Thursday, 20 July 2023: 9:15 AM
Madison Ballroom B (Monona Terrace)
Dan Butler, Univ. of Nebraska Lincoln, Lincoln, NE; and A. L. Houston, G. Limpert, C. K. Potvin, and C. A. Kerr
Mesoscale ensemble numerical weather prediction is now readily available for operational forecasting guidance at convection-allowing grid spacing. However, time between forecast initializations means idle observations are not used but could improve forecasts of high impact weather. Ensemble systems can utilize observations in the free forecast period and ensemble sensitivity analysis (ESA) to potentially increase the lead time of convection initiation. ESA adds value to adjusting ensemble guidance through a weighted or subsetted ensemble since it highlights areas where model errors would highly degrade the predictability of a forecast by ranking individual members by the least error in sensitive regions. Adjusting ensemble guidance is not a new concept within a WoF framework, but examining the sensitivities of the parameter space of an ESA-based weighting/subsetting system is crucial to assess its potential operational value.
The purpose of this study is to design and evaluate a method for developing an ESA-based weighted ensemble/subset for use in forecasting thunderstorm initiation using a WoF-like system. Four severe local storm events were used to optimize an ESA-based weighting/subsetting system. Weighted ensembles led to negligible forecast improvement while subsets had meaningful forecast improvement compared to the original ensemble mean. More cases need to be considered to optimize an ESA-based subsetting system but such a system shows promise for achieving meaningful gains in forecasting skill on watch and warning lead time scales in rapidly evolving thunderstorm environments using statistical post-processing of probabilistic forecast guidance.

- Indicates paper has been withdrawn from meeting

- Indicates an Award Winner