Mapping urban expansion and landslide susceptibility over Solan District of Himachal Pradesh, India, using Random Forest approach integrating LULC dynamics and Sentinel-1 data

The Himalayan region is becoming increasingly susceptible to landslides due to unplanned construction and rapid urbanization, particularly in Uttarakhand and Himachal Pradesh, India. Road widening and infrastructure development on steep, unstable slopes have triggered land subsidence in Joshimath (Awasthi et al., 2024) and in the Char Dham Highway landslide. Landslides are also occurring in Himachal Pradesh's Solan district. In order to better understand how rapid changes in land-use and land-cover (LULC) are increasing the risk of landslides in this vulnerable and urbanizing district of Himachal Pradesh, this study proposes an integrated Remote Sensing (RS) and Earth Observation (EO). In Google Earth Engine (Gorelick et al., 2017), multi-temporal Sentinel-2 images from 2019 to October 2025 were analysed using cloud-masking and spectral indices (NDVI, NDBI, and NDWI) to precisely identify land cover types. Change detection analysis using this processed dataset showed that built-up areas increased by 11% between 2019 and 2024 and a remarkable 16% growth between 2024 and October 2025, indicating increased urbanisation during the most recent period (2024-2025). The analysis identifies a transition in urbanization areas. In the LULC change map, we observed that Baddi, Nalagarh and Barotiwala constitute established urban centers, however, the 2024-2025 duration shows maximum expansion within the Chamba (northeast), Arki (central-east), and Kasauli (southeast) regions. The northeastern and southeastern regions of the Solan district are emerging as the new urban expansion zones. Our ongoing research focuses on developing a Random Forest-based landslide susceptibility model that combines multi-sensor Earth Observation data with these LULC dynamics through an optical–SAR fused framework. In order to develop a framework for identifying high-risk areas, we are investigating Sentinel-1 SAR images (the GRD products) to measure coherence and backscatter change and combining with topographic parameters derived from SRTM (slope, curvature, aspect) (Sharma et al., 2024). To improve this optical-SAR fused  model accuracy, additional geological data from the Geological Survey of India and rainfall data from CHIRPS (Climate Hazards Group Infrared Precipitation with Stations) will be incorporated.  This integrated method allows for a quantitative assessment of susceptibility in this vulnerable Himalayan terrain by correlating land-use transitions with slope instability indicators.