The science behind tracking hurricanes

[unome]

found a good article from Corpus Christi's caller.com on hurricane models

http://www.caller.com/news/2011/jun/18/ ... urricanes/

[wxman57]

"The National Weather Service's cone of probability, the large three-day and five-day spreads showing the likely area for a hurricane landfall, are taken from an aggregation of the top-performing models."

HUH?? The cone represents a 5-year 66.7th percentile, meaning that the center of the storm will track within this cone 66.7% of the time based on errors over the past 5 years. It is not derived from models. Generally good article, though.