How quickly can we adapt to change? An assessment of hurricane damage mitigation efforts using forecast uncertainty

Our ability to adapt to extreme weather is increasingly relevant as the frequency and intensity of these events alters due to climate change. It is important to understand the effectiveness of adaptation given the uncertainty associated with future climate events. However, there has been little analysis of short-term adaptation efforts. We propose a novel approach of using errors from hurricane forecasts to evaluate short-term hurricane damage mitigation efforts. We construct a statistical model of damages for all hurricanes to strike the continental United States since 1955. While we allow for many possible drivers of damages, using model selection methods we find that a small subset explains most of the variation. We also find evidence supporting short-term adaptation effects prior to a hurricane landfall. Our results show that the 67 percent improvement in hurricane forecasts over the past 60 years is associated with damages being 16-63 percent lower than they otherwise would have been. Accounting for outlying observations narrows this range to 16-24 percent.