Vegetated strips or biofilter strips provide many benefits to water qu - PDF document

Vegetated strips or biofilter strips provide many benefits to water quality treatmentof storm water, including increased infiltration, decreased sedimentation, and decreasederosion. Biofilter strips can be applied on various slopes from 5 to 52% and still provideTypically, the Revised Universal Soil Loss Equation (RUSLE) is used to estimateoptimum factors in slope design, including vegetation coverage. RUSLE2 has beenintroduced as an improved computer-based model which can be applied to disturbed sitesin the urban environment. RUSLE2 can be customized for site features using updated soils,climate, cover, and practices information. Factors for urban settings have been determinedto allow this model to be applied to construction sites or other disturbed environments,A two-year monitoring study was performed on biofilter strips adjacent to the highwayby the California Department of Transportation (Caltrans). Eight sites with vegetated stripsin increasing widths from 1.1 to 13 m from the right-of-way were established. Two to five, 30m long concrete collection trenches were installed to collect sheet flow passing through thebiofilter. Storm water was sampled using automated sampling equipment. Water quality wasassessed to determine the optimum width in which treatment occurred over two stormStatistical analysis of the water quality data indicated that most of the treatmentoccurred within approximately 3 m of the right-of-way. Additional biofilter widths greater than3 m did not provide significantly more treatment. However, treatment effectiveness wasaffected by percent vegetation cover. It appeared that at least 65% cover was needed toachieve significant pollutant removal. Erosion rates were not originally estimated in the study.A comparison of biostrip widths may help assess the optimum treatment design forbiostrips. Using RUSLE and RUSLE2 to determine strip width and vegetation coverage can 161 unless they were part of the routine maintenance ofthat particular area. To the maximum extent possible,Storm water was sampled using automatedsampling equipment. Flow-weighted compositestorms per season when possible. Samplers wereprepared prior to predicted storm events. Water qualitywas assessed to determine the optimum width in whichtreatment occurred over two storm seasons. Sampleswere analyzed for water quality parameters such as 162 metals, organics and sediments following thedepartments storm water monitoring protocolsStatistical analysis of the water quality dataindicated that most of the treatment occurred withinapproximately 3 to 4 m of the right-of-way. Additionalbiofilter widths greater than 3 m did not providesignificantly more treatment. For example, the averagetotal suspended solids (TSS) concentration wasreduced to 25 mg/L, total zinc was reduced to 25 ug/L,and dissolved zinc was reduced to 12 ug/L. Othermetals concentrations were reduced to less than 10ug/L (Scharff et al., 2004). Concentration reductionsvaried by site and storm. However, treatmentcover. It appeared that at least 65% cover was neededto achieve significant pollutant removal. Erosion ratesAlthough erosion rates were not measured as partof the original study, they can be estimated. Data onvegetative cover, soils, climate, and slope werecollected as part of the original study (Table 1). RUSLEfactors were then estimated from the original data set.The R factor was determined using site locationinformation and isoerodent maps. The K factor wasestimated using the soil erodibility nomograph. The LSfactor was determined using published LS tables withactual slope length and percent slope measurements.The C factor was estimated using the percent covermeasured at the site. The P factor was set to emulatea slope that was scraped with a bulldozer up and downbeen upon constructioncompletion. RUSLE was computed using the RUSLESiteGravelSacramento 1.1 m51.836.911.31.13.6293Sacramento 4.6 m31.936.531.64.615.13384Sacramento 6.6 m32.536.5316.621.73392Sacramento 8.4 m39.235.8258.427.63390Cottonwood 9.3 m4441.614.49.330.65273Redding 2.2 m39.648.811.62.27.21080Redding 4.2 m47.242.510.34.213.81085Redding 6.2 m34.752.812.56.220.41087San Rafael 8.3 m40.638.620.88.327.35084Irvine 3.3 m24.959.915.23.310.91170Irvine 6 m16.759.523.8619.71163Irvine 13 m20.146.533.41342.81162Yorba Linda 1.9 m28.153.418.51.96.31461Yorba Linda 4.9 m25.353.521.24.916.11482Yorba Linda 7.6 m17.260.622.27.625.01474Yorba Linda 13 m34.249.616.21342.81476Moreno Valley 2.6 m20.361.518.22.68.6133Moreno Valley 4.9 m29.75317.34.916.11316Moreno Valley 8 m16.559.124.4826.31322Moreno Valley 9.9 m13.770.216.19.932.61318San Onofre 1.3 m1963.817.21.34.3881San Onofre 5.3 m27.156.816.15.317.41074San Onofre 9.9 m21.755.722.69.932.61669 164 Differences in the RUSLE A, or estimated soil loss,can be attributed to the values chosen for the RUSLEfactors. The hand computation method uses publishedtables and charts to allow the user to determine theprogram is designed to ease use by allowing the userto select from database values in the program.However, these values may not be as accurate. Forin the RUSLE2 database, but 0.17 by the handcomputation method. This site also has a C factor of0.00002 in the database, and 0.38 by the handcomputation method. It is up to the user to determinethe most appropriate RUSLE factor values for the site.It is difficult to determine which method provides amore accurate model without comparison to actual soilIt appears that either RUSLE or RUSLE2 willprovide the user with viable alternatives to selectpractices that minimize site erosion. Both methods areregarding climate, soils, vegetation, and practices.However, RUSLE2 does require familiarity with thecomputer program, and updates to the database. Italso requires the user to assume some geographicuniformity to select a location close to the site, if thesite city is not represented in the database. Similarassumptions must be made with regard to vegetationAnalyzing the water quality data from the biofilterstudy reveals that the least amount of potentialerosion, and export of particles, occurs within 3 to 4 mof the ROW. Strips larger than 4 m did not appear toreduce erosion significantly more. This is an importantfinding in the transportation environment where spaceis often limited. Even in a site only 4 m wide, waterquality treatment and erosion loss is minimized. Thiscould be potentially due to the development ofconcentrated flow with longer slope lengths, or tochanges in vegetation coverage. Although it appearedthat sites with 65% percent vegetation coverage ormore had the most reduction in water quality pollutantsThe use of biofilter strips as water quality treatmentbest management practices (BMPs) is an importantspaces where larger storm water BMPs may be 157