Interpretation of Sankey diagrams

[dbantje]

Hi all!

I'm having trouble understanding the Sankey diagram, in particular the impact fractions shown for edges in the diagram.

In the subgraph above, how can I make sense of the fact that the top edge has a higher associated impact than the downstream one? I would have expected that downstream edges must always have a larger associated impact.

Thanks!

[marc-vdm]

Hi, The Sankey works by calculating an LCA for each visible activity (node) in the graph, scaled to how much of that activity is used by the previous one. This may sound a bit abstract, but I hope the simple example on the Sankey wiki page helps clarify.

As systems can become very large (e.g. ecoinvent has >20000 activities), the graph traversal is cut-off at some amount of steps (250 by default), or when the impact becomes lower than some percentage of the total score (5% by default), whichever comes first.

One explanation for what you are seeing could be that the market for scrap copper also has a negative impact for your impact category that is not shown because one of the cutoff. You could calculate the results for the market for scrap copper activity and explore why your results are what they are.

Hope this helps!