Identifying the natural and anthropogenic factors influencing the spatial disparity of population hollowing in traditional villages within a prefecture-level city

In developing countries, the phenomena of rural depopulation have been an intense continuing, which have become a major bottleneck for the sustainable revitalization of traditional villages. However, the factors influencing the spatial disparity of population hollowing (SDPH) in traditional villages within a prefecture-level city have not been fully quantitatively researched. Based on the factors that influence general villages, this study incorporated historical and cultural factors related to traditional village characteristics to construct a targeted influencing factor index system and then identified the key factors by applying the geo-detector method. With the percentage of resident population (PRP) used as a metric, this study examined Lishui, one of China’s traditional village agglomeration regions, as an example to explore SDPH in traditional villages. The results of this study were revealed in the following. (1) The average PRP value in traditional villages in Lishui was 0.68, with clear spatial disparities between the northern region (0.73) and the southern region (0.57). (2) The factors driving the SDPH included both natural and anthropogenic factors; of these, altitude, the number of public facilities, and the number of communication base stations were the most significant influencing factors. In contrast, historical and cultural factors have relatively low impacts. (3) The interaction relationships of pair factors were often enhanced on a bivariate basis, with the highest enhanced impact occurring from the interaction of two variables: the degree of intangible cultural inheritance and altitude. (4) The intervals of the variables leading to the hollowing of the population above a moderate level can be detected. This method can effectively analyze the factors influencing SDPH in traditional villages; can help reveal the interaction impact of pair factors; and can help identify the factors’ risk intervals.