Methods were developed for estimating unit hydro- graphs at ungaged sites in Montana using either the Clark or dimensionless unit-hydrograph method. Flood hydrograph data for 26 U.S. Geological Survey streamflow-gaging stations and rainfall data were used together with a rainfall-runoff simulation model (HEC-1) to derive unit hydrographs and important unit-hydrograph variables. A multiple- regression analysis relating four unit-hydrograph variables (time of concentration, basin-storage coefficient, Snyder standard lag, and dimensionless peak discharge) to basin characteristics showed a significant (95 percent confidence level) relation only with drainage area for time of concentration, basin-storage coefficient, and Snyder standard lag. In the regression relation for dimensionless peak discharge, the only significant basin characteristic was basin factor, a function of channel length, distance from the basin centroid to mouth, and channel slope. An alternative equation based only on drainage area was almost as reliable. Regression equations for estimating basin-storage coefficient and dimensionless peak discharge had coefficients of determination (r sq) ranging from 0.19 to 0.47. An average dimensionless unit hydrograph was determined for the 26 sites, and a method was developed for adjusting its magnitude and shape to account for site-specific information. The 26 derived unit hydrographs were compared with those calculated by the described estimation methods. The two methods performed about equally well in matching derived unit-hydrograph peaks and shapes. (USGS)
