Assessing human footprint within protected areas using the maximum entropy model: an example from the Pearl River Delta

Protected areas in rapidly urbanizing regions are increasingly exposed to anthropogenic pressure. It is essential to quantify the human footprint and its spatial pattern within these areas. Previous studies have encountered challenges such as incomplete factor selection, subjective weighting, and coarse resolution. To address these issues, we applied the maximum entropy (MAXENT) model to assess human footprint in protected areas at different scales by combining multisource spatial data. Then, we conducted a comparative analysis to identify the optimal modeling scheme in the Pearl River Delta (PRD), China. We found that the scheme incorporating GPS trajectories of outdoor activities and Points of Interest (POIs) at 30 m resolution outperformed the conventional and coarser-resolution scheme in both accuracy and visual quality. Although the random forest model achieved a higher validation AUC, MAXENT was retained as the primary framework because it better matched the presence-only design and fine-scale pressure-gradient mapping. The MAXENT model further revealed that the dominant factors of human footprint were: nighttime light, population, and distance to cropland. The findings highlight significant spatial conflicts between conservation targets and adjacent transportation or agricultural expansion. Moreover, the results underscore the high vulnerability of sensitive ecosystems to intense light pollution and unmonitored recreational disturbances. Overall, this study provides a spatially explicit basis for assessing anthropogenic pressure in the PRD and a transferable framework for human footprint mapping in other rapidly urbanizing regions.