Configuration for sst restoring/pacemaker run on esm

Technical
, ,
1

Hi there,

It’s been discussed, but I can’t find it in the resources. So I’m posting my question again here.

It’s for esm 1.5, by using payu, and want to restore sst or sst pacemaker experiment.

I guess we need to restore the sst each year the job finishes and need a script to replace the restore file.

so that we need to add get_restore_sst_file.sh in config.yaml

userscripts:

postscript: -v PAYU_CURRENT_OUTPUT_DIR,PROJECT -lstorage=${PBS_NCI_STORAGE} ./scripts/NetCDF-conversion/UM_conversion_job.sh

add something in here

postscript: -v PAYU_CURRENT_OUTPUT_DIR,PROJECT -lstorage=${PBS_NCI_STORAGE} ./scripts/sst_restore/get_restore_sst_file.sh

and also copy to ./scripts/sst_restore/get_restore_sst_file.sh

is that right? More details are welcome, or please point out my mistake.

3

Hi, @arnoldsu , @spencerwong has some experience with a similar issue and is happy to help, I’ve already assign him to this topic.

4

Hi @arnoldsu,

@georgyfalster has previously run some pacemaker simulations using ESM1.5, and has shared the experiments in the branches of the GitHub repository here.

The commit here shows the changes used to activate the sst restoring, in particular:

  • Adding the sst restoring file temp_sfc_restore.nc and mask restore_mask.nc to the ocean inputs in the config.yaml
  • Activating the restoring in the MOM5 settings with the changes in ocean/input.nml

It’s possible to use a single SST restoring file containing many years of data rather than swapping to a new file each year. As long as the model’s internal time falls within the date range for the restoring file’s time coordinate, it should pick up the correct data for each month, and so you should be able to skip adding an extra script.

MOM5 has some strict requirements about the format of the restoring file, and so if you’re setting up a new one there are a few recommendations in the topic here.

Let me know if you have any questions about the above. @georgyfalster feel free to chime in if you have any additional advice!

5

@arnoldsu let me know if you have any trouble creating the input files and I can either share examples or the script we used to make them.

6

@spencerwong @georgyfalster Dear all, many thanks for your help. I mainly try to support the other colleagues’ issues. So can I double check @georgyfalster does your file contain 500 years data rather than swap each time when job finish? Can I have access to your files? My concern will be the year, say if I started from 1950, so does the file I provide should be also follow that? Many thakns again.