Isaac reorganizing as it blows through the Lesser Antilles

Tropical Storm Isaac continues to maintain top winds of just 45 mph as its center prepares to move through the Lesser Antilles islands late this afternoon and early this evening. The entire Lesser Antilles chain of islands is receiving heavy rains from Isaac, with Martinique picking up 1.46" of rain as of 2 pm EDT, and St. Lucia receiving 1.49". However, Isaac is not yet generating much in the way of tropical storm-force winds, and none of the islands had received winds in excess of 30 mph as of 4 pm EDT. During their storm penetration to obtain their 2 pm EDT center fix, an Air Force Reserve aircraft measured top surface winds of just 40 mph, and a central pressure of 1004 mb. Top winds at their 5000 foot flight altitude were 49 mph. Isaac is undergoing significant changes to its structure. The plane found the center had become a broad, elongated oval that extended 40 miles from NW to SE. The old center, fixed at 2 pm near 16.1°N, and closer to the dry air to Isaac's north, is being challenged for dominance by a new center that is attempting to form near a burst of heavy thunderstorms at 15.5°N. The Hurricane Hunters' latest fix at 3:50 pm EDT put the center near 15.9°N, a southwards shift of about 17 miles. The resulting battle between centers is giving Isaac a rather odd spiral rectangular shape, as seen on visible satellite loops. The Hurricane Hunters did not observe an eyewall trying to form, and recent microwave satellite images also show no signs of an eyewall forming. A large area of dry air to the north of the storm, as seen on water vapor satellite loops, continues to interfere with development, and Isaac will not be able to begin strengthening until it resolves the battle between it two competing centers, and casts out the dry air infiltrating it. This is going to take at least a day, since Isaac is a very large storm, and it takes more time to spin up a big chunk of the atmosphere. Radar imagery from Barbados and Martinique show plenty of heavy rain showers, mostly on the south side of Isaac where it is moister. There has been a modest increase in spiral banding since this morning.


Figure 1. Afternoon satellite image of Isaac, showing its odd spiral rectangle shape.

Latest model runs for Isaac
The latest set of 12Z (8 am EDT) model runs are very similar to the previous set of runs, which I discussed in detail in this morning's post. The models show a westward track to a point on the south coast of Hispaniola. Isaac's center shift to the south may require some modest adjustments to the south and west for the models. This would result in the storm spending a few hours less time over Hispaniola, and more time over or just south of Cuba. This would slightly decrease the risk to the Dominican Republic, the east coast of Florida, and the Bahamas, but increase the risk to the west coast of Florida. The ECMWF--our best performing model over the past two years--continues to be an outlier among the models. It predicts that Isaac will track just south of Cuba, cross the western tip of Cuba on Monday, then head north towards an eventual landfall in Louisiana. However, this model is keeping Isaac weaker than the other models, and thus predicts the storm will have a weaker response to the trough of low pressure over the Southeast U.S. If the official NHC intensity forecast is right and Isaac becomes a hurricane on Thursday, the more southerly track of the ECMWF is not going to verify, and Isaac will spend considerable time over Cuba on Saturday and Sunday. Where Isaac pops off the coast of Cuba will be critical in determining its future path and intensity, and at this point, we don't know it its more likely that Isaac will go up the east coast of Florida, the west coast, or straight up the peninsula over land. At this point, I'd put the odds at:

40% chance of a track through the Gulf of Mexico, west of Florida
25% chance of a landfall in South Florida, and a track mostly over the Florida Peninsula
35% chance of a track along the east coast of Florida

Which model should you trust?
Wunderground provides a web page with computer model forecasts for many of the best-performing models used to predict hurricane tracks. So which is the best? The best forecasts are made by combining the forecasts from three or more models into a "consensus" forecast. Over the past decade, NHC has greatly improved their forecasts by relying on consensus forecast models made using various combinations of the GFS, GFDL, NOGAPS, UKMET, HWRF, and ECMWF models. If you average together the track forecasts from these models, the NHC official forecast will rarely depart much from it, and the NHC forecast has been hard to beat over the past few years. The single best-performing model over the past two years has been the ECMWF (European Center model). The GFS model has been a close second. You can view 7-day ECMWF and 16-day GFS forecasts on our wundermap with the model layer turned on. Ten-day ECMWF forecasts are available from the ECMWF web site. The European Center does not permit public display of tropical storm positions from their hurricane tracking module of their model, so we are unable to put ECMWF forecasts on our computer model forecast page that plots positions from the other major models. As seen in Figure 2, the HWRF and GFDL were well behind the ECMWF and GFS in forecast accuracy in 2011, but were still respectable. The BAMM model did very well at 4 and 5-day forecasts. The UKMET, NOGAPS, and CMC models did quite poorly compared to the ECMWF, GFS, GFDL, and HWRF. For those interested in learning more about the models, NOAA has a great training video (updated for 2011.)


Figure 2. Skill of computer model forecasts of Atlantic named storms 2011, compared to a "no skill" model called "CLIPER5" that uses just climatology and persistence to make a hurricane track forecast (persistence=a storm will tend to keep going in the direction it's current going.) OFCL=Official NHC forecast; GFS=Global Forecast System model; GFDL=Geophysical Fluid Dynamic Laboratory model; HWRF=Hurricane Weather Research Forecasting model; NOGAPS=Navy Operational Global Prediction System model; UKMET=United Kingdom Met Office model; ECMWF=European Center for Medium Range Weather Forecasting model; TVCA=one of the consensus models that lends together several of the above models; CMC=Canadian Meteorological Center (GEM) model; BAMM=Beta Advection Model (Medium depth.) Image credit: National Hurricane Center 2011 verification report.

I'll have a new post in the morning.

Jeff Masters