RECON!!!!

[cycloneye]

http://www.storm2k.org/Weather-Central/ ... _Recco.htm

But it is for sunday at 1800z or 2:00 PM EDT not today nor tommorow but things can change and missions can be made earlier.

[chadtm80]

but things can change and missions can be made earlier.

Or canceled all together

[wx247]

Interesting... the NHC must be thinking this thing has a chance?

[Amanzi]

HMM that is interesting.. guess we will just have to hang on and see if they do go and when. It does tell me though that there is interest down there!

[GulfBreezer]

It has raised alot of eyebrows, our local mets are mentioning it during every weather update now................hmmmmm........ :?

[ColdFront77]

I wouldn't be surprised if recon. pushes up the time to investigate this system. Thirteen days ago they scheduled a recon. into the system in the western Caribbean and it was cancelled a number of hours later. This area of convection has persisted this week, but has remained disorganized.

It was mentioned on my local weather reports as a tropical wave that may bring "at least some moisture" to the state (Florida) on Saturday. This is what one of the Orlando meteorologists said on Tuesday and/or Wednesday.

[Scott_inVA]

Not to be a cold splash of water...

but possible flight is there solely to alert maintenance and flight crew. It does not imply a belief development is eminent or RECCO will be scheduled.

Also, trop models have dropped both GOM and ATL.

Here's the AVN for Sunday night:


Note the GOM system beelining into the Panhandle. FWIW, I am *not* sold on that. Upper climo does not seem right for this to verify.

Invest 93L is a Big Fizzle and the follow-up wave (~11N/37W) has closed and is recurving.

All this actually means little except +RA lower MS Valley/Gulf/SE CONUS...dependant on if and where this comes in.

my spin.

Scott

[BreinLa]

hmmmm this one looks pretty good to me, we'll see what happens when it gets in the gulf

[ColdFront77]

Makes sense, Scott. However, this area of convection does need to be watched.

[Steve]

Really. The AVN? I'm not -removed- development here, but there's not much tropical I trust out of the AVN or GFS. I'll trust my own eyes and some of the better models.

Steve

[Scott_inVA]

Really. The AVN? I'm not -removed- development here, but there's not much tropical I trust out of the AVN or GFS. I'll trust my own eyes and some of the better models.

Steve

Steve,
Based on error range inside 96 hours, AVN was great with TCs last year...with several storms, it was in fact the best.

I often overlay just the AVN and EC... if they are close (especially if the UKie is on board), I start honking.

This all comes second to climo but yes, the AVN has been quite reliable.

Scott

Oh...and the Ted DiBaise quote... cool

[Steve]

>>Oh...and the Ted DiBaise quote... cool

Skandar Akbar: "Let me tell you something Debyousee!"

The AVN did throw together a few decent tracks last year, but I'm skeptical. The model that surprised me most last year in a 1-night stand kind of way was tropical model A98 (sp?). It was closest on Lili.

But yeah, the AVN didn't even appear to be initializing anything down there. But in the 12z run, it puts 'whatever comes up' through central or SE LA and sits it around coastal MS thereafter. So it's trending westward as far as surface flow. At 850 (zonal shear), it takes the 'blob' due north up through the panhandle and Big Bend.

I agree that some energy could end up that way as the trof lifts out, but I still think the place to look is the SW Gulf/BOC on Sunday. We'll see.

Steve

[wxman57]

There really isn't any more AVN - it's the GFS in disquise. Many web sites just haven't updated their graphics ro read "GFS" instead of AVN. The AVN and MRF are now the GFS. Other products might sill mention AVN or MRF, just out of habit.

As for the NHC 98 model, I don't think you want to actually use that for forecasting a storm. It just imports CLIPPER (climatology & persistence) and only very slightly looks at what the atmosphere around the storm is doing. Basically, the NHC-98 model isn't much better than climo. The GFS (formerly AVN & MRF) hit Lili's 5-day track dead on. Lili passed right over th 120hr point of the GFS, but about 8 hours ahead of schedule.

Now the GFS does sometimes have a hard time initiailizing storms. From what I've been hearing the new & improved GFDL is the model to watch this year. The folks at the NHC were saying it has undergone major revisions for 2003. One revision should make the identification of the surface circulation much better. They re-ran the "new" GFDL with Michelle and it hit the track dead-on, as opposed to driving Michelle toward Tampa as it did in 2001(?). I assume they are re-running it with other storm data as well. It'll be interesting to observe the GFDL this year.

[Steve]

wx,

No, I don't rely on that A-98. But heading to landfall, it was the only one on plotted on the WREL map that stuck to its guns and the WC LA landfall.

As for still calling the AVN/MRF = GFS, yeah, it's still a bad habit of mine too. It did okay for Lili, but how many times does it recurve waves run after run a few degrees further west. As for Lili, I think it was somewhat luck. I don't have the model runs saved anywhere, but most of them (GFS included) were all over the place. We'll see this year based on what happens.

As for the GFDL, that remains a big question mark for me. I applaud the enhancements, but I'm wait-and-see mode. I still favor the ECMWF as having the best handle on mid-range scenarios.

TPS

[wxman57]

... regarding the GFDL enhancements. I'll reserve judgment until AFTER this season. I've never been much of a fan for the ECMWF for tropical, and the ETA is absolutely horrible since they "improved it" a year ago. The Canadian is pretty worthless too (the model, not the people). I am a big fan of the GFS, but with <b>one BIG caveat</b> - you have to understand its weaknesses/biases and ignore certain things. It's consistent, if nothing else. Knowing its failures, you can compensate and figure out what a system will "really do". So it isn't always right, but it's wrong in predictable ways (for the most part).

[Steve]

I hear what you're saying there. As to the ECMWF, it seems to do a better job with general patterns if not so much with actual tropical tracks. I've read up on the known model biases, and that which I understand, I can throw out the bathwater without the baby. But if you have time as the season gets rolling, drop some comments/critiques on the models as warranted. That would rule.

Steve

[Scott_inVA]

... regarding the GFDL enhancements. I'll reserve judgment until AFTER this season. I've never been much of a fan for the ECMWF for tropical, and the ETA is absolutely horrible since they "improved it" a year ago. The Canadian is pretty worthless too (the model, not the people). I am a big fan of the GFS, but with <b>one BIG caveat</b> - you have to understand its weaknesses/biases and ignore certain things. It's consistent, if nothing else. Knowing its failures, you can compensate and figure out what a system will "really do". So it isn't always right, but it's wrong in predictable ways (for the most part).

Well, I am not in the GFS fan club!

I will agree the GFS has slightly more value when factoring in blatent bias. It is, however, a non-persistent model meaning it with substantially alter Hts, literally on a run-to-run basis. While reviewing GFS 240 - 168 hours for possible trends (which it often drops only to grab anew inside D3) it makes me crazy inside 5 days. Click here: http://sgi62.wwb.noaa.gov:8080/STATS/html/mnz.html.

The EC for tropicals does honk excessively...but my data (all TC's I was able to ingest the Euro) shows it was in the top 50%...meaning it had a better mean margin of error than over half of the models I ran.

No one lives and dies on any model. As has been stated here, it's *much* better to learn about weather and watch models than it is to learn models and watch the weather.

Scott

[wxman57]

I just finished my 10th talk on hurricane forecasting, and one point I made was that you don't just follow one model. Some models do well with one type of storm, or location, and others do well elsewhere. Some seasonal patterns may favor one model over another as well. You just have to stand back and look at the big picture.

[Lindaloo]

One question. If these models are inaccurate.. then why depend on them or even use them?

[wxman57]

Sometimes one particular model may be <b>very</b> accurate. The key is knowing which model and at what time. Model forecasts are only as good as the data used to initialize them. Out in the tropics, we have very little (relatively) data (actual observations of the environment) to input into the models. Therefore, those who run the models have to use an educated guess as to what values of a number of atmospheric variables to use for the initialization. Guess right, an the model may do well. Miss something (like a weak upper-level low), and the forecast may be way off. That's why it's important to look closely at the 00hr (initialization) map for a model and then compare it to observational data (maybe a WV satellite loop, too) to make sure the model is properly initialized.

Only though experience can one know what model to trust more and when to trust it. Sometimes, like with Bret in 1999, they're ALL bad because the upper-level low in Mexico (which was steering Bret north to Texas) was not seen by any of the models at the 00hr. The models had no clue a large upper-level low was sitting right over northern Mexico, so they didn't realize that they were trying to drive Bret perpendicular to the upper-level wind flow. To make a proper forecast (which we did), we had to analyze the past 24-36 hrs of water vapor satellite imagery, deterimine the upper-low would most likely be stationary, then make our own track taking Bret to the lower TX coast while the NHC kept it heading to Mexico.

Experience is the key!