An analysis of coastal and inland fatalities in landfalling US hurricanes

Abstract

Improvements in hurricane forecasts allowing for more timely evacuations from storm-surge zones are credited with reducing lethality of US landfalling hurricanes. The deadly reach of a hurricane, however, is not limited to storm-surge zones. About 80% of direct US hurricane fatalities since 1970 occurred outside of landfall counties, with most of these fatalities caused by inland flooding. We construct a geographic information system database combining the location and cause of fatalities, estimated wind speeds, and rainfall amounts for the entire track of the storm for landfalling US hurricanes between 1970 and 2007. We analyze the determinants of total fatalities and deaths due to freshwater drowning and wind. Inclusion of inland fatalities results in no downward trend in lethality over the period, in contrast to prior research. Local storm conditions significantly affect lethality, as one-inch and one-knot increases in rainfall and wind increase total fatalities by 28 and 4%. Rainfall significantly increases freshwater-drowning deaths and is insignificant for wind deaths, while the opposite relation holds for wind speed. While coastal counties do not exhibit a significantly higher amount of lethality risk versus inland counties for total or wind-driven fatalities, freshwater-drowning fatalities occur most frequently in inland counties along the center of the storm path and its outer county tiers as we have defined them.

Combining Surge and Wind Risk from Hurricanes Using a Copula Model: An Example from Galveston, Texas

“…The active season of 2005, with Katrina, Rita, and Wilma, demonstrated that hurricanes can inflict vast amounts of damage, and the need for hurricane research and preparedness has escalated (USGS 2005). The principal hurricane hazards causing these extreme losses are the violent winds and widespread inundation from storm surges and heavy rainfall (Czajkowski, Simmons, and Sutter 2011). This study focuses on the combined risk of high winds and storm surge at a location on the U.S. Gulf Coast.…”

Reply to Maley: Yes, appropriate modeling of fatality counts confirms female hurricanes are deadlier

“…Gaussian models assume that the variance is constant. Count models are based on the probability distributions appropriate to count data, e.g., Poisson and negative binomial, for which the variance increases with the mean (3,4). The negative binomial model is commonly used to adjust for overdispersion when modeling count data such as hurricane deaths (4), as in our data, and was used to construct a well-fitted model (1).…”

“…It is well established that >50% of US deaths by tropical cyclones occur inland (4,5). A total of 89% of Hurricane Floyd's 56 deaths were from inland flooding (1999).…”

A new methodology for classification of tropical cyclones: the importance of rainfall

“…Czajkowski et al (2011) show that the total rainfall and predicted fatalities have an exponential relation, with fatalities starting to occur in areas with more than 5 inches total rainfall (Figure 5.10). Figure 5.11 shows the total rainfall for hurricane Rita.…”

“…The literature review (Chapter 2) revealed that none of the current scales and indexes consider rainfall. The absence of rainfall was identified as a major issue based on data showing that a considerable part of tropical cyclone damage and death toll arise from rainfall (Rappaport, 2014, Czajkowski et al, 2011, Rappaport, 2000.…”