BLOM-iHAMOCC offline spin up

[adagj]

Hi,
@tjiputra @AleksiNummelin @matsbn @TomasTorsvik @ChunchengGuo

We wish to include a description on how to run a BLOM-iHAMOCC offline spin up in the documentation. Can you please help us with that?

• What kind of forcing files are needed? What do we add to the namelists to provide the forcing files?
• How many years spin up do you recommend?
• Do you run the spin up in cycles?
• What is the recommended compset? and set up in general?

I think a description can be added to this file:

https://github.com/NorESMhub/NorESM/blob/master/doc/configurations/omips.rst

[ChunchengGuo]

Hi, I remember @JorgSchwinger ran the OMIP style BLOM-iHAMOCC spinup (correct me if I'm wrong) and would be the most knowledgeable person here.

[tjiputra]

Hi @adagj,

I meant to discuss this with you last week. I had a quick chat with @monsieuralok today, who also helped with the offline blom-ihamocc spin up last time. I think it was @oyvindseland who produced the CPLHIST forcing fields needed to run the offline spinup. Perhaps he can direct us to the case directory and we can start look into this (i believe we need to simply modify the user_nl_cpl files). Alok will also dig into the case of the offline spinup run to gather more details on the offline setup (e.g., some salinity relaxation is required, among others).

To answer your questions, we should aim for at least 1000 years of spin up. And if I am not mistaken, we're using some 100 years of CPLHIST forcing, so yes, it was cycled.

best,
jerry

[adagj]

Hi @tjiputra, @monsieuralok, @matsbn
I think the CPLHIST forcing fields can be generated by modifying the cpl namelist:

- histaux_a2x: .false. -> .true.
- histaux_a2x1hr: .false. -> .true.
- histaux_a2x1hri: .false. -> .true.
- histaux_a2x24hr: .false. -> .true.
- histaux_a2x3hr: .false. -> .true.
- histaux_a2x3hrp: .false. -> .true.
- histaux_r2x: .false. -> .true.

i.e. adding to the user_nl_cpl:

&seq_infodata_inparm
histaux_a2x = .true.
histaux_a2x1hr = .true.
histaux_a2x1hri = .true.
histaux_a2x3hr = .true.
histaux_a2x3hrp = .false.
histaux_a2x24hr = .true.
histaux_l2x = .true.
histaux_l2x1yrg = .true.
histaux_r2x = .true.

But that is all I know, I have no idea how to set up and run BLOM-iHAMOCC stand alone or offline :)

Ada

[JorgSchwinger]

Hi Ada

I'm currently also experimenting with running BLOM with cplhist-forcing. I hope I'll get this running during the next days. Once I have this I can tell you more (there is a dedicated compset for running BLOM/iHAMOCC with the cpl-forcing)

Jörg

[adagj]

Great @JorgSchwinger! Let's discuss this soon then. Would be good to document it before the NorESM user workshop and the INES meting. As long as I get some help with the setup, I can update the documentation myself ( I hope :)

[matsbn]

It's a while ago, but I created CPLHIST output for f09_tn14 configuration for accelerated CLM and HAMOCC pre-industrial spinup. The case used to produce the output is on Fram here:

/cluster/projects/nn2345k/matsbn/NorESM/cases/N1850_f09_tn14_20190726

The CPLHIST output can be found here:

/cluster/shared/noresm/inputdata/cplhist/N1850_f09_tn14_20190726_751-850

We created a compset, NOICPLHISTOC, to run MICOM/HAMOCC with CPLHIST forcing and a test of this compset can be found on Fram here:

/cluster/projects/nn2345k/matsbn/NorESM/cases/NOICPL_f09_tn14_20190812

The NOICPLHISTOC compset is still defined in components/blom/cime_config/config_compsets.xml. Yuanchao used this CPLHIST forcing for offline spinup of CLM that we later incorporated in the continuation of the spinup.

[JorgSchwinger]

There is also a post processing script involved that gathers daily output into monthly file.

I managed to locate it and will use it. I'll let you know how it works and what to do. Probably this script should be stored somewhere in the NorESM repository(?)

[matsbn]

Indeed, @JorgSchwinger. If useful, the scripts I used for post-processing are here on Fram:

/cluster/projects/nn2345k/matsbn/NorESM/concat_cpl_hist_mon

Great if you are able to go through the production and application of CPLHIST with a recent codebase.

[monsieuralok]

Hi @adagj,

I meant to discuss this with you last week. I had a quick chat with @monsieuralok today, who also helped with the offline blom-ihamocc spin up last time. I think it was @oyvindseland who produced the CPLHIST forcing fields needed to run the offline spinup. Perhaps he can direct us to the case directory and we can start look into this (i believe we need to simply modify the user_nl_cpl files). Alok will also dig into the case of the offline spinup run to gather more details on the offline setup (e.g., some salinity relaxation is required, among others).

To answer your questions, we should aim for at least 1000 years of spin up. And if I am not mistaken, we're using some 100 years of CPLHIST forcing, so yes, it was cycled.

best, jerry

@tjiputra we used 50 years of CPLHIST forcing and approx. we have run it for 1800 years. But, we could check all data on Monday together.

[adagj]

Hi all,
For the INES meeting, Kjetil Aas in the CLM/CTSM group @uio is going to present something on offline spin up of CLM and some new methods for speeding up the spin up of the clm component. He has not tested it him self so it will be more a discussion, than a presentation. I think the offline spin up of BLOM/iHAMOCC fits very well into that session (WP3), so if we gather compsets, scripts, experiences, recommendations by then we can at least discuss what will be a more efficient method to calibrate a new NorESM version - more efficient meaning less than 9 months or so :-)

[tjiputra]

@matsbn thank you for your information. In your stand-alone micom-ihamocc run, you also use sss relaxation toward the climatology from the coupled run outputs, e.g. "sss_climatology_N1850_f09_tn14_20190726_751-850_classic.nc". Can you show me where I can find this file and how to generate it? Blom file ~/blom/phy/rdcsss.F also needs to be modified to allow reading this sss climatology file. How should we deal with this?

[matsbn]

@tjiputra, I used the script /cluster/projects/nn2345k/matsbn/NorESM/sss_climatology/sss_climatology.py (I think @monsieuralok made it initially) to create a SSS climatology from the simulation used to produce the CPLHIST forcing. The climatology can be found on Fram here:

/cluster/shared/noresm/inputdata/ocn/micom/tnx1v4/20170601/sss_climatology_N1850_f09_tn14_20190726_751-850_classic.nc

In the NOICPL_f09_tn14_20190812 case, I modified rdcsss.F in SourceMod/src.micom to be able to read it. Not a very transparent workflow I must admit. Fortunately, in BLOM one specifies the SSS climatology file to be used by the namelist variable "SCFILE", so setting "SCFILE = <filename>" in "user_nl_blom" should do the trick.

[adagj]

I have included a recipe from @tjiputra here: #314
Please comment.

[adagj]

One more thing, will it be useful to have these scripts in a shared folder/repository?

/cluster/projects/nn2345k/matsbn/NorESM/concat_cpl_hist_mon/concat_cpl_hist_mon.csh
/cluster/projects/nn2345k/matsbn/NorESM/sss_climatology/sss_climatology.py

And is is only the sea surface salinity which needs to be relaxed towards a climatology?

[tjiputra]

@adagj I agree, perhaps in: /cluster/shared/noresm/blom_standalone/tools?
Yes, so far we only see the need to relax sss.

[tjiputra]

It's a while ago, but I created CPLHIST output for f09_tn14 configuration for accelerated CLM and HAMOCC pre-industrial spinup. The case used to produce the output is on Fram here:

/cluster/projects/nn2345k/matsbn/NorESM/cases/N1850_f09_tn14_20190726

The CPLHIST output can be found here:

/cluster/shared/noresm/inputdata/cplhist/N1850_f09_tn14_20190726_751-850

We created a compset, NOICPLHISTOC, to run MICOM/HAMOCC with CPLHIST forcing and a test of this compset can be found on Fram here:

/cluster/projects/nn2345k/matsbn/NorESM/cases/NOICPL_f09_tn14_20190812

The NOICPLHISTOC compset is still defined in components/blom/cime_config/config_compsets.xml. Yuanchao used this CPLHIST forcing for offline spinup of CLM that we later incorporated in the continuation of the spinup.

@matsbn what and where is the case where you recoupled the output from "NOICPL_f09_tn14_20190812" (875 years) and continue the spin up in coupled mode?

[matsbn]

@tjiputra, I do not recall that we recoupled the "NOICPL_f09_tn14_20190812" simulation for continuation in coupled mode. For the NorESM2-MM spinup, it was only the CPLHIST forced CLM spinup that was recoupled.

[JorgSchwinger]

Regarding the scripts, would it be an idea to include them into the BLOM repository?

[adagj]

@JorgSchwinger I think that is a good idea. Maybe the BLOM team can fix?
@matsbn since you are the owner of these scripts, maybe you can decide how to make these scripts available and then enlighten the rest of us? :-)

[adagj]

One more thing, will it be useful to have these scripts in a shared folder/repository?

/cluster/projects/nn2345k/matsbn/NorESM/concat_cpl_hist_mon/concat_cpl_hist_mon.csh /cluster/projects/nn2345k/matsbn/NorESM/sss_climatology/sss_climatology.py

And is is only the sea surface salinity which needs to be relaxed towards a climatology?

@tjiputra @JorgSchwinger @matsbn

Is this issue solved? If not, should we maybe finish it before the new NorESM2 release on Friday 9th ?