WikiJournal Preprints/Life Cycle Assessment methodology/Review: An open source dataset and Ontology for product footprinting (Awarded best poster - ESWC 2019)

A review of An open source dataset and Ontology for product footprinting (Awarded best poster - ESWC 2019)

Rudy Patard

(RQ : Just a quick paste for the time being... I'm in a rush)

As a "post-review" I'd say :

Lacking challenges :
"Current sustainability information on products – “product footprints” – is:"
 * Inconsistent (at least on subjectivity integration and according to current standards)
 * "potentially" biased due to conflicts of interests
 * methodo-logically wrong (posing axiological order as "observation and calculation")
 * although incomplete, already overloading human cognitive capacities

Need :

Ontological part for judgment integration (axiological structure) (NB it's 4 more dots in the list, but isn't "incomplete = Lacking in detail " a lead to save a line ?)

Determining flow
(As I see it, the "reference flow" from other vocabulary).

Stating in the data base "the flow that matters most", the "determinant one" leads to incorporating value judgment inside a priori "objective data" (so mixing observation and opinion). An idea that seems in the lineage of "ecoinvent 2 VS 3" controversies and that still seems to avoid serious reviewing. As I see it, a loop between an axiological structure and the present ontology free of the "primary" distinction between flows should enable an adaptable multi-functionality treatment AND other requested judgment. (As a personal instance, I'm wondering the share of "Saw" vs "Axe" I'll use for my woodcutting as I intend to use a dry toilet - so a litter. Is burning wood or sawdust the referential flow, is there a need for any "leader" flow in assessing the impacts of my activity ?). Instances, the use of a "human being" inside the instance graph (as mentioned to the left, "human activities" ; "person or thing"), may put the ruler on this particular point of determining flow, human being have the particularly nasty characteristic of being widely applied, versatile and multi-function... Is the metal worker in action with the furnace A) A furnace operator B) A father C) A husband D) A friend E) A food consumer F) A house inhabitant...

"Determining flow" = "Reference flow" in ecoinvent language: No value judgement is intended here. The definition "Product output of an activity for which a change in demand will affect the production volume of the activity" implies an observation of a change in production volume as a response to a change in demand. The intricacies of this identification are described in detail here: https://consequential-lca.org/clca/determining-or-dependent-co-products/.

I have no objection that "demand" is a determinant aspect of a process. Though I deny it is the single one and thus that no judgment are required in "modeling" and "assessing" impacts. Particularly, this seems to stress out the loss in goal of an assessment, i.e. a decision. I'll use the wheat example from "type 1 situation" in. "wheat grain that can only come from a wheat field. All the other co-products are then dependent co-products (by-products)." Wheat is a staple food, and also an input in starch industries. There are many products labelled as "staple food" with ranging properties. And there is also a range "inside" wheat Wheat.

Wheat used to come with long and strong straw. We modified that, "playing" (by selection, and-or modification) with the share between co- and by- products. Thus, the inputs of wheat production may range according to 'a range of properties' from "size of the straw" to rate of gluten etc.

So you may answer "Yeah, no problem, it's the "technical-spatial uniqueness issue" related to "data availability issue", but anyway they are always led by market demand and prices gives us scaling factors".

Sadly here again, price is no attribute of the object, it's an attribute of the "transaction". So it varies according to ranging bargaining power, volumes, market-shares, lead-times, speculations : the date to the nanosecond, currencies and there ranging "relative positions to goods". So "what price" for the rice sir ?

Is the impact of the wheat inside the cargo really changing because the "Börsen" said so (Börsen, stock exchange) ?

How many question require an answer and where lies judgments, 'value'-judgments ?

"As far as I can see this does not require any value judgements, just observation. I believe the description can be converted to computer code, but I have not seen anyone doing that yet."

Even though it requires many judgements (and I agree that "demands" request addressing inside these ones), I consider them "manageable" according to the conditions "judgments matrices" are inputs of the assessment system. And as previously stated, multiple judgments are needed while modelling and assessing. Managing consistency leads me to advocating for a specific step and compartment specific to 'opinions'.

Judgement in general
"What is the amount of flow x produced as output during the time period y due to the production of flow z?

What is the amount of CO 2 output in 2018 to produce steel in Germany?

- Answer: 254,32 Mt CO 2 to produce 787,65 Mt steel" These questions imply judgments, so their current presentation as a simple "database request" is probably the worst part I could point out in this poster. The "Vision" of a "calculation" is deeply in opposition with the concept of "assessment" of eVALUation.
 * imply "attribute to product" function judgment (i.e. defining any functional unit)
 * imply cut-off judgement
 * imply time horizon judgement
 * imply uncertainty judgment (for use on impact assessment, Yes I've read the "uncertainty next step" but I suppose it's pointing toward incorporating data distributions and sparql adapted requests, for instance, a box plot (boite à moustache), here I mean selecting environmental models.
 * imply multi-functionality treatment judgements

"While the choice of system model is an obvious value choice, the competency question in case already states this choice implicitly as consequential in the use of the term 'due to'. Once this question/choice is posed/made, no further value judgements should be necessary, since it follows logically from the question that the system boundaries must be global, infinite, and not applying by-product cut-offs. Of course, the database can be used to answer many other questions that imply different system boundaries. In this respect, the specific competency questions are only exemplary."

I believe the 'boundary' question to be 'technological' and 'spatial' as I doubt you mean Bo "the system boundaries must be global, infinite," including time dimension. I esitated while writting "cut-off" but I'll specifically treat this one down further. As I'm not sure I was clear. I'll try to "reformulate" the sparql interrogation exposed as "competency question" :

"What is the amount of CO 2 output in 2018 to produce steel in Germany?"

What does it mean ? How many questions unfold ? And are there judgments ?

" What is the amount (I doubt there is ambiguity it's a mass interrogation)
 * of CO2 (? GHG eq. or inventory effective carbone dioxyde ? overwise if CO2 eq. was intended + "according to IPCC2007 GWP100 and respective factors and value judgments ") imply time horizon judgement (GWP100), why not 500y ; imply uncertainty judgment (meaning IPCC2007 gives us the "best knowledge" between models "we know for sure - we aren't sure but it's damned important we have to take it into account")
 * output in 2018
 * ? is the "time horizon from 1st January 2018 to 31 December 2018, meaning all activities, all processes "usually" allocated to steel production, (of course I doubt it as you mention in your reply "no cut-off infinite boundaries", it's just explanatory development for other readers)
 * or is the time horizon the usual 100years "egalitarian perspective" applied to 2018 steel production, and with what control on "past emissions" and "incoming due to rest of life cycle product emission with prospective scenario 'PS-ref?', thus judgment on "what will append" "imply time horizon judgement" "
 * to produce steel * in Germany?
 * why I stated "imply "attribute to product" function judgment (i.e. defining any functional unit)"
 * if I take the primary route (with coke) what prospective scenarii are taken into account for allocating products in use for other than steel production ? How are handled emissions that may in future, with ranging demand be used for 'cast iron' instead of 'steel' or that will be used in steels though it is not necessary the single output for them (Cu, Ni, Cr ...)
 * Meaning the "steel products sold from a factory in Germany", or a specific-limited 'semi-product' list (bloom, brame, billette, coils... ? Are "rebars" considered semi-products of concrete buildings?
 * Does it include, or exclude, purely "commercial" activities on steel as considering 'not producing them' though acting on "demand", a surface treatment, a forge operation, simple machining, final product such as cars... as 'part of the steel net of production - value - chain, inside German borders ?
 * as the supply chain we (Bo and I) want infinite (with good reason, it's why LCA was developed to avoid impacts 'rebound', we'll cross all human activities thus crossing many times multiple - in-commensurable - outputs in or out of markets, thus it implies multi-functionality treatment judgements (as such as the one explain in this graph.

So one simple competency question in my opinion would look like :

What is, with a boxplot result so uncertainty appears:

the amount (mass) of CO2 emitted from

 ferrous metallurgy factories

 inside German borders

 during the 2018 year

 to produce steel semi-products : brame, bloom, coils as specified in -name of the major standard steel designation-;

 allocating C02 (for non-steel by- or co-products) according to products prices, using median prices (sales in 2018) in euros , with prospective scenario that unsold-yet 2018 steel production will be sold to 2016-2018 linear regression prices and that 'technical factors' will remain as 2018 ones 

Here, I believe (maybe I missed some choices), all would be consistently in the question and no further judgment would be required, to reply to the question indeed. But the result is not even LCI, 'inventory level'. One emission, on a specified list of activities with a specified limited boundary (I'd one figure of a steel-brame, bloom, coils-Germany U- ecoinvent inventory data set). What to do with it ?

"cut-off" and consequential LCA
There is a particular point about "cut-off" and consequential LCA.

While working on a straw-bale house, I discussed the "Determining or dependent co‑products?" theme with a consultant that append to work on 'building' studies. I believe it can contribute to clarify methodological debate and practical solution for 'practitioners', so I'll relate the "story".

I start to explain that in consequential LCA, we address decision that affect the "supply chain tissue" the production capacities in place, i.e. productive infrastructures.

Then I point to the straw-bale-house we're working on. And I ask: So a few set of products (buildings) inside the time horizon of regular impact methods (ex: IPCC2007 PRG100), influenced the demand for a shape of bale, but also the "thermic" machines inside, insulation sector, possibly 'fire-protection' products and services thanks to fire resistance, health-care while reducing/altering "houses ambiances" and building sector work ...
 * "How long will it stand ?"
 * "When was the first "modern" straw-bale-house erected ?" "Between 1896 and 1945, an estimated 70 straw-bale buildings, including houses, farm buildings, churches, schools, offices, and grocery stores had been built in the Sandhills.[8]Straw-bale_construction"
 * "How many was there (obviously one or few), so not big a change in demand. (?)"
 * "Has the 'straw-bales' changed since, and their related machines (from paved - "square" to cylinder - "round" here in France from what I see in the country-side) ?"
 * Does demand in "square ones" rise and how it affect "agriculture machines" ? "Since the 1990s straw-bale construction has been substantially revived, particularly in North America, Europe, and Australia"

I doubt there are any models for Consequential-LCA that can absorb in modelling such schema. Would the first builders be credited for the reduction in concrete and rebars in a change that could appear in less than a time-horizon of LCIAM ? If I had been the consultant doing LCA on the first building, I doubt I'd have anything to deploy consequential LCA. I'd have cut it off out of an opinion, a prejudice (particularly while taking note that the cows were eating the house).

Cut-off meaning cutting from the model what doesn't influence conclusions seems highly doubtful. And in the same time, required prospective seems to avoid our grasp. So we 'cut' consciously or not, possible scenarii. While working in industry "bei eine Pumpen Firma" that depended on "sanitation market", I proposed while working on their "carbon bilanz" that they develop in "dry sanitation" (possibly using no-pump systems). It was ten years ago and to the date, I'm still an exception using dry-toilet and their market hasn't disappear (yet). But still, I was convinced during this study and I've seen others in France doing so. If by ten years from now 'their' market drop seriously, what will result in "sanitation" - "building" - "water" - "health" sectors ? If a couple of notorious actors go vegan and attract their "fan-zone" toward veganism.

To my opinion, "sensitivity" analysis in consequential LCA can very be "exploratory" and be a selection of scenarii cut-offs. And in reverse, it would be very interesting to see "what change in "value matrices" affects the supply chains and production capacities and 'respectively' (meaning for instance, how the SUV segment affect the 'positive' retro-action loop towards 'SUV's' by advertisement and sales, i.e. presence in markets, screens and streets).

BR

--RP87 (discuss • contribs) 13:09, 28 November 2019 (UTC)

Question and perspective :
A step I consider an important part of the "EHO" design I stressed out. For this part of my work, this glimpse should draw light on the matter (https://framagit.org/RP87/these/blob/master/ILCD_LCA_LCIAM_extended_principles.pdf).
 * Are the authors searching into licenses attributes for their database (as "open access" / "open source" does not state user rights apart from "free to read").
 * I see in the converging definition of "Activity" : "... both human activities and environmental mechanisms..." as a significant evolution.

Eager to read related articles.

Best Regards

and continuations to all.

Rudy

(Former, maybe future again, PhD candidate and still a LCA methodologist keeping watch on "sustainability assessment".) https://cv.archives-ouvertes.fr/rudy-patard