Model: Static Normals (data-cutoff: 2024-12-31)
Emissions per province and territory, and per greenhouse gas, modelled as non-time-varying Normal distributions. National totals are modelled as the sums of provincial and territorial totals. This is a baseline model, not intended to be accurate.
Model Estimates of Past and Future Emissions
Emissions are summarized by year, and in terms of total Mt per non-aggregate NIR sector. The black lines reflect reported historical emissions (the dotted line is the reported mean, the two thin solid lines are upper and lower 95% credible interval for the mean). The coloured elements reflect modelled emissions (the dotted line is the modelled mean, the shaded region is the model-estimated 95% credible interval for national emissions per sector). Colours correspond to high-level emissions categories:
- dark blue for energy used in stationary combustion (e.g. for public electricity, oil and gas extraction, etc.)
- light green for energy used in vehicles (e.g. light-duty, heavy-duty, off-road, marine, etc.)
- purple for fugitive sources(e.g. venting, flaring)
- wheat for agriculture-related emissions (e.g. soil, manure)
- pink for waste-related emissions (e.g. landfills)
- other colours for various industrial activities (mineral products, chemicals, metals)
Prediction Quality
Prediction Task: Pre-NIR-2025-m04
Pre-NIR-2025-m04 is about predicting the last year of NIR-2025 data (calendar year 2023), 4 months ahead of publication.
The predictions themselves are shown above. The table below lists KL divergences between the reference and predicted emissions at the level of national totals. For more detail, see post Static Normals: a baseline predictive model, appendix on NIR-Prediction.
Overall weighted KL-divergence over all sectors, regions, and greenhouse gases: 3.322.
KL Divergence of predictions from reference national emissions (lower is better)
Posts developing this page
- © James Bergstra, 2026.
- Licensed under CC-BY.
- Statements on this site are not necessarily true or fit for any purpose.
- Design: HTML5 UP