Here is the link:
http://met.psu.edu/tropical/tcgengifs/
Maybe I am just tired & need a rest
Models Question
Moderator: S2k Moderators
Forum rules
The posts in this forum are NOT official forecasts 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. For official information, please refer to products from the National Hurricane Center and National Weather Service.
Models Question
Can someone please look at this for me and tell me what is the deal? If you run the GFDL for Karl it shows Jeanne mixing with remains of Ivan? and then going south? But if you run the GFDL for Jeanne it shows her mixing the remains of Ivan and going Northeast? HUH? How can it be both?
Here is the link:
http://met.psu.edu/tropical/tcgengifs/
Maybe I am just tired & need a rest
Here is the link:
http://met.psu.edu/tropical/tcgengifs/
Maybe I am just tired & need a rest
0 likes
Wow, that is a wonderful observation.
Hmm, I would have thought the GFDL was the GFDL.
Apparently there is no provision for more than 1 "focal storm"
Hmmm, maybe because the designers of the model were guided to track a storm, they have only one set of variables available to plug-in such specific data.
Clearly a major flaw, that they cannot put in the more finite data of a tropical system, more than once.
Cool revelation!
Hmm, I would have thought the GFDL was the GFDL.
Apparently there is no provision for more than 1 "focal storm"
Hmmm, maybe because the designers of the model were guided to track a storm, they have only one set of variables available to plug-in such specific data.
Clearly a major flaw, that they cannot put in the more finite data of a tropical system, more than once.
Cool revelation!
0 likes
-
orion
- S2K Supporter
- Posts: 165
- Joined: Mon Aug 02, 2004 5:44 pm
- Location: Indian Harbour Beach, FL
- Contact:
Hmmm... interesting. I though maybe they were run at 2 different times... but both are 12Z17Sept.
I posted this a few minutes ago in a different topic and when I went back to see it, it said the topic didn't exist. So since its about models, I'll try again here...
Just a few thoughts on the models, technology, and 5 day forecasts...
The complex equations used by the models have numerous parameters, such as SSTs, pressure, etc. If you run the same model twice, but the second time with just one of these parameters a little different... even out to several decimal places... the outcome will be similar for the first 2 or 3 days, but then it will diverge more and more as time passes. By the time you get out to 5 days you have seen how unpredictable they can be. The technology, while getting better all the time, still is not accurate enough to make the models better out to 5 days. As a simplified example, a temp of 27.4 C could make a huge difference in the output at 5 days than the same model run with a temp of 27.5 C.
If you are a math-type person like me, check out chaos and fractals in the weather. A name to look for is Lorenz, many of these findings came from his work while trying to come up with equations to model the weather.
Also, if you haven't seen them, the animated 5 day forecasts can be interesting to look at.... for Ivan, they are at:
http://www.nhc.noaa.gov/archive/2004/IV ... hics.shtml
You can also animate the 3-day, the strike probability, and the wind swaths.
Jeff
I posted this a few minutes ago in a different topic and when I went back to see it, it said the topic didn't exist. So since its about models, I'll try again here...
Just a few thoughts on the models, technology, and 5 day forecasts...
The complex equations used by the models have numerous parameters, such as SSTs, pressure, etc. If you run the same model twice, but the second time with just one of these parameters a little different... even out to several decimal places... the outcome will be similar for the first 2 or 3 days, but then it will diverge more and more as time passes. By the time you get out to 5 days you have seen how unpredictable they can be. The technology, while getting better all the time, still is not accurate enough to make the models better out to 5 days. As a simplified example, a temp of 27.4 C could make a huge difference in the output at 5 days than the same model run with a temp of 27.5 C.
If you are a math-type person like me, check out chaos and fractals in the weather. A name to look for is Lorenz, many of these findings came from his work while trying to come up with equations to model the weather.
Also, if you haven't seen them, the animated 5 day forecasts can be interesting to look at.... for Ivan, they are at:
http://www.nhc.noaa.gov/archive/2004/IV ... hics.shtml
You can also animate the 3-day, the strike probability, and the wind swaths.
Jeff
0 likes
-
Stormsfury
- Category 5
- Posts: 10549
- Age: 53
- Joined: Wed Feb 05, 2003 6:27 pm
- Location: Summerville, SC
Stormsfury wrote:The GFDL for each individual storm is run on different gridpoints/sectors .. and will come up with different results ...
results may vary ...
Why is that?
Aren't all the factors the same for each grid on the terrestrial map?
I still think that there is still only one data base/set of detailed variables designed in for tropical cyclones for this model, and since they didn't design the model to handle the number of variables that are routinely measured for a tropical cyclone to be available for EVERY grid-point, they have to make 2 seperate runs!
Since the grid points for Ivan contain ALL of it's very detailed data only for it's run, but there's "no room" for that same amount of data for the run focused on Jeanne.. there would be 2 seperate runs, and therefore 2 seperate results.
Does anyone else get that idea?
0 likes
Who is online
Users browsing this forum: aspen, crownweather, NotSparta, Team Ghost and 107 guests