Extended Abstract

Statement of the Problem: Landslides are among the most important geomorphological hazards in mountainous areas. Large-scale landslides often have significant social and economic impacts, making them one of the costliest natural disasters worldwide, causing considerable damage to individuals and infrastructure each year. Nowadays, due to climate change and human activities, the frequency, intensity, and impact of landslides have increased, posing significant threats to life and property. In recent years, with rapid socio-economic development, the spatial extent of human activities has also gradually expanded. The occurrence of heavy rainfall has increased the frequency of landslides in mountainous areas and the resulting economic losses have gradually increased, turning landslides into barriers to human environment development. Given the complex physics and diverse nature of triggering mechanisms (such as rainfall and earthquakes), dynamic landslide monitoring is vital for risk assessment and management. While small landslides constitute the majority of landslide events annually, large landslides are usually responsible for a significant portion of casualties and damages. In most parts of the world, identifying their locations and effects still faces many challenges due to the diversity of landslide types and morphologies, as well as the difficulty in collecting and updating survey forms. Therefore, the development of tools and techniques for identifying and monitoring these widespread gravitational processes is of great importance. Spatial assessment and preparation of hazard maps are fundamental strategies in landslide risk management. In the present study, spatial assessment and zoning of landslide risk in the Zemkan watershed in Kermanshah have been addressed. This watershed drains parts of the western slopes of the Zagros Mountainous area and has a high potential for landslides.

Methodology: In order to assess the spatial risk of landslides in the watershed, studies of 13 factors affecting landslide occurrence were conducted, including elevation, slope, aspect, the Melton ruggedness number, surface convexity, length of slope, valley depth, TWI, precipitation, geological formations, distance from streams, distance from roads, and vegetation cover. In line with the research objectives, the necessary data were obtained from 1:25,000 and 1:50,000 topographic maps; 1:100,000 and 1:250,000 geological maps; Aster satellite images with a resolution of 27 meters; Sentinel satellite images (with a resolution of 10 meters) and Google Earth (GeoEye with an approximate resolution of 1 meter); climate data; field studies; and library resources. ArcGIS, ENVI, and SAGA GIS software were used for the preparation and preparation of thematic layers and the execution of research models. A catastrophe theory-based approach was employed for zoning and spatial prediction of landslide occurrence risk in the Zemkan watershed area. Functions of catastrophe theory were used in GIS environment to combine and overlay these factors. The model employed was used to address uncertainties associated with decision-making and data classification. In this model, the weight of criteria is determined by the internal mechanism of the system, which has a mathematical nature.



Results and discussion: The results of the research demonstrate the satisfactory performance of catastrophe theory functions in zoning the risk of landslides. The model accuracy, calculated using field mapping of landslide zones and receiver operating characteristic (ROC) curves, was approximately 90%. Catastrophe theory-based modeling showed that slope factors with a coefficient of 1.3, precipitation with a coefficient of 1.2, elevation with a coefficient of 1.1, and geological formations with a coefficient of 1 are the most important variables affecting landslide occurrence in the Zemkan watershed. Over 32.4% of the study area fell into high and very high susceptibility classes. Several conditions have led to significant portions of the central and southern parts of the Zemkan watershed having a high landslide potential. In this regard, factors such as high elevation and slope, significant precipitation, the presence of landslide-sensitive geological formations such as marl, clayey limestone, and shale (Kazhomi formation), westward orientation perpendicular to westward winds, long slopes, high drainage density, and human interventions are noteworthy.

Conclusion: In the present study, a model based on catastrophe theory was utilized to mitigate uncertainties associated with decision-making and data classification. In this model, the weights of criteria are determined by the internal mechanism of the system, which has a mathematical nature. The results of the model are repeatable and generalizable, thereby reducing decision-making uncertainties and enabling the integration of multiple quantitative and qualitative data. The research findings indicate the effectiveness of catastrophe theory functions based on GIS in spatial assessment and zoning of landslide risk in the study area. The alignment of landslide field mapping with landslide risk zoning maps and receiver operating characteristic (ROC) curves indicates a model accuracy of approximately 90%. The model results show that slope, precipitation, elevation, and geology are the most important factors influencing landslide occurrence in the Zemkan watershed. A significant portion of the study area falls into high and very high susceptibility classes of landslides, indicating a considerable potential for landslides in the Zemkan watershed and the need for planning to mitigate the risks associated with landslide occurrence.