Cloud-Based Ecological Monitoring of Post-Flood Vegetation Dynamics in the Himalayan Melamchi River with Sentinel-2 and Google Earth Engine

Extreme hydrometeorological events increasingly threaten Himalayan mountain systems, where sediment-laden flash floods cause severe ecological disruption. This study develops a reproducible Google Earth Engine–based framework using Sentinel-2 imagery to map flood extent and assess long-term vegetation dynamics along Nepal’s Melamchi River following the catastrophic June 2021 debris flood. Among tested spectral indices (NDVI, NDWI, MNDWI, WRI), NDVI achieved the highest flood detection accuracy (83%, Kappa = 0.76) with 11% commission error, outperforming conventional water indices under highly turbid conditions. Annual mean NDVI declined from 0.83 pre-flood to 0.63 in 2022, with only partial recovery by 2024. Pixel-level slope analysis (2019–2024) revealed persistent vegetation degradation across ~90% of the affected area, particularly in forested zones, while only 3% showed recovery. The results demonstrate prolonged ecological impacts of sediment-rich floods and present a cost-effective, transferable framework for monitoring cascading hazards and guiding ecosystem restoration in data-scarce mountain regions.