FSU "Superensemble" Approach to Predicting Hurrica

[Guest]

Neat Article

http://powerful.fsu.edu/scicol/3/teest.html

[Ripopgodazippa]

Thanks for the link!
I know they keep a tight lid on who can get a peak at the results.

[tropicstorm]

So how has the FSU superensemble performed this season with Charley, Frances & other tropical storms? Why don't we see their track models anywhere? If it's so great, why isn't the NHC jumping on the FSU model bandwagon? Is is top secret & reliable (like in '99) or experimental & suspect?

[WeatherNole]

Here is a more recent article from last week.

http://www.tallahassee.com/mld/tallahassee/9573637.htm

Mike

--

[Storminole]

Nice to see that my alma mater is on the cutting edge of hurricane forecasting. Unfortunately, the approach of Ivan this weekend might require yet another postponement of FSU's game with the UM Hurricanes, now set for Friday night. If that happens, finding another mutually workable date for the game might be a real problem.

[WeatherNole]

Naahhhhhh. (Hopefully)

The current forecast slows the forward speed enough that it should still be south of Cuba on Saturday. My feeling is the game will likely be played.

By the way - tonight's NHC forecaster is Jack Beven - who is an FSU grad too. He was a grad student while I was an undergrad. He is an EXCELLENT forecaster.

Mike

--

[Huckster]

Here is a more recent article from last week.

http://www.tallahassee.com/mld/tallahassee/9573637.htm

Mike

--

"Coming ashore south of Fort Pierce on Saturday. Across the peninsula and exiting into the Gulf of Mexico north of New Port Richey on Sunday. hen wading ashore again near Apalachicola Monday.

That's the path Hurricane Frances will take, Florida State meteorology professor T. N. Krishnamurti said Thursday afternoon."

Sounds interesting. This article was written last Friday.

I just read all the Ivan discussions from TPC. Unless I missed one, it looks like the FSU Superensemble was mentioned in 5 different discussions. Let me quote the context for each one...


11:00 am 9/04

THE OFFICIAL TRACK FORECAST IS SIMILAR
TO THAT FROM THE PREVIOUS ADVISORY. THE GFDL AND FSU SUPERENSEMBLE MODEL TRACKS ARE A LITTLE TO THE SOUTH OF...AND A LITTLE FASTER THAN...THE OFFICIAL FORECAST.

11:00 am 9/06

THE CURRENT NHC TRACK FORECAST IS
VERY CLOSE TO THE LATEST DYNAMICAL MODEL CONSENSUS. IT SHOULD BE NOTED THAT THE LATEST FSU SUPERENSEMBLE RUN IS TO THE SOUTH OF THE OFFICIAL TRACK BEYOND 3 DAYS.

11:00 pm 9/06

THE GFS...UKMET...AND THE FSU SUPERENSEMBLE BRING IVAN
WELL SOUTH OF CUBA TOWARD THE YUCATAN PENNINSULA. NOGAPS AND THE GFDL BRING IVAN MUCH CLOSER TO THE SOUTHERN COAST OF CUBA. THE OFFICIAL FORECAST SPLITS THE DIFFERENCE AND IS LINE WITH THE CLOSELY PACKED MODEL CONSENSUS POINTS.

11:00 am 9/07

HOW FAR NORTH THE HURRICANE WILL MOVE DEPENDS ON
THE INTENSITY OF TROUGH OVER THE GULF OF MEXICO...AND THIS VARIES WITH MODELS. THE OFFICIAL FORECAST HAS BEEN SHIFTED SOUTHWARD A LITTLE BIT AT THE END OF THE FORECAST PERIOD...TO BE IN BETTER AGREEMENT WITH THE GFDL AND THE FSU SUPERENSEMBLE.

5:00 pm 9/07

A TROUGH OVER
THE CENTRAL UNITED STATES AND THE GULF OF MEXICO. THIS TROUGH IS FORECAST TO ERODE THE RIDGE LEAVING A LARGE AREA OF WEAK STEERING CURRENTS IN THE NORTHWESTERN CARIBBEAN AND SOUTHEASTERN GULF OF MEXICO. BY THEN...THE HURRICANE COULD MOVE NORTH OVER WESTERN CUBA AND SOUTH FLORIDA AS INDICATED BY THE GFS...THE UK AND NOGAPS...OR CONTINUE ON A SLOW WEST-NORTHWEST TRACK AS INDICATED BY THE GFDL AND THE FSU SUPERENSEMBLE.

It is interesting to note that all the references to this model indicate a track more to the left or south. This has been the overall trend with the TPC track too. Though I cannot see the FSU superensemble track, from the sounds of it, it's been rather consistent with this storm. If it's track record is as good as many seem to think, given the apparent consistency with Ivan, I suspect Ivan crosses into the south central Gulf via the Yucatan Channel or the Yucatan Peninsula. After that, it's anyone's guess. Whatever the case is, I cannot figure out why so many people hang on every move of these models, especially in the day 4 and 5 time frame. Most of these things have been flip-flopping like crazy. I think a good meteorologist could take an educated GUESS and be more consistent than the GFDL and GFS. At the very least, maybe his forecast would not flip 200 miles every 6 hours. Anyway, hope this is helpful, and if anyone has more info on this, or corrections to my analysis or take, by all means, pass on the information. Just be nice

[bev]

--[/quote]

I think a good meteorologist could take an educated GUESS and be more consistent than the GFDL and GFS. At the very least, maybe his forecast would not flip 200 miles every 6 hours. Anyway, hope this is helpful, and if anyone has more info on this, or corrections to my analysis or take, by all means, pass on the information. Just be nice [/quote]

I think a good amateur could take an educated GUESS and be more consistent than the models. What I've noticed is that the smallest change in data plugged into a model run can make it go haywire.

The models are still a work in progress as evidenced by Charley and Frances.

My main complaint isn't with the models though, it is with the way the NHC is using the models. Their forecasts invariably seem to take the average (middle of the road) of all the models available. It seems to me that they seem to know what the weakness of each model is. Wouldn't it be more prudent to assess the model weaknesses and come up with a track based on the obvious errors in the models?

For instance. Stewart of the NHC seems to do exactly that, while Avilla always takes it right smack down the middle of the models.

Wouldn't want to stick that neck out.... or think too much.

-Bev

______________
7th Generation Floridian

[nolecaster]

I'm an undergrad at FSU right now, and let me tell you there is a LID on the superensemble. It's pretty interesting the way that the model is run, and generally it's damn good.

Can't wait for Firday night vs the Canes, go Noles!

[vbhoutex]

Any idea when they are going to take th elid off? Superensemble on the new computer is now 3 years old if I remember correctly.

Go NOLES!!! Diehard fan since the 60's!!!!

[bev]

Any hackers here?

The URL for the FSU superenesemble is:

http://lexxy.met.fsu.edu/rtnwp/forecasts/ao/index.html

[Greg]

I'd start with trying to guess the meaniing of "lexxy", with two Xs.

[weatherfsu]

The super tight lid on this thing robably won't come off anytime soon. The NHC has access to the forecasts that it spits out so there isn't a rush to make it available to the public.

The thing you have to keep in mind is that it is privately funded. Those companies funding the project don't want it made public just like any company working on something new.

[alicia-w]

It's a public university. How can that be privately funded? I dont know the answer to the question, honestly looking for an answer.

[clearwater]

"Private Sector Funding" and/or "Sponsored Research Grants" are quite common at the university level.

[MBryant]

It's like this:

Take 1 million of private funding.

Mix with 10 million in Gov't grants. (funded by the American Taxpayers)

If something of commercial value is discovered, 100% of commercial value is given to private funder.

Reality 101 (can substitute Cynicism 101)