[jeffseif]

Background

The dataset provides a global ~5km resolution raster for the cultivation area of 173 crops. It is modeled from 27 other crop maps, including MAPSPAM -- not a remote sensing model. It will inform the area-share allocation underlying JD proxy (in the next PR).

This is honestly a pretty straightforward ingestion.

Notes

1. It arrives as a zip file of 173 global netcdf's and we massage it into a single global 615 MiB COG
2. The new 173 enums will form the taxonomy (modulo livestock) of the JD proxy emissions factors
3. A couple of the crops have typos in them, but we leave them as-is in order to match the paths in the published zip file
4. I ran the ingest and the data looks pretty good. There are clearly artifacts due to it being a statistical model -- like hotspots at the intersection of a jurisdictional boundary and an ecosystem.

Oil palm in oceania

CROPGRIDS

MAPSPAM

Coffee in central america

CROPGRIDS (arabica and robusta)

MAPSPAM (arabica)

Soy in south america

CROPGRIDS

MAPSPAM

Testing

Claude confirmed that the sums over national jurisdictions matches those reported by FAOSTAT