Open BenPortner opened 4 years ago
A few thoughts, as someone with some experience in this area:
location
entries to only be countries.activities
and flow_objects
, for geographical relations (e.g. inside, overlaps, contains). This is also the approach taken by LCA libraries. We would then have a separate repo with a PostGIS database which would store the actual GIS info for all locations used in Bonsai. Such separation of concerns is very much in the philosophy of Bonsai.
@cmutel
There is nothing in the schema limiting location entries to only be countries.
That's true. But the given entries don't give any clue how to implement non-country locations.
"RoW" locations were a mistake. I am strongly opposed to including this concept in Bonsai. We never have real data for "RoW" in any case, just global or country-specific data.
I agree that there should be no RoW in bonsai. But there is RoW in ecoinvent. If we want to include ecoinvent in Bonsai we will have to find a way to translate from ecoinvent to bonsai. And that means we will have entries like "world minus region x, y, z"
It would be much simpler to store a correspondence table, as we do for activities and flow_objects, for geographical relations (e.g. inside, overlaps, contains).
First time I hear about correspondence tables in this context. Can you provide an example?
Right now, entries in the location table are countries according to https://www.geonames.org/countries/. This won't do for the future. Here are two problematic examples:
1) ecoinvent RoW activities are not countries. If anything they are logical expressions of the scheme "world minus region x, y, z"
2) regional LCA can demand locations on sub-country or cross-country scale
How to deal with this? I think we need GIS. Maybe @cmutel has some advice on this based on bw_regional?