SST perturbation experiment

[PappuP]

Hello, I am currently planning to run SST perturbation experiment (--compset F2010 --res ne30pg2_r05_IcoswISC30E3r5). So my plan is to modify the default (1 degree) SST input file in different location such as West Pacific , East Pacific, Southern Ocean, and many other location.

For each of the modification do I need to create new case & build every time? As The only thing I want to change the path of SSTICE_DATA_FILENAME. I already create, build and run West Pacific (WP) experiment. Can I do create_clone from WP case instead of creating new case and build for all the experiment? which saves a lot of time in my case as I do have a lot of different SST modification. Or hybrid run or branch run would be better?

My understanding is that model use SST data only when it runs? So it is time-consuming and unnecessary as the same executable may be used for multiple SST experiment?

[mahf708]

Yes, the SST setting in question is a runtime parameter. Would you be willing to share your workflow script(s) with us to help you further?

I think you are most concerned about the build process, correct? Creating new cases (without building them) is relatively painless.

If I were you, I would follow the following workflow. I would create a baseline case and build it as normal; if desired, run it. Let's assume this case has the path /path/to/my/base/case/. Then, I would set up new cases (the sensitivity cases you're looking to conduct) where I would set do_case_build=false (here) and set the readonly OLD_EXECUTABLE=/path/to/my/base/case/build/e3sm.exe (here). It may help to use a script that is similar to the one we provide in the root of the repo: https://github.com/E3SM-Project/E3SM/blob/master/run_e3sm.template.sh. Also be sure to include other edits as desired in your sensitivity runs. The namelist setting SSTICE_DATA_FILENAME is the one you likely have in mind.

Let us know if this helps, and best of luck!

[PappuP]

Hi I am little bit confused for the third step, I used F2010 compset and if I run my baseline case for 21 years. What should be my START_DATE and RUN_REFDATE in this part of the script for other cases?

readonly START_DATE="?"
# Additional options for 'branch' and 'hybrid'
readonly GET_REFCASE=TRUE
readonly RUN_REFDIR="/path/to/my/base/case_name/archive/rest/2020-01-01-00000"
readonly RUN_REFCASE="case_name"
readonly RUN_REFDATE="?"   # same as MODEL_START_DATE for 'branch', can be different for 'hybrid'
`````````

[mahf708]

What are your MODEL_START_TYPE, START_DATE, STOP_OPTION, STOP_N, REST_OPTION, and REST_N in the baseline case?

Generally, for "branch" run, the date will be the one that you will see in archive/rest. If you started your F2010 case from 0001-01-01, that will be 0022-01-01 (after 21 years).

[PappuP]

readonly MODEL_START_TYPE="initial"  # 'initial', 'continue', 'branch', 'hybrid'
readonly START_DATE="0001-01-01"
  readonly STOP_OPTION="nyears"
  readonly STOP_N="1"
  readonly REST_OPTION="nyears"
  readonly REST_N="1"
  readonly RESUBMIT="20"
  readonly DO_SHORT_TERM_ARCHIVING=false

[mahf708]

Yeah, so it should be 0022-01-01. Please remember to do ./case.st_archive for the above to work. You should be able to see the restart dates under archive/rest and then you know you have the right numbers, etc.

[PappuP]

Great, thank you so much, I will try it when my baseline case run end. Then I will mark it as answered