Keep the structure or rebuild? How to make a fair carbon comparison

Same brief, or it is not a comparison

A retain-or-rebuild study can mislead before any carbon is counted. One option may provide more area, a different use or a higher performance target. Those can be legitimate benefits. They must be visible in the brief, rather than hidden inside an apparently like-for-like carbon comparison.

Fix these items before modelling, and keep them identical for both options:

• the required use, service and usable floor area, with gross floor area and net-to-gross differences reported
• the performance targets: energy, comfort, daylight and any rating
• the study period and the decision horizons within it
• the weather data, occupancy schedules and operating assumptions
• the emissions factor scenario, stated year by year for electricity and any fuels.

If the existing frame cannot deliver the brief, that is a finding. Record it rather than quietly changing the brief for one option. If the feasible options genuinely differ in area, show absolute emissions for each option, emissions per unit of function, and any unmet or extra service separately.

Rating rules are not a complete decision boundary

Keep two outputs distinct: the assessment required by the rating scheme, and the option study used to decide what to build. Their boundaries need not be identical.

For example, NABERS Embodied Carbon Rules v2.1 allow materials retained on site in their original state to be treated as zero embodied carbon. They exclude demolition of previous buildings from a new-building rating but include demolition associated with a partial rebuild [1]. The rating result therefore needs interpretation before it is used to compare retention and replacement.

For the decision study, count the future activities caused by each option once: demolition, strip-out, strengthening, new materials, construction, operation, maintenance, replacements and end of life. Do not charge either option again for emissions from the original construction that have already occurred. State exclusions, including refrigerants and temporary works where relevant. Reconcile this decision boundary with the scheme calculation so the team can explain differences. Keep module D benefits separate from upfront emissions.

Green Star also has version-specific treatment. Its published Buildings v1 reuse FAQ is a defined credit route, not a general retain-versus-rebuild method [3]. GBCA has since introduced Buildings v1.1 [4]. Use the requirements for the project's registered tool and revision. Do not transfer an earlier credit route to another version or assume a rating advantage settles the carbon decision.

The equal-brief comparison

The cumulative comparison

For each option, place future emissions in the years in which they are assumed to occur. Report the initial works separately. Use the same unit, such as tonnes of CO2e, throughout. At each decision horizon t:

Difference at year t = (rebuild initial emissions − retain initial emissions) + Σ(rebuild operational emissions − retain operational emissions) to year t + Σ(rebuild other included emissions − retain other included emissions) to year t

A positive difference means retention is lower; a negative difference means rebuilding is lower. Plot both cumulative curves at agreed horizons, such as 10, 20 and 30 years. Also show the full study-period result and the treatment of remaining service life. A crossing is scenario-dependent and can reverse after a later replacement. The simple payback calculation - extra initial emissions divided by annual operational savings - works only with a positive, constant saving and no other differences. The three examples below use deliberately invented tonnes of CO2e to test that logic; they are not forecasts.

Hypothetical A. Retain starts at 1,200 because of strengthening, then adds 80 a year. Rebuild starts at 1,000, then adds 100 a year. The rebuild has no operational advantage, yet it is lower until year 10, when the two are equal at 2,000. Retain is lower after that. A short horizon favours the rebuild here.

Hypothetical B. Retain starts at 1,000, then adds 100 a year, with a 1,000 replacement in year 10. Rebuild starts at 1,500, then adds 80 a year. The simple screen says 25 years. The cumulative comparison shows retain lower until year 9 and rebuild lower from year 10 (2,300 against 3,000), because of the replacement.

Hypothetical C. Rebuilding adds 1,000 tonnes more upfront and saves 100 tonnes in year 1. The saving falls by 15% each year thereafter. The constant-saving screen says 10 years. But the actual 30-year sum is 100 × (1 − 0.85³⁰) / (1 − 0.85) = 661.58 tonnes. If that decline continues indefinitely, total savings approach 666.67 tonnes and never repay the extra 1,000 tonnes.

Test the result against at least these four assumptions:

1. Emissions factors. Show the electricity and fuel scenarios year by year, separately from energy demand. Where rebuilding has higher initial emissions but uses less electricity, a declining grid factor reduces its future carbon savings for the same kWh difference. With other impacts unchanged, this delays or removes its payback. If energy demand, fuel type or replacements also differ, use the full curves; a blanket rule about grid decarbonisation is not enough.
2. Façade on the retained frame. If a new high-performance façade can go on the retained structure, much of the rebuild's operational advantage may also be available to retention.
3. Strengthening quantity. New steel and concrete for strengthening, transfers or extra storeys can erode the retention advantage quickly.
4. Study period and horizons. A longer period gives operational savings more time. A shorter one does not always favour retention: if strengthening makes the retained option's initial impact higher, the rebuild can be lower early, as in hypothetical A.

Our recommended decision rule is simple. Call the carbon preference robust only if it survives the credible cases tested. Call it conditional if a named assumption reverses it, and identify the survey, design work or data needed to resolve that assumption. If uncertainty is larger than the apparent difference, report no clear carbon winner. Energy cost, comfort, heritage, programme and functional benefits remain separate decision criteria.

When each option loses

Retention can be unsuitable before carbon is considered: the structure may not safely deliver the intended use, or the feasible upgrade may fail the brief. Establish condition, capacity and upgrade scope before comparing quantities. The ABCB's handbook provides a framework for developing the scope of work to an existing building [5]. Have the structural engineer and building surveyor record the project-specific constraints; do not treat an untested retention sketch as a feasible option.

Rebuild is not automatically right either. Retention at scale is possible. The GBCA reported that Quay Quarter Tower in Sydney retained 65% of the original 1976 structure and 95% of its core, and saved over 12,000 tonnes of CO2 emissions during construction [6]. That is a project selected by the GBCA and Lendlease to show what is possible. It proves feasibility, not a general rule.

Next step

At feasibility, commission one coordinated option study. Obtain a condition and capacity assessment, a defined compliance scope, comparable cost and material schedules, and energy models for both options. Ask the LCA lead to maintain one decision boundary. The team briefing should end with five items: preferred option; cumulative emissions at agreed horizons; assumptions that can reverse the result; unresolved evidence with an owner and deadline; and the decision that must wait. Preserve the retention option until that evidence has been considered.

If the option comparison uses product declarations, first check whether the EPDs are comparable on the chosen functional and life-cycle basis.

Sources

1. NABERS, Embodied Carbon – The Rules, version 2.1 (February 2026), s3.2.2, s7.6, s7.7 and s7.8.
2. ABCB, Embodied carbon handbook (July 2026), s5.5.5. Voluntary guidance, not NCC requirements.
3. GBCA, Green Star FAQ 356, Green Star Buildings v1, credit 24 Upfront Carbon Reduction (updated 15 July 2026).
4. GBCA, Green Star Buildings v1.1, transition dates.
5. ABCB, Upgrading existing buildings handbook, section 1.
6. GBCA, Reducing upfront carbon at scale (31 May 2023).