JAWRA - Stormwater Runoff - Laserfiche WebLink

Home Browse Search

STORMWATER RUNOFF QUALITY AND QUANTITY FROM TRADITIONAL AND IoW IMPACT DEVELOPMENT WATERSHEDS TABLE 2. Monitoring Schedule for the Jordan Cove Project, Connecticut. Watershed Period Calibration Construction Postconstruction Control 11/1/1995 to 6/30/2005 Traditional 4/4/1996 to 10/8/1997 10/8/1997 to 6/19/2003 6/19/2003 to 6/30/2005 LID 1/18/1996 to 3/23/1999 3/23/1999 to 8/1/2002 8/1/2002 to 6/30/2005 Sample Analysis Acidified, filtered samples were analyzed for nitrate + nitrite - nitrogen (NO + NO -N) using EPA Method 353.2 (USEPA, 1993b) and ammonia - nitrogen (NH -N) using EPA Method 350.1 (USEPA, 1993c). Acidified, unfiltered samples were analyzed for total Kjeldahl nitrogen (TKN) (Method 351.2) (USEPA, 1993d) and total phosphorus (TP) (Method 365.4) fol- lowing persulfate digestion (USEPA, 1983a). These analyses were performed using a LaChat colorimetric flow injection system. Nonacidified samples were ana- lyzed gravimetrically for total suspended solids (TSS) using EPA method 160.2 (USEPA, 1983a). Acidified, unfiltered monthly composites of samples were ana- lyzed for total copper (Cu), and zinc (Zn) using plasma emission spectroscopy (EPA Method 200.7) (USEPA, 1994). Pb analysis used atomic absorption (EPA Method 239.2). Fecal coliform membrane filter (Method 9222D) (Clesceri et al., 1998) and BOD (Method 521013) (Clesceri et al., 1998) analyses were performed immediately after returning from weekly field collections. A USEPA approved Quality Assur- ance Project Plan was followed during the study in accordance with the USEPA (1998, 2001). The Plan provided for preservation and storage protocols. Duplicates and spikes were analyzed every 20 samples and a check standard every 10 samples. Household Survey A one -page survey was mailed to each household in all three watersheds from 1999 -2004. The overall response rate was 55 %. Statistical Analysis Analysis of covariance (ANCOVA) was used to deter- mine changes in the slopes and intercepts of treatment and control watershed regressions for water quality and quantity variables between the calibration and treatment periods. For cases when regressions were not significant, paired t -tests were used to determine differences in the mean peak discharge and the mean TP, BOD, and FC concentrations between the tradi- tional and BMP watersheds. Data were analyzed using SAS version 9.1 software (SAS Institute Inc., 2002). Mass export (g/ha/month) was calculated as the prod- uct of weekly cumulative flow and weekly sample con- centration, divided by the watershed area. The presentation of results from paired watershed studies includes the means observed during each period, a pre- dicted mean, and percent change due to the treatment. Means reported for the control watershed during the calibration period may be different in Tables 2 and 3 because the calibration period length differed for the traditional and LID watersheds. Predicted treatment watershed means were calculated from calibration regression equations based on observed control means during each treatment period. Percent change was cal- culated as the difference between the observed mean and the predicted mean for the appropriate treatment watershed using the equation: % change = [(Observed — Predicted)/Predictedl *100. A Chi - square test was used to compare household survey counts in a two -way contingency table. RESULTS AND DISCUSSION Precipitation The normal annual precipitation in Groton, Con- necticut, located 7 km from the study area, was 1,237.5 mm (NOAA, 2002). Annual departures from normal ranged from —24 to +14% during this study. Overall, the precipitation during this study averaged 5% below normal. Storm Flow Traditional Watershed. An increase in storm flow from the traditional watershed was observed during the postconstruction period as compared with the calibration period. Mean flow depth increased by JOURNAL OF THE AMERICAN WATER RESOURCES ASSOCIATION 1001 JAWRA