Satellite-Based Framework for Detecting Seabed Dynamics from Shoreline Variability along Spit Coasts—A Pilot Study from the Hel Peninsula

Abstract

This study presents a pilot, data-driven framework for detecting nearshore seabed dynamics based on shoreline variability derived from high-resolution satellite imagery, integrated with in situ bathymetric measurements along the Hel Peninsula in the southern Baltic Sea. The approach focuses on assessing whether satellite-derived shoreline displacement can serve as an indirect indicator of changes in seabed morphology within the sedimentologically active coastal zone. Four cross-shore transects were analyzed using Pléiades satellite imagery and field bathymetric data collected during two monitoring campaigns conducted in 2019 and 2022. Changes in seabed elevation were quantified through cross-sectional profile analysis, while key morphodynamic parameters, including closure depth and the length of the active seabed, were estimated using empirical formulations proposed by Hallermeier, Birkemeier, and Houston. To provide hydrodynamic context, wave and current conditions were characterized using Copernicus Marine Service reanalysis data. The results reveal pronounced spatial variability in seabed response along the Hel Peninsula, with erosion-dominated conditions in high-energy open-coast sectors and limited morphological change in semienclosed, low-energy environments. Among the tested formulations, Birkemeier’s equation showed the closest agreement with field-derived closure depths, while all empirical models systematically overestimated closure depth in low-energy settings. The findings demonstrate the potential of satellite-based shoreline observations as a supporting tool for coastal morphodynamic assessment while highlighting the limitations imposed by measurement uncertainty, short observation windows, and simplified process representation. Overall, the proposed framework should be regarded as a proof-of-concept application requiring longer observational time series, improved uncertainty quantification, and local calibration before operational implementation can be considered