Akademik Veri Yönetim Sistemi
Turkish Journal of Earth Sciences, cilt.33, sa.2, ss.208-223, 2024 (SCI-Expanded)
Landslides constitute recurrent natural hazards worldwide with greater socioeconomic impacts in low-income places. Landslide susceptibility studies are carried out to determine the potential for landslides to occur in a region, to predict possible risks in the future and to integrate this information into sustainable planning and management strategies. Slip-type landslides occur in limited areas of less than 1% of the Asi watershed in Southern Türkiye. Therefore, meaningful estimations cannot be made with classical statistical models. So, using the maximum entropy approach (MaxEnt), a technique for estimating based on a small number of observations, a landslide susceptibility assessment was constructed. The association between the locations where landslides are detected and environmental variables is modeled using the best probability density in the maximum entropy technique. The analyses were repeated with three different random selection methods using two-thirds of the landslide dataset to analyse it and one-third as validation data. The accuracy and precision of the obtained models were evaluated with the area under the receiver operating characteristic curves and success prediction curves. Hence, we performed a susceptibility assessment based on a maximum entropy method in the 7893 km2 Asi watershed. The area under the receiver operating characteristic curves was 0.76 as a result, and 72% of the landslides occurred in areas with a total area of 28% that were high and extremely high susceptible areas. The determination of the landslides that will develop in the region over time and the susceptibility assessments to be repeated will contribute to the more effective planning of mitigation studies. This method can be applied in other watersheds with localized landslide processes that affect human activities. Landslide susceptibility studies are a fundamental tool for understanding landslide risks at a regional scale and predicting future hazards. These studies have a central role in regional planning, infrastructure management, crisis response, and the formulation of sustainable development strategies.