Composite Climatology of Florida Summer Thunderstorms

Patrick J. Michaels Department of Environmental Sciences, University of Virginia, Charlottesville, VA 22903

Search for other papers by Patrick J. Michaels in
Current site
Google Scholar
PubMed
Close
,
Roger A. Pielke Department of atmospheric Science, Colorado State University, Fort Collins, CO 80523

Search for other papers by Roger A. Pielke in
Current site
Google Scholar
PubMed
Close
,
J. T. Mcqueen Systems and Applied Science Technologies Corporation, 5809 Annapolis Road, Hyattsville, MD 20784

Search for other papers by J. T. Mcqueen in
Current site
Google Scholar
PubMed
Close
, and
D. E. Sappington Department of Environmental Sciences, University of Virginia, Charlottesville, VA 22903

Search for other papers by D. E. Sappington in
Current site
Google Scholar
PubMed
Close
Full access

Abstract

In an attempt to produce an objective climatology of peninsular Florida thunderstorms that does not suffer from observer bias, we composited 9088 hours of high-resolution manually digitized radar (MDR) data and 28 days of daytime satellite imagery. Both indicated maximum activity over the southwestern corner of the peninsula. Radar shows much higher frequencies than are found in other studies, but this is due in part to the method of data collection. We then decomposed the MDR data into Principal Components (PC) to isolate independent spatial and temporal patterns. The most important pattern had a strong diurnal component that was very congruent with coastline geography. Additional important PCs generate the most likely daily regime.

A temporal analysis of the daily march of the radar PC composites over the area of highest activity indicates two discrete diurnal maxima. Over southwestern south Florida (near Flamingo), activity peaks sharply around 2000 UTC. There is another discrete peak in activity approximately 200 km to the northwest two hours later.

Satellite composites also demonstrated that deep cumulonimbus activity over south Florida on undisturbed summer days is strongly focused by the peninsula.

In a statistical analysis at the synoptic scale, interdiurnal variability in the radar PCs is more related to variation at the 850 mb level than variation at the surface.

Abstract

In an attempt to produce an objective climatology of peninsular Florida thunderstorms that does not suffer from observer bias, we composited 9088 hours of high-resolution manually digitized radar (MDR) data and 28 days of daytime satellite imagery. Both indicated maximum activity over the southwestern corner of the peninsula. Radar shows much higher frequencies than are found in other studies, but this is due in part to the method of data collection. We then decomposed the MDR data into Principal Components (PC) to isolate independent spatial and temporal patterns. The most important pattern had a strong diurnal component that was very congruent with coastline geography. Additional important PCs generate the most likely daily regime.

A temporal analysis of the daily march of the radar PC composites over the area of highest activity indicates two discrete diurnal maxima. Over southwestern south Florida (near Flamingo), activity peaks sharply around 2000 UTC. There is another discrete peak in activity approximately 200 km to the northwest two hours later.

Satellite composites also demonstrated that deep cumulonimbus activity over south Florida on undisturbed summer days is strongly focused by the peninsula.

In a statistical analysis at the synoptic scale, interdiurnal variability in the radar PCs is more related to variation at the 850 mb level than variation at the surface.

Save