Which variable to extract from wrf-output for analysis puposes.

[zemega]

Hello.

I'm currently running a case study about a strong wind event in Antarctica. I would like to know, in other's opinion, what variable that is needed for analysis?

For your information, I'm trying a hourly 1km resolution run. The file size for 1 time step for the smallest domain is around 2.7GB. While I currently have unlimited storage in the scratch directory of the cluster I'm using, I would not have enough space for personal storage.

I plan to use wrf-python, and selecting only variable that is needed for an analysis of a strong wind event. I will keep the wrfbdy and wrfinput files though.

 

[Ming Chen]

If you are only interested in winds, I would suggest that you keep wind fields (U, V, W, U10, V10) and other related caribe's such as coordinate variables and pressure, temperature, geopotential height etc. You can get rid of irrelevant variables like RAIN and tendency terms (e.g. RTHCUTEN etc. This will save you storage space. To remove unnecessary variables, you can delete 'h' in the 8th column of the variable in Registry. Below is an example:
Suppose you want to remove SNOWNC from wrfout, you change change the line in Registr.EM_COMMON from:

state real SNOWNC ij misc 1 - rhdu "SNOWNC" "ACCUMULATED TOTAL GRID SCALE SNOW AND ICE" "mm"

to
state real SNOWNC ij misc 1 - rdu "SNOWNC" "ACCUMULATED TOTAL GRID SCALE SNOW AND ICE" "mm"

Then please type ./clean -a and recompile the code. After this, wrfout will not include SNOWNC. You can follow the similar way to remove other variables from wrfout and save your storage space.