Research Center for Geoscience and Social Studies, Hakim Sabzevari University, Sabzevar, Iran , m.boroughani@hsu.ac.ir
Abstract: (12 Views)
- Introduction
Landslide is one of the most important natural hazards in mountainous and steep areas, which causes significant human, financial, and environmental losses. Given the numerous complexities involved in landslide formation, accurate and multifactor modeling of this phenomenon is of particular importance. Identifying and analyzing the effective factors can help improve prediction, reduce risk, and facilitate optimal management of this hazard. Landslides are a phenomenon that has significant impacts on the environment and human infrastructure. This phenomenon leads to disruption of the water and soil system, destruction of roads, highways, residential areas and engineering structures. Landslides also destroy vegetation, agricultural lands and increase sedimentation in nature. This process accelerates erosion and causes sediment to be transported behind dams and changes the landscape. This phenomenon transforms the natural landscape and affects local ecosystems. Landslides are one of the most important natural hazards in mountainous and steep areas, causing significant loss of life, property, and environmental damage. The Tarom-Khalkhal basin is known as one of the areas prone to landslides due to its geological characteristics, steep slopes, heavy rainfall, and climate change. The increase in human activities such as land use change, road development, and construction operations without proper planning has increased the sensitivity of this area to landslides.
- Methodology
The present study aims to assess vulnerability using multiple modeling approaches to zone landslide risk in the Tarom-Khalkhal basin. In this study, 13 main factors were considered, including slope, elevation, slope direction, land use, distance from fault, geology, distance from waterway, distance from road, surface curvature, precipitation, soil, vegetation cover index (NDVI), and topographic wetness index (TWI). This study was conducted by simultaneously applying two advanced approaches: hybrid modeling based on deep learning and cluster analysis (MFSDSM) and a model based on multiple factor simulation using dynamic systems modeling (DLCAM) in landslide analysis and prediction in the Tarom-Khalkhal basin. The efficiency of the presented methods was analyzed and evaluated using the parameters of coefficient of determination, mean square error, and Nash-Sutcliffe index.
- Results
The results indicate that the slope factor has the highest importance with a weight of 0.22, which indicates the fundamental role of this factor in the landslide phenomenon. Slope, as a direct physical factor, has a significant impact on the dynamics and stability of the land; the greater the slope of the land, the greater the probability of landslides, because the sloping surface has a greater tendency to slide. After that, the height and soil type factors are in the next rank with weights of 0.13 and 0.14. Height indicates factors related to the intensity of the influence of other factors such as slope and rain. Soil type also plays a decisive role in the sensitivity of the land surface to water, density and stability, considering its physical and chemical properties. The factors of slope direction, TWI index, NDVI index, geology and land use with weights in the range of 0.07 to 0.10, have a moderate to small role in modeling. The factors of distance from the road and curvature have the lowest importance (0.04 to 0.03). In fact, these factors may have a short-term and limited role in a specific range or under specific conditions. The results indicate that the combination of factors such as slope and precipitation, especially in high-altitude areas, have the greatest impact on landslide occurrence. Steep slopes and heavy precipitation provide suitable conditions for landslide occurrence in the basin. The DLCAM model uses deep learning algorithms and cluster analysis to identify complex patterns in spatial data. This model has been able to simulate areas with high landslide risk with high accuracy. In the MFSDSM algorithm, high risk zones (19.21) and very high risk zones (59.33) correspond to the western, eastern, and southern parts. In the DLCAM model, high risk zones (23.5) and very high risk zones (48.66) correspond to the northern, southern, and western areas. The performance analysis of the algorithms with statistical indices indicates the superiority of the MFSDSM model with an RMSE of 271, a coefficient of determination of 0.91, and an NSE index of 0.81 over DLCAM.
- Discussion & Conclusions
These results confirm the ability of both algorithms to accurately identify areas at risk and are effective tools in managing natural risks. Finally, landslide modeling in the Tarom-Khalkhal basin is an effective tool for environmental protection, improving water resources management, and achieving sustainable development.
Type of Study:
Research |
Received: 2025/09/23
Send email to the article author
