Dr. Jeff Masters explains the atmospheric forces that are driving the computer models’ disagreement over Hurricane Gustav’s eventual track:
The ECMWF and NOGAPS models foresee a landfall in the Cancun/Cozumel region on Mexico’s Yucatan Peninsula, followed by a second Mexican landfall south of Brownsville, Texas, early next week. This solution assumes the trough of low pressure moving across the Midwest U.S. late this week will not be strong enough to turn Gustav to the north. The other models predict that this trough will be strong enough to turn Gustav northward, and foresee a landfall on the Gulf Coast between the Florida Panhandle and Texas border 6-8 days from now. The GFDL is the fastest, bringing Gustav to New Orleans on Sunday afternoon. This is a plausible forecast, but at this point, virtually any point along the Gulf Coast has a roughly equal chance of a direct hit by Gustav.
There’s much more in Dr. Masters’s post, including the intensity forecast, the outlook for Haiti and Cuba, and the latest on Invest 95L and the “as yet hypothetical” Invests 96L and 97L. “It is possible we will have three or four simultaneous named storms in the Atlantic a week from now,” he writes. Read the whole thing.