Hurricane Sandy

[piemma]

They don't have a solid feel at all.

The latest set of 00Z (8 pm EDT) and 06Z (2 am EDT) computer model runs are in substantial agreement for the next 3 days, but Sandy's future is as clear as mud after that. Sandy will continue to punish the Bahamas today and Friday, as it tracks north to north-northwest. Sandy will probably come close enough to the Southeast U.S. on Saturday afternoon to spread heavy rains to the coasts of South Carolina and North Carolina. However, the 4 - 6 day computer model forecasts for Sunday - Tuesday diverge widely. The GFS model, which has been one of our two top models for predicting hurricane tracks the past two years, has been very inconsistent with its handling of Sandy. Runs of the GFS model done 6 hours apart, at 8 pm last night and 2 am EDT this morning, were 300 miles apart in their position for Sandy on Tuesday, with the latest run predicting a landfall in Maine on Wednesday morning. On the other hand, the ECMWF model, our other top model for predicting hurricane tracks, has been very consistent in its handling of Sandy. The ECMWF model has Sandy hitting Delaware on Monday afternoon, the same forecast it has had for three consecutive runs. The other models tend to follow one extreme or the other, and NHC is picking a solution somewhere in the middle of these two extremes. An extra set of balloon-borne radiosondes is going to be launched at 2 pm EDT Thursday all across the U.S., which should help this evening's model runs. Extra radiosondes will be launched every 6 hours through Saturday afternoon.