Allow CLM by PFT history outputs with FATES nocomp

[JessicaNeedham]

I'm running CLM-FATES and would like to output some CLM variables by vegtype using hist_dov2xy = .false. and hist_type1d_pertape = 'PFTS'. This is incompatible with FATES at the moment, because itypes are not set if FATES is on. This makes sense in full FATES but could actually work in nocomp, where each PFT is on its own patch.

Discussing with @mvdebolskiy the possibility of adding an 'if nocomp' statement where itype gets set in clmfates_interfaceMod.

[JessicaNeedham]

Noting this was implemented fairly recently: #2935

@rgknox - can you remind me why we can't have itype set even in nocomp mode? Why can't we associate itype with the FATES PFT via nocomp_pft_label_pa?

[JessicaNeedham]

Chatted more with @ckoven and @rgknox and it sounds like this is possible but might not be a quick fix I can get in in time for WIEMIP simulations. Would be nice down the line but is not high priority.

[wwieder]

At the CLM-SE meeting @rgknox noted this seems like a significant infrastructure change that's needed to enable this feature since the HLM and FATES pft mapping isn't the same (and itype refers to the HLM mapping).

@adrifoster asked why PFT output can't be written on the FATES side already?
@JessicaNeedham, what are the specific CLM variables you're interested in writing at the PFT level?

[mvdebolskiy]

@wwieder

why PFT output can't be written on the FATES side already?

Because some of the variables that are associated with clm patch arrays are not passed into fates. They are are purely clm-side variables, but are calculated on patches.

[wwieder]

OK, naive question here, but can you provide an example of this?

[JessicaNeedham]

This came up in the context of WIEMIP where some of the variable requests are by vegtype but are not FATES PFT history variables. My understanding, from discussion with @ckoven and @rgknox are that some are not possible because they are column level. But there are others like FSR, FSDS, FSH, FIRA which are by patch. These already get passed to FATES for land use level history variables so those might not be so complicated to process on the FATES side. For WIEMIP the others are BVOCs by PFT (MEG_isoprene, MEG_methanol, MEG_acetone), heterotrophic respiration, NBP (not sure if that is patch level?) and veg type level carbon in soil.