Abstract
Dam breach models and inundation mapping have become increasingly required at both state and federal levels. Many modeling programs exist and vary widely in the equations used, numerical schemes, computational methods, and format of results. While both HEC-RAS and DSS-WISE Lite simulate dam breaches and the resulting inundation using the two-dimensional (2D) shallow water equations, the forms of the equations used are different and they are discretized using different numerical schemes. Additionally, modification of the terrain data to more accurately capture realistic conditions can have significant effects on the inundation results. This study intends to explore the effects of the differing numerical schemes of HEC-RAS and DSS-WISE Lite as well as the effects of terrain modifications in HEC-RAS on the inundation results. A breach of Rancho Cielito Dam in Chino Hills, CA and the resulting flood inundation was simulated using both HEC-RAS and DSS-WISE Lite models. Many sources of data were required to run the models. Terrain data is from InterMap Technologies NEXTMap 5 Digital Terrain Model with a posted resolution of 5 m, resampled to a resolution of 10 m and modified to include reservoir bathymetry by the National Center for Computational Hydroscience and Engineering. Land cover data is from the National Land Cover Database at 30 m resolution, with Manning’s n for different land covers from DSS-WISE Lite. Capacity information for the selected reservoir is from the California Department of Water Resources publicly available data for jurisdictional dams and elevation data collected on site by the Office of Water Programs at Sacramento State. For a Rancho Cielito Dam breach scenario, the HEC-RAS and DSS-WISE Lite “equivalent” models produced similar results for inundation boundaries, maximum depths, and maximum velocities throughout the inundation extent. Some minor differences in inundation boundaries were observed in flat areas with low depths, however the differences are on the order of a few hundred feet at their maximum and the differing boundaries extend and recede across one another at many points – neither model produces a greater inundation boundary at all points. For this case, the differing numerical schemes used by each model do not appear to have a significant effect on the inundation results. The DSS-WISE Lite simulation ran over 100 times faster than the HEC-RAS model. However, several potentially flow-altering features of the terrain were not captured in the pre-loaded DSS-WISE Lite terrain file, which cannot be modified by the user except to remove bridges. These features included freeway sound walls, culverts, and levees which, when included in the HEC-RAS simulation, had a pronounced effect on the inundation boundaries, resulting in a vastly different flow path than that predicted by DSS-WISE Lite and HEC-RAS without terrain modifications. While DSS-WISE Lite runtime is incredibly fast and accuracy appears comparable to that of HEC-RAS, thus making it the model of choice for many dam breach scenarios, the lack of ability to make terrain modifications must be taken into account by the modeler. For areas where significant terrain features are not captured in the terrain data, HEC-RAS offers the modeler a wide variety of options to mitigate this issue.