ATL IKE: Models Discussion

[pablolopez26]

7PM



4PM


It seems the GFDL stayed put. The BAMMS model is also around the GFDL run.

The rest are right above CC. There is still a significant split. So, its all up to the GFDL now.

[TTheriot1975]

can you post the spaghetti model

[lonelymike]

Most models have shifted north at 00Z.

All models now b/w Aransas Pass and GLS.

Houston...we have a problem.

[Air Force Met]

It seems the GFDL stayed put. The BAMMS model is also around the GFDL run.

The rest are right above CC. There is still a significant split. So, its all up to the GFDL now.

If by right above Corpus you mean 50 miles...then you are right. the closest interpolated model at 00z is 50 miles north of Corpus.

[A1A]

[Air Force Met]

Steve now showing the old models on the TV....

"Incestuous models..."

I like that. The term...not the incest.

[RBDnhm]

Which of the sophisticated models are not "incestuous?" GFS/GFDL obviously are. Are UKMET/Euro based on others?

[pablolopez26]

Look at the NAM model track. Looks like that supercomputer went to lunch.

GFDL has stayed put for the last 2 runs. I dont like that, i see the GFDL as a pretty consistent model.

Oh well, i dont know much about these models anyways!

[sphelps8681]

Well Jim C. is in Galveston if I heard correctly. We know what that means. Ha Ha.

[Serenity]

Well Jim C. is in Galveston if I heard correctly. We know what that means. Ha Ha.

Well how far off is he going to be though?

[TreasureIslandFLGal]

Well Jim C. is in Galveston if I heard correctly. We know what that means. Ha Ha.

Ha ha! I immediately asked where Stephanie Abrahms(?) was, since we saw her keep getting smacked a few years back whenever she went into the field. It seemed they placed her is less likely landfalls than Cantore, but wherever she wound up, they would hit!

-But Nikki just reminded me that she was in the studio tonite. But wherever they do decide to put her....watch out.

[sphelps8681]

Well Jim C. is in Galveston if I heard correctly. We know what that means. Ha Ha.

Well how far off is he going to be though?

I believe he said he went from Port Lavaca to Freeport to Galveston.

It is quite surprising that we N. of Beaumont don't even have School Closures or Evacuations. I did hear that the lower lying areas of Jefferson, Hardin, and I believe Orange county was forcasting Voluntary Evacs.

By listening to the local mets just forcasting TS winds. Seems quite odd. But I could be wrong.

[pablolopez26]

The GFDL model, then, missed the 24-hour forecast by an average of 55.4 miles, doing the best of all the models.

GFDL NWS/Geophysical Fluid Dynamics Laboratory model
HWRF NWS Hurricane Weather Research and Forecasting
UKM United Kingdom Meteorological Service model
NGPS Navy Operational Global Prediction System
GFSO NWS/Global Forecast System (formerly Aviation)
EMX ECMWF (European) global model
OCD5 An early model shown for comparison


http://blogs.chron.com/sciguy/archives/2008/09/an_ike_update_a.html

For the year 2007, then, the UKMET model did demonstrably better at five-day forecasts and three-day forecasts. The obvious conclusion to draw is that, for Hurricane Gustav, we should have been paying the closest attention to the UKMET model.

So let's do that retrospectively. Here's the model's projection at three days before landfall:



Oops. I don't think Gustav made landfall near the Texas/Mexico border. So we see that the model that did the absolute very best with long-term forecasts last year failed with Gustav by 700 miles.

[TampaFl]

2100 UTC steering flow - High pressure over the Texas coast appears to be hanging strong. Ike continues in the weakness between the two highs.

[TTheriot1975]

wow..if that were to stay..then Florida would be back in the picture...lol..just kidding folks...but that would really be a bad thing

[terrapintransit]

wow..if that were to stay..then Florida would be back in the picture...lol..just kidding folks...but that would really be a bad thing

It would certainly catch the Central Gulf by surprise.........

[thetruesms]

AFM,

Why are initializations off by so much. This is frustrating, so many local mets here in Houston still counting on the Corpus landfall and so many in government not pulling the trigger to force anything to happen. I'm worried that in the end many people will be hurt and/or trapped on highways they didn't think they would be on for more than a few hours. Why can't we erre on the side of caution????

In short, a lack of high-density, quality data. All the models have to work with over open waters are satellite data, recon flights when they're out there, and some widely spaced buoy and ship reports. From that, the model has to create its analysis. With all of the interpolation and estimation it has to do, there are bound to be errors

[Stormcenter]

Middle Tx. coast.

[Sabanic]

Middle Tx. coast.

What SC?

[jrbaytown]

Is there one particular model that seems historically to be more accurate than another? Just curious as I have said before I am new to the Gulf / Atlantic Storm Basin Scene.

Maybe this will help you or lead you in the right direction. It gives a summary of the top six models and also talks about the HWRF, which according to the website is supposed to replace the GFDL in 2009:

http://www.wunderground.com/hurricane/models.asp