM's Statistics tool, a frequency analysis of the peak
outflows from the site's detention pond will be made. The analysis will show how often
and for what periods of time the pond has a certain discharge and how the magnitudes of
the peak discharges during these periods are distributed. To begin, one first opens the
Statistics Selection dialogue by clicking the Statistics icon (ฃ) on SWMM's main toolbar
or by selecting Report | Statistics from the main menu bar.
Figure 9-7 shows how the dialog should be filled in for this particular query. The
object to be analyzed is node J_out since it directly receives the outflow that exits the
pond from each of its outlet devices. The variable to be analyzed for this node is Total
Inflow. The time period used to define events is Event-Dependent, meaning that
separate events will be defined by consecutive reporting periods where certain event
threshold conditions are met. Within each such event period, the statistic to be analyzed is
the Peak value of total inflow to the node (which is equivalent to the peak pond
discharge through all of its outlet structures). And finally, the threshold criteria state that
a new event begins whenever an inflow of at least 0.005 cfs occurs at least 6 hours after
the last inflow of at least this amount was recorded. The selection of these values will
affect how many events are counted. In general, larger values of the minimum flow or
larger values of the time of separation will produce a fewer number of events.
165
-------
Statis
Object Category
Object Name
Variable Analyzed
Event Time Period
Statistic
Event Thresholds
Total Inflow
Event Volume
Separation Time
Node
J out
Total Inflow
Event-Dependent
Peak
v
vl
0.005
OK
Cancel
Help
Figure 9-7. Statistics Selection dialog for analyzing peak pond outflows
The report generated by running this statistical query contains four sections as
shown in Figures 9-8 through 9-11. The first section is the Summary Statistics, shown in
Figure 9-8. According to this summary, 295 events were identified in the record based on
the thresholds values for flow and inter-event time defined for the frequency analysis;
these events comprise 13.7 % of the total simulation time. Note the high positive
skewness coefficient (4.93) which implies a large number of low discharges and few
large ones. This is confirmed by the small value of the mean peak flow, 0.4 cfs.
Figure 9-9 shows a portion of the second section of the report, the Event Listing
for the variable under study. The events are listed in order of decreasing value of the
event statistic (the peak value) for the variable being analyzed (total inflow rate at node
J out). Five fields are included: Start date of the event, duration, value of the variable
under study (in this case peak flow), the exceedance frequency and an estimation of the
corresponding return period. Note the length of the events listed in this table; the majority
of them last longer than 24 hours. This shows that discharges are released quite slowly
from the pond, as expected based on its design criteria. Moreover, a similar frequency
analysis for duration instead of the peak discharge would show that the mean event
duration is 40.7 hours.
166
-------
Summar]i]| Events | Histogram | Frequency Plot
SUMMARY
STATISTICS
Obj ect Node J_out
Variable Total Inflow (CFS)
Event Period Variable
Event Statistic Peak (CFS)
Event Threshold Total Inflow > O.OOEO (CFS)
Event Threshold Event Volume > 0.0000 (ft3)
Event Threshold Separation Time >= 6.0 (hr)
Period of Record 01/01/1968 to 01/01/1378
Number of Events 295
Event Frequency* 0.137
Hinimum Value 0. DOS
Maximum Value 8.743
Mean Value 0.401
Std. Deviation 1.010
Skewness Coeff 4.933
*Fraction of all reporting periods belonging to an event.
Figure 9-8. Summary statistics for peak inflow to node J_out (same as peak pond outflow)
Summary I Events! Histogram Frequency Plot
Rank
Start Date
Event
Duration
(hours)
Event
Peak
(CFS)
Exceedance
Frequency
(percent)
Return
Period
(years)
A
I
1 90.8 8.743 0.34 11.00
2
3
4
5
E
7
S
9
10
11
12
07/23/1977
07/19/1970
05/23/1971
06/07/1974
07/27/1970
06/10/1970
05/05/1973
05/27/1975
04/19/1971
05/04/1969
06/05/1972
89.3
53.3
54.8
69.3
54.3
80.3
64.5
144.3
226.3
7.723
5.163
4.988
4.266
4.131
3.909
3.906
3.686
3.148
0.68
1.01
1.35
1.68
2.03
2.36
2.70
3.04
3.38
120.5 3.016 3.72
5.50
3.67
2.75
2.20
1.83
1.57
1.38
1.22
1.10
1.00
56.5 2.400 4.05 0.92
V
Figure 9-9. Event listing of peak inflow to node J_out (same as peak pond outflow)
167
-------
Summary II Events h Frequency Plot
NodeJ out Total Inflow
345
Event Peak Total Inflow (CFS)
Figure 9-10. Histogram of peak inflow to node J_out (same as peak pond outflow)
Summary | Events || Histogram |[
100
NodeJ out Total Inflow
456
Event Peak Total Inflow (CFS)
Figure 9-11. Cumulative frequency of peak inflow to node J_out (same as peak pond outflow)
The exceedance frequency and return period displayed in the Event Listing are
both computed using the Weibull formula for plotting position. Therefore, for a specific
event the exceedance frequency F and the return period in years T are calculated using
the following equations:
m
nR+\
(9-1)
168
-------
r = (9-2)
m
where m is the event's rank, HR is the total number of events and n is the number of years
under analysis. For example, for the 4th event occurring on 05/23/1971, the exceedance
frequency is equal to ฅ4 = 4 / (295 + 1) = 0.0135= 1.35 %, and the return period is equal
to T4 = (10 + l)/4 = 2.75 years.
The event listing also shows that during the 10 years of simulation there were 2
events where the peak discharge was greater than the 7.34 cfs value that was computed in
Example 3 when the model was run with the 10-yr design storm. This small discrepancy
likely reflects the fact that the 10 years of record used here is only one of 29 different
consecutive 10-year sequences that could have been selected from the 30 years worth of
rainfall data available.
The third section of the Statistics report contains a histogram of the event statistic
being analyzed, as shown in Figure 9-10. For this particular example it shows what
fraction of all the events had a peak flow of a given size. Note how the figure confirms
what was said earlier, that the distribution of peak inflows is highly skewed towards the
low end of the flow scale. Finally, the fourth section of the report, as shown in Figure 9-
11, presents a cumulative frequency plot of the event statistic under study. In this
particular example, only 10 % of the peak discharges over the 10-yr period exceed 0.5
cfs, and only 6 % exceed 2 cfs. Note that both the histogram and cumulative frequency
plots are just graphical representations of the same information as provided in the event
listing.
Detention Pond Water Depth
A second application of the SWMM Statistics tool to this example will analyze
the maximum depth in the detention pond. This will help verify if the Water Quality
Capture Volume (WQCV) of the pond was effective in capturing the more frequently
occurring storms. Recall from Example 3 that the first 1.5 feet of storage volume was
designated for this purpose. The statistical query used to answer this question is shown in
Figure 9-12. Note that this time events are defined to consist of all days where the water
depth in the pond was at least 0.05 ft deep. The "Daily" option is chosen for the Event
Period because the drawdown time for the WQCV is 40 hours; thus a daily analysis of
the depth in this portion of the pond is sufficient. Figure 9-13 shows the frequency plot
that results from this query. On only 7 % of the days when the pond is wet does its depth
exceed the WQCV. Thus, one can conclude that the large majority of all storms are
captured within the WQCV and the pond will function as an effective BMP control.
169
-------
<: Statistics Selection
Object Category
Object Name
Variable Analyzed
Event Time Period
Statistic
Event Thresholds
Depth
olume
TI Tim*
OK 1 I
Node
SLI1
Depth
Daily
Peak
0.05
Cancel
Help
Figure 9-12. Statistical query for daily peak water depth in the detention pond
0.5
1 1.5 2 2.5
Daily Peak Depth (ft)
3.5
Figure 9-13. Frequency plot of daily peak water depth in the detention pond
170
-------
Rainfall Statistics
Another variable typically analyzed using statistics is rainfall. Several quantities
that characterize storm events for the rainfall record used in this example will be
examined and compared. These are the event duration, mean intensity, total volume, and
peak intensity. As before, the record will be separated into a sequence of independent
storm events of varying magnitude and duration using SWMM's Statistic tool. Note that
a rainfall analysis can be made without having to run a complete runoff/routing
simulation for the entire SWMM model. See the sidebar "Analyzing Stand-Alone Rainfall
Files with SWMM" for details on how to do this.
Analyzing Stand-Alone Rainfall Files with SWMM
Modelers often have a need to statistically analyze a long-term rainfall
record independently of using it within a particular SWMM model. For example,
one might want to identify a sub-period of the record that is representative of the
entire record or identify particular events, extreme or otherwise, that could be used
for single event analyses. It is quite simple to set up a general SWMM project for
just this purpose and run it independently for any rainfall file. The steps involved
are as follows:
1. Create a new project with SWMM that consists of just a single subcatchment
and associated rain gage and set the outlet of the subcatchment to itself.
2. Save the project with a meaningful name, such as RainStats.inp.
3. Whenever a rainfall file needs to be analyzed, launch SWMM and open this
special project. (Note that a new SWMM session can be launched while
other SWMM sessions are still active).
4. Edit the Rain Gage to use the file of interest (remembering to enter
appropriate values for the Rain Format, Rain Interval, Station ID, and Rain
Units if a file with the User-Supplied format is being used).
5. Edit the Dates in the Simulation Options dialog to cover the period of
interest within the rainfall file. Also set the Reporting and Wet Runoff time
steps to be the same as the Rain Interval used for the rainfall data.
6. Run a simulation and then use the Time Series Graph, Tabular Report, and
Statistics tools to analyze the System Rainfall variable.
Regarding the final step of this process, time series plots are useful for
providing an overall picture of how rainfall varies within a period of several years
or months, tabular reports can retrieve rainfall hyetographs for specific events that
can be pasted directly into a SWMM Time Series created in another SWMM
session, and examples of using the Statistics tool to analyze the rainfall record are
shown elsewhere in this chapter.
171
-------
In rainfall analysis, the separation time used to decide when one event ends and
another begins is referred to as the Minimum Inter-Event Time (MIT). This is the smallest
number of consecutive dry periods that must occur before the next wet period is
considered as a separate event. There is no established "correct" value for the MIT,
although 3 to 30 hours are often used for rainfall data (Hydroscience, 1979). See Adams
and Papa (2000) for a detailed discussion on this subject. When storm events are
characterized as a Poisson process, the time between events follows an exponential
distribution for which the mean equals the standard deviation (i.e., the coefficient of
variation (CV) is 1). Thus one suggested approach to choosing a MIT is to find a value
that produces a CVof 1 for the resulting collection of inter-event times.
Figure 9-14 shows the statistical query used to test how well a MIT (separation
time) of 12 hours produces a sequence of events whose inter-event times have a CV of 1
for the rainfall record used in this chapter. The resulting CV (standard deviation divided
by the mean) is 211.33 / 195.28 = 1.08, indicating that 12 hours is a reasonable MIT to
use for this rainfall record.
Object Category
Object Name
Variable Analyzed
Event Time Period
Statistic
Event Thresholds
Rainfall
Event Volume
Separation Time
System
Statistics System Rainfall
[Summary]) Events | Histogram | Frequency Plot |
Rainfall
Event-Dependent
Inter-Event Time
OK
Cancel
Help
SUHHAEY
STATISTICS
Object .............. System
Variable ............. Eainfall (in/hr)
Event Period ......... Variable
Event Statistic ...... Inter-Event Time (hours)
Event Threshold ...... Eainfall > 0.00 (in/hr)
Event Threshold ...... Event Volume > 0.00 (in)
Event Threshold ...... Separation Time >= 1Z.O (hr)
Period of Record ..... 01/01/1963 to 01/01/1978
Number of Events ..... 446
Event Frequency* ...... 0.043
Minimum Value ........ 14. 000
Haxirnum Value ........ 1771 . ฃ00
He an Value ........... 195.288
Std. Deviation ....... 211.333
Skewness Coeff ....... 2.337
*Fraction of all reporting periods belonging to an event.
Figure 9-14. Selection of an MIT for analyzing a rainfall record
Next a frequency analysis is made for each of the following rainfall quantities:
duration, mean intensity, total volume, and peak intensity. Each analysis uses a Statistics
Selection dialog that looks the same as that in Figure 9-14, except that a different choice
of event statistic is used for each. Table 10-2 lists the summary statistics found for each
frequency analysis. The first two properties (number of events and event frequency, or
percentage of total time in which rainfall is registered) are the same for each rainfall
property since all of the frequency analyses used the same thresholds to define an event.
172
-------
Table 9-2. Summary statistics for various rainfall event properties
Property
Number of Events
Event Frequency
Minimum Value
Maximum Value
Mean Value
Std. Deviation
Skewness Coeff.
Duration
(h)
446
0.043
1.000
61.000
8.466
10.557
2.278
Event
Mean Intensity
(in/hr)
446
0.043
0.006
0.6900
0.042
0.058
5.642
statistic
Total Volume
(in)
446
0.043
0.01
4.44
0.247
0.445
4.733
Peak Intensity
(in/hr)
446
0.043
0.006
1.470
0.078
0.125
5.550
Figure 9-15 shows the frequency plot produced by this analysis for event duration
and total rainfall. (This figure was generated by using SWMM's Edit | Copy To menu
option to copy the data associated with each frequency plot to the Windows Clipboard
from which it was pasted into a spreadsheet program and combined together on a single
graph.) This plot can also be used as a less direct indicator than the pond depth frequency
analysis performed earlier to see if the Water Quality Capture Volume (WQCV) of 0.23
inches is sufficient to capture the majority of runoff events.
0
0.5
Event total rainfall (in)
1.5 2 2.5 3 3.5
4.5
Total rainfall _
Duration
100
90
80
70 td
n
n
60 ฃ.
50 |
n
40 "c
n
n
30 ^
20
10
0
16
24 32 40 48 56
Event total duration (hr)
Figure 9-15. Frequency plots for event duration and depth
64
72
80
173
-------
It is seen (blue line) that about 70% of the rainfall events have a total volume
smaller than 0.23 inches. Even if the watershed were totally impervious with no surface
retention capacity, the pond would control around 70% of the events according to the
model. In reality there are infiltration and storage losses so that a larger percentage of the
rainfall events are controlled by the pond's volume. Finally, note also that long storms are
not very frequent, with 50% of them being shorter than 5 hours, and 80% shorter than 14
hours. Storms longer than one day are very uncommon at this recording station located in
the Colorado foothills.
Another use of this type of multi-variate statistical analysis is to study in more
detail the correspondence between the frequencies of the events based on different
rainfall characteristics. The largest event is not necessarily the longest one or the most
intense. It is useful to determine the degree of dependence among storm event
characteristics in order to see if these characteristics are correlated or not, and to identify
a subset of the most critical events that could be used for design purposes, depending on
the objective of the analysis. The intensity of the events is significant in determining peak
discharge, but the magnitude may be more important when storage control structures are
designed or evaluated.
Table 9-3 shows the ten most extreme events according to the four characteristics
under analysis: duration, mean intensity, volume and peak intensity. The table reveals
that the peak and mean intensities are closely correlated; seven of the ten events with the
largest peak intensity are also in the groups of events with a large mean intensity (dates of
these common events are underlined). There is also a correspondence between duration
and total volume; four of the longest events are also in the group that contains the largest
ones (the dates of these common events are in blue). A weaker correlation is observed
between the mean intensity and volume (one common event in a red box), and between
volume and peak intensity (two common events in green).
Table 9-3. Ten most severe events based on duration, depth and intensity
Rank
1
2
3
4
5
6
7
8
9
10
Tr
(year)
11
5.5
3.67
2.75
2.2
1.83
1.57
1.38
1.22
1
Duration
Start Date
5/5/1969
10/23/1975
10/10/1969
10/3/1969
3/29/1970
8/2/1976
5/27/1975
5/20/1975
3/9/1968
6/7/1968
(hr)
Value
61
56
51
51
50
49
49
48
46
46
Mean (in/hr)
Start Date
7/27/1970
7/20/1973
7/16/1969
7/19/1970
5/23/1971
7/28/1974
8/8/1970
5/8/1971
17/24/1977
8/14/1975
Value
0.69
0.39
0.38
0.36
0.36
0.33
0.27
0.22
0.20
0.19
Volume
Start Date
|7/24/1977
5/27/1975
5/5/1969
6/10/1970
6/7/1974
8/12/1975
10/3/1969
10/10/1969
4/24/1971
4/24/1973
(in)
Value
4.44
3.29
2.98
2.78
2.75
2.25
1.97
1.66
1.62
1.57
Peak (in/hr)
Start Date
8/12/1975
7/19/1970
5/23/1971
7/27/1970
7/24/1977
8/14/1975
6/17/1975
7/28/1974
9/11/1973
7/20/1973
Value
1.47
0.97
0.85
0.69
0.65
0.55
0.52
0.44
0.40
0.39
174
-------
Table 9-4 shows the ranks of the common events identified in Table 9-3. Numbers
before and after the "&" symbol indicate the rank according to the variables defined in
the first row of the table. For example, from the second column, the longest event (61
hrs) is also the third largest event (2.98 inches), the third longest event (51 hrs) is the
eighth largest event (1.66 inches), and so on. Note that the two common events identified
among the most severe in terms of the volume and peak intensity are the same two events
associated with the largest peak discharges released by the storage unit discussed earlier;
these are the events of 07/47/1977 and 8/12/1975. Finally, there are no common events
among the ten most severe in terms of both duration and intensities (both mean and
peak). This analysis corresponds to a preliminary step in evaluating the correlation
between storm event characteristics. Other methodologies including the use of scatter
plots between variables and correlation coefficients can be used. These methods are
applied to all the events and not only the most severe ones.
Tabtej)^^
Duration &
Mean
-
-
-
-
-
-
-
Duration & Duration & Mean &
Volume Peak Volume
1&3 - 9&1
3&8
4&7
7&2
.
.
.
Mean &
Peak
1&4
2&10
4&2
5&3
6&8
9&5
10 & 6
Volume &
Peak
1&5
6&1
-
-
-
-
-
9.5 Summary
This example demonstrated how to use continuous simulation along with
statistical analysis to evaluate the long-term behavior of a drainage system. Ten years of
continuous rainfall data (1-hr resolution) and monthly average evaporation rates were
used to analyze the performance of the conveyance system and detention pond designed
in Example 3. The key points illustrated in this example were:
1. Continuous simulation allows modelers to more faithfully represent the behavior of
drainage systems because it subjects them to a long sequence of actual rainfall events
of varying magnitudes and durations and also accounts for the variability of
antecedent conditions that exist from one event to the next.
2. Evaporation is an important component of the long-term water budget that should be
considered in continuous simulation. In this example it accounted for 19 % of the
total rainfall input to the catchment.
3. Model setup for continuous simulation is quite straightforward in SWMM, providing
that reliable long-term rainfall records and evaporation data are available.
175
-------
High time resolution on rainfall data is required if accurate predictions of peak runoff
flows are needed. The hourly time interval of the rainfall data used in this example
may cause peak discharges to be underestimated.
SWMM's Statistics tool is a valuable aid in interpreting the large amount of output
data that can be generated from a long-term continuous simulation. It was used here
to determine how effective the detention pond was in reducing peak discharges and in
capturing the majority of runoff events within its water quality control volume. The
tool was also used to characterize the properties of the most severe rainfall events
occurring over the 10-year simulation period.
176
-------
References
Adams, BJ. and Papa, F. (2000). Urban Stormwater Management Planning with
Analytical Probabilistic Models., John Wiley & Sons, Inc, New York, NY.
Akan, A.O. and Houghtalen, RJ. (2003). Urban Hydrology, Hydraulics, and Stormwater
Quality, John Wiley & Sons, Inc, Hoboken, NJ.
American Society of Civil Engineering (ASCE) (1992). Design and Construction of
Urban Stormwater Management Systems., ASCE Manuals and Reports on
Engineering Practice No. 77 and Water Pollut. Control Fed. Manual of Practice
RD-20, New York, NY.
Bedient, P.B. and Huber, W.C. (2002). Hydrology and Floodplain Analysis, 3rd Ed.,
Prentice-Hall, Upper Saddle River, NJ.
City of Fort Collins (1984). Storm Drainage Design Criteria and Construction
Standards, Utilities Department, Stormwater Division, Fort Collins, CO.
(http://www.fcgov.com/stormwater/pdf/sw-manual07.pdf)
City of Fort Collins (1997). Memorandum: Update to the Stormwater Drainage Design
Criteria. Memorandum to Storm Drainage Design Criteria Users. April 29, 1997,
Utilities Department, Stormwater Division, Fort Collins, CO.
City of Fort Collins (1999). Memorandum: New Rainfall Criteria. Memorandum to Storm
Drainage Design Criteria Users. April 12, 1999, Utilities Department,
Stormwater Division, Fort Collins, CO.
Clark County Regional Flood Control District (CCRFCD) (1999). Hydrological Criteria
and Drainage Design Manual, Las Vegas, NA.
(http://acequia.ccrfcd.org/pdf_archl/hcddm/hcddm.pdf)
Douglas County (2008). Storm Drainage Design and Technical Criteria Manual,
Douglas County, Department of Public Works-Engineering Division, Castle
Rock, CO.
(http://www.douglas.co.us/publicworks/engineering/Storm_Drainage_Design_and
Technical Criteria Manual .html)
Driscoll, E.D., Palhegyi, G.E., Strecker, E.W. and Shelly, P.E. (1989). Analysis of Storm
Events Characteristics for Selected Rainfall Gauges throughout the United States,
U.S. Environmental Protection Agency. Washington, DC.
Field, R. and Tafuri, A.N. (1973). "Storm Pollution Control in U.S". Combined Sewer
Overflow Seminar Papers. EPA 670/2-73-077, U. S. Environmental Protection
Agency, Edison, NJ.
(http://www.epa.gov/ednnrmrl/publications/conferencepapers/epa670273077)
Grigg, N.S. (1996). Water Resources Management: Principles, Regulations, and Cases,
McGraw-Hill, New York, NY.
177
-------
Guo, J.C.Y. (2001). "Design of Infiltration Basins for Stormwater." Stormwater
Collection Systems Design Handbook. Edited by L. Mays, McGraw-Hill, New
York, NY.
Guo, J.C.Y. and Urbonas, B. (1995). Special Report to the Urban Drainage and Flood
Control District on Stormwater BMP Capture Volume Probabilities in United
States. Denver, CO.
Guo, J.C.Y. and Urbonas, B. (1996). "Maximized Detention Volume Determined by
Runoff Capture Ratio." Journal of Water Resources Planning and Management,
122(l):33-39.
Guo, J.C.Y. and Urbonas, B. (2002). "Runoff Capture and Delivery Curves for Storm-
water Quality Control Design." Journal of Water Resources Planning and
Management, 128(3):208-215.
Hydroscience, Inc. (1979). A Statistical Method for the Assessment of Urban Stormwater.
EPA-440/3-79-023, U.S. Environmental Protection Agency, Cincinnati, OH.
Huber, W., Clannon, L. and Stouder, M. (2006). BMP Modeling Concepts and
Simulation. EPA/600/R-06/033, U.S. Environmental Protection Agency (EPA).
Cincinnati, OH.
Manning, M.J., Sullivan, R.H. and Kipp T.M. (1977). Nationwide Evaluation of
Combined Sewer Overflows and Urban Stormwater Discharges - Vol. Ill:
Characterization of Discharges. EPA-600/2-77-064c (NTIS PB-272107), U.S.
Environmental Protection Agency, Cincinnati, OH.
Metcalf and Eddy, Inc. (1991). Wastewater Engineering: Treatment, Disposal, Reuse.,
McGraw-Hill, Inc., New York, NY.
Nicklow, J.W., Boulos, P.P. and Muleta, M.K. (2004). Comprehensive Sewer Collection
System Analysis Handbook for Engineers and Planners, MWH Soft, Inc.
Pasadena, CA.
Rossman, L. (2008). Storm Water Management Model User's Manual Version 5.0.
EPA/600/R-05/040, U.S. Environmental Protection Agency, National Risk
Management Research Laboratory, Cincinnati, OH.
Sansalone, J.J. and Hird, J. (2003). "Treatment of Stormwater Runoff from Urban
Pavement and Roadways." Wet-Weather Flow in the Urban Watershed,
Technology and Management. Edited by R. Field and D. Sullivan. CRC Press
LLC, Boca Raton, FL.
Sartor, J.D. and Boyd, G.B. (1972). Water Pollution Aspects of Street Surface
Contaminants. EPA-R2-72-081 (NTIS PB-214408), U.S. Environmental
Protection Agency, Washington, DC.
Urban Drainage and Flood Control District (UDFCD) (2001). Urban Storm Drainage
Criteria Manual, 2007 revision. Denver, CO.
(http: //www .udfcd. org/downl oads/down_critmanual. htm).
178
-------
U.S. Environmental Protection Agency (1983). Results of the Nationwide Urban Runoff
Program, Volume I Final Report. NTIS PB84-185552, U.S. Environmental
Protection Agency, Washington, DC.
U.S. Environmental Protection Agency (1986). Methodology for Analysis of Detention
Basins for Control of Urban Runoff Quality. EPA 440/5-87-001, Office of Water,
Washington, DC.
U.S. Environmental Protection Agency (1999). Storm Water Technology Fact Sheet
Infiltration Trench. EPA 832-F-99-019, U.S. Environmental Protection Agency.
Washington, DC.
(http://www.epa.gov/owm/mtb/infltrenc.pdf).
Vanoni, V.A. (1975). Sedimentation Engineering., ASCE Manual and Report on
Engineering Practice No. 54, New York, NY.
Water Environment Federation (1998). Urban Runoff Quality Management, WEF
Manual of Practice No. 23., ASCE Manual and Reports on Engineering Practice
No. 87, Water Environmental Federation, Alexandria, VA.
179
------- |