Hurricane forecasting: No longer a Gray area
Posted by Paul Nunn | Archive on Tuesday 13 July 2010, 2:27PM
Competition in hurricane forecasting methods continues to drive innovation
It wasn’t so many years ago that pioneering Dr Bill Gray of Colorado State University had a monopoly when it came to capturing underwriters’ attention as he announced his forecast for hurricane activity in the upcoming Atlantic season (June – November).
Fast forward to today and things have changed considerably. Dr Gray has passed on the baton to Phil Klotzbach (who continues to refine theCSU approach), while a host of other organisations have entered the fray with their own techniques. As is often the way, competition drives innovation, both in forecasting methods (typically statistically-based but increasingly dynamic computational models, or a blend) but also in measures of forecasting “skill” – as each team attempts to demonstrate more reliable forecasts than the others.
A relative “new-kid-on-the-block” is Matt Huddleston at the UK Met Office who is throwing the phenomenal computing power of their GloSea model at the problem (click here for good explanation of what they do and why it has such promise from Dr Jeff Masters at Weather Underground). If the approach is promising, the 2010 forecast is horrible – calling for 20 named tropical storms (70% chance of 13 – 27 storms) and an ACE Index (Accumulated Cyclone Energy) of 204, both measures around double the climatological averages.
NB: While it is nice to have some organisations like the Met Office that make me proud to be British, I kinda hope they’re wrong!