Update config_compsets.xml and config_pes.xml on release-1.6 branch?

[TomasTorsvik]

There is a suggestions to port back relevant changes from the v1.7 tag to the release-1.6 branch, which is the current release branch for NorESM2.3. One such issue was #431.

I see that both config_compsets.xml and config_pes.xml files have evolved quite a lot between 1.6 and 1.7, so I wonder if there are parts of these files that should be ported.

[JorgSchwinger]

I know this has been discussed before (but I couldn't follow), why don't we just use 1.7 for the NorESM2.3 release?

If this is not possible, I don't see why we should update these two files.

[jmaerz]

In how far is the v1.7 still CMIP6 compatible? For the v1.6, some settings via nml allow to get back the original behavior (https://github.com/NorESMhub/BLOM/wiki/new-BLOM-with-CMIP6-settings). Admittedly, I struggle to follow the BLOM changes a bit (as in: in how far do the new developments change the answer in the isopycnic model version and which switches would allow to recover CMIP6 behavior - whether bfb or at least the physics).

One of the reasons that I also suggested previously to have tag 1.8 which still supports MCT - so being still compatible with NorESM2.3 (to be scientifically useful - the v1.7 tag has still issues in the hybrid coordinates with the virtual freshwater fluxes, as far as I am aware of - which makes it a bit less useful for NorESM2.3).

[TomasTorsvik]

In NorESM2.3 we rely on the checkout_externals script to fetch the source code. This has been replaced by git submodules for NorESM2.5, so it is necessary to add an externals_BLOM.cfg file (this could maybe be included if we make a release-1.7 branch?). Other than that, I thought it should be possible to run NorESM2.3 + v1.7.0 with the namelist changes for v1.6. The default is now hybrid coordinates, so an additional xmlchange BLOM_VCOORD="isopyc_bulkml" is also required. Running a test now.

[matsbn]

If only BLOM appears in the compset (not BLOM%HYB), then isopycnic should still be selected in current master. In 2.3 I think only isopycnic will be tested and primarily used, and none of the later BLOM changes should make a impact with isopycnic coordinate.