Prediction of Landslide Susceptibility in the Polroud Watershed Using Machine Learning Methods

Abstract :

Landslides are among the most significant geomorphological hazards, posing a serious threat to natural resources, infrastructure, and human settlements. The aim of this study was to perform landslide susceptibility zonation in the Polrud Watershed using machine learning algorithms and to evaluate their performance comparatively. For this purpose, three machine learning algorithms—Random Forest (RF), Support Vector Machine (SVM), and Decision Tree (DT)—were employed, and 19 landslide conditioning factors, categorized into topographic, hydrological, and environmental–anthropogenic groups, were utilized in the modeling process. The landslide inventory data were randomly divided into training (70%) and testing (30%) datasets for model evaluation. The performance of the models was evaluated using metrics including the Area Under the Curve (AUC), overall accuracy, sensitivity, specificity, precision, correlation coefficient (R), F1-score, CV-ROC, and Cohen's kappa coefficient. The results showed that the RF algorithm outperformed the other models in identifying the spatial pattern of landslide susceptibility and, by providing a more realistic distribution of high-susceptibility zones, classified approximately 24% of the watershed area as high susceptibility. Accordingly, the AUC value of 0.75 and CV-ROC of 0.97 indicate appropriate model stability, although the overall accuracy of 0.62 reveals the impact of data imbalance. In contrast, the SVM model showed a greater tendency to classify areas as low susceptibility, whereas the DT model exhibited lower stability. Variable importance analysis confirmed the prominent role of the morphometric indices TRI and TPI, as well as the hydrological variables SPI and CN, in controlling slope instability. Overall, the research results confirm the superiority of the RF algorithm and its greater efficiency and reliability in landslide susceptibility zonation.

References:

باباربیع، نگار، لرستانی، قاسم، و اسماعیلی، رضا (1404). پهنه‌بندی وقوع خطر زمین‌لغزش با استفاده از مدل‌های جنگل تصادفی و ماشین بردار. پژوهش‌های دانش زمین، 16(1)، 152-168. doi.org/10.48308/esrj.2025.236839.1234
بابلی مؤخر، حمید، تقیان، علی رضا، و شیرانی، کورش (1397). ارزیابی نقشه پهنه‌بندی حساسیت زمین‌لغزش با استفاده از روش تلفیقی فاکتور اطمینان و رگرسیون لجستیک با به‌کارگیری شاخص‌های ژئومورفومتریک. پژوهش‌های ژئومورفولوژی کمّی، 7(3)، 91-116. dor.isc.ac/dor/20.1001.1.22519424.1397.7.3.6.4
حجازی، سید اسدالله (1393). پهنه‌بندی خطر زمین‌لغزش در حوضه گویجه بل اهر با استفاده از سیستم اطلاعات جغرافیایی (GIS). جغرافیا و برنامه‌ریزی، 18(50)، 135-152.
دسترنج، علی، و کریمی سنگچینی، ابراهیم (1401). پیش‌بینی حساسیت وقوع زمین‌لغزش با الگوریتم یادگیری ماشین بیشینه آنتروپی (حوزه آبخیز بار نیشابور). پژوهش‌های دانش زمین، 13(51)، 76-96. doi.org/10.48308/esrj.2022.102965
دسترنج، علی، کریمی سنگچینی، ابراهیم، و نور، حمزه (1403). ارزیابی ریسک زمین‌لغزش در حوزه آبخیز بار نیشابور، استان خراسان رضوی. علوم و مهندسی آبخیزداری ایران، 18(67). doi.org/10.22034/18.67.4
ساداتی، سید حمید، موسوی، سید رمضان، وهاب‌زاده کبریا، قربان، و روشان، سید حسن. (1404). ارزیابی مدل‌های جنگل تصادفی و ماشین بردار پشتیبان در نقشه‌برداری خطر زمین‌لغزش (مطالعه موردی: حوضه تجن، استان مازندران). مجله مخاطرات طبیعی محیطی، 14(45)، 133-154. doi.org/10.22111/jneh.2025.50031.2071
طالبی، علی، نفرزادگان، علیرضا، و ملکی نژاد، حسین (1388). مروری بر مدل‌سازی تجربی و فیزیکی زمین‌لغزش‌های ناشی از بارندگی. پژوهش‌های جغرافیای طبیعی (پژوهش‌های جغرافیایی)، 41(70)، 45-64.
فیض‌نیا، سادات، حسن‌زاده، محمد، و احمدی، حسن (1380). پهنه‌بندی خطر زمین‌لغزش حوزه آبخیز شلمان رود در استان گیلان. منابع طبیعی ایران، 54(3)، 207-220.
مددی، عقیل، پیروزی، الناز، و فعال‌نذیری، مهدی (1399). ارزیابی مقایسه‌ای الگوریتم‌های تصمیم‌گیری چندمعیارة MABAC و CODAS در پهنه‌بندی خطر زمین‌لغزش، نمونة پژوهش: شهرستان کوثر. جغرافیا و برنامه‌ریزی محیطی، 31(4)، 1-24. doi.org/10.22108/gep.2020.124723.1348
نظریانی، نسترن، و فلاح، اصغر (1402). مدل‌سازی خطر وقوع زمین‌لغزش با استفاده از داده‌کاوی در جنگل‌های هیرکانی. پژوهش‌نامه مدیریت حوزه آبخیز، 14(27)، 123-134. doi.org/2010.61186/jwmr.14.27.123
Abdo, H. G., & Richi, S. M. (2024). Application of machine learning in the assessment of landslide susceptibility: A case study of mountainous eastern Mediterranean region, Syria. Journal of King Saud University-Science, 36(5), 103174. doi.org/10.1016/j.jksus.2024.103174.
Akbar, T. A., & Ha, S. R. (2011). Landslide hazard zoning along Himalayan Kaghan Valley of Pakistan—by integration of GPS, GIS, and remote sensing technology. Landslides, 8(4), 527–540. doi.org/10.1007/s10346-011-0260-1.
Allegri, E., Zanetti, M., Torresan, S., & Critto, A. (2024). Pluvial flood risk assessment for 2021–2050 under climate change scenarios in the Metropolitan City of Venice. Science of the Total Environment, 914, 169925. doi.org/10.1016/j.scitotenv.2024.169925.
Arabameri, A., Saha, S., Roy, J., Chen, W., Blaschke, T., & Tien Bui, D. (2020). Landslide susceptibility evaluation and management using different machine learning methods in the Gallicash River Watershed, Iran. Remote Sensing, 12(3), 475. doi.org/10.3390/rs12030475.
Cui, J., Kou, P., Tao, Y., Jin, Z., Huang, Y., & Zhang, J. (2024). Enhanced Landslide Susceptibility Mapping Using Machine Learning and InSAR Integration: A Case Study in Wushan County, Three Gorges Reservoir Area, China. doi.org/10.21203/rs.3.rs-4766128/v1.
Fatemi Aghda, S. M., Bagheri, V., & Razifard, M. (2018). Landslide susceptibility mapping using fuzzy logic system and its influences on mainlines in lashgarak region, Tehran, Iran. Geotechnical and Geological Engineering, 36(2), 915–937. doi.org/10.1007/s10706-017-0365-y.
Haftlang, K. K., & Lang, K. K. H. (2003). The book of Iran: a survey of the geography of Iran. Alhoda UK.
Kühnl, M., Sapena, M., Wurm, M., Geiß, C., & Taubenböck, H. (2023). Multitemporal landslide exposure and vulnerability assessment in Medellín, Colombia. Natural Hazards, 119(2), 883–906. doi.org/10.1007/s11069-022-05679-z.
Lin, L., Lin, Q., & Wang, Y. (2017). Landslide susceptibility mapping on a global scale using the method of logistic regression. Natural Hazards and Earth System Sciences, 17(8), 1411–1424. doi.org/10.5194/nhess-17-1411-2017, 2017.
Meles, M. B., Younger, S. E., Jackson, C. R., Du, E., & Drover, D. (2020). Wetness index based on landscape position and topography (WILT): Modifying TWI to reflect landscape position. Journal of Environmental Management, 255, 109863. doi.org/10.1016/j.jenvman.2019.109863.
Mirzaei, S., Vafakhah, M., Pradhan, B., & Alavi, S. J. (2021). Flood susceptibility assessment using extreme gradient boosting (EGB), Iran. Earth Science Informatics, 14(1), 51–67. doi.org/10.1007/s12145-020-00530-0.
Mojaddadi, H., Pradhan, B., Nampak, H., Ahmad, N., & Ghazali, A. H. bin. (2017). Ensemble machine-learning-based geospatial approach for flood risk assessment using multi-sensor remote-sensing data and GIS. Geomatics, Natural Hazards and Risk, 8(2), 1080–1102. doi.org/10.1080/19475705.2017.1294113.
Norallahi, M., & Seyed Kaboli, H. (2021). Urban flood hazard mapping using machine learning models: GARP, RF, MaxEnt and NB. Natural Hazards, 106, 119–137. doi.org/10.1007/s11069-020-04453-3.
Pandey, A., Dabral, P. P., Chowdary, V. M., & Yadav, N. K. (2008). Landslide hazard zonation using remote sensing and GIS: a case study of Dikrong river basin, Arunachal Pradesh, India. Environmental Geology, 54(7), 1517–1529. doi.org/10.1007/s00254-007-0933-1.
Patel, M., Chavda, M., Patel, R., Goswami, A., & Mevada, J. (2024). Predicting Landslide Using Machine Learning Techniques. ITM Web of Conferences, 65, 3012. doi.org/10.1051/itmconf/20246503012.
Pradhan, B. (2011). Use of GIS-based fuzzy logic relations and its cross application to produce landslide susceptibility maps in three test areas in Malaysia. Environmental Earth Sciences, 63(2), 329–349. doi.org/10.1007/s12665-010-0705-1.
Sahin, E. K., Colkesen, I., & Kavzoglu, T. (2020). A comparative assessment of canonical correlation forest, random forest, rotation forest and logistic regression methods for landslide susceptibility mapping. Geocarto International, 35(4), 341–363. doi.org/10.1080/10106049.2018.1516248.
SELLAMI, E. M., & Rhinane, H. (2024). Google Earth Engine and Machine Learning for Flash Flood Exposure Mapping—Case Study: Tetouan, Morocco. Geosciences (Switzerland), 14(6). doi.org/10.3390/geosciences14060152.
Shahabi, H., Ahmadi, R., Alizadeh, M., Hashim, M., Al-Ansari, N., Shirzadi, A., Wolf, I. D., & Ariffin, E. H. (2023). Landslide susceptibility mapping in a mountainous area using machine learning algorithms. Remote Sensing, 15(12), 3112. doi.org/10.3390/rs15123112.
Towfiqul Islam, A. R. M., Talukdar, S., Mahato, S., Kundu, S., Eibek, K. U., Pham, Q. B., Kuriqi, A., & Linh, N. T. T. (2021). Flood susceptibility modelling using advanced ensemble machine learning models. Geoscience Frontiers, 12(3). doi.org/10.1016/j.gsf.2020.09.006
Wang, Y., Fang, Z., & Hong, H. (2019). Comparison of convolutional neural networks for landslide susceptibility mapping in Yanshan County, China. Science of the Total Environment, 666, 975–993. doi.org/10.1016/j.scitotenv.2019.02.263.
Yu, H., Pei, W., Zhang, J., & Chen, G. (2023). Landslide susceptibility mapping and driving mechanisms in a vulnerable region based on multiple machine learning models. Remote Sensing, 15(7), 1886. doi.org/10.3390/rs15071886.
Zare, M., Pourghasemi, H. R., Vafakhah, M., & Pradhan, B. (2013). Landslide susceptibility mapping at Vaz Watershed (Iran) using an artificial neural network model: a comparison between multilayer perceptron (MLP) and radial basic function (RBF) algorithms. Arabian Journal of Geosciences, 6(8), 2873–2888. doi.org/10.1007/s12517-012-0610-x
Zhuo, L., Huang, Y., Zheng, J., Cao, J., & Guo, D. (2023). Landslide susceptibility mapping in Guangdong Province, China, using random forest model and considering sample type and balance. Sustainability, 15(11), 9024. doi.org/10.3390/su15119024.