Semi-quantitative method to identify the vulnerable areas in terms of building aggregation for probable landslide runout at the regional scale: a case study from Soacha Province, Colombia

Landslide susceptibility and vulnerability maps are key components for urban planning and risk management. The main objective of this research was spatial vulnerability mapping in the probable landslide runout zone in Soacha Province, Colombia. This study included three major steps: identification of a landslide susceptible area, identification of its runout zone, and vulnerability assessment using an area damage index method. The landslide-prone area was identified through a susceptibility analysis using a logistic regression model. In total, 182 landslide locations were collected and randomly distributed as training data (70%) and validation data (30%). The final landslide susceptibility map was validated using the area under the curve method. The validation result showed success and prediction rates of 88.71% and 89.96%, respectively. The Flow-R model was applied to identify the runout zone, and a back-propagation analysis approach was applied to estimate two essential input data for the model, i.e., the travel angle and velocity. From seven locations, the back-propagation analysis showed an average travel angle of 14.6° and an average velocity of 11.4 m/s. A total of 3777 buildings were identified within the probable runout zone. A physical vulnerability assessment was done by finding the ratio between area of buildings and area of runout zone in each small unit boundary. The physical vulnerability was classified as low, moderate, extensive, and complete on the basis of building exposure. The final result revealed that most of the village areas are in null or moderate vulnerability zones. In contrast to the village areas, the city areas include zones of extensive and complete vulnerability. This study showed that about 52% of the area of the city of Cazuca is completely vulnerable, i.e., in areas where abandoned quarry sites are present. The map of vulnerable areas may assist planners in overall landslide risk management.