Abstract Accurate forecasting of regional building stock distribution is a crucial prerequisite for large-scale seismic risk assessment and the enhancement of regional disaster resilience. Marked regional disparities in China’s existing building stock and limitations in current macro-level forecasting methods necessitate improved prediction capabilities. This article introduces a regional classification-based framework for forecasting building stock. A four-dimensional variable system (comprising geographic, administrative, economic, and demographic variables) and Ward’s hierarchical clustering were used to classify 372 cities in China’s mainland into nine distinct regions (I–IX). Analysis of these regions revealed significant heterogeneity and a distinct spatial gradient in existing building stock levels (higher values in the East/Central regions compared to the West/Northeast). Regression analyses subsequently identified both GDP growth and construction industry value-added growth as strong predictors of building stock growth ( r > 0.85) and uncovered an inverted U-shaped relationship between economic growth and stock growth rates. Based on these findings, region-specific, multivariate regression models were developed and validated against data from 2016 to 2020. Projections to 2025 using these models suggest continued stock growth across all regions, maintaining the observed spatial patterns. Collectively these findings provide a quantitative basis for differentiated urban planning, targeted policy interventions, and seismic risk assessments. Specifically, the framework supports urban renewal in high-stock regions, infrastructure optimization in less-developed regions, strategic planning for low-carbon building transitions, and targeted seismic risk assessments to enhance disaster mitigation for vulnerable urban systems.
Wu et al. (Tue,) studied this question.