Flash flood monitoring and prediction is considered integral to severe weather operations of National Weather Service (NWS) Weather Forecast Offices (WFO) across the intermountain Western United States. Complex terrain features – which can include dry washes and narrow canyons – are capable of inducing flash flooding with relatively small rainfall amounts. The challenge is becoming familiar with these features in order to better recognize the flash flood threat. An important tool used by NWS WFOs in the flash flood warning and decision making process is the Flash Flood Monitoring and Prediction (FFMP) program, which determines precipitation accumulations by drainage basins rather than by radar bins. A key component of FFMP is its comparison of these accumulations, or Average Basin Rainfall (ABR), against basin-specific Flash Flood Guidance (FFG) provided by NWS River Forecast Centers (RFC). Unless detailed and accurate FFG is used by FFMP, basins with similar ABR will show similar FFG-exceeding values, despite the possibility that these basins may have quite different physiographic and hydrologic characteristics. Unfortunately, the current methods used to generate FFG in FFMP can create inaccurate or smoothed values in the West, thereby limiting the usefulness of this FFMP component. WFOs can locally modify FFG values to better represent basin characteristics, though this can be an arduous task when there are thousands of basins within any one radar umbrella.

To address these deficiencies, the NWS Colorado Basin RFC has developed a set of GIS-based Flash Flood Potential (FFP) indicators for several WSR-88D radar umbrellas in the Western United States, including for those within the NWS Salt Lake City WFO (SLC) County Warning Area (CWA). A variety of static GIS raster data layers that include information about basin terrain features, vegetation, forest cover, land use (specifically for urbanization effects), and soil characteristics have been created to produce ten discrete levels corresponding to increased risk of flash flooding. The newly classified layers are weighted and combined, resulting in a static set of flash flood indicators that describe an area’s relative potential for flash flooding. Although currently available on a hardware platform separate from the operational workstation platform used to run FFMP, forecasters at SLC have benefited from this additional information in the flash flood warning decision making process for the past two flash flood seasons.

This paper will document the added value of using the FFP to enhance the information provided by FFMP, especially to reduce flash flood warning false alarm rates. Several case studies are presented to demonstrate its usefulness in the flash flood warning decision making process at WFO SLC. Plans for future enhancements to the FFP will also be described, including adding dynamic layers to account for other flash flood factors such as antecedent basin soil moisture and wildfire burn impacts. Also possible is the addition of a dynamic atmospheric flash flood potential layer currently under development at WFO SLC using neural network techniques.