Hurricane Followup

I posted a comment the other day on hurricanes and prediction. One reader responded with a newsletter (PDF) from his firm put out in fall 2005. An excerpt:

As successful as the meteorological agencies have been in predicting the eventual scale of a cyclone, there is still a random element in this type of weather pattern. All the current early-warning models use a statistical distribution to model the hurricane characteristics and hence will always be vulnerable to statistical outliers. When this does occur, as in 2005, there is a tendency to seek explanations for these outliers from current events, for example global warming. People place too much faith in the predictive models used to forecast hurricanes and should rather be setup to react appropriately when these storms arrive.