Forecasting Heavy Rainfall Events through the Synthesis of Ingredients-Based Diagnostics

Forecasting Heavy Rainfall Events through the Synthesis of Ingredients-based Diagnostics

Michael Pletcher¹, Mark Klein², David Roth², Andrew Orrison², Jim Nelson²

¹Department of Atmospheric and Oceanic Science, University of Maryland, College Park, Maryland

²National Oceanographic and Atmospheric Administration, Weather Prediction Center, College Park, Maryland

Abstract

Excessive rainfall events, including flash flooding, have consistently been some of the deadliest weather phenomenon in the United States. As such, forecasters are on high alert before, during, and after excessive rainfall events. The Weather Prediction Center (WPC) performs a critical role within the National Weather Service (NWS) providing outlooks and short-term forecasts of such events through the issuance of Mesoscale Precipitation Discussions (MPDs) as well as Excessive Rainfall Outlooks (EROs). In this role, WPC forecasters typically examine a multitude of physical ingredients from observations and short-term numerical model guidance related to the production of heavy precipitation to identify potential significant rainfall events. However, operational forecasters are under time constraints and are unable to examine all parameters before making a decision.

This leads to a discussion of what can be done to ensure forecasters can both accurately and quickly predict significant rainfall events. This research aims to provide WPC forecasters with a “recommender” to identify areas of heavy to excessive rainfall. This entails determining thresholds for the ingredients forecasters examine for excessive rainfall over various geographic areas in the United States. Diagnostics would then be used to synthesize the ingredients and their thresholds to provide forecasters a model-driven index or “first guess” as to where heavy rainfall is most likely to occur. WPC forecasters could then utilize the index to discern whether or not to issue an MPD or ERO for an area.