Skip to content

Conversation

seanmcilroy29
Copy link
Contributor

Proposed Spec update

| geolocation | decimal degrees | Static | Yes |
| power-usage-effectiveness | ratio | Annual | Yes |
| grid-carbon-intensity | gCO₂eq/kWh | Annual | Yes |
| provider-cfe-annual | proportion | Annual | Yes |

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

hi @seanmcilroy29 - I appreciate I missed this on the last call, that new levels of conformance had been introduced to this - both A1. (mandatory) and A2. (optional).

Was this what came out of the conversation in #109 when speaking to folks at the ISO? I'm asking in the context of people who might build tooling for future validation against different levels of conformancy.


A user of the cloud region metadata can specify which cloud provider and region they use to run a workload and get all the relevant metadata about that region. Cloud region metadata is published annually and lags by 6 months to 18 months, so the year shall be specified, or the latest data should be used.

The annual average location-based marginal grid-carbon-intensity value required for SCI-o is provided when available. Because of differences between cloud providers, data providers and reporting methodologies, there are several possible carbon models, and data may not be available. Attempting to consume a not-available or blank metric shall cause any calculations to fail.
Copy link

@mrchrisadams mrchrisadams Jun 19, 2025

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@seanmcilroy29 - can I check I understand here? I'm having a bit of trouble parsing annual average location-based marginal grid-carbon-intensity .

I understand the following term:

  1. annual average location-based grid-carbon-intensity - for this, I understand it to be the mean of all 8760 hourly carbon intensity figures over the year. And for hourly carbon intensity, I understand to be a weighted average of the different forms of generation used (so if there was were two units of coal, and one of gas you would use one unit for gas, plus two units for coal).

This is available as open data at a country level from various government bodies, but also groups like Ember, amnd Our World In Data, and at higher geographic and temporal resolution from groups like Electricity Maps.

  1. annual average location-based marginal grid-carbon-intensity - for this, I am less clear about what this refers to, and who the providers are. I know Watttime is a provider of marginal carbon intensity figures, but as I understand it the mechanism used to work out marginal intensity differs from provider to provider, so there is much less fungibility.

In the SCI, there is a term called Region specific carbon intensity, which has been used to avoid needing to be prescriptive about which one to use for SCI calculation.

If we want to be consistent with the SCI, it seems like annual average region specific carbon intensity would cover both cases without being too prescriptive. Is there a reason we can't use that? We have columns for both kinds of grid carbon intensity in the schema.

The annual average location-based marginal grid-carbon-intensity

This was referenced Oct 3, 2025
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants