Base-flow samples were collected seasonally from 20 small streams in the 372-square-mile Pepacton Reservoir watershed to evaluate the effects of hydrogeologic factors and nonpoint sources of contamination on the chemical composition of ground-water discharge. The reservoir provides part of New York City?s water supply. The subbasins represented one of three general types of land use, each with at least 45 percent forested area (mostly on the hillsides): farmed (dairy) land, formerly farmed land with low-density residential development, or forested land with little or no development. The subbasin areas ranged from 0.38 to 10.23 square miles. All streams were sampled in December 2000 and in May, July, and October 2001. Three of the sites were designated as landuse- index sites and were sampled as many as five additional times during the study. No samples exceeded state or federal drinking-water standards for chloride, sodium, nitrate, orthophosphate, herbicides, or herbicide degradates.

The chemical composition of base-flow samples was classified into major-ion water types, which were broadly defined as naturally occurring and road-salt-affected water types. About one-third of the base-flow samples were roadsalt- affected types. Natural water types were differentiated as dilute or evolved. Dilute waters have bicarbonate and sulfate as dominant anions and evolved waters have only bicarbonate as a dominant anion. Dilute water types indicate relatively short ground-water residence times or contact with unreactive aquifer material. Evolved waters have either longer ground-water residence time or contact with more reactive aquifer material than dilute ground waters. The larger subbasins with wider valley-bottom areas were more likely to have evolved water types than small subbasins with little floodplain development.

Positive correlations between selected constituents and the intensity of nonpoint sources emphasize the connection between land use, shallow ground-water quality, and stream base-flow water quality. Chloride and sodium, which are relatively conservative constituents, showed strong linear relations with annual estimates of road-salt application during all four sampling periods. Nonconservative constituents, such as the nutrients nitrate and orthophosphate, showed linear relations with manure production rate among farmed basins, but only at specific times of the year because of losses through biologic activity. Nitrate showed the strongest relation in winter because losses to biological activity were at a minimum. Orthophosphate showed the strongest relation in early summer, when hydrologic and chemical conditions appear to favor release from sediments. Atmospheric nitrogen deposition is an additional source of nitrogen that can be released from mature or stressed forested basins.

Detections of herbicides (atrazine, metolachlor, simazine) and herbicide degradates ( Metolachlor ESA, alachlor ESA, deethylatrazine) in base flow were closely correlated with subbasins in which corn was grown during the study. Atrazine was detected at the farmed index site only in early summer, after application and two rain storms. This detection corresponded to the peak orthophosphate concentration. In contrast, metolachlor ESA was detected in nearly all farmedindex- subbasin samples and peaked in late summer, when percent base-flow contributions from farmed valley-bottom areas were likely highest.

The implications of this study are that seasonal and more frequent base-flow surveys of water chemistry from small stream basins can help refine the understanding of local hydrogeologic systems and define the effects of nonpointsource contamination on base-flow water quality. The concentration of most nonpoint sources in valley-bottom or lower-hillside areas helped indicate the relative contributions of water from hillside and valley-bottom areas at different times of year. The positive correlations between the intensity of nonpoint-source activities and nonpoint-source constituents in base flow underscores the link between land use (nonpoint sources), ground-water quality, and surface-water quality.