MPAS model forecast stopped right after hour 0, segmentation fault

[czhou]

I am trying to run MPAS model forecast on the North America 3-km domain but my forecast stopped just a few time steps after hour 0, with the values of the global min, max w and global min, max u way too big. I am attaching the log file and namelist file here. Can somebody provide some clues on how to fix this? I have been playing around with the time steps and damping coefficients but haven't got any luck yet. Thank you in advance for any comments/suggestions!

 

[mgduda]

The min/max velocities reported in the log file are unreasonably large as you said. It could be helpful to understand where the min/max values are occurring, and you can use these namelist settings to get information on the latitude, longitude, and k-level of the extreme values:

&printout
    config_print_detailed_minmax_vel = true
    config_print_global_minmax_sca = true
/

From your namelist.atmosphere file it also looks like you might be doing DA with MPAS-JEDI (config_jedi_da = true)? Have you tried a cold-start simulation (no DA) to see whether that works?

 

[czhou]

Thank you for your reply! I will add the namelist for the printout and update it here.
The namelist option (config_jedi_da = true) was set so to use mpasout instead of the restart files, and it worked fine for our CONUS-12km runs (cold start or DA runs). This is a cold start simulation.

 

[czhou]

I am attaching the log file with the printouts here. Looks like the unreasonable wind appeared at the SE corner of the domain. Thank you!