Data-driven cross-regional climate zone classification for building codes: A future-compatible framework

The University of Hongkong, Department of Architecture, Hong Kong SAR, China

^* Corresponding authors: This email address is being protected from spambots. You need JavaScript enabled to view it. , This email address is being protected from spambots. You need JavaScript enabled to view it.

Abstract

Climate zone classification (CZC) provides a foundational layer for building energy codes and performance standards. However, nation-specific classification schemes pose challenges to cross-border building design and construction, as well as to the alignment of performance criteria in fields such as thermal comfort and energy efficiency. Traditional climate classification systems, exemplified by the Köppen-Geiger (KG) scheme, are mainly derived from ecological and environmental characteristics, and are therefore not tailored to the requirements of building regulations. Accelerating climate change further challenges the static nature of traditional schemes with higher variability and more intensive and frequent extremes. This study proposes a data-driven CZC framework that integrates spatial heterogeneity and temporal adaptability, enabling both historically grounded and future-oriented climate zoning. The framework constructs a unified weather feature space using long-term historical sequences consolidated from ISD, ERA5-Land, and NSRDB datasets. From these sequences, we derive both Typical Meteorological Year-based features that characterize mean climatic conditions and extreme-event-based features that captures variability and tail risks. These features are then clustered through K-means to identify climate groups. Cluster quality is evaluated using a suite of EnergyPlus-simulated metrics, ensuring that zones reflect not only climatic similarity but also performance relevance for building design. Across simulated energy metrics, the proposed framework demonstrates substantially improved clustering performance relative to the KG system, as reflected by an over threefold increase in the average F-statistic (7.20 vs. 2.14). Using CORDEX-derived future weather projections calibrated to the historical feature space, we further show that the framework supports trajectory-based analysis of climatic momentum. Among ten cities in North and South America, six are projected to shift into different climate zones in future periods (2025-2045, 2046-2070, 2071-2095), highlighting the material impact of climate change on building code applicability. Overall, the framework provides a scalable and future-compatible basis for climate zoning, enabling more consistent building energy modeling and supporting regulatory adaptation under evolving climate conditions for more targeted improvement in occupant-centric indoor environmental quality and energy efficiency.