Integrating Physical Monitoring and Social Media Analytics for Multi-dimensional Perception of Landslide Hazards

Abstract

Landslide disasters are characterized by sudden onset, limited predictability, and severe consequences, with deformation and failure processes governed by complex interactions among geological structures, hydrological conditions, and engineering disturbances. Recent advances in remote sensing, sensor technologies, and artificial intelligence have facilitated the transition of landslide monitoring from single-parameter observations to multi-source information fusion. However, existing approaches primarily characterize the physical evolution of slopes and provide limited insight into post-disaster societal impacts and public responses. This study reviews landslide classification schemes and recent advances in multidimensional monitoring technologies within the “space–sky–ground–underground” framework and discusses their applications in deformation detection and early warning. Social media data are further incorporated as a complementary source of disaster-related information, and a Weibo-based method for analyzing public responses is developed through data collection, text processing, keyword frequency analysis, and word-cloud visualization. By integrating physical monitoring data with social sensing information, a multi-source fusion framework is proposed to enhance landslide monitoring and assessment throughout the disaster lifecycle and support emergency response.