There are a number of datapoints in OneClick which are sourced from the ecoinvent database.
For example:
I am interested to better understand the GWP value associated with some of these inputs (like the one above) – how they were determined, and what the scope is. There’s no EPD in many cases though. Does anyone have any guidance?
I see this more as an ecoinvent database question than a Oneclick question, hence posting here rather than OneClick support.
A little funny that a metadata field is labeled “EPD program” as I believe that what we’re seeing there is a ‘unit process’ and not an EPD. I don’t know OneClick, but have used ecoinvent for LCA. If my description below doesn’t apply to OneClick for some reason, I’m happy to stand corrected!
A unit process is the smallest unit in a life cycle inventory (LCI, the process model behind a life cycle assessment) and will describe a process of material transformation. In a unit process, the inputs are raw and processed materials. The outputs are the ‘reference product’ (for which the process is typically named), emissions to air/water/etc., and any co-products. These inputs and outputs describe physical/energetic flows, so there’s no GWP directly attached to a unit process.
In the life cycle impact assessment (LCIA) phase of an LCA, all of the physical flows from the LCI are assigned to ‘midpoint categories’ (like GWP or ODP) using a set of characterization factors (like those provided by the EPA or IPCC, for instance).
tl;dr - An EPD describes the results of an entire LCA and reported as environmental impacts, a unit process is a teensy part of an life cycle inventory and reported as a set of physical/energetic flows. Ecoinvent is a collection of the latter.
Metadata for Ecoinvent processes are available for free online at ecoinvent.org. Create a guest account, then use their ecoQuery service to get dataset details.
Curious to hear your takeaways from the investigation and your thoughts on the data’s applicability. In my opinion, the use of Ecoinvent data in wbLCA is probably worthy of a bit more skepticism. It can be useful as background data, but we are often using it to model the foreground system. And the provenance of the data is sometimes hard to understand in both Gabi (behind Tally) and OneClick.
My thoughts on the data applicability: If the stated scope of the data matches what I’m assessing, I am reasonably comfortable with using it in wbLCA; but we do clearly state the limitations of wbLCA practice, including the eracity of the available data. It may turn out, in future, that the number was very incorrect. However, if I have no alternative option What else can I do?