Excellent model agreement this far out for Dennis

[tampastorm]

I think in Tampa we are learned last year to look at the cone not the exact pinpoint. Also it is 100% that the track will change many times. And each time everyone will for the most point change there opinion.

[HouTXmetro]

JUST GREAT!!!!!! More hot air will be pumped into Texas!!!!!

[Wthrman13]

Rainband,

I have not glorified the GFS one bit here All you have to do is ask me about the GFS and Frances last year and you will hear some blunt things (please don't ask )

That is where forecasting and understanding patterns come into play. The GFS solution is a very viable one. Another thing going for it is that is has other model agreement, in particular the esteemed ECMWF.

The GFS is not perfect by any means, but it is not as horrible as some would like to make you think. Every event is different.

The GFS is a pretty good model for tropical cyclone prediction. In recent years it's performance has been great. I'm not really sure how well the ECMWF stacks up agains the GFS as far as track prediction goes. I do know that it performs somewhat better than the GFS in the medium to long range as far as overall performance is concerned. This is mostly due, in my opinion, to the sophisiticated 4DVAR (short for 4-dimensional variational data assimilation) initialization procedure. The GFS (and other NCEP models) uses a simpler procedure. NCEP has to develop and maintain several model systems, both short and long-range, whereas ECMWF is focused on only one. 4DVAR is very difficult and costly to develop, and is also computationally very intensive.

[Brent]

JUST GREAT!!!!!! More hot air will be pumped into Texas!!!!!

Maybe Hell has moved to Texas.

[loon]

Its always been in Texas, just usually way more out west....maybe its on vacation

[HouTXmetro]

JUST GREAT!!!!!! More hot air will be pumped into Texas!!!!!

Maybe Hell has moved to Texas.

I guess it's going to take some strong fronts to break this pattern and you know what that means........ More East Gulf and FLA threats.

[Brent]

JUST GREAT!!!!!! More hot air will be pumped into Texas!!!!!

Maybe Hell has moved to Texas.

I guess it's going to take some strong fronts to break this pattern and you know what that means........ More East Gulf and FLA threats.

Don't hold your breath for a cold front in JULY!!!

They usually die just north of here.

The MASSIVE trough that forced Charley into SW Florida caused all-time August record lows to be shattered across this area.

[loon]

Rainband,

I have not glorified the GFS one bit here All you have to do is ask me about the GFS and Frances last year and you will hear some blunt things (please don't ask )

That is where forecasting and understanding patterns come into play. The GFS solution is a very viable one. Another thing going for it is that is has other model agreement, in particular the esteemed ECMWF.

The GFS is not perfect by any means, but it is not as horrible as some would like to make you think. Every event is different.

The GFS is a pretty good model for tropical cyclone prediction. In recent years it's performance has been great. I'm not really sure how well the ECMWF stacks up agains the GFS as far as track prediction goes. I do know that it performs somewhat better than the GFS in the medium to long range as far as overall performance is concerned. This is mostly due, in my opinion, to the sophisiticated 4DVAR (short for 4-dimensional variational data assimilation) initialization procedure. The GFS (and other NCEP models) uses a simpler procedure. NCEP has to develop and maintain several model systems, both short and long-range, whereas ECMWF is focused on only one. 4DVAR is very difficult and costly to develop, and is also computationally very intensive.

This is probably not the place, but I was curious if one of you mets could enlighten me on the horsepower (computers) used for models and generally all aspects of weather met use. I am a 7 yr Sysad with most of my knowledge in UNIX but I admin many windows servers too (you have to yanno) but I'd like to someday make it into a weather office of some sort working on their computers, and I imagine there are specific hardware and software knowledge bases that I could really start attacking now and possibly by my 10 year mark be able to try that move in my career.

I realize this is off-topic and would accept PM's for answers if any of you Mets have time, its much appreciated. Again, sorry to hi-jack the thread and if its a problem mod's please delete or move to the appropriate board, thanks!

cheers,
loon

[Air Force Met]

JUST GREAT!!!!!! More hot air will be pumped into Texas!!!!!

Maybe Hell has moved to Texas.

You know JB is using 1954 as an analogue year right? Have you ever seen the JUnction Boys or read the book (great book BTW)...that was 1954. HOT.

Yeah...I hate it when storms go into LA...all that sinking air...gets downright balmy.

[dhweather]

Rainband,

I have not glorified the GFS one bit here All you have to do is ask me about the GFS and Frances last year and you will hear some blunt things (please don't ask )

That is where forecasting and understanding patterns come into play. The GFS solution is a very viable one. Another thing going for it is that is has other model agreement, in particular the esteemed ECMWF.

The GFS is not perfect by any means, but it is not as horrible as some would like to make you think. Every event is different.

The GFS is a pretty good model for tropical cyclone prediction. In recent years it's performance has been great. I'm not really sure how well the ECMWF stacks up agains the GFS as far as track prediction goes. I do know that it performs somewhat better than the GFS in the medium to long range as far as overall performance is concerned. This is mostly due, in my opinion, to the sophisiticated 4DVAR (short for 4-dimensional variational data assimilation) initialization procedure. The GFS (and other NCEP models) uses a simpler procedure. NCEP has to develop and maintain several model systems, both short and long-range, whereas ECMWF is focused on only one. 4DVAR is very difficult and costly to develop, and is also computationally very intensive.

This is probably not the place, but I was curious if one of you mets could enlighten me on the horsepower (computers) used for models and generally all aspects of weather met use. I am a 7 yr Sysad with most of my knowledge in UNIX but I admin many windows servers too (you have to yanno) but I'd like to someday make it into a weather office of some sort working on their computers, and I imagine there are specific hardware and software knowledge bases that I could really start attacking now and possibly by my 10 year mark be able to try that move in my career.

I realize this is off-topic and would accept PM's for answers if any of you Mets have time, its much appreciated. Again, sorry to hi-jack the thread and if its a problem mod's please delete or move to the appropriate board, thanks!

cheers,
loon

well Loon, I think it's appropriate - it is significantally tropical weather realted.

Most of the models are running on some Unix-based architecture.
Some small shops will run, for example, Linux with 8 processors and
8GB of memory. That will usually fit their needs.

When you get into larger areans, particularly the global models,
you are on giant Unix based supercomputers. The Navy and ECMWF
both have massive IBM Power 4 plus systems for thier climate/weather/ocean modeling. Ditto for NCEP.

FNMOC runs on SGI, I can't remember what AFWA runs on, but I think it
is SG as well. A great deal of wave research is done on Cray systems.

So to answer your question, scalable unix systems and clusters is
an area to focus on if that is where you want to go.

[dhweather]

JUST GREAT!!!!!! More hot air will be pumped into Texas!!!!!

Maybe Hell has moved to Texas.

You know JB is using 1954 as an analogue year right? Have you ever seen the JUnction Boys or read the book (great book BTW)...that was 1954. HOT.

Yeah...I hate it when storms go into LA...all that sinking air...gets downright balmy.

Yeah, and JB runs his models on a UNIVAC too.

[loon]

Thanks dhweather, excellent response. I'm mostly into Sun boxes and AIX currently, and of course Linux (which I cut my teeth on early on). Thanks again, great response.

cheers,
loon

[Cape Verde]

Could this wind up being a worst case scenerio for the people of New Orleans?

<p>Potentially, most any hurricane entering the Gulf could do that. New Orleans is vulnerable from two different angles of approach.<P>But it takes a BIG storm and just the right approach to doom New Orleans, which makes the chances of it happening with any specific storm very small.<P>Over the long run, it becomes rather likely. I'd have to guess that the perfect storm will strike New Orleans in the next 50 years, without a doubt.<P>Should the people of New Orleans move? The answer to that is yes,...[edited by moderator].

[Cape Verde]

Sorry if my attempt at humor was offensive. New Orleans is a great town and nobody realistically wants any harm to come to it.

[dhweather]

No offense taken, New Orleans does have some major issues.


The new mayor Ray Nagin seems to be addressing them, so there is
hope!

[timeflow]

http://www.nhc.noaa.gov/archive/2004/CH ... hics.shtml

Since there was a little discussion on possible similarities between the historical track of Charley (2004) and the projected track of Dennis (2005), I thought I would link in this "Graphics Archive" of the former, which is basically an animation of all the NHC 3-Day forecast tracks.

On August 10, 2004 - between the 11am and 5pm Charley updates, the projections depicted a significant shift to the right in the 2-3 day range. It would be interesting to see maps that show that massive trough... which ultimately pulled Charley up on almost the exact projected path of the August 11th, 11am image.

There may be no reason for Dennis to make such a pronounced turn. But we'll see how this all pans out, it's really interesting, and strange for July.