Making it easy to track the emissions intensity of production and consumption 

by Seton Stiebert and Dave Sawyer  

This document presents an overview of the methodology used to estimate the emissions intensities contained in the Canadian Emissions Intensity Database.   

Tracking Canada’s progress towards its climate goals requires accurate, granular data on the emissions intensity of each sector of the economy. These data can be used to estimate Scope 1, 2, and 3 (upstream) emissions, to determine financed carbon (the emissions associated with financing businesses), to estimate facility level carbon emissions, and to estimate the emissions intensity of particular goods and services.

To develop the Canadian Emissions Intensity Database, we constructed a model to flow emissions through Canadian supply chains and economic relationships to determine embodied carbon in production, final demand (such as household personal expenditures), and investment. Two steps were followed in developing the model and generating the data contained in the Canadian Emissions Intensity Database:

1. Allocate national emissions to economic sectors.  
2. Pass embodied carbon through supply chains and allocate to domestic expenditures and exports.

Each step is discussed below. 

1. Allocate national emissions to economic sectors  

The model first allocates all emissions from the National Inventory Report (2021 emissions) to 68 industry sectors aligned with the detailed industry categories in Canada’s Symmetric Input-Output Tables (2021). These emissions data are then organized by the North American Industry Classification System (NAICS), as well as to a few categories of household expenditures where fuels are purchased and consumed and result in direct emissions from households. 

The emissions allocation model maps National Inventory Report (NIR) emission categories to industry sectors and household expenditures. Direct mappings of emissions are first established where there is a perfect correlation between the emissions category in the NIR and one of the 230 industry sectors in Statistics Canada’s Input-Output Tables. This approach accounts for 44.6 per cent of all NIR emissions in 2020 excluding land use, land-use change, and forestry. The remaining 55.4 per cent of emissions require that NIR emission categories be mapped to multiple industry sectors. This indirect mapping of emissions is based on different available datasets:  

• Energy-related emissions were mapped using the Canadian Energy and Emissions Data Centre’s energy database, which estimates industry-level fuel demand and related emissions.  
• Unmapped industrial process emissions can be mapped based on expenditure data for different equipment that gives rise to these emissions.  
• Fugitive emissions and non-energy products from fuels and solvent use can be mapped using NIR Tables A10-3 with additional expenditure data for allocation to industry sub-sectors.   

Indirect mappings have a somewhat lower confidence level or a higher uncertainty then direct mappings; however, note that total NIR emissions and the subsequently mapped total industry and household emissions are exactly equal, and so the model allocates all emissions to expenditures and exports without concern that there are either missing emissions or double counting of emissions.

There is a further mapping of some emission categories to personal household expenditures, notably energy emissions that arise directly from household operations such as for passenger vehicle operation, heating of dwellings, off-road recreational vehicles, household generators and other gasoline or diesel powered equipment, or other residential equipment. There are also embodied industrial process emissions associated with the operation of household refrigerators as well as some use of household products.

2. Pass embodied carbon through supply chains and allocate to domestic expenditures and exports

This step links industry embodied carbon to consumed goods and services across the entire economy. 

With all 2021 emissions from industry allocated to 68 industries, the model uses the 2021 National Symmetric Input-Output Tables from Statistics Canada (detailed level) to understand how output from each sector was consumed either by other domestic industrial sectors as intermediate inputs, exported from Canada, or used and consumed by one of 51 final demand categories. Final demand includes 100 personal expenditure categories (emissions by households), as well as various expenditure categories in government, construction, machinery and equipment, and intellectual property categories.  

The result of this step is that 50.6 per cent of embodied emissions are purchased by industry, 32.4 per cent are exported, and 16.9 per cent are allocated to the 51 final demand expenditure categories.

The model then considers that the 50.6 per cent of embodied emissions used as intermediate inputs must be ultimately passed on by the purchasing industry sectors as embodied emissions in their final products. The Symmetric Input-Output Table is then employed in a second interaction to allocate these emissions, where it allocates 21.4 per cent of total embodied emissions as purchased by industry, while the remainder is exported or allocated to final demand expenditure categories. Further iterations of the model are conducted until all emissions are finally allocated to either exports or final demand expenditure categories.

Carbon intensities  

This model estimates Scope 1, Scope 2, and Scope 3 (upstream) emissions and carbon intensities associated with any industry output (tCO2e/$ output or $ GDP) or any expenditure (tCO2e/$ expenditure). Note, there is a difference between carbon emissions released upstream and embodied in production and consumption and embodied carbon that is unburnt and passed through supply chains.  Practically, this means that the Scope 3 emissions intensity includes emissions that are released upstream in production but does not include embodied carbon that is eventually released by another end-use downstream.

Scope 1 emissions in the model are simply the total emissions first allocated to all sectors. Scope 2 emissions are the emissions associated with the first purchase of electricity output from the electric power generation, transmission, and distribution sector.  Scope 3 emissions are all other emissions that are embodied in output that is purchased by industries or ends up in final consumption expenditures or exports.

Industry emissions intensities are revealed by dividing the total Scope 1, 2, or 3 emissions by the 2020 output of the sector or by the market GDP of the sector.  

Expenditure emissions intensities (for products consumed) are revealed by dividing the total Scope 1, 2, or 3 emissions by the total 2020 expenditures of the consumption category.

Caveats to modelling

1. The model reveals national average emissions intensities and is therefore not representative of output from specific provinces or facilities that may have a very different emissions intensity profile.  Therefore, the results of the model are less useful for estimating project level or facility level emissions and more useful for generating a view of emissions from a portfolio of investments or purchases in Canada. As such, the results have little use for any regulatory purpose.
2. The model only includes Canadian emissions. When considering the embodied emissions of products, it will only be useful for estimating Scope 1, Scope 2, and Scope 3 emissions for products manufactured and sold in Canada. However, many domestically manufactured products will have inputs that are imports or components that are imports, thus potentially underestimating Scope 3 emissions. Most of the 60 industries have more than one product or commodity that they sell. Emissions may be far more intensive in one commodity than another. As we track only the emissions intensity average of the entire production, we lose perspective on these sub-products and how they may increase or decrease the embodied carbon of purchases by different industries or by final consumption expenditure categories. These differences could lead to substantially different emissions intensities if they were tracked.