Hydrologic and Climatic Variability in and Modeling of Streamwater Sulfate Concentrations at Panola Mountain Research Watershed, Georgia, U.S.A.

Abstract:

Variability in streamwater sulfate concentrations at Panola Mountain Research Watershed (PMRW), a small 41-hectare forested watershed near Atlanta, Georgia, U.S.A., was assessed for the period 1996–2012. The source of sulfate at PMRW is predominantly atmospheric deposition of which about 85% is retained within the watershed. Sulfate concentrations increase with increasing streamflow due to an increasing contribution of soil water, which has higher concentrations than that of groundwater. Sulfate concentrations also increased when an intermittent stream in the upper part of the watershed with higher sulfate concentrations contributed larger portions to total streamflow. The highest streamwater sulfate concentrations were observed in hydrologic events that occurred during periods of wetting up after long periods of drought, which were most evident during July through December. These high sulfate concentrations presumably are the result of desorption of sulfate from the shallow soils that accumulated during droughts. Simple concentration-discharge models were fairly weak, with a model R² of about 0.35, but improved to an R² of about 0.4 when adding the ratio of streamflow between the upper part of the watershed and the outlet as an independent variable. Additional model improvements were attempted by separating the samples into various groups based on: (1) time of year that high sulfate values were observed; (2) current drought conditions, and; (3) drought conditions during the previous growing season. The largest improvements occurred when separate models were made based on the drought conditions during the previous growing season with model R²s ranging from 0.43 to 0.67 and improvement was observed across all prior drought conditions. The use of hydrologic and climatic variables and categorization led to an enhanced understanding of the factors that affect water-quality variability, and can strengthen predictions of concentrations and estimates of loads.