Evaluating the Tool used to Issue Flash Flood Warnings in the US

The primary tool used by National Weather Service (NWS) forecasters to issue flash flood warnings in the US is called Flash Flood Monitoring and Prediction (FFMP). The algorithm compares radar-estimated rainfall rates at 1-, 3-, and 6-hour accumulation periods to thresholds, referred to as flash flood guidance (FFG). FFG values are produced at River Forecast Centers (RFCs), and are based on soil moisture states produced from a lumped-parameter hydrologic model. Recent studies have pointed out the scale mismatch between basin FFG values and their comparison to rainfall values on a much finer radar grid. As a result, many RFCs have developed newer approaches in deriving gridded flash flood guidance (GFFG) values.

This study is the first of its kind to evaluate GFFG using FFG as a benchmark. Flash flood observations are provided by three different, independent databases: Storm Data reports collected by the NWS, streamflow observations for basins less than 260 km², and reports that were collected at the National Severe Storms Laboratory in the Severe Hazards Analysis and Verification Experiment (SHAVE). A statistical analysis will be presented using each of the observation databases. A major outcome from this analysis will be optimum rainfall/GFFG exceedance thresholds, which can be directly transferred to NWS forecasters.