Incorporating Surface Storage and Slope to Estimate Clark Unit Hydrographs for Ungauged Indiana Watersheds.

Application of Soil Conservation Service's (SCS) dimensionless unit hydrograph method for ungauged basins in Indiana yields very high peak flows and short time to peaks for the northern region, thus producing unrealistic flow estimates for design purposes. It is hypothesized that the overestimation of peak flows in northern region using SCS method is due to the flat terrain and high surface storage caused by the Wiconsinan glaciations. To incorporate the slope and storage characteristics, the application of Clark synthetic unit hydrograph (SUH), which incorporates time of concentration (t_c) and a storage parameter (R) to produce runoff hydrograph, is explored for Indiana. A statistical analysis of 29 geomorphic attributes and past storm hydrographs for thirty watersheds in Indiana show that watersheds in north, central and southern regions are statistically different in terms of slope and storage characteristics. A statistical comparison of Clark parameters (R and t_c) estimated for past storm events also support the effect of storage (higher R) in the north, and slope (higher t_c) in the south. Linear and nonlinear regression models of R and t_c against all geomorphic attributes for all watersheds yielded storage and slope attributes as significant independent variables, thus providing a way to incorporate slope and storage into hydrograph predictions for ungauged watersheds through Clark SUH. Additionally, regional regression of R and t_c with geomorphic attributes produced equations that included attributes related to storage and slope for north, and attributes related to land use, slope and stream network for central and southern regions. Validation of regression equations using new storm events for seven watersheds show that the performance of Clark SUH is the best for northern region followed by central and southern regions. The peak estimated by Clark SUH is more than 60% lower compared to SCS method for northern watersheds, but comparable to observed peak. Overall, the use of R and t_c estimated through regression for Clark SUH yield better results compared to SCS method for the entire state including central and southern regions, thus providing more confidence in hydrograph prediction for ungauged basins in Indiana. [ABSTRACT FROM AUTHOR]