• Hello, please take a minute to check out our awesome content, contributed by the wonderful members of our community. We hope you'll add your own thoughts and opinions by making a free account!

Misc 2022-23 Fall/Winter Whamby Thread

Status
Not open for further replies.
TheBatman said:
Easy now. They are just as happy for the western members.
Click to expand...
I am, kind of. I feel as a central NC'linian, I can drive just 4 hours and see snow multiple times a year if necessitated. Lower gulf states don't have that luxury, but having said that, if I lived there it wouldnt be anything I'd ever expect as well. Here in NC my expectation of seeing snow is a little higher.
 
There has to be some feedback errors on this. Ive never seen a wall of high pressure like this depicts.
bouncycorn said:
Thanks. Here’s the total snowfall, crazy! I adjusted the color table and the white strip is still 24”+. .. 950mb low!!!
mslp.000061.png
View attachment 127142
Click to expand...
 
WeatherNC said:
Thank you, it's still hard to dial in the actual SLP based on color scale but I assume around Culpepper VA, similar to the 18z GFS. Apologize if already shared but can you link any details/specs on documentation running in house. My assumption is mesoscale with some type of extension, you can take the reply to Whamby don't want to add clutter here.

View attachment 127143
Click to expand...

I'll fix the color table so it goes over a larger range of values, but the frame minimum is on there above the color scale at the bottom.

It's a highly modified version of the open source WRF (most models are based off of wrf, such as HRRR before the latest upgrade, NAM12, NAM3, GRAF, RPM, etc)... most of the physics configurations are also custom and honed in for the southeastern US (custom local PBL, single moment NSSL-like microphysics, modified CAM sw/lw radiation). There's also AI integration which is part of the reason it takes so long to run. At each timestep, an AI analyzes the global grid and makes bias-corrections based on 20 years of analyzed GFS runs and observations. This version of the MMFS-R (regional) is a global model with a 12km nest over the eastern US (pictured) and the ability to add 4km and 1km nests as needed (usually done in closer range since they are more computationally expensive).

There is also an MMFS-H which is for hurricane applications.. completely different source code modifications and configurations for that one, it also has AI integration.

If you have any questions about anything specific, I'm happy to answer them. This has been in the works for about 10 years, since I was in high school and it was very low resolution (1 degree horizontal resolution) and was based on the CAM open source model.
 
bouncycorn said:
I'll fix the color table so it goes over a larger range of values, but the frame minimum is on there above the color scale at the bottom.

It's a highly modified version of the open source WRF (most models are based off of wrf, such as HRRR before the latest upgrade, NAM12, NAM3, GRAF, RPM, etc)... most of the physics configurations are also custom and honed in for the southeastern US (custom local PBL, single moment NSSL-like microphysics, modified CAM sw/lw radiation). There's also AI integration which is part of the reason it takes so long to run. At each timestep, an AI analyzes the global grid and makes bias-corrections based on 20 years of analyzed GFS runs and observations. This version of the MMFS-R (regional) is a global model with a 12km nest over the eastern US (pictured) and the ability to add 4km and 1km nests as needed (usually done in closer range since they are more computationally expensive).

There is also an MMFS-H which is for hurricane applications.. completely different source code modifications and configurations for that one, it also has AI integration.

If you have any questions about anything specific, I'm happy to answer them. This has been in the works for about 10 years, since I was in high school and it was very low resolution (1 degree horizontal resolution) and was based on the CAM open source model.
Click to expand...
Do you have any verification scores on how this stacks up against the other short-range models?
 
Ollie Williams said:
Do you have any verification scores on how this stacks up against the other short-range models?
Click to expand...
Tests on the current version of the MMFS-R were done at 120 hrs and 500mb heights.. it scored on par with UKMET, which was better than GFS and CMC but slightly worse than ECMWF.
 
I have a serious question for those of you who actually know what’s kind of going on with this pattern. Have we lost our good pattern that was showing a few days ago? Or are the models just waffling back and forth as usual? It just feels like everything is getting pushed back as usual.
 
Stevo24 said:
I have a serious question for those of you who actually know what’s kind of going on with this pattern. Have we lost our good pattern that was showing a few days ago? Or are the models just waffling back and forth as usual? It just feels like everything is getting pushed back as usual.
Click to expand...
Models are not keying in on a Gulf Low anymore and rather a Colorado Low miller B solution.
 
Stevo24 said:
I have a serious question for those of you who actually know what’s kind of going on with this pattern. Have we lost our good pattern that was showing a few days ago? Or are the models just waffling back and forth as usual? It just feels like everything is getting pushed back as usual.
Click to expand...
IMHO, there’s been absolutely nothing getting pushed back. We’re still going into a good pattern, and yes there are still signals for a southeast winter storm. However like has been said, being in favorable pattern doesn’t guarantee anything.
 
bouncycorn said:
I'll fix the color table so it goes over a larger range of values, but the frame minimum is on there above the color scale at the bottom.

It's a highly modified version of the open source WRF (most models are based off of wrf, such as HRRR before the latest upgrade, NAM12, NAM3, GRAF, RPM, etc)... most of the physics configurations are also custom and honed in for the southeastern US (custom local PBL, single moment NSSL-like microphysics, modified CAM sw/lw radiation). There's also AI integration which is part of the reason it takes so long to run. At each timestep, an AI analyzes the global grid and makes bias-corrections based on 20 years of analyzed GFS runs and observations. This version of the MMFS-R (regional) is a global model with a 12km nest over the eastern US (pictured) and the ability to add 4km and 1km nests as needed (usually done in closer range since they are more computationally expensive).

There is also an MMFS-H which is for hurricane applications.. completely different source code modifications and configurations for that one, it also has AI integration.

If you have any questions about anything specific, I'm happy to answer them. This has been in the works for about 10 years, since I was in high school and it was very low resolution (1 degree horizontal resolution) and was based on the CAM open source model.
Click to expand...

Essentially you have coded a custom mesoscale, CAM, similar to a tropical model? With semi custom & off the shelf GFS bias correction now being down automatically; AI? I know nothing about coding, nor weather models beyond the previous decade, really just want to affirm my understanding and you did it very well in your post, fingers crossed.

Almost like a DGEX+++ (looking at 3 generations beyond)
 
DadOfJax said:
There is one more geared for the South, but it lost all of its members from being over moderated…..
Click to expand...
If anything this forum is very very lax with moderation which is why I am still here and every banned member has had multiple rounds on here … the mods are the most laid back ever. I thank them every day for it , even meanie pants Mcmeanager mean faces @metwannabe @packfan98 ! Even they are very relaxed and cool lol
 
WeatherNC said:
Essentially you have coded a custom mesoscale, CAM, similar to a tropical model? With semi custom & off the shelf GFS bias correction now being down automatically; AI? I know nothing about coding, nor weather models beyond the previous decade, really just want to affirm my understanding and you did it very well in your post, fingers crossed.

Almost like a DGEX+++ (looking at 3 generations beyond)
Click to expand...
CAM stands for Community Atmospheric Model in this case, not convection allowing model. It's an older model framework and I use their radiation scheme in the MMFS. DGEX was essentially an extended NAM, it was a north american regional mesoscale model that was fed by the GFS.. MMFS is different in that mmfs is a global model fed directly by observations and not just based over north America or fed by any other model. The tropical model is separate (there's MMFS-R and MMFS-H. two separate models coded differently). Hope all that makes sense. Thanks
 
Absolute garbage discussion on here about NC vs SE. The board exists due to regular posters analyzing the weather. If you guys want a broader discussion on other regions put in the work and don't go poof when the threat does. The regular posters do a great job at not being too imby with their posts
 
Status
Not open for further replies.
Back
Top