Sources of Forecast Errors for Extreme-Rain-Producing Mesoscale Convective Systems

Mesoscale convective systems (MCSs) produce a large proportion of the extreme rainfall events in the United States and around the world. Considerable progress has been made in the understanding and prediction of these MCSs, but their details remain very challenging to forecast. This presentation will investigate the reasons for errors in the timing, location, and rainfall production in model forecasts of heavily raining MCSs. In particular, the MCSs of 26-28 July 2022, which led to devastating flooding in eastern Kentucky, will be examined. As is often the case for extreme-rain-producing MCSs, numerical weather prediction models correctly indicated that heavy rain would occur somewhere in the general region where it was observed, but with large differences in the location and precipitation amounts. These differences were largely connected to uncertainties in the water vapor in the inflow to the MCS: model forecasts with ample low-level moisture correctly predicted extreme precipitation, but those with relatively dry inflow produced only modest rainfall accumulations. These findings are consistent with other past studies of MCS forecast errors, and underscore the importance of accurately observing boundary-layer moisture in predicting warm-season heavy rainfall.