Mary Schoen

Carnegie Mellon University

This study examines the relationship between changes in precipitation and temperature and the properties of low streamflow to estimate the potential impact of climate change on design-period low flows and associated Total Maximum Daily Loads (TMDLs) of primary pollutants. Stepwise linear regression is used for predicting the future low-flow statistic Q710 using the physiographic and climatic characteristics of 160 watersheds in the Mid-Atlantic region of the United States. We would like to use the MMD data to update the climate predictions of future increases in temperature and changes in precipitation. Previous model results show a decrease in low flow over the 21st century. Using Latin Hypercube sampling of parameter estimates, the fractional change in low flow and the resulting change in TMDL of a point-source primary pollutant are estimated for GCM climate predictions; for most GCM climate predictions, a future reduction in contaminant load will be necessary to meet current water quality standards.