Error while running WRF-Chem V4.5.1 with CBMZ and MOSAIC chemistry

[antoinep]

Hello,

I wish to use WRF-Chem V4.5.1 together with the CBMZ/MOSAIC chemistry option on two nested domains over Eastern Europe for a simulation of 54 hours. However, the run stops after some time whenever I select aerchem_onoff = 1 for my domains in order to include aerosol chemistry. The run stops later if I increase the update times chemdt or photdt, but the result is ultimately the same. The error file always shows the same kind of error message associated with mosaic_aerchem_driver:

d02 2022-03-26_18:17:18 aerosols_driver calling mosaic_aerchem_driver
*** Error in `./wrf.exe': malloc(): smallbin double linked list corrupted: 0x00000000196f64b0 ***

I tried to set as few options in the chemistry section of the namelist as possible, and could not find any additional option that would solve the problem. See the attached files if you would like to check my namelist and the corresponding error file. I also tried different chem_opt options such as 7 - 10, 31, and 170 without finding any improvment. I am assuming that 32 -34 will lead to the same error too. Note that I could successfully run WRF with gas and aerosol chemistry turned off in this namelist.

Can you help me to solve this issue?

Thank you in advance.

Antoine

 

[Jiayu Xu]

Hi Antoine,
I met just the same error when running WRF-Chem V4.5.1 and CBMZ-MOSAIC chemistry as you. Have you figured it out?
Thank you very much!
Jiayu

 

[antoinep]

Hi Jiayu,

I did not find a solution to this problem, so I abandoned the idea of using these chemistry schemes and opted for MOZART/GOCART instead, which worked properly. The corresponding namelist option is chemp_opt = 112.

Best,
Antoine

 

[Jiayu Xu]

Hi Jiayu,

I did not find a solution to this problem, so I abandoned the idea of using these chemistry schemes and opted for MOZART/GOCART instead, which worked properly. The corresponding namelist option is chemp_opt = 112.

Best,
Antoine

Thank you very much for your reply, Antoine.

I guess this issue might be caused by specific "bad" nodes on the HPC platform or by using too many cores at once. Because when I submitted the same task multiple times without making any changes, it successfully ran through on one occasion. Additionally, when I reduced the number of cores used, the model can also ran stably.

However, considering you mentioned that the model runs successfully after changing the chem_opt, my guess might not be entirely accurate. Anyways, I wish you the best of luck with running the model :）

Jiayu