StormHound -- actually, the FSU Superensemble is nothing more than a least squares method of finding the biases in each of the member models (GFS, NOGAPS, etc) and determining a track forecast based off of the expected least errors in each of the tracks. A least squares fit isn't exactly an average, but it's a means of getting a weighted average sort of thing.

There is no special data that goes into it that doesn't go into the other models; that's what makes the left bias it has exhibited for several days now so striking, in that it picked up on the tendencies of the models to the point that any leftward shift in those tracks will send the Superensemble a bit further left as well.

That graphic on the front page about the crossing tracks is very telling. You see where they intersect? That's directly over my place in Orlando. Reports from that region say that items are at a high premium even now, particularly at Wal-Mart where they can't even get things onto the shelf before people take it out.

To clarify a point about the poor initializations of the GFS/GFDL models...the 00z runs will have that new data into them, but it's still up to the models to initialize it correctly. It may well assign a minimal weight to that data and go with something lower anyway -- that's what they have been doing. It shows why you cannot rely on the models solely and must use meteorological intuition -- several here have noticed the greater strength of the ridge and kept it on a more westerly/southerly track. Unfortunately, the NHC is very much tied to the models these days; it's partially the trend in forecasting and partially political/governmental pressure. But, the past two discussions have been very nice departures from a straight-up model discussion, and I would not be shocked to see the 5a track shifted further west...and a much slower weakening of the storm over land. It'll be moving fast and will take a decent amount of time to spin down; I think 40mph near Gainesville is severely underestimating things.

More as I read through the thread...