Evidence Hub

Impacts of Urban Three-Dimensional Morphology on Seismic Vulnerability and its Driving Mechanisms: A XGBoost-SHAP Explainable Analysis in Jiangyou City, Southwestern China

Liu Y, Zhou B, Zheng H

2026

Abstract

Abstract Understanding the interaction between urban urban three-dimensional morphology (urban 3D morphology) and seismic vulnerability is crucial for optimizing urban spatial planning and disaster risk reduction. However, existing seismic vulnerability assessments have primarily focused on individual buildings or conventional structural and socioeconomic factors, with limited attention to the nonlinear relationships between urban-scale 3D morphological configurations and vulnerability patterns. To address this limitation, this study develops an integrated and explainable assessment framework that combines XGBoost, SHAP analysis, Pearson correlation-based feature selection, and Moran’s I spatial clustering to quantify the contributions of urban 3D morphological characteristics and reveal their spatial heterogeneity in relation to building seismic vulnerability. Results show that among all feature categories, Spatial Crowding Degree is the most influential feature in the model, returning the widest range of SHAP value variations and the highest average influence. The influence of different urban 3D characteristics on seismic vulnerability exhibits diverse patterns, in which high-density and compact layouts can reduce vulnerability, while features such as building height variation and plot ratio significantly affect prediction contributions within specific ranges. Moran’s I–based spatial clustering shows that High–Low and Low–High areas reflecting transitional patterns and…

Read the paper

Related studies

Go deeper in Evidence Hub

Evidence Hub adds a study data sheet (design, sample size, evidence level, key numbers), an AI summary and a chat grounded in the paper's text.

Open this study Sign up free