Landslide Susceptibility Mapping in the Vicinity of Dams in Mountainous Areas Using Remote Sensing and Geographic Information System (Case Study: Cheragh Veys Dam, Saqqez County)

The movement of sedimentary layers that occurs due to different reasons such as earthquakes, volcanoes, anthropogenic activities, heavy rainfall or loose soil is called landslide. Landslides sometimes cause irreparable financial and human losses. The aim of this research is to predict the location of the risk of landslides around the newly established Cheragh Veys dam using fuzzy overlap and weighted aggregation models in QGIS environment. For this purpose, 11 factors of, slope, aspect, height, distance to the road, distance to the river, road density, river density, curvature, topographic wetness index, stream power index and vegetation cover were used. The results of the study show that both models performed somewhat similar in detecting areas of low sensitivity to very high sensitivity. In both models, more than half of the area is prone to landslides. The findings of this study can be used by decision-makers and managers to reduce the risks of landslides.
 
Extended Abstract
 
Introduction
Landslides involve the slow to rapid movement of materials down slopes caused by a wide range of natural processes and human activities. Every year, these events result in economic losses and numerous casualties in the province and the country, posing a significant threat to people living in these areas. In general, in Iran and specifically in Kurdistan province, landslides are a constant threat to infrastructure, agriculture, other natural resources, and tourism. Local administrations are sometimes under financial and logistical pressure to address these issues. Most landslides in Iran are caused by heavy rainfall. Certain areas, such as those around newly established dams and roads in mountainous regions, are particularly susceptible to landslides and may become more unstable due to human interventions. This research addresses the same issue and aims to use remote sensing techniques and geographic information systems to zone landslide-sensitive areas around the newly established Cheragh Veys Saqqez dam. Over the past few decades, there have been tremendous advances in remote sensing science and geographic information systems, facilitating the preparation of landslide susceptibility maps. These maps, as comprehensive resources, can be used by policymakers and decision-makers to mitigate financial losses caused by landslides. A wide range of models and methods, including hierarchical analysis, the entropy index, geographically weighted principal components, machine learning methods, artificial neural networks, support vector machines, fuzzy logic, and others, have been proposed for preparing landslide susceptibility maps. It should be noted that since the Cheragh Weis dam is newly established, no studies have been conducted in this regard for the area under study so far. This present study, a comprehensive effort, will zone the areas prone to landslides around the aforementioned dam for the first time. Due to the complexity of predicting landslide risks, many researchers have proposed a hybrid model approach. Among these, the alternative decision tree (ADTree) is noteworthy. Both simple and hybrid models have been employed in studies of floods, fires, droughts, gully erosion, land subsidence, and landslides. This study uses fuzzy overlay and weighted aggregation models to create a map of landslide-prone areas around the new Cheragh Weis dam.
 
Methodology
The first step in landslide studies is identifying historical locations and the factors that influence landslides. This research focuses solely on zoning areas sensitive to landslides. In the present study, 11 factors influencing landslides were selected based on previous research, specialized knowledge, and the physical characteristics of the study area. Additionally, two models—fuzzy overlap and weighted aggregation—were used to produce a landslide susceptibility map. A digital elevation model (DEM) was employed within a geographic information system (GIS) environment to generate the 11 factors. After reclassification, these factors were used to create the landslide susceptibility map. It should be noted that the spatial resolution of the 11 selected factors is 30 meters, which is suitable for producing a landslide susceptibility map of the study area. Since the DEM, from which most of the layers were derived, had a resolution of 30 meters, its derivatives naturally have the exact spatial resolution. Landsat satellite imagery, with a multi-spectral band resolution of 30 meters, was also used to create the vegetation map. Additionally, the road layer was extracted from Google Earth with high accuracy, and the distance from roads and road density maps were resampled to produce a suitable final map with a spatial resolution of 30 meters. The layers extracted from the DEM in the GIS included slope direction, slope, elevation, topographic wetness index, stream power index, and curvature. Furthermore, river density, road density, distance from roads, distance from rivers, and vegetation were extracted in the GIS to produce the final map. In this study, the fuzzy overlap model and the weighted aggregation model were used to achieve the desired objectives.
 
Results and Discussion
The research findings indicate that in the weighted accumulation model, about 10 square kilometers of the area fall into the low sensitivity group, while 17 square kilometers are at very high risk. Additionally, 26 square kilometers of land around the Cheragh Weis Dam are classified as having medium sensitivity. Finally, a large and significant portion of the area, totaling 32 square kilometers, belongs to the high-sensitivity group. The quantitative results demonstrate that the area is highly prone to landslides and should be carefully monitored by authorities to prevent potential financial losses and risks to life. The weighted overlap model also predicts a similar trend with some variations. In this model, about 2 square kilometers have been added to the areas with low sensitivity. The medium sensitivity group covers 25 square kilometers. Meanwhile, 34 square kilometers of land around the dam is associated with a high risk of landslides. In the fuzzy overlap model, 3 square kilometers of land are exposed to a very high risk of landslide occurrence, which is less than in the weighted accumulation model.
 
Conclusion
Landslides, if they occur in residential, tourism, and agricultural areas, can cause significant destruction. Without identifying high-risk areas and under the right conditions, they can result in irreparable financial and human losses. Landslide risk zoning maps enable organizations and officials to monitor high-risk areas in a targeted manner and, if necessary, implement preventive measures to avert accidents. Globally, landslides cause millions of dollars in financial losses yearly, resulting in numerous injuries and fatalities. This risk also leads to economic and human losses in our country. The results of this study can assist authorities in securing landslide-prone areas and preventing potential damages. One limitation of the current research was the lack of access to the southern parts of the dam due to the flooding of the connecting bridges after the dam was drained. Future research could utilize InSAR or DInSAR models to identify landslides, providing better and more accurate modeling for such studies.