An environmental product declaration is evidence about a product. A building carbon calculation is an assessment. Neither becomes a better building until someone uses the information to make, document and verify a design decision.
The next phase of carbon accounting is about decisions
Construction teams increasingly receive environmental product declarations (EPDs), model-based quantity schedules and life-cycle assessment (LCA) reports. Yet the critical question often remains unanswered: which decision changed because of the carbon evidence?
The European Commission’s revised Energy Performance of Buildings Directive (EPBD) introduces life-cycle global warming potential disclosure for new buildings above 1,000 m² from 2028, extending to all new buildings from 2030. These are EU requirements with national implementation pathways; they must not be presented as automatically identical to Norwegian law [1].
Norway already has a specific climate-accounting requirement under TEK17 §17-1 for new construction and major refurbishment of apartment buildings and non-residential buildings, using NS 3720:2018. The minimum scope includes A1–A4, B2 and B4 for specified building elements, together with construction-site waste [2]. This is not the same thing as an exhaustive whole-life A–C assessment. Project teams should state which scope they are reporting, rather than calling every calculation “whole-life carbon.”
The practical challenge is to connect product evidence to quantities, design alternatives, responsibility and revision history.
An EPD is a source, not a verdict
EN 15804+A2 establishes core product-category rules for construction-product EPDs, including indicators, life-cycle stages and conditions for comparisons [3]. It does not say that every EPD is directly interchangeable with every other EPD.
Before comparing two products, check at least the following:
- Declared or functional unit: one kilogram, one cubic metre and one square metre of installed assembly are not equivalent bases.
- Functional performance: compare alternatives that meet the same structural, thermal, acoustic, durability and fire requirements.
- Life-cycle modules: A1–A3 manufacturing results cannot simply be equated with an A–C whole-life result.
- Product and geography: generic, manufacturer-specific and project-specific datasets carry different relevance.
- Version and validity: record the EPD identifier, programme operator, publication date, validity and applicable PCR.
- Scenarios: transport, maintenance, replacement and end-of-life assumptions may change the ranking.
A low A1–A3 number can be a useful signal, but not proof that a complete building alternative has the lowest lifetime impact.
Why a building model is not a carbon model
A BIM model can provide element types, geometry and quantities. It does not necessarily contain reliable material compositions, waste factors, replacement intervals, transport distances or end-of-life scenarios.
A wall object labelled “timber façade” may hide several layers of cladding, membranes, insulation, fasteners and substructure. Quantity extraction is therefore an evidence problem as much as a software problem. An apparently precise result can conceal a missing material layer or an incorrect unit conversion.
Teams should distinguish three categories of input: verified project quantities, estimated quantities and proxies. Every proxy should have an owner and a plan for replacement with better evidence. The carbon model should also record what is excluded.
A proposed carbon decision register
Jarlhalla proposes a simple, version-controlled decision register as a working method — not a new regulatory requirement.
For each consequential material or design decision, retain a compact evidence record:
| Field | What to capture |
|---|---|
| Decision ID | Stable identifier tied to a design issue |
| Model revision | IFC/model version, date and element references |
| Alternatives | Functionally equivalent options |
| Quantities | Unit, quantity source and measurement assumptions |
| Carbon factors | EPD or dataset ID, version, modules and unit |
| Scenario | Transport, installation, replacement and end-of-life assumptions |
| Uncertainty | Known gaps, proxies and sensitivity tests |
| Outcome | Selected option, reviewer, rationale and approval date |
The record should be readable by engineers and auditable without access to a proprietary dashboard. It should also allow a rejected option to be reconstructed: the reason for rejection is part of the project’s institutional memory.
Compare systems, not isolated products
Imagine two façade options that both satisfy the project brief. Option A has a lower product-stage emissions factor, while Option B requires less supporting structure or fewer replacements over the assessment period.
A credible comparison should use the same reference study period, service conditions, boundary, functional requirements and calculation method. If the alternatives differ in operational energy, maintenance or replacement, those differences must be included where relevant to the decision.
For illustration only, consider the equation:
Scenario GWP = Σ(quantity × matching impact factor) + scenario-specific processes
This simplified expression is not a substitute for NS 3720, EN 15978 or a compliant LCA tool. Its purpose is to show that every multiplication requires matching units, boundaries and assumptions. A numerical result without those links is difficult to audit.
Make uncertainty visible
A single figure such as 285 kg CO₂e/m² can look definitive, even when it contains generic products, provisional quantities and uncertain transport distances. Rather than inventing confidence intervals, teams should test defensible alternatives: shorter and longer transport routes, different replacement assumptions, and plausible quantities for poorly defined assemblies.
A practical register can mark each input as verified, estimated or proxy and show which decisions are sensitive to uncertain inputs. Where a choice changes under reasonable scenarios, the design team should avoid claiming a clear winner.
Keep module D separate
Benefits and loads beyond the system boundary, commonly reported in module D, can be valuable contextual information. They should not silently be subtracted from the building’s A–C impacts to create a more flattering headline. Report them separately and follow the applicable methodology and disclosure rules.
The same caution applies to biogenic carbon: the timing and treatment of carbon uptake, storage and release depend on the assessment framework and product data. “Carbon stored” does not automatically mean “net-zero building.”
What the EU transition changes
The European Commission describes the 2028 and 2030 disclosure milestones under the revised EPBD and a roadmap process for national limit values [1]. Its 2026 guidance on life-cycle approaches encourages integrating design, material selection, renovation and end-of-life considerations [4]. The EU Level(s) framework offers a shared language for sustainability performance, including life-cycle global warming potential [5].
For Norwegian teams, the right response is not to assume that the EPBD has replaced TEK17. It is to maintain a clear regulatory matrix: current Norwegian requirements, contractual client requirements, voluntary whole-life assessments and prospective EU-driven obligations should be separately identified.
This distinction is especially important for international projects and suppliers serving several markets.
A 30-day implementation pilot
The following is a proposed Jarlhalla editorial workflow, not evidence of a completed pilot.
Week 1 — Define the baseline. Choose one live design package, document the applicable reporting scope, freeze a model revision and identify the three highest-impact material categories in the preliminary calculation.
Week 2 — Audit the inputs. Trace quantities back to model objects or schedules. Check units, EPD versions, module coverage and data gaps. Assign owners to uncertain inputs.
Week 3 — Test alternatives. Compare two or three functionally equivalent options using aligned boundaries. Run sensitivity checks where transport, service life or replacement assumptions could alter the outcome.
Week 4 — Review and close. Have a responsible designer and LCA practitioner review the evidence, document the selected option and archive the calculation with its model revision. Record whether the decision reduced the estimated impact, was inconclusive, or was rejected for other technical reasons.
A pilot should be judged on traceability, decision quality and reproducibility — not on an arbitrary percentage of claimed carbon savings.
The human approval boundary
Automation can flag missing EPDs, incompatible units, stale model versions and unusually high-impact components. It can also prepare alternative comparisons. But engineering responsibility, statutory compliance, procurement choices and contractual approvals remain with authorized professionals.
The strongest digital workflow is therefore not the one that automatically picks the “greenest” material. It is the one that makes evidence, assumptions, alternatives and accountable decisions visible.
Conclusion: carbon data must survive design changes
A building’s carbon estimate will change as its design develops. That is normal. The failure is not revision; it is losing the explanation of why the result changed.
A version-controlled carbon decision register can connect environmental data with the reality of architectural and engineering practice. It gives project teams a way to distinguish genuine improvements from changed boundaries, new datasets or corrected quantities.
For Jarlhalla, the opportunity is to connect BIM, circular material strategies and responsible AI around one principle: sustainability claims should be reproducible, reviewable and tied to decisions.
Sources and verification notes
- European Commission, Global warming potential of buildings. Official EPBD implementation overview; 2028 and 2030 disclosure scope.
- Direktoratet for byggkvalitet, TEK17 §17-1: Klimagassregnskap fra materialer. Current Norwegian regulatory guidance and NS 3720 reference.
- NEN, EN 15804:2012+A2:2019 standard description. Product EPD core rules and comparison conditions; full licensed standard not reproduced.
- European Commission, Commission publishes new guidance to decarbonise buildings and increase their circularity, 27 March 2026.
- European Commission, Level(s): EU sustainability assessment framework. Includes new Level(s) publication listed 23 October 2026.
- European Commission, Life-cycle GWP calculation framework announcement, 16 December 2025.
- SINTEF Byggforsk, Klimagassregnskap for bygninger — krav, utarbeidelse og bruksområder, November 2024. Overview of methodology and reporting.
Source limitations: Official public guidance and accessible standard descriptions were checked. Full licensed text of NS 3720, EN 15804 and EN 15978 was not reviewed. National implementation details must be checked at publication and on a project-by-project basis.