1- Kharazmi University
2- Kharazmi University ,karamA.khu@yahoo.com
2- Kharazmi University ,
Abstract: (3374 Views)
Abstract
The catchment area of Alamut River in Qazvin province is witnessing numerous landslide hazards and landslides every year, which cause significant economic and sometimes life-threatening losses. Diagnosing the unstable areas of slopes through soil texture characteristics is a difficult task due to the difficulties of obtaining soil samples in mountainous areas. For this reason, in the present study, by using Sentinel A1 radar data, by determining the percentage of clay and sand in the soil, the soil texture map at the depths of 10, 60, 100 and 200 cm with two random forest (RF) and support vector machine (SVM) algorithms was produced in the eastern Alamut region, which was verified with soil profile samples. The results indicated that the Kappa index was more accurate in the RF model at three depths of 10, 60 and 100 cm. Then, by extracting the soil moisture map from Sentinel 2 data, at the same time as examining the internal friction angle of the types of soils in the region, comparing the slope and profile of the slopes and the shape of the convex (divergent) and concave (convergent) slopes, the unstable areas of slope movements, RF and SVM models were specified and validated with GPS data, field visits and Google Earth. Research findings show that the instability map resulting from the RF model has a higher accuracy (AUC=0.93) than the instability map resulting from the SVM model (AUC=0.90) and there is more instability in areas with medium to high slope and with soil texture of sandy clay loam and sandy loam. . This method has many advantages in preparing the soil texture map, determining the unstable areas of the slopes against mass movements and landslides, determining the vulnerable areas in mountainous areas and reducing financial and human losses.
The catchment area of Alamut River in Qazvin province is witnessing numerous landslide hazards and landslides every year, which cause significant economic and sometimes life-threatening losses. Diagnosing the unstable areas of slopes through soil texture characteristics is a difficult task due to the difficulties of obtaining soil samples in mountainous areas. For this reason, in the present study, by using Sentinel A1 radar data, by determining the percentage of clay and sand in the soil, the soil texture map at the depths of 10, 60, 100 and 200 cm with two random forest (RF) and support vector machine (SVM) algorithms was produced in the eastern Alamut region, which was verified with soil profile samples. The results indicated that the Kappa index was more accurate in the RF model at three depths of 10, 60 and 100 cm. Then, by extracting the soil moisture map from Sentinel 2 data, at the same time as examining the internal friction angle of the types of soils in the region, comparing the slope and profile of the slopes and the shape of the convex (divergent) and concave (convergent) slopes, the unstable areas of slope movements, RF and SVM models were specified and validated with GPS data, field visits and Google Earth. Research findings show that the instability map resulting from the RF model has a higher accuracy (AUC=0.93) than the instability map resulting from the SVM model (AUC=0.90) and there is more instability in areas with medium to high slope and with soil texture of sandy clay loam and sandy loam. . This method has many advantages in preparing the soil texture map, determining the unstable areas of the slopes against mass movements and landslides, determining the vulnerable areas in mountainous areas and reducing financial and human losses.
Keywords: Slope Instability, Soil Texture, Remote Sensing, Random Forest, Support Vector Machine, Eastern Alamut.
Type of Study: Research |
Subject:
Geomorphology
Send email to the article author
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
Journal of Applied researches in Geographical Sciences are fully compliant with open access mandates, by publishing its articles under Creative Commons Attribution 4.0 International License (CC BY 4.0).
