search

ESCOMP / CTSM

Community Terrestrial Systems Model (includes the Community Land Model of CESM)
http://www.cesm.ucar.edu/models/cesm2.0/land/
Other
302 stars 307 forks source link

Some collapse options should be disabled with FATES #1780

Closed ekluzek closed 1 month ago

ekluzek commented 2 years ago

@rgknox points this out in his most recent PR #1766. The collapse options that work on landunit should work provided FATES is using create_crop_landunit=F and 16-PFT datasets are being read and you don't have FATES managing crop landunits. But, "n_dom_pfts" is incompatible with FATES, except maybe for FATES-SP mode. I think we should disable it for all FATES modes though. "toosmall_soil" is also problematic and likely should be disabled.

When FATES can manage crop landunits, the "toosmall_crop" might be problematic. But, for the case when we run FATES for natural veg and BGC-CROP for CFT's, "toosmall_crop" should be fine.

It also a question on if we should test FATES with the collapse options. I don't think this is a use case that would at all be common so I'm thinking it shouldn't be.

ekluzek commented 2 years ago

We had some discussion about this last week. We figure we should allow options that should work with FATES to exist, but won't add explicit testing for it, as we don't think this is a normal way of working. Perhaps if people start to use it down the line, we'll need to add testing.

We did disable the use of n_dom_pfts with FATES in #1766, but this should likely also going into the FORTRAN code and apply to the other things pointed out above as well.

glemieux commented 9 months ago

Just a quick note that this came up in testing fates landuse v1; see https://github.com/ESCOMP/CTSM/pull/2076#issuecomment-1858405373. In summary, collapse_to_dominant was being passed numpft bounds to set the dummy weight argument bounds inside the procedure, even though surfpft bounds are technically used to allocate the actual input weight argument. Currently these two bounds will only ever diverge if fates is used. The short term fix proposed and implemented in the above PR is to simply align the input argument bounds with the allocation argument bounds.

I should note that this appears to only be an issue for the nag compiler. This passed through the gnu compiler on derecho.

ekluzek commented 1 month ago

I'm converting this to a discussion. If we work on this we should map out exactly what needs to be done.