Comparing Waterline and Water occurrence approaches for satellite‑derived intertidal topography

Intertidal zones are situated at the boundary between land and sea and are one of the most important natural buffer zones for the protection of coastal regions, characterized as highly dynamic areas under constant redistribution of sediment. Despite their importance, monitoring the morpho-sedimentary dynamics of intertidal zones is still a challenge. However, satellite measurements enable efficient ways to monitor intertidal areas, providing wide coverage and frequent observation. This study focuses on assessing the performance of two space-based methods in mapping intertidal topography using MultiSpectral (MS) and Synthetic Aperture Radar (SAR) imagery: (i) the waterline method, which uses elevation from waterlines extracted at different tidal stages, with (ii) the water occurrence method, which estimates elevation based on the frequency of inundation of each pixel. To compare the results of these methods, three locations were selected: Bay of Veys, Utah beach and Seine estuary. These sites, located in Normandy region in France, represent different intertidal environments, ranging from a shallow estuarine system to an open beach and an anthropogenically modified intertidal area. DEMs are generated utilizing Sentinel-2 (MS) and Sentinel-1 (SAR) satellites with water level information obtained from two model outputs: the HYbrid Coordinate Ocean Model (HYCOM) and the Finite Element Solution ocean tide model (FES2022). To evaluate the performance of these methods, DEMs were generated using Sentinel-2 data with two different indices (Normalized Difference Water Index - NDWI and Optimized Water Index for Coastal Zones - SCOWI), each of them combined with both water level models. A combination of Sentinel-2 and Sentinel-1 was also tested. All these data combinations were applied to both waterline and water occurrence methods. Compared with LiDAR derived DEMs, preliminary results across all sites show that the waterline method generally achieves Mean Absolute Error (MAE) values in the 0.23 - 0.35 m range, while the water occurrence MAE ranges from 0.33 to 0.57 m. Different intertidal environments and validation data show that both methods have solid performance in different intertidal environments, with opportunities for further improvement. Unlike the waterline method, the water occurrence method can be fully automated, which makes it a promising option for large scale applications.