The key to this 4 to 5 day forecast and I won't speculate beyond that is the strength of the 500 mb trough over the central states in 5 days. This mornings GFS had that trough a little stronger than yesterdays runs, all the way down from the northern plains to TX. This feature, as well as the actual forward motion of Felix, are the two key variables to this forecast. One run that bothered me, and I'm sure it was just an anomalous one, was the 06Z HWRF. This model this year was introduced as on par with the GFDL - we'll see after Felix. Anyway, it stalls the storm just east of Belize and then slowly tracks it north along the east coast of the Yucatan. If this storm slows considerably or stalls, in the NW caribbean, then we are in for some major changes with the future track. I don't think that is likely now but still a reasonable possibility. We have to remember that the models are not real good when troughs interact with ridges because the timing is so important to directional movement (i.e. Charley). I don't see a dramatic situation like Charley but again we need to see how sharp that 500 mb trough becomes on Sept 7th.