Idea for handling Scope 3 data and passenger km data for Autos

[MichaelTiemannOSC]

As explained in this 12MB PDF file (https://deepblue.lib.umich.edu/bitstream/handle/2027.42/148647/Scope?sequence=1), in the automotive sector, S1+S2 emissions are dwarfed by S3 Emissions. The ratio is over 10:1 for Ford Motor Company, and 25:1 according to the 2022 OECM benchmark. S3 emissions come in many shapes and sizes, but for Ford, 75% of S3 emissions are tailpipe emissions (and fit within "Use of Sold Products").

While Auto companies do report S1 and S2 emissions related largely to manufacturing, most report only emissions intensity (EI) when it comes to tailpipe emissions--which makes computing absolute S3 numbers difficult.

For Autos we could fill in S3 data as follows: If we have "Use of Sold Products S3", use that as tailpipe emissions, which, when divided by EI gives passenger km "produced". If we don't have an S3 number, use OECM benchmark average 25:1 multiplier to convert S1+S2 to S3, then (based on to Ford paper) take 75% of total S3 as tailpipe emissions. Divide derived tailpipe emissions by EI to get passenger km as an output.

Thoughts on 25:1 ratio (could be as high as 100:1 looking at BMW, Toyota, and other reports)?
Thoughts on 70%-85% as fraction of total S3 to derive tailpipe emissions (again, looking at actual reports)?

We can enhance the tool so that for Autos we work with S1+S2+S3 (or S3 alone). I will take a crack at that.

@LeylaJavadova @ImkeHorten @joriscram @kmarinushkin

[kmarinushkin]

Thoughts on 25:1 ratio (could be as high as 100:1 looking at BMW, Toyota, and other reports)?

We could have this ratio as a column in the input table. It will allow us to set specific values per company, and adjust it on demand w/o a need to touch the code

[MichaelTiemannOSC]

Another consideration for this task is the fact that fleet EI on a per-country basis can vary by more than a factor of two. In the case of Toyota, their EI average is 96 g CO2/(passenger km) whereas their EI average for South Africa is 194 g CO2/(passenger km). Fortunately, it doesn't hurt older versions of the tool to add columns of data to input tables--the tool ignores columns it doesn't understand. Consequently we can add a Country column to the targets information, which allows us to set country-specific base year or target-specific data.

We already have a country column for the fundamental data, and if reported we can provide country-specific data (including country-specific LEIs) for fundamental data, which could be used to weight the target data.

Country-specific LEIs that roll up to an ultimate parent could be used for financial information that is not reported or analyzed on a country-specific level, such as cash. But many companies do report country-specific production and emissions data, and the tool could show country-level contributions in stacked charts (bars) or pie charts (circles).

[MichaelTiemannOSC]

@LeylaJavadova what priority would you give to country-specific metric and target analysis?

[kmarinushkin]

In this context, by "country" we understand "region", right?

Also, do different companies report different regions? Do we want to make country columns configurable per company? Columns could be " France", "Italy", ... Or could be " Region1", "Region2", and each company provides it's own definition of Region1 and Region2

[MichaelTiemannOSC]

So it turns out that the very first datapoint: North American financial results vs. separate US and Canadian EI numbers suggests that there's a whole upstream data problem we have to solve before we can guide batter on input for the tool.

[kmarinushkin]

@MichaelTiemannOSC i think we discussed today, that i will approach the "S3 split and recombine" task.
I traced it down to this "issue". Should we assign it to me?

[MichaelTiemannOSC]

I've assigned to you. Please build a test_s3_allocation unit test that we can add to so we can check the case where S3 data is known good and correct as far as the sector is concerned, the case where S3 comes in aggregated and we need to disaggregate, keeping what belongs to the sector and discarding the rest.

There is a concept in this world of emissions analysis of which emissions are "covered" or not covered by a given scope definition or methodology. You may want to keep in mind the idea of having a bucket where uncovered emissions go, just so that the math always adds up. Or, it may be OK to simply lose uncovered emissions. We'll see!