Satellite-Derived Bathymetry in Tropical Shallow Waters: Methods, Sensors, Accuracy, and Applications

Abstract

Accurate bathymetry is fundamental to coastal navigation, habitat assessment, shoreline management, hydrodynamic modelling, and fisheries planning, yet conventional hydrographic surveys remain costly and difficult in remote or very shallow tropical waters. Satellite-derived bathymetry (SDB) provides a complementary approach by estimating water depth from optical reflectance, wave properties, or spaceborne laser altimetry. This review synthesizes methodological developments in SDB with emphasis on tropical shallow-water environments and studies published from 2018 to August 2026, while retaining foundational algorithms required to explain current practice. The literature was organized according to sensor, preprocessing, calibration data, retrieval model, validation metric, and environmental limitation. Sentinel-2 has become the principal freely available optical source because its 10-m visible bands, five-day revisit, and global coverage are well matched to reef flats, lagoons, and island coasts. Landsat remains important for historical reconstruction, whereas PlanetScope and very-high-resolution imagery improve representation of narrow channels and small geomorphic features. ICESat-2 has substantially changed SDB by supplying independent along-track depth observations that can calibrate empirical and machine-learning models where field soundings are unavailable. Band-ratio and log-linear approaches remain robust baselines, but Random Forest, Support Vector Regression, neural networks, and multi-temporal fusion increasingly improve performance in complex waters. Accuracy nevertheless depends more strongly on optical depth, turbidity, bottom heterogeneity, sun glint, tide, temporal mismatch, and reference-data quality than on algorithm complexity alone. For tropical archipelagic regions such as Indonesia, an operational workflow should combine aquatic atmospheric correction, glint and cloud screening, tidal normalization, spatially independent validation, and explicit uncertainty reporting. SDB should therefore be treated as a scalable complement to hydrographic surveying rather than an unrestricted replacement for echo sounding.