An evaluation of methods for the estimation of tributary mass loads

Tributary loading estimation methods were evaluated by conducting retrospective studies with comprehensive sets of field data for flow rates, nutrients, heavy metals, and polychlorinated biphenyls. Three broad classes of loading estimation methods were investigated: simple averaging methods, ratio estimation methods, and regression methods. Estimators were evaluated using Monte Carlo sampling studies in which random subsamples of complete loading records were used to estimate annual loadings. These estimates were then compared to “true” loadings determined by calculations using the entire record. No group of estimators were found to be superior for all test cases considered. However, individual estimation approaches within each group often provided low error estimates. Results were inconsistent among test cases and these inconsistencies appeared to be related to specific test case characteristics such as the strength and form of the flow-concentration relationship and the nature of the annual hydrograph. Ratio estimators appeared to be more robust to sources of bias than other estimation approaches.