Evaluation and design of a streamflow-data network in Washington

Abstract

A network-design technique developed by Moss and Karlinger (1974) is described and applied in the evaluation and design of a streamflow-data network in Washington State. The technique evaluates the changes in the accuracy of regionalization models that result from changes in lengths of periods of record, or number of stations. The regionalization models for natural-flow streams in Washington are the regression equations for means and standard deviations of the annual mean-flow and peak-flow series and floods with exceedance probabilities of 2 percent. The net work-design analyses indicate that relatively little improvement in the accuracy of estimation at streamflow data--neither in terms of added record length nor added streamflow stations. (Woodard-USGS)