Can GEOSldas use cube-sphere-based forcing inputs for corrected precipitation?

[gmao-rreichle]

Question: In GEOSldas, could the corrected-precipitation forcing inputs be on the cube-sphere grid while all other "lfo" forcing inputs would be on the lat-lon grid? What would it take to make this work?

Background:
GEOSldas forcing is typically from GEOS-generated instantaneous and time-average "land forcing" (lfo) nc4 files. For precipitation, GEOSldas also accepts pre-processed files that include observation-corrected precipitation.
To support the coupled LDAS-ADAS system, functionality was added that lets GEOSldas used cube-sphere-based "lfo" forcing inputs, as long as the GEOSldas tile space matches the cube-sphere grid of the forcing data. For example:

GEOSldas/src/Components/GEOSldas_GridComp/GEOSmetforce_GridComp/GEOS_MetforceGridComp.F90

Line 619 in a019db8

if(trim(grid_type) == "Cubed-Sphere" ) then

GEOSldas/src/Components/GEOSldas_GridComp/GEOSmetforce_GridComp/LDAS_Forcing.F90

Line 4123 in a019db8

logical :: isCubed ! forcing on cs grid: true/false

cc: @gmao-qliu

[weiyuan-jiang]

If the simulation is on EASE grid but the forcing is on cubed grid, the interpolation of the forcing is not easy. That is why we force the simulation on identical cubed grid when forcing is on cubed grid.

[gmao-rreichle]

@weiyuan-jiang: Sorry, I wasn't clear. The simulation would be on the cube-sphere tile space that is associated with the cube-sphere grid of the obs-corrected precip data. That is, the EASE grid does not come into play at all.

[weiyuan-jiang]

If so, GEOSldas should work. If it doesn't, I believe the change should be easy.

[gmao-qliu]

@weiyuan-jiang does it work if other forcing variables are on lat/lon grid while precipitation variables are on cubed sphere grid?

[weiyuan-jiang]

I think it should work. @gmao-qliu But if if see the error, please let me know.