Reexamining Best Management Practices for Improving Water Quality in Urban Watersheds

Reexamining Best Management Practices for Improving Water Quality in Urban Watersheds
of 15
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
   ABSTRACT: Municipalities will be implementing structural bestmanagement practices at increasing rates in their effort to complywith Phase II of the National Pollutant Discharge Elimination Sys-tem (NPDES). However, there is evidence that structural best man-agement practices (BMPs) by themselves may be insufficient toattain desired water quality standards. This paper reports on ananalysis of the median removal efficiencies of structural BMPs andcompares them to removal efficiencies estimated as being necessaryto attain water quality standards in the Rouge River in Detroit,Michigan. Eight water quality parameters are reviewed using datacollected from 1994 to 1999 in the Rouge River. Currently, five of the eight parameters in the Rouge River including bacteria, bio-chemical oxygen demand, and total suspended solids (TSS) exceedthe required water quality standards. The reported analysis of structural BMP efficiencies reveals that structural BMPs appearcapable of reducing only some of the pollutants of concern to accept-able levels.(KEY TERMS: nonpoint source pollution; storm water manage-ment; urban water management; water treatment; water quality;watershed management.) Pennington, Stephen R., Michael D. Kaplowitz, and Scott G. Witter, 2003. Reex-amining Best Management Practices for Improving Water Quality in UrbanWatersheds. Journal of the American Water Resources Association (JAWRA)39(5):1027-1041. INTRODUCTIONMunicipalities have been responsible for ensuring local water quality in their rivers and streams foryears (Berry  et al., 1996). Section 401 of the CleanWater Act of 1972 (CWA) prohibits the discharge of pollutants from a point source into waters of the Unit-ed States (U.S.) unless the discharge has been autho-rized by a National Pollutant Discharge EliminationSystem (NPDES) permit. Congress mandated a stormwater permitting program in the 1987 amendments tothe CWA. Under Phase I of the program, communitiesbegan specifically addressing storm water manage-ment. Phase I of the CWA uses NPDES permits toaddress storm water runoff from: (1) medium andlarge municipal separate storm sewer systems (MS4s)serving populations of 100,000 or more; (2) construc-tion activities that disturb five acres of land or more;and (3) ten categories of industrial activities. Morerecently, the U.S. Environmental Protection Agency(USEPA) issued final regulations for the next phase of the CWA. USEPA’s NPDES Revisions Addressing Storm Water Discharges (USEPA, 1999) (commonlyreferred to as Phase II) specifically addresses non-point source pollution and is the next step in fulfilling the CWA’s mandate of restoring and maintaining thechemical, physical, and biological integrity of thenation’s waters. Phase II requires operators of MS4sin urban areas with populations less than 100,000and operators of construction sites of less than fiveacres to obtain NPDES permits and implement prac-tices to control polluted storm water runoff. U.S.municipalities must meet Phase II requirements formanaging storm water before March 2003 (USEPA,1999).There are six elements to storm water manage-ment under Phase II: public education and outreach,public participation, illicit discharge detection andelimination, construction site runoff control, post-construction runoff control, and pollution prevention/ good housekeeping. However, the most commonresponse for many communities attempting to 1 Paper No. 02006 of the  Journal of the American Water Resources Association. Discussions are open until April 1, 2004. 2 Respectively, Ph.D. Candidate, Department of Resource Development, Michigan State University, 323 Natural Resource Building, EastLansing, Michigan 48824; Associate Professor, Department of Resource Development, Michigan State University, 311a Natural ResourceBuilding, East Lansing, Michigan 48824; and Professor and Department Chair, Department of Resource Development, Michigan State Uni- versity, 308 Natural Resource Building, East Lansing, Michigan 48824 (E-Mail/Kaplowitz: J OURNAL OF THE  A MERICAN  W ATER  R ESOURCES  A SSOCIATION  1027 JAWRA JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION OCTOBERAMERICAN WATER RESOURCES ASSOCIATION2003 REEXAMINING BEST MANAGEMENT PRACTICES FORIMPROVING WATER QUALITY IN URBAN WATERSHEDS 1  Stephen R. Pennington, Michael D. Kaplowitz, and Scott G. Witter 2  eliminate and prevent storm water pollution has beenthe implementation of structural BMPs (Roesner  etal., 2001; Urbonas, 2001). Structural BMPs for stormwater (sometimes called structural storm water treat-ment practices) are physical undertakings and con-struction projects used to reduce the levels of pollutants in storm water runoff to improve waterquality. Six typical storm water structural BMPs are:dry and wet ponds, wetlands, filtering and infiltrationpractices, and swales (ASCE, 2000; CWP, 2000). Thispaper evaluates the efficacy and potential benefit of these six structural BMPs for improving water quali-ty in an urban watershed.SOME PREVIOUS RESEARCHIn response to Phase I storm water pollutionrequirements, an increased number of structuralBMPs were implemented by communities in theirattempt to control pollution from urban runoff (Bar-rett, 2000; Roesner  et al., 2001). The objective of anumber of studies has been to assess the ability of storm water treatment practices such as wet ponds,grass swales, wetlands, sand filters, and dry deten-tion basins to reduce pollutant discharges associatedwith storm water runoff (CWP, 1997, 2000; ASCE,2000; Urbonas, 2001). However, inconsistent studymethods, lack of detailed design information, and thefailure to adequately report protocols have made widescale assessment of structural management practicesdifficult (Jones, 2001; Strecker  et al., 2001). Recently,the second edition of the National Pollutant RemovalPerformance Database (CWP, 2000), prepared by theCenter for Watershed Protection (CWP), raised doubtsabout the ability of structural BMP to protect watersdownstream from urban discharges (Roesner  et al., 2001). In line with such findings, Tom Schueler, exec-utive director of CWP, shared his belief that struc-tural BMP do not adequately protect downstreamaquatic environments from urban runoff (Roesner  etal., 2001).RESEARCH QUESTIONSThe reported research tests the hypothesis thatstructural BMPs do not adequately reduce the concen-trations and/or mass of pollutants of concern in therunoff to the Rouge River to levels necessary to meetcurrent surface water quality standards. Further-more, the research tests the hypothesis that as thepercentage of impervious area in a watershed increas-es (i.e., an area becomes more urbanized), the abilityof structural BMPs to remove the required concentra-tions and mass of pollutants of concern decreases.RESEARCH METHODThe reported research is based on an analysis of water quality data from the Rouge River NationalWet Weather Demonstration Project (RRNWWDP,2000). The reported analysis uses the Rouge Riverwater quality data to estimate the percentage of removal of pollutants necessary based on currentstandards for eight pollutants of concern (POC). Thisanalysis compares the necessary removal rates to thegenerally accepted removal efficiencies of six commonstructural BMPs. Doing so allows for an examinationof whether the structural BMP seem likely to meetthe water quality standards for the Rouge River. Fur-thermore, the analysis stratifies the results by theresearch area’s prevailing percentage of imperviousland cover in order to examine any variability associ-ated with different levels of urbanization.The RRNWWDP undertook an extensive effort toaccurately determine levels of impervious area andthe percentage of directly connected area (DCIA) foreach of 10 land use categories within the RougeRiver’s 11 subwatersheds (RRNWWDP, unpublished,1994, Camp Dresser McKee Memorandum: Determi-nation of Impervious Area and Directly ConnectedImpervious Area). The most recent digital land usedata (1990) were obtained by RRNWWDP and used tocreate 10 land use geographic information system(GIS) layers that were consistent with land use cate-gories in other national studies of pollutant loads.The percentages of impervious area and DCIA foreach of the land use categories, together with knowl-edge of the total acres of each land use type withineach subwatershed, were used to calculate areaweighted values of percent impervious area and per-cent DCIA for each subwatershed (RRNWWDP,unpublished, 1994, Camp Dresser McKee Memoran-dum: Determination of Impervious Area and DirectlyConnected Impervious Area). The GIS model and datawere ground truthed and verified by examining atotal of about 300 sample areas using aerial pho-tographs and field observations (RRNWWDP, unpub-lished, 1994, Camp Dresser McKee Memorandum:Determination of Impervious Area and Directly Con-nected Impervious Area). As a result, the reportedanalysis is undertaken at the watershed and subwa-tershed levels, while BMPs themselves generallyoperate at much smaller scales.The method used in this analysis of comparing median POC values with general BMP removal rates,while useful, should be approached with caution. JAWRA 1028  J OURNAL OF THE  A MERICAN  W ATER  R ESOURCES  A SSOCIATION P ENNINGTON , K APLOWITZ , AND  W ITTER  Limitations in such an approach relate primarily tospatial, temporal, and other variables. Variabilitywithin the watershed is an issue because spatial, tem-poral, or loading data do not exist to help frame BMPperformance against the range of local conditions.Therefore, we present the first and fourth quartile foreach POC in the study area as one way to capture anddescribe POC variability. Also, it must be pointed outthat the results of the reported analysis are directlycomparable to the srcinal RRNWWDP watershedmodel, which does account for spatial and loading sce-narios. The reported results are in general agreementwith the RRNWWDP model result. Therefore, itseems reasonable to use the median pollutant levelsfor the eight POC and the median removal efficienciesfor the six BMPs in the reported study. Additionallimitations associated with the RRNWWDP data andmonitoring results as well as the removal efficienciesof BMPs are discussed later in this paper.  Rouge River National Wet Weather Demonstration Project (RRNWWDP) The Rouge River National Wet Weather Demon-stration Project is a USEPA grant funded comprehen-sive program to manage wet weather pollution and restore water quality in the Rouge River, a tribu-tary of the Detroit River in southeastern Michigan(RRNWWDP, 2001a). The Rouge River is more than126 miles in length with four separate branches cov-ering an area of approximately 466 square miles (Fig-ure 1). The Rouge River watershed encompasses all orpart of 48 municipalities in three counties and is con-nected to more than 400 lakes, impoundments, andponds. With a population of more than 1.5 million, theRouge River watershed is the state’s most urbanizedland area, with only 25 percent of the land in thewatershed remaining undeveloped (RRNWWDP,2001a).The state of Michigan has designated all surfacewaters to be protected for the following uses: (1) agri-culture, (2) industrial water supply, (3) public watersupply at the point of intake, (4) navigation, (5) warmwater fisheries, (6) other indigenous aquatic life andwildlife, (7) partial body contact recreation, (8) totalbody contact recreation, and (9) cold water fisheries(Brown  et al., 2000). Those uses that apply to allreaches of the Rouge River include Items 5 through 8and Item 9 in Johnson Creek. However, severe pollu-tion problems in the Rouge River have limited the uses of the river throughout the watershed (RRNWWDP, 2001b).The early focus of the RRNWWDP aimed at controlling combined sewer overflows (CSOs) in older urban core portions of downstream areas. Theregulatory approach of issuing National PollutantDischarge Elimination System (NPDES) permitsmandating corrective action worked relatively well(Murray  et al., 1999). However, additional monitoring of the river after the Phase I permits were issuedshowed that storm water pollution needed to be con-trolled before full restoration of the river could beachieved. In response, a total of 60 pilot storm watermanagement projects were implemented throughoutthe Rouge River watershed (Murray  et al., 1999). Thestructural storm water treatment practices imple-mented in the Rouge River area include: wetland cre-ation and restoration, grassed swales and detentionponds, erosion controls, stream bank stabilization,and habitat restoration (RRNWWDP, 1996). Thereported analysis focuses on the efficacies of six struc-tural BMP (dry ponds, wet ponds, wetlands, filtering practices, infiltration practices, and swales) usedalong the Rouge River.  Rouge River Data  As mentioned above, the reported analysis usespublicly available data from the EPA sponsored RougeRiver National Wet Weather Demonstration Project.The RRNWWDP reports following generally accepteddata collection and water quality monitoring proce-dures (RRNWWDP, 1998). Monitoring of Rouge Riverwater quality began in the fall of 1993 and continueson an annual basis. Field based data collection pro-grams have been conducted along the Rouge on a sea-sonal basis, typically when air temperatures remainabove freezing and the water temperature remainsabove 12˚C (RRNWWDP, 1998).The long term average annual precipitation for theRouge watershed is 32.62 inches. Fifty-seven percent,or roughly 18.8 inches, of precipitation falls in the wetseason, from April through September. The remaining 13.8 inches (43 percent) falls from October throughMarch. Between 1994 and 1999, the year with themost rainfall was 1998, with 34.13 inches, while 1996had the least rainfall, with 27.39 inches. The largestsingle event occurred on July 8, 1998, and was 4.34inches. Runoff from this event was not sampled. Meanannual water temperature was 48.6˚F and rangedfrom a low of 48.3˚F in 1996 to a high of 53.5˚F in1998.This paper focuses on eight pollutants of concern inthe Rouge River: bacteria, biochemical oxygendemand, total Kjeldahl nitrogen (TKN), nitrate andnitrite nitrogen, total phosphorous, total suspendedsolids (TSS), total copper, and total zinc. These POCconsistently constitute eight of the nine principal POC in urban runoff (Roesner  et al., 2001). Lead isnot considered in the analysis. The reported research J OURNAL OF THE  A MERICAN  W ATER  R ESOURCES  A SSOCIATION  1029 JAWRA R EEXAMINING  B EST  M ANAGEMENT  P RACTICES FOR  I MPROVING  W ATER  Q UALITY IN  U RBAN  W ATERSHEDS  uses data from the Rouge River Project’s chemical andbacterial sampling programs that monitor instreamlevels of selected pollutants under both dry and wetconditions. The data used in the reported analysis aredrawn from 22 continuous water quality sampling sites so that specific CSO sampling sites were avoid-ed. Two to seven storm events are typically monitoredat each site in a given season.  Removal Needed The median concentrations for the eight pollutantsof concern addressed in this paper were calculatedbased on the Rouge River data from 1998 as well ascumulative data from 1994 to 1999. The 1998 calcula-tions are intended to show the variability across POCover a one-year period. The year 1998 was selectedbecause data for that year were the most recent andcomplete sampling available. The cumulative dataanalysis attempts to show the general health of theriver and is based on sampling data from long-termmonitoring of the Rouge River.The POC median concentration levels for both the1998 and cumulative year data were used to estimatethe levels of pollutant removal necessary to bring theRouge River into compliance with current standards.In the case of E. coli, where the geometric mean is the JAWRA 1030  J OURNAL OF THE  A MERICAN  W ATER  R ESOURCES  A SSOCIATION P ENNINGTON , K APLOWITZ , AND  W ITTER Figure 1. Rouge River Watershed.  usual measurement method, the researchers chose touse the median figure since some data were missing.There were not five sampling events per month foreach reach of the Rouge River from May throughOctober in the years 1994 to 1999, and therefore itwas not appropriate to calculate the geometric mean.Where specific water quality standards have beenestablished for the Rouge River, they were used to cal-culate the percentage of removal necessary to attainthe required water quality (MNREPA, 1994). In othercases, published water quality literature was used todetermine the target ambient water quality levels forPOC for the river (Harte  et al., 1991; Chapman, 1996;CWP, 2000). The percentage of pollutant removalneeded was calculated asRemoval Needed (%) = [(MedConc - Std)/MedConc]* 100where MedConc = the median concentration of thepollutant of concern; and Std = an established govern-ment standard, an irreducible concentration, or a rec-ognized level for unpolluted surface water.Equation (1) assumes that all inputs, including those not being treated, will have the same medianconcentration and therefore avoids the need to calcu-late loads. This assumption allows comparisonsbetween concentrations. Total maximum daily loadmeasures were not used as a standard because theyhave not yet been established for the Rouge River.To avoid misrepresenting and oversimplifying thewater quality of a heterogeneous 126-mile river, thedata were stratified based on relative levels of imper- vious cover. For each sampling station, levels of imperviousness for the surrounding area were calcu-lated from estimates of the percent impervious coverfor each subwatershed. These estimates wereobtained from the Rouge River Project (Carl Johnson,PE, Camp Dresser McKee, personal communication,July 17, 2001). Each sampling station was categorizedbased on the average impervious cover from its con-tributing area. By stratifying the sampling stations inthis way, a rough analysis of the effects of increasing urbanization on BMP removal efficiency is possible.Table 1 illustrates the placement of the water qualitydata and removal percentages needed into threegroups based on the percentage of imperviousnessassociated with the sampling sites (i.e., less than 20percent, 20 to 40 percent, and more than 40 percent).  Best Management Practices Removal Efficiencies The storm water management practices examinedin this study are: wetland creation and restoration,grassed swales, dry ponds, wet ponds, filtration practices (e.g., surface sand filters), and infiltrationspractices (e.g., porous pavement). These BMPs repre-sent an array of conventional storm water controlpractices commonly undertaken in urban and exur-ban watersheds (CWP, 1997, 2000; ASCE, 2000; Bar-rett, 2000; Roesner  et al., 2001; Strecker  et al., 2001). As such, they represent a reasonable set of stormwater treatment practices for the reported examina-tion of the ability of structural BMPs to remove thelevels of POC necessary for meeting the applicablewater quality standards in an urban watershed.Pollution removal efficiencies, usually referred toas percentages, describe the ability of a managementpractice to reduce pollutant levels between the inflowand the outflow of the management system. As men-tioned before, the projected pollutant removal efficien-cies associated with the structural practices in use onthe Rouge River (RRNWWDP, 1996) are generallyconsistent with removal efficiency estimated for struc-tural practices in the NPRPD (CWP, 2000). Becausethere is insufficient and inadequate data on struc-tural practice removal efficiencies in the Rouge River,the reported research uses generally accepted struc-tural practice removal efficiencies from the NPRPDdata for its the analysis. The NPRPD data are basedon completed studies and include 139 studies of stormwater treatment practices (STPs) over a 20-year peri-od (CWP, 2000).BMP pollutant removal efficiencies are notstraightforward, and a wide variety of methods havebeen employed in calculated removal efficiencies(Strecker  et al., 2001). Despite the variety in ways tocompute removal efficiencies for storm water manage-ment practices, the STP Database does not adjustremoval efficiencies in its database. It is beyond thescope of this paper to compute and analyze alterna-tive removal efficiency computational forms. There-fore, the reported research relies on the removalefficiencies of structural storm water treatment prac-tices reported in the STP Database (CWP, 2000).However, it must be noted that percent BMP removalefficiencies have been criticized as inappropriate mea-sures of BMP effectiveness not only due to the incon-sistent method, but also because they may potentiallymischaracterize some practices as less effective(Strecker  et al., 2001).The Center for Watershed Protection (CWP, 2000)notes that STP removal efficiency results should beused to examine the general removal capability of  various methods as well as designs. The reportedmedian removal values are based on the broad spec-trum of studies that make up the database. Further-more, the median removal percentages represent theBMP’s removal capacity under a variety of climaticand physiological conditions. The data used by theauthors to determine general removal capabilities for J OURNAL OF THE  A MERICAN  W ATER  R ESOURCES  A SSOCIATION  1031 JAWRA R EEXAMINING  B EST  M ANAGEMENT  P RACTICES FOR  I MPROVING  W ATER  Q UALITY IN  U RBAN  W ATERSHEDS (1)
Similar documents
View more...
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!