Radar and Remote Sensing

The State of Flash Flood Prediction Research at the National Weather Center

Zachary Flamig
SoM/ARRC

09 April 2015, 1:15 PM

National Weather Center, Room 1350
120 David L. Boren Blvd.
University of Oklahoma
Norman, OK

The creation of the Multi-Radar Multi-Sensor project, which produces fine resolution quantitative precipitation estimates (QPE) on scales of 1-km2 2-min, has enabled new possibilities for hydrologic forecasting including for flash flood events at any location in the conterminous U.S. This talk will detail the transformation of the Coupled Routing and Excess Storage (CREST) distributed hydrologic model from a global model running at 1/8th degree 3- hour resolution into a fine-scale distributed hydrologic model, called the Ensemble Framework for Flash Flood Forecasting (EF5), forced with the QPE from MRMS. EF5 features water balance components from CREST and the Sacramento Soil Moisture Accounting model. These water balance models are linked to kinematic wave or linear reservoir distributed streamflow routing. EF5 was set up as a quasi-operational system running over the CONUS at 1-km2 5-min resolution. A-priori parameters were derived for the water balance and routing models over the CONUS and then validated on flash flood events on basins monitored by USGS stream gauges.
This system will be included in the MRMS-Flood Locations and Simulated Hydrographs (FLASH) transition to the NWS over the coming year. The software and hardware architecture used for this purpose will be discussed. Details will be given on how the products were evaluated in the first Hazardous Weather Testbed Hydro experiment and future plans for testbed experiments.
EF5 is now being transitioned back to global flood modeling work, including intensive training workshops with our partners in Africa. Future products from our distributed hydrologic modeling suite such as inundation, drought monitoring and ensemble forecasts will also be presented.

Radar and Remote Sensing Seminar Series website