Global model runs discussion

[Riptide]

18z GFS does show a couple of storms in the long range but quiet until then.

Somehow, it loses the CV low at 108 hours. This model is really bugging me out, literally.
Lets take a look at the shear in the area.


As you can see, all is well and it should continue for the foreseeable future. I think the GFS will only be useful for winter and severe weather now.

[srainhoutx]

There are many new parameters to the GFS. I suspect strongly that the NHC/TPC pushed to see this new deployment. If you recall this is the second round of improvements/upgrades to the GFS since December. I choose to us all the 'tools' in the tool bucket myself.

[cycloneye]

The 00z run of GFS shows nothing in the tropics for 192 hours. Definitly,I think the old one has to be put on again and this new one taken out to fix the problem it has in the tropical areas.

[Ivanhater]

Intense hurricane on the 00z run in the long range

[Scorpion]

Well at least we're finally getting something intense on the long range . Good sign IMO

[cycloneye]

But anything at short nor medium range.

[Ivanhater]

But anything at short nor medium range.

Actually it does Luis, check out my post in the wave emerging off Africa thread...it is very similar to last night's Euro on the 850 vort. Some reason the surface reflection do not coincide which may be associated with the parameter problems mentioned earlier..

[ROCK]

looks like we are stuck with the CMC, NOGAPS, UKMET and the ECM...for now....which sucks because all of these run only twice a day......still feel the CMC will be the one to watch for developement then use the ECM for tracking...JMO..

[Stormcenter]

Sorry but intense hurricanes are not good signs for anyone. IMO

Well at least we're finally getting something intense on the long range . Good sign IMO

[Jagno]

Sorry but intense hurricanes are not good signs for anyone. IMO

Well at least we're finally getting something intense on the long range . Good sign IMO

Stormcenter; I honestly don't think they were referring to the storms itself being a good sign but rather the fact that this particular model is finally picking it up despite some issues it's having. These are the wrong folks to get uptight and defensive with because they'll be the same ones to see you through each and every one of them. I've been around a few years as you have and have learned first hand that I'd rather have this board than MSM any day when a storm threatens my area. Sometimes we let our unknown fears guide our thoughts and keyboard strokes. Let's try to step back and see things from a positive perspective okay. it's gonna be a long season otherwise. LOL

[somethingfunny]

Did the new GFS miss Alex? It had a different solution than the old GFS but if I recall they both developed Alex and busted the track in the mid-long range. None of the models had a very firm latch on Bonnie. So far the new GFS is performing just fine. I think that what we have here is a case of confirmation bias now that the NWS went ahead and said there's a problem.

My thinking is if the new GFS consistently underestimates windspeeds in the tropics by 20kts, then it's probably not going to have very many false positives at all. Maybe some false "all-clear" runs though...as I suspect the current quiet run is false. If the new GFS develops a tropical system despite its weak bias, that could be a very clear indication something big will be happening. Don't think of it as the death of an old and trusted model, but rather the birth of a new and exciting model.

I do wish they'd keep running the old GFS, but I understand that it's not easy to double the amount of supercomputing power devoted to GFS overnight. They have to choose one or the other.

[cycloneye]

For a second day in a row, HPC talks about the not so good new GFS. I hilighted the most revealing parts of their thinking about this.

http://www.hpc.ncep.noaa.gov/discussions/fxca20.html

TROPICAL DISCUSSION - INTERNATIONAL DESKS
NWS HYDROMETEOROLOGICAL PREDICTION CENTER CAMP SPRINGS MD
836 AM EDT THU JUL 29 2010

PRELIMINARY DISCUSSION FOR PUERTO RICO AND THE USVI. A NEW GFS
BECAME OPERATIONAL DURING THE 12 UTC CYCLE YESTERDAY. AT UPPER
LEVELS IT RESOLVES A TUTT LOW TO THE NORTH/NORTHEAST OF THE
ISLANDS...WHERE IT IS EXPECTED TO MEANDER DURING THE NEXT FEW
DAYS. THE TUTT ALOFT IS TO ALLOW FOR MOISTURE TO SURGE ACROSS THE
NORTHEAST CARIBBEAN EARLY IN THE CYCLE...BUT THROUGH TOMORROW
MORNING THE PWS ARE TO DROP. IN CONTRAST TO THE OLD GFS...THE NEW
ONE DOES NOT FORESEE MUCH MOISTURE REACHING THE FORECAST AREA
DURING THE WEEKEND. AT LOW LEVELS THIS NEW MODEL SEEMS TO BE
HAVING PROBLEMS WITH PERTURBATIONS IN THE LOW LEVEL
EASTERLIES...AS IT IS NOT RESOLVING THESE WAVES TO THEIR FULL
EXTENT. I AM LOOKING FOR ANOTHER ALTERNATIVE TO THIS MODEL...AND
COMPLETELY OPEN TO SUGGESTIONS.

THE NEW MODEL IS VERY DRY BIASED...AND THIS SEEMS TO BE HAVING AN
IMPACT ON THE REGIONAL NAM...AS IT DEPENDS ON THE GFS TO
INITIALIZE THE BOUNDARY CONDITIONS. EXCEPT FOR SOME CONVECTION
TODAY...BOTH MODELS THEN TREND TOWARDS A VERY DRY PATTERN OVER THE
NEXT THREE TO FOUR DAYS.

[blp]

I cannot understand how you would make such a major upgrade to such an important model during the heart of hurricane season with these issues. What makes this even more frustrating is that they ran it in parallel and somehow now they are noticing these differences.

Did the new GFS miss Alex? It had a different solution than the old GFS but if I recall they both developed Alex and busted the track in the mid-long range. None of the models had a very firm latch on Bonnie. So far the new GFS is performing just fine. I think that what we have here is a case of confirmation bias now that the NWS went ahead and said there's a problem.

My thinking is if the new GFS consistently underestimates windspeeds in the tropics by 20kts, then it's probably not going to have very many false positives at all. Maybe some false "all-clear" runs though...as I suspect the current quiet run is false. If the new GFS develops a tropical system despite its weak bias, that could be a very clear indication something big will be happening. Don't think of it as the death of an old and trusted model, but rather the birth of a new and exciting model.

I do wish they'd keep running the old GFS, but I understand that it's not easy to double the amount of supercomputing power devoted to GFS overnight. They have to choose one or the other.

[srainhoutx]

For a second day in a row, HPC talks about the not so good new GFS. I hilighted the most revealing parts of their thinking about this.

OT: HPC isn't the only one discussing 'issues'...

AREA FORECAST DISCUSSION
NATIONAL WEATHER SERVICE FORT WORTH TX
306 AM CDT THU JUL 29 2010

.DISCUSSION...
GFS MEX MOS VIEWABLE ON LINE IS COOLER THAN THE GFS MEX MOS WE
RECEIVE IN OUR WORKSTATIONS. THE DAY SHIFT WILL HAVE TO FIND OUT
WHAT IS GOING ON...BUT THERE WAS A NEW VERSION OF THE GFS MODEL
ROLLED OUT ON WEDNESDAY AND THERE MAY BE A PROBLEM WITH THE MOS.
THAT BEING SAID...I HAVE OPTED TO LEAN TOWARD THE INTERNET
TEMPERATURES WHICH ARE COOLER...AND MORE IN LINE WITH THE MOS
GUIDANCE RECEIVED YESTERDAY. AT ANY RATE HIGHS SHOULD REACH TRIPLE
DIGITS MANY AREAS SATURDAY THROUGH NEXT THURSDAY.

[Blown Away]

Sorry but intense hurricanes are not good signs for anyone. IMO

Well at least we're finally getting something intense on the long range . Good sign IMO

Stormcenter, intense hurricanes happen whether Storm2k talks about them or not. An intense hurricane that stays out at sea and transfers heat out the tropics is good. Is there a reason why you feel the need to scold folks who would like to see a hurricane develop and discuss it in Storm2k, you have a track record of doing that? You have 4000+ posts so something tells me you would like to see a hurricane develop but you just can't admit it? Maybe this forum is an opportunity for you set your conscience free and admit you love hurricanes and want to see one develop now!!!!
Sorry off topic, I'm done!

[cycloneye]

Ok, lets focus on the models and especially on the critics for the new version of GFS that HPC and others have made.

[KWT]

The GFS seems to have been made somewhat worse in terms of actual development, indeed it appears that because of the trade issue the model may now underdo system development!

[cycloneye]

IMO, they would have waited until November to make the change to allow the old version to continue thru the peak of the season but that is only my take.

[KWT]

Yeah, I agree with the idea that the NHC pushed through this and it looks like in order to totally elminate the convective feedback issue they've managed to create another problem.

[Macrocane]

Can someone tell me why the 192 h images look so weird on the new GFS?