Monitoring surface exposure and land cover shifts on barrier islands using the bare soil index

This study presents a cloud-based remote sensing approach for monitoring geomorphic surface dynamics of barrier islands in the Cigu District, Tainan, Taiwan. Using Sentinel-2 Surface Reflectance data (2018-2025), we conducted a time series analysis of the Bare Soil Index (BSI), Normalized Difference Water Index (NDWI), and Normalized Difference Vegetation Index (NDVI) to detect long-term trends in surface exposure. Linear regression of BSI, NDWI, and NDVI values was applied to identify potential erosion and accretion areas. The combined use of these indices captures the complexity of biophysical processes in the dynamic interface between land and sea. This approach identified a relatively balanced extent of erosion and accretion across the barrier island. The analysis of bare soil dynamics showed that a substantial portion of the area remained relatively unchanged, while other parts experienced either an increase or decrease in exposure. Additionally, the remaining areas exhibited variable patterns, indicating a high degree of spatial variability during the study period. The indices were then used to generate land cover map using Support Vector Machine classifier. The results showed very good classification performance, with high overall accuracy and strong agreement indicated by kappa coefficients for all indices: combined BSI and NDWI, NDWI alone, and NDVI. These results confirm the robustness of BSI in detecting exposed, sediment-dominated zones and demonstrate its value as a scalable, spectral proxy for geomorphic monitoring in data-limited coastal systems. Ultimately, this methodology offers a lightweight, cloud-native framework for tracking landscape change in barrier islands, supporting adaptive coastal management under climate-driven stressors such as sea-level rise, sediment imbalance, and ecosystem transformation.