strange problem with wrf output in variables 3d

[aricky84]

Hi all, I am a ‘long-standing’ WRF user, and I am facing a really strange problem (for me).

I am using WRF in the COAWST framework, coupled model, but this is not important because it also happens with WRF standalone (pure installation).

Here is the error with the discussion.

WRF 4D variables saved incorrectly. · Issue #296 · DOI-USGS/COAWST

Hi John, I ran the runs with version 3.8 of COAWST. The WRf, WRf+ROMS and WRF+ROMS+SWAN runs ran perfectly and the ground output is consistent. Unfortunately, the output of WRF, of the 4D variables...

github.com

To summarise, the 3D variables that WRF contains contain huge (and tiny) values.
Obviously this is a compilation problem, because I don't get this error if I compile with intel.

Something happens in this application that blows everything up. i attach the file i use to automate the installation, and the environment variables. With this environment (and gfortran) I have tried to compile in every way, but all runs work, but BAD.
I have narrowed it down to 3 fundamental pieces of information:
1) COAWST test cases work fine, even with this environment.
2) my runs work badly in COAWST (with this environment), they work badly if I compile and use WRF with this environment, they work well if I use them in WRF compiled with intel.

I have worked on numerous machines, but always have the same problem.

any help are welcome


Antonio

 

[kwerner]

Hi Antonio,
Can you share the met_em* files you're using to run WRF? These files will be too large to attach here, so see the home page of this forum for instructions on sharing large files. You only need to share the files up through the time when you first start seeing the issue. Please also share the namelist.input file you're using, and let me know which version of WRF you've been trying. Thanks!