Investigation of the Low Short-Range Predictability Associated with the 3 May 2020 Derecho Using a Convection-Allowing MPAS Ensemble

Abstract:

During the morning of 3 May 2020, elevated storms in southern Kansas transitioned to a bow echo mesoscale convective system (MCS) that moved across southern Missouri, northern Arkansas, southwest Kentucky, and Tennessee throughout the day, causing widespread wind damage consistent with derecho criteria. The evolution of elevated convection to a progressive bow echo was not well anticipated by forecasters and convection-allowing models (CAMs) even a few hours before the organization of the bow echo. In this study, a convection-allowing ensemble of the Model for Prediction Across Scales (MPAS) is used to assess the physical processes related to the bow echo formation and to explore why the predictability of the event was low.

A short-wave, mid-level trough embedded in zonal flow was moving eastward near the Kansas/Nebraska border between 0600 and 1200 UTC. A surface stationary front extended from central Oklahoma east-northeastward to Pennsylvania. Elevated convection initiated north of the surface front in southern Kansas near 0900 UTC. Storms initiated and grew upscale into an organized, persistent bow echo MCS with an intense cold pool. The MCS propagated east-southeastward, crossed into the cyclone warm sector, and caused widespread wind damage before weakening in middle Tennessee.

A global MPAS mesh with spacing of 60 km over most of the globe decreasing to nearly 3 km where the derecho occurred is used. Forty 30-h MPAS forecasts are initialized at 0000 UTC 3 May 2020 using 20 initial conditions from the Global Ensemble Forecast System (GEFS) analyses and 20 GEFS 6-h forecasts.

The bow echo formation is highly dependent on the development of elevated convection over southern Kansas in the early morning of 3 May. More accurate MPAS members were able to produce multiple cold pool-driven intense storms and later organized into a bow echo. These members were characterized by a deeper surface low in west-central Oklahoma, which in turn caused greater convergence and warm advection along the stationary front. This case also illustrates decision-making and risk communication challenges faced by severe weather forecasters when utilizing CAM ensemble guidance that provides a solution ranging from a minimal severe weather threat to high-end significant severe potential.