Global model runs discussion

[Alyono]

as of now, the Canadian is significantly more reliable in terms of genesis than is Model Uccelleni

[jlauderdal]

another phantom from Model Uccelleni

Why do you call it that?

NWS model and the NWS director has done some things so incredibly absurd (such as not being allowed to use wind data to estimate wind speed). Plus, the same NWS director used to be NCEP director and he had a chance to rectify some of the glaring faults with the GFS (hint... the data assimilation scheme is the least technologically advanced in the world using 3DVAR. That old school way of thinking needs to adapt with the times and go to 4DVAR)

alonyo is president of the ulcellini fan club

[blp]

Well the GFS is back on its own. The FIM and CMC have dropped the idea of the C.Atlantic development. The Euro shows nada so it shaping up as another possible phantom. At least it is providing some entertaining runs for those slow times.

[tropicwatch]

It is kind of odd. There is quite a bit of disturbed weather in areas that would normally be a hot bed for tropical development but it is just not happening.

[Dean4Storms]

The Shear Monster is raging all over!

[blp]

another phantom from Model Uccelleni

Why do you call it that?

NWS model and the NWS director has done some things so incredibly absurd (such as not being allowed to use wind data to estimate wind speed). Plus, the same NWS director used to be NCEP director and he had a chance to rectify some of the glaring faults with the GFS (hint... the data assimilation scheme is the least technologically advanced in the world using 3DVAR. That old school way of thinking needs to adapt with the times and go to 4DVAR)

Thank you for the insight. That is quite amazing since many including myself considered it to be among the best in the world with the latest technology. What I don't get is why this year has it been so bad, it had the same 3DVAR scheme last year and did not perform as bad but did some other upgrade this year mess this up. This year so far it has produced phantoms in the E. Atlantic, C. Atlantic, Gulf and W. Carribean and I can understand the W. Carribean one which is a long known problem, but it has gone off the rails everywhere else this year. Has it performed this bad in the other basins?

[tolakram]

Forget the tropics, it's been horrible inland as well. Our forecast up here has failed miserably when the local mets decide to side with the GFS instead of the euro. Since it's software I blame both decisions and version control. You never let a piece of software get worse and if it does you roll back and do an analysis to figure out what went wrong.

Now I'm not sure it's worse, it may just be stumbling over conditions it's particularly bad at, like a slow tropical season, but it certainly seems worse.

[cycloneye]

blp,GFS has performed better in EPAC and WPAC as it has shown developing systems and later development has occurred. Here is an example with Marie at EPAC which turned into one of the most powerful systems of 2014 in the tropics not only in EPAC. GFS was magnificent forecasting the development of Marie days before it became a TD and developing into a powerful hurricane as I am showing below.

[Alyono]

the reason 3DVAR is bad (and to be fair, there is a hybrid scheme in the GFS that also uses an Ensemble Kalman Filter), is that it maps ALL obs to synoptic time. That means all obs, even if they are taken at 0900 UTC, are mapped to either 0600 or 1200 UTC

4DVAR has a process that allows for the time component to be included in the data assimilation.

Even the Canadian had 4DVAR. NCEP/Uccellini is stuck in the 1980s in this regard (as is his wind data shall not be used to determine wind speeds edict from last year)

[Alyono]

and our crazy model is at it again. Tries to develop the wave that just moved off of Africa into a storm before striking the Caribbean. It them really blows it up in the eastern Caribbean

[cycloneye]

Well,the San Juan NWS folks mention the GFS and GEFS members in the afternoon discussion

A TROPICAL WAVE NEAR THE
COAST OF AFRICA IS LIKED BY GFS AND SEVERL GEFS MEMBERS. BY
MIDDLE OF NEXT WEEK...OP GFS AND GEFS SHOW A TROPICAL CYCLONE
APPROACHING THE SRN WINDWARD ISLANDS. ABOUT 10 OF 20 ENSEMBLE
MEMBERS SHOW A WEAK LOW PRESSURE WITH MSLP`S 1005 MB OR WEAKER. AS
THE CURRENT TUTT PATTERN WEAKENS...THIS WAVE MAY HAVE A BETTER
CHANCE TO ORGANIZE INTO A TROPICAL CYCLONE.

[Fego]

So far, GFS has been moving the time frame. As per the future path, your choice; south, northeast or across the big islands.

[Yellow Evan]

blp,GFS has performed better in EPAC and WPAC as it has shown developing systems and later development has occurred. Here is an example with Marie at EPAC which turned into one of the most powerful systems of 2014 in the tropics not only in EPAC. GFS was magnificent forecasting the development of Marie days before it became a TD and developing into a powerful hurricane as I am showing below.

http://i60.tinypic.com/2ur1huf.png

The GFS has been showing plenty of phantom storms in the WPAC. EPAC makes the models look good this year.

[Hurricaneman]

Well,the San Juan NWS folks mention the GFS and GEFS members in the afternoon discussion

A TROPICAL WAVE NEAR THE
COAST OF AFRICA IS LIKED BY GFS AND SEVERL GEFS MEMBERS. BY
MIDDLE OF NEXT WEEK...OP GFS AND GEFS SHOW A TROPICAL CYCLONE
APPROACHING THE SRN WINDWARD ISLANDS. ABOUT 10 OF 20 ENSEMBLE
MEMBERS SHOW A WEAK LOW PRESSURE WITH MSLP`S 1005 MB OR WEAKER. AS
THE CURRENT TUTT PATTERN WEAKENS...THIS WAVE MAY HAVE A BETTER
CHANCE TO ORGANIZE INTO A TROPICAL CYCLONE.

The 18zGFS no longer develops this wave but it could be one of those systems that waits until the western Caribbean to develop like a typical tevelopment in October

[Siker]

The 18zGFS no longer develops this wave but it could be one of those systems that waits until the western Caribbean to develop like a typical tevelopment in October

In fact that's exactly what happens on this run. It could do that, or this could be the start of the GFS dropping the system. We'll see.

[Hurricaneman]

The 18zGFS no longer develops this wave but it could be one of those systems that waits until the western Caribbean to develop like a typical tevelopment in October

In fact that's exactly what happens on this run. It could do that, or this could be the start of the GFS dropping the system. We'll see.

It could be interesting, but we'll probably have to wait 5 more days to see if any other models come aboard

[blp]

00z GFS has it again and the 18z Navgem is also bitting again and shows its strongest run.

I think the 18z GFS dropped it because it crashed into S. America. It has it very close on the 00z but enough offshore to stick around. I will give credit the GFS keeps moving timeframe down and now has the low appearing in 132hrs.

[Hurricaneman]

00z GFS has it again and the 18z Navgem is also bitting again and shows its strongest run.

I think the 18z GFS dropped it because it crashed into S. America. It has it very close on the 00z but enough offshore to stick around. I will give credit the GFS keeps moving timeframe down and now has the low appearing in 132hrs.

the 6zGFS does pretty much the same thing but what is telling is that the pattern might lead to development in the Western Caribbean after day 10 from the wave moving off of Africa in about 48hrs which a few models are starting to pick up on so it may not be out of the realm of possibility

The posts in this forum are NOT official forecast and should not be used as such. They are just the opinion of the poster and may or may not be backed by sound meteorological data. They are NOT endorsed by any professional institution or storm2k.org. For official information, please refer to the NHC and NWS products

[LarryWx]

as of now, the Canadian is significantly more reliable in terms of genesis than is Model Uccelleni

The blind M.U. (Lol) will find its nut again at some point. It admittedly has been rather insistent on this Caribbean development and there's some chance this is the nut it will happen to find. However, it was also previously insistent on development near the Yucatan and subsequently landfalling in the US for something like 12 of 13 runs around 9/20 and nothing tropical ever verified. Also, the Euro says that M.U. is on crack.
As noted, the 6z M.U. still has something develop and move into the W. Caribbean near the end of the month. Out of curiosity, I may later today do some research on analogous tracks. I do already know that a development of a TD+ east of 55W and south of 20N that late then hitting the CONUS would be unprecedented since at least 1851 per official tracks. However, there were some that formed between 55 and ~60 W and south of 20N that did make it.

Edit: M.U. dropped it at 12Z.

Edit: Checked through 1950 so far and found these that formed south of 20N and within 55-62W corridor that later hit CONUS after these formation dates from last few days of Sept an later: 10/9/1879 (56W) FL Panhandle, 10/9/1887 (60W) SE LA, 10/9/1897 (62W) NC Outer Banks

Edit 2: Only one more to add but it was a real doozy: Hazel of 10/5/1954 formed at 59W and hit near the SC/NC border as a ferocious storm

So, to summarize: No official geneses after 9/25 east of 55W and south of 20N later hit the CONUS since 1851.
However, four formations between 56W and 62W later did. Interestingly enough, these four were much later than 9/25 as they all formed not til the 10/5-10/9 timeframe,with three right on 10/9! Two hit as a H and two hit as a TS. Three of the four were within 1879-1897. Only 1954's Hazel has followed that through 2013! So, fwiw, this kind of thing is very long overdue to say the least!

[LarryWx]

Folks,
It's back on the 18Z M.U. lol.