Work

Design and Evaluation of an ESA-based Method of Ensemble Subsetting for a WOFS (Warn on Forecast-like System)

Year: 2024
Type: article
Abstract: Forecasting severe thunderstorm environments in the southeastern United States can be challenging due to mesoscale heterogeneities such as shortwave troughs, pre-existing airmass boundaries, cold fronts aloft, low-level jets, dry air intrusions, and mesoscale lows. To combat these challenges, ensemble sensitivity analysis (ESA) may be applied to a Warn-on-Forecast (WOF)-like ensemble to improve forecasts of severe convection through ensemble weighting and subsetting. Ensemble-based weighting and subsetting uses ensemble members that most accurately represent the thunderstorm environment in areas of mesoscale heterogeneity. This study creates and evaluates the ensemble-based weighting and subsetting in four cases of severe thunderstorm occurrence. The open parameter space for ESA-based ensemble subsetting is also explored within the study which includes subsetting versus weighting, subset size, forecast response box size, p-value in the gridded ESA field, sizes of neighborhood maximum ensemble probabilities, and weighting observations by proximity. The results indicate that ensemble weighting does not lead to any forecast improvement but ensemble subsetting shows promise for forecast improvement. More cases are needed to make further conclusions about other free parameters, but the system shows promise for meaningful forecast improvement which may help improve warning lead times. Advisor: Adam L. Houston
Author: Butler, Daniel J.
Language: English
Open Access status: green
OpenAlex ID: W7037155836
FWCI: 0.133
Cites: 19
Cited by: 1
Related to: 10