Streamflow generalization in an area of the central United States

Abstract

Streamflow records of a sample of 41 natural flow stations in an area of the Central Plains were analyzed by electronic computer for the statistics of annual high-flow and low-flow volumes, flow duration, and monthly- and annual-flow discharges. Peak discharges and high- and low-flow volumes for different durations were computed at selected recurrence intervals from statistics of the distributions of the annual values. These dependent variables, discharges and statistics of discharges, were regressed against independent variables of topographic, geomorphic, climatic, and mapped-flow characteristics. The results provide a base for investigating the adequacy of the stream-gaging program and a set of relationships for estimating natural flow characteristics at ungaged sites in the region under study. Comparisons were made of the results of regressions which used only topographic and climatic independent variables with those which also used mapped-flow characteristics as independent variables. The standard error of estimate of the low-flow characteristics was quite large, indicating that for this discharge range, other methods of obtaining this information, may be more accurate or that a quantification of the effect of other factors such as geology on low flow is probably needed to explain part of the present error. Annual and monthly mean flows were defined with standard errors of 18 to 49 percent. Flood peaks and high-flow volumes were defined with standard errors of about 50 percent and 30 percent, respectively, where a significant portion, the residual error, is probably due to sampling error in the dependent variable caused by a short period of record.