More specific cropland definition

[flohump]

Crop yields can be based on physical or harvested area. This PR refines the definition of the existing Yield|Cropland|{Crop Types} variables (physical) and adds new variables for yield based on harvested area Yield|Cropland Harvested|{Crop Types}.

[jonathandoelman]

Hi Florian, all,

Thanks for creating the issue. I am not sure if it is useful to have both variants of yield represented in the common definitions template, it may actually be quite confusing. I would argue that yield on the basis of harvested area is the most common and sensible indicator. On the other hand, we also don't report harvested area per crop type, but physical area as part of the Land Cover category, even though this is a bit of a constructed variable as well. In AGMIP the focus is also for area information per crop type on harvested area.

I dont want to create a lot of changes at this late stage, but it would be good to hear what others think. For sure we need to be clear about the definition. @IAMconsortium/common-definitions-land please let us know what you think.

Jonatahn

[flohump]

I think clarity is essential here. I suggested including both definitions because it helps make transparent that different interpretations of "yield" exist and are in use across modeling teams. Including both doesn't necessarily mean that all teams need to report both. Rather it can serve to highlight the distinction and encourage more precise usage. Re-defining Yield|Cropland|{Crop Types} as yield based on physical area doen't necessarily mean making it a default. But if most teams would indeed report yields based on harvested as default, we could of course do it the other way around, i.e. re-defining Yield|Cropland|{Crop Types} as yield based on harvested area, and adding a new variable for physical area. @IAMconsortium/common-definitions-land

[realxinzhao]

Echoing what's been discussed above, @flohump and @jonathandoelman, agreed that adding both is useful.

But also adding a quick note that FAO yield is usually in wet tonnes per unit of harvested area, so we won't have a direct comparison anyway. Also, if models assume constant harvest frequency and water content over time (implied in GCAM), the growth rate would be the same for an individual crop.

[danielhuppmann]

For the record, it may be preferable to keep the variable as is and add a method-explainer in square brackets, e.g.

Yield|Cropland|{Crop Types} [Dry Matter per Harvested Land Area]

But I can also go ahead and merge for now and revisit later (which probably won’t happen)

[flohump]

Thanks for the suggestion. Indeed that seems the better solution. I will adjust the PR accordingly.