Evaluation of Flash Flood Guidance in the US

This study quantifies the skill of the National Weather Service's (NWS) Flash Flood Guidance (FFG) product. FFG is generated by River Forecast Centers (RFCs) across the US and then transmitted to local NWS Forecast Offices. Radar estimates and gauge measurements of rainfall are used to determine when FFG is exceeded; this is the NWS's primary method of predicting, monitoring and warning the public about flash flood events. Various methods of generating FFG are available for use by the RFCs.

A national flash flood observation database consisting of NWS Storm Data reports and United States Geologic Survey stream gauge measurements is used to determine the skill of FFG.

FFG skill is calculated at several different precipitation-to-FFG ratios, for each RFC, and for each method of FFG generation (i.e., lumped FFG versus the newer gridded FFG methods [GFFG], and flash flood potential index [FFPI]). Depending on the exact area, FFG generation method, and verification dataset used, the critical success index of FFG ranges from 0.00 to 0.25 (where 0 indicates no skill and 1 indicates perfect skill). This US-wide, detailed analysis of the operational tools used for flash flood monitoring and prediction will serve as benchmarks in future research to improve, modify, or replace the current FFG system. Overall, results indicate little to no skill with FFPI in the western region, although these results should be interpreted with caution given inadequacies in the observations of flash floods and rainfall estimates there. GFFG was statistically similar to the lumped FFG, but offers higher resolution. The best performance using the NWS Storm Data Reports was found with the continuous antecedent precipitation index method that runs operationally in the Middle Atlantic RFC.