Saturday, August 30, 2008

Beyond cable news: Hurricane watch

The wonkish way to keep track of hurricanes is from the folks at Colorado State University. Their current compilation of model predictions for Hurricane Gustav has in one picture more information that you will get from any cable channel.



Anonymous Anonymous said...

I don't know, dude. The National Hurricane Center has a lot of wonkish detail, and it's map complies all of the computers into one "likely" path--with a cone showing other probabilities.

Sunday, August 31, 2008 7:34:00 AM  
Blogger C. Briem said...

sure... but that composite path takes away a lot more info than it adds. Look at the NHC's historical projections of where Katrina was heading and you will see the wide range of where it was predicted to make landfall. You lose all that insight when looking at the one projection.

That and not be callous about the human trajedy, but that other site has something worth looking at which is a model overlaying the hurricane track on top of oil production to make a prediction of how much oil production will be shut in and for how long. The financial news of next week will all be reacting to that as much as anything.

and it's not just oil production. That composite track right now has the hurricane on a collision course with the LOOP, the only supertanker port for the US which if shut in for any amount of time will put a hit on oil imports as well as production.

Sunday, August 31, 2008 10:58:00 AM  
Blogger Bram Reichbaum said...

I've always been partial to The wonk in me likes the many options available under tracking maps, but the kid in me likes Dr. Jeff Master's weather blog.

Sunday, August 31, 2008 1:15:00 PM  
Anonymous Anonymous said...

All three sites are good. I wasn't dissing the good researchers at CSU (or You just need to keep in mind that the NWS track is a composite. The cone it shows is the range of the available computer models; i.e., overlay the CSU tracks and you the get the range of the cone. That's why a hurricane "watch" or "warning" zone can be so long (in terms of linear coastline).

And don't get me started on the lack of isobars on weather maps printed for the public. Anyone with a rudimentary knowledge of weather patterns could predict what was going to happen a few days out by studying patterns and isobars. It is somewhat strange that we know have an overload of weather forecasts today via the internet and TV, but little in the way of raw data. I suppose it is just a ploy by and its ilk to commoditize the product.

Monday, September 01, 2008 8:08:00 AM  
Anonymous Anonymous said...

Oops, I got so lathered that I forgot to say that the NWS site gives you a visual of the probabilities--the projected composite path is the most likely path, but it also shows what could happen. I didn't see that at the CSU site.

Monday, September 01, 2008 8:10:00 AM  
Blogger Bram Reichbaum said...

Well, you don't want to leave it at composite average of those projections, either. If you want to "look inside the numbers" you can probably find reasons for trusting certain models in certain situations ... or if you want to get really addictive, view and listen to descriptions of the most updated jet-airplane photos, and speculate about the eyewall.

Monday, September 01, 2008 5:50:00 PM  

Post a Comment

<< Home