CTSM5.2 regional simulation: h0/h1 outputs show north–south flipped land mask; target study region filled with NoData

[kylin]

I am running CTSM5.2 regional spinup simulation on a 0.01°×0.01° grid covering the Qinling Mountains.

The model completes successfully and produces h0/h1 history files. Visualization reveals a strange artifact:

The Qinling target region is entirely NoData in outputs.

This NoData zone matches the geometry of my target region flipped north–south.

Grid cells within the rectangular domain but outside the target zone have valid simulated values.
Checks I have performed:

domain.lnd.nc: mask = 1 for the target region, mask = 0 elsewhere.

surfdata: PFTDATA_MASK = 1 across the study domain, PCT_OCEAN = 0.

mask_mesh.nc: elementMask derived from the identical regional land mask.
I suspect inconsistencies in latitude ordering among domain file, SCRIP grid and mask_mesh, but I cannot identify which component causes the mismatch.

Has anyone encountered comparable issues of north–south mask flipping or output NoData problems in CTSM regional configurations?

 

[oleson]

Check to see that your latitude coordinates (e.g. nodeCoords or centerCoords) are oriented south to north in your mesh file

 

[kylin]

Check to see that your latitude coordinates (e.g. nodeCoords or centerCoords) are oriented south to north in your mesh file

Thanks. I have fixed the north-south flip error of the study area. Unfortunately, all variables inside my target region remain NoData in the exported h0 and h1 files, as shown in the attached figure:

I wonder if you have any suggestions to resolve this problem.

 

[oleson]

There must still be some problem with the mask and the associated lat/lon coordinates in one or more of the associated files. I would also look at the surface dataset and the atmospheric forcing as well.