Unraveling the evolution of landslide susceptibility: a systematic review of 30-years of strategic themes and trends

:Landslide susceptibility mapping (LSM) research is critical for preventing and mitigating regional landslide disasters. Despite its importance, few researchers have systematically analyzed the key areas of LSM's development. SciMAT, a scientometric tool, offers the possibility of a graphically displaying of landslide susceptibility hotspot themes and their evolutionary trends. In this study, We searched the Web of Science core collection database for literature on LSM published from 1993 to 2022 with the search term “TI=(landslide susceptibility)”. The literature type and language were limited to “article” and “English”. After removing duplicate and irrelevant data, a total of 1661 papers were obtained. To analyze the retrieved literature, we employed bibliometric VOSviewer and SciMAT. Innovatively, we conducted cluster analysis and thematic evolution analysis using SciMAT, which revealed popular themes and evolutionary trends in landslide susceptibility. The results showed an upward trend in publications over the past 30 years. Landslide susceptibility modeling methods, geological information, and landslide-triggering factors were key topics of interest. The evolution of landslide susceptibility modeling methods is the primary knowledge path, with related topics appearing most frequently as essential nodes in the evolutionary map. There is a notable and widespread trend towards utilizing machine learning and deep learning techniques to achieve precise risk zonation in landslide susceptibility research. The application of artificial intelligence (AI)-based modeling methods has gained significant popularity due to their consistently high accuracy rates, often surpassing 90 percent, as evidenced in numerous studies. Particularly in recent years, with the advent of the big data era, Convolutional Neural Network (CNN)-based approaches for landslide susceptibility modeling have emerged as a dominant theme, showcasing exceptional fitting capabilities and robust predictive performance. The study provides valuable references for scholars to identify gaps in the literature, highlight key research directions, and inform policy and decision-making related to landslide susceptibility.