Characterizing the Regionality of Concrete Sustainability Using Hierarchical Clustering on Principal Components

Abstract

While local in its usage, concrete carries global implications due to the sheer demand for concrete worldwide. Previous work has laid out the general principles and directions for improving the sustainability of concrete, but there is a need to understand and evaluate the relative importance of these issues at the regional and local levels to formulate effective and issue-specific solutions. This study applies hierarchical clustering on principal components, a machine learning technique, to characterize the regionality of concrete sustainability among the 47 prefectures of Japan using a set of eight regional variables that describe pressing sustainability issues for concrete construction. The analysis identified and quantitatively described five distinct clusters of prefectures using statistical metrics, then examined these patterns in the context of the socio-economic characteristics of the prefectures in each cluster. The need to effectively utilize concrete waste or increase utilization of blended cements were found to be critical issues in some prefectures, while other prefectures should aim to reduce cement consumption or water and energy consumption and CO₂ emissions. The results may be used to inform more sustainable decision-making by revealing local sustainability issues that should be prioritized. Furthermore, although the regional patterns found in this study are unique to Japan, the analytical approach serves as a demonstration of how regional differences in concrete sustainability may be investigated, and it is expected that future studies will build on this study to develop a more comprehensive understanding of the regionality of concrete sustainability.