Veriye Dayalı Heyelan Duyarlılık Haritalarının Üretiminde Sismik Etki Kullanımına İlişkin Bir Çalışma

Abstract

Realistic mapping of landslides and landslide-prone areas is of great importance for mitigating the direct and indirect negative impacts of these hazards on human life. In this context, the preparation of landslide inventories and the generation of inventory maps represent a dynamic and ongoing process. Regular mapping of newly occurring landslides, particularly those triggered by extreme rainfall or earthquakes, is essential for the effective management of nationwide disaster risk reduction efforts. The presence (1) and absence (0) information in landslide inventory maps serve as outputs in modeling tools used to produce landslide susceptibility maps, while the input parameters typically include geological, topographical, and environmental preparatory factors, as well as triggering parameters such as extreme rainfall and earthquakes. However, many existing landslide inventories often lack sufficient data regarding triggering factors. In regions such as Turkey, where seismic activity is high and heavy rainfall is also common, the fact that moderate to large earthquakes generally have shorter recurrence intervals than extreme rainfall events raises the question of whether the influence of earthquakes on susceptibility is adequately reflected in current susceptibility maps. The reduction factor (rf), defined as the ratio between pseudostatic and static safety factors, can be expressed in relation to slope angle and the peak horizontal ground acceleration (PGA) of a given earthquake scenario. Nevertheless, there is no existing study in the literature in which a landslide susceptibility map generated by conventional methods is directly integrated with seismic influence using this relationship. Existing approaches often assume that seismic effects invariably increase landslide susceptibility, and thus apply these effects directly to the final susceptibility map. In contrast, this study adopts a novel approach by incorporating seismic effects as an input parameter rather than applying them post hoc to the result map. To this end, seismic effect equations were reformulated to yield meaningful values for all slope angles, leading to the definition of the Modified Reduction Factor (MRF). In the selected study area, landslide susceptibility maps were produced for three cases: without considering seismic effects (DUY1), and with earthquake scenarios of Mw=6.5 (DUY2) and Mw=7.5 (DUY3), where MRF was used as an input parameter. The area under the curve (AUC) values for DUY1, DUY2, and DUY3 were found to be 0.620, 0.638, and 0.641, respectively. These results indicate that all three susceptibility maps exhibit acceptable levels of accuracy; however, the map corresponding to the DUY3 scenario demonstrates the best overall performance.