AMIP-style sims in UFS/SFS

[barlage]

Looking for any information on the preferred (or non-preferred) method for running AMIP-style simulations, specifically in an SFS configuration. I would like to perform a small set of prescribed ocean/ice simulations to test some land modifications. 

To familiarize myself a little with what seems to be out there, I ran the atm_ds2s_docn_pcice regression test (it seems atm_ds2s_docn_dice for some reason is commented in rt.conf). From what I can tell, the data models (both ocean and ice) are using: 

/scratch4/NAGAPE/epic/role-epic/UFS-WM_RT/NEMSfv3gfs/input-data-20251015/DOCN_DICE_ERA5/sst_siconc_skt.era5.20210321_20210331.nc

specifically the sst and ice_cov fields. Presumably these are coming from ERA5. Do these files exist somewhere for other time periods? Or is there some documentation on creating them? Are there other options for sst/ice inputs?

It looks like the data in this file are hourly, but are really daily (no change for 24 hours) so can we just create daily timestep files? Since these will be one-year sims, it would be preferred for smaller input files if the hourly information is not used/needed.

[NeilBarton-NOAA]

FYI @taozhang2026

[barlage]

There is a related global_workflow issue.

[NickSzapiro-NOAA]

Hi @barlage . For the RT, I downloaded those from ECMWF, like https://cds.climate.copernicus.eu/datasets/reanalysis-era5-single-levels-timeseries?tab=overview

Any other datasets for SST and sea ice concentration should work too. Just also need to create a corresponding ESMF mesh description (e..g., https://ufs-weather-model.readthedocs.io/en/develop/InputsOutputs.html#mesh-generation).
Please let me know if more detailed support would help.

atm_ds2s_docn_dice is fine too...that is dependent on inputs of coupling variables from separate run.
It's commented to reduce the number of "feature" RTs (without more operational relevance) being tested on every commit.

[barlage]

OK, thanks, @NickSzapiro-NOAA. I think it would be pretty straightforward to create this sst/ice_cov file from data on the NCAR RDA as well. It looks like @taozhang2026 may have an AMIP solution close to working in global_workflow using the old global cycle method.