Land-mask bug in sfcsub.F

[GeorgeGayno-NOAA]

Description

The UFS_UTILS global_cycle program is being updated to:

• Link to the sfcsub.F from the ccpp repository.
• Work with fractional grids.

During initial testing, a failure occurred in sfcsub.F routine newice. My test (C768, tile1 only) was not run with a fractional grid. Here is the error message:

FATAL ERROR: inconsistency in slifcs or slianl.
  at lat=-34.9 lon=305.0 slifcs= 1.0 slimsk= 0.0 tsffcs=280.0 set to tsfanl=280.0
abort:

The problem point has a mask of '1' (land) on input to sfcsub.F (sfcccycle). The error indicates the land mask is changing from '1' (land) to '0' (open water). That can't be correct.

I am using sfcsub.F at 8e670b0

After some investigation, and consultation with @SMoorthi-emc, he suspects there is a bug in this loop:

ccpp-physics/physics/sfcsub.F

Line 2020 in 8e670b0

Updating the loop as follows eliminates the failure:

ufs-community/UFS_UTILS#549 (comment)

Additional Context

See UFS_UTILS issue ufs-community/UFS_UTILS#549 for more details.

[junwang-noaa]

@GeorgeGayno-NOAA would you please create a CCPP PR for this fix? We hope to get this fix committed soon.

[GeorgeGayno-NOAA]

@GeorgeGayno-NOAA would you please create a CCPP PR for this fix? We hope to get this fix committed soon.

Yes, I will create PR today.

[GeorgeGayno-NOAA]

@GeorgeGayno-NOAA would you please create a CCPP PR for this fix? We hope to get this fix committed soon.

@junwang-noaa See #721.