Spatial Heterogeneity of Traditional Villages in Southern Sichuan, China: Insights from GWR and K-Means Clustering

Understanding the spatial heterogeneity and driving mechanisms of traditional villages is critical for their tailored preservation and revitalization. Existing studies often overlook intra-regional variations shaped by historical and cultural contexts. In addition, the lack of systematic quantitative approaches limits the formulation of effective conservation strategies. This study addresses these gaps by examining 71 nationally listed traditional villages across five prefectures in southern Sichuan, China. We first mapped spatial patterns using ArcGIS10.5 and Geodetector. Then we applied GWR (adjusted R 2 = 0.70), K-means clustering, and Kruskal–Wallis tests to examine the spatial heterogeneity. This workflow resulted in three different village clusters related to historical migration: S1-Indigenous ( n = 14)—Villages established before the Ming Dynasty, primarily inhabited by indigenous Sichuan residents. S2-Huguang migrants ( n = 30)—Villages formed during the late Ming to early Qing “Huguang Migration to Sichuan,” facilitated by proximity to rivers and transport routes. S3-Refugees ( n = 27)—Villages settled by war refugees from northern and eastern Sichuan, often located in secure, high-elevation areas. Based on these findings, we propose tailored conservation strategies: preserving historical layout and architectural integrity in S1; maintaining migration-shaped forms and highlighting cultural imprints in S2; and balancing spatial conservation with improved mountain road accessibility in S3.