Bayesian network-GIS-based livability assessment of rural settlements in the Upper Min River Basin, China

Rural settlements are fundamental components of the human-land system and play a pivotal role in rural revitalization. Evaluating their livability helps identify key factors influencing rural development and enables targeted implementation of revitalization strategies. This study focuses on the Upper Minjiang River, China. We constructed a livability evaluation index system covering five dimensions: Site Safety, Resource Endowment, Environmental Suitability, Accessibility of Life, and Affluence of Economy. A Bayesian network-GIS probabilistic model was established to quantitatively analyze the spatial pattern of livability and identify key drivers. The results show that: (1) the overall livability of the study area trends high in the southeast and low in the northwest, with a clear pattern of distribution along rivers and roads. (2) The number of rural settlements decreases with the reduction in livability, and the number of rural settlements in the medium and high livability areas accounts for 81.33% of the total settlements. (3) Resource endowment and environmental suitability are the primary determinants of a region's overall livability. Among these, natural geographic factors such as water network density and temperature-humidity indices play a pivotal role in establishing a region's livability level. Socioeconomic factors like transportation accessibility further influence regional livability. The results provide a scientific basis for planning beautiful, livable villages and promoting rural revitalization in the Upper MinJiang River and similar regions.