How do natural and socio-humanistic factors influence the spatial distribution of Tibetan Buddhist monasteries?

Tibetan Buddhism is a branch of Buddhism with strong regional and ethnic characteristics, widely distributed across the Qinghai–Tibet Plateau in China. Monasteries, as key bearers and historical–cultural heritage of Tibetan Buddhism, have spatial distributions that are deeply influenced by the plateau’s unique geographical environment. Traditional studies mostly rely on qualitative analysis of historical literature, making it difficult to quantitatively capture the nonlinear characteristics and spatial heterogeneity of the influencing factors. This study constructs an interpretable machine learning model to quantitatively analyze the impacts of physical geographical and socio‑humanistic factors on the spatial distribution of Tibetan Buddhist monasteries on the Qinghai–Tibet Plateau. The results show that: (1) the model integrating LightGBM and SHAP effectively identifies nonlinear relationships between monastery distribution and its driving factors. A binary color visualization technique is innovatively employed to express these relationships spatially, significantly enhancing interpretability; (2) village/town density, precipitation, distance to ancient routes, temperature, and elevation are the main factors affecting monastery distribution. The importance of these factors varies among the three major Tibetan dialect regions: natural factors dominate in Amdo and Kham, whereas socio‑humanistic factors exert a stronger influence in Ü‑Tsang; (3) the study proposes a multi‑scale perspective on influencing factors: macro‑scale factors drive the large‑scale spread and diffusion of monasteries, meso‑scale factors shape intra‑regional spatial heterogeneity, and micro‑scale factors determine local site selection. This research provides new data support and methodological tools for quantitatively understanding the formation mechanisms underlying the spatial distribution of Tibetan Buddhist monasteries.