The computer models are in pretty good agreement upon the evolution of the upper-level weather pattern across the United States through the next 72 hours. The three key features will be the following:
1) The westward extent of the subtropical ridge axis across the southeast Atlantic Ocean
2) The depth of the trough across the eastern/northeastern United States
3) The evolution of the cut-off upper low currently across the intermountain west of the United States

The first two are somewhat related to one another; a deeper trough, even by just a hundred miles, will force the western edge of the subtropical ridge to be weaker and/or shifted a bit further east. This will allow for a turn further east, such as is currently shown by the Canadian and ECMWF models. No model is currently showing a full capture and recurvature, however, given the large distance between the storm and trough, with each drastically slowing the storm down as a ridge builds in across the eastern United States. Conversely, a weaker trough will allow for a stronger ridge, allowing the storm to be steered further west before turning northward.

What should we watch for with these things? Take a look at the upper air data from Florida, the Bahamas (if available), Bermuda, and truthfully the entire eastern seaboard. How are observations every 12 hours comparing to the model analyses and forecasts? If they are comparing poorly, in which direction are they doing so? We are still likely a day or two away from getting a high altitude recon surveillance flight ahead of the storm, making what observations we have now of the utmost importance from an analysis standpoint.

The seasonal trend all summer has been for a deeper east coast trough. The normally reliable and trustworthy ECMWF, the best forecast model in the world, has been insistent on one of the most eastward tracks of the entire model suite. It is tough to discount that possibility given those two factors, but it is tough to put much stock in it as well given the large latitudinal difference between Fay and the northeast US trough and the number of other forecast models suggesting a track further west. We (and I) don't want to model hug, but the models are tools that can and should be used with the proper understanding.

The evolution of the cut-off low across the intermountain west will be particularly interesting as well. We have a blocking structure in place, somewhat of a cross between an omega and Rex block, across MT/WY/CO this early morning. Blocks by definition are tough to forecast; models often tend to break them down too early and eject their energy toward the east too early. There is pretty good agreement, however, in this cut-off moving into the south-central Plains by Monday and then eastward from there. How far east this feature makes it by Monday will play a large role in how far west Fay makes it: a cut-off that is slower in ejecting will result, all other factors being equal, in a storm that makes it further west -- perhaps on the western edge of the model guidance. A faster ejecting cut-off will result, all other factors being equal, in an earlier turn. Watching the water vapor satellite loop over the weekend will give us all a good idea how this is going to unfold.

All of this discounts storm-scale processes -- land interaction, convective organization, and so on and so forth -- that will play an important role. A deeper storm in the short term is going to be more likely to "feel" the tug of the troughs and find any weakness on the periphery of the subtropical ridge. A weaker storm is going to be more likely to be steered with the low level flow, implying a further west track at least initially. So far, Fay has generally gone south of model forecasts given the slow organization process partially caused by land interaction and center reformation within the deeper convection. Short story long here, let's see what Haiti and Cuba do with the storm before making any drastic calls.

Intensity-wise, discounting land effects both with Cuba and Florida, the upper level pattern is likely to be favorable for intensification in the Caribbean and Gulf with a well-defined upper ridge northeast of the storm, weak upper-level low just ahead of the storm enhancing upper level diffluence, and upper-level low about 15 degrees east of the storm providing an outflow channel. Heat content is very high in the northwest Caribbean Sea and reasonably high in the eastern Gulf of Mexico.

That's the setup for now. Wanna discuss a long range forecast? Try the Forecast Lounge