Statistical characteristics and scenario analysis of embodied carbon emissions of multi-story residential buildings in China

Residential buildings are major contributors to carbon emissions from the building and construction industry. Although previous studies have comprehensively investigated the methods for embodied carbon assessment and proposed measures for carbon reduction, it remains challenging to screen the features and influencing factors of embodied carbon intensities from a statistical perspective. This study explored the characteristics of embodied carbon emissions based on 438 multi-story residential building samples. The influences of building features, such as the number of floors, number of basements, structural form, seismic design requirement, delivery type, and building area, were analyzed using clustering and significance tests. The average embodied carbon intensity was evaluated as 467.6 kg CO_2e/m² and relevant coefficient of variation was approximately 20 %. Further discussion on building material weight, cost, and embodied carbon intensities indicated that a simplified system boundary considering 95 % thresholds of both weight and cost can provide satisfactory results for embodied carbon assessment with relative errors below 3 %. Moreover, three linear models were proposed for predicting embodied carbon intensities. The results showed that the carbon intensities can be predicted based on five types of primary materials with good accuracy. The proposed system boundary, predictive models, and identified characteristics have practical applications for assessing, estimating, and reducing embodied carbon. Accordingly, it is possible to facilitate low-carbon design of multi-story residential buildings.