A time-variable topo-bathymetry from coastal remote sensing observations

Sea level rise is commonly associated with retreating shorelines. However, shoreline evolution is the result of the complex interaction between several groups of processes: Changes in inundation from changing water levels, vertical land motion and morphodynamics. Our goal is to quantify and separate the influence of these processes on the shoreline geometry by using remote sensing data. In a case study for the barrier island of Terschelling (the Netherlands), we found that between 1992 and 2022 morphodynamics had the largest impact on shoreline changes: Inundation by sea level rise, corrected for vertical land motion, accounted for -0.3 m/year shoreline retreat on average while the total average shoreline trend was -3.2 m/year.

These results are very site-specific and cannot be easily transferred to other places. The main limitation for upscaling this method lies in the availability of land elevation data. Local high-quality elevation datasets from airborne LiDAR or ship-based bathymetry are ideal but usually limited to countries which invest in regular observations campaigns. On the other hand, global Digital Elevation Models (DEMs) either lack the required vertical accuracy or horizontal resolution, they often cover only either the topography or the bathymetry, or they mix several data sources resulting in a mean elevation model spanning time periods of several years to decades.

Here we present a technique to derive a time-variable elevation grid that 1) can be applied globally, 2) has a high temporal resolution, 3) covers the intertidal area around the shoreline (foreshore and upper shoreface), and 4) has sufficient vertical accuracy and horizontal resolution. Additionally, we will address the question which accuracies are considered "sufficient" for certain problems.

To create such a time-variable topo-bathymetry model with yearly resolution for the years 1993-present, we combine existing global DEMs (e.g. DeltaDTM or CoastalDEM) with satellite remote sensing observations in a Kalman filter scheme. The observations are yearly 2.5D point clouds (x,y,h) of the intertidal zone that we generate by assigning sea surface heights from coastal altimetry to shoreline contours from optical remote sensing ("waterline method"). First, we incrementally update the global DEMs with these point clouds in a forward Kalman filter. Then, we use a backward smoother to derive the final elevation grid that best represents the topo-bathymetry at one point in time.

For validation, we apply this technique to sandy beaches in the Netherlands, Duck (USA) and Narrabeen (Australia), where high-quality elevation dataset are available. We hope to find that this method increases the accuracy of global DEMs and allows us to study temporal variations in coastal morphology and the role of sea level rise in data-sparse regions worldwide.