Rainfall-runoff modeling is the predominant method used by hydrologic engineers for determining runoff hydrographs. Due to uncertainty associated with predicting future runoff hydrographs used as part of the design of hydraulic structures, historical rainfall-runoff events are used to calibrate hydrologic input parameters. One of the inherent assumptions used when calibrating a model is that the rainfall is uniformly distributed through the watershed at each time step through the storm. This assumption has obvious shortcomings as rainfall intensities are rarely uniform over the spatial extents of a watershed. While the errors associated with the assumption of uniform rainfall may be minimal for small watershed areas (acres), the errors over large areas (tens or hundreds of square miles) make calibration efforts tenuous at best. Version 6 of HEC-RAS has the ability to input spatial variation of rainfall based on radar measurements. The program will then use 2D hydrodynamic equations/methodology to route the runoff through the watershed. Radar rainfall data in hourly increments for Tropical Storm Ida (September 2021) was obtained from the Iowa State’s Multi-Radar Multi-Sensor (MRMS) database and used to analyze the runoff effects throughout multiple basins within New Jersey. Rainfall over eighteen basins ranging in size between 1.99-44.9 square miles as well as one additional large basin (490 square miles) was simulated. Hydrologic abstractions were accounted for using spatial variation of Curve Numbers (CN) based on soil type / land use conditions. Land use conditions were also used to approximate Manning’s n values to account for surface resistance as the excess precipitation volume was routed through the watershed. Multiple simulations were performed with each simulation using different initial abstraction (Ia) ratio’s as follows: no CN, Ia = 0, Ia = 0.05, Ia = .1, Ia = .2. Hydrographs were determined at the outlet to each of the basins and compared to USGS flow records at those same locations. Both peak flow and runoff volumes were compared to USGS data. Peak flow errors ranged from 1.1-261.4% with an average of 38%. Volume of runoff errors ranged from 0.4-200% with an average of 28%. The results of this study show that using spatially distributed rainfall in combination with spatially distributed CN’s and Manning’s n values as part of a HEC-RAS model which uses 2D hydrodynamic equations/methodology to route the excess precipitation is a viable and possibly better alternative to using unit hydrograph theory associated with standard hydrologic models.
