Hurricane Risk Variability along the Gulf of Mexico Coastline

Abstract

Hurricane risk characteristics are examined across the U. S. Gulf of Mexico coastline using a hexagonal tessellation. Using an extreme value model, parameters are collected representing the rate or λ (frequency), the scale or σ (range), and the shape or ξ (intensity) of the extreme wind distribution. These latent parameters and the 30-year return level are visualized across the grid. The greatest 30-year return levels are located toward the center of the Gulf of Mexico, and for inland locations, along the borders of Louisiana, Mississippi, and Alabama. Using a geographically weighted regression model, the relationship of these parameters to sea surface temperature (SST) is found to assess sensitivity to change. It is shown that as SSTs increase near the coast, the frequency of hurricanes in these grids decrease significantly. This reinforces the importance of SST in areas of likely tropical cyclogenesis in determining the number of hurricanes near the coast, along with SSTs along the lifespan of the storm, rather than simply local SST. The range of hurricane wind speeds experienced near Florida is shown to increase with increasing SSTs (insignificant), suggesting that increased temperatures may allow hurricanes to maintain their strength as they pass over the Florida peninsula. The modifiable areal unit problem is assessed using multiple grid sizes. Moran’s I and the local statistic G are calculated to examine spatial autocorrelation in the parameters. This research opens up future questions regarding rapid intensification and decay close to the coast and the relationship to changing SSTs.

Event-Based Climatology of Tropical Cyclone Rainfall in Houston, Texas and Miami, Florida

“…Increased SSTs allow for Miami to have a longer season, but higher SSTs may cause a confounding effect on those TCs impacting Miami [37]. Rainfall from TCs in Houston averaged nearly 50 mm of rain more than in Miami.…”

North Atlantic Hurricane Winds in Warmer than Normal Seas

“…North Atlantic hurricane climatology is described in a wide variety of articles [39][40][41][42]. All results in this study are displayed using a hexagonal tessellation because hexagons are a geometric representation of a circle (similar to the shape of a hurricane), and hurricane characteristics and risk have been successfully shown using them in the past [16,17,43]. Each of the 36 hexagons has an area of 626,147 km 2 , which is a little less than the size of Texas.…”

“…For a more complete description of the statistical theory supporting this model please refer to Kotz and Nadarajah [51] and Coles [49], and for more applications of this theory with hurricane data using hexagonal tessellations see Trepanier [16] and Trepanier et al [17].…”

“…Hurricane scientists around the world have been attempting to assess the frequency and intensity of hurricanes in a changing climate for over three decades [9][10][11][12]. The approaches to gain insight into this include empirical [13], numerical [14,15], and statistical [16,17], all having their rightful place in the pursuit of understanding. The models show some disagreement due to varying approaches, but strong influences on the variability of hurricanes from multi-decadal and interannual forces, such as the Altantic Multidecadal Oscillation [18] and the El Niño Southern Oscillation [19], are generally agreed upon.…”

“…More sophisticated statistical approaches exist to calculate risk, including generalized Pareto probabilities [24] and quantile regression [25]. Some emphasize the importance in the spatial variability of hurricane risk and utilize methods appropriate to calculate changes across space, i.e., extreme value theory [26] and geographically weighted regression [16,17].…”