R.M. Summers S. D. Preston 1997 Water-quality data collected at stream sites in the Patuxent River Basin from 1986 through 1990 were used to estimate loads of nutrients and otherconstituents. Studies were performed to determine the adequacy of the water-quality data for load estimation and to evaluate load estimation methods.A regression-based estimator and a ratio estimator were used to estimate loads. Comparisons indicated that the estimators provided similar levels of accuracy when constituent concentration data were available from the entire discharge range.When high-discharge concentration data were not available, it appeared that the regression-based estimator could overestimate loads of some constituents, whereas the ratio estimator appeared to underestimate some loads. The ratio estimator was selected for application in this study because the temporal inconsistencies in the sampling frequencies and patterns represented violations of the assumptions of the regression-based method.Ratio estimator load-estimate quality varied because high-flow concentration data were not available during some years. Preliminary estimation of the base-flow percentages of total loads was performed by calculating conservatively high and conservatively low base-flow load estimates, to provide limits for the actual base-flow percentage. The highest base-flow percentages--at the Unity, Savage, and Killpeck Creek sites--were for total nitrogen, because nitate from ground-water input isthe largest percentage of total nitrogen at those sites. application/pdf 10.3133/wri964175 en U.S. Geological Survey ; Branch of Information Services [distributor], Estimation of nutrient and suspended-sediment loads in the Patuxent River basin, Maryland, water years 1986-90 reports