Building Disaster Early Warning Capacity in Cali, Columbia: Extending DisasterAWARE® capabilities, assessing landslide risks to lifelines, and developing sustained procedures and skills to support disaster manager decision making in Columbia
Dr. Nishan Kumar Biswas, Goddard Earth Sciences Technology and Research II (GESTAR II) Center.
This project is a sub-award from the University of Hawaii. UMBC team will be responsible for two tasks, 1) landslide inventory development using satellite images, and 2) landslide susceptibility map generation. Based on previous discussions with Cali city officials, there does not appear to be a comprehensive record of landslide incidents, but information that is located will be incorporated into a machine-learning system that uses satellite imagery to map landslides. The system, known as Semi-Automatic Landslide Detection (SALaD) has been developed and applied by PDC-partner (UMBC) in collaboration with NASA. The team will use freely available Sentinel-2 satellite imagery for mapping. The landslide data will be generated from this imagery, covering the 2016-2025 timeframe (coincident with the record from Sentinel-2) and along lifelines serving the city of Cali. This analysis will create a consistent spatial and temporal record of landslide occurrences for the area. By combining the landslide inventory with environmental and infrastructure datasets, a probabilistic susceptibility model will be developed to identify the most likely locations of future landslides along key infrastructure corridors. Population exposure, demographic characteristics, and the vulnerability of critical assets will also be assessed using the PDC All-Hazards Impact Model (AIM), which will include estimates of affected populations and infrastructure replacement costs, and institutional capacity.