[Action/SCI] Metrics for Emissions factor (I) for Falco

[nikimanoledaki]

SCI = (E * I) + M per R

Carbon emission intensity per geographical region.

Acceptance Criteria

• Identify the electricity grid of the Equinix geo we are using
• Replace carbon coefficient in SCI dashboard with the Equinix-specific one
• Add Emissions (I) to docs

[raymundovr]

I'd love to help with this one :)

[rossf7]

Hi @nikimanoledaki @raymundovr,
for the dynamic emissions factor I'd like to suggest using the grid-intensity-go tool that @mrchrisadams and I maintain. https://github.com/thegreenwebfoundation/grid-intensity-go

It has a prometheus exporter that integrates with the Electricity Maps API free tier. It provides this carbon intensity metric we could use in the dashboard.

curl -s http://localhost:8000/metrics | grep grid_intensity_carbon_average
# HELP grid_intensity_carbon_average Average carbon intensity for the electricity grid in this location.
# TYPE grid_intensity_carbon_average gauge
grid_intensity_carbon_average{location="FR",node="kind-control-plane",provider="ElectricityMap",region="",units="gCO2e per kWh"} 39

We have a little bit of work to do as we need to push the container image and helm chart to a registry. We should be able to add that pretty quickly if we want to use grid-intensity-go for this. WDYT?

[raymundovr]

Thank you @rossf7 !
I was thinking a bit futher on the argument between "live" intensity with electricy maps and the static measurement. Could you please remind me which source / measurement was the one that you mentioned during the last call to use as static value and allow comparisons?

[rossf7]

@raymundovr My thinking was to use the IEA average carbon intensity figure for France. As I thought this was a common approach.

However looking at their site it looks more complex :( They have both paid and free datasets and I can't see a way to access the data. https://www.iea.org/data-and-statistics/data-product/emissions-factors-2022

In grid-intensity-go we embed a dataset from Ember Climate from 2021 as they publish data with a CC-BY-SA 4.0 license. https://ember-climate.org/ we could follow a similar approach.

@mrchrisadams it would be great to get your thoughts on this.

[mrchrisadams]

hi @rossf7! For annual data, the easiest option I can think of might be to use the helpfully formatted data we have in CO2.js.

I've linked to the json file below, with France specifically:

https://github.com/thegreenwebfoundation/co2.js/blob/main/data/output/average-intensities.json#L422-L427

These are updated every month on a cronjob from Ember, to reflect any changes to their methodology, and represent the most up-to-date, easy to use open data that we can find.

The changes are also checked by Fershad Irani (@fershad) and myself to catch surprises in the numbers, and query them upstream. You can see the most recent example below:

thegreenwebfoundation/co2.js#190

(As an aside, if there is a simple way to trigger an update to grid-intensity-go, when PRs like are merged in, I would be all ears - maybe a build step in the go library when cutting new releases, or something smarter than that)

[rossf7]

Thank you @mrchrisadams using the value you have in CO2.js makes sense to me but I will defer to @nikimanoledaki and @raymundovr as they are driving this one!

I'm working on the embodied emissions factor for #40 using Boavizta's dataset.

[raymundovr]

Thank you @mrchrisadams and @rossf7 for the great resources, really appreciated!
Linking the script that generates the json for reference: https://github.com/thegreenwebfoundation/co2.js/blob/main/data/functions/generate_average_co2.js

It seems reasonable to me to set this value for France in the board, at least to start with. What do you think @nikimanoledaki ?

[nikimanoledaki]

Thank you @raymundovr @rossf7 @mrchrisadams :)
(I) is in the dashboard + docs have been added. Marking as completed! 🎉