This time of year, a lot of what I do is follow hurricanes. Many of the tools that are useful the rest of the year fail miserably with tropical systems.
There are a number of problems. Hurricanes… even big ones like Dennis, are often relatively small enough to fall between the cracks of the numerical weather prediction programs. So the computer models I’d normally follow aren’t particularly helpful.
Hurricanes can be very interesting when they’re far from land – away from radar and surface observations. Our government’s NEXRAD network is worthless until the storm is poised to hit land.
Here are some of the secondary sites I follow to try and get more info than would normally be available.
The spinning radar on the left side¹ is from one of Cuba’s network of weather radars. On any given day, half of them might be out of service. In Cuba, that’s not unusual.
On the other hand, there are seven. That’s a lot for an island of Cuba’s size.
Even though the south coast of Cuba is within range of the Key West radar, there are mountains in the way. I think the Cuban radar does a better job at this position. It’s always surprised me that the Cuban images are on the net. I’ve used their sites for at least three years.
The College of DuPage, a two year college where you wouldn’t expect big time meteorology, has one of the best sites for ‘domestic’ imagery like satellites and radar images. DuPage has the full NIDS suite, meaning you can see the Doppler portion of Doppler radar – winds!
For prediction, I have been paying close attention to the MM5 model being run at Florida State by Bob Hart (originally of North Branford, CT). The MM5 was initially formulated at Penn State and runs on off the shelf hardware (though still beefier than what you’ve got at home).
What makes FSU’s iteration of the MM5 so special is its superior ability to properly see topography, ‘bogused’ data from the actual hurricane (to better set the initial parameters, and sophisticated physics.
Bob is among the smartest people I’ve ever known. It’s no surprise this forecast tool is run under his supervision.
Hurricanes are so difficult to accurately predict, especially using conventional methods. Any improvement in the state of the art becomes a possible life saver! What could be sweeter than doing your job and saving lives?
Finally, I look at the Hurricane Center‘s output. For me, seeing just the Hurricane Center’s work product removes the fun of doing it myself.
I do religiously read their technical forecast discussions. The link, unfortunately, changes each time they issue a new one. Links to NHC products are also on the right side of this webpage and on the Hurricane Center homepage.
¹ – Because this webpage will live on long after Dennis is gone, this is a captured radar image. The link goes to the real thing.