Estimates of the quantity and quality of ground-water discharge from valley-fill deposits were calculated for nine valley reaches within the Pepacton watershed in southeastern New York in July and August of 2001. Streamflow and water quality at the upstream and downstream end of each reach and at intervening tributaries were measured under base-flow conditions and used in mass-balance equations to determine quantity and quality of ground-water discharge. These measurements and estimates define the relative magnitudes of upland (tributary inflow) and valley-fill (ground-water discharge) contributions to the main-valley streams and provide a basis for understanding the effects of hydrogeologic setting on these contributions. Estimates of the water-quality of ground-water discharge also provide an indication of the effects of road salt, manure, and human wastewater from villages on the water quality of streams that feed the Pepacton Reservoir. The most common contaminant in ground-water discharge was chloride from road salt; concentrations were less than 15 mg/L.

Investigation of ground-water quality within a large watershed by measurement of stream base-flow quantity and quality followed by mass-balance calculations has benefits and drawbacks in comparison to direct ground-water sampling from wells. First, sampling streams is far less expensive than siting, installing, and sampling a watershed-wide network of wells. Second, base-flow samples represent composite samples of ground-water discharge from the most active part of the ground-water flow system across a drainage area, whereas a well network would only be representative of discrete points within local ground-water flow systems. Drawbacks to this method include limited reach selection because of unfavorable or unrepresentative hydrologic conditions, potential errors associated with a large number of streamflow and water-quality measurements, and limited ability to estimate concentrations of nonconservative constituents such as nutrients.

The total gain in streamflow from the upper end to the lower end of each valley reach was positively correlated with the annual-runoff volume calculated for the drainage area of the reach. This correlation was not greatly affected by the proportions of ground-water and tributary contributions, except at two reaches that lost much of their tributary flow after the July survey. In these reaches, the gain in total streamflow showed a negative departure from this correlation.

Calculated ground-water discharge exceeded the total tributary inflow in each valley reach in both surveys. Groundwater discharge, as a percentage of streamflow gain, was greatest among reaches in wide valleys (about 1,000-ft wide valley floors) that contain permeable valley fill because tributary flows were seasonally diminished or absent as a result of streambed infiltration. Tributary inflows, as a percentage of streamflow gain, were highest in reaches of narrow valleys (200-500-ft wide valley floors) with little valley fill and high annual runoff.

Stream-water and ground-water quality were characterized by major-ion type as either (1) naturally occurring water types, relatively unaffected by road salt, or (2) road-salt-affected water types having elevated concentrations of chloride and sodium. The naturally occurring waters were typically the calcium-bicarbonate type, but some contained magnesium and (or) sulfate as secondary ions. Magnesium concentration in base flow is probably related to the amount of till and its carbonate content, or to the amount of lime used on cultivated fields within a drainage area. Sulfate was a defining ion only in dilute waters (with short or unreactive flow paths) with low concentrations of bicarbonate. Nearly all tributary waters were classified as naturally occurring water types.

Ground-water discharge from nearly all valley reaches that contain State or county highways had elevated concentrations of chloride and sodsodium. The mean chloride concentrations of ground-water discharge--from 8 to 13 milligrams per liter--did not exceed Federal or State standards, but were about 5 times higher than naturally occurring levels. Application of road salt along a valley bottom probably affects only the shallow ground water in the area between a road and a stream. The elevated concentrations of chloride and sodium in the base-flow samples from such reaches indicate that the concentrations in the affected ground water were high enough to offset the low concentrations in all unaffected ground water entering the reach.

Nutrient (nitrate and orthophosphate) concentrations in base-flow samples collected throughout the valleyreach network could not generally be used to estimate their concentrations in ground-water discharge because these constituents can be transformed or removed from water through biological uptake, transformation, or by adsorption on sediments. Base-flow samples from streams with upgradient manure sources or villages served by septic systems consistently had the highest concentrations of these nutrients.