Development of a flash flood potential index using physiographic data sets within a geographic information system

A Flash Flood Potential Index (FFPI) grid representing hydrologic response characteristics to heavy rainfall was developed for use in the flash flood warning decision-making process. Physical and empirical relationships exist between rainfall infiltration into the soil and the physiographic properties of soil type, forest cover, landuse, and terrain slope. Datasets representing these four properties were acquired and analyzed within a Geographic Information System (GIS) to produce a grid-based index representative of a flash flood threat. Historical flash flood event locations were used to compare against the gridded index for each data type. The geographic study area used for this comparison was the State of Utah. The FFPI grid was summed to pre-existing flash flood basins that are utilized by the Flash Flood Monitoring and Prediction System (FFMP). Operational testing of the FFPI in the flash flood warning process in Utah resulted in an improvement in flash flood warning verification statistics. Further testing in other geographic regions is desired to evaluate the geographic extent of FFPI applicability. Relationships between the FFPI grid and individual datasets suggest modifying the analysis from the scale of a single grid cell to a river basin scale comprised of several grid cells and testing the sensitivity to the various input datasets to achieve better results.