How to choose better stormwater BMPs


    Tools backed by reliable scientific studies are emerging to help civil engineers.

    Which best management practices (BMPs) are most effective? How does one BMP compare to another in controlling specific pollutants? Which BMP should I choose to protect my local waterbody? Is this BMP going to be more or less cost-effective over the long term?

    These are all important questions that stormwater professionals have been debating and discussing for a long time. Not too many years ago, it was relatively easy to find opinions on these topics, but much more difficult to find hard facts. Today, the science of stormwater management has advanced to the point where it can at least begin to answer some of these questions and, ultimately, help us choose better BMPs. (This article focuses on the performance of permanent, or post-construction, BMPs that treat or control urban stormwater.)

    So, who really needs to know about BMP performance? The obvious answer is: “Everyone involved in stormwater management”—to some degree. Information about BMP performance is critical to those who are making day-to-day decisions about the stormwater BMPs that are being designed and installed. This includes developers and builders who are responsible for designing stormwater management systems and the municipal or county officials who review these designs.

    Of course, many, if not most, of the people developing stormwater designs and reviewing plans are engineers. Engineers play an often underappreciated role in stormwater management and need access to the latest information on BMP performance to choose the most effective BMP for their site. Beyond these “key constituents,” there are many others who have an interest in BMP performance, including NPDES permit writers, watershed planners, total maximum daily load (TMDL) writers, researchers, and many others.

    Performance factors
    Defining the performance of BMPs is not an exact science. BMPs, unlike other treatment technologies—such as those used in wastewater treatment systems—are not static systems that deliver constant or even predictable results. The performance of stormwater BMPs is significantly influenced by such factors as design, installation, rainfall patterns and characteristics, soil types, slopes, land use in the drainage area, and age of the system. Figure 1 illustrates the variability of BMP performance in removing total suspended solids. Keeping this in mind, one important piece of advice we can offer is: Don’t latch on to one number and expect a stormwater BMP to deliver that level of performance each and every time.

    Figure 1: BMPs are not static systems that deliver constant or even predictable results. Source: International BMP Database


    While we can’t now, and might never, be able to predict BMP performance with 100-percent accuracy, we do have information—as a result of the diligent efforts of hundreds of researchers and practitioners—that will help detect trends in BMP performance.

    Percent removal: The stormwater urban myth
    Spend any time at all discussing BMP performance or looking at articles and studies and you will soon encounter the concept of percent removal. This widely (mis)used concept, which is also called percent effectiveness and other names, uses an influent and effluent concentration to obtain a simple percentage.

    Experts in the field of BMP performance, including those at EPA, find this measure more misleading than helpful. Among the many reasons to avoid this measure is that by relying on just one percentage value, we lose the context of the conditions from which it was derived. For instance, the influent in one study may have been particularly dirty, in another, significantly less so. Looking at the percentage alone does not provide that information. In fact, percent removals are highly influenced by how dirty the site (influent) is and don’t tell much about how well the BMP worked. An excellent article on this subject, “15 Reasons You Should Think Twice Before Using Percent Removal to Assess BMP Performance,” appeared in the January/February 2008 issue of Stormwater magazine (

    Total load reduction
    Rather than relying on a percentage to describe BMP performance, you should consider several factors, including event mean concentrations (EMCs), volume reduction information (see below), and total load reduction. One of the better ways to compare the performance of BMPs across types is to compare total load reduction. A load-reduction measure takes into account the volume of stormwater both coming into and leaving the BMP, as well as the concentration of pollutants. Using a load-reduction measure (usually pounds or kilograms) becomes particularly relevant when examining BMPs that reduce the volume of stormwater.

    Figure 2 illustrates the concept of examining pollutant reduction potential by looking at a load-reduction measure. In this example, calculating a simple percent removal using the concentration numbers (as shown in the beakers) results in a 50-percent reduction of the pollutant, while using the total load results in a 75-percent reduction. The 50-percent removal number is misleading because it does not account for the large volume of stormwater infiltrated by the BMP. It is only when we look at total load reduction that we see the true performance of the BMP.

    Figure 2: Illustration of total load reduction


    Importance of volume reduction
    As we learn more about the relationship between urban and suburban development and watershed health, it becomes clear that one of the major factors affecting the condition of our rivers, lakes, coastal waters, and wetlands is the increased volume and velocity of stormwater runoff reaching these water bodies. As we add impervious surfaces—including roads, driveways, homes, shopping malls, and all of the other facets of urbanization—to our watersheds, we create changes in how these watersheds function hydrologically. Increases in impervious surfaces result in increases in stormwater flowing off the land and corresponding reductions in the rain and snow melt that would otherwise soak into the ground. These changes can have profound impacts on many of our waters, particularly sensitive headwater streams, wetlands, and small lakes. Much has been written elsewhere on this topic.

    The key message is that when planning and designing systems to manage urban stormwater, we need to weigh options that encourage infiltration of stormwater and help to maintain or restore the natural hydrologic balance of the land that is being developed. Another important point to keep in mind (clearly seen in Figure 2) is that when we choose stormwater BMPs that reduce the volume of stormwater, we also reduce the load of pollutants entering surface waters.

    A watershed context
    When selecting stormwater BMPs, perhaps the most important consideration is the condition and characteristics of the watershed in which they will be placed. View decision-making criteria (pollutant reduction, volume reduction, costs, maintenance, et cetera) in the context of the condition of the watershed that we are trying to protect or restore. Unfortunately, there are too many examples of “inappropriate” BMPs being diligently designed and installed with limited benefit to the watershed. Sometimes these inappropriate BMPs can actually cause water quality to decline rather than improve.

    Before beginning the stormwater BMP selection and planning process, gather some basic facts about the watershed in which you will be working. Information about water quality standards and designated uses, historical monitoring data, existing impairments and water quality problems, and any existing or planned TMDLs can help create a picture of the issues facing a particular river, lake, wetland, or coastal water and help develop and refine the criteria used to select BMPs.

    Other factors
    Pollutant and volume reduction potential are two key considerations always to keep in mind when selecting or approving stormwater BMPs. However, many other considerations also have a role in BMP selection. Costs are always a significant factor. Total capital costs (meaning cost of installation), as well as factors such as cost per acre treated, can provide valuable information. Another cost-related factor that is frequently overlooked is the cost of maintenance of the BMP over the long term. All BMPs require regular, and sometimes frequent, maintenance. Considering ease of maintenance and overall maintenance costs should help inform our choices.

    Other factors to consider include the function of the BMP in regional drainage and flood management efforts and whether the BMP has secondary benefits, including habitat values, carbon sequestration, and water conservation, among others. Aesthetics and public acceptance are also very important, as is the useful life of the BMP. Of course, everyone involved in stormwater management also needs to consider and mitigate any potential safety issues.

    The Center for Watershed Protection (with funding from EPA) recently published a valuable manual called Urban Stormwater Retrofit Practices. While this manual focuses on retrofitting existing development with better stormwater BMPs, the concepts, decision-making structures, and performance and cost information is useful for anyone selecting stormwater BMPs, whether in a new development or redevelopment context. (The manual is available for free download at

    Urban BMP Performance Tool
    So, how do you select a “better” BMP for your site? In January 2008, EPA released the Urban Stormwater BMP Performance Tool (available at that presents BMP performance data from more than 220 scientific studies. This tool contains many of the studies collected by the International Stormwater BMP Database ( and allows a user to search BMP performance data by pollutant of interest, by type of BMP, or by volume of stormwater reduction.

    The Urban BMP Performance Tool results are displayed showing both the influent and effluent concentrations of each BMP, with the effluent concentrations sorted from low to high. Volume reduction information is also displayed, where available. The can be used in many ways. Two common approaches to using the Urban BMP Performance Tool include the following:

    • finding the best BMP for controlling a specific pollutant, and
    • determining how a specific BMP performs across a range of pollutants and parameters.

    Users who are concerned about a specific pollutant—for example, those who are in an impaired watershed with a TMDL—should select “search studies by the pollutants that were measured.” This allows users to identify a specific pollutant or pollutant group and see the most effective BMPs. Users who are interested in how effective a particular BMP is in controlling different pollutants should select “search studies by the BMPs examined.” Users then select the BMP and can further narrow their search by selecting an appropriate pollutant group.

    This tool provides valuable information to help users select a better BMP for their site. As discussed above, however, there are many factors that go into proper BMP selection; therefore users should not just select the first BMP on the list. EPA will continue to update the Urban BMP Performance Tool and add additional studies and BMPs as more data becomes available.

    Choosing effective BMPs is a significant challenge for all of us in the stormwater community. As discussed in this article, there are still many unknowns, and the many considerations that we need to balance can make the task challenging. Fortunately, tools, backed by reliable scientific studies, are emerging that will help improve the process of choosing better BMPs.

    Nikos Singelis is a senior program analyst for the U.S. Environmental Protection Agency. He can be reached at John Kosco is a principal engineer with Tetra Tech, Inc. He can be reach at


    Terms and phrases used to describe BMP performance can be confusing and sometimes even misleading. The list below includes some of the key terminology used in this article. Event Mean Concentration (EMC), volume, and total load are the key measures we rely on to help describe the performance of BMPs.

    concentration—influent and effluent concentrations, usually expressed as mg/L or ug/L

    event mean concentration (EMC)—total pollutant load in a runoff event divided by the volume of that event. In real world conditions, this is difficult or impossible to measure directly, so it is commonly estimated using flow-weighted (automatic) sampling

    volume—in liters, coming into (influent) and out of (effluent) a BMP

    percent removal—using an influent and effluent concentration to obtain a simple percentage (not recommended!)

    total load—a calculation of the amount of a pollutant by weight (e.g., pounds or kilograms) coming into and out of a BMP during a storm event or series of events (a better overall measure)


    Key references


    EPA Stormwater Webcast Series (archived webcasts are available free at

    • BMP Performance, January 2008
    • The Art and Science of Stormwater Retrofitting, March 2008

    Manuals and articles

    • Impacts of Impervious Cover on Aquatic Systems, CWP, March 2003,
    • Better Site Design: A Handbook for Changing Development Rules in Your Community, CWP, August 1998,
    • Urban Stormwater Retrofit Practices, CWP, August 2007,
    • “15 Reasons You Should Think Twice Before Using Percent Removal to Assess BMP Performance,” Jones, Clary, Strecker, and Quigley, Stormwater, January/February 2008,