Carbon accounting and GHG inventories.

Treating emissions data like financial data

An emissions figure has the same properties as a financial one. It comes from a source, it is transformed by a methodology, it is reviewed by someone, and it ends up in a published document that people rely on. The difference is that finance has spent decades building controls around that chain, and emissions data usually has not.

The consequence shows up under assurance. A verifier asks how a number was produced and the answer is a spreadsheet whose author has left, referencing an emissions factor whose vintage nobody recorded, applied to activity data that came from an email. The number may well be right. It cannot be demonstrated to be right, which is the same thing as being unsupported.

Building the inventory

Boundaries

Setting the organizational boundary, whether by equity share or control, and the operational boundary determining which sources are included. These decisions drive everything downstream and need to be documented, applied consistently, and reconciled to the reporting entity used in the financial statements.

Scope 1 and Scope 2

Direct emissions from owned and controlled sources, and indirect emissions from purchased energy. Generally the most tractable part, though it still requires a complete source inventory, reliable activity data, and both location-based and market-based Scope 2 calculations where energy attribute certificates are involved.

Scope 3

Value chain emissions, and the part where most inventories are weakest. The work is deciding which of the categories are relevant and material, choosing an approach for each, and documenting the basis. Spend-based estimation is legitimate and often the only feasible starting point, but the methodology and its limitations need to be stated, along with a path toward better data over time.

• Category-by-category relevance and materiality screening
• Method selection per category, with the rationale recorded
• Supplier data collection where primary data is achievable
• Estimation methodology and documented assumptions where it is not
• Consistency of approach across periods, so trends mean something

Methodology and evidence

Every calculation needs a documented method: the formula, the emissions factor source and vintage, the unit conversions, and the treatment of gaps and exclusions. Every input needs a retained source. Every output needs a review step with evidence that the review happened.

This is unglamorous and it is the entire difference between an inventory that passes assurance and one that generates findings.

Where the inventory gets used

One inventory typically feeds several obligations: CSRD climate disclosures, IFRS S2, California SB 253, customer questionnaires, and voluntary reporting. Building it once, to the highest standard any of them requires, is far cheaper than maintaining several versions that eventually disagree with each other in public.

Making it repeatable

The first inventory is a project. Every subsequent one should be a process: defined sources with named owners, a collection calendar, calculation logic that runs the same way each period, and a review workflow. Platform support and automation of the collection and transformation layer both help considerably here, particularly where data arrives from dozens of sites in inconsistent formats.