Short term ways to address interpolation spinup stability issue

[wwieder]

Brief summary of bug

SpinupStability check fails when interpolating between f09 and ne30 initial conditions.

---

• [x] test using init_interp_fill_missing_with_natveg flag
  • see /glade/work/wwieder/ctsm5.3.0/ctsm5.4_5.3.068_PPEcal115f09_117b_pSASU
  • This still fails, after 160 years in pSASU 37.3% of grids are still in disequilibrium, poor image below shows grids are all over

• [ ] Initial Conditions. I suggest that for supported compsets we require initial conditions for each grid, maybe this is done anyway.
• [ ] Documentation. We have to make this limitation clear to users who may be running with unsupported setups or initial condition files

[wwieder]

@linniahawkins and @djk2120 what was your procedure for initializing PPE runs on the sparse grid. Was it still an AD, SASU, pSASU series from a spun up initial state? I wonder if that would help here?

[linniahawkins]

Our spin-up protocol was designed for paramfile changes (obviously), but may be relevant here:

• Start from initial conditions from 2-degree cold start with default parameters.
• Change paramfile and run 40y AD, 120y SASU, 40y pSASU at same resolution.

We had less stringent spin-up stability thresholds and this was sufficient for moderate paramfile changes for insufficient for large changes, particularly in the Arctic.

[slevis-lmwg]

From conversation with @wwieder @ekluzek @slevis-lmwg an investigation of the scope of the problem (i.e. is it just fv to se grids or more widespread?) will likely take a day or two.