Federal technical guidance for implementing stormwater-runoff requirements

Technical Guidance on Implementing the
Stormwater Runoff Requirements for
Federal Projects under Section 438 of the
Energy Independence and Security Act

United States Office of Water (4503T) EPA 841-B-09-001
Environmental Washington, DC 20460 December 2009
Protection Agency www.epa.gov/owow/nps/lid/section438

Foreword
Stormwater runoff in urban and developing areas is one of the leading sources of water pollution in the United States. In recognition of this issue, Congress enacted Section 438 ofthe Energy Independence and Security Act of 2007 (EISA) to require federal agencies to reduce stormwater runoff from federal development projects to protect water resources. More recently, the President signed Executive Order 13514 on "Federal Leadership in Environmental, Energy, and Economic Performance" calling upon all federal agencies to "lead by example" to address a wide range of environmental issues, including stormwater runoff. The Executive Order required the U.S. Environmental Protection Agency (EPA), in coordination with other federal agencies, to publish this Technical Guidance. EPA worked closely with many federal agencies to develop this Technical Guidance to help federal agencies in implementing EISA Section 438. The guidance provides a step-by-step framework that will help federal agencies maintain pre-development site hydrology by retaining rainfall on-site through infiltration, evaporation/transpiration, and re-use to the same extent as occurred prior to development. The Technical Guidance provides background information, key definitions, case studies, and guidance on meeting the new requirements. Federal agencies can comply with Section 438 by using a variety of stormwater management practices often referred to as "green infrastructure" or "low impact development" practices, including, for example, reducing impervious surfaces, using vegetative practices, porous pavements, cisterns and green roofs.
One of the most exciting new trends in water quality management today is the movement by many cities, counties, states, and private sector developers toward the increased use of this next generation stormwater-management practices to help protect and restore water quality. Many federal agencies, including EPA, are already using a full spectrum of stormwater management practices to reduce the impact of federal facilities on local watersheds. These projects have produced results such as reductions in site runoff volumes and increased stormwater quality, which ultimately lead to more sustainable facilities.
EPA enjoyed the opportunity to work with a number of federal agencies to develop this state-of-the art, technical guidance and appreciate all their input. We look forward to continuing the dialogue as we all work to implement this guidance.
assistant Administrator office of Water
U.S. Environmental Protection Agency
Section 438 Technical Guidance December 2009

Technical Guidance on Implementing the Stormwater Runoff Requirements for Federal
Projects under Section 438 of the Energy Independence and Security Act

INTRODUCTION

In December 2007, Congress enacted the Energy Independence and Security Act of 2007.
Section 438 of that legislation establishes strict stormwater runoff requirements for federal
development and redevelopment projects. The provision reads as follows:

“Storm water runoff requirements for federal development projects. The
sponsor of any development or redevelopment project involving a Federal facility
with a footprint that exceeds 5,000 square feet shall use site planning, design,
construction, and maintenance strategies for the property to maintain or restore, to
the maximum extent technically feasible, the predevelopment hydrology of the
property with regard to the temperature, rate, volume, and duration of flow.”

The intent of Section 438 of the Energy Independence and Security Act of 2007 (EISA) is to
require federal agencies to develop and redevelop applicable facilities in a manner that maintains
or restores stormwater runoff to the maximum extent technically feasible. Until recently,
stormwater programs established to address water quality objectives have been designed to
control traditional pollutants that are commonly associated with municipal and industrial
discharges, e.g., nutrients, sediment, and metals. Increases in runoff volume and peak discharge
rates have been regulated through state and local flood control programs. Although these
programs have merit, knowledge accumulated during the past 20 years has led stormwater
experts to the conclusion that conventional approaches to control runoff are not fully adequate to
protect the nation’s water resources (National Research Council, 2008).

Implementation of Section 438 of the EISA can be achieved through the use of the green
infrastructure/low impact development (GI/LID) infrastructure tools described in this guidance.
The intention of the statute is to maintain or restore the pre-development site hydrology during
the development or redevelopment process. To be more specific, this requirement is intended to
ensure that receiving waters are not negatively impacted by changes in runoff temperature,
volumes, durations and rates resulting from federal projects. It should also be noted that a
performance-based approach was selected in lieu of a prescriptive requirement in order to
provide site designers maximum flexibility in selecting control practices appropriate for the site.

Section 14 of the Executive Order 13514, Federal Leadership in Environmental, Energy,
and Economic Performance

On October 5, 2009, President Barack Obama signed Executive Order 13514,
“Federal Leadership in Environmental, Energy, and Economic Performance.”
Section 14 of the Executive Order provides:

Stormwater Guidance for Federal Facilities. Within 60 days
of the date of this order, the Environmental Protection
Agency, in coordination with other Federal agencies as
1
Section 438 Technical Guidance December 2009

appropriate, shall issue guidance on the implementation of
section 438 of the Energy Independence and Security Act of
2007 (42 U.S.C. 17094).

This provision contains two significant elements. First, for the first time, EPA is formally
assigned the responsibility to write and issue the Section 438 guidance, in coordination with
other federal agencies. Second, it establishes a deadline for EPA to do so by December 5, 2009.

Purpose and Organization of this Guidance

The purpose of this document is to provide technical guidance and background information to
assist federal agencies in implementing EISA Section 438. Each agency or department is
responsible for ensuring compliance with EISA Section 438. The document contains guidance on
how compliance with Section 438 can be achieved, measured and evaluated. In addition,
information detailing the rationale for the stormwater management approach contained herein
has been included.

This document is intended solely as guidance. This document is not a regulation nor does it
substitute for statutory provisions or regulations. This guidance does not impose any legally
binding requirements on federal agencies and does not confer any legal rights or impose legal
obligations upon any member of the public. This document does not create a cause of action
against the EPA, other federal agencies, or the United States.

The following information is presented within this document:

Part I: Implementation Framework
A. Background
B. Benefits and outcomes of the new stormwater performance requirements
C. Applicability and definitions
D. Tools to implement the requirements of Section 438
E. Calculating the 95th percentile rainfall event

Part II: Case Studies on Capturing the 95th Percentile Storm Using Onsite Management
Practices
Case studies representing typical federal installations have been included. The case studies were
selected to demonstrate the feasibility of providing adequate stormwater control for a range of
site conditions and building designs. To the maximum extent technically feasible, each case
study includes a description of a method that can be used to determine the design objectives of
the project based on retaining the 95th percentile storm. Examples of onsite technologies and
practices have also been provided. The case studies are intended to provide examples of
modeling procedures that can be used to quantify treatment system performance and processes
for assessing sites and determining appropriate control techniques to the maximum extent
technically feasible.
2


Section 438 Technical Guidance December 2009

3

Part I: Implementation Framework

A. BACKGROUND

This section contains background on the causes and consequences of stormwater discharges,
solutions that can be used to address the causes and consequences of stormwater discharges and
how to implement those solutions to comply with Section 438 of EISA.

Alterations to Natural Hydrology and the Impact on Stormwater Runoff
In the natural, undisturbed environment rain that falls is quickly absorbed by trees, other
vegetation, and the ground. Most rainfall that is not intercepted by leaves infiltrates into the
ground or is returned to the atmosphere by the process of evapotranspiration. Very little rainfall
becomes stormwater runoff in permeable soil, and runoff generally only occurs with larger
precipitation events. Traditional development practices cover large areas of the ground with
impervious surfaces such as roads, driveways, sidewalks, and buildings. Under developed
conditions runoff occurs even during small precipitation events that would normally be absorbed
by the soil and vegetation. The collective force of the increased runoff scours streambeds, erodes
stream banks, and causes large quantities of sediment and other entrained pollutants to enter the
water body each time it rains (Shaver, et al., 2007; Booth testimony, 2008).

As watersheds are developed and impervious surfaces increase in area, the hydrology of the
watersheds fundamentally changes over time which results in degraded aquatic ecosystems. In
recognition of these problems, stormwater managers employed extended detention approaches to
mitigate the impacts of increased peak runoff rates. However, wet ponds and similar practices
are not fully adequate to protect downstream hydrology because of the following inherent
limitations of these conventional practices (National Research Council, 2008; Shaver, et al.,
2007):

 Poor peak control for small, frequently-occurring storms;
 Negligible volume reduction; and
 Increased duration of peak flow.

Detention storage targets relatively large, infrequent storms, such as the two and 10-year/24-hour
storms for peak flow rate control. As a result of this design limitation, flow rates from smaller,
frequently-occurring storms typically exceed those that existed onsite before land development
occurred and these increases in runoff volumes and velocities typically result in flows erosive to
stream channel stability (Shaver, et al., 2007). Section 438 is intended to address the
inadequacies of the historical detention approach to managing stormwater and promote more
sustainable practices that have been selected to maintain or restore predevelopment site
hydrology.

A 2008 National Research Council report on urban stormwater confirmed that current
stormwater control efforts are not fully adequate. Three of the report’s findings on stormwater
management approaches are particularly relevant (National Research Council, 2008).

Section 438 Technical Guidance December 2009

4
1. Individual controls on stormwater discharges are inadequate as the sole solution to
stormwater in urban watersheds;
2. Stormwater control measures such as product substitution, better site design,
downspout disconnection, conservation of natural areas, and watershed and land-use
planning can dramatically reduce the volume of runoff and pollutant load from new
development; and
3. Stormwater control measures that harvest, infiltrate, and evapotranspire stormwater are
critical to reducing the volume and pollutant loading of small storms.


Pre-development Hydrology. Courtesy of C. May,
University of Washington. Post-Development Hydrology. Courtesy of C. May, University of Washington.



Figure 1. Pre-Development and Post-Development Hydrology. (USDA).

Figure 1 contains two sets of diagrams depicting the water balances at undeveloped and
developed sites. Runoff patterns will vary based on factors such as geographic location, local
meteorological conditions, vegetative cover and soils. The first set of figures represents
conditions in the Pacific Northwest where storms have a long duration and low intensity, i.e., the
volume of rain in an individual storm is small. The second set of figures from the U.S.
Department of Agriculture represents a more generalized set of conditions, but was included to
illustrate that heavily urbanized areas typically cause large increases in runoff.
Section 438 Technical Guidance December 2009

Land cover changes that result from site development include increased imperviousness, soil
compaction, loss of vegetation, and loss of natural drainage patterns, which result in increased
runoff volumes and peak runoff rates. The cumulative impacts of the land cover changes result
in alterations of the natural hydrology of a site, which disrupts the natural water balance and
changes water flow paths. The consequences of these impacts include:

1. Increased volume of runoff. With decreased area for infiltration and evapotranspiration
due to development, a greater amount of rainfall is converted to overland runoff which
results in larger stormwater discharges.
2. Increased peak flow of runoff. Increased impervious surface area and higher connectivity
of impervious surfaces and stormwater conveyance systems increase the flow rate of
stormwater discharges and increase the energy and velocity of discharges into the stream
channel.
3. Increased duration of discharge. Detention systems generate greater flow volumes and
rates. These prolonged higher discharge rates can undermine the stability of the stream
channel and induce erosion, channel incision and bank cutting.
4. Increased pollutant loadings. Impervious areas are a collection site for pollutants. When
rainfall occurs these pollutants are mobilized and transported directly to stormwater
conveyances and receiving streams via these impervious surfaces.
5. Increased temperature of runoff. Impervious surfaces absorb and store heat and transfer it
to stormwater runoff. Higher runoff temperatures may have deleterious effects on
receiving streams. Detention basins magnify this problem by trapping and discharging
runoff that is heated by solar radiation (Galli, 1991; Schueler and Helfrich, 1988).

The resulting increases in volume, peak flow, and duration are illustrated in the hydrograph in
Figure 2, which is a representation of a site’s stormwater discharge with respect to time. The
hydrograph illustrates the impacts of development on runoff volume and timing of the runoff.
Individual points on the curve represent the rate of stormwater discharge at a given time. The
graph illustrates that development and corresponding changes in land cover result in greater
discharge rates, greater volumes, and shorter discharge periods. In a natural condition, runoff
rates are slower than those on developed sites and the discharges occur over a longer time period.
The predevelopment peak discharge rate is also much lower than the post-development peak
discharge rate due to attenuation and absorption by soils and vegetation. In the post-
development condition there is generally a much shorter time before runoff begins because of
increased impervious surface area, a higher degree of connectivity of these areas and the loss of
soils and vegetative cover that slow or reduce runoff.

5
Section 438 Technical Guidance December 2009

6
Q
Post-Development Condition
Pre-Development Condition
t

Figure 2. Post-Development Hydrograph.
(Q = volumetric flow rate; t = time)


Figure 3. Stream Displaying the Effects of Stormwater Runoff and Channel Downcutting.

The Solution: Preserving and Restoring Hydrology
A new approach has evolved in recent years to eliminate or reduce the amount of water and
pollutants that run off a site and ultimately are discharged into adjacent water bodies.

Section 438 Technical Guidance December 2009

7
The fundamental principle is to employ systems and practices that use or mimic natural
processes to: 1) infiltrate and recharge, 2) evapotranspire, and/or 3) harvest and use precipitation
near to where it falls to earth.

GI/LID practices include a wide variety of practices that utilize these mechanisms. These
practices can be used at the site, neighborhood and watershed/regional scales. In this document
the focus is on site-level practices, which is most consistent with the terms used in Section 438:
“project,” “facility,” and “property.” Although these performance requirements apply at the
project site-level, flexibility exists to utilize nearby areas or areas directly adjacent to the facility
to manage the runoff, i.e., evapotranspirate, infiltrate or harvest and use. Where justifiable, it
also may be appropriate to evapotranspirate, infiltrate or harvest and use an equivalent or greater
amount of runoff offsite as long as the runoff is discharged or used in the same receiving
subwatershed or watershed.

The purpose of EISA Section 438 is to replicate the pre-development hydrology to protect and
preserve both the water resources onsite and those downstream. For example, if prior to
development, twenty five (25) percent of the annual rainfall runs directly into the stream and the
remainder infiltrates into the ground or is evapotranspired into the air, then the post-development
goal should be to limit runoff to twenty five (25) percent of the annual precipitation while
maintaining the correct aquifer recharge rate. This has the benefit, in most cases, of delivering
water to the stream at approximately the same rate, volume, duration and temperature as the
stream had naturally evolved to receive prior to development. The result will be to eliminate or minimize the erosion of streambeds and streambanks, significantly reduce the delivery of many pollutants to water bodies, and retain historical instream temperatures.

Restoring or maintaining pre-development hydrology has emerged as a control approach for several reasons. Most importantly, this approach is intended to directly address the root cause of impairment.
Current control approaches have been selected in an attempt to control the symptoms (peak flow, and excess pollutants), but this strategy is not fully adequate because of the scale of the problem, the cumulative impacts of multiple developments and the need to manage both site and watershed level impacts.
With current approaches, it is also difficult to adequately protect and improve water quality because the measures employed are not addressing the main problem which is a hydrologic imbalance.

Designing facilities based on the goal of maintaining or restoring pre-development hydrology provides a site specific basis and an objective methodology with which to determine appropriate practices to protect the receiving environment.
Figure 4. Parking lot bioswale and permeable pavers in Chicago.
Section 438 Technical Guidance December 2009

8
Using pre-development hydrology as the guiding control principal also allows the designer to
consider climatic and geologic variability and tailor the solutions to the project location. Thus the need for a one size fits all approach is rendered unnecessary since the design objective is dictated by the pre-development site conditions and other technicalities of the project site and facility.
Instead of prescribed approaches dictating discharge volumes or flow rates, site assessments of historical infiltration and runoff rates will inform the designer and provide the basis for a suitable design. The use of this approach will minimize compliance complications that may arise from prescriptive design approaches which do not account for the variability of precipitation frequencies, rainfall intensities and pre-development land cover and soil conditions that influence infiltration and runoff.


B. BENEFITS AND OUTCOMES OF THE NEW STORMWATER PERFORMANCE REQUIREMENTS

Implementation of these new stormwater performance requirements in EISA Section 438
provides numerous environmental and economic benefits in addition to reducing the volume of
stormwater runoff:

Benefits to Water Resources:

 Cleaner Water. The use of plants, soils and water harvesting and use practices can reduce stormwater runoff volumes and pollutant loadings and the frequency and magnitude of combined sewer overflows (volume and pollutant loading reductions). These practices are part of a larger set of practices called green infrastructure/low-impact development.
 Clean and Adequate Water Supplies. GI/LID approaches using soil based vegetated infiltration
systems can be used to recharge ground water and maintain stream base flow. By recharging ground water aquifers, aquatic ecosystem health is maintained and base flows are increased which helps ensure more constant flows for drinking water withdrawals. Harvesting and reusing rainwater also reduces the need to use potable water for all uses and can reduce both the infrastructure and energy needed to treat and transport both drinking water and
stormwater.
 Source Water Protection. GI/LID practices provide pollutant removal benefits, thereby providing some protection for both ground water and surface water sources of drinking water. In addition, GI/LID provides ground water recharge benefits.


GI/LID approaches are a set of management approaches and technologies that utilize and/or
mimic the natural hydrologic cycle processes of infiltration, evapotranspiration and use. GI/LID practices include green roofs, trees and tree boxes, rain gardens, vegetated swales, pocket wetlands, infiltration planters, porous and permeable pavements, vegetated median strips, reforestation and revegetation and protection of
riparian buffers and floodplains.
These practices can be used almost anywhere soil and vegetation can be worked into the urban or suburban landscape.
They include decentralized harvesting approaches such as rain barrels and cisterns that can be used to capture and re-use rainfall for watering plants or flushing toilets.
Section 438 Technical Guidance December 2009

9
Other Social and Environmental Benefits:

 Cleaner Air. Trees and vegetation improve air quality by filtering many airborne pollutants
and can help reduce the amount of respiratory illness (Vingarzan and Taylor, 2003).
 Reduced Urban Temperatures. Summer city temperatures can average 10oF higher than nearby suburban temperatures (Casey Trees, 2007). High temperatures are also linked to higher ground level ozone concentrations. Vegetation creates shade, reduces the amount of heat absorbing materials and emits water vapor – all of which cool hot air (Grant, et al., 2003). Reductions in impervious surface and the use of light colored pervious surfaces (e.g.,
permeable concrete) also can mitigate urban temperatures.

regionally, but GI/LID techniques can provide adaptation benefits for a wide array of circumstances. They can be used to conserve, harvest and use water, to recharge ground waters
and to reduce surface water discharges that could contribute to flooding. In addition, there are mitigation benefits such as reduced energy demand and carbon sequestration by vegetation.
 Increased Energy Efficiency. Green space helps lower ambient temperatures and, when incorporated on and around buildings, helps shade and insulate buildings from wide temperature
swings, decreasing the energy needed for heating and cooling. Diverting stormwater from
wastewater collection, conveyance and treatment systems can reduce the amount of energy needed
to pump and treat the water. Energy efficiency not only reduces costs, but also reduces generation of greenhouse gases.
 Community Benefits. Trees and plants improve urban aesthetics and community livability by
Moderate the Impacts of Climate Change. Climate change impacts and effects vary providing recreational and wildlife areas. Studies show that property values are higher when trees and other vegetation are present. Increased green space also has public health benefits and has been shown to reduce crime and the associated stresses of urban living.


C. APPLICABILITY AND DEFINITIONS

Applicability

1. Who is a “Sponsor” of a project?

Section 438 applies to the “sponsor of any development or redevelopment project involving a
Federal facility . . .” Section 438 requires that the “sponsor . . . shall use . . . strategies for the property to maintain or restore . . . the predevelopment hydrology. . .” The “sponsor” should
Figure 5. Rain water cistern.
Section 438 Technical Guidance December 2009

10
generally be regarded as the federal department or agency that owns, operates, occupies or is the primary user of the facility and has initiated the development or redevelopment project. If the federal agency hires another entity to perform activities such as site construction or maintenance, the agency should nonetheless be regarded as the sponsor and be responsible to assure compliance with the requirements of Section 438. The agency sponsor is free to contract out various duties and responsibilities that are associated with achieving compliance.

2. What is a “Federal facility”?

Section 438 provides that its requirements apply to the “sponsor of any development or
redevelopment project involving a Federal facility . . .” Section 401(8) of EISA states: “The
term `Federal facility' means any building that is constructed, renovated, leased, or purchased in part or in whole for use by the Federal Government.”

3. What is a “footprint”?

Section 438 applies to a federal facility “with a footprint that exceeds 5,000 square feet.” For the purposes of this guidance, any project involving a federal facility that disturbs 5,000 square feet or more of ground area is covered by this guidance. Existing facilities that have an overall footprint of 5,000 square feet or greater that disturb less than 5,000 square feet of land area as part of any single development or redevelopment project are not subject to Section 438 requirements. Consistent with the purpose of Section 438 to preserve or restore pre-development hydrology, the term “footprint” includes all land areas that are disturbed as part of the project.

4. What is “the property”?

Section 438 provides that the project sponsor “shall use site planning, design, construction, and maintenance strategies for the property to maintain or restore, to the maximum extent
technically feasible, the predevelopment hydrology of the property.” This clause has been
interpreted to mean that the land surrounding the project site is available to implement the
appropriate GI/LID practices where optimal.
Although the performance requirements of EISA Section 438 apply only to the project footprint,
the flexibility exists to utilize the entire federal property in implementing the stormwater
strategies for the project.

Definitions

95th percentile rainfall event. The 95th percentile rainfall event represents a precipitation amount which 95 percent of all rainfall events for the period of record do not exceed. In more technical terms, the 95th percentile rainfall event is defined as the measured precipitation depth accumulated over a 24-hour period for the period of record that ranks as the 95th percentile rainfall depth based on the range of all daily event occurrences during this period.

The 24-hour period is typically defined as 12:00:00 am to 11:59:59 pm. In general, at least a 20- 30 year period of rainfall record is recommended for such an analysis. This raw data is readily Section 438 Technical Guidance December 2009 available and collected by most airports across the county. Small rainfall events that are 0.1 of an
inch or less are excluded from the percentile analysis because this rainfall generally does not
result in any measureable runoff due to absorption, interception and evaporation by permeable,
impermeable and vegetated surfaces. Many stormwater modelers and hydrologists typically
exclude rainfall events that are 0.1 inch or less from calculations of rainfall events of any storm from their modeling analyses of rainfall event frequencies. See, for example, the Center for Watershed Protection's Urban Subwatershed Restoration Manual 3 (available at www.cwp.org).

Federal facility. The term “federal facility” means any buildings that are constructed, renovated,
leased, or purchased in part or in whole for use by the federal government as defined in section
401(8) of the Energy Independence and Security Act.

Development or re-development. For the purposes of this provision this term applies to any
action that results in the alteration of the landscape during construction of buildings or other
infrastructure such as parking lots, roads, etc, (e.g., grading, removal of vegetation, soil
compaction, etc.) such that the changes affect runoff volumes, rates, temperature, and duration of
flow. Examples of projects that would fall under “re-development” include structures or other
infrastructure that are being reconstructed or replaced and the landscape is altered. Typical
patching or resurfacing of parking lots or other travel areas would not fall under this requirement.


D. TOOLS TO IMPLEMENT THE REQUIREMENTS OF SECTION 438

Section 438 of the Energy Independence and Security Act of 2007 reads as follows:
Section 438. Storm water runoff requirements for federal development projects.
The sponsor of any development or redevelopment project involving a Federal facility
with a footprint that exceeds 5,000 square feet shall use site planning, design,
construction, and maintenance strategies for the property to maintain or restore, to the
maximum extent technically feasible, the predevelopment hydrology of the property with
regard to the temperature, rate, volume, and duration of flow.

The intention of EISA Section 438 is to preserve or restore the hydrology of the site during the
development or redevelopment process. To be more specific, this requirement is intended to
ensure that aquatic biota, stream channel stability, and historical aquifer recharge rates of
receiving waters are not negatively impacted by changes in runoff temperature, volumes,
durations and rates resulting from federal projects. A performance based approach was selected
in lieu of a prescriptive requirement in order to provide site designers maximum flexibility in
selecting control practices appropriate for the site.

To meet these performance objectives, technically feasible stormwater control practices that are
effective in reducing the volume of stormwater discharge should be used. To implement EISA
Section 438, this guidance recommends that the federal facility use all known, available and
reasonable methods of stormwater retention and/or use to the maximum extent technically
feasible (METF). Tools to implement the requirements of Section 438 are described below and
illustrated in Figure 8.

Section 438 Technical Guidance December 2009

12
Establishing Section 438 Performance Design Objectives
Described below are options site designers can use to comply with Section 438. There may be
situations where Option 1 (retaining the 95th percentile rainfall event) is not protective enough to
maintain or restore the predevelopment hydrology of the project (for example, in some
headwater streams). In these cases, Option 2 (site-specific hydrologic analysis) could be used to
determine the types of stormwater practices necessary to preserve predevelopment runoff
conditions. Option 2 could also be used if predevelopment runoff conditions can be maintained
by retaining less than the 95th percentile rainfall event. Because a performance based approach
was selected in lieu of a prescriptive requirement in order to provide site designers maximum
flexibility in selecting control practices appropriate for the site, Option 2 was provided in
recognition that there are established methodologies that can be utilized to estimate the volume
of infiltration and evapotranspiration based on site-specific hydrology and thus establish the
predevelopment hydrology performance design objectives.



One approach to establishing the performance design objectives is to design, construct, and
maintain stormwater management practices that manage rainfall onsite, and prevent the off-site
discharge of the precipitation from all rainfall events less than or equal to the 95th percentile
rainfall event to the maximum extent technically feasible (METF). This objective should be
accomplished by the use of practices that infiltrate, evapotranspire and/or harvest and use
rainwater. The 95th percentile rainfall event is the event whose precipitation total is greater than
or equal to 95 percent of all storm events over a given period of record. For example, to
determine what the 95th percentile storm event is in a specific location, all 24 hour storms that
have recorded values over a 30 year period would be tabulated and a 95th percentile storm would
be determined from this record, i.e., 5% of the storms would be greater than the number
determined to be the 95th percentile storm. Thus the 95th percentile storm would be represented
by a number such as 1.5 inches, and this would be the design storm (example 95th percentile
storm events for selected cities are presented in Table 1). The designer would then select a
system of practices, to the METF, that infiltrate, evapotranspire or harvest and use this volume
multiplied by the total area of the facility/project footprint. Methods and data used to estimate
the 95th percentile event are discussed in Part II of this document.

For the purposes of this guidance, retaining all storms up to and including the 95th percentile
storm event is analogous to maintaining or restoring the pre-development hydrology with respect
to the volume, flow rate, duration and temperature of the runoff for most sites. This 95th
percentile approach was identified and recommended because this storm size represents the
volume that appears to best represent the volume that is fully infiltrated in a natural condition and thus should be managed onsite to restore and maintain this pre-development hydrology for duration, rate and volume of stormwater flows. In general, only large storms generate significant runoff. In addition, this approach was identified because it employs natural treatment and flow attenuation methods that are presumed to have existed on the site before construction of infrastructure (e.g., building, roads, parking lots, driveways,) and is intended to infiltrate or evapotranspirate the full volume of the 95th percentile storm. Because this approach necessitates the use of practices that generally preclude extended detention, it will also typically address the
Option 1: Retain the 95th Percentile Rainfall Event
Section 438 Technical Guidance December 2009
13
issue of maintaining predevelopment temperatures. However, in cases where there are discharges
to cool water streams or other sensitive receiving waters, additional strategies may be needed to ensure that stormwater discharges do not result in greater thermal impacts than would occur in pre-development conditions (Schueler and Helfrich, 1988).
Where technically feasible, the goal of Option 1 is that one hundred percent (100%) of the
volume of water from storms less than or equal to the 95th percentile event over the footprint of the project should not be discharged to surface waters. In some cases, runoff can be harvested and used and ultimately may be discharged to surface waters or a sanitary treatment system; such direct or indirect discharges must be authorized or allowed by the regulatory authority. For example, if runoff is captured for nonpotable uses such as toilet flushing or other uses that are not irrigation related, these waters potentially could be discharged into the sanitary sewer system.
Preferred mechanisms for retaining discharges from storms greater than the 95th percentile event are through overflow or diversion for the volume that exceeds the 95th percentile amount.
Because standard underdrains typically discharge from smaller storms as well, underdrain designs, if employed, should ensure adequate retention capacity for the 95th-percentile event
volume. For structures such as roofs and paved surfaces that can increase the temperature of
stormwater runoff, materials that minimize temperature increases (e.g., concrete vs. asphalt;
vegetated roofs) should be considered and used as appropriate.
Retaining 100 percent of all rainfall events equal to or less than the 95th percentile rainfall event was identified as Option 1 because small, frequently occurring storms account for a large proportion of the annual precipitation volume, and the runoff from those storm events also significantly alters the discharge frequency, rate and temperature of the runoff.
The runoff produced by these small storms and the initial portion of larger storms has a strong
negative cumulative impact on receiving water hydrology and water quality. In areas that have
been developed, runoff is generated from almost all storms, both small and large, due to the
impervious surfaces associated with development and the loss of soil and vegetation.
In contrast, natural or undeveloped areas discharge little or no runoff from small storms
because the rain is absorbed by the landscape and vegetation. Studies have shown that
increases in runoff event frequency, volume and rate can be diminished or eliminated through the use of GI/LID designs and practices, which infiltrate, evapotranspire and capture and use
stormwater.

Figure 6. Bioretention facility in Oregon.

Section 438 Technical Guidance December 2009

14
Option 1 was identified because it is a simplified approach to meet the intent of Section 438 in contrast to Option 2 which requires the designer to conduct a hydrologic analysis of the site based on site-specific conditions.

Table 1. Example 95th Percentile Storm Events for Select U.S. Cities
(adapted from Hirschman and Kosco, 2008).
City
95th Percentile
Event Rainfall
Total (in) City
95th Percentile
Event Rainfall
Total (in)
Atlanta, GA 1.8 Kansas City, MO 1.7
Baltimore, MD 1.6 Knoxville, TN 1.5
Boston, MA 1.5 Louisville, KY 1.5
Buffalo, NY 1.1 Minneapolis, MN 1.4
Burlington, VT 1.1 New York, NY 1.7
Charleston, WV 1.2 Salt Lake City, UT 0.8
Coeur D’Alene, ID 0.7 Phoenix, AZ 1.0
Cincinnati, OH 1.5 Portland, OR 1.0
Columbus, OH 1.3 Seattle, WA 1.6
Concord, NH 1.3 Washington, DC 1.7
Denver, CO 1.1
0.0
1.0
2.0
3.0
4.0
5.0
6.0
7.0
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Percentile
R
a
in
fa
l
l
D
ep
th
(
in
ch
e
s
)
95%

Figure 7. Rainfall Frequency Spectrum showing the 95th percentile rainfall event for Portland, OR
(~1.0 inches)


Section 438 Technical Guidance December 2009

15
Calculating the 95th Percentile Rainfall Event
Section E of this guidance contains information on how to calculate the 95th percentile rainfall
event for a specific area. A long-term record of daily rainfall amounts (ideally, at least 30 years)
is needed to calculate the 95th percentile rainfall.

Designers opting to use Option 1 need to do the following:

1) calculate or verify the precipitation amount from the 95th percentile storm event (this number
would be typically expressed in inches, e.g., 1.5”, and

2) employ onsite stormwater management controls to the METF that infiltrate, evapotranspire
or harvest and use the appropriate design volume.

The 95th percentile event can be calculated by using the following procedures below
(summarized from Hirschman and Kosco, 2008, Managing Stormwater in Your Community: A
Guide for Building an Effective Post-Construction Program, Center for Watershed Protection):

 Obtain a long-term rainfall record from a nearby weather station (daily precipitation is
fine, but try to obtain at least 30 years of daily record). Long-term rainfall records can be
obtained from many sources, including NOAA at
http://cdo.ncdc.noaa.gov/pls/plclimprod/poemain.accessrouter?datasetabbv=SOD&countr
yabbv=&georegionabbv=.
 Remove data for small rainfall events that are 0.1 inch or less and snowfall events that do
not immediately melt from the data set. These events should be deleted since they do not
typically cause runoff and could potentially cause the analyses of the 95th percentile
storm runoff volume to be inaccurate.
 Using a spreadsheet or simple statistical package, sort the rainfall events from highest to
lowest. In the next column, calculate the percentage of rainfall events that are less than
each ranked event (event number/total number of events). For example, if there were
1,000 rainfall events and the highest rainfall event was a 4” event, then 999 events (or a
percentile of 999/1000, or 99.9%) are less than the 4” rainfall event.
 Use the rainfall event at 95% as the 95th percentile storm event.



Another approach to establishing the performance design objective is to design, construct, and
maintain stormwater management practices that preserve the pre-development runoff conditions
following construction. Option 2 allows the designer to conduct a site-specific hydrologic
analysis to determine the pre-development runoff conditions instead of using the estimated
volume approach of Option 1. Under Option 2, the pre-development hydrology would be
determined based on site-specific conditions and local meteorology by using continuous
simulation modeling techniques, published data, studies, or other established tools. If the
designer elects to use Option 2, the designer would then identify the pre-development condition
of the site and quantify the post-development runoff volume and peak flow discharges that are
equivalent to pre-development conditions. The post-construction rate, volume, duration and
Option 2: Site-Specific Hydrologic Analysis
Section 438 Technical Guidance December 2009

16
temperature of runoff should not exceed the pre-development rates and the predevelopment
hydrology should be replicated through site design and other appropriate practices to the
maximum extent technically feasible. These goals should be accomplished through the use of
infiltration, evapotranspiration, and/or rainwater harvesting and use. Defensible and consistent
hydrological assessment tools should be used and documented. Additional discussions of
appropriate methodologies to use in assessing site hydrology have been included in the technical
sections of this document. See, for example, the discussion of spreadsheet versions or curve
numbers based on the Natural Resource Conservation Service Technical Release 55 (TR-55)
Method in Appendix A of this document.

Development
The pre-development hydrologic condition of the site is the combination of runoff, infiltration
and evapotranspiration rates and volumes that typically existed on the facility site before
"development" on a greenfields site (meaning any construction of infrastructure on undeveloped
land such as meadows or forests). In practice, determining the pre-development hydrology of a
given site can be difficult if there is no suitable reference site. As a result, reference conditions
for typical land cover types in the locality often are used to approximate what fraction of the
precipitation ran off, soaked into the ground or was evaporated from the landscape. The use of
reference conditions can be problematic if suitable data are not available or unique site
conditions exist that do not fit within a typical land use cover type for the area, e.g., meadow or
forest. In cases where suitable data from comparable conditions cannot be found or is otherwise
inadequate to be used in conducting an Option 2 analysis for the specific area being considered
for development or redevelopment, the project sponsor should use the Option 1 analytical
framework.

Re-development
For re-development sites, existing site conditions and uses of the site can influence the amount of
runoff that can be managed on site through infiltration, evapotransporation and harvest and use
and thus the performance design objective. In these cases the design process in Figure 8 and
Scenario 9 illustrate the decision processes that can be used.

In the context of some re-development projects, fully restoring predevelopment hydrology can be
difficult to achieve and Congress recognized this potential difficulty by including the METF
language in the statute. In these cases, Congressional intent can be best carried out by using a
systematic METF analysis to determine what practices can be implemented at the site to
maintain or store the hydrologic condition of the site. Scenarios 1-8 provide examples of METF
analyses that demonstrate that pre-development hydrology can be achieved. Scenario 9 provides
an example of an METF analysis that demonstrates that pre-development hydrology cannot be
fully achieved and illustrates the extent to which pre-development hydrology can be restored.

Note: It should also be emphasized that the performance based approach in Option 1 is intended
to be a surrogate for determining the pre-development reference condition and this standard is
intended to be used in cases where it is more practical, cost effective, and/or expeditious than
Option 1, or where it is difficult or infeasible to identify the relevant reference conditions for the
site.

Section 438 Technical Guidance December 2009

17
Determination of Maximum Extent Technically Feasible

Compliance with Section 438 requires that stormwater management measures are implemented
to the maximum extent technically feasible (METF) to maintain or restore the pre-development
hydrology conditions specifically with respect to temperature, rate, volume, and duration of flow.

Performance or design goals based on the pre-development hydrology can be established by
using options such as the following: Retention of the 95th percentile rainfall event (Option 1), or
through a site-specific hydrologic analysis that estimates the volume of infiltration,
evapotranspiration or onsite stormwater harvesting and use based on site-specific hydrologic
conditions (Option 2).

Technical Infeasibility

For projects where technical infeasibility exists, the federal agency or department sponsoring the
project should document and quantify that stormwater strategies, such as infiltration,
evapotranspiration, and harvesting and use have been used to the METF, and that full
employment of these types of controls are infeasible due to site constraints. Some western states
place restrictions on harvesting and use due to water rights, however, these requirements do not
necessarily preclude the sponsor of the project from implementing strategies such as infiltration
and evapotranspiration. Documentation of technical infeasibility should include, but may not be
limited to, engineering calculations, geologic reports, hydrologic analyses, and site maps. A
determination that the performance design goals cannot be met on site should include analyses
that rule out the use of an adequate combination of infiltration, evapotranspiration, and use
measures. Examples of where site conditions may prevent the full employment of appropriate
management techniques to the METF include a combination of:

 The conditions on the site preclude the use of infiltration practices due to the presence of
shallow bedrock, contaminated soils, near surface ground water or other factors such as
underground facilities or utilities.
 The design of the site precludes the use of soil amendments, plantings of vegetation or other
designs that can be used to infiltrate and evapotranspirate runoff.
 Water harvesting and use are not practical or possible because the volume of water used for
irrigation, toilet flushing, industrial make-up water, wash-waters, etc. is not significant
enough to warrant the design and use of water harvesting and use systems.
 Modifications to an existing building to manage stormwater are not feasible due to structural
or plumbing constraints or other factors as identified by the facility owner/operator.
 Small project sites where the lot is too small to accommodate infiltration practices adequately
sized to infiltrate the volume of runoff from impervious surfaces,
 Soils that cannot be sufficiently amended to provide for the requisite infiltration rates,
 Situations where site use is inconsistent with the capture and use of stormwater or other
physical conditions on site that preclude the use of plants for evapotranspiration or
bioinfiltration.
 Retention and/or use of stormwater onsite or discharge of stormwater onsite via infiltration
has a significant adverse effect on the site or the down gradient water balance of surface
waters, ground waters or receiving watershed ecological processes.
Section 438 Technical Guidance December 2009

18
 State and local requirements or permit requirements that prohibit water collection or make it
technically infeasible to use certain GI/LID techniques.
 Compliance with the Section 438 requirements would result in the retention and/or use of
stormwater on the site such that an adverse water balance impact may occur to the receiving
surface waterbody or ground water.

Please note that a single one of these characteristics is very unlikely to preclude meeting the
performance standard, but a combination of factors may.

In cases where the facility has a defensible showing of technical infeasibility and can provide
adequate documentation of site conditions or other factors that preclude full implementation of
the performance design goal, the facility should still install stormwater practices to infiltrate,
evapotranspire and/or harvest and use onsite the maximum amount of stormwater technically
feasible. Note: Facilities must still comply with all other applicable federal, state and local
requirements.


Section 438 Technical Guidance December 2009


19
Figure 8. Section 438 Implementation Process
Section 438 Technical Guidance December 2009

20
Documenting EISA Section 438 Implementation
Each agency or department is responsible for ensuring compliance with Section 438. It is
recommended that: 1) the final design and as-built drawings of each facility shall be reviewed
by a registered professional engineer and 2) the agency or department develop and maintain
documentation of the following design criteria for each project subject to Section 438:

 Site evaluation and soils analysis
 Calculations for the 95th percentile rainfall event or the pre-development runoff volumes
and rates to identify the volume of stormwater requiring management
 Documentation of modifications to the performance design objective based on technical
constraints (site-specific METF determination)
 The site design and stormwater management practices employed on the site
 Design calculations for each stormwater management practice employed
 The respective volume of stormwater managed by each practice and the system as a
whole
 Operations and maintenance protocols for the stormwater management system

The information should provide the necessary documentation and detail to demonstrate
compliance and operation of stormwater management practices for the entire site.

Common Green Infrastructure/Low Impact Development Tools to Implement Section 438
Although Congress did not prescribe specific practices to be used to implement Section 438 it
can be inferred that one of the goals of the Act was to promote the use of innovative stormwater
management approaches, designs and practices that better protect receiving water quality, flow
regimes and provide other important environmental benefits. GI/LID are preferred practices, to
be supplemented with or replaced with conventional controls when site specific conditions
dictate.

The GI/LID management approaches and technologies that federal agencies would typically use
enhance and/or mimic the natural hydrologic cycle processes of infiltration, evapotranspiration,
and use. Federal agencies can also use footprint reduction practices (e.g., building up instead of
out) to reduce their stormwater impact. GI/LID approaches include biological systems and
engineered systems. These include but are not necessarily limited to:

 Rain gardens, bioretention, and infiltration planters
 Porous pavements
 Vegetated swales and bioswales
 Green roofs
 Trees and tree boxes
 Pocket wetlands
 Reforestation/revegetation using native plants
 Protection and enhancement of riparian buffers and floodplains
 Rainwater harvesting for use (e.g., irrigation, HVAC make-up, non-potable indoor uses).

GI/LID practices are recommended to implement EISA Section 438 for the following reasons:
 cost savings in many cases
Section 438 Technical Guidance December 2009

21
 overall environmental performance
 pollutant loading reduction capability
 pollution prevention focus
 effectiveness in managing runoff
volumes and rates
 energy efficient and energy
conservative
 appropriate in a wide range of site
condition and locations
 appropriate for new development and
redevelopment projects
 appropriate at multiple scales of
development, e.g., site, neighborhood,
region

For more information on specific GI/LID
practices and how they function, visit:
www.epa.gov/greeninfrastructure and
www.epa.gov/nps/lid.

Cost of Compliance
The cost of complying with Section 438 may
require the use of approaches and techniques
that initially may be more costly to design and implement. It is anticipated that as the expertise
of the implementing agency or department increases and the demand for GI/LID materials and
equipment increases that the overall costs of the projects will be lower or equivalent to the costs
of constructing conventional stormwater practices. Initial studies conducted by EPA and others
suggest that the use of GI/LID practices can be cost competitive. Recent evaluations of GI/LID
projects have identified opportunities for cost savings because of reduced infrastructure and site
preparation demands. In addition, longer term studies have indicated that GI/LID practices are
continuing to gain cost efficiency as they are adopted more widely and with greater frequency
thus reducing overall implementation costs.

In Reducing Stormwater Costs through LID Strategies and Practices (EPA 841-F-07-006,
December 2007 - available for download at www.epa.gov/nps/lid), EPA examined 17 case
studies in which conventional development costs were compared to GI/LID costs. In the great
majority of cases, the GI/LID approach was between 15 and 80 percent less expensive than
conventional control measures because implementation of GI/LID practices can offset costs of
conventional construction and stormwater management approaches. Significant cost savings that
were identified in the report include:

 Elimination or reduction of detention ponds
 Elimination or reductions of stormwater and CSO treatment and conveyance systems
such as pipes, storage structures, stormwater treatment devices, and other related
stormwater infrastructure
 Narrower streets with reduced material demands
Figure 9. Disconnected downspout
discharging to planter box.
Section 438 Technical Guidance December 2009

22
 Fewer square yards of sidewalks
 Reduced land purchases for stormwater control structures

In addition, other benefits were achieved through the use of GI/LID such as more beneficial uses
of land previously dedicated to stormwater devices, increased livability and higher property
values.

There are many different combinations of practices that can be employed at particular sites to
achieve pre-development hydrology. In selecting the appropriate set of practices to be used at
the site, project sponsors should consider a broad range of factors, including cost-effectiveness of
particular combinations of practices as applied to the site, as well as the potential for ancillary
cost savings or community benefits (e.g., elimination or reduction of infrastructure costs, or the
creation of attractive green spaces). EPA encourages project sponsors to include these factors in
the planning and design phases of their projects so as to maximize triple bottom-line (economic,
environmental, and social) results.

E. CALCULATING THE 95TH PERCENTILE RAINFALL EVENT

A long period of precipitation records, i.e., a minimum of 10 years of data, is needed to
determine the 95th percentile rainfall event for a location. Thirty years or more of monitoring
data are desirable to conduct an unbiased statistical analysis. The National Climatic Data Center
(NCDC) provides long-term precipitation data for many locations of the United States. You can
download climate data from their Web site (www.ncdc.noaa.gov) or by ordering compact discs
(NOTE: The NCDC charges a fee for access to their precipitation data). Local airports,
universities, water treatment plants, or other facilities might also maintain long-term
precipitation records. Data reporting formats can vary based on the data sources. In general,
each record should include the following basic information:

 Location (monitoring station)
 Recording time (usually the starting time of a time-step)
 Total precipitation depth during the time-step

In addition to the above information, a status flag is sometimes included to indicate data
monitoring errors or anomalies. Typical NCDC flags include A (end accumulation), M (missing
data), D (deleted data), or I (incomplete data). If there are no flags, the record has passed the
quality control as prescribed by the NCDC and has been determined to be a valid data point.

There are several data processing steps to determine the 95th percentile rainfall event using a
spreadsheet. These steps are summarized below:

1. Obtain a long-term 24-hr precipitation data set for a location of interest (i.e., from the
NCDC website).

2. Import the data into a spreadsheet. In MS Excel [Data / Import External Data / Import Data]

Section 438 Technical Guidance December 2009

23
3. Rearrange all of the daily precipitation records into one column if the original data set has
multiple columns of daily precipitation records.



4. Review the records to identify if there are early periods with a large number of flagged
data points (e.g., erroneous data points). Select a long period of good recording data that
represents, ideally, 30 years or more of data. Remove all of the extra data (if not using the
entire dataset).

5. Remove all flagged data points (i.e., erroneous data points) from the selected data set for
further analysis.

6. Remove small rainfall events (typically less than 0.1 inches), which may not contribute to
rainfall runoff. These small events are categorized as depressional storage, which, in
general, does not produce runoff from most sites.



Note: Steps 4 through 6 can be processed by applying data sort, delete and
re-sort spreadsheet functions. In MS Excel [Data / Sort]

7. Calculate the 95th percentile rainfall amount by applying the PERCENTILE spreadsheet
function at a cell. In MS Excel [=PERCENTILE(precipitation data range,95%)]


Section 438 Technical Guidance December 2009

24

Note: The PERCENTILE function returns the nth percentile of value in the entire
precipitation data range. This function can be used to determine the 95th percentile storm
event that captures all but the largest 5% of storms.

8. The 95th percentile was calculated in the previous step. However, if the user would like
to see this information represented graphically and get a relative sense of where
individual storm percentiles fall in terms of rainfall depths, the following methodology
can be used. Derive a table showing percentile versus rainfall depth to draw a curve as
shown below. The PERCENTILE spreadsheet function can be used for each selected
percent. It is recommended to include at least 6 points between 0% and 100% (several
points should be between 80% and 100% to draw an accurate curve).





Use the spreadsheet software to create of plot of rainfall depth versus percentile, as shown above.
The 95th percentile storm event should correlate to the rainfall depth calculated in step 7,
however the graph can be used to calculate rainfall depths at other percentiles (e.g., 50%, 90%).
Section 438 Technical Guidance December 2009

25
Part II: Case Studies on Capturing the 95th Percentile Storm
Using Onsite Management Practices

INTRODUCTION

This section contains nine case studies that are intended to be representative of the range of
projects that are subject to the requirements legislated in Section 438 of the Energy
Independence and Security Act. The facility examples in the case studies were selected to
illustrate project scenarios for differing geographic locations, site conditions, and project sizes
and types. As noted in Part I, all projects with a footprint greater than 5,000 square feet must
comply with the provisions of Section 438. What this means is that both new development and
redevelopment projects should be designed to infiltrate, evapotranspirate, and/or harvest and use
runoff to the maximum extent technically feasible (METF) to maintain or restore the pre-
development hydrology of the site. Scenarios 1-8 are examples of sites where it was technically
feasible to design the stormwater management system to retain the 95th percentile storm onsite.
Scenario 9, however, was provided as an example of an METF analysis where site constraints
allowed the designers to retain only 75% of the 95th percentile storm.

Given the site-specific nature of individual projects, the case study scenarios described herein do
not include site specific design features such as runoff routing, specific site infiltration rates, the
structural loading capacity of buildings, etc. in terms of stormwater practice selection.

It should be noted that an example of Option 2, which requires a site-specific hydrologic
analysis, has not been provided in this document because of the complexity of factors and the
lack of general applicability such an analysis would have.

Background
Numerous approaches exist for determining the volume of runoff to be treated through
stormwater management. Retaining stormwater runoff from all events up to and including the
95th percentile rainfall event was identified as Option 1 because small, frequently-occurring
storms account for a large proportion of the annual precipitation volume. Using GI/LID
practices to retain both the runoff produced by small storms and the first part of larger storms can
reduce the cumulative impacts of altered flow regimes on receiving water hydrology, e.g.,
channel degradation and diminished baseflow. For the purposes of this guidance, retaining all
storms up to and including the 95th percentile storm event is analogous to maintaining or
restoring the pre-development hydrology with respect to the volume, flow rate, duration and
temperature of the runoff for most sites.

Determination of the 95th Percentile Rainfall Event
The 95th percentile rainfall rainfall event was determined using the long-term daily precipitation
records from the National Climate Data Center (NCDC, 2007). By analyzing the frequency and
rainfall depths from daily rainfall records over 24-hour periods, the 95th percentile storm event
can be determined. From a frequency analysis viewpoint, the 95th percentile event is the storm
event that is greater than or equal to 95% of all storms that occur within a given period of time.
Regional climate conditions and precipitation vary across the U.S. Because of local values, it is
essential that the implementing agency or department establish the 95th percentile storm event for
Section 438 Technical Guidance December 2009

26
the project site since the control volume may vary depending on local weather patterns and
conditions.

Onsite Stormwater Management Practice Determinations
For the purposes of the case study scenarios, the following four categories of practices were
selected as the most appropriate practices for implementing Section 438 requirements:
bioretention, permeable pavements and pavers, cisterns, and green roofs. These practices were
selected based on known performance data and cost. For each case study, the same hierarchy of
selection criteria was used, i.e., the most cost effective practices were considered before other
practices were considered. Bioretention practices were considered first because these systems
generally have the lowest cost per unit of stormwater treated (Hathaway and Hunt, 2007). Thus,
if the bioretention system could not be designed to adequately capture the desired runoff volume,
permeable pavement and pavers, cisterns, and green roofs were considered in that order based on
relative cost. In most cases a combination of practices was selected as part of an integrated
treatment system. It should be noted that all treatment systems were designed to accomplish the
goal of capturing the 95th percentile rainfall event onsite. Examples of onsite stormwater
management practices selected for each site are presented in the results section. For the Boston,
MA site, it was assumed that bioretention was not feasible in order to simulate a situation where
space was severely limited; as a result, interlocking modular pavers were selected as the most
cost-effective stormwater management to capture the requisite design volume. To further
illustrate the range of site conditions designers may encounter, and how site conditions impact
the selection of appropriate control options, Scenario #3 (Cincinnati, OH) was re-analyzed as
Scenario #8. In Scenario #8, it was assumed that the site had clay soils and low infiltrative
capacity. Given these site conditions, the range of potential control options was more limited
and a combination of modular paving blocks, a green roof, and cisterns was ultimately selected
based on cost and site suitability factors.

For purposes of these modeling exercises, a number of assumptions were associated with each
category of practice. These assumptions are not necessarily an endorsement of a particular
design paradigm, but rather were used to keep a somewhat conservative cap on the scenarios in
order to demonstrate the feasibility of the approach. For example, bioretention retrofits can and
should often be located in prior impervious locations; however, in all modeled scenarios
bioretention was restricted to currently landscaped areas. The assumptions were:

 Bioretention areas: On-lot retention of stormwater through the use of vegetation, soils,
and microbes to capture, treat and infiltrate runoff.

It is assumed bioretention practices would be installed within currently landscaped
pervious areas or that pervious areas would be created for bioretention cells. While
termed bioretention, these systems are designed to provide infiltration as well as
temporary storage. Bioretention areas would be designed to accept up to a depth of 10
inches of water across the surface of the bioretention cell (see Appendix A). The
conceptual design of this storage depth would occur within the media and/or could be
included as ponded storage. Further design storage beyond the 10 inches would be
acceptable (and encouraged) above the media on a site-by-site basis with ponded depth
generally not to exceed 12 inches.
Section 438 Technical Guidance December 2009

27

Uniform infiltration was assumed across the entire base of the bioretention cell. No
additional media underneath the amended soils were included in the designs with
infiltration rates in this layer governed by the in situ soils. Underdrains were not modeled
directly but could be applied at the point of storage overflow such that no overflow
occurs until the design depth of 10 inches is saturated. This approach was selected to
maximize the storage and infiltration benefits of these systems. Designs utilizing
underdrains at the base of the bioretention cell do not store the requisite volumes because
the media is permeable and the underdrain conveys the runoff offsite through the
underdrain before it can be infiltrated. Because standard underdrains typically discharge
from smaller storms as well, underdrain designs, if employed, should ensure adequate
retention capacity for the 95th percentile event volume.

The bioretention footprint for modeling purposes was calculated as one uniform area that
did not include side slopes. There is an expectation that actual bioretention cell
construction would be distributed throughout the site with targeted locations based on
hydrology (natural flow paths) and soils with greater infiltrative capacity. Side slopes
may increase the surface excavation area required to accommodate the footprint and
freeboard of these systems depending on the design or the bioretention system.

 Porous/permeable pavement: Transportation surfaces constructed of asphalt, concrete
or permeable pavers that are designed to infiltrate runoff.

Infiltration was modeled for the entire porous pavement area with drainage pipes used
only as overflow outlets. This design was chosen to maximize infiltration capabilities of
the system. While many types of porous pavement systems can be used, modular block
type pavers were generally applied in this design category under the assumption that they
typically include sufficient volumetric storage in the media layer. [Note: Other types of
porous pavement applications are available that support heavy loads and can be designed
to temporarily store and infiltrate runoff beneath the surface of the pavement.]

For these systems, an equivalent of 2 inches of design storage depth was assumed. This
design depth could be achieved by specifying 10 inches of media depth that had 20%
void space. Similarly, this could be achieved by designing six inches of media depth
above the bottom surface, with specified media containing 33% void space. This
alternative would have the overflow outlet at the 6 inch depth providing an equivalent
water storage depth of 2 inches.

The soils under the paver blocks may require or be subjected to some compaction for
engineering stability. As a result, infiltration into underlying soils was modeled
conservatively by applying the minimum infiltration rate for each soil type
(see Appendix A).

Generally, porous pavement is not recommended for high traffic areas or loading bays
Because of this the scenarios assumed that only a percentage of total parking and road
areas on a site can be converted to porous pavement. The assumed maximum percentage
Section 438 Technical Guidance December 2009

28
applied in the scenarios was set at 60% of the total paved area. Guidance on porous
pavements is available at:
http://cfpub.epa.gov/npdes/greeninfrastructure/technology.cfm#permpavements

 Cistern: Containers or vessels that are used to store runoff for future use.

Cisterns were modeled in cases where green roofs were not feasible or where it was
necessary to include additional storage volume to meet the goal of onsite rainfall runoff
capture. The sizes of cisterns would be calculated based on site-specific rainfall, site-
specific spatial and structural conditions, use opportunities and rates, and consideration of
cost per volume of storage. For simplicity, cistern volume was reported as a total
volume. This total volume could be subdivided into any number of cisterns to provide
the total necessary storage but should be based on the impervious area and runoff
quantities which will flow to the cistern. The most efficient cost per volume storage
would need to be considered on a site-by-site basis (see Appendix A).

 Green roof: Roof designed with light weight soil media and planted with vegetation.

Frequently, green rooftop area is limited by structural capacity. In addition, other rooftop
equipment may need to be accommodated in this space including HVAC systems and air
handlers. For this reason, and to provide a somewhat conservative rate of application, it
was assumed for these modeling analyses that up to 30% of a roof’s impervious area
could be converted into a green roof. Green roof area was assumed to have 1 inch of total
effective stormwater storage, i.e., a 2.5 inch media depth with 40% void space (see
Appendix A).

General Approach
Using site aerial photos, spatial analysis should be conducted to estimate the land cover types
and areas for each site. The surface conditions of each site can be digitized using geographic
information systems (GIS) techniques. Alternatively, computer-aided design (CAD) drawings
can be used to estimate the surface area of each land cover type. The schematic in Figure 10
illustrates the processes used for selecting and determining the overall size of stormwater
management practices for each site.
Section 438 Technical Guidance December 2009

29

Stormwater
Management Analysis &
Design Process
Collect spatial data from aerial
photos (determine pervious
and impervious areas)
Collect historic rainfall
data from nearest station
Determine the 95th percentile
24-hour rainfall event
Estimate the current runoff
Select onsite control measure options
Check whether control
measure options
meet performance
goals performance
Determine the size(s) of control measure(s)
Yes
1. Select alternative control measures
using METF analysis and site
limitations to determine appropriate
runoff control measures if
performance goals cannot be
achieved
and/or
2. Exercise optional offsite runoff
management approach and select
appropriate control measures
No
Determine location and size(s) of
onsite or off-site control measures

Design and implement control measure(s)
Yes
No Select control
measure(s) to fit the site
and confirm performance
Figure 10. Flow chart depicting the process for determining control measures
using the 95th percentile, 24-hour, annual rainfall event.

Section 438 Technical Guidance December 2009

30
The following steps provide more detailed information on acquiring and calculating the
necessary data to complete the processes indicated in Figure 10. This methodology was used in
the scenario analyses that follow.

Collecting spatial data for a site
1. Collect an aerial orthophotograph for the desired site.
2. Digitize land use/land cover conditions using GIS techniques. If CAD drawings of the
site exist, they can be used to estimate land cover area (pervious, impervious).
3. Categorize the digitized or planned land use/land cover based on surface hydrologic
conditions, e.g., rooftop, pavement, and pervious/landscaped area.
4. Estimate the size of each land use/land cover category (by polygon).

Determining the 95th percentile, 24-hr rainfall event
1. Obtain a long-term 24-hr precipitation data set for the location of interest (i.e., from the
NCDC Web site or other source).
2. Import the data into a spreadsheet. In MS Excel [Data / Import External Data / Import Data]
3. Rearrange all of the daily precipitation records into one column if the original data set has
multiple columns of daily precipitation records.
4. Remove all flagged data points (i.e., erroneous data points) from the selected data set for
further analysis.
5. Remove small rainfall events (typically less than 0.1 inches) that may not contribute to
rainfall runoff. These small storms often produce little if any appreciable runoff from
most sites and for modeling purposes are typically considered as volume captured in
surface depression storage.
6. Calculate the 95th percentile rainfall volume by applying the PERCENTILE spreadsheet
function to a range of data cells. The PERCENTILE function returns the nth percentile
value in the specified precipitation data range. This function can be used to determine the
95th percentile storm event that captures all but the largest 5% of storms. In MS Excel
[PERCENTILE(precipitation data range,95%)]

Estimating Current Runoff and Placing onsite control measures to capture the 95th percentile
rainfall event
1. Collect spatial data for a site, e.g., rooftop, pavement, and pervious areas as above.
2. Check soil type (USDA mapping, borings, or onsite testing) for the site to determine
infiltration parameters. For this modeling, many of the assumptions that pertain to
generalized soils groups and their infiltration properties come from the EPA Stormwater
Management Model (SWMM 4.x) manual (see Appendix A).
3. Determine the current runoff volume that would occur during a 24 hour period by
applying the 95th percentile rainfall to the existing site conditions (land use and soil
properties) as above using a hydrologic model (such as TR-55 or SWMM). For this
analysis, it is assumed that the rainfall amount is distributed over a 24 hour period.
Actual rainfall event duration (and intensity) was not considered for determining rainfall
runoff (however, timing was considered when modeling infiltration).
4. Determine flow paths so that management practice placements are in locations where
flows can be intercepted and routed to practices. Because this is a site specific effort and
may require detailed topographic information or further surveys this would be a task to be
Section 438 Technical Guidance December 2009

31
completed onsite and therefore is not included as a part of the modeling scenario
exercise.
5. Select onsite control practices to capture the current 95th percentile runoff event; base the
selection of appropriate options on site conditions, areas available for treatment options,
and other factors such as site use and other constraints.

Note: The steps above have been generalized for the purposes of this guidance. It is
recommended that a qualified professional engineer determine or verify that stormwater
management practices are sized, placed, and designed correctly. It should also be noted that the
methodology to determine rainfall amount used a 24 hour time period based on daily records.
Actual rainfall events may have occurred over shorter or longer time periods. Similarly, for
modeling purposes, the 24 hour rainfall amount was distributed to pervious and impervious areas
(and management practices) as a uniform event occurring during a 24-hour period. A large
dataset (greater than 50 years) was used to reasonably represent rainfall depth on a daily bases.
It stands to reason that more frequent, shorter duration precipitation events are better represented
than less frequent, longer duration precipitation events.

Scenarios
Eight locations were selected for the 9 case studies as shown in Figure 11 and Table 2. Case
study numbers 3 and 8 were both developed based on the Cincinnati, Ohio facility, although the
site parameters were altered to represent differing site conditions and design constraints. Annual
average rainfall depths for these locations range from 7.5 inches to 48.9 inches. Analyses of the
95th percentile rainfall events for these locations produced rainfall depths that range from 1.00
inch to 1.77 inches (Table 2).


Figure 11. Locations for Analyzing Onsite Control Measures.
Section 438 Technical Guidance December 2009

32

The government facilities in the 8 case study locations were selected because they represent
generic sites from the major climatic regions of the U.S. These facilities also were selected
because the sites have a range of site characteristics that can be used to illustrate different site
designs and stormwater management options, e.g., pervious, roof, and pavement areas (Table 3).
Site sizes ranged from 0.7 to 27 acres with percent site imperviousness area ranging from 47% to
95% of the site. Aerial photos of the sites are included along with site specific rainfall runoff
and soil results.

Table 2. Summary of Rainfall Data for the Seven Locations.
No Location NCDC Daily Precipitation Data Rainfall Depth (inches)
Period of record Coverage Annual
average 95
th
percentile
rainfall event
1 Charleston, WV 1/1/1948 - 12/31/2006 (59 yrs) 99% 43.0 1.23
2 Denver, CO 1/1/1948 - 12/31/2006 (59 yrs) 96% 15.2 1.07
3 Cincinnati, OH 1/1/1948 - 12/31/2006 (59 yrs) 96% 36.5 1.45
4 Portland, OR 1/1/1941 - 12/31/2006 (66 yrs) 98% 35.8 1.00
5 Phoenix, AZ 1/1/1948 - 12/31/2006 (59 yrs) 99% 7.5 1.00
6 Boston, MA 1/1/1920 - 12/31/2006 (87 yrs) 99% 41.9 1.52
7 Atlanta, GA 1/1/1930 - 12/31/2006 (77 yrs) 100% 48.9 1.77
8 Norfolk, VA 1/1/1957 - 12/31/2006 (50 yrs) 99% 45.4 1.68

The results of the spatial analyses were summarized and divided into three land cover categories;
rooftop, pavement, and pervious area, as shown in Table 3.

Table 3. Summary of Land-use Determinations of the Study Sites.
No Location Facility Spatial Info (acres) Site
Imperviousness Rooftop Pavement Pervious Total
1 Charleston, WV 0.1 0.4 0.2 0.7 73%
2 Denver, CO 0.5 1.9 2.0 4.5 55%
3 Cincinnati, OH 1.6 8.0 9.4 19 51%
4 Portland, OR 8.8 16.9 1.3 27 95%
5 Phoenix, AZ 0.2 0.7 1.1 2 47%
6 Boston, MA 0.9 1.5 1.1 3.5 69%
7 Atlanta, GA 3.9 10.8 6.2 21 70%
8 Norfolk, VA 0.9 0.55 0.15 1.6 91%

Methods for Determining Runoff Volume

Direct Determination of Runoff Volume
Runoff from each land cover was estimated using a simplified volumetric approach based on the
following equation:

Runoff = Rainfall – Depression Storage – Infiltration Loss

Again, this methodology does not consider routing of runoff; therefore slope is not considered
when calculating on a volumetric basis.
Section 438 Technical Guidance December 2009

33

Infiltration loss is calculated only in pervious areas (e.g., there is no infiltration in impervious
areas). In this analysis, infiltration was estimated using Horton’s equation:

Ft = f - k tmin + (fmax – fmin) e

where, Ft= infiltration rate at time t (in/hr)
fmin = minimum or saturated infiltration rate (in/hr)
fmax = maximum or initial infiltration rate (in/hr)
k = infiltration rate decay factor (/hr) and
t = time (hr) measured from time runoff first discharged into infiltration area

Infiltration loss for the 24-hr rainfall duration was estimated by the following equation with
assumptions of a half hour ∆t and uniform rainfall distribution in time:

Infiltration Loss = ∑ (f ·∆t)

To more accurately describe the dynamic process of infiltration associated with Horton’s
equation, infiltration loss was integrated over a 24-hour period using a half hour time step while
applying the maximum and minimum infiltration rates (in/hr) with time using the appropriate
soil decay factor. The results of this process are further illustrated in Appendix A.

Once runoff from each land cover was estimated, the total runoff from a site can be obtained
using an area-weighted calculation as shown below:

Runoffsite ={(Runoffroof ×Aroof)+(Runoffpavement ×Apavement)+(Runoffpervious ×Apervious)}/Asite

Where Runoffsite = total runoff from the site (inches); Asite = site area (acres); Runoffroof = runoff
from rooftop (inches); Aroof = rooftop area (acres); Runoffpavement = runoff from pavement area
(inches); Apavement = pavement area (acres); Runoffpervious = runoff from pervious area (inches);
and Apervious = pervious area (acres).

An example demonstrating how to calculate runoff by applying the Direct Determination method
is presented below using the Charleston, WV (Scenario #1) site condition presented in Tables 2
and 3.

Runoffroof = 95th Rainfall – Depression Storage
= 1.23 – 0.1 = 1.13 inches

Runoffpavement = 95th Rainfall – Depression Storage
= 1.23 – 0.1 = 1.13 inches

Runoffpervious = 95th Rainfall – Depression Storage – Infiltration Loss
= 1.23 – 0.1 – 9.73 = 0 inches (i.e., no runoff because the result is a
negative number)

Section 438 Technical Guidance December 2009

34
Runoffsite ={(Runoffroof ×Aroof)+(Runoffpavement ×Apavement)+(Runoffpervious ×Apervious)}/Asite
={(1.13 ×0.10)+(1.13×0.41)+(0 ×0.19)}/0.7 = 0.82 inches

Infiltration loss was estimated based on soil type B by applying the Horton equation as described
above. Because the volume removed from surface runoff through infiltration was substantial, no
runoff occurred from the pervious area.

In cases where sites had limited physical space available for stormwater management, a series of
practices was used (e.g., treatment train) to simulate the runoff and infiltrative behavior of the
system. For example, if there was inadequate area and infiltrative capacity to infiltrate 100
percent of the 95th percentile storm event within a bioretention system another onsite
management practice was selected to manage the runoff that could provide the necessary
capacity. In this manner, excess runoff was routed to another management practice in the series
of treatment cells where possible.

Two types of soils were considered for every site: hydrologic soil group B and C (except for
scenario 8 in which hydrologic soil group D was used). Group B soils typically have between 10
percent and 20 percent clay and 50 percent to 90 percent sand and either loamy sand or sandy
loam textures with some loam, silt loam, silt, or sandy clay loam soil textures placed in this
group if they are well aggregated, of low bulk density, or contain greater than 35 percent rock
fragments. Group C soils typically have between 20 percent and 40 percent clay and less than 50
percent sand and have loam, silt loam, sandy clay loam, clay loam, and silty clay loam soil
textures with some clay, silty clay, or sandy clay textures placed in this group if they are well
aggregated, of low bulk density, or contain greater than 35 percent rock fragments (USDA-
NRCS, 2007). The application of these hydrologic soil groups was intended to give reasonable
and somewhat conservative estimates of infiltration capacity.

General hydrologic parameters in this analysis were assumed as follows (see Appendix A for
citations of assumptions):

 Depression storage (or initial abstraction)
o Rooftop: 0.1 inches
o Pavement: 0.1 inches
o Pervious area: 0.2 inches

 Horton Infiltration parameters
o Hydrologic Soil Group B
 Maximum infiltration rate: 5 in/hr
 Minimum infiltration rate: 0.3 in/hr
 Decay factor: 2 /hr

o Hydrologic Soil Group C
 Maximum infiltration rate: 3 in/hr
 Minimum infiltration rate: 0.1 in/hr
 Decay factor: 3.5 /hr
Section 438 Technical Guidance December 2009

35

 Design storage assumptions of control measures
o Bioretention: up to 10 inches (but variable based on balancing necessary storage
volume, media depth for plant survivorship, and surface area limitations)
o Green roof: 1 inch (2.5 inches deep media with 40% void space)
o Porous pavement: 4 inches (10 inches deep media with 40% void space)

Other Methods for Estimating Runoff Volume
Runoff from a site after applying the 95th percentile storm can be estimated by using a number of
empirical, statistical, or mathematical methods. Several methods were considered in this
analysis. The Rational Method can be used to estimate peak discharge rates and the Modified
Rational Method can be used to develop a runoff hydrograph. The NRCS TR-55 model can be
used to predict runoff volume and peak discharge. TR-55 can also be used to develop a runoff
hydrograph. The EPA Stormwater Management Model (SWMM) can be used to simulate
rainfall-runoff, pollutant build-up and wash-off, transport-storage-treatment of stormwater flow
and pollutants, backwater effects, etc. for a wide range of temporal and spatial scales. The
SWMM model can be fit to model a small site with a distributed system. Hydrologic Simulation
Program – Fortran (HSPF, USDA) is a watershed and land use based lumped model that can be
used to compute the movement of water and pollutants when evaluating the effects of land use
change, reservoir operations, water quality control options, flow diversions, etc. In general,
regionally calibrated modeling parameters are incorporated into HSPF. QUALHYMO is a
complete hydrologic and water quality model, which can be used to factor in snowmelt or soil
moisture conditions or to simulate system behavior based on infiltration and ET, ground water
storage tracking, baseflow and deep volumetric losses, and other variables.

Many of the existing tools for analyzing distributed systems use some part or all of the principles
or formulae of the modeling approaches highlighted above. For example, the Emoryville
spreadsheet control measure model (Emoryville, CA) uses a runoff coefficient (i.e., Rational
Method) for analyzing lot-level to neighborhood-scale control measure sizing. The Green
Calculator (Center for Neighborhood Technologies) estimates the benefit of onsite GI/LID
options on a neighborhood-scale by applying the curve numbers (i.e., TR-55) and the Modified
Rational Method. The Northern Kentucky Spreadsheet Tool uses a TR-55 based approach for
control measure sizing on neighborhood or site level spatial scales. The WWHM (Western
Washington Hydrology Model) is a regionally calibrated HSPF model intended for use in sizing
stormwater detention and water quality facilities to meet the Washington State Department of
Ecology standards. WBM-QUALHYMO is a Canadian model used in conjunction with the
Water Balance Model (WBM). This model can be used to continuously simulate stormwater
storage routing, stream erosion, drainage area flow routing, and snowmelt runoff (and ultimately
freeze-thaw). Table 4 contains a summary of these different methods based on generic modeling
features.

Section 438 Technical Guidance December 2009

36

Table 4. Potential Methods for Analyzing Control Measures.
Model Considerations Rational Method TR-55 SWMM Direct Determination HSPF QUALHYMO
Temporal
scale
Single Event Yes Yes Yes Yes Yes Yes
Continuous
Simulation No No Yes Possible Yes Yes
Spatial
scale
Lot-level Yes Yesb Yes Yes No No
Neighborhood Yes Yes Yes Yes Possible Possible
Regional Yes Yesc Yes No Yes Yes
Outputs
Peak
Discharge Yes Yes Yes No Yes Yes
Runoff
Volume Yes Yes Yes Yes Yes Yes
Hydrograph Yesa Yes Yes No Yes Yes
Water Quality No No Yes Possible Yes Yes
a
Modified Rational Method
b
No less than 1 acre.
c
No more than 25 square miles (up to 10 subareas).

From the viewpoint of modeling both lot-level and neighborhood scale projects, the Rational
Method, NRCS TR-55, SWMM, and Direct Determination approaches were selected for use in
scenario analyses. Strength and weakness of these methods are presented below:

Table 5. Comparison of approaches for determining runoff volume.
Method Strengths Weaknesses
Direct
Determination



Methodology for runoff determination
is same as SWMM
Models basic hydrologic processes
directly (explicit)
Simple spreadsheet can be used


Direct application of Horton’s method may
estimate higher infiltration loss, especially at
the beginning of a storm
Does not consider flow routing

Rational
Method   Method is widely used Simple to use and understand  Cannot directly model storage-oriented onsite control measures
TR-55 
 Method is widely used Simple to use and understand
 May not be appropriate for estimating runoff
from small storm events because depression
storage is not well accounted for
SWMM

 Method is widely used Can provide complete hydrologic and
water quality process dynamics in
stormwater analysis


Needs a number of site-specific modeling
parameters
Generally requires more extensive experience
and modeling skills

Each method requires specific parameters for estimating runoff from a site. Runoff coefficients
for the Rational Method are assumed to be 0.9 for rooftop and pavement areas, and 0.1 and 0.135
for Group B and C soil pervious areas, respectively (Caltrans, 2003). The slope of the pervious
area was assumed to be an average of 2%. Applying these runoff coefficients for each surface,
the overall area-weighted runoff coefficient can be determined.

Section 438 Technical Guidance December 2009

37
When applying the NRCS TR-55 method, Curve Numbers (CNs) should be determined for each
drainage area. For rooftop and pavement areas the CN was assumed to be 98, and pervious area
CN was determined on the basis of the hydrologic soil group and the status of grass cover
condition. Curve numbers for pervious areas were assumed to be 61 and 74 for Group B and C
soils, respectively, with an assumption of over 75% grass cover. The overall CN can be
estimated by using an area-weighted calculation (USDA-SCS, 1986).

In SWMM modeling, infiltration was modeled using Horton’s equation. The same infiltration
parameters and depression storage values used in the direct determination method of runoff
treatment volume described earlier were applied to the SWMM analyses. The average slope of
the pervious area was again assumed to be 2%. The same uniform rainfall distribution and time
step was applied for the SWMM model runs.

Runoff Methodology Results
Stormwater management practice sizes (and depth) were determined using the Direct
Determination approach to capture the volume of runoff generated in a 95th percentile rainfall
event at each location. Total acreage, impervious area, the 95th percentile rainfall event, the
current expected runoff for the 95th percentile rainfall event, and the future runoff with
stormwater management controls were reported for each site. Results were summarized for the
two soil types (three soil types for scenarios #3 and #8 in Cincinnati). The spatial location of
onsite control measures was also illustrated in the site aerial photo figures. Note that site
practices were placed only on undeveloped or landscaped areas without regard for true flow
paths or technical feasibility. It may be preferred to place practices in existing impervious areas,
if possible. For the purposes of this modeling exercise, the least cost and most practical solutions
were used, i.e., locating bioretention systems on undeveloped or landscaped areas. On an actual
site, flow paths would be determined and berms and swales might be used to route runoff to
areas that are most suitable for infiltration. In other cases, areas that are currently impervious
could be modified to accept runoff, e.g., impermeable pavements removed and replaced by
permeable, sidewalks could be redesigned to include sidewalk bioretention cells and streets
could be designed with flow through or infiltration curb bumpouts/raingardens.

To compare other approaches of runoff estimation, alternate methodologies were also employed
for three scenarios. TR-55 was used for Scenario #1 (Atlanta), the Rational Method was applied
to Scenario #2 (Denver), and the SWMM was run for Scenario #7 (Charleston).

Although flood control is not the focus of this guidance, most localities have flood control
requirements that will need to be considered in designing control measures to comply with
Section 438. For flood control purposes, TR-55 was used to model the 10 year frequency design
storm for each site under the assumption that all stormwater management practices were in place.
The 10-year design storms were selected from the NRCS TR-55 Manual (USDA, 1986) for both
the Eastern U.S. and the Western U.S. Precipitation Frequency Maps
(www.wrcc.dri.edu/pcpnfreq.html). The 10-year frequency design storm was selected because it
represents a common design standard used by state and local governments in order to manage
peak rates of runoff and prevent flooding.

Section 438 Technical Guidance December 2009

38
COST ESTIMATES FOR SELECTED SCENARIOS
Scenarios #2 and 7 include cost estimates comparing the capital costs for a design to comply
with Section 438 (retention of the 95th percentile rainfall event) and capital costs for a traditional
stormwater management design (e.g., typical curb and gutter, off-site pond for stormwater
management). These costs are based on average unit costs to construct both traditional and
GI/LID controls.

Section 438 Technical Guidance December 2009

39
Scenario #1 - Charleston, WV
A 0.7-acre site with 73% impervious area was selected from Charleston, West Virginia (Figure
12). If the 95th percentile rainfall event (1.23 inches) occurred on the existing site (i.e., with no
control measures), 0.82 inches of runoff using the Direct Determination method would be
generated and require management. The runoff from the 95th percentile rainfall event could be
retained by the installation of bioretention systems totaling 0.03 acres if hydrologic soil group B
is present, or 0.06 acres if hydrologic soil group C (Table 6) is the predominant soil type on the
site. Assuming that bioretention practices are placed in areas that are currently pervious or
landscaped, a total of 0.2 acres of pervious area would be available for the placement of
bioretention systems. The effective design storage depth within the designated bioretention area
was 8 inches.

Figure 12. Actual Site and Onsite Control Measures (Charleston, WV)

Table 6. Estimated Sizes of Onsite Control Measures for Scenario #1 (Charleston, WV)
Total Area (acres) 0.7
Estimated Imperviousness (%) 73%
95th Percentile Rainfall Event (inches) 1.23
Expected Runoff for the 95th Percentile Rainfall Event (inches) 0.82
Stormwater Management Area Required

Hydrologic Soil Group
B C
Bioretention estimated by Direct Determination method (acres) 0.03 0.06
Bioretention estimated by SWMM (acres) 0.03 0.05
Off-site storage necessary to control the 10-yr
event of 3.9 inches (acre-ft) With onsite controls 0.10 0.12 Without onsite controls 0.16 0.17
Note: The two hydrologic methods used (direct determination and SWMM) estimated similar
bioretention sizes.
Section 438 Technical Guidance December 2009

40
Scenario #2 - Denver, CO
A 4.5-acre site with 55% impervious area was selected from Denver, Colorado (Figure 13). If
the 95th percentile rainfall event (1.07 inches) occurred on the existing site (i.e., with no control
measures), 0.53 inches of runoff from the site would be generated and require management. The
runoff from the 95th percentile rainfall event could be retained by the installation of bioretention
systems totaling 0.16 acres if the hydrologic soil group B is present or 0.3 acres if hydrologic soil
group C (Table 7) is the predominant soil type on the site. Assuming that bioretention practices
are only placed in areas that are currently pervious or landscaped, a total of 2 acres of pervious
area is available for the placement of bioretention systems. The design storage depth of media
within the designated bioretention area was 6 inches.


Figure 13. Actual Site and Onsite Control Measures (Denver, CO)

Table 7. Estimated Sizes of Onsite Control Measures for Scenario #2 (Denver, CO)
Total Area (acres) 4.5
Estimated Imperviousness (%) 55%
95th Percentile Rainfall Event (inches) 1.07
Expected Runoff for the 95th Percentile Rainfall Event (inches) 0.53
Stormwater Management Area Required

Hydrologic Soil Group
B C
Bioretention estimated by the Direct Determination method (acres) 0.16 0.3
Bioretention estimated by Rational Method (acres) 0.16 0.28
Off-site storage necessary to control the
10-yr event of 3.2 inches (acre-ft) With onsite controls Without onsite controls 0.35 0.64 0.52 0.64
Section 438 Technical Guidance December 2009

41
Cost estimates were also developed for this scenario (Table 8) to compare the costs of installing
onsite control measures to retain the 95th percentile rainfall event versus the costs to install
traditional stormwater management controls (e.g., curbs and gutters combined with off-site
retention such as extended detention wet ponds). In a GI/LID scenario, the bioretention cell
would occupy a specified area. This same area in a traditional design would be covered in turf
since the pond would typically be offsite and not occupy the area planted in turf. Table 8
includes this cost under the traditional column. Note: typical land development practices involve
mass clearing and grading so little or no pre-existing vegetation is typically retained. It is also
assumed that the use of GI/LID practices would require less underground infrastructure because
the traditional design typically routes stormwater underground to an off-site pond via pipes or
culverts while GI/LID practices are designed to manage runoff onsite and as close to its source as
possible. They are also dispersed across the site and routing occurs through surface drainage via
bioswales and overland flow. As a result GI/LID practices do not require as much or any hard or
grey infrastructure. The cost estimates were developed for Hydrologic Soil Group B.

Table 8. Estimated Costs for Scenario #2 (Denver, CO)
Sizes of Onsite Control Practices
Controls for 95th Percentile Event Traditional Stormwater Controls
Rainfall depth (in) 1.07
Bioretention (acres) 0.1
Paver blocks (acres) 0
Green roof (acres) 0
Off-site Pond WQV (ac-ft) - 0.18
10-Yr Fld Cntr (ac-ft) 0.15 0.14
Total Off-Site Requirement (ac-ft) 0.15 0.32
Land Area (assumes avg 3 ft depth) 0.05 0.11
% of the site 2.8%

Costs of Onsite Control Practices
Biorention/alternative $32,495 $4,187
Off-site Pond WQV (ac-ft) $14,833
10-Yr Fld Cntr (ac-ft) $10,073 $9,527
Infrastructure Pipe $8.990 $16,982
Inlet $9.920 $14,880
Land Area (assumes $300K/acre) $14,500 $31,500
Sum $75,978 $91,909
% difference from Traditional -17.3%












Section 438 Technical Guidance December 2009

42

Scenario #3 - Cincinnati, OH
A 19-acre site with 51% impervious area was selected in Cincinnati, Ohio (Figure 14). If the
95th percentile rainfall event (1.45 inches) occurred on the existing site (i.e., no control measures
were in place), 0.68 inches of runoff from the site would be generated and require management.
The runoff from the 95th percentile rainfall event could be retained by the installation of
bioretention systems totaling 0.8 acres if the hydrologic soil group B is present or 1.3 acres if
hydrologic soil group C (Table 9) is the predominant soil type on the site. Assuming that
bioretention practices are only placed in areas that are currently pervious or landscaped, a total of
9.4 acres of pervious area is available for the placement of bioretention systems. The design
storage depth of media within the designated bioretention area was 8 inches.


Figure 14. Actual Site and Onsite Control Measures (Cincinnati, OH)

Table 9. Estimated Sizes of Onsite Control Measures for Scenario #3 (Cincinnati, OH)
Total Area (acres) 19
Estimated Imperviousness (%) 51%
95th Percentile Rainfall Event (inches) 1.45
Expected Runoff for the 95th Percentile Rainfall Event (inches) 0.68
Stormwater Management Area Required

Hydrologic Soil Group
B C
Bioretention estimated by the Direct Determination (acres) 0.8 1.3
Off-site storage necessary to control the 10-yr
event of 4.2 inches (acre-ft) With onsite controls Without onsite controls 2.42 3.29 3.24 3.73
Section 438 Technical Guidance December 2009

Scenario #4 - Portland, OR
A 27-acre site with 95% impervious area was selected in Portland, Oregon (Figure 15). If the
95th percentile rainfall event (1.0 inches) occurred on the existing site (i.e., no control measures),
0.86 inches of runoff would be generated and require management. This site has the greatest
imperviousness among the 7 sites.

Given these site conditions, there is not enough pervious area to manage the entire runoff volume
discharged by the 95th percentile rainfall event with bioretention. As a result, other practices
were evaluated and selected. The practices integrated into the design included a green roof,
cisterns, and porous pavement. Based on the technical considerations of constructing and
maintaining control measures at the site, it was assumed that approximately 30% of the available
pervious area could be converted into bioretention cells; 20% of total rooftop area could be
converted into green roofs; 40% of paved area could be converted into paver blocks; and 50,000
gallons of total volume could be captured in cisterns for use on this urbanized site. Using this
system of four different practices, all runoff for the 95th percentile rainfall event would be
retained (Table 10).
43


Figure 15. Actual Site and Onsite Control Measures (Portland, OR)














Section 438 Technical Guidance December 2009

44
Table 10. Estimated Sizes of Onsite Control Measures for Scenario #4 (Portland, OR)
Total Area (acres) 27
Estimated Imperviousness (%) 95%
95th percentile Rainfall Event (inches) 1.00
Expected Runoff for the 95th Percentile Rainfall Event (inches) 0.86
Stormwater Management Area Required

Hydrologic Soil Group
B C
Paver block area estimated by Direct Determination (acres)
Bioretention estimated by Direct Determination (acres)
Green Roof estimated by Direct Determination (acres)
Cistern volume estimated by Direct Determination(gallons)
1.4 3.5*
0.4
1.7
50,000
Off-site storage necessary to control the 10-yr
event of 3.7 inches (acre-ft) With onsite controls 5.37 5.62 Without onsite controls 7.70 7.71
*The size of porous pavement area was increased because the other control options were maximized based on
the site-specific design assumptions.

A total of 1.3 acres of the site is pervious area or landscaped of which, 0.4 acres (30% of the
pervious area) could be converted to bioretention cells that have a storage depth of 10 inches. Of
the 8.8 acres of current rooftop area, 1.7 acres (20% of the rooftop area) could be retrofitted into
green roof areas. Of the 16.9 acres of paved area, 1.4 acres (8% of the paved area) for hydrologic
soil group B, or 3.5 acres (20% of the paved area) for hydrologic soil group C, of paver block
systems could be implemented. One or more cisterns (as indicated in Figure 15) could be used to
capture up to 50,000 gallons of runoff from rooftop areas. Note: The high percentage of
imperviousness of the site (95%) requires that all infiltration designs be based on resident soil
type and design volumes, or with adequate sub-bases or amended soils.





















Section 438 Technical Guidance December 2009

45
Scenario #5 – Near Phoenix, AZ
A 2-acre site with 47% impervious area was selected near Phoenix, Arizona (Figure 16). If the
95th percentile rainfall event (1.0 inches) occurred on the existing site (i.e., with no control
measures), 0.42 inches of runoff would be generated and require management. The runoff from
the 95th percentile rainfall event could be retained by installing bioretention systems totaling 0.06
acres if the hydrologic soil group B is present or 0.1 acres if hydrologic soil group C (Table 11)
is the predominant soil type on the site. Assuming that bioretention practices are only placed in
areas that are currently pervious or landscaped, a total of 1.1 acres of pervious area is available
for the placement of these practices. The design storage depth of media within the designated
bioretention area was 6 inches. Note: If the design storage depth were increased to 10 inches,
the off-site storage necessary for the 10-year event could be reduced to 0.03 acre-ft for type B
soils and 0.08 acre-ft for type C soils.


Figure 16. Actual Site and Onsite Control Measures (Phoenix, AZ)

Table 11. Estimated Sizes of Onsite Control Measures for Scenario #5 (Phoenix, AZ)
Total Area (acres) 2
Estimated Imperviousness (%) 47%
95th

Percentile Rainfall Event (inches) 1.00
Expected Runoff for the 95th Percentile Rainfall Event (inches) 0.42
Stormwater Management Area Required

Hydrologic Soil Group
B C
Bioretention estimated by the Direct Determination (acres) 0.06 0.1
Off-site storage necessary to control
the 10-yr event of 2.4 inches (acre-ft) With onsite controls Without onsite controls 0.05 0.18 0.12 0.18
Section 438 Technical Guidance December 2009

46
Scenario #6 - Boston, MA
A 3.5-acre site with 69% impervious area was selected in Boston, Massachusetts (Figure 17). If
the 95th percentile rainfall event (1.52 inches) occurred on the existing site (i.e., with no control
measures), 0.98 inches of runoff would be generated and require management. Given these site
characteristics, there is adequate area to place appropriately sized bioretention cells to capture the
95th percentile storm event. However, for the purposes of this analysis, unspecified conditions
preclude the use of bioretention. As a result, a paver block system was selected as the best onsite
control measure and the system was designed such that the necessary design parameters could be
achieved by storing some of the volume in the paver media and by infiltrating the remainder of
the volume. The runoff from the 95th percentile rainfall event could be retained by installing a
paver block area totaling 0.4 and 0.8 acres assuming soil types B and C, respectively (Table 12).
For the purposes of this case study, a total of 1.5 acres of parking lot was made available to
accommodate the paver block system. The area retrofitted with paver blocks would primarily be
dedicated for use as parking stalls.

Figure 17. Actual Site and Onsite Control Measures (Boston, MA)

Table 12. Estimated Sizes of Onsite Control Measures for Scenario #6 (Boston, MA)
Total Area (acres) 3.5
Estimated Imperviousness (%) 69%
95th Percentile Rainfall Event (inches) 1.52
Expected Runoff for the 95th Percentile Rainfall Event (inches) 0.98
Stormwater Management Area Required

Hydrologic Soil Group
B C
Paver block area estimated by Direct Determination (acres) 0.4 0.8
Off-site storage necessary to control
10-yr event of 4.5 inches (acre-ft) With onsite controls 0.59 0.71 Without onsite controls 0.89 0.96
Section 438 Technical Guidance December 2009

47
Scenario #7 - Atlanta, GA
A 21-acre site with 70% impervious area was selected in Atlanta, Georgia (Figure 18). If the
95th percentile rainfall event (1.77 inches) occurred on the existing site (i.e., with no control
measures), 1.17 inches of runoff would be generated and require management. The runoff from
the 95th percentile rainfall event could not be adequately retained solely with bioretention
systems. Based on the technical considerations of constructing and maintaining control
measures at the site, it was assumed that up to 15% of the pervious area could be converted into
bioretention cells and up to 40% of paved area could be converted into a paver block system. If
the stormwater management techniques used on the site includes both bioretention and paver
blocks as presented in Table 13, then all runoff for the 95th percentile rainfall event would be
controlled.

Figure 18. Actual Site and Onsite Control Measures (Atlanta, GA)

Table 13. Estimated Sizes of Onsite Control Measures for Scenario #7 (Atlanta, GA)
Total Area (acres) 21
Estimated Imperviousness (%) 70%
95th Percentile Rainfall Event (inches) 1.77
Expected Runoff for the 95th Percentile Rainfall Event (inches) 1.17
Stormwater Management Area Required

Hydrologic Soil Group
B C
Bioretention estimated by the Direct Determination (acres)
Paver block area estimated by the Direct Determination (acres)
0.9
0.9 3.2*
Bioretention estimated by TR-55
Paver block area estimated by TR-55
0.8** 0.9
0** 1.84
Off-site storage necessary to control
10-yr event of 6.0 inches (acre-ft) With onsite controls 5.85 6.62 Without onsite controls 7.25 8.49
*The size of porous pavement was increased because the bioretention already reached its
maximum size based on the site-specific design assumptions.
**Because TR-55 estimated smaller runoff in this scenario, bioretention can retain all of the 95th
percentile runoff if the site has soil group B.
Section 438 Technical Guidance December 2009

For the example site in Atlanta, GA, areas of 1.8 acres for hydrologic soil group B, and 4.1 acres
for hydrologic soil group C, would be required to manage the runoff discharged from a 95th
percentile rainfall event. Assuming that bioretention practices are only placed in areas that are
currently pervious or landscaped, a total of 6.2 acres of pervious area is available for the
placement of bioretention systems. The design storage depth of media within the designated
bioretention area was 10 inches. Permeable pavement systems could be used to treat the
remaining volume on the 10.8 acres of existing paved area.

In applying the TR-55 model, the overall curve numbers for the site were 87 and 91 for Group B
and C soils, respectively. TR-55 was used to estimate 0.73 inches of runoff for soil group B and
0.97 inches for soil group C, which are smaller numbers than the 1.17 inches of runoff estimated
by the Direct Determination method. As a result, the sizes of the onsite control measures
designed using the TR-55 model were smaller than those designed using the Direct
Determination method. Note: It is recommended that caution be exercised when using TR-55 to
model storms less than 0.5 inches per event. See application of TR-55 in Table 5.

Cost estimates were also developed for this scenario (Table 14) to compare the costs to install
onsite control measures to retain the 95th percentile rainfall event, and costs to install traditional
stormwater management controls (e.g., primarily curb and gutter with off-site retention). The
cost estimates were developed for Hydrologic Soil Group B.

48
Table 14. Estimated Costs for Scenario #7 (Atlanta, GA)
Sizes of Onsite Control Practices
Controls for 95th Percentile Event Traditional Stormwater Controls
Rainfall depth (in) 1.77
Bioretention (acres) 0.94
Paver blocks (acres) 0.86
Off-site Pond WQV (ac-ft) - 1.75
10-Yr Fld Cntr (ac-ft) 0.84 0.0
Total Off-Site Requirement (ac-ft) 0.84 1.75
Land Area (assumes avg 3 ft depth) 0.28 0.58
% of the site 8.5%

Costs of Onsite Control Practices
Biorention/alternative $232,923 $30,617
Paver block/alternative $236,878 $88,409
Off-site Pond WQV (ac-ft) $0 $72,888
10-Yr Fld Cntr (ac-ft) $39,648 $0
Infrastructure Pipe $54,827 $191,095
Inlet $52,080 $79,360
Land Area (assumes $300K/acre) $84,000 $175,000
Sum $700,356 $637,368
% difference from Traditional 9.9%






Section 438 Technical Guidance December 2009

49
Scenario #8 - Cincinnati, OH
A 19-acre site with 51% impervious area was selected in Cincinnati, Ohio (Figure 19). If the
95th percentile rainfall event (1.45 inches) occurred on the existing site (i.e., with no control
measures), 0.68 inches of runoff would be generated and require management. The runoff from
the 95th percentile rainfall event could be retained by the installation of bioretention systems
totaling 0.8 acres if the hydrologic soil group B is present or 1.3 acres if hydrologic soil group C
(Table 9) is the predominant soil type on the site. Assuming that bioretention practices are only
placed in areas that are currently pervious or landscaped, a total of 9.4 acres of pervious area is
available for the placement of bioretention systems. The design storage depth of media within
the designated bioretention area was 8 inches.

Scenario #8 represents an alternative to the Cincinnati, scenario in #3 (Figure 14). In this case,
hydrologic soil group D was selected to represent the soil characteristics present for the entire
site. Alternatively, simulations could have been run under the assumption that the use of
infiltration practices were precluded by contaminated soils or high ground water tables. Under
these site conditions, bioretention options are severely limited and cannot be used to adequately
capture the entire 95th percentile storm event. As a result, options such as cisterns and green
roofs were considered. In the absence of management practices, the 95th percentile rainfall event
discharges 1.45 inches of stormwater and 0.53 inches of this runoff is captured by onsite
depression storage. The difference, 0.92 inches of runoff, would then require capture and
management. Based on the technical considerations of constructing and maintaining controls at
the site, it was assumed that up to 20% of pervious area can be converted into bioretention areas;
up to 30% of paved area can be converted into porous pavement; and up to 30% of the rooftop
area can be converted into green roofs. Cisterns can be added to the system if additional storage
volume is required. It should be noted that green roofs were selected lowest in the hierarchy of
practices evaluated because of cost and potential structural issues associated with design and
placement on existing buildings. By using the four onsite control options as presented in Table
15, all runoff for the 95th percentile rainfall event would be retained. From a management
perspective, it was assumed that the design storage depth within the designated bioretention area
was 6 inches because of the low infiltration rates adopted for this scenario.

Section 438 Technical Guidance December 2009

50

Figure 19. Actual Site and Onsite Control Measures (Cincinnati, OH)

Table 15. Estimated Sizes of Onsite Control Measures for Scenario #8 (Cincinnati, OH)
Total Area (acres) 19
Estimated Imperviousness (%) 51%
95th Percentile Rainfall Event (inches) 1.45
Expected Runoff for the 95th Percentile Rainfall Event (inches) 0.92
Stormwater Management Applied
Bioretention estimated by Direct Determination (acres)
Paver block area estimated by Direct Determination (acres)
Green Roof estimated by Direct Determination (acres)
Cisterns estimated by Direct Determination(gallons)
Hydrologic Soil Group D
1.9
2.4
0.5
13,000

This site contains a total of 9.4 acres of pervious area, 8.0 acres of paved area, and 1.6 acres of
rooftop area. If 1.9 acres (20%) of the pervious area were converted to bioretention cells; 2.4
acres (30%) of parking lot converted to paver blocks; and 0.5 acres (30%) of rooftop area were
retrofitted to green roof areas for this site, then 97% of stormwater runoff from the 95th percentile
storm would be captured on site. By also adding one or more cisterns (as indicated in Figure 19),
an additional 13,000 gallons could be captured, thus illustrating that 100% of the rainfall from
the 95th percentile event can be managed onsite with GI/LID practices.

Section 438 Technical Guidance December 2009

51
Scenario #9 – Norfolk, VA
A 1.6 acre site with 91% impervious area was selected from Norfolk, Virginia. Table 16
contains the land use categories for the site. Figure 20 depicts the site and associated facilities.
Site specific factors based on an METF analysis allow management of 75 % of the 95th percentile
storm onsite (1.27 inches). The remaining portion of the 95th percentile rainfall event (0.41
inches would be discharged off of the site.

Table 16. Land Use Determination After Redevelopment
Land Use Acres Site Coverage Percent
Building 0.90 56.3
Parking 0.35 21.9
Streets/Sidewalks 0.20 12.5
Undeveloped 0.15 9.3
Total 1.60 100%


























Figure 20. Proposed Redevelopment Scenario





Section 438 Technical Guidance December 2009

52





















Figure 21. Location of Facility (Norfolk, VA)

Site conditions and intended uses limited the number of practices that were technically feasible
to use onsite to manage runoff. For example, the use of a green roof was not feasible because the
project includes the construction of an airplane hanger which lacks the structural strength to
support a green roof. Cisterns were also not included in the set of suitable practices based on the
analysis, which considered the number of people and amount of daily water use at the site, i.e.,
40 people x 3.5 toilet flushes per day would use only 280 gallons of runoff per day or 2,000
gallons per week. Stormwater use for HVAC make-up would also be negligible based on the
typical cooling system design. To put things in perspective, if the hanger rooftop covers the
entire building footprint, 41,000 gallons of runoff would be generated from a 1.68 inch rainfall.
Assuming a drawdown of 2,000 gallons per week based on toilet flushing, the users would only
use 5% of the 95th percentile event. Because of the relatively large volume of water that would
need to be collected and used, cisterns were not considered a feasible option to manage a
significant volume of runoff at the site.

However, site conditions did allow for the use of both permeable pavement and bioretention
practices. Approximately 0.15 acres (6,500 sf) of the proposed site is undeveloped and available
for bioretention. Based on Department of Defense facility requirements, ten percent of the
parking area is designed with landscaping, usually around the perimeter and in landscaped
islands. If this ten percent were designed as bioretention cells, then 0.035 acres of bioretention
would be achieved. If bioretention cells were also placed in about 30% of the undeveloped area
of the project, then an additional 0.045 acres of bioretention could be implemented. Note: not all
undeveloped land was assumed to be available for bioretention because of conflicts with site
Redevelopment Site
Section 438 Technical Guidance December 2009

53
utilities, security and anti-terrorism requirements and slopes that limited the use of infiltration
practices directly adjacent to the hanger.

Figure 22. Actual Site and Onsite Control Measures (Norfolk, VA)

Table 17. Estimated Sizes of Onsite Control Measures for Scenario #9 (Norfolk, VA)
Total Area (acres) 1.6
Estimated Imperviousness (%) 91%
95th Percentile Rainfall Event (inches) 1.68
Expected Runoff for the 95th Percentile Rainfall Event (inches) 1.50
Stormwater Management Area Required
Porous Pavement estimated by Direct Determination method (acres)
Bioretention estimated by Direct Determination method (acres)
Hydrologic Soil Group D
0.21
0.08

The bioretention cells were designed with an effective storage depth of 10 inches, which
included a depth from media surface to outlet of 10 inches. In this case study, state regulations
precluded the project from taking credit for the storage potential provided by the void space
within the bioretention cell media. Similarly, approximately 0.55 acres of the proposed site is
impervious due to parking lots, streets, and sidewalks. Due to manufacturer’s recommendations
that permeable pavement materials not be used in applications subject to heavy loads and
potential pollutant exposure the access roads and parking lot access isles were assumed to be
constructed from conventional impervious concrete or asphalt. Thus 60% of the parking area
(primarily parking stalls and sidewalks), which is about 38% of the entire paved area, is assumed
to be suitable for paver blocks. A high water table at the site limited the modeled net storage
depth under paver blocks placed in the parking areas and sidewalks to four inches. This storage
was calculated using the assumption that the pavement sub-base of 12 inches would have a
minimum void space of approximately 30%.

Section 438 Technical Guidance December 2009

54
COMPARISON OF THE RUNOFF ESTIMATION METHODS
As illustrated in each of the case studies above, runoff of the 95th percentile storm was estimated
in order to size onsite control measures. These estimates were produced by applying four
different methods: the Direct Determination method, the Rational Method, the NRCS TR-55, and
the EPA SWMM. The results comparing each of these methods for scenarios 1-7 are presented
in Table 18.

Table 18. Comparison of the estimated runoff (unit: inches)
Direct
Method Determination Rational Method TR-55 SWMM
Soil Groups B C B C B C B C
1 Charleston, WV 0.82 0.82 0.83 0.84 0.36 0.53 0.82 0.83
2 Denver, CO 0.53 0.53 0.57 0.59 0.12 0.26 0.53 0.53
3 Cincinnati, OH 0.68 0.68 0.73 0.76 0.26 0.46
4 Portland, OR 0.86 0.86 0.86 0.86 0.63 0.71
5 Phoenix, AZ 0.42 0.42 0.46 0.48 0.06 0.17
6 Boston, MA 0.98 0.98 0.99 1.00 0.51 0.70
7 Atlanta, GA 1.17 1.17 1.17 1.19 0.73 0.97 1.19 1.23

As shown in the above table, the estimated runoff results from direct determination, the Rational
Method, and SWMM are relatively similar. Runoff volumes using TR-55 are lower than the
other estimates. SWMM modeling results using NRCS 24-hour rainfall distributions were nearly
identical to the results based on uniform distribution.

Table 19. Applicability of the methods for analyzing onsite control measures
Direct
Purpose Determination Rational Method TR-55* SWMM
Planning Tool Applicable Applicable Applicable Applicable
Preliminary Design Applicable Applicable Applicable Applicable
Detailed Design Not applicable Not applicable Not applicable Applicable
Actual Assessment (Long-term) Not applicable Not applicable Not applicable Applicable
Water Quality Not applicable Not applicable Not applicable Applicable
*Use with caution when applying this method for small storms


CONCLUSIONS

Although sites varied in terms of climate and soil conditions, in most of the scenarios selected,
the 95th percentile storm event could be managed onsite with GI/LID systems. There are other
infiltration, evapotranspiration and capture and use stormwater management options available
than those used in these analyses. These options provide site managers additional flexibility to
choose appropriate systems and practices to manage site runoff.

Section 438 Technical Guidance December 2009

55
REFERENCES

Booth, Derek Direct Testimony. 2008. Pollution Control Hearings Board for the State of
Washington, Puget Soundkeeper Alliance and People for Puget Sound; Pierce County Public
Works and Utilities Department; City of Tacoma; The Port of Seattle; Snohomish County; Clark
County; and Pacificorp and Puget Sound Energy, Appellants, vs. Department of Ecology,
Respondent, and King County; City of Seattle; Port of Tacoma, and Washington State
Department of Transportation, Intervenors, August 2008.

California Department of Transportation (Caltrans). 2003. Caltrans Storm Water Quality
Handbooks. California Department of Transportation.

Casey Trees. 2007. The Case for Trees – Relief from Summer Heat,
www.caseytrees.org/resources/casefortrees.html.

Galli, John. 1991. Thermal Impacts Associated with Urbanization and Stormwater Best
Management Practices in Maryland, Anacostia Restoration Team for the Maryland Department
of the Environment, Washington, DC.

Grant, Gary, Luke Engleback and Barry Nicholson. 2003. Green Roofs: Their Existing Status
and Potential for Conserving Biodiversity in Urban Areas, Report Number 498, English Nature
Research Reports.

Hathaway, J., and W.F. Hunt. 2007. Stormwater BMP Costs. North Carolina Department of
Environment and Natural Resources.
www.bae.ncsu.edu/stormwater/PublicationFiles/DSWC.BMPcosts.2007.pdf.

Hirschman, David and John Kosco. 2008. Managing Stormwater in Your Community: A Guide
for Building an Effective Post-Construction Program, Center for Watershed Protection,
www.cwp.org/postconstruction.

Holz, Thomas Written Direct Testimony. 2008. Pollution Control Hearings Board for the State
of Washington, Puget Soundkeeper Alliance and People for Puget Sound; Pierce County Public
Works and Utilities Department; City of Tacoma; The Port of Seattle; Snohomish County; Clark
County; and Pacificorp and Puget Sound Energy, Appellants, vs. Department of Ecology,
Respondent, and King County; City of Seattle; Port of Tacoma, and Washington State
Department of Transportation, Intervenors, August 2008.

Horner, Richard Direct Testimony. 2008. Pollution Control Hearings Board for the State of
Washington, Puget Soundkeeper Alliance and People for Puget Sound; Pierce County Public
Works and Utilities Department; City of Tacoma; The Port of Seattle; Snohomish County; Clark
County; and Pacificorp and Puget Sound Energy, Appellants, vs. Department of Ecology,
Respondent, and King County; City of Seattle; Port of Tacoma, and Washington State
Department of Transportation, Intervenors, August 2008.

Section 438 Technical Guidance December 2009

56
National Climatic Data Center (NCDC). 2007. NCDC precipitation data, CD-ROM, National
Climatic Data Center.

National Research Council. 2008. Urban Stormwater Management in the United States, The
National Academies Press, Washington, DC.

Shaver, E., R. Horner, J. Skupien, C. May, and G. Ridley. 2007. Fundamentals of Urban Runoff
Management: Technical and Institutional Issues – 2nd Edition, North American Lake
Management Society, Madison, WI.

Schueler, T. and M. Helfrich. 1988. “Design of Wet Extended Detention Pond Systems,” Design
of Urban Runoff Controls, L. Roesner and B. Urbonas eds., American Society of Civil
Engineers, New York, NY.

Schueler, T and H. Holland. 2000. The Practice of Watershed Protection: Techniques for
Protecting our Nation’s Streams, Lakes, Rivers, and Estuaries. Center for Watershed Protection,
Ellicott City, MD.

U.S. Department of Agriculture, Natural Resources Conservation Service. 2007. National
Engineering Handbook, title 210–VI. Part 630, chapter 7. Washington, DC.
http://directives.sc.egov.usda.gov/.

U.S. Department of Agriculture, Soil Conservation Service. 1986. Urban Hydrology for Small
Watersheds. Technical Release No. 55. Second Edition. Washington, D.C.

U.S. Environmental Protection Agency, Managing Wet Weather with Green Infrastructure,
www.epa.gov/greeninfrastructure.

Vingarzan and Taylor. 2003. Trend Analysis of Ground Level Ozone in the Greater Vancouver/
Fraser Valley Area of British Columbia, Environment Canada – Aquatic and Atmospheric
Sciences Division.

Wisconsin DNR. 2008. Impact of Redevelopment on TSS Loads, Runoff Management, available
at http://www.dnr.state.wi.us/runoff/pdf/rules/nr151/Impact_of_RedevTSSLoads_021308.pdf.


Section 438 Technical Guidance December 2009

57
APPENDIX A: Runoff Methodology Parameter Assumptions

Runoff from each land cover was estimated by the following equation:

Runoff = Rainfall – Depression Storage – Infiltration Loss (1)

Depression Storage
Reference depression storage (inches)

Reference Impervious Pervious
1 0.05 - 0.1 0.1 - 0.3
2 0.01 - 0.11 0.02 - 0.6
3 0.1 0.2

1. ASCE, (1992). Design & Construction of Urban Stormwater Management Systems. New York, NY.
2. Marsaleck, J., Jimenez-Cisreros, B., Karamouz, M., Malmquist, P-R., Goldenfum, J., and Chocat, B.
(2007). Urban Water Cycle Processes and Interactions. Urban Water Series, UNESCO-IHP, Tyler &
Francis.
3. Walesh, S. G. (1989).Urban Surface Water Management. John Wiley & Sons, Inc.

Based on the above reference data, depression storage (or initial abstraction, the rainfall required
for the initiation of runoff) to the direct determination method was assumed as follows:
 Rooftop: 0.1 inches
 Pavement: 0.1 inches
 Pervious area: 0.2 inches

Infiltration
Infiltration loss occurs only in pervious areas. In this analysis, infiltration was estimated by
Horton’s equation:

Ft = fmin + (f - k tmax – fmin) e (2)

where, Ft = infiltration rate at time t (in/hr),
fmin = minimum or saturated infiltration rate (in/hr),
fmax = maximum or initial infiltration rate (in/hr),
k = infiltration rate decay factor (/hr), and
t = time (hr) measured from time runoff first discharged into infiltration area

Reference infiltration parameters
Maximum infiltration rate (in.hr), fmax
Partially dried out with Dry soils with Infiltration
(in/hr) No vegetation Dense vegetation No vegetation Dense vegetation
Sandy 2.5 5 5 10
Loam 1.5 3 3 6
Clay 0.5 1 1 2
Reference: Huber, W. C. and Dickinson, R. (1988). Storm Water Management Model User’s Manual,
Version 4. EPA/600/3-88/001a (NTIS PB88-236641/AS), U.S. Environmental Protection Agency,
Athens, GA.
Section 438 Technical Guidance December 2009

58

Minimum infiltration rate (in/hr), fmin
Hydrologic
Soil Group Infiltration (in/hr)
A 0.45 - 0.30
B 0.30 - 0.15
C 0.15 - 0.05
D 0.05 - 0
A: well drained sandy; D: poorly drained clay
Reference: Huber, W. C. and Dickinson, R. (1988). Storm Water Management Model User’s Manual,
Version 4. EPA/600/3-88/001a (NTIS PB88-236641/AS), U.S. Environmental Protection Agency,
Athens, GA.

Decay coefficient, k
Soils k (sec-1) k (hr-1)
0.00056 2
0.00083 3
0.00115 4
Sandy


Clay 0.00139 5
Reference: Huber, W. C. and Dickinson, R. (1988). Storm Water Management Model User’s Manual,
Version 4. EPA/600/3-88/001a (NTIS PB88-236641/AS), U.S. Environmental Protection Agency,
Athens, GA.

Based on the above reference data, infiltration parameters to the direct determination method
were assumed as follows:
 Hydrologic Soil Group B
 Maximum infiltration rate: 5 in/hr
 Minimum infiltration rate: 0.3 in/hr
 Decay factor: 2 /hr
 Hydrologic Soil Group C
 Maximum infiltration rate: 3 in/hr
 Minimum infiltration rate: 0.1 in/hr
 Decay factor: 3.5 /hr
 Hydrologic Soil Group D
 Maximum infiltration rate: 1 in/hr
 Minimum infiltration rate: 0.02 in/hr
 Decay factor: 5 /hr

Infiltration loss for the 24-hr rainfall duration was estimated by the following equations with
assumptions of a half hour ∆t:

Infiltration Loss at the nth time-step = (f ·∆t) = })2/){( 1 tff nn  (3)
Integrated Infiltration Loss for 24 hours = ∑ (f ·∆t) (4)

Section 438 Technical Guidance December 2009

59
Integrating infiltration loss during 24 hours with a half hour ∆t
Infiltration rate (in/hr) a Infiltration volume (inches) b time-
step t (hr) Soil B Soil C Soil D Soil B Soil C Soil D
0 0 5 3 1 0 0 0
1 0.5 2.03 0.60 0.100 1.757 0.901 0.275
2 1 0.94 0.19 0.027 0.741 0.198 0.032
3 1.5 0.53 0.12 0.021 0.368 0.076 0.012
4 2 0.39 0.10 0.02 0.230 0.054 0.01
5 2.5 0.33 0.1 0.02 0.179 0.05 0.01
6 3 0.31 0.1 0.02 0.161 0.05 0.01
7 3.5 0.30 0.1 0.02 0.154 0.05 0.01
8 4 0.3 0.1 0.02 0.15 0.05 0.01
9 4.5 0.3 0.1 0.02 0.15 0.05 0.01
10 5 0.3 0.1 0.02 0.15 0.05 0.01
11 5.5 0.3 0.1 0.02 0.15 0.05 0.01
12 6 0.3 0.1 0.02 0.15 0.05 0.01
13 6.5 0.3 0.1 0.02 0.15 0.05 0.01
14 7 0.3 0.1 0.02 0.15 0.05 0.01
15 7.5 0.3 0.1 0.02 0.15 0.05 0.01
16 8 0.3 0.1 0.02 0.15 0.05 0.01
17 8.5 0.3 0.1 0.02 0.15 0.05 0.01
18 9 0.3 0.1 0.02 0.15 0.05 0.01
19 9.5 0.3 0.1 0.02 0.15 0.05 0.01
20 10 0.3 0.1 0.02 0.15 0.05 0.01
21 10.5 0.3 0.1 0.02 0.15 0.05 0.01
22 11 0.3 0.1 0.02 0.15 0.05 0.01
23 11.5 0.3 0.1 0.02 0.15 0.05 0.01
24 12 0.3 0.1 0.02 0.15 0.05 0.01
25 12.5 0.3 0.1 0.02 0.15 0.05 0.01
26 13 0.3 0.1 0.02 0.15 0.05 0.01
27 13.5 0.3 0.1 0.02 0.15 0.05 0.01
28 14 0.3 0.1 0.02 0.15 0.05 0.01
29 14.5 0.3 0.1 0.02 0.15 0.05 0.01
30 15 0.3 0.1 0.02 0.15 0.05 0.01
31 15.5 0.3 0.1 0.02 0.15 0.05 0.01
32 16 0.3 0.1 0.02 0.15 0.05 0.01
33 16.5 0.3 0.1 0.02 0.15 0.05 0.01
34 17 0.3 0.1 0.02 0.15 0.05 0.01
35 17.5 0.3 0.1 0.02 0.15 0.05 0.01
36 18 0.3 0.1 0.02 0.15 0.05 0.01
37 18.5 0.3 0.1 0.02 0.15 0.05 0.01
38 19 0.3 0.1 0.02 0.15 0.05 0.01
39 19.5 0.3 0.1 0.02 0.15 0.05 0.01
40 20 0.3 0.1 0.02 0.15 0.05 0.01
41 20.5 0.3 0.1 0.02 0.15 0.05 0.01
42 21 0.3 0.1 0.02 0.15 0.05 0.01
43 21.5 0.3 0.1 0.02 0.15 0.05 0.01
44 22 0.3 0.1 0.02 0.15 0.05 0.01
Section 438 Technical Guidance December 2009

60
45 22.5 0.3 0.1 0.02 0.15 0.05 0.01
46 23 0.3 0.1 0.02 0.15 0.05 0.01
47 23.5 0.3 0.1 0.02 0.15 0.05 0.01
48 24 0.3 0.1 0.02 0.15 0.05 0.01
Sum: Infiltration loss during 24 hours c 9.743 3.430 0.769
a
Calculated infiltration rate at each time by Equation (2)
b
Calculated infiltration volume from the previous time to the current time by Equation (3)
c
Integrated infiltration volume for 24 hours with a half hour ∆t by Equation (4)

Based on the above calculation, 24-hr infiltration losses for pervious areas and bioretention areas
were modeled as follows:
 Soil Group B: 9.743 inches
 Soil Group C: 4.430 inches
 Soil Group D: 0.769 inches

Infiltrations of underlying soils at paver blocks were modeled conservatively by applying the
minimum infiltration rate for each soil type (Infiltration loss = fmin · 24) because the soils under
the paver blocks may require or be subjected to some compaction for engineering stability. The
estimated infiltration losses for each soil are presented below:
 Soil Group B: (0.3 in/hr) · (24 hrs) = 7.2 inches
 Soil Group C: (0.1 in/hr) · (24 hrs) = 2.4 inches
 Soil Group D: (0.02 in/hr) · (24 hrs) = 0.48 inches

Design Storage of Management Practices

Bioretention
Reference Ponding
(inches)1 Mulch (inches) Soil media (ft) Soil Media Porosity Underdrain
1 up to 12 2 - 4
(optional) 1 - 1.5 about 40% bioretention systems utilize infiltration rather than an underdrain
2 6 - 12 2 - 3 2.5 - 4 about 40% recommended, especially if initial
testing infiltration rate < 0.52 in/hr
3 6 - 12 2 - 4
4 2 - 3 1.5 - 4 if necessary
5 up to 6 1.5 - 2 30 - 40% Optional
6 6 - 18 as
needed 2 - 4 if necessary

1. State of New Jersey. (2004). New Jersey Stormwater Best Management Practices Manual
www.nj.gov/dep/stormwater/tier_A/pdf/NJ_SWBMP_9.1 print.pdf.
2. Maryland Department of the Environment (MDE), (2000). 2000 Maryland Stormwater Design Manual,
Volumes I & II, prepared by the Center for Watershed Protection and the Maryland Department of the
Environment, Water Management Administration, Baltimore, MD.
www.mde.state.md.us/Programs/WaterPrograms/SedimentandStormwater/stormwater_design/index.asp.

1
Ponding is a measure of retention capacity
Section 438 Technical Guidance December 2009

61
3. Clar, M. L. and R. Green, (1993). Design Manual for Use of Bioretention in Storm Water Management,
prepared for the Department of Environmental Resources, Watershed Protection Branch, Prince
George's County, MD, prepared by Engineering Technologies Associates, Inc. Ellicott City, MD, and
Biohabitats, Inc., Towson, MD.
4. U.S. Environmental Protection Agency. (1999). Storm Water Technology Fact Sheet: Bioretention.
EPA 832-F-99-012. Office of Water. US Environmental Protection Agency. Washington, D.C.
www.epa.gov/owm/mtb/biortn.pdf.
5. Prince George's County. Bioretention Design Specifications and Criteria. Prince George's County,
Maryland.
www.co.pg.md.us/Government/AgencyIndex/DER/ESG/Bioretention/pdf/bioretention_design_manual
.pdf.
6. City of Indianapolis. (2008). Indianapolis Stormwater Design Manual.
www.sustainindy.org/assets/uploads/4_05_Bioretention.pdf.

Paver Blocks
Reference Media (inches) Void Space
1 12 or more 40%
2 9 or more 40%
3 12 - 36 40%

1. Univ. of California at Davis. (2008). Low Impact Development Techniques: Pervious Pavement.
http://extension.ucdavis.edu/unit/center_for_water_and_land_use/pervious_pavement.asp.
2. AMEC Earth and Environmental, Center for Watershed Protection, Debo and Associates, Jordan Jones
and Goulding, and Atlanta Regional Commission. (2001). Georgia Stormwater Management Manual
Volume 2: Technical Handbook www.georgiastormwater.com/.
3. Subsurface Infiltration Bed. www.tredyffrin.org/pdf/publicworks/CH2 - BMP4 Infiltration Bed.pdf.

Green Roofs
Reference Media (inches)
1 3 - 4
2 1 - 6
3 2 - 6

1. Charlie Miller. (2008). Extensive Green Roofs. Whole Building Design Guide (WBDG).
www.wbdg.org/resources/greenroofs.php.
2. Great Lakes WATER Institute. Green Roof Project: Green Roof Installation.
www.glwi.uwm.edu/research/genomics/ecoli/greenroof/roofinstall.php.
3. Paladino & Company. (2004). Green Roof Feasibility Review. King County Office Project.
http://your.kingcounty.gov/solidwaste/greenbuilding/documents/KCGreenRoofStudy_Final.pdf.

Based on the above reference data, design storages to the direct determination method were
assumed as follows:
 Bioretention: up to 10 inches (depending on practice used, site conditions, etc.)
 Green roof: 1 inch (2.5 inches deep media with 40% void space)
 Porous pavement: 4 inches (10 inches deep media with 40% void space)

Factors that influence total storage available include, ponding depth, available media void space,
and supplemental storage if the system is designed with gravel or open pipes underneath the
media.