Date of Award

1994

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Robert A. Muller

Abstract

The abundance of recent heavy rainstorms in the southeastern United States has raised concerns that excessive rainstorms may be increasing in frequency and magnitude across the region. For this reason, the heavy rainfall climatology was examined to determine whether long-term trends exist within the historic climate record. In addition, the synoptic conditions leading to these events were examined. The synoptic weather patterns that generated storm events over a 75 mm (3 inches) threshold were classified using an all-inclusive manual weather typing system. Heavy rainfall events in the region were generated primarily by frontal weather systems (79 percent) followed by tropical disturbances (13 percent) and airmass storms (8 percent). Regarding seasonality of heavy rainfall events, the western and central portion of the study area had peaks in frequencies in fall, winter, and spring while the eastern portion had summer peaks in frequencies. Time series of the number of annual events over the 75 mm threshold were used to test for trends. Significant results at a number of sites were found and spatial cohesion in the direction of the Spearman correlation coefficients (number of events and year) was found. The East Coast was dominated by negative associations (suggesting a temporal decrease in events) and areas farther west had positive correlations primarily (suggesting a temporal increase in events). This spatial pattern may be related to fluctuations in the strength and migration of the Bermuda High. Temporal trends in heavy rainfall events by season and by synoptic weather type were also investigated. Trends were found in at least one season at each site analyzed. Also, linear regression of the interarrival times between events (for sites with Poisson distributions) revealed significantly increasing trends in frontal and tropically-disturbed events along the central Gulf coast. Temporal trends in storm magnitudes were investigated through analysis of annual maximum storm series data. Spearman correlations between annual maximum storm events and year found only three significant associations, but the dominance of positive correlation coefficients suggested a tendency toward increasing values region-wide. Thus, recent flood losses may be as much the result of climate as in poor land management and zoning practices.

Pages

194

Share

COinS