Estimating peak discharges, flood volumes, and hydrograph shapes of small ungaged urban streams in Ohio

Abstract

Methods are presented for estimating peak discharges, flood volumes and hydrograph shapes of small (less than 5 square miles) urban streams in Ohio. Examples of how to use the various regression equations and estimating techniques also are presented. Multiple-regression equations were developed for estimating peak discharges having recurrence intervals of 2, 5, 10, 25, 50, and 100 years. The significant independent variables affecting peak discharge are drainage area, main-channel slope, average basin-elevation index, and basin-development factor. Standard errors of regression and prediction for the peak discharge equations range from +37 percent to +41 percent. An equation also was developed to estimate the flood volume of a given peak discharge. Peak discharge, drainage area, mainchannel slope, and basin-development factor were found to be the significant independent variables affecting flood volumes for given peak discharges. The standard error of regression for the volume equation is +52 percent. A technique is described for estimating the shape of a runoff hydrograph by applying a specific peak discharge and the estimated lagtime to a dimensionless hydrograph. An equation for estimating the lagtime of a basin was developed. Two variables mainchannel length divided by the square root of the main-channel slope (L/VsL) and basin-development factor have a significant effect on basin lagtime. The standard error of regression for the lagtime equation is +48 percent. The data base for the study was established by collecting rainfall-runoff data at 30 basins distributed throughout several metropolitan areas of Ohio. Five to eight years of data were collected at a 5-minute record interval. The U.S. Geological Survey rainfall-runoff model A634 was calibrated for each site. The calibrated models were used in conjunction with long-term rainfall records to generate a long-term streamflow record for each site. Each annual peak-discharge record was fitted to a Log-Pearson Type III frequency curve. Multiple-regression techniques were then used to analyze the peak discharge data as a function of the basin characteristics of the 30 sites.