Statistical landslide susceptibility assessment of the Mansehra and Torghar districts, Khyber Pakhtunkhwa Province, Pakistan

Natural hazards greatly impact human life and the development of infrastructure, especially in poorer countries. To tackle this issue and develop sufficient strategies with which to assess hazards under local governing conditions, open-source approaches have been pursued in many aid development and technical cooperation projects. In this study, the landslide susceptibility of the Mansehra and Torghar districts is assessed within the framework of the multiphase German-Pakistani technical cooperation project “Georisk Assessment Northern Pakistan” (GANP), which was initiated after the devastating Kashmir Earthquake occurred on October 8, 2005. This landslide susceptibility assessment is strongly aligned with an open data approach that utilizes free, accessible data and data sources. The landslide inventory was collected using high-resolution imagery provided by Google Earth. The landslide-controlling parameters were derived from a series of open access data, including published data, free and accessible satellite imagery (e.g., Landsat), and digital elevation models. A mid-level complexity statistical approach was used to assess relative landslide susceptibility patterns on a regional scale. Stepwise data integration with subsequent evaluation was then applied. The evaluation of the model quality was performed using receiver operation characteristic curves. The established workflow, which also incorporated sensitivity studies and an uncertainty assessment, produces reliable landslide susceptibility patterns that can be used for regional spatial planning.