Constructing Variable Tables for CARRA Reanalysis

[kevinboyd]

Hello everyone,

I'm attempting to run WRF (actually, Polar WRF) using pressure level data from CARRA Reanalysis (Info). I need to construct a custom Vtable for the dataset, but the GRIB codes corresponding to some variables are abnormal, with 6 characters instead of the usual 3. I'll attach the output of some commands for parsing the dataset (e.g., wgrib2, grib_ls, g2_print.exe). Each of the variables and their corresponding GRIB code can be found within the ECMWF's GRIB database (Parameter Database).

I've never come across a Vtable containing variables with 6-character codes, so I was wondering if this was doable and if so, if there is any kind of formatting restrictions for the Vtable that I should keep in mind. I suppose this is related to how ungrib.exe parses the information in each column.

Thanks in advance for any advice or insights!

Kevin

 

[kevinboyd]

Quick update: some preliminary testing suggests that ungrib.exe parses the data fine, so long as the Vtable is configured properly. I'll be sure to upload my final Vtable for future individuals undergoing similar experiments!

 

[kevinboyd]

To all future observers, I'll attach a couple variable tables to be used with CARRA pressure level data. Be sure to remove the .txt extension before utilizing these files. You'll need to use the soil data from ERA5, and I recommend converting geopotential to geopotential height in CARRA before running ungrib.exe. Also, I wasn't able to get ungrib.exe to recognize the snow depth in CARRA, so unless you're more adept than I, you'll need to import that from ERA5 as well.

One day I'll create a GitHub page with scripts and such that I used along the way, as I believe this may be the first time someone has used this dataset in conjunction with WRF! Until then, good luck!