wrf.exe crashes after a few days, no helpful error message found in rsl files

[marcusdwilliams]

Hello all,

I am attempting to run a multi-year simulation using WRF-Chem 3.8. For this particular run the chemistry options are turned off. I am able to complete WPS and real.exe with minimal issues. When I run wrf.exe the model crashes a few minutes into the run on day 5 of the simulation. I am generating restart files every 24 hours, but cannot successfully continue the run from a restart file. I am running in an HPC environment using 60 CPUs. I have attached the first and last rsl.error files as well. I have also completely removed the &chem section from the namelist. This stops the can't open aux5 error but the model still crashes on day 5. Any help would be greatly appreciated.

Thank you

 

[Ming Chen]

Your rsl files indicate that WRF_CHEM is compiled ad run for this case. Can you compile WRF-ARW, then try again? If WRF_ARW works fine, then you can move on to build ad run WRF_chem with those chemical options turned on.

 

[marcusdwilliams]

Thanks for taking the time to respond. You are correct that WRF_CHEM was compiled for this case. I was under the assumption that if chem_opt=0 the chemistry portion of the model was turned off. And I have compiled WRF-ARW and have run into a Segmentation fault - invalid memory reference error. I am testing a 1 year segment of the run instead of doing the full 5 years. The namelist and rsl.error file are attached. I changed the time step from 6*dx to 4*dx and that did not help

 

[marcusdwilliams]

***** Solved******

Found a CFL violation in on of the later rsl files. Changed the time step to 3*dx and it is working

 

[Ming Chen]

Now that WRF-ARW works fine, then I guess you can move on to WRF_chem and see how it works. If you still have problems, then it is possibly because of something wrong in chemistry. please post your question under the topic of WRF_CHEM.