Matthew now has peak winds of 75 mph, making it the fifth hurricane of the season. It is centered about 300 miles south of San Juan, and tracking west at 15 mph.
The storm has encountered unexpectedly high wind shear, which has slowed down its intensification rate, but it is still managed to achieve hurricane intensity. It is the season's fifth hurricane, and only the fifth hurricane to form in the Caribbean in the past five years (the others were Rina '11, Ernesto '12, Sandy '12, and Gonzalo '14).
A band of strong wind shear lies ahead of it now, which is actually not uncommon in the central Caribbean... but it should pass through it by Saturday.
As mentioned yesterday, the long-range forecast track is rather similar to Hazel (1954), but if we start the comparison from the point where Matthew should turn northward, the forecast track out to 5-7 days also resembles Sandy (2012).
Matthew is still expected to continue its westward track through the Caribbean for another 2-3 days, then the model guidance starts to diverge. Some models immediately commence with a northward turn, while others leave it lingering in the Caribbean for at least another day or two, delaying the turn and reducing the predictability.
In this plot of model runs from late last night, the tracks appear to be somewhat clustered, but the timing is quite different. I used this map rather than the most recent because it includes a leading global model, labeled ECMF on here (dark red line). Notice that both the ECMF and UKM lines show a track that is further south and much slower than the others. Any time a tropical cyclone is moving very slowly, it is a sign of weak steering currents and relatively low predictability.
In terms of intensity, nearly all models agree that it will maintain hurricane intensity through the weekend, and then strengthen some more as it heads for Haiti/Jamaica/Cuba. The official NHC 5-day forecast is shown here:
Some key "cone of uncertainty" refreshers:
1) The cone does not indicate the actual level of confidence or predictability. The cone is a fixed size all season long, for all storms and all forecasts. In some situations, the realistic uncertainty in the track forecast is greater than what the cone portrays, and sometimes it's less.
2) The cone is only designed to enclose the storm's position with 2/3 probability... there is historically a 1/3 chance the center will track outside of the cone.
3) The cone does not indicate where the impacts will be experienced. Impacts from a hurricane will extend beyond the cone, even for a perfect forecast.
For now, any U.S. impacts are too uncertain to worry about, and would be at least 5-6 days away at the earliest (south Florida would be the closest). But in the shorter term, Jamaica, Haiti, and Cuba are in Matthew's sights early-mid next week, and heavy rain is a risk even far from where the center tracks.