The letter comments on a paper that received widespread media attention. Engineers from Carnegie Mellon University found that a hypothetical wind farm located offshore Galveston, TX, would have a 30% chance of losing half of its wind turbines and a 41% chance of losing ten or more over a typical 20-year wind-farm lifetime. This study was covered by USA Today, the New York Times, and other media outlets, and cast uncertainty on the nascent offshore wind-energy industry in areas susceptible to hurricanes.
An error and several poor assumptions in the original approach were found. In a new analysis of the same hypothetical wind farm, we found an average loss of 2 turbines per 20-year period, compared to 24 using their methods. In their response to the letter, the revised approach downgrades the risk from a 41% to a 2% chance of losing 10 or more turbines over a 20-year period. This is over an order of magnitude downgrade in the risk, but still relies on a flawed approach using fits to statistical extreme value distributions. We recommend the use of industry-standard risk models similar to those used to estimate risk in the residential insurance market in Florida.
This letter can be accessed at http://www.pnas.org/content/early/2012/07/23/1206189109.full.pdf