As I write this, we’re wrapping up a streak of record-setting rains in southern California. After last fall’s wildfires, which left residual ash blanketing hundreds of acres, some are worried about the potential for erosion, and about the detriment to our lakes, rivers, and beaches from not only the exposed soil but also the ash itself. Some of us, as the rains began, believed we were seeing some vivid examples of what might be termed the first-flush phenomenon as all that ash was picked up by the first good storm. Is that really what we were seeing—or is “first flush” actually an outmoded, or at least overrated, concept?

First flush—the concentration of pollutants that builds up during dry periods, which is carried by runoff during the initial part of a storm—is surprisingly controversial for something that, on the face of it, seems intuitive. Stormwater managers and regulators and BMP designers have debated the question of the first-flush phenomenon for years: what it is, whether it even exists, and if it does what its implications are for BMP design and selection. How much runoff, in other words, do we really need to capture and treat?

Some people have interpreted the first-flush concept to mean that treating only the initial runoff from a storm—some say half an inch, others an inch—is really all that’s necessary; most of what we needed to catch would be contained in that first flush. That isn’t always the case, though. As various researchers have pointed out, the size of the watershed or basin under study, the time of concentration, the sources of pollutants, and types of land use within the area, including the amount of impervious surface, all have an effect. The data can be vastly different for total pollutants versus just dissolved pollutants, and for different types of constituents in the runoff.

The article on page 32 of this issue, “First-Flush Characterization for Stormwater Runoff Treatment” by Masoud Kayhanian and Michael K. Stenstrom, helps to clarify many of these issues. It reports on a multiyear study of first flush, and, as the authors state, one of their goals is to improve first-flush characterization studies in general. The article addresses sampling methods and offers a concise definition of first flush—not only for a single storm event but also what the authors term seasonal first flush. Finally, they examine how the results are relevant to BMP selection and design.

That’s really what it comes down to for most of us: the practical application of the concept. Few of us will have the chance, or the resources, to conduct such studies on our own, but understanding the best and most comprehensive research on the subject helps in setting policy and in designing a stormwater treatment program to capture the most pollutants for the buck.

SW March/April 2008