Probabilistic Precipitation Forecasts and Hydrologic Modeling for Improving Flash Flood Warnings: The 2018 HMT-Hydro Experiment (79)

Steven Martinaitis, University of Oklahoma/CIMMS and NOAA/OAR/National Severe Storms Laboratory, Norman, Oklahoma

J. J. Gourley, NOAA/OAR/National Severe Storms Laboratory

Pamela Heinselman, NOAA/OAR/National Severe Storms Laboratory

Katie Wilson, University of Oklahoma/CIMMS and NOAA/OAR/National Severe Storms Laboratory

Nusrat Yussouf, University of Oklahoma/CIMMS and NOAA/OAR/National Severe Storms Laboratory

 

Abstract:

Flash flood warning decisions have been based on quantitative precipitation estimates (QPEs) and flash flood guidance for many years, and more recently, the hydrologic modeling and QPE comparison products within the Flooded Locations and Simulated Hydrographs (FLASH) system. These products provide deterministic output and lead to a binary warning decision by forecasters. The next evolution of flash flood prediction and warnings focus on two advancements: 1) Probabilistic information and 2) high-resolution short-term forecasting.

The 2018 edition of the Multi-Radar Multi-Sensor (MRMS) Hydrometeorology Testbed - Hydro (hereinafter denoted as HMT-Hydro) experiment will have the ability to evaluate the latest advancements of flash flood prediction using experimental probabilistic hydrologic modeling within the FLASH system and introduces the coupling of the Warn-on-Forecast (WoF) quantitative precipitation forecasts (QPFs). The final output results in a gridded probabilistic flash flood product that conveys the uncertainty of the potential occurrence of flash flooding.

During the experiment, participating forecasters evaluated QPE-forced probabilistic hydrologic outputs through simulated real-time operations to assess the benefits of incorporating hydrologic model parametric and structural uncertainties when issuing flash flood warnings. The use of WoF QPFs within the FLASH system was explored for the first time through a series of archived cases. Flash flood predictability will be investigated using both the deterministic and probabilistic FLASH data and then with the incorporation of 0-3 hour forecasts from the WoF system.

Structured evaluations and group discussions of the various probabilistic data sets will provide further understanding on the usability and effectiveness of these products, as well as guide future efforts of the Forecasting a Continuum of Environmental Threats (FACETs) and WoF projects for flooding hazards. The HMT-Hydro experiment testbed design, evaluated products, and initial results will be presented.