EXTRACTION OF SLOPE VARIATION EXTENT USING OPTICAL SENSORS AND SAR FOR THE MANASLU CIRCUIT TRAIL, NEPAL

Abstract

Frequent slope deformation areas induced by intense monsoonal rainfall in Nepal threaten trekking routes and local infrastructure, especially in high-altitude regions like the Manaslu Circuit Trail. This study evaluates the effectiveness of satellite-based landslide detection methods using both optical and synthetic aperture radar (SAR) data. Two optical-based techniques (land cover change classification and Normalized Difference Vegetation Index (NDVI)- Gray Size Index (GSI) differencing) and five SAR-based approaches (texture analysis, Normalized Difference Polarization Index (NDPI) differencing, dual-polarization image interpretation, coherence difference analysis, and interferogram interpretation) were applied to pre- and post-monsoon Sentinel-1 and Sentinel-2 datasets. A field survey conducted in September 2024 provided ground-truth data for validating remote sensing results. Among all methods, optical approaches produced more reliable detection results, with NDVI–GSI differencing demonstrating the highest accuracy. SAR-based techniques faced challenges due to terrain-induced distortions and coherence loss, though coherence difference analysis showed relative promise. Limitations in optical imagery, such as misclassification caused by seasonal land use changes, suggest that SAR data can complement optical analysis to improve reliability. This study concludes that a hybrid approach integrating optical and SAR datasets enhances landslide detection in mountainous regions.