Extended Range Predictability of Severe Convective Storm Environments in Global Ensemble Forecast System Version 12 Reforecasts

Prediction of severe convective storms (SCSs) on sub-seasonal timescales has become a topic of increasing interest to the research and forecast communities in recent years. While several observational studies have assessed phenomena such as the Madden-Julian Oscillation and atmospheric angular momentum budgets for potential links with SCS activity, studies using operational forecast data have been comparatively less explored. In this study, we evaluate 20 years of weekly Global Ensemble Forecast System Version 12 (GEFSv12) Reforecasts over the CONUS for variables relevant in the prediction of SCSs. As this data is provided for lead times out to 35 days, it allows us to establish a climatology of forecast skill in the extended range for these variables. Using both deterministic and probabilistic verification methods, we find that typical variables used in the prediction of SCSs such as convective available potential energy and particularly composite parameters have difficulty at longer lead times, especially at daily timescales. In contrast, better results are found for synoptic variables such as 500 hPa geopotential height, with skill exceeding climatology extending well into week 2. The aggregation of forecasts into 3, 5, or 7-day windows using rolling means may improve objective verification. We additionally use weekly summations of composite parameters to probabilistically analyze the potential for patterns favorable for SCSs in the extended range. This evaluation of reforecast data will assist in selecting variables that may be helpful in training for sub-seasonal prediction of SCSs in later machine learning experiments and provide a baseline to exceed in said experiments.