Problems running ndown.exe for 1km domain

[EeshanBasu]

Hi all,
I am running WRF/Chem with chemical option 133 (a new chemical option designed by me/my colleague) and trying to use WRF v3.9.1 to simulate 36km, 6km, and 1km model results. I had no problems running wrf.exe for 36km, no problems running ndown for 6km, and no problems running wrf for 6km. But when I am trying to run ndown.exe for the 1km initial and boundary condition files, I am only getting wrfinput_d02 with no displayable variables. The job is running for the entire time without updating the wrfinput_d02 anymore and no wrfbdy_d02 file is getting generated. I have no error message even if I update the namelist.input to debug_level = 1000. I have attached the namelist.input, rsl.error.0000, and rsl.output.0000 files to this post.

Just to check that everything is correct, I even tried running real.exe for the child domains and that is also running fine. The problem is ndown.exe. Has anyone faced this error before?

Any sort of help is appreciated.

 

[jordanschnell]

Hi, I have not seen this before, but is it possibly a memory issue? Also, just a side note, in general we do not recommend using an even number for the parent-child grid size ratio.

 

[EeshanBasu]

Hi Jordan, thank you for replying. I am using 4 nodes and 192 cores for two hours and not getting wrfbdy_d02 for one day. So, it's highly unlikely that there is a memory issue.
I previously used the same namelist.input but for chem_opt = 17 (for GHGs), everything worked perfectly fine.

Any help is appreciated.

 

[llsalazardom]

Hi @EeshanBasu
Have you been able to solve the problem?
I am working with WRF 4.4.2 and I have the same problem, it does not generate the wrfbdy_d02 file

 

[EeshanBasu]

Hi @llsalazardom
I could not solve the issue on Stampede2, but the issue got automatically resolved when I used Cheyenne.

 

[llsalazardom]

Thank

 

[llsalazardom]

Hello @EeshanBasu
Could you tell me what configuration you have in Cheyenne?