Detection of Severe Midwest Thunderstorms Using Geosynchronous Satellite Data

Abstract

The potential and limitations of detecting severe thunderstorms in the Midwest region of the United States with short?interval (?5 min) geosynchronous satellite data are explored. Thunderstorms are defined in the infrared (IR) data as points of relative minimum in brightness temperature (TB) that have good time continuity and exhibit a period of rapid growth (TB decrease or TB isotherm expansion). Thunderstorms so defined are tracked and monitored and parameters related to storm (updraft) intensity, are extracted from the satellite data. These parameters are rates of TB decrease (both in the upper troposphere and the stratosphere), rate of TB isotherm expansion (at TB≤ 226 K), and storm lifetime minimum TB (Tmin, a measure of storm maximum height). Each parameter is shown to be statistically related to the occurrence of severe weather (tornadoes, hail) on four case study days. The four parameters are combined into a Thunderstorm Index (TI), varying from values of one to nine. Storms with TI≥ 6 have a much higher probability of having severe weather reports and there is a potential warning, lead time of 15 min for the first report of hail and 30 min for the first tornado report. The results are confirmed with an independent case. The appearance of a ?V?shaped? IR cold feature with an embedded warm point (a cold?warm couplet) is also shown to be correlated with severe weather reports and with the satellite intensity estimates. Most storms (75%) with the V?shape have severe weather, but many severe storms (45%) do not have the feature. Limitations of using satellite IR data to detect severe thunderstorms are detailed, including the difficulty of storm identification during certain stages, limitations due to the coarse field of view on current geosynchronous satellites and limitations due to cloud top TB?height ambiguities.