Extracting hydrologic information from high-resolution rainfall forecasts

Operational meteorological models will soon attain the ability to explicitly simulate mesoscale convective events. Except for the very short range, these forecasts will presumably be unreliable with respect to individual convective elements, but may contain useful and accurate information regarding the overall distribution and nature of precipitation.

This paper will explore a way to extract information on the spatial distribution of extreme rainfall that is analogous to a common way hydrologists depict precipitation risk - the area-amount diagram. With this diagram, forecasters can immediately determine the flood and flash flood threat across an area according to numerical weather prediction models. This method of presenting forecast data also suggests a method for verifying forecasts of extreme precipitation which emphasizes correct forecasts of peak amounts instead of correct forecasts of peak location.