Sampling Strategies for Mass‐Discharge Estimation

Abstract

Methodology derived from sampling theory is presented for the estimation of in‐stream contaminant mass‐discharge or load. The historical record is stratified into relatively homogeneous units in order to approach normality in the subpopulations thereby reducing the inaccuracies introduced by the highly skewed populations. Within strata a ratio estimator is employed to estimate the load and variance of the subpopulations. Strata are also defined to minimize other forms of inaccuracy. Estimates for individual strata are combined to produce pooled estimates for the desired period of interest. With a prior data base, sampling theory is employed to devise sampling strategies designed to produce load estimates of a specified target precision in future studies. Case studies demonstrate the significance of flow events as the primary transport mechanism for suspended sediment and the need to sample such events intensively in order to achieve reliable load estimates. Small “flashy” streams require sampling of much greater intensity during events than larger rivers with attenuated event response in order to produce load estimates of comparable precision.