Close Window

Colorado State University (Jonathan Erdman, COMET graduate student fellow): "Characteristics and parameters enabling improved forecast of cloud-to-ground lightning, heavy surface precipitation, and wet microbursts"

Final Report

It has been well established that a great majority of thunderstorms in Colorado and elsewhere undergo a pronounced electrical lifecycle in which in-cloud flashes dominate the storm's developmental stage, followed by a transition to cloud-to-ground (CG) lightning at the storm's mature state. It has been proposed that CG lightning near a convective core is initiated by an ice mass aloft descending toward the surface; therefore, enhanced CG lightning rates may be a precursor to increased surface rain or hail rates. In addition, several studies suggest a possible coexistence between positive polarity CG flashes and the growth of large hail aloft before storm maturity. Other studies (Seimon 1993, for example) question whether positive CG lightning may be used as a general severe storm precursor. This fellowship program sought to determine if these hypotheses could be applied to nowcasting convection.

The morphological and electrical lifecycles of 11 thunderstorms in N.E. Colorado were examined during the 1995 and 1996 convective seasons. CG lightning flash rates were compared to WSR-88D products, such as the vertically-integrated liquid (VIL), and to public reports of large hail and tornadoes. Two of the hailstorms were observed by the CSU-CHILL polarimetric Doppler radar.

For the nine cases observed by the WSR-88D in July 1996, Erdman presents four correlations of CG flash rates, WSR-88D products, and reports of large hail repeatedly observed in the analysis. Negative CG flash rates and maximum VIL increased together early in the storm's development; however, as the updraft strengthened and VIL continued to increase, negative CG flash rates diminished rapidly and in most cases, positive CG flashes appeared. In six cases, reports of large hail were temporally correlated with non-zero positive CG flash rates, with a mean lead time of 8.5 minutes from the first positive CG flash to the hail report. Nonetheless, peak positive CG flash rates lagged the report of hail in three cases by an average of 8.3 minutes. No persistent trends were found in CG lightning associated with tornado reports.