Abstract
Understanding the dynamics of soil moisture in rapidly evolving coastal landscapes is crucial for assessing hydrological variability, ecosystem responses, and agricultural resilience. The western Sundarbans Delta, influenced by monsoon rainfall, tidal movements, and diverse vegetation, poses significant challenges for conventional monitoring. To bridge this gap, we developed a multi-temporal retrieval system utilizing Sentinel-1 C-band SAR data processed within the Google Earth Engine (GEE) environment. This system seamlessly integrates radar backscatter with vegetation indices and precipitation data. Monthly soil-moisture maps for 2021 were generated and rigorously validated against a comprehensive network of in-situ measurements. The retrievals effectively captured the region's hydrological seasonality, showcasing pronounced wet conditions during the monsoon and significant moisture depletion prior to summer. The strong correlation between radar-derived soil moisture and field observations (r
= 0.90) underscores the efficacy of the SAR-based change-detection methodology in complex coastal environments. Sensitivity analyses highlighted the reliability of radar backscatter as an indicator of near-surface soil water, particularly in areas with limited vegetation cover or roughness interference. Correlation assessments revealed that soil moisture has a more coherent relationship with NDVI than with rainfall, underscoring the role of vegetation as a stabilizing intermediary that integrates water availability over time. Lag analysis provided additional insights, revealing that vegetation responses extend beyond immediate precipitation events, reflecting broader ecosystem-level water regulation mechanisms. This methodological framework provides a scalable and transferable strategy for soil-moisture monitoring in vulnerable deltas, offering crucial insights for sustainable agriculture, ecological conservation, and climate adaptation planning. It also contributes to UN SDG 13, SDG 14, and SDG 15.