Smart Operational Systems for Disaster Response: A Landslide Monitoring Perspective

Effective disaster response relies on the rapid collection, verification, and dissemination of accurate information from affected areas. In the context of landslides, timely access to high-quality data is vital not only for immediate response but also for hazard assessment, model development, and the improvement of early warning systems. Despite advances in Earth Observation (EO) and Artificial Intelligence (AI), landslide detection and mapping still largely depend on manual efforts involving multiple actors and inconsistent methodologies, resulting in slow, fragmented, and often incomplete situational awareness. This presentation discusses the transition from research-focused landslide models to operational monitoring platforms that can deliver actionable intelligence in near real-time. It highlights key technical and organizational challenges, such as robust data pipelines, model deployment, benchmarking, and long-term maintenance, and emphasizes the importance of collaboration across research, policy, and operational domains. As a practical example, the Landslide Hunter platform, developed under the NWO Open Competition project “The first fully automated AI-based platform to map and monitor landslides remotely”, demonstrates a scalable, open, and sustainable approach. By integrating EO data streams, machine learning models, and rule-based reasoning, the platform enables automated detection and monitoring of landslides, even under obstructed conditions. Results are openly shared through a web-based catalog and API, promoting transparency, reproducibility, and interoperability in landslide hazard monitoring.