365953 Developing Metrics for Mesoscale Precipitation Discussions

Tuesday, 14 January 2020
Emily J. Blumenauer, NCEP, College Park, MD; University of Maryland, College Park, MD; and J. A. Nelson

Heavy rainfall across the country poses a risk of flash flooding, which could lead to catastrophic damage if not accurately forecast. To prepare for flash flooding, the National Weather Service has established tiers of different products: Excessive Rainfall Outlooks, Mesoscale Precipitation Discussions, and Flash Flood Watches and Warnings. A Mesoscale Precipitation Discussion (MPD) is issued 4-6 hours ahead of an anticipated flash flood and is intended to aid a Weather Forecast Office (WFO) in distributing warnings. A challenge faced by the forecasters in the NOAA NWS Weather Prediction Center is the lack of verification for the MPDs they issue. Without MPD verification, it is difficult to distinguish whether a given MPD is effective or ineffective in predicting flash flooding occurrence. To determine the effectiveness of MPDs for 2018, flash flood warning data was compared to MPD issuances. Using Python, the possibilities for effective MPDs and their corresponding flash flood warnings were compiled based on the valid dates, valid times, and affected WFOs. ArcMap will be used to overlay MPDs and flash flood warnings to determine which warnings occurred within an MPD area. Because some MPDs will only contain a portion of a flash flood warning, MPD confidence levels and forecaster feedback will be considered when deciding the criteria for an effective MPD. This project will provide insight for how MPD forecasting can be measured and lead to improved understanding of accuracy by season or situations.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner