Extended Abstract
Introduction
Landslides are the sliding of a mass of soil or rock or a combination of them on a slope due to gravity, topographic factors, and human activities, which can occur suddenly and locally (Das et al., 2012; Kumar et al., 2017). Landslides pose serious intra-regional and extra-regional risks in mountainous regions around the world (Goetz et al., 2011), and cause serious damage to settlements, railways, power lines, roads, gas pipelines, and agricultural lands (Lin et al., 2012; Yana et al., 2016; Mousavi and Niazi, 2016). In addition, this phenomenon also causes damage by wasting large amounts of fertile soil and filling dam reservoirs downstream (Gutierrez et al., 2015). Among the areas subject to this phenomenon, Iran has always been exposed to various types of landslides due to the presence of the Alborz mountain range in the north and the Zagros mountain range in the northwest to southeast, slopes above 45 degrees, formations with sticky clay soils, and intermittent rain and snow fall, and is no exception to this issue (Rifahi, 2009).
Materials and Methods
The data used in this study include Sentinel-1 radar images from 2017-2021 (Table 1) and 8 geomorphological parameters including slope, slope direction, elevation, distance from road, distance from river, land use, distance from fault, and geological map, which were processed and analyzed using the Analysis Network Model (ANP) and Geographic Information Systems (GIS). Sentinel-1 images are one of the important sources of radar data for interferometric analysis (InSAR). These satellites, developed by the European Space Agency (ESA), are equipped with synthetic aperture radar (SAR) that can measure changes in the earth's surface with high accuracy. In this study, after collecting the required data and information, data analysis was carried out in three stages, each of which is described below:
Research Method
1- Evaluation of the rate of vertical displacement changes
In the interferometric analysis process (InSAR) using Sentinel-1 data, first two radar images of the study area are selected, one before and the other after the desired event (such as an earthquake or landslide). These images must overlap in time and space. Then, using the interferometric technique, the phase difference between these two images is calculated, which indicates minor changes in the height of the ground surface. Subsequently, various processes are performed such as removing atmospheric noise, smoothing the data, and converting the phase difference into accurate displacements. Finally, the results are presented in the form of displacement maps or elevation models that are used to analyze land surface changes. In this section, the results of the rate of change of land surface displacements in the study area are examined in order to evaluate and identify landslide-prone areas.
2- Potential assessment of landslide-prone areas
In this section, first, in order to prepare a landslide vulnerability map using the multi-criteria decision-making method, criteria that indicate the vulnerability of the area to landslides were used. The criteria used in this section include height, slope and slope direction, land use, distance from the river, distance from the fault, geology, and distance from the road. In this stage, the relationships between the criteria were determined using the EDPSIR method, and then the weight of each criterion was determined using the Analytical Network Process (ANP) method through expert opinion (questionnaire). Finally, after fuzzing the layers, by combining them in the Geographic Information System (GIS) environment based on the weight of each criterion, a landslide vulnerability map was produced at the level of the study area.
3- Final analysis of the results of the method
After evaluating the changes in ground surface displacements and identifying areas prone to landslides, the hazardous zones in the study area were finally identified. For this purpose, first, images with an appropriate time interval from the landslide event were selected to accurately record and analyze the ground surface displacements. For example, for the image of July 22, 2018 in the area, the previous image was recorded on July 4, 2018 and the next image was recorded on August 2, 2018. This 18 to 20-day interval between images is a good choice for accurate analysis of ground surface changes. Considering the strength of the earthquake and weather conditions, these images can provide appropriate data for investigating landslides. Finally, a landslide hazard zoning map was prepared using five hazard classes (very low, low, medium, high, and very high). Different colors on the map represent the level of landslide hazard in each part of the area. This map helps identify areas prone to landslides and is of great importance for environmental risk management and urban and rural planning. The map is produced based on network analysis models and environmental data, and the percentage of areas under different risks is well defined.
Discussion and Results
According to the objectives of this study, first the amount of vertical displacements in the study area has been investigated. Then the landslide-prone areas have been identified and evaluated, which are described below:
1- Evaluation of the amount of vertical displacement changes
2- Evaluation of landslide-prone areas
3- Analysis of results
Figure (12) shows the map of landslide-prone areas using five risk classes (very low, low, medium, high and very high). Different colors on the map represent the level of landslide risk in each part of the area. This map helps identify landslide-prone areas and is of great importance for environmental risk management and urban and rural planning. The map was generated based on analytical models and environmental data, and the percentage of areas at different risk is well defined. Very low (23.95%): This class, with 195,567 pixels, covers about 3055.73 hectares of the area, which is approximately 24% of the total area. These areas have the lowest risk of landslides and are usually located in areas with low slopes and away from faults. Low (21.06%): This class, with 171,939 pixels, covers about 2686.55 hectares of the area, which is 21% of the total area.