Message to Tampa, FL, Regarding Katrina and Weather Forecasting

Here is a very helpful blog: Central Florida Hurricane Center, which seems perfectly-suited for your current needs.

In regards to certain questions regarding how often forecasts are generated, and where one can check out the NOGAPS model....

Updates occur every twelve hours, 365 days a year. The U.S. National Weather Service sends up radiosondes by balloon every twelve hours from its network of stations, as do all Weather Services for all nations worldwide. The times are noon (labeled 12 Zulu, or 12Z) and midnight (labeled 00 Zulu, or 00Z) Greenwich Mean Time, which works out to 5 a.m. and 5 p.m. Eastern Standard Time. The data are quickly gathered, computers run, and several hours later, maps are updated to account for the new data.

I get NOGAPS (also abbreviated as NGP) model results at Monterey, California's Fleet Naval Meteorological and Oceanographic Center (FNMOC) Web Site. I access the loops by clicking on North America NGP in the upper-left-hand corner (GFS is another model), then click on the "loop" button for "Previous 12-hr Precipitation Rate [mm/12hr] and Sea Level Pressure [hPa]." At a minimum, I also check out the 500 millibar maps. My understanding is that NOGAPS is a baroclinic model and GFS is a barotropic model: barotropic models make some simplifying approximations regarding temperature gradients and are thus prone to error in some regards. My experience is that, even in the tropics, where temperature gradients are small, NOGAPS does a superior job with respect to the timing of events, which is really what hurricane path forecasting is all about. NOGAPS doesn't jump and swerve, like UKMET and some of the other models do. I notice that the National Hurricane Center (NHC) tends to poll models in order to make their path forecasts, which isn't a very good idea, since some models are simply better than others. It's like asking 50 people on the street where they think the storm is going to go, and weighing everyone's opinions equally. My experience is the GFS model watches too much TV: the NOGAPS model is better, and probably better than most of the other models.

Another helpful Web Site is Calvert Maryland's Center for Ocean-Land-Atmospheric Studies' (COLA's) Web Site. I worry a bit about the ETA model's tendency to have hurricanes wobble all over the place, but second opinions are always welcome.

Today, NOGAPS seems to be pushing Hurricane Katrina, west, towards the Mississippi delta, which is good for Tampa (and bad for Biloxi). For some reason, the Monday 12Z part of the loop is missing, making it hard to evaluate when the NOGAPS model run says Katrina begins angling NE (which it is bound to do sometime), so its hard to say how different it is than the NHC consensus. Nevertheless, it is probably reasonable to say that NOGAPS is west of the NHC consensus.

The question is, is the NOGAPS western forecast reasonable? By looping the NOGAPS loops repeatedly, and comparing them to the GFS loops, I think what NOGAPS is trying to say is that the midlatitude storm system currently moving through the northern Mississippi Valley (here is the Weather Channel map) is moving too rapidly eastwards to help Katrina angle NE - Katrina is simply too far away to be much affected by this fast-moving storm system. On the other hand, GFS says the storm system is moving at a stately pace, and will strongly affect Katrina's path. Catching the speed and magnitude of midlatitude storm systems is NOGAPS strength, and GFS' comparative weakness, and so right now I'd say - watch out Biloxi!

But, as always with all weather forecasting, the proof will be in the pudding!