Environ. Sci. Technol. 2002, 36, 4441-4446
Treatment of Parking Lot Stormwater Using a StormTreat System REBECCA S. SONSTROM, JOHN C. CLAUSEN,* AND DAVID R. ASKEW† Department of Natural Resources Management and Engineering, University of Connecticut, 1376 Storrs Road, Storrs, Connecticut 06238-4087
The effectiveness of a StormTreat system in treating stormwater from a commercial parking lot in Connecticut was evaluated. Flow-weighted composite samples were collected from StormTreat inflow and outflow during a 2-yr study. Bypass flow was not monitored. The StormTreat significantly (P < 0.05) reduced total suspended solids, total phosphorus, total Kjeldahl-N, total zinc, total copper, and fecal coliform bacteria on a concentration basis. The StormTreat system retained 49% total suspended solids, 74% total phosphorus, 44% total Kjeldahl-N, 45% total zinc, 29% total copper, 2% total lead on a mass basis, and 99% fecal coliform on a concentration basis. Treatment efficiency was not associated with storm size, chamber stage, discharge rate, or hydraulic retention time (r < 0.355). The system retained ammonia-N more efficiently during the summer than during the winter (P < 0.01) and retained total zinc less efficiently during the summer than during the winter (P < 0.05). Season did not significantly (P > 0.05) affect the treatment of other monitored water quality variables. The StormTreat system reduced the concentrations of stormwater pollutants commonly found in parking lot runoff.
Introduction Commercial parking lots are known to be sources of nonpoint pollutants in stormwater runoff (1-5). Such lots have been characterized as hotspots, which produce elevated loads of hydrocarbons and metals (3). Commercial parking lots in Wisconsin have been found to be a critical source of contaminants, including solids, nutrients, heavy metals, and bacteria and should be targeted for best management practices in order to most cost-effectively control pollutant loads (1). In Michigan, commercial parking lots produced higher concentrations of polycyclic aromatic hydrocarbons in stormwater than rooftops, residential lawns, and streets (5). In response to these findings, considerable attention has been directed toward designing new technologies to treat parking lot stormwater runoff. The StormTreat system was designed in 1994 to capture the first flush of stormwater runoff. The system is an inground modular, 2.9 m diameter unit with a series of chambers that receives runoff from a catch basin. Stormwater * Corresponding author fax: (860)486-5408; phone: (860)486-2840; e-mail:
[email protected]. † Present address: Tolland County Soil and Water Conservation District, 24 Hyde Ave., Vernon, CT 06066. 10.1021/es020797p CCC: $22.00 Published on Web 09/14/2002
2002 American Chemical Society
enters the system through a grit bag, which filters fine solids, and proceeds through a series of sedimentation chambers. An inverted elbow transfers water from the first chamber to the next, potentially retaining oils and floatables in the previous chamber. The remaining chambers are fitted with skimmers, which decant water from one chamber to the next by drawing water from approximately 7.6 to 10 cm below the surface (6). The skimmers should hinder sediment and floatables from passing through the system (6). The partially treated water then enters the root zone of the fringing constructed wetland filled with 1-cm gravel. On the basis of potential treatment processes, StormTreat is expected to provide filtration and sedimentation within the central chamber and filtration, adsorption, and biochemical reactions in the surrounding wetland. However, the actual treatment processes are not known. Mean concentration data and treatment efficiencies using five discrete samples collected at the first StormTreat installation site in Kingston, MA, have been reported (7). However, no long-term, continuous-monitoring studies of treatment efficiency for the StormTreat system were available. StormTreat literature indicates that the system should remove 99% of total suspended solids, 90% of phosphorus, 77% of total dissolved nitrogen, 97% of fecal coliform, 90% of total petroleum hydrocarbons, and 90% of zinc based on client data (8). This study was conducted to evaluate the performance of the StormTreat system in treating solids, nitrogen, phosphorus, heavy metals, bacteria, and total petroleum hydrocarbons in runoff from a commercial parking lot. In addition, this study evaluated the effect of climate and hydrologic factors, including season, storm size, chamber stage, discharge, and hydraulic residence time on treatment efficiency.
Experimental Section Field Site. Runoff from a 0.27-ha commercial, parking lot and rooftop in East Hartford, CT, was sampled prior to flow into the StormTreat system (Figure 1). The parking lot was exposed to an average of 650 vehicles day-1 (9) and moderate automotive duration (10-30 min vehicle-1). The asphalt surface had a uniform 3.4% slope and was in good condition with few cracks. Sand and salt were applied to the parking lot during winter storms. Vegetation bordered the site but was not in the watershed. Runoff originated from the parking lot, contributing rooftop area, and the periodic washdown of the trash and grease dumpsters (10). StormTreat System. In June 1997, two parallel StormTreat tanks were installed in the parking lot. Together the tanks treated the first 0.46 cm of runoff from the drainage area based on engineering designs (10). Normally, StormTreat systems are sized to treat the first 1.27 cm of runoff, which would require five tanks at this site. Space limitations imposed by the owner and costs only allowed installation of two units. Inflow to the StormTreat system came from a double catch basin with sump fed from preexisting drainage structures. An overflow weir and baffle wall allowed excess stormwater flow to bypass the StormTreat (Figure 2A). Each tank was 1.2 m high and 2.9 m in diameter, with a storage capacity of 5261 L (Figure 3). The outflow from the StormTreat was controlled by a ball valve at a maximum of 0.76 L min-1 and discharged via a 5.1-cm pipe (Figure 2B). Each unit cost $3300 and installation cost $24 600. The wetland substrate consisted of No. 7 ASTM gravel approximately 1.00-1.25 cm in size (10). The surface was planted in June 1997 with woolly sedge (Scirpus cyperinus L.), tussock sedge (Carex stricta Lam.), umbrella sedge VOL. 36, NO. 20, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4441
FIGURE 1. Site plan for the StormTreat system at a parking lot location, East Hartford, CT (CB ) catch basin). (Cyperus diandrus Torr.), and soft rush (Juncus effusus L.). The plants were hand-watered in the first few weeks to encourage survival. During maintenance of the system, the plant rooting depth was determined to be 1 m (10). Sampling. For this study, only the StormTreat system was studied. There was no attempt to study or measure bypass loads or catch basin collection. Stage in the catch basin was measured by a pressure transducer and recorded on a Campbell Scientific CR-10 data logger. Effluent discharge was measured using a calibrated tipping bucket, with tips recorded by a mechanical counter. Weekly flow-weighted samples were collected at the StormTreat inlet and outlet. The samples were split into one nonacidified and two acidified composite bottles using a passive splitter in situ. An ISCO automated sampler collected a flow-weighted composite sample at the inlet. A passive, slotted splitter at the outlet directed flow into composite bottles when the bucket tipped. Discrete within-event samples were not obtained using this technique. Bottles were pre-acidified with 2 mL of H2SO4 or HNO3 L-1 for subsequent nutrient or metals analyses, respectively, and located in coolers with ice packs. Holding times for acidified samples range from 28 d to 6 month depending on the variable analyzed (11). Sample bottles were transported in a cooler with ice packs. Precipitation was measured with a tipping bucket rain gauge and recorded by the data logger. One 5-cm monitoring well was installed in the wetland of each StormTreat system to track wetland water levels. Analytical Methods. Weekly composite samples were analyzed for total suspended solids (TSS), total phosphorus (TP), total Kjeldahl nitrogen (TKN), ammonia-nitrogen (NH34442
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 20, 2002
N), nitrate+nitrite-nitrogen (NO3-N), and pH. Weekly insitu dissolved oxygen measurements, taken in the monitoring wells, were made using a YSI dissolved oxygen meter. Monthly composite samples were analyzed for total copper (Cu), total lead (Pb), and total zinc (Zn). Preserved metals have a 6-month holding time (11). Grab samples were analyzed for fecal coliform bacteria (FC) and total petroleum hydrocarbons (TPH). Grab samples were collected during events by opening the appropriate container and collecting overland flow from the pavement surface as it flowed into the catch basin and from the discharge of the StormTreat. Grit bags were removed from the StormTreat chambers annually. Bags were dried and weighed, and collected sediment was analyzed for residue gravimetrically (12). Subsamples were analyzed for total N using a LECO FP-2000 nitrogen analyzer and for total P using perchloric digestion followed by colorimetric analysis (11). TP, TKN, NH3-N, and NO3-N concentrations were analyzed by colorimetric flow injection analysis on a Lachat autoanalyzer using U.S. EPA methods (11). Nutrient samples were analyzed within 28 d of collection. Analyses of pH, beginning in May 1998, were performed using a pH meter (12). TSS analyses were performed on nonacidified samples using gravimetric methods (12). Cu and Zn concentrations were analyzed using plasma emission spectroscopy (ICP), and Pb was analyzed using atomic absorption furnace (GFAA) methods (11). FC analyses were conducted using a membrane filter (12). TPH for lump sum diesel range organics was analyzed using methylene chloride extraction (GC/FID) (13). Samples were collected and analyzed in accordance with a U.S. EPA approved quality assurance project plan. Standard quality control samples included check standards every 10 samples, duplicates and spikes every 20 samples, and equipment field blanks every 20 samples. Statistical Methods. Statistical analyses were performed using JMP (14). Means for censored data were determined using UNSENSOR (15). The Shapiro-Wilks test was used to assess normality. Most data were found to follow log-normal distributions; therefore, the log transformation was used in those cases. Percent retention values were normalized with a square root transformation. Significant differences between influent and effluent concentrations and percent retention efficiency between winter and summer seasons were determined using the Student’s t test. Comparisons for Cu and Pb were made both between values just above detection limits and between all values by replacing the values below the detection limit with half the detection limit. Pearson correlation coefficients were used to examine the relationships between percent retention and storm size, chamber stage, discharge rate, and hydraulic retention time. Correlation analyses represent only weeks when samples were collected from both the inlet and the outlet. Mean concentrations, mass loading, and cumulative mass retention values represent all data obtained. Hydraulic residence time (HRT) was calculated as the volume of the StormTreat system, adjusted for porosity, divided by the discharge. HRT was determined to have a mean (anti-log) of 9 d (s ) 4.8 d). Therefore, effluent concentrations were compared to influent concentrations from the preceding week in statistical analyses. Percent concentration retention efficiency was calculated from efficiency ) 1 - (mean effluent concentration/mean influent concentration) and converted to a percentage.
Results and Discussion During the 2-yr study period, 80 influent and effluent composite samples were collected; no runoff occurred during 16 weeks. Average annual precipitation at the Hartford WSO Airport, located approximately 7.5 km north of the site, was 1312 mm yr-1 in year one and 850 mm in year two. These precipitation amounts represent +17% and -24% departures
FIGURE 2. (A) Cross-section view of catch basin. (B) Cross-section view of outlet from the StormTreat.
FIGURE 3. Cross-section view of StormTreat system chamber and constructed wetland.
TABLE 1. Median Concentrations in Commercial Runoff, Mean (anti-log) Water Quality Concentrations, and Student’s t for Paired t-Tests for the StormTreat System (July 1997-July 1999) variables TSS (mg L-1) TKN (mg L-1) NO3-N (mg L-1) NH3-N (mg L-1) TP (mg L-1) Pb (µg L-1) Cu (µg L-1) Zn (µg L-1) FCU TPH (mg L-1) pH
n
mediana
influent mean
effluent mean
student’s tb
79 69 14 7 4.581*** 80 1.179 1.6 1.0 4.494*** 81 0.572 0.2 0.2 1.001 80 0.02 0.13 0.17 -1.341 80 0.201 0.143 0.048 6.089*** 16 104 5 5 0.184 18 29 7 4 2.711* 18 226 151 53 2.695* 16 12,000 590 0.30) and total Pb (χ2 ) 0.03, P > 0.30). Mean effluent StormTreat concentrations were below the U.S. EPA drinking water action levels for copper and lead of 1300 and 15 µg L-1, respectively (18). The U.S. EPA has not set an action level for zinc in drinking water. NH3-N concentrations were well below the U.S. EPA ambient water quality criteria at the temperature and pH ranges observed (19). Zn influent concentrations exceeded the U.S. EPA recommended water quality criteria of 120 µg L-1 (20), but effluent concentrations were lower than the criteria (Table 1). Mean concentration data and treatment efficiencies using five samples collected at the first StormTreat installation site in Kingston, MA, have been reported (8). The system treated a 1740-m2 drainage area, consisting of a road and parking lot. Influent samples were collected in the sedimentation chamber. Percent retention efficiencies were calculated from mean influent and effluent concentrations as 99% TSS, 89% TP, 44% total dissolved nitrogen (NH3-N + NO3-N), and 97% FC. Additionally, the Kingston, MA, StormTreat system retained 90% TPH, 77% Pb, and 90% Zn. Using the same method for calculating percent retention, the StormTreat system in East Hartford, CT, retained 50% TSS, 66% TP, 0% TDN, and 99% FC. This StormTreat system also retained 60% TPH, 0% Pb, and 74% Zn. Differences in percent retention values between the two studies are partly due to differences in mean influent concentrations, which were larger at the Kingston, MA, site. Mean effluent concentrations were similar at both sites. Mass Retention. Mass loadings to the StormTreat system were lower than mass export values for commercial land uses estimated by NURP (17) (Table 2). NURP found that land-use categories including residential, commercial, and mixed did not explain site-to-site variability in pollutant contributions and concluded that the best general characterization of urban runoff would be provided by pooling the data from all urban sites (17). The median concentrations for urban runoff along with an annual rainfall value (1016 mm) and a land-use runoff coefficient for commercial sites (0.8) were used to estimate annual export from commercial areas. Parking lot loading to the StormTreat system was low relative to the values estimated by NURP because the mean influent concentrations at this site were lower than the median concentrations for urban runoff used by NURP. Modest TSS retention (48.8%) by the StormTreat system was attributed to low influent TSS concentrations and the consistent presence of iron bacteria in the effluent. The catch basin retained solids prior to the StormTreat system. Using results from prior studies, the U.S. EPA (21) reported TSS removal by catch basins to be 10-25%. The iron bacteria 4444
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 20, 2002
increased the amount of TSS in the effluent, thereby reducing the percent retention. The grit bag stored 50.7% of TSS input. Most of the nitrogen retained by the StormTreat system was organic (Table 2). The main mechanisms for nitrogen removal in constructed wetlands include sedimentation, volatilization, and denitrification (22). Filtration by the grit bag removed 2.2% of total N. Low mass retention of NH3-N resulted from concentrations often being near the detection limits. Also, reducing conditions within the StormTreat system hindered the oxidation of NH3-N to NO3-N, sometimes resulting in higher NH3-N concentrations in the effluent than in the influent. Similar results have been documented elsewhere for wetland systems (23, 24). Surplus NH3-N is thought to result from the decomposition of organic-N in combination with insufficient oxygen required for nitrification (25, 26). Anoxic conditions, like those found in this system, are typically conducive to denitrification (22); however, mass retention of NO3-N was not observed. In this system, denitrification likely was limited by an inadequate NO3-N source. Despite favorable anaerobic conditions, insufficient NO3-N (27) and carbon (27-29) sources can limit denitrification. As the system ages, carbon supplies may increase due to the buildup and decay of organic materials (22), yet denitrification will continue to be hampered by low influent NO3-N. The primary mechanisms for TP removal in constructed wetland systems are plant uptake, settling, adsorption onto substratum, precipitation, and complexation reactions (22, 30). The main retention mechanisms operating in the StormTreat system were sedimentation and filtration, followed by adsorption to the gravel. The grit bag trapped 5.8% of TP as particulate-bound P, leaving the dissolved P to undergo further treatment in the fringing wetland. Phosphorus taken up by plants during the summer months was released back into the system during plant senescence. Phosphorus removal may become limited as the system ages because of saturation of adsorption sites (31) and reduced uptake by mature plants (22). Sedimentation and filtration were the likely retention mechanisms for particulate-bound Cu, Pb, and Zn in this system. The dissolved forms may have been removed through precipitation and adsorption. Anaerobic environments, as in this system, have been found to immobilize Cu and Zn through binding to sulfides (32, 33). Petroleum hydrocarbons, which have been found associated with particles (34, 35), likely were removed by sedimentation and filtration. High FC retention can be attributed to entrapment, filtration, and pathogen die-off (22). Generally, there is at least a 90% reduction of FC in effluent from constructed wetlands (36, 37). Differences were observed when retention was calculated based on concentrations versus mass. Such differences are expected since concentrations often vary with flow. The authors recommend that mass retention be generally used to evaluate stormwater treatment devices except for variables such as bacteria and pH. Climatic and Hydrologic Effects. There was no significant (P ) 0.05) correlation between weekly precipitation, chamber stage, discharge, HRT, and percent retention for all water quality variables (r ) 0.032-0.383). Comparing winter to summer, only Zn and NH3-N retention varied between seasons (Table 3). The percent retention of Zn was greater (P ) 0.05) in winter than in summer. Less retention of NH3-N occurred in winter than in summer (P ) 0.01) because winter concentrations were often near detection levels, resulting in 0% retention. Although cooler temperatures lower biological reaction rates, longer HRTs observed during the winter months may compensate for slower reaction rates (22). Possibly the mean HRT of 9 d within the StormTreat system
TABLE 3. Seasonal Comparison of Mean Weekly Percent Retention and Student’s t for a Non-Paired t-Testa retention % variables
annual
summer
winter
Student’s t
TSS TKN NH3-N NO3-N TP total Zn
49 54 45 23 69 75
58 56 81 21 66 67
41 52 19 28 72 85
1.360 0.377 3.438** -0.541 -0.553 -2.193*
a Summer ) May-October; winter ) November-April. *, P ) 0.05; **, P ) 0.01.
TABLE 4. Summaries of Percent Mass Retention for Alternative Stormwater Treatment Systems mass retention (%) suspended solids total phosphorus total Kjeldahl N Zn fecal coliform bacteria TPH concentration a
Ref 21.
b
constructed Storm- Stormwetlands & wet pondsa ceptor b Treat Vortechnicsc 60-80 25-65 20-55 20-60
25-34 19-29 32 21-60 -15 12
49 74 44 45 99 37
77-88 67 18 85 -7 16
Refs 39 and 40. c Refs 39 and 41.
was sufficient to counteract for decreased reaction rates during the winter. In other studies of wetland systems, the treatment efficiency has been found to level off after a 1-d HRT for TSS (26) and a 1.5-d HRT for Cu, Pb, or Zn (38). However, at least 6-8 d has been cited as the minimum necessary for effective treatment of NH3-N (26). These studies imply that beyond an HRT of 1.5 d the treatment efficiency for TSS, metals, and other adsorbed pollutants will not increase. Mean HRT within the StormTreat system was 9 d and was greater than 1.5 d for 95% of the study period. HRT would not have been a significant control factor in the removal of solids and metals in the StormTreat system given these circumstances. The effects of climate and hydrologic factors likely were dampened by StormTreat design controls, including a maximum allowable inflow rate, regulated via an overflow weir, and a maximum discharge rate controlled by a valve. Comparison with Alternative Treatment Technologies. There are numerous alternative management practices to the StormTreat system for reducing pollutants in stormwater runoff from parking lots. The common practices include various types of constructed wetlands and ponds and several types of oil-grit separators (21). Direct comparison of these alternatives should acknowledge that many systems are designed to function differently. For example, the StormTreat system is not designed to collect all of the stormwater runoff from a parking lot, rather surface water in a catch basin is decanted and there can be considerable bypass flow. Likewise, some wetland systems are designed to treat the first-flush of stormwater and bypass higher flows. Also, some oil-grit separators bypass high flows. Selection of alternatives is also affected by space for the treatment system. The StormTreat system and oil-grit separators are often being used where space is limited as compared to ponds and wetlands, which may require more space. In comparison to some of the oil-grit separators, such as the Stormceptor and the Vortechnics, the StormTreat system was found to retain less TSS but more TP, TKN, and TPH (Table 4). Reduction of FC bacteria was much greater in the StormTreat system than in the oil-grit separators. However, constructed wetlands and wet ponds may retain
as much, and in some cases more, stormwater pollutants as the StormTreat system (Table 4). The StormTreat system appears to offer additional treatment of stormwater runoff well beyond that provided by a catch basin. The system was particularly effective in treating bacteria and TP. TSS treatment would probably be higher if the system did not produce iron bacteria solids in the effluent.
Acknowledgments This research was funded in part by the Connecticut Department of Environmental Protection through a U.S. Environmental Protection Agency Clean Water Act §319 nonpoint source grant and also by the Tolland and Hartford County Soil and Water Conservation Districts. This is Storrs Agricultural Experiment Stations Scientific Contribution No. 2109.
Literature Cited (1) Bannerman, R. T.; Owens, D. W.; Dodds, R. B.; Hornewer, N. J. Water Sci. Technol. 1993, 28, 241-259. (2) Rabinal, F. I.; Grizzard, T. J. Concentrations of Selected Constituents in Runoff from Impervious Surfaces in Four Urban Catchments of Different Land Use; Occoquan Watershed Monitoring Monitoring Laboratory, Department of Civil Engineering, Virginia Tech: Manassas, VA, 1995. (3) Schueler, T. Water Prot. Tech. 1994, 1, 1-5. (4) Shepp, D. L. Watershed ‘96 Proc. 1996, 220-223. (5) Steuer, J.; Selbig, W.; Hornewer, N.; Prey, J. An Urban Basin in Marquette, Michigan and an Analysis of Concentrations, Loads, and Data Quality; USGS WRI Report 97-4242; U.S. Geological Survey: Reston, VA, 1997. (6) Allard, L. A.; Platz, W.; Carr, R.; Wheeler, J. Watershed ‘96 Proc. 1996, 463-465. (7) Horsley, S. W. Water Prot. Tech. 1995, 2, 304-306. (8) StormTreat Systems, Inc. StormTreat System Infocard. (9) Lopez, A. Personal Communication, 1999. (10) Askew, D. Personal Communication, 1998. (11) U.S. Environmental Protection Agency. Methods for Chemical Analysis of Water and Wastes; EPA-600/4-79-020; U.S. EPA: Cincinnati, 1983. (12) American Public Health Association. Standard Methods for the Examination of Waste and Wastewater, 17th ed.; American Public Health Association: Washington, DC, 1989. (13) U.S. Environmental Protection Agency. (SW-846) Method 8015B. Nonhalogenated organics using GC/FID and Method 3510. Test Methods for Evaluating Solid Waste. Physical and Chemical Methods; Laboratory Manual Vol. 1b, revision 2; Office of Solid Waste and Emergency Response: December 1996. (14) SAS Institute, Inc. Jmp Statistics and Graphics Guide Version 3.1; SAS Institute: Cary, NC, 1995. (15) Newman, M. C.; Dixon, P. M.; Looney, B. B.; Pinder, J. E., III. Water Res. Bull. 1989, 25, 905-915. (16) National Oceanic and Atmospheric Administration. Climatological Data Annual Summary: New England; Goverment Printing Office: Washington, DC, 1995; Vol. 107, pp 1-2. (17) U.S. Environmental Protection Agency. Results of the Nationwide Urban Runoff Program; Water Planning Division: Washington, DC 1983. (18) United States Environmental Protection Agency. National recommended water quality criteria. Office of Water, Washington, DC, 1998. (19) U.S. Environmental Protection Agency. 1999 Update of Ambient Water Quality Criteria for Ammonia; Office of Water: Washington, DC, 1999. (20) U.S. Environmental Protection Agency. National Recommended Water Quality Criteria-Correction; Office of Water: Washington, DC, 1999. (21) U.S. Environmental Protection Agency. Guidance Specifying Management Measures for Sources of Nonpoint Pollution in Coastal Waters; EPA 840-B-92-002; Government Printing Office: Washington, DC, 1993. (22) Reed, S. C.; Crites, R. W.; Middlebrooks, E. J. Natural Systems for Waste Management and Treatment; McGraw-Hill: New York, 1995. (23) Herskowitz, J.; Black S.; Lewandowski W. In Aquatic Plants for Water Treatment and Resource Recovery; Reddy, K. R., Smith, W. H., Eds.; Magnolia Press, Inc.: Orlando, FL, 1987; pp 247254. (24) Newman, J. M.; Clausen, J. C. Wetlands 1997, 17, 375-382. VOL. 36, NO. 20, 2002 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
4445
(25) Reed, S. C.; Brown, D. S. Water Environ. Res. 1992, 64, 776-781. (26) Reed, S. C.; Brown, D. S. Water Environ. Res. 1995, 67, 244-248. (27) Groffman, P. M. Cur. Top. Wetland Biogeochem. 1994, 1, 1535. (28) Van Oostrom, A. J.; Russell, J. M. Water Sci. Technol. 1994, 29, 7-14. (29) Zhu, T.; Sikora, F. J. Water Sci. Technol. 1995, 32, 219-228. (30) Mann, R. A.; Bavor, H. J. Water Sci. Technol. 1993, 27, 107-113. (31) Tanner, C. C.; James, P.; Sukias S.; Upsdell, M. J. Environ. Qual. 1998, 27, 448-458. (32) Gersberg, R. M.; Lyon, S. R.; Elkins, B. V.; Goldman, C. R. Proceedings of the Future of Water Reuse, San Diego, CA; AWWA Reseach Foundation: 1984; Vol. 2, pp 1639-1645. (33) Scholze, R.; Novotny, V.; Schonter, R. Water Sci. Technol. 1993, 28, 215-224. (34) Hoffman, E. J.; Latimer J. S.; Mills, G. L.; Quinn, J. G. J. Water Pollut. Control Fed. 1982, 54, 1517-1525. (35) Pitt, R.; Field R.; Lalor M.; Brown, M. Water Environ. Res. 1995, 67, 260-275. (36) Kadlec, R. H.; Knight, R. L. Treatment Wetlands; Lewis Publishers: Boca Raton, FL, 1996. (37) Ottova, V.; Balcarova, J.; Vymazal, J. Water Sci. Technol. 1997, 35, 117-123.
4446
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 36, NO. 20, 2002
(38) Crites, R. W.; Dombeck, G. D.; Watson, R. C.; Williams, C. R. Water Environ. Res. 1997, 69, 132-135. (39) Clausen, J. C.; Belanger, P.; Board, S.; Dietz, M.; Phillips, R.; Sonstrom, R. Final Report Stormwater treatment devices section 319 project. Connecticut Department of Environmental Protection: Hartford, CT, 2002. (40) Waschbusch, R. J. Evaluation of the Effectiveness of an Urban Stormwater Treatment Unit in Madison, Wisconsin, 1996-1997; USGS Water Resources Investigations Report 99-4195; U.S. Geological Survey: Middleton, WI, 1999. (41) West, T. A.; Sutherland, J. W.; Bloomfield, J. A.; Lake, D. W., Jr. A study of the effectiveness of a Vortechs stormwater treatment system for removal of total suspended solids and other pollutants in the Marine Village watershed, Village of Lake George, New York; NYS Department of Environmental Conservation, Division of Water: Albany, NY, 2001.
Received for review June 21, 2002. Revised manuscript received July 12, 2002. Accepted July 25, 2002. ES020797P
