Hurricane season is upon us and with it comes the recurring question: what will the damage be this year? Natural disasters can cause a huge amount of destruction and, in 2017, the US alone spent a record $322 billion (CPI-adjusted) dealing with the damages caused directly by these events. Hurricanes now account for seven of the ten costliest natural disasters in the US and are furthermore responsible for the highest number of deaths caused by natural disasters since 1980.
Two recent papers by Andrew Martinez seek to better understand these destructive storms. Martinez examines how measures of hurricane damages can be improved and estimates the benefits of better forecasts in the context of climate change.
In his first paper, “Improving Normalized Hurricane Damages”(2020), Martinez analyses the historical damages of all hurricanes to make landfall in the continental United States since 1900. He finds that after accounting for building cost inflation, recent damage from individual hurricanes is considerably less than the costliest storms in the early 20th century. Martinez finds that the decline in cost of damages is most likely caused by a combination of adaptation through improving building techniques and the construction of sea walls, better forecasts, and recent hurricanes not directly striking large and vulnerable population centres.
However, because of the effects of climate change, the probability of more extremely damaging hurricane seasons, like the one in 2017, may be higher than previously calculated. To determine whether the destruction can be minimised, Martinez’s second paper “Forecast Accuracy Matters for Hurricane Damage” (2020), evaluates the accuracy of predicting these devastating events and the risk that comes with getting forecasts wrong.
There is no doubt forecast accuracy has improved over time, driven by improved understanding of storm dynamics, the use of satellites to improve data availability, and supercomputers for modelling. Tracking the path of a hurricane as well as the key factors influencing how destructive a storm will be, such as rainfall, storm surge, and wind speed, allows forecasters to tell communities how to respond to an incoming storm. Even small errors in hurricane forecasts can have significant ramifications for storm damage.
Martinez finds that an increase in the forecast error of the path of the hurricane by just one standard deviation (about 25 miles for an average storm) translates on average to up to $9,000 in additional damage per affected household. To put this in context, if an additional 1,000 households were unexpectedly impacted because a storm hit in a location other than was forecasted, it could result in additional costs of up to $9 million, as well as possible loss of life.
Martinez estimates that improvements in forecasting have resulted in approximately $82 billion in avoided damages between 1970 and 2015. Of course, there are costs that come with producing and improving the forecasts, but when these are accounted for, the savings remain in the range of $30-$71 billion. The benefits of better forecasts far outweigh the costs incurred to achieve them.
Despite the advances that have been made, now is not the time for complacency. Climate change is expected to lead to increased intensity and unpredictability, making improved forecasts more important than ever. The 2017 hurricane season was already among the top seven most intense ever recorded and the 2018/19 season (Dorian, Florence and Michael) also broke a variety of records. These early indicators of future storm severity show that maintaining investment in research, as well as long-term adaptation, is likely to be crucial, not only for saving livelihoods, but also lives.
 National Centers for Environmental Information, National Oceanic and Atmospheric Administration (NOAA) Billion-Dollar Weather and Climate Disasters: Summary Stats