Recommended spinup sequence with Matrix-CN / sasu

[hfduarte]

What version of the code are you using?
ctsm5.3.033

Have you made any changes to files in the source tree?
No

Describe your problem or question:
It seems that SASU is the new recommended approach for model spinup since the ctsm5.3 release. I've been trying to find some documentation on how to proceed with this new approach, but could not find very clear instructions... and some info I found seemed dated (4-step sasu).
From CLMBuildNamelist.pm, it seems that the recommended spinup sequence is:

CLM_accelerated_spinup=on [starting from cold start]
CLM_accelerated_spinup=sasu
CLM_accelerated_spinup=off
[then production runs]

Is this sequence correct? I found some info on a "post-sasu" step but I think that is dated now.
Based on your experience, how many years of simulation do you typically shoot for each spinup phase (on, sasu, off), for a typical CLM60%BGC compset?
Any advice would be super helpful.
Many thanks!

 

[oleson]

Sorry that the documentation is behind the model development. The sequence we've identified is AD, SASU, pSASU (for post-SASU). Typically, we'll run AD about 300 years, SASU about 320 years, and pSASU about 160 years. AD is with CLM_ACCELERATED_SPINUP=on. SASU is with CLM_ACCELERATED_SPINUP=sasu and reseed_dead_plants = .true., and pSASU is basically a normal run with CLM_ACCELERATED_SPINUP=off and reseed_dead_plants = .true.

 

[hfduarte]

Thank you, @oleson, for the information. This helps a lot! Just one more quick question: during pSASU, do you guys recommend keeping the matrix-cn on (i.e., use_matrixcn = .true. and use_soil_matrixcn = .true. and spinup_matrixcn = .false.) in the namelist? If that is the case, should the production runs also keep the matrix-cn configuration on?

 

[oleson]

No, the use_matrixcn and use_soil_matrixcn are .false., and the production runs have matrix-cn off.

 

[hfduarte]

Got it, thank you so much for the help!