Since the early 1990s, Doppler radar has been the primary tool for warning decisions for severe thunderstorms and tornadoes. As shown by national warning verification statistics, increased experience with Doppler radar led to a gradual improvement in warning issuances ahead of severe thunderstorm and tornado events, but with no improvement noted in false alarm rate statistics.

Since early 1999, training and operations at the National Weather Service (NWS) office in Wichita have focused on the incorporation of near-storm environment information and spotter reports, as well as radar data, into the warning decision-making process. Mesoscale analysis became a new and important role in warning operations. In addition, warning verification efforts were increased, especially for storms in mostly rural areas.

Results at NWS Wichita have been impressive. For the three consecutive seasons from 1999 through 2001, verification statistics for NWS Wichita were improved over prior seasons. The greatest improvement was in the false alarm rates, which were reduced to levels less than half of the national average, with no negative impact on probability of detection scores.

This presentation provides quantitative evidence of the value of three important steps for improving the warning decision-making process and the resultant verification statistics: incorporation of available near-storm environment data, greater incorporation of spotter observations, and enhanced verification efforts. The 2001 convective season for the NWS Wichita office will be reviewed to illustrate the benefits and potential pitfalls of these steps.