The Stormwater Blogs

Others believe that this isn’t the case and cite several pilot studies (Selbig and Bannerman 2007, Center for Watershed Protection 2008) in which the analyses of the collected data and modeling efforts seem to support the conclusion that street cleaning shows limited effectiveness in reducing the concentrations of pollutants found in urban runoff (Geosyntec Consultants 2008). However, the TSS data that have been collected and the analytical methods used as part of these other studies were flawed, since they did not measure all of the particulate material being transported by the runoff. In addition, these projects did not use modeling tools that can accurately simulate the sediment accumulation and washoff behaviors and their interaction with cleaning practices (Sutherland and Minton, manuscript in press).

How can they be flawed? First, the withdrawal water velocities of automatic samplers used to collect the water quality samples limits pickup capabilities with regard to particle size. Silt and sand particles larger than 100 to 200 microns have been rarely sampled even when present. Second, the standard laboratory procedures for TSS testing do not accurately measure these larger particles, assuming they were sampled at all. Protocols for water-quality sampling that specify the use of a pipette for splitting a sample into smaller aliquots for analysis and/or rapid hand mixing and pouring fail to properly move larger material that may have been sampled. The good news is that newer procedures (e.g., churn splitters, separate analysis of larger material by prescreening, full sample rather than sub-sample analysis) have likely reduced these laboratory procedures biases, but sampling techniques still remain a problem.

How do we know this is a problem? A 1998 study of runoff from an interstate highway in Cincinnati, Ohio, used gravimetric-based sampling techniques to collect all of the sediment instead of using automatic samplers. The study also filtered and analyzed the entire volumes of discrete samples obtained throughout each sampled runoff event. The study concluded that 20% by mass of the particulate material transported in the runoff ranged from 600 to 1,000 microns and 30% from 1,000 to 10,000 microns (Sansalone et al. 1998). Recent discrete runoff sampling of eight storms captured from an elevated section of the I-10 freeway in Baton Rouge, Louisiana, used both gravimetric sampling techniques and whole effluent analyses. The study found particles in transport that ranged from 1 to 24,500 microns in size (Kim and Sansalone 2008).

Given the results of Sansalone’s research, one must conclude that the concentrations of sediments and other associated pollutants found in runoff from highway pavements have been routinely under sampled and thus understated. Metals, phosphorus, petroleum and related hydrocarbons, and pesticides are all hydrophobic and, therefore, sorb to larger particles and were also understated. Sansalone and Cristina (2004) found that more than 60% of the particulate-bound metal mass observed in highway pavement runoff (i.e. Cd, Cu, Pb, and Zn) was associated with particles greater than 250 microns.

Those that have been critical of street cleaning as an effective BMP often lack an understanding of the complicated processes that relate to both the accumulation and transport by runoff of particulate material from urban streets. Therefore they often don’t grasp how the effective removal of this accumulated material by cleaning practices can essentially deplete its available supply. The pilot studies that concluded street cleaning is not an effective BMP used very simple models or models whose washoff components were not based on sediment transport principles (Selbig and Bannerman 2007, Center for Watershed Protection 2008). These modeling techniques cannot accurately simulate the washoff processes or the complex interactions of accumulation and removal by cleaning.

Pacific Water Resources (PWR) has developed sediment transport based models (Sutherland and Jelen 1996) that have accurately reproduced these complicated processes and their interactions. One interaction these models include is called wet-weather washon, which is the contribution of significant particulate and associated pollutant loadings from adjacent paved and unpaved areas to the street and parking lot surfaces. Some have mistakenly believed that PWR’s modeling studies do not account for any “off-street” loadings sources. Geosyntec Consultants (2008) stated incorrectly that PWR’s models rely solely on street build up and washoff equations for the introduction of pollutants into runoff and that all other processes for how pollutants are entrained into runoff are assumed to be negligible. In fact, PWR has proven that only when the process of wet-weather washon is accounted for in their unique accumulation function can sediment and pollutant washoff be accurately modeled storm by storm, one season to another, year after year (Sutherland and Jelen 1996).

The general lack of understanding on how PWR’s models actually work has led to a mistaken belief that PWR’s conclusions regarding the pollutant reduction effectiveness of street cleaning are wrong. In reality, it is the complex interaction of these “off-street” loadings along with the sediment transport of these and other direct street accumulations that determine the time-varying pollutant concentrations and mass loadings found in urban runoff. If proper sampling and lab techniques are used then these behaviors would be observed.

PWR seems to be the only researchers in the country that emphasize the importance of a street cleaner’s ability to pick up and contain the entire range of accumulated particulate material. PWR studies using a unique sediment transport based washoff model that includes wet-weather washon called SIMPTM (Sutherland and Jelen 1998) estimated, for example, that regenerative air sweeping on single family residential areas in Livonia, Michigan, once every two weeks would remove from stormwater an estimated 63% of the TSS annual mass loadings (Sutherland et. al. 2001).

Some have stated that these street sweeping pollutant removal estimates “are highly questionable” (Geosyntec Consultants 2008). Let’s examine some numbers to see whether that statement is true. A recent street sweeping pilot study of an urban residential watershed in Baltimore, Maryland, estimated that monthly regenerative air sweeping would only reduce TSS by 22% (Center for Watershed Protection 2008). The Livonia, Michigan, modeling study using SIMPTM calibrated to accurately simulate measured street dirt accumulations on a residential site collected over a six month period concluded that monthly residential sweeping with regenerative air would reduce TSS stormwater washoff by 48% (Sutherland et. al. 2001). Errors in both sampling and analytical methods are estimated to limit the observed TSS washoff to only half of the sediment mass actually being transported (personal communication with Dr. Sansalone 2007). So if we assume that was the case in Baltimore, then the 22% TSS washoff reduction due to monthly regenerative air sweeping really applies only to the 50% of the sediments that were actually measured. If monthly regenerative air street sweeping were to remove from the washoff 75% of the other 50% of sediments that wasn’t measured, then a 48% overall washoff reduction for TSS would be an accurate estimate.

So the remaining question is whether monthly regenerative air street sweeping reduce the transported TSS sediments that aren’t being observed by traditional practices by 75%. Since the runoff sediments that aren’t being observed are likely greater than 200 microns, then a 75% reduction in the washoff of this courser fraction is very reasonable. Recent testing of an Elgin Crosswind regenerative air machine operating at 5 miles per hour under real-world sweeping conditions found that the machine picked up 96.4% of the initial accumulated material (Pacific Water Resources 2008). The pick-up performance of the particulates greater than 250 microns was measured at 97.5%. Granted, street sweeper pick-up is not the same thing as sediment washoff reduction, but they are closely related, especially for these coarser sediments for which washoff occurs at a much higher rate during more intense and generally higher-depth storms. These higher-depth storms have longer return intervals that can generally equal or exceed the frequency of the monthly cleaning operations in our example. So the likelihood of keeping the available supply of these coarser sediments low is quite good, especially when the pick-up efficiency for this particle size group is so high. Therefore, a 75% reduction in the washoff of these coarser sediments that aren’t being measured is realistic. So what may appear to be highly questionable estimates to some are in fact very reasonable estimates when the complex processes are understood and all the sediments in transport are counted.

For years, the focus has been on the removal or containment of the finest fraction of these accumulated and transported particulates. These fine fractions are certainly important due largely to the extremely toxic pollutants that have been associated with them. But based on Sansalone’s work regarding the mass of sediments, metals, phosphorus, hydrocarbons, and pesticides, it should now be clear to public works and stormwater management staff that effective stormwater pollutant reduction through improved street cleaning practices is possible. But it requires an examination of the removal of a much larger range of accumulated particulate material (i.e., up to 2000 microns).

One important question that should be addressed is whether these larger particulates that are estimated to make up half of the sediments in highway runoff also dominate in runoff from other urban areas like single-family residential and commercial land uses, and the impact the fine sediment sampling bias has had on the stormwater quality data already collected from these other areas. Most importantly, we must ask what impact has this had on the interpretation of this information.

Given the serious problems with the bias in data collection and data analyses that have been used by recent street cleaning pilot studies (Selbig and Bannerman 2007; Center for Watershed Protection 2008), PWR strongly believes that these data sets should be reevaluated using physically based explicit models like SIMPTM. This model can simulate the complex interactions of accumulation, washon, washoff, and the removal by street cleaners and catch basins for the entire range of sediments sizes. Data analyses using these types of tools will result in a much better understanding of how effective urban cleaning practices truly are. Until this occurs, we at PWR believe it would be prudent to recognize that the conclusions of these recent studies are flawed. If we truly care about the maximum extent practicable (MEP) and cost effective removal of pollutants from urban runoff, then the resources must be found to undertake a re-analysis of these and other data sets. Until this occurs, the controversy surrounding the true effectiveness of urban street cleaning practices will continue.

References

Center for Watershed Protection. 2008. Deriving Reliable Pollutant Removal Rates for Municipal Street Sweeping and Storm Drain Cleanout Programs in the Chesapeake Bay Basin, funded by a US EPA Chesapeake Bay Program grant.

Geosyntec Consultants. 2008. BMP Effectiveness Assessment for Highway Runoff in Western Washington – Appendix 3, Highway Sweeping, prepared for Washington State Department of Transportation.

Kim, J.Y., and J.J. Sansalone. 2008. “Event-Based Size Distributions of Particulate Matter Transported During Urban Rainfall-Runoff Events.” (Manuscript In Press).

Pacific Water Resources, Inc. 2008. Street Cleaner Pick-Up Performance Tests, prepared for the Elgin Sweeper Company, Elgin, Illinois.

Sansalone, J.J., and C.M. Cristina. 2004. “Prediction of Gradation-Based Heavy Metal Mass Using Granulometric Indices of Snowmelt Particles,” in Journal of Environmental Engineering. ASCE. 130(12): 1488-1497.

Sansalone, J.J., J.M. Koran, J.A. Smithson, and S.G. Buchberger. 1998. “Physical Characteristics of Urban Roadway Solids Transported During Rain Events,” in Journal of Environmental Engineering. ASCE. 124(5): 348-365.

Selbig, W.R., and R.T. Bannerman. 2007. “Evaluation of Street Sweeping as a Stormwater-Quality-Management Tool in Three Residential Basins in Madison, Wisconsin.” US Geological Survey, Middleton, Wisconsin, Water Resource Investigations Report 2007-5156.

Sutherland, R.C., G.R. Minton, and U. Marinov. 2006. “Stormwater Quality Modeling of Cross Israel Highway Runoff,” in Intelligent Modeling of Urban Water Systems, Monograph 14 (Edited by W. James, K.N. Irvine, E.A. McBean and R.E. Pitt). CHI. Guelph, Ontario, Canada: Chapter 8.

Sutherland, R.C., R.J. Myllyoja, and S.L. Jelen. 2001. “Quantifying the Stormwater Pollutant Reduction Benefits of Traditional Public Works Maintenance Practices,” in Best Modeling Practices for Urban Water Systems, Monograph 10 (Edited by W. James). CHI. Guelph, Ontario, Canada: 127-150.

Sutherland, R.C., and S.L. Jelen. 1996. “Sophisticated Stormwater Quality Modeling is Worth the Effort,” in Advances in Modeling the Management of Stormwater Impacts, Volume 4, (Edited by William James). CHI. Guelph, Ontario, Canada: 1-14.

Sutherland, R.C., and S.L. Jelen. 1997. “Contrary to Conventional Wisdom: Street Sweeping Can be an Effective BMP”. In Advances in Modeling the Management of Stormwater Impacts, Volume 5, (Edited by William James) CHI. Guelph, Ontario, Canada: 179-190.

Sutherland, R.C., and S.L. Jelen. 1998. Simplified Particulate Transport Model User’s Manual, Version 3.2. Pacific Water Resources, Inc., Beaverton, Oregon.