I'm looking at the 18Z GFDL model, which, along with the WV imagery has a truly frightening scenario appearing to be possible.

Now follow this one along folks, because I don't think I'm completely off-base with it.

The WV imagery clearly shows the two low-pressure centers - the foci of the trough(s) involved - over Alaska and just east of southern California. These two features are moving at a decent clip towards the east, with the Alaskan feature, as is the usual for such things, diving south and the other moving slight north of east.

I do not see these combining as some of the models suggest - rather, I see an amplification effect rather than a combinatorial one setting itself up. For them to combine one must basically "run over" the other, with the stronger being the one who does the running over (the Alaskan). It looks to me like the southern feature is going to get here first, before it gets overtaken though. This is consistent with what happens frequently this time of year - you get a first impulse that drives a trough towards the gulf, and a second impulse of energy rides "shotgun" and forces it further south.

In addition the WV loop clearly shows the high near the BOC that has been suppressing northward and west motion backing off, as predicted.

The GFDL appears to think that Wilma gets picked up by a frontal boundary that is approximately N-S off the MS/LA coast by the 23rd, and that the weakness in front of that is what drives the storm to the north and east.

However, the GFDL doesn't "see" the obvious low over the NV/CA border right now, looking at this model run. So....

If we get the "impulse" model, then the storm should run somewhat more north of the predicted track. This puts it in closer to Tampa than Sarasota. What's potentially ugly about this scenario, however, is that it will likely exit closer to Jacksonville and ride up the coast - and if the intensity forecast is to be believed, the GFDL shows (weak) Cat 3 windfields (to the east of center, natch) off Long Island at 00Z 24 October, with a 976 mb central pressure and a wild-looking hybrid system off the east coast.

Plausible? Maybe. Much worse though if its about 75nm west of the forecast position.... and if that combination of troughs does not happen, but we get the "impulse" scenario instead, I'd expect a somewhat left and northward shift..... which could set that up. Given the pounding from the weather that this part of the nation has taken over the last week, a strong tropical system off the NE coast is exactly what they don't need come early next week.

The bad news is the forward motion - this sucker is going to be really hauling butt once it gets picked up, so there will not be a lot of warning in this sort of situation, and a few degrees of error in the turn - or a few hours of error in the timing - could make the difference in terms of endpoint as well.

This one bears watching folks - that same GFDL model shows a small area of Cat 4 winds (not sure I believe that!) on the east coast just as the storm is about to exit back over water. Certainly, a 6-hour transit time over the peninsula looks plausible.... quick - too quick for it to lose much definition and strength in the process.