Consensus decision on which extensions to include

[cmutel]

EXIOBASE has the following extensions:

• 39 resources, e.g. Blue water consumption, building stone, anthracite
• 5 land use types, e.g. arable land
• 66 emissions, e.g. NOx, Mirex, lead
• 66 emissions avoided by the use of manure
• 66 emissions from "unregistered waste", of which ~10 are non-zero
• 10 emissions from unregistered waste, drawn from same emissions list as above
• 8 "Other auxilliary account not elsewhere considered", e.g. "other non market feed", "use as feed of crop residues"

In addition, there are separate supply and use tables for 19 categories of waste/materials for the following:

• Total waste
• Packaging materials
• Materials in heterogeneous products
• Stock addition (only supply)
• "Supply of waste from materials accumulated previously in the society" (only supply)

Please provide your input on what is needed for a) a minimum viable product, and b) the final input of EXIOBASE into BONSAI. Each additional dataset represents a nontrivial burden to a group already stretched thin.

[cmutel]

In my opinion, the most important are primary emission, land use, and resources. I would be quite curious to get some input from @Stefano-MRC on how the was SUTs could be used in our model to help produce the IO table, unless the answer is boring in that they are just separate activities and flow objects.

[massimopizzol]

a) the first three listed (resources, land use types, emissions)
b) ideally all of them but pragmatically only those for which somebody makes a specific request

[Stefano-MRC]

Hi Chris, sorry but I am not sure to have understood what you mean. Are you referring to the method to convert the SUTs into IOT? If yes, think that the choice of the approach to move from SUTs to IOT also depends on how we solve the equations of the IOT. If we opt for the matrix inversion, then the IOT must be square. Therefore I don't know other options than aggregating/disaggregating. Instead if we opt, for example, for linear programming we can keep the SUTs format (then we just need to define the objective function) However, in Barcelona we started bulding MOJO with the aim to have squared IOT with the Stone's method as product modeling.

[mfastudillo]

I think many of these extensions come from mass balances, (e.g. unregistered waste come from discrepancies of statistics of waste treatment and what you would expect from material composition of products). it would be nice to have them but I don't see it as a must. I would go for the first three categories for a first version

[MicDr]

1. The choice of the extensions depends on the objective and resources, e.g. we could have only GHG emissions for a proof of concept/minimum viable product i.e. for carbon foot printing. Currently this is what EXIOBASE is mostly used for as far as I know. It is also disputable whether other extensions are sufficiently substantiated by the data quality.
2. For the "final Deliverable", as EXIOBASE has the extensions listed above, why would we want to leave some of those out?
3. Obviously, based of the relevance for specific impact categories, some extension are worth to keep while others could be ignored. E.g. in the often-mentioned "ideal world", the 5 land use types could be further disaggregated, because five land use types are barely enough to distinguish between different forest land types (primary, secondary, planted, broadlealf, deciduous, tropical, boreal and combination of those) with very different carbon stock. Moreover, talking about land use types, I know that in EXIOBASE they are a distinct category, but IMHO they are actually just another "resource"

[mfastudillo]

do we have any uri's for extensions ? I see it as an option in the arborist readme (here) but I could not find them. I'd be interested to know also your thoughts about "compartments". Characterization factors are dependent on these and I don't know how this is represented in the ontology.