A systematic investigation into the methodological variables of embodied carbon assessment of buildings

Embodied carbon accounts for an increasing share of the life cycle emissions of new buildings. Despite many studies have been prompted to analyze buildings’ embodied carbon, reported results display large unjustifiable variations, whereas the different variables have seldom been examined systematically. This study aims to investigate how embodied carbon assessment is affected by different variables and what the exact numerical influences are. A framework is first established to examine variables in four methodological dimensions, i.e. temporal differences, spatial disparities, procedural inconsistencies, and physical diversities. Then, using this framework, a three-step normalization is conducted to minimize the discrepancies in different variables among 244 cases studies published between 2000 and 2020 to ensure they can be compared on the same scale. Eleven variables affecting the results are identified. After the normalization, the average embodied carbon in the manufacturing, transportation, and construction stages decreases significantly (to 443 kg CO2e/m2) compared with the initial samples (513 kg CO2e/m2). Remarkable variations in the assessment results are seen as a result of changes in modeling approaches (e.g. 200% increases from process-based to hybrid method), emission factor databases (e.g. 282% for steels in ICE) and building structures (e.g. 166 kg CO2e/m2 for concrete and timber buildings). The study provides a theoretical foundation, using eleven variables, for explaining variations in the reported carbon of different buildings, which better facilitates cross-case comparisons and benchmarking in future studies. The variables that exert the most influence are quantitatively identified following the three-step normalization, bridging the gap where only qualitative analysis was previously conducted.