New definitions of emissions in AFOLU sector for comparison with NGHGIs

[tgasser]

This is the first of two PRs to initiate a change in the AFOLU-related reporting, following a community-wide consultation carried out within the NewPathways HEU project.

It introduces CO2-specific variables, thereby splitting CO2 and non-CO2 definitions of AFOLU emissions. It also attempts to clean up land-related subcategories, notably fires.

It does not change anything regarding gross emissions variables, as this should be part of a broader effort that includes all sectors.

There's an obvious loss of backward compatibility, although only through renaming as all previously existing subcategories should still be available.

Discussed notably with @flohump, @strefler, but final suggestions my own.

Resulting definitions file attached: definitions_luc_em_def.xlsx.

---

List of changes:

1. Removed Agricultural Waste Burning and Agriculture subcategories for CO2.
   Rationale: these should be zero by definition for CO2 because climate-neutral, removing prevents adding possibly non-zero values to anthropogenic emissions.

2. Renamed Land subcategory to LULUCF, and introduced sub-level categories following IPCC 2006 guidelines' land types) + Deforestation.
   Rationale: comparison with Grassi et al.'s NGHGI database for CO2 requires at least Forest Land and Deforestation; HWPs already present; Wetlands important for non-CO2; other subcategories introduced for completeness but can always be ignored.

3. Fires were moved two levels up in the tree to Emissions|{species}|Fires and removed for CO2.
   Rationale: only explicit fire management should be included in AFOLU; since it is very unclear whether these are anthropogenic or natural fires we want to force the user to think about it; CO2 removed altogether because either climate-neutral (grassland), already in LULUCF (deforestation fires), or supposed to be in the natural response (other forest or peat fires); non-CO2 may still be necessary for ScenarioMIP (@jkikstra?).

4. Introduced CO2|AFOLU [NGHGI]|Natural Flux Over Managed Land as the correction term that allows translation from model/scientific definition to NGHGIs', optionally split over the same land categories as the LULUCF term.
   Rationale: since this represents only natural processes (response to CO2, climate change, etc.), only a few IAMs can estimate it currently, but it is a necessary component of the total CO2|AFOLU [NGHGI] flux.

5. Introduced CO2|AFOLU [NGHGI] and Kyoto Gases|AFOLU [NGHGI] aggregates that are CO2|AFOLU and Kyoto Gases|AFOLU + CO2|AFOLU [NGHGI]|Natural Flux Over Managed Land.
   Rationale: for completeness and partially solves issue 374.

File structure changes:

• split out the AFOLU sector into a new file (from emissions to emissions-afolu)
• created a level-2-species-no-CO2 tag file

Other change:

• corrected level 0 species' units (because used in a previous version, but not used in the end -- thought I'd leave it here anyway)

[flohump]

@tgasser — this PR carries Gross Removals|CO2|AFOLU|Intentional/Unintentional
over unchanged, but I think those two sub-categories should be dropped rather than
relocated: they've been reported only patchily across ScenarioMIP and the
distinction doesn't hold up (reasoning and a State of CDR reference in #224).
Ideally we'd handle this as part of this PR, to avoid merge conflicts on the same
AFOLU block — but I'm happy to open a separate PR if you prefer. Also tagging @strefler and @danielhuppmann here.

[strefler]

Agree on removing the intentional / unintentional categories. I can't judge though in which PR this would be placed best.