1NOAA/National Severe Storms Laboratory - U.S.A.
2NOAA/Storm Prediction Center - U.S.A.
3University of Oklahoma - U.S.A.
Direct observations of severe weather events depend upon many factors and are clearly not a complete record of actual severe weather. Relatively high-quality records tend to be short in almost all locations. The best record for severe thunderstorm occurrences is in recent years (~<10 years) in the central United States, although the record for the strongest events is likely to be more complete for a somewhat longer period, perhaps 50 years or so. As a result, our picture of the occurrence of severe thunderstorms is incomplete at best.
One way to address this incomplete record is to develop relationships between atmospheric parameters, derived from vertical sounding profiles, and severe weather occurrence in locations and times when the observational record is relatively complete. For example, we have found that strong low-level shear and low lifted condensation levels are good discriminators between environments that produce strong and violent tornadoes and environments that produce less severe weather. Then, assuming that the relationships between observed atmospheric conditions and observed severe weather hold true, we can derive probabilities of severe weather occurrence for time and locations where the observed record of severe weather is incomplete. The result is a "synthetic" climatology of severe thunderstorms. We are using observed soundings in the United States to attempt to create such a climatology for time periods when our observational record is poor. The NCAR/NCEP reanalysis data provide an opportunity to do the same thing on an international scale for over 50 years. We hope to present results of a preliminary study to address this problem.