Emission factors need a stated basis

An emission factor is a relationship defined on a particular basis. It is not a universal property of electricity that can be moved between regions, periods and reporting questions without changing meaning. A factor may attribute emissions from a specified generation mix, represent a contractual reporting basis or estimate a response to a change in load. The analyst should identify that basis before applying the factor. A precise value with an unidentified denominator, boundary or purpose is less useful than a clearly stated provisional assumption whose limits can be challenged.

The selection begins with the intended claim. A corporate purchased-energy statement, a facility screening comparison and a power-system consequence require different evidence. The energy quantity must also be identified: electricity generated, delivered, bought or used can involve different boundaries. Losses and upstream effects may be included or excluded. The reviewer should not infer those choices from the unit alone. A factor labelled per unit of electricity is still incomplete if the relevant electricity stage and emissions perimeter are unresolved.

Geography and time matter because the factor relates to a specified system. A historical regional value may be appropriate for an inventory question under a particular method while remaining insufficient to predict the consequences of a future intervention. A supplied annual value can conceal variation during the facility's operating hours. That variation may or may not be material to the decision. The analyst should record the selected geography, period and temporal resolution, then explain why they fit the question. The existence of an official dataset does not remove the need to choose an appropriate field and application.

Contractual attribution introduces evidence beyond the physical annual balance. A supply arrangement can contain attributes, claims and conditions that matter to a reporting method. An annual quantity of renewable energy in the screen does not demonstrate those contractual requirements. It also does not establish exclusive ownership of a reporting attribute. The reviewer should ask which instrument, supplier record or programme rule supports the intended statement. The commercial label on an energy product should not be allowed to answer an unassessed accounting question.

A marginal or consequential question needs a counterfactual about what changes in the energy system when load or supply changes. Such a question can depend on timing, operating constraints and the horizon over which the system responds. A simple annual multiplication can be a preliminary approximation under an explicitly appropriate basis, but it is not an automatic dispatch estimate. More resolved methods also have scopes and assumptions. The analyst should choose the resolution that fits the decision and preserve the method's limits rather than treating complexity as a substitute for relevance.

The eGRID and AVERT sources illustrate why factor selection should be application-specific. They are retained here as public methodological context for electricity emissions information and temporally resolved change estimation within their stated jurisdictions. The study does not import a numerical factor from either source and does not run a power-system model. It would be inaccurate to describe all eGRID information as incapable of supporting avoided-emissions approximations, just as it would be inaccurate to treat every published rate as a complete marginal estimate. The relevant dataset field, method and scope determine the interpretation.

The annual Transition lab instead accepts supplied grid and additional-renewable factors. Its arithmetic assigns each factor to the corresponding annual energy used. The model cannot infer whether the factors have consistent lifecycle coverage, whether they describe the same electricity stage or whether a contractual method permits their use. The input names describe their role in the illustrative calculation, not a certification of accounting basis. A user must supply that basis explicitly before applying the screen to a real reporting or investment question.

Consistency between factors is essential to the comparison. A generation-stage grid factor and a lifecycle renewable factor can produce a difference partly driven by unlike boundaries. An analyst could then interpret that difference as a physical transition benefit even though part of it comes from accounting selection. The factor record should state a common basis or explain why different bases are appropriate for the intended question. If the difference is unresolved, the dossier should retain separate interpretations rather than present a single preferred result as self-evident.

The factor should also retain provenance and edition. Public guidance can be revised, historical datasets can be updated and a contract's attributes can change. The evidence record should name the retained source and the time to which the value applies. A future analyst needs to know whether a result changed because demand changed, the factor changed or the method changed. These are different mechanisms and may require different governance responses. Retaining the edition makes historical replication possible without implying that the same factor remains appropriate for every future decision.

Factor sensitivity should be separated from physical performance. Changing the grid factor in the screen changes attributed emissions while leaving annual electricity demand and energy charges unchanged. A lower abatement result in such a test is not evidence that the efficiency measure performs worse. It is evidence that the same supplied physical change has a different attributed consequence under the alternative factor. The distinction becomes especially important when a financial saving remains identical while cost per tonne moves sharply. The denominator can change without a change in engineering or cash benefit.

The record should explain who can approve a factor change. An engineer may update a demand estimate, while a reporting specialist may interpret a contractual method and a system analyst may develop a marginal scenario. Allowing any participant to substitute a preferred factor can create an unstable comparison whose recommendation depends on presentation incentives. A factor change should retain its reason, applicability and effect on the intended claim. This does not require a bureaucratic process for every concept assumption; it requires an accountable interpretation when the result supports a material obligation or public statement.

Where factor uncertainty spans different methods, a simple numerical range may be inadequate. A location-based interpretation and a consequential interpretation are not always endpoints of the same variable. They may answer different questions and use different boundaries. The dossier should present them as separate statements when appropriate, rather than place them in a range that suggests one unknown true factor is waiting to be discovered. Sensitivity is strongest when the tested alternative preserves the meaning of the claim. A method change requires its own explanation.

The reviewer should finally resist the idea that lower attributed abatement automatically weakens the investment. The package may still release cash, improve service or meet a legitimate obligation. Conversely, a large attributed reduction does not establish favourable economics or executability. A factor challenge reveals which proposition moves. The committee can then decide whether the change affects the investment rationale, a reporting statement, or only the interpretation of a screening metric. Keeping that consequence explicit is the purpose of the factor-selection record.