A request for CO2 fossil fuel emissions global totals

[ChrisJones-MOHC]

Dataset source ID(s)

CO2 fossil fuel emissions

This might exist somewhere, but I can’t spot it, and not heard mention of it – it would be really useful as a QC measure for model groups to be able to check their processing of the fossil fuel CO2 emissions to ensure the correct totals.

The single top priority is that models get the right global annual total. CO2 is well mixed and long lived, so the exact spatial pattern and seasonal cycle are not so crucial. But we have to have the right total amount emitted.

Given the various splits into sectors and vertical levels, there is a risk some sources will get missed.

So a timeseries, of annual global total from the CMIP7 forcings would be really useful. Then model groups can test their input files for correct inclusion of all components.

[durack1]

This is a really good suggestion @ChrisJones-MOHC.

We have @znichollscr who's producing the CR-CMIP-1-0-0 concentrations dataset, and @ssmithClimate producing the CEDS-CMIP-2025-* emissions.

It would be an excellent sanity check to have these emissions as a globally aggregated cumulative, and then we could attempt to eyeball (or quantitatively) compare these against the concentration evolution too..

@ssmithClimate @rhoesly do you guys have such integrated totals in the CEDS repo somewhere?

[vnaik60]

This https://github.com/JGCRI/CEDS/tree/master/documentation has plots including CO2 emissions timeseries.

[ChrisJones-MOHC]

great - many thanks. is it poissible to publish the data timeseries from those plots?

[vnaik60]

Sure. Great question. Over to @ssmithClimate @rhoesly for an answer.

[ssmithClimate]

Yes, we do publish that. We publish checksum files that are global totals directly from our netCDF generation routines. Those are available on zenodo.

We should, indeed, put a link to those in the documentation.

We also published time series at various sectoral and spatial resolutions (and do internal checks that are gridded files added up to the same totals within tolerance).

[durack1]

@ssmithClimate if these global totals are already generated in netcdf format, and follow a similar/same metadata, and file naming format, we could publish these alongside the existing published netcdf data and simplify access for @ChrisJones-MOHC and like others who are chasing the same datasets

[ssmithClimate]

Checksum files for the 0.5° gridded data are now on zenodo.

https://zenodo.org/records/17693020

@durack1 - good idea. We don't have these in netCDF, but can try to also do that in the future.

[ChrisJones-MOHC]

that's great - thanks @ssmithClimate . If i could push my luck, as this is still very cpomplex with lots of layers and sectors, the QC request really is a single timeseries of the total CO2 emissions. That's important too to make sure groups haven't missed a sector or set of files. Is that possible please?

[znichollscr]

@ChrisJones-MOHC if you just want the sums quickly, then as a very short-term solution we had to do this for the harmonisation exercise (https://github.com/iiasa/emissions_harmonization_historical), file here: global-workflow-history_202511040855_202511040855_202511040855_202511040855_0002_0002.csv. If you just want CO2 fossil, look at the Emissions|CO2|Energy and Industrial Processes variable. (Note: not directly checked with the CEDS team, but we have checked ourselves and if this is wrong, then good to find out now because we need to update harmonisation if so)

Zenodo record where this will be published: https://zenodo.org/records/17527153

Long term I think @durack1's suggestion of publishing global sum files alongside the detailed files is a good one (let's see whether long-term is longer than is useful for the fast track or not)

[ChrisJones-MOHC]

thanks @znichollscr - i will have a look and double check your check! This sounds really useful. anecdotaly i know of at least one attempt to create ESM inputs that missed the aircraft emissions for example on the first draft. Such a file will be useful both for modelling groups to QC their runs but also afterwards for analysis that people can easily see what was input

[znichollscr]

also afterwards for analysis that people can easily see what was input

Great. All of this will come out on the scenario explorer eventually, so that will be the best long-term reference, but for now we have to go with workarounds

[ssmithClimate]

QC request really is a single timeseries of the total CO2 emissions

Noted. We will generate that as well (as quick fix I'll update the description to note what files sum to the total)