Machine Learning for GIS-Based Flood Susceptibility Mapping: A Global Bibliometric and Systematic Review of Research Trends and Future Directions (2005-2026)

Dr Kazi Abdul Abdul Mannan , Ismat Zerin

Flood susceptibility mapping (FSM) has become an essential component of disaster risk reduction and sustainable environmental management, particularly with the growing frequency and severity of flood events under changing climatic conditions. The integration of Geographic Information Systems (GIS) and machine learning (ML) has significantly improved the accuracy and efficiency of flood susceptibility assessment by enabling the analysis of complex spatial and environmental relationships. This study presents a global bibliometric and systematic review of research on machine learning applications in GIS-based flood susceptibility mapping covering the period from 2005 to 2026. Using an illustrative bibliometric framework for manuscript development, the review examines publication growth, leading journals, influential countries, institutional collaboration, keyword evolution, and emerging methodological trends. The findings indicate a substantial expansion of scientific output over the past two decades, accompanied by increasing interdisciplinary collaboration and the widespread adoption of advanced machine learning algorithms, including ensemble learning and deep learning approaches. The review further identifies emerging research directions involving explainable artificial intelligence, GeoAI, cloud-based geospatial computing, and multi-source Earth observation data integration. By synthesizing the intellectual development and methodological evolution of the field, this review provides a comprehensive foundation for future research and supports the development of more accurate, interpretable, and scalable flood susceptibility models for disaster risk management and climate adaptation.