WRF-ELEC: quick termination at first time step when elec_physics on

[weny]

Dear colleagues,

I encountered a problem while running WRF-ELEC and have taken some measures, but the issue remains unresolved. I am writing to seek your advice.

Problem description:
The case runs successfully when elec_physics is turned off, and the simulation completes. However, when I turn it on, during the wrf.exe stage, the first time step of d01 finishes almost immediately. Details are provided in the attached rsl.error.0000 file (uploaded as Attachment 1). My server runs KylinSec Advanced Server OS, which is similar to RedHat.

Attempt 1:
The error always occurs at:

d01 2022-07-27_00:00:00 call rk_update_scalar
d01 2022-07-27_00:00:00 call rk_scalar_tend

Following the guidance at CM1 Documentation - NSSL microphysics, I modified the advection options and set the shell stack memory limit:

• moist_adv_opt = 3, 3, 3 (5th-order WENO for moisture)
• scalar_adv_opt = 3, 3, 3 (5th-order WENO for scalars, including charge)
• ulimit -s unlimited
• OMP_STACKSIZE=16M
  Result: The same error persists.

Attempt 2:
Since the model runs without the electrical process, could the issue be in the electric calculation module? Is it possible that BoxMG4WRF is not linked correctly?
I checked the executable files (e.g., wrf.exe) and found that they contain the relevant links. Moreover, no errors were reported when compiling WRF.


Attempt 3:
check real.exe , wrfinput


Now I don't know what else to try. Attached are the namelist and rsl files. I would greatly appreciate any insights or suggestions. Thank you very much.

 

[kwerner]

It looks like you're running this with only a single processor. It's possible that when using elec_physics, more computation is required. If you have more processors available, can you try building and running this with more processors and see if that makes any difference?

Addtionally, a lot was updated with the NSSL scheme(s) between WRFv4.6.0 and WRFv4.7.1 (which is the most-recently released code). Can you try with the newest version to see if there is any difference? Note that in the newer version, NSSL microphysics have been consolidated into a single scheme (option 18).

If it still fails after those 2 tests, please attach your latest namelist.input file, and please package all rsl* files into a single *.tar or zipped file and attach that, as well. Thanks!