ATL: IRMA - Models

[Evil Jeremy]

I think a lot of people are missing the forest for trees....

The model guidance is pretty tightly clustered at the 5-6-7 day range considering we are looking that far ahead. They are all predicting a left turn somewhere around the southern tip of the Florida Penninsula...If we were looking at a track 6 days out that was headed straight for the coast that would be the difference between a Jacksonville and Savannah landfall. With the angle of attack this storm is expected to take, a difference in 75 miles (nothing in a 5 day forecast) is the difference between riding up the east coast or west coast of Florida with impacts up the eastern seaboard or inland to Pensacola/Tallahasee.

While this is not a technical post (I will leave those to the experts) the bottom line is this storm has been very well modeled and forecasted by the NHC. On that note, I am extremely curious to see where the final forecast point is at the 11:00 update, seeing how it hasn't moved for the last 3-4 updates...

I agree. I think Florida's unique geography makes the spread in the models seem larger than it actually is this far out. I'm seeing a narrative that there's a large spread in the models 5 days out, and I'm thinking, compared to Matthew last year, and other troubling storms in recent years, the models are relatively consolidated.

[chris_fit]

The NHC track is closest to the ECMWF, actually. Trend in the ensembles is a little bit of an earlier northward turn. GFS is east of the Canadian & European (ensemble mean). Both the EC & CMC indicate a hit on the southern FL Peninsula. Timing of the turn is the key, and we may not be certain of the turn until it begins this weekend.

wxman, do you find the HWRF run that just went through as plausible or maybe even probable? Just thinking we might get the dirty side.

HWRF has never impressed me at all. They say it's much better (modelers), but I disagree.

What's your call on Irma in RE: to FL 57?

[Blinhart]

They need to stop posting these, these don't mean anything.

What I mean by that it is only 18Z there is no way to know what the 00Z will show, I think these are just on the climatology at that time.

Wait, huh?

None of the 0z models (GFS, Euro, NAM, Canandian, UKMET, COAMPS) have run yet. So what is the 0z model runs plot showing? It just shows the 18z models with the 6hr forecast (or 12z with a 12 hour forecast) as the initial point and it shows the 0z statistical models (BAMS, CLPR, TABS) which are completely useless. Those spaghetti plots show useless models and the useful models are old runs. They give a terrible sense of the uncertainly and I wish they didn't exist.

Thank you for explaining it better than I could.

[Blinhart]

00Z SFWMD plots:

Just another set that isn't up to date info included.

[Blinhart]

I know these are from the 12Z but since the 0Z, there is definitely a move to the West.

http://www.weathernerds.org/tc_guidance ... ng?8695221

[MidnightRain]

NHC's track almost a outlier now.

[Ken711]

NHC's track almost a outlier now.

How so? Euro and HWRF model runs look similar.

[RL3AO]

NHC's track almost a outlier now.

Except it doesn't include a key piece of information. The Euro and its ensembles.

[AutoPenalti]

NHC's track almost a outlier now.

Umm, no it's not.

[Jag95]

Keep in mind I have less than a dozen posts ever. Any chance this left hook shown at 168 hours could happen earlier around North Gulf Coast than later? Pressure looks still pretty low. Any chance of getting the hook early?

Right now I'm more concerned with it taking the predicted north hook and keeping it out of the GOM. Images of Ivan always comes to mind with these types. Getting the left hook is better than the right uppercut. A little different scenario here as always, but until it takes that north direction I'm concerned, even this far west.

[chris_fit]

NHC's track almost a outlier now.

How so? Euro and HWRF model runs look similar.

And JMA I believe

[Hammy]

They need to stop posting these, these don't mean anything.

What I mean by that it is only 18Z there is no way to know what the 00Z will show, I think these are just on the climatology at that time.

Wait, huh?

None of the 0z models (GFS, Euro, NAM, Canandian, UKMET, COAMPS) have run yet. So what is the 0z model runs plot showing? It just shows the 18z models with the 6hr forecast (or 12z with a 12 hour forecast) as the initial point and it shows the 0z statistical models (BAMS, CLPR, TABS) which are completely useless. Those spaghetti plots show useless models and the useful models are old runs. They give a terrible sense of the uncertainly and I wish they didn't exist.

Are the spaghetti plots the so-called early cycle ones?

[ronjon]

Here is what causes the GFS to keep Irma offshore the east coast. Focus on the yellow/orange little circle in southern Alabama at 114 hrs. Thats a little shortwave vorticity at 500mb that erodes the ridge and causes Irma to turn north-northeast abruptly. The large long wave trough is long gone.

https://www.tropicaltidbits.com/analysis/models/?model=gfs&region=seus&pkg=z500_vort&runtime=2017090518&fh=114

[Kingarabian]

Not sure how much weight to put in the Euro since it initiates Irma too weak and then keeps Irma weak for the duration of the track. It has Irma at 962mb which means its being steered by the 300mb-850mb layer when in reality the storm is stronger and at 910mb and being steered by the 200-700mb layer.

[Ken711]

Here is what causes the GFS to keep Irma offshore the east coast. Focus on the yellow/orange little circle in southern Alabama at 114 hrs. Thats a little shortwave vorticity at 500mb that erodes the ridge and causes Irma to turn north-northeast abruptly. The large long wave trough is long gone.

https://www.tropicaltidbits.com/analysis/models/?model=gfs&region=seus&pkg=z500_vort&runtime=2017090518&fh=114

Interesting. So are the Euro and other models now accounting for this shortwave or minimizing its affect?

[tgenius]

Not sure how much weight to put in the Euro since it initiates Irma too weak and then keeps Irma weak for the duration of the track. It has Irma at 962mb which means its being steered by the 300mb-850mb layer when in reality the storm is stronger and at 910mb and being steered by the 200-700mb layer.

Even though Euro may get strength wrong in terms of intensity, track-wise it's much better.

[Blinhart]

Not sure how much weight to put in the Euro since it initiates Irma too weak and then keeps Irma weak for the duration of the track. It has Irma at 962mb which means its being steered by the 300mb-850mb layer when in reality the storm is stronger and at 910mb and being steered by the 200-700mb layer.

And with a system like this, who knows which winds will directly effect her. This storm is going into the history books no matter what, another name that will be retired (no matter the amount of damage or deaths).

[gigabite]

NHC's track almost a outlier now.

How so? Euro and HWRF model runs look similar.

And JMA I believe

The NHC track looks to be focused on the warmest sea surface temperatures, for the Saturday & Sunday points. Katrina & Charley tracked warm water like that was what they were supposed to do. Fortunately the outflows from the Mississippi and Peace Rivers are full.

[MatthewsRevenge]

Doubt that AL shortwave will do much. GFS, NAVGEM and UK are overdoing its power. First it was a trough, now it's a shortwave? OK.

I mean I GET IT, they don't want this to hit ANY land but nothing they do will push this away from any US landfall. Might as well go back to what they were doing earlier...

[HDGator]

Not sure how much weight to put in the Euro since it initiates Irma too weak and then keeps Irma weak for the duration of the track. It has Irma at 962mb which means its being steered by the 300mb-850mb layer when in reality the storm is stronger and at 910mb and being steered by the 200-700mb layer.

Even though Euro may get strength wrong in terms of intensity, track-wise it's much better.

I think the question is does that track accuracy extend to storms of this intensity and extremely low pressure.
The Euro may have performed well with normal storms but what about one that's under 920mb now?
I've had the same questions and I'm speaking from an uneducated stance so any enlightenment is welcomed.