Experimental, Satellite, Microphysically-Based, Early Alerts of Severe Convective Storms, Part 2: Evaluation of the Experimental Application

A team of research scientists has been working under the auspices of Woodley Weather Consultants (WWC) on the Second Phase of a Grant from NOAA's Small Business Innovative Research Program (SBIR-2) to develop and test a method to provide satellite, microphysically-based, “Early Alerts” (EA) of severe convective storms (tornadoes, hail and strong straight-line winds). The method objective is to predict when and where a severe weather is most likely to occur one to two hours prior to the actual event. As such, an EA comes after a severe weather watch and before the warning when the event is underway. Part 1 described the methodology, its underlying physics and its application. The evaluation and skill of the methodology is addressed here (Part 2).

The testing and evaluation of this Early Alert method was done at the Storm Prediction Center (SPC) in Norman, Oklahoma in April, May and early June 2008 in a severe weather environment among scientists who are recognized experts in severe convective storms. GOES 11 and GOES 12 multispectral satellite imagery were used in the tests. The intent was to obtain a thorough objective evaluation of the methodology from individuals who are tasked with the forecasting of severe storms. Some of the interactions with the SPC forecasters took place in real time. The presentation will focus on severe storms that occurred over the United States during the period of testing. It will be shown that there is a strong relationship between Early Alerts and subsequent severe weather. Initial impressions are the magnitude of the EA score may be helpful in distinguishing between tornadic and hail events In most cases the EAs occurred in regions already delineated by a Severe Weather Watch. In some cases the EAs and the subsequent severe weather occurred in regions where no Severe Weather Watch had been issued. As such the EAs in these cases might have served as one contributing factor for the issuance of a Severe Weather Watch. This will be quantified in the presentation by determining how often there were severe weather events that were not associated with EA values and SPC Watches and conversely how often there were EA values and Watches without subsequent severe storms (False Alarms). We shall also present evidence that suggests, in some circumstances, EA's may also serve as a significant contributing input to advanced severe thunderstorm and/or tornado warnings, up to one hour before the event.

The ramifications of the findings will be discussed. Early indications are that the EA system will be useful in identifying heightened watch areas, in some cases prompting the issuance of a severe storms watch when none might have been issued without the EA information.