Modified compsets with JRA interannual forcing to use year range 1958-2018

[matsbn]

With this both NUOPC and MCT based NorESM will use the same year range for JRA interannual forcing (1958-2018). This is then consistent with what was used for NorESM2 in CMIP6/OMIP2.

This PR also modifies the processor configuration for TL319_tnx1v4 grid combination to be consistent with or without iHAMOCC enabled. This avoids the PIO_STRIDE related crash on Betzy reported in NorESMhub/NorESM#593.

[JorgSchwinger]

With MCT, JRA v1.3 is used /cluster/shared/noresm/inputdata/ocn/jra55/v1.3_noleap

The new compset names suggest that JRA v1.4 is used? Is this version different from 1.3?

Otherwise, looks good to me

[matsbn]

With MCT, JRA v1.3 is used /cluster/shared/noresm/inputdata/ocn/jra55/v1.3_noleap

The new compset names suggest that JRA v1.4 is used? Is this version different from 1.3?

Otherwise, looks good to me

In an email from Hiroyuki Tsujino to an OMIP-CMIP email-list, he mentions:

Version 1.3.2 and 1.4.0 differ only in runoff around Greenland: version 1.3.2 uses climatological (1961-1990) runoff from Greenland (Bamber et al.2012) and version 1.4.0 uses interannually varying (1958-2016) runoff from Greenland and Canadian Arctic Archipelagos (Bamber et al.2018)

For NorESM2, we actually use v1.4 for the OMIP2 runoff forcing, combined with v1.3 for the atmospheric forcing. So I think the combined forcing in NorESM2 should be quite consistent with JRA v1.4, although I have not checked carefully. Anyway, with NorESM2.5 and beyond, identical results is not expected with NorESM2 and also we now use CICE6 compared to CICE5 in OMIP type experiments. Therefore, I think it is fine to use "pure" JRA v1.4 with NorESM2.5. Furthermore, JRA v1.5 is now supported by CDEPS (NorESMhub/CDEPS#9), extending the JRA forcing to 2023, and we should soon consider to try out this protocol which might be relevant for CMIP7/OMIP2.

[JorgSchwinger]

Thanks for the info @matsbn, I agree