I posted a blog on August 12, 2009 saying there is a good reason that the National Hurricane Center only issues a 5-day forecast. That August 12th blog showed 10-day forecasts from the European Center for Medium Range Weather Forecasts (ECMWF) model and from the U.S. Global Forecast System (GFS). The ECMWF model indicated Hurricane Bill would be a little to the east of the Bahamas while the GFS indicated a position in the Bahamas. These positions were valid for 8:00 am today, August 22, 2009.
The center of Hurricane Bill at 8:00 am EDT from the National Hurricane Center advisory was at Latitude 34.0N Longitude 68.4W or about 235 miles west-northwest of Bermuda. That position is well to the northeast of the model forecasts verifying at that time. I estimate the 10-day position error to be near 620 miles for the ECMWF and about 800 miles for the GFS. My previous blog had estimated the ‘average’ 10-day forecast error (over a large sample) to be near 600 miles.
I don’t see much use to a 10-day forecast given such large errors. But that is not going to stop anyone from looking at the long range forecasts from the models that are available on the Internet. We need to remember that the 5-day average forecast error from the NHC is roughly 300 miles. That is a tremendous amount of uncertainty even at the 5-day tme period. If a given location comes within the large cone of uncertainty at the 5-day period, that doesn’t mean that we want the typical person to take action at that point in time. If the 5-day cone starts to move over your community, we just want you to start thinking about what you would do if the threat continues to increase. Let’s learn how to use the 5-day forecast properly before we get too excited about even longer range forecasts.