Water-quality trends in the Scituate reservoir drainage area, Rhode Island, 1983-2012

Abstract

The Scituate Reservoir is the primary source of drinking water for more than 60 percent of the population of Rhode Island. Water-quality and streamflow data collected at 37 surface-water monitoring stations in the Scituate Reservoir drainage area, Rhode Island, from October 2001 through September 2012, water years (WYs) 2002-12, were analyzed to determine water-quality conditions and constituent loads in the drainage area. Trends in water quality, including physical properties and concentrations of constituents, were investigated for the same period and for a longer period from October 1982 through September 2012 (WYs 1983-2012). Water samples were collected and analyzed by the Providence Water Supply Board, the agency that manages the Scituate Reservoir. Streamflow data were collected by the U.S. Geological Survey. Median values and other summary statistics for pH, color, turbidity, alkalinity, chloride, nitrite, nitrate, total coliform bacteria, Escherichia coli ( E. coli ), and orthophosphate were calculated for WYs 2003-12 for all 37 monitoring stations. Instantaneous loads and yields (loads per unit area) of total coliform bacteria and E. coli , chloride, nitrite, nitrate, and orthophosphate were calculated for all sampling dates during WYs 2003-12 for 23 monitoring stations with streamflow data. Values of physical properties and concentrations of constituents were compared with State and Federal water-quality standards and guidelines and were related to streamflow, land-use characteristics, varying classes of timber operations, and impervious surface areas. Tributaries in the Scituate Reservoir drainage area for WYs 2003-12 were slightly acidic (median pH of all stations equal to 6.1) and contained low median concentrations of chloride (22 milligrams per liter [mg/L]), nitrate (0.01 mg/L as nitrogen), nitrite (0.001 mg/L as nitrogen), and orthophosphate (0.02 milligrams per liter as phosphorus [mg/L as P]). Turbidity and alkalinity values also were low with medians of 0.57 nephelometric turbidity units and 5.1 mg/L as calcium carbonate, respectively. Total coliform bacteria and E. coli were detected in most samples from all stations, but median concentrations were generally low-43 colony-forming units per 100 milliliters (mL) and 15 colony-forming units per 100 milliliters, respectively. Median values of several physical properties and median concentrations of several constituents correlated positively with the percentages of developed land and negatively with the percentages of forest cover in the drainage areas above the monitoring stations. Median concentrations of chloride correlated positively with the percentages of impervious land use in the subbasins of monitoring stations, likely reflecting the effects of deicing compounds applied to roadways during winter maintenance. Median concentrations of alkalinity also correlated positively with the percentage of impervious land use, which may be related to the deterioration of fabricated structures containing calcium carbonate. Median values of color correlated positively with the percentage of wetland area in the subbasins of monitoring stations, reflecting the natural sources of color in tributaries. Streamflows were negatively correlated with turbidity and concentrations of total coliform bacteria and E. coli, possibly reflecting seasonal patterns in which relatively high values of these properties and constituents occur during warmer low-flow conditions late in the water year. Similar seasonal patterns were observed for pH, alkalinity, and color. Negative correlations between concentrations of chloride and streamflow also were significant, indicating that deicing salts from roadways and other impervious surfaces that lack direct connection to the tributaries are likely infiltrating to the groundwater and discharging to some of the tributaries late in the water year. While salt-laden runoff directly enters some of the tributaries at roadway crossings, most of the roadway runoff infiltrates into the adjacent berms throughout the drainage area. Statistically significant correlations were not identified between various degrees of tree-canopy reduction caused by timber operations in the subbasins and median values or concentrations of water-quality properties. Loads and yields of chloride, nitrate, nitrite, orthophosphate, and bacteria varied at monitoring stations in the Scituate Reservoir drainage area in WYs 2003-12. Loads generally were greater at stations in the Barden Reservoir and the Regulating Reservoir Subbasins that have larger drainage areas than in subbasins with smaller drainage areas. Subbasin yields of fecal-indicator bacteria and orthophosphate generally were largest in the Westconnaug Reservoir Subbasin, and subbasin yields for chloride, nitrate, and nitrite were largest in the Moswansicut Reservoir Subbasin in the northeastern part of the drainage area. Upward trends in pH were identified for nearly half of the monitoring stations for WYs 1983-2012 and may reflect regional reductions in acid precipitation. Many upward trends in alkalinity also were identified for both the WYs 1983-2012 and for WYs 2003-12 periods and are likely related to the natural weathering of structures containing concrete or, in some cases, the application of lime or fertilizers on agriculture lands. Significant trends in chloride concentrations at most stations during WYs 1983-2012 were upward; however, results for WYs 2003-12 substantiate few significant upward trends and, in a few cases, downward trends were identified in several tributary drainage areas.