Enter your email address
Submit
Write your message
Volume 12, Issue 3 (12-2022)
Disaster Prev. Manag. Know. 2022, 12(3): 253-268 |
Back to browse issues page
Download citation:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
BibTeX | RIS | EndNote | Medlars | ProCite | Reference Manager | RefWorks
Send citation to:
Dastranj A, Shadfar S. Predicting the susceptibility of landslide occurrence in order to manage landslide risk in Bar Neyshabur watershed. Disaster Prev. Manag. Know. 2022; 12 (3) : 1
URL: http://dpmk.ir/article-1-510-en.html
URL: http://dpmk.ir/article-1-510-en.html
1- 1- Corresponding Author: Soil Conservation and Watershed Management Department, khorasan Razavi Agricultural and Natural Resources Research and Education Center, AREEO, Mashhhad. Iran
2- Associate Prof, Soil Conservation and Watershed Management Research Institute (SCWMRI), Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran.
2- Associate Prof, Soil Conservation and Watershed Management Research Institute (SCWMRI), Agricultural Research, Education and Extension Organization (AREEO), Tehran, Iran.
Abstract: (1214 Views)
Background and objective: Landslide susceptibility zoning using different methods is one of the landslide management strategies. The purpose of this study is to evaluate the landslides susceptibility in the Bar watershed in Khorasan Razavi province using the support vector machine (SVM) algorithm.
Method: First, the landslide layer of the area was corrected through field visits and Google Earth satellite imagery. Finally, 73 landslides were identified and the layer related to these landslides was prepared in GIS environment. Of these landslides, 70% were used for model training and the remaining 30% for modeling validation. Then, according to the review of extensive sources and expert opinions, 16 factors affecting the occurrence of landslides in the study area were identified and layers related to these parameters were prepared. Then, using the SVM algorithm in ModEco software environment, a ground sensitivity map was prepared and finally this map was divided into five classes with very high, high, medium, low and very low sensitivity. Finally, the performance of this algorithm was evaluated using the ROC curve.
Findings: Based on the results, the area under curve (AUC) was obtained using training data (0.87) and validation data (0.85). Therefore, SVM algorithm has a very good performance to evaluate landslide sensitivity in the study area. Also, model evaluation based on kappa index showed that slope length (LS) and slope indices have the greatest impact on slope instability in this region. The results of sensitivity zoning also showed that 27.6% of the lands in the region, which were mainly located in the west and northwest of the basin, were in the class with high and very high sensitivity.
Conclusion: Based on the results obtained in the model training and validation stage, indicates that the SVM algorithm is approved in terms of accuracy and validity of modeling. As a result, the landslide susceptibility map obtained in the studied area can be used as a management tool as a cornerstone of landslide research in critical times, land use planning, damage reduction and landslide risks.
Method: First, the landslide layer of the area was corrected through field visits and Google Earth satellite imagery. Finally, 73 landslides were identified and the layer related to these landslides was prepared in GIS environment. Of these landslides, 70% were used for model training and the remaining 30% for modeling validation. Then, according to the review of extensive sources and expert opinions, 16 factors affecting the occurrence of landslides in the study area were identified and layers related to these parameters were prepared. Then, using the SVM algorithm in ModEco software environment, a ground sensitivity map was prepared and finally this map was divided into five classes with very high, high, medium, low and very low sensitivity. Finally, the performance of this algorithm was evaluated using the ROC curve.
Findings: Based on the results, the area under curve (AUC) was obtained using training data (0.87) and validation data (0.85). Therefore, SVM algorithm has a very good performance to evaluate landslide sensitivity in the study area. Also, model evaluation based on kappa index showed that slope length (LS) and slope indices have the greatest impact on slope instability in this region. The results of sensitivity zoning also showed that 27.6% of the lands in the region, which were mainly located in the west and northwest of the basin, were in the class with high and very high sensitivity.
Conclusion: Based on the results obtained in the model training and validation stage, indicates that the SVM algorithm is approved in terms of accuracy and validity of modeling. As a result, the landslide susceptibility map obtained in the studied area can be used as a management tool as a cornerstone of landslide research in critical times, land use planning, damage reduction and landslide risks.
Article number: 1
Type of Study: Research |
Subject:
Special
Received: 2022/06/19 | Accepted: 2022/07/30 | ePublished: 2023/01/3
Received: 2022/06/19 | Accepted: 2022/07/30 | ePublished: 2023/01/3
| Rights and permissions | |
 |
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License. |
Copyright © The Author(s);
This is an open access article distributed under the terms of the Creative Commons Attribution License (CC-By-NC), which permits use, distribution, and reproduction in any medium, provided the original work is properly cited and is not used for commercial purposes.
Contact Information
