What version of the code are you using?
CESM2.2.2
Have you made any changes to files in the source tree?
No CAM source-code changes are required to reproduce the problem described here.
The test case is a clean FXmadHIST case using the default initial condition selected by preview_namelists.
Describe every step you took leading up to the problem:
I created a WACCM-X + MAD chemistry case:
The generated chemistry mechanism confirms that the case is using MAD chemistry:
User-given Tag Description: WACCM_MAD_MAM4
Tag database identifier : MZ280_MAD_MAM4_20210122
The mechanism contains MAD / D-region ion species such as:
The default ncdata selected by preview_namelists is:
The model starts successfully and writes some history output, but then the CAM chemistry solver repeatedly fails to converge. The typical log sequence is:
After these imp_sol failures, the model eventually crashes with a segmentation fault in WACCM-X electrodynamics:
Describe your problem or question:
I would like to ask whether this is a known issue with FXmadHIST / WACCM_MAD_MAM4 in CESM2.2.2, especially with the f09_f09_mg17 grid.
The main issue is that FXmadHIST starts successfully but later fails in the MAD implicit chemistry solver:
My questions are:
Best,
Tianliang
CESM2.2.2
Have you made any changes to files in the source tree?
No CAM source-code changes are required to reproduce the problem described here.
The test case is a clean FXmadHIST case using the default initial condition selected by preview_namelists.
Describe every step you took leading up to the problem:
I created a WACCM-X + MAD chemistry case:
And submit
The generated chemistry mechanism confirms that the case is using MAD chemistry:
User-given Tag Description: WACCM_MAD_MAM4
Tag database identifier : MZ280_MAD_MAM4_20210122
The mechanism contains MAD / D-region ion species such as:
The default ncdata selected by preview_namelists is:
The model starts successfully and writes some history output, but then the CAM chemistry solver repeatedly fails to converge. The typical log sequence is:
After these imp_sol failures, the model eventually crashes with a segmentation fault in WACCM-X electrodynamics:
Describe your problem or question:
I would like to ask whether this is a known issue with FXmadHIST / WACCM_MAD_MAM4 in CESM2.2.2, especially with the f09_f09_mg17 grid.
The main issue is that FXmadHIST starts successfully but later fails in the MAD implicit chemistry solver:
and only afterward crashes in:
PS: The log files are too big (hundreds of GB) due to the repeated imp_sol faileed to converge
My questions are:
- Is FXmadHIST with f09_f09_mg17 expected to be stable in CESM2.2.2?
- Are there known issues with the MAD implicit solver for not-transported D-region ion species such as CO3m, CO4m, and NOp_CO2?
- Since the default FXmadHIST initial condition is named like an FXHIST initial file and does not contain these MAD-specific ion species, how are these not-transported implicit MAD species initialized internally? Or shall I just DIY the initial file for FXmadHIST?
Best,
Tianliang