Update default fire_emis_factors_file

[samsrabin]

As described in ESCOMP#2734.

Tested by building namelist in a case with -fire_emis included. If you'd like me to run aux_clm or some subset of it, let me know.

One potential issue is that it's going from a 16-PFT file to a 78-PFT one; not sure if that might cause problems with certain compsets/settings.

[ekluzek]

@samsrabin I looked at the files a tiny bit, just to see what made sense to me. From the meta-data it looks like Fang work with Louisa and others on getting this right which also resulted in some papers. So that seems really good to me.

As you point out it does expand the fields out from 16-pft to all 78 with all crops. So I wonder if the natural PFT's are the same or if any of them are different. But, that would take a little more analysis beyond cursory.

But, the number of compounds is expanded as well, which makes me think that this is likely to affect natural PFT's as well. So it seems likely to change answers even for natural PFT's.

This is also a bigger update than just a cleanup. I'll update the issue to talk about that.

I do suggest running the following tests:

ERP_D_Ld5.f10_f10_mg37.I2000Clm50BgcCru.dercho_gnu.clm-fire_emis
SMS_D_Ld5.f10_f10_mg37.I2000Clm60BgcCrop.dercho_gnu.clm-fire_emis
SMS_D_Ld5.f10_f10_mg37.I2000Clm60Sp.dercho_gnu.clm-fire_emis

so it's exercised for both Bgc, Bgc-Crop and Sp so we should see that it at least doesn't blow up for any of the main spectrum of case types. The top one is the only one we run for aux_clm.

[samsrabin]

@ekluzek @slevis-lmwg I get the following error with those tests, which doesn't seem to have anything to do with the emissions file:

dec2122.hsn.de.hpc.ucar.edu 88:   ERROR: glacier_region_behavior not specified for ID            1
dec2122.hsn.de.hpc.ucar.edu 88:  You may need to extend the glacier_region_behavior namelist array,
dec2122.hsn.de.hpc.ucar.edu 88:  or you may be running with an old/incompatible surface dataset.
dec2122.hsn.de.hpc.ucar.edu 88:  ENDRUN:
dec2122.hsn.de.hpc.ucar.edu 88:  ERROR:  ERROR: glacier_region_behavior not specified for ID ERROR in /glade/u/home/samrabin/ctsm_emission-factors-20240829/src/main/glcBehaviorMod.F90 at line 413

[ekluzek]

@samsrabin thanks for doing those tests. It occurs to me that those tests should fail because we don't have ctsm5.3.0 f10 files in place. Sorry I didn't remember that.

But, it also occurs to me that we should do those tests in ctsm5.2.027 with just the fire-emissions dataset updated, so that the first one can be compared to the ctsm5.2.027 baseline. So can you repeat the tests doing that? Thanks so much.

[samsrabin]

@ekluzek Two of the tests work, but I get a failure in SMS_D_Ld5.f10_f10_mg37.I2000Clm60Sp.derecho_gnu.clm-fire_emis:

dec2249.hsn.de.hpc.ucar.edu 60:  htapes_fieldlist ERROR: M_LEAFC_TO_FIRE in fincl(           1 ) for history tape            1  not found
dec2249.hsn.de.hpc.ucar.edu 60:  ENDRUN:
dec2249.hsn.de.hpc.ucar.edu 60:  ERROR: ERROR in /glade/u/home/samrabin/ctsm_emission-factors-20240829/src/main/histFileMod.F90 at line 885

However, this is true with ctsm5.2.027 as well, probably because cnveg_carbonstate_inst%Init() isn't called when use_cn is false.

[ekluzek]

Ahhh, yes the fire_emis test mod, must assume that at least Bgc is on.

I think we might want to run with fire-emissions more often, so I'm going to flag this with next to discuss tomorrow.

[ekluzek]

Oh, and thinking about it -- fire emissions doesn't make sense without BGC. D'oh

Actually that means we should change build-namelist to not allow --fire_emis on without BGC.

[samsrabin]

Yeah, makes sense. For posterity, though, splitting the fire_emis testmod into fire_emis_sp and fire_emis_bgc—where only the latter includes C-related variables—makes all 3 tests work.