Getting WRF to work with ERA-Interim data
I actually started a blog as part of another class, but lets pretend that never happened and mark this as the official start.
I’m slowly going to start putting myself out there, I’m trying not to be a hypocrite – I preach open-source but I keep myself closed.
I’m in the process of running WRF. It’s not been without challenges though, when running WRF locally compiling can be a major time consumer, you have to know some GNU/Linux basics, which is actually pretty cool. Okay, I’m all over the place, but let me write about my current problem and as the blog evolves hopefully I can get more ordered.
I’m currently running WRF on a convective scale over the South-Africa Highveld, the area is also home to Jo’burg and Pretoria, two of South-Africa major cities, I’m doing this to try to understand hailstorms better. WRF has various input datasets that can be used to initialize the model, I’m specifically using ERA-Interim data, but the ERA data requires some extra steps, I’m not going to try and give an in-depth explanation of all the steps, but instead I’m just doing them here in the order that was successful eventually. I also assume that you’ve already complied WRF successfully and was to lazy to RTFM (like me) and used Google (like me).
Last thing, I use vim, but replace that with whatever you want, on a remote machine nano and vi are good alternatives, Emacs is pretty cool (I’m super slowly learning that) but I don’t know about its availability on all remote computers.
First get the data ready in WPS:
Create ecmwf_coeffs file:
Look for PRES:2013-11-*
Look for met_em.d0*
If this worked you will now go back and change directory to WRF:
Look for wrfbdy_d01, wrfinput_do1, wrfinput_d02, …
Look for wrfout_do1_2013* until complete for the run time series
Again, Google around for ecmwf_coeffs table I’m assuming you check, double-check and triple check your namelist.wps and namelist.input. If using more than two domains, check fourth time, its easy to miss something when looking at to long.