|Abstract:||The source water area for the drinking-water supply of the city of Cambridge, Massachusetts, encompasses major transportation corridors, as well as large areas of light industrial, commercial, and residential land use. Because of ongoing development in the drinking-water source area, the Cambridge water supply has the potential to be affected by a wide variety of contaminants. The U.S. Geological Survey (USGS) has monitored surface-water quality in the Hobbs Brook and Stony Brook Basins, which compose the drinking-water source area, since 1997 (water year 1997) through continuous monitoring and discrete sample collection and, since 2004, through systematic collection of streamwater samples during base-flow and stormflow conditions at five primary sampling stations in the drinking-water source area. Four primary sampling stations are on small tributaries in the Hobbs Brook and Stony Brook Basins; the fifth primary sampling station is on the main stem of Stony Brook and drains about 93 percent of the Cambridge drinking-water source area. Water samples also were collected at six secondary sampling stations, including Fresh Pond Reservoir, the final storage reservoir for the raw water supply. Storm runoff and base-flow concentrations of calcium (Ca), chloride (Cl), sodium (Na), and sulfate (SO4) were estimated from continuous records of streamflow and specific conductance for six monitoring stations, which include the five primary sampling stations. These data were used to characterize current water-quality conditions, estimate loads and yields, and describe trends in Cl and Na in the tributaries and main-stem streams in the Hobbs Brook and Stony Brook Basins. These data also were used to describe how streamwater quality is affected by various watershed characteristics and provide information to guide future watershed management. Water samples were analyzed for physical properties and concentrations of Ca, Cl, Na, and SO4, total nitrogen (TN), total phosphorus (TP), caffeine, and a suite of 59 polar pesticides. Values of physical properties and constituent concentrations varied widely, particularly in samples from tributaries. Median concentrations of Ca, Cl, Na, and SO4 in samples collected in the Hobbs Brook Basin (39.8, 392, 207, and 21.7 milligrams per liter (mg/L), respectively) were higher than those for the Stony Brook Basin (17.8, 87.7, 49.7, and 14.7 mg/L, respectively). These differences in major ion concentrations are likely related to the low percentages of developed land and impervious area in the Stony Brook Basin. Concentrations of dissolved Cl and Na in samples, and those estimated from continuous records of specific conductance (particularly during base flow), often were greater than the U.S. Environmental Protection Agency (USEPA) secondary drinking-water guideline for Cl (250 mg/L), the chronic aquatic-life guideline for Cl (230 mg/L), and the Commonwealth of Massachusetts, Executive Office of Energy and Environmental Affairs drinking-water guideline for Na (20 mg/L). Mean annual flow-weighted concentrations of Ca, Cl, and Na were generally positively correlated with the area of roadway land use in the subbasins. Correlations between mean annual concentrations of Ca and SO4 in base flow and total roadway, total impervious, and commercial-industrial land uses were statistically significant. Concentrations of TN (range of 0.42 to 5.13 mg/L in all subbasins) and TP (range of 0.006 to 0.80 mg/L in all subbasins) in tributary samples did not differ substantially between the Hobbs Brook and Stony Brook Basins. Concentrations of TN and TP in samples collected during water years 2004–07 exceeded proposed reference concentrations of 0.57 and 0.024 mg/L, in 94 and 56 percent of the samples, respectively. Correlations between annual flow-weighted concentrations of TN and percentages of recreational land use and water-body area were statistically significant; however, no significant relation was found between TP and available land-use information. The volume of streamflow affected water-quality conditions at the primary sampling stations. Turbidity and concentrations of TP were positively correlated with streamflow. In contrast, concentrations of major ions were negatively correlated with streamflow, indicating that these constituents were diluted during stormflows. Concentrations of TN were not correlated with streamflow. Twenty-five pesticides and caffeine were detected in water samples collected in the drinking-water source area and in raw water collected from the Cambridge water-treatment facility intake at the Fresh Pond Reservoir. Imidacloprid, norflurazon, and siduron were the most frequently detected pesticides with the frequency of detections ranging from about 24 to 41 percent. Caffeine was detected in about 37 percent of water samples at concentrations ranging from 0.003 to 1.82 micrograms per liter (μg/L). Although some of the detected pesticides degrade rapidly, norflurazon and siduron are relatively stable and are able to immigrate though the serial reservoir system. Concentrations of 2,4-D, carbaryl, imazaquin, MCPA (2-methyl-4-chlorophenoxyacetic acid), metsulfuron-methyl, norflurazon, siduron, and caffeine were detected more frequently in stormflow samples than in base-flow samples. Concentrations of pesticides did not exceed USEPA drinking-water guidelines or other health standards and were several orders of magnitude less than the lethal exposure level established for several fish species common to the drinking-water source area. Imidacloprid, an insecticide, was the only pesticide with a concentration exceeding available long-term aquatic-life guidelines. Several pesticides correlated significantly with the amount of recreational, residential, and commercial area in the tributary subbasins. Mean annual base-flow concentrations of caffeine correlated significantly with parking-lot land use. For most tributaries, about 70 percent of the annual loads of Ca, Cl, Na, and SO4 were associated with base flow. Upward temporal trends in annual loads of Cl and Na were identified on the basis of data for water years 1998 to 2008 for the outlet of the Cambridge Reservoir in the Hobbs Brook Basin; however, similar trends were not identified for the main stem of Stony Brook downstream from the reservoir. The proportions of the TN load attributed to base flow and stormflow were similar in each tributary. In contrast, more than 83 percent of the TP loads in the tributaries and about 73 percent of the TP load in main stem of Stony Brook were associated with stormflow. Mean annual yields of Ca, Cl, Na, and SO4 in the Stony Brook Reservoir watershed, which represents most of the drinking-water source area, were 14, 85, 46, and 9 metric tons per square kilometer, respectively. Mean annual yields among the individual tributary subbasins varied extensively. Mean annual yields for the respective constituents increased with an increase in roadway and parking-lot area in the tributary subbasins. Mean annual yields of TN in the tributary subbasins ranged from about 740 to more than 1,200 kilograms per square kilometer and exceeded the yield for the main stem of Stony Brook at USGS station 01104460 upstream from the Stony Brook Reservoir. Mean annual yields estimated for the herbicides 2,4-D and imidacloprid ranged from 34 to 310 grams per square kilometer (g/km2) and 3 to 170 g/km2, respectively. Annual loads for 2,4-D were entirely associated with stormflow. The largest annual load for imidacloprid was estimated for the main stem of Stony Brook; however, the highest annual yield for this pesticide, as well as for benomyl, carbaryl, metalaxyl, and propiconazole, was estimated for a tributary to the Stony Brook Reservoir that drains largely residential and recreational areas. Mean annual yields for the herbicide siduron ranged from 6.9 to 35 g/km2 with most of the loads associated with stormflow. Mean annual yields for the insecticide diuron ranged from 2.1 to 4.4 g/km2. Annual yields of caffeine ranged from 11 to 410 g/km2.