Abstract
Tornado outbreaks are high-impact events, often causing significant loss of life and property. This study evaluated the forecast skill of U.S. tornado outbreak activity using the Global Ensemble Forecast System (GEFS), version 12, at lead times up to 2 weeks. Tornado outbreak activity is represented in GEFS using an outbreak index, which relates the likelihood of outbreak-level tornadoes to GEFS forecasts of convective precipitation (CP), storm-relative helicity (SRH), and convective available potential energy (CAPE). GEFS forecasts of the outbreak index are verified against smoothed Storm Prediction Center report data. Since the performance of the outbreak index depends on how well GEFS predicts the index constituents, we also evaluated the climatology and forecast skill of CP, SRH, and CAPE, as well as their covariability. We found that GEFS has a systematic low-CAPE bias and that the forecast skill of the outbreak index is most limited by GEFS forecast skill of CP. We corrected the low-CAPE bias and index seasonality errors via a seasonally and regionally dependent scaling of CAPE and the index, which improved the seasonal cycle and forecast skill of CAPE and tornado outbreak activity in GEFS. Overall, on average, GEFS has its highest forecast skill of tornado outbreak activity during winter and spring—in some cases, positive skill extends beyond week 1 forecast leads—and has its lowest forecast skill during summer.
Significance Statement
Tornado outbreaks—when multiple tornadoes occur over a short time span—are one of the most extreme forms of severe convective storms. Individual tornadoes are typically regarded as unpredictable except at very short lead times, but broad tornado activity or likelihood might be predictable past the weather time scale (a week or more in advance). We evaluated the forecast skill of U.S. tornado outbreak activity as well as the forecast skill of environmental conditions relevant to the favorability of tornado outbreaks in a state-of-the-art forecast model. We found that one environmental condition that describes storm instability or “fuel”—convective available potential energy—is often too low in the forecast model. We also found that tornado outbreak activity as represented by model environmental conditions has seasonal cycle errors. After correcting for these issues, we determined that the forecast skill of tornado outbreak activity is highest in winter–spring and is lowest in summer. Overall, winter, spring, and fall forecasts of U.S.-wide tornado outbreak activity are skillful past the weather time scale.
Malloy’s current affiliation: Department of Geography and Spatial Sciences, University of Delaware, Newark, Delaware.
© 2025 American Meteorological Society. This published article is licensed under the terms of the default AMS reuse license. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).
