Global model runs discussion

[tolakram]

... why shouldn't I delete the last 10 or so posts?

jeesh.

[northtxboy]

... why shouldn't I delete the last 10 or so posts?

jeesh.

Its good entertainment

[CYCLONE MIKE]

I enjoyed reading it. Gave some really good insight

[USTropics]

If by some series of unfortunate events the previous 10 posts are deleted , here are some of the popular opinions along with some factual information:

Model wars are pointless. Use a consensus of all models and know the strengths/weaknesses of particular ones.

The ECMWF is headquartered in Reading, England on one of the largest super computer complexes (that's right, it's not just run on a super computer, but a complex of super computers ) in the world with 230 employees. It was initialized in 1975. The ECMWF is renowned worldwide as providing the most accurate medium-range global forecasts and uses a constant input of satellite, weather stations, aircraft, ships, and weather balloon data. With that said, the ECMWF is terrible at cyclonegenesis (possibly due to a low resolution issue), but overall atmospheric pattern recognition and eventual tracks of formed systems are its strengths. If the ECMWF shows tropical genesis in the deep tropics, you best know things are about to get real. The ECMWF also tends to have a westward bias, and over does ridges at times.

The GFS has recently been updated and is also considered one of the better performing long range models along with the ECMWF. It was initialized in 2000 (as 2 seperate models, the AVN and the MRF before they were merged in 2002 and took the name GFS). It has a tendency at times to over develop systems or to be overzealous with development of tropical identities that never materialize (although not to the degree of other models mentioned below). It tends to have a right bias, and overdoes troughs.

The CMC(GEM)/NAVGEM/FIM also have either had updates or are new to game (when considering the time frame that these models have been available compared to the GFS/ECMWF; The NAVGEM is the replacement for the NOGAPS and just went operational this year and the FIM is projected to be the eventual replacement for the GFS). While they appear to be turning the corner this year, it'll take a few years of data to have an accurate opinion of their place in the model landscape.

The GFDL model was initialized in 1950 and uses the GFS data. the GFDL actually has its own dedicated laboratory inside the NOAA with 300+ employees (The GFDL use to be run on a supercomputer located at Princeton, but now runs on a massively parallel Cray supercomputer with over 30,000 processor cores at Oak Ridge National Laboratory). While their primary focus is for the study of climate change, this model is for tropical means when there is an active tropical invest/system. It typically over strengthens systems.

The HWRF was initialized in 2007 and is the eventual replacement for the GFDL (in fact, it's projected to be the operational backbone for hurricane track and intensity forecasts for the NHC in the future). The model was developed by the NOAA, the NRL, the University of Rhode Island, and FSU. The model is the first of its kind in that it uses data from satellite observations, buoys, and reconnaissance aircraft. The model is planned to run at an even higher resolution in the future and performed decently for 2012. Just like the GFDL, it also only runs when an invest/active tropical system is present and tends to be overzealous on strength.

The TVCN will often be mentioned when an active system is in the Atlantic by the NHC. Unlike the other models, the TVCN uses a consensus of forecast models. It uses interpolated versions of the GFDL, UKMET, NOGAPS (no longer in operation and the consensus models do not use NAVGEM data at this point in time), GFS, and the ECMWF for track guidance. For intensity it uses the LGEM, interpolated GFDL/HWRF, and DSHIPS/ICON consensus.

Also, here is an image of track forecast skill for models:

[cycloneye]

Great post that explains all about each model.

[ninel conde]

GFS still showing nothing out to 384?

[jlauderdal]

GFS still showing nothing out to 384?

nothing at 384=5 systems at 384....means a whole lot of nothing either way, maybe at 240 you could sense a pattern change or something like that

[caneman]

Why would you delete the thread. It is a friendly conversation about model accuracy. This sums if up nicely and show why the NHC uses a blend. There are of course three factors that go into modeling: Development, Track and Strength. With all those things considered, I'm not sure that any one model is all that great. At least one or two models performed better than I thought. Thanks.

[ninel conde]

GFS still showing nothing out to 384?

nothing at 384=5 systems at 384....means a whole lot of nothing either way, maybe at 240 you could sense a pattern change or something like that

finally got my link to work. it does show lower pressure off africa at 384 but that might be just climo.

http://tc.met.psu.edu/tcgengifs/GFS/201 ... /slp36.png

at 240 it does show a different pattern. the ridge is gone in the atlantic. if that can be the pattern for the rest of the season maybe we can see some development east of bermuda.

[cycloneye]

Warning=Is very long range. 12z GFS shows a developing CV system.

[cycloneye]

Here is how it ends.

[blp]

Here is how it ends.

http://oi40.tinypic.com/20zqups.jpg[/img]

First run I see from the GFS where in the long run it holds a system together past 35W. Let's see if the trend continues. It is about that time where we need to start seeing something from that part of the world.

[TeamPlayersBlue]

The 384 IF it panned out would be a beeline to Florida. We should follow the 1016 MB line for TS correct?

[ROCK]

if 384hr GFS panned out it would be the first time in history.....

[cycloneye]

I look more to how the pattern will be on long range than the systems as they change constantly.

[weatherwindow]

The 384 IF it panned out would be a beeline to Florida. We should follow the 1016 MB line for TS correct?

Good afternoon, blp et al....yes, if it persists, but it is a fairly progressive pattern and ask Rock noted it is shall we say "not a forecastable solution" at 384. That said, the run below is the MSLP anomaly solution on the 12Z GFS...weak ridging north and west of the Greater Antilles to the Ohio Valley(follow the 1016mb isobar and no significant trofing to the Rockies..TC tracking to Fla or the SE coast ...the SE coast is about 6 days out from the 384hr TC position so things can certainly change....Grtz from KW, Rich


http://www.tropicaltidbits.com/analysis ... opics.html

[Blown Away]

Here is how it ends.

Actually strengthens as it moves through MDR, wow imagine that!! I looked at the 06z GFS this morning and nothing through August 28th and I started thinking are we in some kind of overall global longterm pattern shift to fewer storms and is related to GW, SUV's, and is Al Gore right!!!! Finally, the 12z clears that insanity out of my mind, well kinda...

[TheStormExpert]

06z GFS had a 1005mb low exiting Africa at 384hrs.

[tolakram]

12Z Euro vorticity and pressure map might actually be showing a storm in the central Atlantic at 144 hours. Nothing was there in the 0Z run.

[tolakram]

12Z Euro at 192H takes this vorticity NW, still in the central Atlantic (around 46W) but now north in latitude of the islands. So it doesn't do much with it, yet this same feature was nowhere to be found in the previous run.

It shows up on this free map as the small circle of lower pressure.



It's pretty much gone at 216 hours.