Issues encountered with RAP and HRRR SDFs in chgres_cube with fractional grids

[MichaelLueken]

While attempting to update the UFS_UTILS hash, it has been noted that the RAP and HRRR WE2E tests fail at hash 7addff5 (associated with PR #873). This update introduced fractional grids to chgres_cube. As part of this update, the field tiice, was added. The issue stems from the fact that both the RAP and HRRR SDFs use RUC LSM. When RUC LSM is used, Model%kice is set to 9 vertical layers. However, tiice only has two vertical layers in the initial conditions. This is causing the RAP and HRRR SDFs to fail to read in the new tiice field. Setting Model%kice to 2 will allow the weather-model to read in the sfc_data file, the weather-model will still fail with a signal 11 segfault while dealing with ice fields at the end of the model run.

As of this time, issue #961 has been created in UFS_UTILS to see if their team can add back the generation of the v1 sfc_data file so that RAP and HRRR can be used with the fractional grid update. Another issue will be opened in the FV3ATM repository to see if something can be done to ensure that tiice will use two vertical layers, regardless of whether RUC LSM is used in the SDFs or not. I am welcome to ideas that we can use as temporary fixes before authoritative repository fixes are made available.

Some ideas that I have as of this time include:

• Update the UFS_UTILS hash and remove all RAP and HRRR WE2E tests - worst of the lot
• Update the UFS_UTILS hash and replace all RAP and HRRR WE2E tests with RRFS_V1BETA (or some other non-RUC LSM based physics suite)
• Update the UFS_UTILS hash to the v1.12.2 tag that is currently used in the rrfs_workflow repository (the latest tag that doesn't include fractional grids)
• Add warm start capability to the SRW App and save the grid, orography, sfc_clio, ICs, and LBCs data for all RAP and HRRR WE2E test cases, update the UFS_UTILS hash, then run all RAP and HRRR WE2E tests as warm start tests

Please feel free to discuss in this GitHub Discussion and be prepared for the next SRW CM meeting on July 11.

Thank you very much for your time!

[MichaelLueken]

GitHub Issue #854 was opened in the FV3ATM subcomponent repository to see about allowing RUC LSM physics suites to work with internal ice temperature (tiice).