Hurricane Dean grazed the southern tip of Martinique overnight. Thankfully, he weakened a bit while passing by, probably lessening the blow suffered by the islands. But now he’s strengthening again: 105 mph and counting, as of 11:00 AM.

Unfortunately, there’s some bad news on the computer-model front. Last night I noted that confidence in the forecast might improve this morning because of newly available data from the NOAA jet. It hasn’t worked out that way. Indeed, the 11:00 AM discussion states, “IT SHOULD BE NOTED THAT GUIDANCE FOR DAY 4 AND 5 IS MORE UNCERTAIN TODAY THAN YESTERDAY.” According to Dr. Jeff Masters, “The NOAA jet mission did not help at all with narrowing down the uncertainty in the computer forecasts for the 4-5 day period, which remain divergent.” See for yourself here.

Why this disagreement? In Dr. Masters’s words, “each model has a different solution for the behavior of an upper-level low pressure system expected to be over the Gulf of Mexico early next week, and there is no is currently no way to guess which model will be right.” It also seems like each weather blogger has a different interpretation of the models. For instance, AccuWeather’s Henry Margusity wrote this morning that “South Texas [is] now becoming the only spot in the U.S. where Dean may hit.” On the other hand, his colleague Jesse Farrell wrote, “This morning’s Model Spread seems to shift the tracks up the coast from southern Texas to evenly cover the entire state.” FLhurricane.com notes the same thing: “the track models have begun to shift a bit north… This means, now more than ever, that the entire Gulf should be watching [Dean].”

Eric Berger, in Houston, looks at the scenarios and their implications for Texas. He sounds worried, and it’s not hard to see why. As Farrell notes, and shows graphically, the GFDL is showing a 916 mb Category 5 monster offshore of Houston on Wednesday. That would be, um, bad. The model image is rather scary:

For his part, however, Alan Sullivan wonders whether we should be trusting the models at all. Looking at “the real world, as opposed to the world of computer models,” he “see[s] something happening that could notch down those intensity forecasts” and alter Dean’s track. Specifically, and I quote his analysis almost in its entirety:

I’m watching the upper low off Florida. It looks mighty sluggish. It’s hesitating in response to a deepening polar trough over eastern Canada. As usual this summer, that polar feature is growing abnormally strong, and digging a bit further south than expected. By holding the cut-off low longer, it changes Dean’s prospects in two ways.

First, intensity: Dean is moving very swiftly. It is overtaking the upper-level cyclonic flow that wraps round the base of the cutoff into the northern Caribbean. Unless the low speeds up, its flow will begin to shear Dean. Already we can see a hint of shear in the storm’s shape. Dean’s envelope of outflow is becoming subtly elongated along a NE-SW axis as the NW flank of its circulation begins to encounter the shear-field. At this point I think Dean will encounter enough adversity in the Caribbean to prevent further intensification until it nears Jamaica. Meanwhile it may even weaken back to category one for a time. OK, this is a wild guess, but I’m a contrarian. (Make sure to include that last sentence if you quote me, Brendan!)

Second, track: Dean will probably continue on an unchanged course as it bucks the shear, but as it gets further west it may begin responding to the pressure by edging right. There is more land to the right. All that much-touted deep warm water means little if the hurricane declines to pass over it. Also, should Dean hit mountainous Jamaica head on, or go even further right and skims near the south coast of Cuba, its intensity will be affected by friction with land. A course close to Cuba would eventually take Dean over the western end of the island and onto the Loop Current in the eastern Gulf. Intensity would probably surge there, and Dean could in time present a serious threat to New Orleans or points east.

Meanwhile, watch that upper low. Run the NW Atlantic vapor loop… That will show you how the Florida low is interacting with the polar flow over the northeastern US. Don’t trust the models. They do not seem to see what is happening.