Long-term data indicate that chloride concentrations in surface waters in the Hudson River basin are directly related to water use by sewage-treatment plants. Median low-flow chloride concentrations for 15 watersheds, all but one with a drainage area greater than 450 km2 (square kilometers), calculated from 1970-90 data are directly related to sewage input per watershed area. The median low-flow chloride concentration for the Hudson River at Corinth, N.Y., is 2 mg/L (milligrams per liter), and the annual sewage discharge in this watershed is less than 20 L/km2 (liters per square kilometer). In contrast, the median low-flow chloride concentration in a nearby tributary, the Hoosic River below Williamstown, Mass., is 17 mg/L, and the annual sewage discharge is nearly 20,000 L/km2. Samples collected from the 15 watersheds also indicate that median low-flow chloride concentration increases with population density.
Plots of chloride concentration as a function of discharge for the 15 watersheds indicate that chloride concentration generally declines with increasing discharge. Therefore, it appears that chloride concentrations are largely controlled by point sources. During the winter (December through March), however, chloride concentrations for a given discharge occasionally exceed those predicted for that discharge in other months. Elevated chloride concentrations observed during the winter are observed during moderate to low flows as well as high flows and probably are attributable to runoff of road
Temporal increases in chloride concentrations in water samples from the Wallkill River, a tributary to the Hudson in southeastern New York, support the correlation between population density and chloride concentrations. The median chloride concentration for this river increased from 16 mg/L during 1970-76 to 36 mg/L during 1980-90; this difference was consistent among all flow conditions and corresponds to a population increase of 20 percent from 1980 to 1990.
Chloride concentrations of stream samples collected throughout the Hudson River basin under low-flow conditions during August 1993 are directly related to population density and fecal coliform counts. Chloride concentrations in samples collected from 42 watersheds throughout the basin during August 1993 ranged from less than 2 mg/L in watersheds with population density less than 0.2 per km2 to more than 100 mg/L in watersheds with a population exceeding 200 per km2. If population density is related to sewage inputs in these 42 watersheds, as it is at the 15 watersheds for which long-term data are available, low-flow chloride concentrations can be used as a regional indicator of sewage input. A positive correlation between chloride concentrations in these samples with fecal coliform counts for watersheds that are less than 30 percent agricultural is further evidence that low-flow chloride concentrations are related to sewage inputs