Usually I write this blog with residents of South Florida in mind, but I can’t help but comment on how consisentent the National Hurricane Center (NHC) has been on the track forecast of Hurricane Gustav. In fact, from the time the track forecast cone first touched the north central Gulf Coast on Tuesday evening at 5:00 pm (Advisory #7) while Gustav was centered over Haiti, nearly every forecast has had the cone centered on Louisiana. This is in spite of a fair amount of scatter in the computer models, especially in the early stages.
The NHC can’t afford to flip flop back and forth like some models occasionally do. This is called the “windshield-wiper” effect and is not a nice thing to do to emergency managers or the marine community. The NHC verifies its own forecasts as well as most of the available computer model forecasts. Forecasters know which models are the best overall after-the-fact, but never know in advance which model on a specific forecast will be the best.
The NHC builds a consensus forecast by combining, in various ways, the U.S. National Weather Service (NWS) Global Forecast System, the United Kingdom Meteorological Office Model, the U.S. Navy Operational Global Atmospheric Prediction System, the NWS Geophysical Fluid Dynamics Laboratory (GFDL) Hurricane Model, the NWS Hurricane Weather Research and Forecasting Model, the Navy Version of the GFDL Model, and the European Center for Medium-range Weather Forecasting Model. Of course, the NHC Hurricane Specialist can and does tweak the consensus forecast, but for the most part, the consensus is hard to beat consistently.
The Consensus Models usually verify better than the individual operational models. In the case if Gustav, this consensus approach has served us well.