Still a mainstay of many stormwater programs, treating first-flush runoff might not be enough.

By Mary Catherine Hager

"First flush" is the runoff that occurs at the beginning of a rainstorm. Generally thought to be more pronounced on impervious surfaces, the first flush carries with it concentrations of pollutants that have accumulated during the period of dry weather between storms, which could be one day or several months. Communities often struggle to adequately define first flush, such as what volume of rain it constitutes and whether or not it is affected by rainfall frequency or intensity, and to provide adequate treatment measures to counter it. First-flush concerns often figure prominently as smaller cities and counties work toward gaining compliance with Phase II of the National Pollutant Discharge Elimination System (NPDES), and with meeting EPA’s total maximum daily load (TMDL) requirements as specified by each state. Communities vary considerably in how they define first flush and how they treat it.

Scott McClelland, a vice president of the international consulting, engineering, and construction firm Camp, Dresser & McKee, describes the phenomenon of first flush: "When things are dry, pollutants tend to store up on land, both pervious and impervious areas. Then when it rains even a little bit, pollutants are entrained and carried off the land, the majority in the first portion of the storm event." McClelland, who is based in Florida, works directly with cities, counties, and states on planning strategies to address their environmental issues. He specializes in stormwater planning and management, including master plans and financial areas of stormwater. In Florida, where it rains about 125 times each year, McClelland says it is important to consider those rainfall events "that occur 90% of the time–generally small, less than an inch [accumulation]." To deal with stormwater-quality issues, McClelland stresses the importance of controlling the first half-inch or inch of rainfall, which occurs in the smaller storm events within the first half-hour. Components of first flush that are particularly easy to visualize are car and truck engine greases and oils that accumulate on roadways. These become major sources of stormwater pollution if they are allowed to flush into surface waters.

Detention ponds hold water a brief amount of time to allow biological, chemical, or physical treatment.

A conveyance pipe routes stormwater from the roof to a drainage area.

Porous pavement helps reduce runoff volume.

McClelland believes that for Florida, measuring first flush is no longer a critical issue because the benefits of controlling first flush for the frequent smaller storms the state receives have already been convincingly documented. McClelland’s confidence stems partly from Florida’s extensive work in the areas of stormwater treatment–it is one of only six or seven states with statewide stormwater regulation–and in its pioneering efforts to define and treat first flush. "Controlling pollution is very important in the state of Florida," McClelland explains, "because it has an environment that is very sensitive to pollution. It’s very easy to tip it over the edge." McClelland believes the focus on defining first flush is shifting to treating it: "I think we’ve gone beyond needing to know if first flush is an issue. We’ve come to a point where we need to know more about the effectiveness of the types of best management practices [BMPs] we use."

McClelland’s sentiments are echoed by a fellow Floridian, Eric Livingston, chief of the Florida Department of Environmental Protection (DEP) Bureau of Watershed Management. Involved since 1978 in the development, implementation, and evolution of Florida’s nonpoint source and watershed management programs, essentially since their beginning, Livingston agrees that Florida has collected enough monitoring data to be confident of its stormwater database. "We have enough data to feel very comfortable that we know what the pollutants are, what the loads are." But Livingston cautions that other states must collect their own data, because environmental considerations and pollutant loads vary considerably between locations: "You have to do your own data collection to come up with your own BMP design criteria that achieve a certain level of pollutant load reduction."

Livingston recounts the history of the first-flush concept, citing its origins in Florida research that stemmed from the Clean Water Act section 208 program, the first nonpoint source program, in the mid-1970s. Because there was very little knowledge about stormwater at that time, Livingston explains, most of the early monitoring attempted to characterize different kinds of runoff from different land uses. "[We conducted] discrete monitoring over a hydrograph, a series of samples as the flow increased and decreased throughout the storm, so you could see what was happening with concentrations, and you could measure flow and take a look at total loads." One of the trends apparent in the data was the occurrence of the first flush, which was factored into designing BMP practices once stormwater treatment became required in Florida as of 1979. With so many small storms occurring in Florida annually, most of them 1 in. or less rainfall, the apparent documentation of first flush led to the concept that "you really only need to capture that small amount of the first storms that come along." Livingston now says this was too simplistic a view, and even in the early days of monitoring, first flush did not hold for drainage basins larger than 100 ac.

Livingston carries his doubts as far as to state, "I don’t think you can talk about first flush as a general characteristic of stormwater anymore." Many factors influence the occurrence and impact of a first-flush event. "First flush depends on a number of site characteristics: the time of concentration of the basin, the imperviousness of the basin, the kind of stormwater routing in the basin, [and] the pollutants of concern. Are they attached to particles? How do they move? How do they exert their impact on the environment?" Larger sites experience greater times of concentration–the time it takes for flow to get from one point in the drainage basin to another–and receive more sources coming in. In addition, larger, more complex sites might have natural mechanisms or depressional storage areas, such as wetlands and floodplains, that cut down on the runoff. In these cases, explains Livingston, "[first flush] just gets hidden by the myriad things going on." Very small sites with large impervious areas, such as office complexes on large parking lots, definitely exhibit a first flush, but otherwise, Livingston finds the trend to be very site-specific. In terms of monitoring, he supports the idea of characterizing a site through discrete sampling over a hydrograph, "to see if there are any discrete pollutants that pop in as the storm goes on," but he believes that stormwater management needs to continue to focus on stormwater loads, not concentrations, and TMDLs. Livingston explains that Florida’s stormwater program, similar to most in the United States, is "designed to get 80% average annual load reduction of total suspended solids." He says, "We’ve used the research we have–our information on loadings, rainfall, and rainfall distribution–to come up with design criteria for infiltration systems, wet detention systems, filter systems, various common kinds of BMPs. These presumptive design criteria are set forth in our rules, and they get periodically fine-tuned as we learn more about the treatment mechanisms of typical BMPs."

Two samples taken every 10 minutes during a storm. Time 0 is at the 10:00 positions. Every 10 minutes, two more samples were taken (going clockwise). The storm ended after 40 minutes, so no samples were pulled in the last two bottles.

Times 0 and 1 on left and every 10 minutes to the right

Roger James is a California water resources management consultant with a long career in that field. He worked with the state’s Regional Water Quality Control Board from 1960 to 1988, including a stint as executive officer, and as an operations and water-quality manager for the Santa Clara Valley Water District from 1988 to 1995. Now working primarily with municipalities, James also sounds a cautionary note when it comes to unqualified acceptance of the first-flush event: "I think anyone looking at pollutant loadings or trying to select and identify BMPs should not just blanketly accept the first-flush theory. I think they need to know an awful lot more about their specific site … about the pollutants that they are dealing with." James finds that the belief of what constitutes the first flush varies considerably: "I think if you put 10 people in a room, you’d get 10 definitions of first flush." Of even greater concern to James is that first flush does not directly deal with pollutant loadings. "Now that we are into the TMDL program so much, I think if you only focus on first flush, you really haven’t dealt with the real pollutant loading."

James cites evaluations of the first-flush event in different parts of the country, which have produced highly variable findings. Groundbreaking work in Austin, TX, by George Chang and colleagues explored the assumption that the first half-inch of rainfall runoff washes off 90% of the pollutants. Investigators found instead that these pollutants make up only about 20% of the annual load and that much greater rainfall volumes–maybe more than 1.25 in.–should be treated. Work by the City of Portland, OR, determined that first flush occurred with small and moderate storms for total pollutants, but only minimally for dissolved pollutants. Portland also found that treating first flush for its design storm of 0.83 in. would treat only about 20% of the pollutant load during major storm events. Bob Pitt at the University of Alabama has demonstrated that it takes containing more than an inch of rainfall in that area to trap most of the pollutant load, and John Sansalone of Louisiana State University has shown in his studies that most of the pollutants in the first flush are associated with large particles. These studies suggest that stormwater treatment BMPs should be designed to capture or treat larger volumes of runoff to be effective in addressing pollutant loads and achieving compliance with TMDLs.

Because of such a range of findings, James agrees with Eric Livingston’s assessment that first flush is very site-specific and is affected by many variables. "[First flush] could depend on many factors, including whether you have acid rain or not. It’s probably very dependent on the storm event itself. What is the intensity? If you get an inch of rain in an hour, versus the Pacific Northwest, where you get an inch over a day and a half, you see entirely different things. It could be the duration and intensity of the storm, or how many antecedent dry days since the last storm, and perhaps more important, the physical characteristics of the pollutants that are being addressed."

Baffle box installation

Pollutant specificity has become a major concern of those evaluating the effectiveness of treating first flush. Flint Holbrook is an associate partner and project director for Woolpert LLP, which provides client services in engineering, architecture, design, and related services. Holbrook, who works in the Charlotte, NC, office, specializes in stormwater work–NPDES compliance, municipal separate storm sewer system permitting, watershed master planning, erosion control activities, and so on–for state and local governments. Holbrook recounts the start of first-flush capture requirements in South Carolina in 1991, which specified the first half-inch of rainfall, a requirement generally interpreted as applying to impervious areas. Holbrook, who helped write the legislation, explains that the intent was always pollutant-specific: "Our focus was simply to capture particulate pollutants, pollutants attached to sediments, which is a postconstruction condition. We never intended for it to capture dissolved pollutants. The sediments wash off the site, [as do] oils and greases, floatables, and so forth." From the standpoint of capturing sediments, particulates, oil and greases, and floatables, Holbrook believes that first-flush requirements are effective, but he cautions, "If you’re trying to get dissolved pollutants, such as dissolved nitrogen and phosphorus in the water column, you’re not going to have much success, particularly if you have a dry basin, and you should have a separation device to separate that first flush from the bypass flow." Holbrook says that some in the stormwater field now believe that the dissolved pollutant concentrations in runoff continue to increase throughout a storm event.

Gordon England, a project manager for Creech Engineers in Melbourne, FL, has accumulated many years of experience in dealing with stormwater issues and currently works with clients in such areas of stormwater management as NPDES permitting, TMDLs, and retrofitting for water-quality purposes and flood control. He believes that although a larger proportion of pollutants are found in the first flush, which he considers the first inch or so of rainfall, "you still see pollutants in rainfall runoff no matter what time [in the storm event] it is, whether first flush or last flush." England supports the idea of pollutant specificity: "You tend to see different pollutants at different times in the storm: greases and oils and sediments probably most at the first part of the storm–the first-flush effect–but the dissolved stuff, like fertilizer in yards, will continue to wash out of the yard throughout the storm. You probably don’t see too much of a first-flush effect with that type of pollutant." England believes that the first-flush effect is likely to be more pronounced in arid locations, where large concentrations of pollutants build up during the long periods between storms, than in places such as Florida, where storms are frequent but accumulation is small for most storms. He feels strongly that first flush must be considered for every storm, not just for those occurring at certain intervals.

In many areas of the country, England acknowledges, the regulatory framework focuses on addressing suspended solids, making exclusive consideration of first flush somewhat more valid for determining treatment options. England and James both caution, however, that the US Geological Survey (USGS) recently published highly critical evaluations of the reliability of total suspended solid data, and they note that the USGS recommended that these data should not be used for design or performance evaluations of sediment-removal BMPs.

In Florida, whose stormwater program has evolved over the years, "We have learned that there’s a lot more to it than the first flush," says England, "and we ought to be looking at bigger storms and more complex criteria. It’s not simple. I wish it was." England explains that the focus in stormwater treatment has shifted to picking BMPs "that will treat the whole storm." Much of Florida’s stormwater activities are paid for through local user or utility fees that have been adopted by more than 100 local governments. "Everybody who contributes to the stormwater burden pays a fee, relative to the amount they contribute," explains McClelland. "If you have controlled stormwater on your site to the point that there’s no burden to the community, then you don’t pay a fee–that’s the concept at least. It’s a win-win situation." Other sources of funding for Florida’s stormwater projects, which are often undertaken through funding partnerships, include taxes levied by regional water management districts, these districts’ own monies, and federal section 319 grant dollars administered through Florida’s DEP.

Baffle box illustration

McClelland provides a short list of approaches to addressing stormwater quality by "treating the whole storm": biological, physical, and chemical. Biological methods allow plants to take up pollutants and make them unavailable to humans. Physical treatment provides opportunities for sediment and particulate matter to settle out of the water column. "If you slow the water down, those particulates settle to the bottom," McClelland says, so that they remain in the bottom sediments "until they are cleaned up or isolated for a long period of time and don’t affect the environment very much." Chemical treatment uses several different approaches, including letting water sit until the pollutants change chemically, or forcing a chemical change by introducing coagulants that force solid material to join together and settle out. McClelland emphasizes that all of these approaches share a common strategy: "The stormwater has to be able to sit somewhere for a period of time, or at least slow down a bit, before it can be treated." Systems built exclusively for flood control through moving water–conveyance pipes, drainage systems, canals–work against this strategy. Stormwater ponds, whether wet or dry, provide better opportunities for biological, chemical, and physical treatment. "Dry ponds are ones where the water goes away over a short period, like two weeks; wet ponds are ones where water stays for a period of time, and the dissolved pollutants have more chance of being biologically or chemically treated." Retention ponds retain water indefinitely so that it either evaporates or settles into the ground, and detention ponds hold water a brief amount of time to allow biological, chemical, or physical treatment.

Effective treatment approaches can also "capture" the dissolved pollutants, such as dissolved metals. Online approaches store runoff temporarily before allowing most of the volume to discharge; first-flush events are masked because all of the stormwater runoff is mixed together. Although online systems are designed primarily to provide flood-control benefit, online wet detention systems also provide water-quality benefits. Offline practices divert the more polluted stormwater found in the first flush, and bypass the remaining rainfall to be managed for flood control. "There’s a first-flush concept that’s built in, and we still use that in a sense," describes Livingston, "but it’s not the first half-inch. We’ve also factored in a requirement that there’s a multiplier, depending upon how much impervious surface is on a site. There’s also a multiplier that depends upon whether you use retention versus detention, and a multiplier to determine the treatment volume, how much water you have to capture and treat." Where retention ponds are used, water is retained on-site and allowed to soak back into the ground with no discharge. "You get 100% treatment for all water you put back into the ground, and you rely upon the natural processes of soil and plant remediation to take care of some of that." When soils are not available or appropriate for infiltration, detention of water is employed before it is discharged. For flood control, explains Livingston, they do dry detention to hold up the water a little before it is released. "That does very little for water quality. For water quality, you have to do wet detention," he emphasizes, explaining that in Florida the criteria for wet detention is that 30% of the surface area on a site should be a shallow littoral zone with wetland plants. "Again," continues Livingston, "you’re relying upon a wide variety of natural processes that occur within a lake to remove those dissolved pollutants." These constructed treatment sites have been required in Florida for decades: "Since 1979, every new development has had to have an onsite stormwater system."

Gordon England explains that these requirements give developers a choice of BMPs–wet pond, dry pond, wetland, underground exfiltration trench, and so on–but he agrees that wet retention ponds are very common in Florida. "Because we have high groundwater tables, just a few feet down, it’s very easy for us to build these." Wetlands are also preferred, though not always easy to build, "and regulators like to see them used." Individual design engineers choose BMPs based on characteristics and restrictions of their particular site. A downtown commercial site might lack room for a pond, so exfiltration trenches under the parking lot might be an appropriate option. When infiltration systems are employed, the diverted first flush is stored rather than discharged, until it is removed through infiltration, evaporation, or evapotranspiration. In an application of biological treatment, vegetation such as grass helps to trap stormwater pollutants in infiltration areas. In arid locations, storms are more catastrophic and thus offer greater flooding threat. McClelland explains that, even in these instances, "you can control first flush by having a little offline system that deals with water quality by biological, chemical, and physical means."

Flint Holbrook admits that, for a number of BMPs, such as grassy swales and grass channels, "we really don’t know yet how effective some of those are in removing constituent [pollutants]. We have some rule-of-thumb guidance, but we haven’t developed a loading function of what exactly are the pollutants from varying land uses … for us to determine the efficiency of these BMPs." Holbrook and Woolpert are involved in a project with the South Carolina Office of Coastal Resource Management to do just that, "to look at those loading functions, the BMPs and their efficiency, and how those pollutants begin to decay in some of the practices that we’ve been using." Holbrook hopes the findings of this work will allow the development of some empirical relationships to give information on the decay functions and how they are affected by BMP performance.

In urban areas, development increases the amount of impervious surface, which also increases conveyance of stormwater, reducing, McClelland says, "the natural ability of the environment to take up pollutants." Acknowledging that improving flood control through increased conveyance works against providing effective stormwater treatment, McClelland says, "The same is true for wetlands. When you reduce wetlands, you reduce the ability of water to slow down [and to] biologically, chemically, and physically deal with pollution." McClelland says the common BMPs for his projects are detention and retention ponds and source control, "not allowing the pollutants to get into the stormwater in the first place." He emphasizes the need to "know more about the effectiveness of those things for urban, industrial, commercial, and agricultural systems so that we can apply them with more confidence." Eric Livingston explains that BMPs will vary from one state to another but that vegetation is "the key aspect" of any stormwater management system. Florida benefits, of course, from a year-round growing season compared to a six-month season in more northern locations. Livingston explains that solutions viable in Florida might not be applicable to areas that freeze, use extensive salts for ice melting, and experience the release of pollutants trapped in snowmelt. He reiterates the need for local data collection to design appropriate BMPs: "You can’t rely on national numbers."

Although stormwater managers might consider it far simpler to treat only first-flush runoff, some benefits of the "whole storm" approach might offset the effort involved. Gordon England describes an "ancillary benefit" of treating the whole storm that people often are not aware of: "When you put in the bigger ponds to treat more than the first flush, you’re storing more water in your system and releasing less to the downstream pipes, so you flood the downstream people less. So when you build the bigger ponds for more than the first flush, you’re helping with flood control throughout the rest of the community."

Mary Catherine Hager is a biologist, writer, and editor in Lafayette, LA.

Home + About + Subscribe + News + Calendar + Glossary
Talk + Images + Advertise + Contact Us + Search + Register + Services

Erosion Control Magazine | MSW Management Magazine
Grading & Excavation Contractor | ForesterPress | Forester Media

© 2000 - 2001 FORESTER MEDIA, INC.