Developing Rainfall Insurance Rates for the Contiguous United States

Abstract

Historical hourly rainfall data (1950?84) were subjected to spatial and temporal analyses to provide information for developing rainfall insurance rates for the contiguous United States. The dimensions of the study illustrate a balance between insurance needs, funding, and adequate climatic analyses. Assessment of the hourly rainfall data from 2092 stations in the United States revealed that only 211 had data deemed useful to this study. Seventeen regions with similar probabilities of rainfall (rate areas) were defined. The average patterns of hourly rainfall closely resemble the nation's average annual precipitation pattern. Separate seasonal rain-hour probabilities were defined throughout the United States because of marked differences between wetter and drier portions of the year. Temporal analyses of rain hours defined the length of record to use in rate determinations, and how often new rates should be calculated. Long-term trends were not present during 1950?84 in any area, and 78% of the 5- and 10-year values were within ±5% of the long-term average, reflecting generally low interannual variability; however, 15% of these short-term fluctuations deviated greatly (>20%) from average. Results led to the recommendation that rerating should be done once every 5 years in most rate areas and that values of the most recent 25 years should be used for rating. Analysis of in-day hourly rain probabilities revealed major diurnal differences existed during the wet seasons in the central and southeastern United States, and different (night and day) rates were recommended.