Stormtrack WRF

[Tim Vasquez]

For those of you interested in model forecasts of Texas and Oklahoma we are running an experimental 10 km WRF here:

http://www.stormtrack.org/wrf/

Tim

 

[Skip Talbot]

That's pretty sweet Tim. Nice set of products and great resolution. Can you provide any technical info on where the model data is coming from and what software you are using to create the plots? Are these products going to be available to us next spring or is this a temporary/changing project?

 

[Tim Vasquez]

The initialization and boundary fields are obtained from the 12 km NCEP WRF.

It's basically the WRF which is public domain and can be downloaded here: http://www.mmm.ucar.edu/wrf/users/download/get_source.html though keep in mind it runs under RH Linux.

Not sure how long we'll keep it running but the main goal is to put an 8 km WRF floater over chase targets this upcoming spring. I'm not a big fan of modelcasting but I've been sold on some of the results I've seen in recent years with Eta, RUC, and WRF output.

Tim

 

[rdale]

I'd recommend changing to the GFS for boundary conditions - basically you're running the WRF based on the WRF for now, you might get better results (at least "a different vantage point") basing it off GFS init.

 

[Jayson Prentice]

I think this is a nice idea... You can never see too many forecast models, right?!?

Tim, what made you decide to go with the parameterizations that you did? As it indicates you are using the GD cumulus, and then the WSM6 precipitation; I notice after viewing an output it also shows your choices for PBL schemes and radiational schemes. Not saying any of them are bad choices...

 

[Tim Vasquez]

Interesting point -- I'll switch this to GFS initialization and see how it looks. Mainly I'm constrained by the resolution, as NAM gives me 12 km and GFS 30 km; the 20 km RUC grids aren't yet working. As far as parameterizations, I'm not up to speed on them but I went with GD on the strength of its ensemble method and the WSM6 in the hopes that a newer technique + more phases/particles might give better results.

Tim

 

[Brian Stertz]

Thanks Tim.....nice job. Looks like all eyes will be on the southeast half of the grid the rest of 2008 as we are nearing the sometimes exciting "cold season" .

 

[Andrea Griffa]

Thanks a lot,Tim, this will be very helpful in the next season! A better resolution is what we need.

 

[Beau.Gjerdingen]

Just beware of a few things... first, initializing off the GFS would give a different result than initializing off the NAM analysis, however at T382 resolution, the true resolution of the GFS over the central US is closer to 40 km... much wider than the NAM 12km, so it may be less than ideal for initializing a mesoscale model without several nests. Also... most operational models run with over 60 vertical levels, granted some of these are stratospheric, but many are located in the troposphere. I ran a study using a similar configuration of the WRF-ARW last year over Minnesota and ran into a smattering of problems generating convection, simply because the vertical resolution was too rough to capture important features. Also I noticed there is a 200 mb top on the model... this might be pretty low and can allow spurious gravity waves to essentially propagate and grow along the top of the model (I had problems with this in my study too). Very cool though, and a tip of the hat to you for setting this up, I know from experience that it can be a royal pain in the butt to even get to the point of outputting something useful.

 

[Tim Vasquez]

I'm taking the WRF down until our next system comes in as I need my CPU cycles for other things. There's a good chance though that it will be updated every day or two.

Tim