A Potential Future for SDB: Addressing Turbidity, Scene Selection, and Labor with the Sentinel-2 Mission

Trabajo presentado en el AGU Fall Meeting 2021, celebrado en New Orleans, LA (Estados Unidos), del 6 al 17 de diciembre de 2021Satellite-Derived Bathymetry (SDB) has significant potential to enhance our knowledge of Earth¿s coastal regions. However, SDB still has limitations when applied to the turbid, but optically shallow, nearshore regions that encompass large areas of the world¿s coastal zone. Environments with transient turbidity are more problematic because turbidity severely biases depths, when it does not completely obscure the bottom, thus constraining SDB for its routine application. Retrieving SDB has typically involved manual screening of the available images to find a scene that can provide the least area of turbidity, as well as having minimal interference from clouds, surface reflectance (sunglint), and transient breaking waves. This screening does not eliminate all turbidity, besides being tedious and highly subjective. We have developed a multi-temporal method with the Sentinel-2 satellites that eliminates manual screening and reduces turbidity effects, while applying a commonly used, physically-based algorithm (log ratio). The proposed methodology incorporates a robust atmospheric correction (ACOLITE), a multi-scene compositing method, and a switching model to improve mapping in shallow water. No manual intervention or scene selection (beyond defining a time range for imagery) is required to go from the individual input scenes to the composited scene. Currently calibration requires selection of only 10-15 reference points obtained from available nautical charts (lidar surveys are not needed for calibration). The semi-automated approach yields accurate and consistent SDB at 10 m spatial resolution along the Eastern coast of USA, with median errors of 0.5 m for depths 0-25 m when validated with lidar surveys, errors that favorably compare to uses of SDB in clear water. This framework substantially decreases time, labor, and the influence of transient turbidity, particularly when using data from a platform, like the Sentinel-2 twin mission, which provides routine and repetitive image acquisition. As a result, SDB can expand on the temporal and spatial scope of more conventional mapping methods, such as lidar or multi-beam. By implementing the multi-temporal correction, SDB may substantially enhance change detection and support operational and recursive coastal monitoring on local to regional scales