19th Conference on IIPS


western region flash flood project

Greg Smith, NOAA/NWS, Salt Lake City, UT

The Western Region Flash Flood Project (WRFFP) is a GIS-based method for producing flash flood potential and guidance (FFG) products in the western region. This project endeavors to produce a more useful and accurate product than recent Modernized FFG methods that have proven to have serious limitations in the West. Initially the project focus is on defining an areas relative potential for flash flooding. This will be accomplished by first asking the question: What physiographic properties of an area make it susceptible to flash flooding and can we identify these? A variety of GIS raster data layers that include information about terrain features, vegetation, forest cover, and soil characteristics will be analyzed in an attempt to answer this question. These layers will be independently re-classified into ten discrete levels corresponding to increased risk of flash flooding. The newly classified layers will then be weighted and combined resulting in a static layer of flash flood indicators (FFI) that describe an areas relative potential for flash flooding. This will then be compared to and verified with actual flash flood observation data. This raster layer of flash flood potential will then be interpolated to FFMP or AMBER basins and field tested at two Weather Forecast Offices in the Colorado Basin River Forecast Centerís (CBRFC) area of responsibility.

Further project enhancement will involve dynamic adjustment of the static FFG grid. Initially we will adjust for seasonal variation (e.g. vegetation state, snowcover), progressing to event variation (e.g. wildfire burn area), and finally to shorter time scales that involve some type of soil moisture component if required. Ideas are currently being postulated how to derive actual flash flood guidance values upon successful creation of a flash flood potential grid. These include the simplest approach of assigning FFG values to each of the FFI classification, spatial regression techniques against data layers of observed flash flood events, incorporation of return frequency precipitation associated with flash flood events, etc. Two important factors of this project are the use of observed flash flood event information and the emphasis that rainfall intensity drives flash flooding more than soil moisture in the West. The project is at the point where field-testing is about to begin. The incorporation of observed flash flood event data layers into the analysis and initial attempts at deriving flash flood guidance values is anticipated by late autumn 2002.

Session 6, GIS Applications
Tuesday, 11 February 2003, 1:25 PM-5:25 PM

Previous paper  Next paper

Browse or search entire meeting

AMS Home Page