An operational approach for integrating multiple high resolution model solutions during a high impact event: the 22 May 2010 EF4 Bowdle tornado

On the evening of 22 May 2010, an outbreak of eight tornadoes occurred over north central South Dakota including a destructive EF4 near Bowdle, South Dakota. The Bowdle tornado was the first EF4 tornado to strike South Dakota since 2003 and only the seventh EF4 recorded in the state since 1950. Operational and short range ensemble forecast (SREF) products leading up to the event did indicate kinematic and thermodynamic parameters favorable for severe weather. However, there was significant uncertainty regarding convective initiation (CI) and convective mode prior to the event. This was partially due to strong convective inhibition (CIN) over the northern plains associated with a strengthening elevated mixed layer (EML). Previous studies have demonstrated that deterministic high-resolution guidance has shown skill in predicting several facets of severe convection, most notably convective initiation and convective mode. More recently, a real-time convection-allowing storm scale ensemble forecast (SSEF) system developed by CAPS has been employed at NOAA Hazardous Weather Testbed (HWT) with the goal of displaying uncertainty information inherent to the SSEF system. However, due to the experimental nature of this project, real-time information from the SSEF is not available to NWS operational forecasters. Subsequently, this study analyzes output from multiple high resolution models leading up to the 22 May 2010 tornado outbreak. The models used were all available in real-time via websites or run locally at the NWS Weather Forecast Office. The goal of the research was to assess the potential value of this “poor man's” ensemble approach in real-time forecasting situations, such as this high impact, tornadic event. This investigation demonstrates that forecasters may acquire many operational benefits if output from multiple models is examined in real-time.