Probability distributions of gust factors in land-falling hurricanes

Anecdotal evidence from forecasters at the 45th Weather Squadron, Cape Canaveral Air Force Station (CCAFS), suggests that gust factors substantially higher than those found in the literature may occur in tropical cyclones affecting Cape Canaveral and vicinity. This is significant because operational decisions about what hurricane protection steps to take may depend not only on the forecast sustained wind but also on the forecast peak wind. In addition, post-storm damage assessment and analysis has frequently required an estimate of the peak wind experienced by a facility or structure. This study presents probability distributions of gust factors derived from wind tower measurements in 2004 hurricanes Frances and Jeanne. The results facilitate replacing subjective judgments based on anecdotal evidence with quantitative risk assessments.

The data were collected during September 2004 during the landfall of Hurricanes Frances and Jeanne. The two storms followed nearly identical tracks across the Florida peninsula, approaching within less than 180 km (100 n. mi.) of Kennedy Space Center (KSC) and CCAFS. Both storms produced boundary layer winds approaching hurricane force over the Cape. Despite loss of power to many of the sites during the most severe portion of the storms, many hours of wind speed and direction measurements were recorded from a network of 44 wind towers extending from the Cape to well inland. Most of these towers were instrumented at a height of 16.5m (54 ft). This permits examination of the behavior of the gust factor statistics as a function of distance from the coastline. Three of the towers were instrumented at multiple levels from 3.7m (12 ft) to 62.2m (204 ft), and one was instrumented at seven levels from 3.7m (12 ft) to 150m (492 ft). This permits examination of the behavior of the gust factor statistics as a function of height above ground level.

Because of the configuration of the data acquisition system, the windspeed was averaged over five minute periods and the peak one second wind speed within each such period was reported. The gust factor for each period is defined in this study as the ratio of the one second peak windspeed to the five minute mean windspeed for the period. The mean, standard deviation, skewness and kurtosis are computed over selected time intervals at each height on each tower. These statistics are examined to select candidate probability distributions to which the data are then fit. The resulting parameters of the selected distributions are then examined for systematic variations with height, mean wind speed, and distance from the coastline. The work is in progress as this abstract is being written. The extended abstract and conference presentation will present the results of the analysis.