On the use of satellite information to detect coastal change: Demonstration case on the coast of Spain

Abstract

Recent developments in satellite processing tools allow low-cost, fast and automatic processing of a large amount of information from Earth observation, enhancing the capability of detecting coastal changes from space at different temporal scales. Some works have assessed the quality of these data and applied it to detect coastal evolution locally, most of them focusing on mid-term and long-term changes in the coastline. In this work, we evaluate the capability to monitor changes in coastal morphology at various temporal and spatial scales using 1D (coastlines) and 3D (bathymetry) satellite-derived data obtained from site-specific processing methods. Local characteristics were included in several phases of the development of the satellite products used here: i) geolocated very high resolution images from each pilot site were used in the coregistration process to enhance geolocation accuracy in images from different missions, ii) different spectral indices were tested at each pilot site to obtain more reliable detection of the coastline at all sites and iii) measured topobathymetry data were used to obtain datum-based satellite shorelines and bathymetry. The accuracy and skill of those satellite products were assessed at several pilot sites in Spain. The results indicated high horizontal accuracy (RMSE < pixel size), with errors on the order of half of the pixel size (RMSE = 5.0 m and for Sentinel-2 and 18.8 m for Landsat5). Furthermore, the coastlines used here presented errors comparable to those obtained from the widely used open-source tool CoastSat. Time-series analysis using satellite-derived shorelines showed that coastal change processes can be detected at several temporal and spatial scales, such as short-term erosion and accretion events on a small beach, seasonal beach rotation, and long-term trends at local and regional scales. However, the results from satellite-derived bathymetry indicated that the quantitative assessment of the coastal morphology with 3D products is still limited. Some in situ measurements are necessary to obtain satellite data that represent site-specific conditions. However, the quantity of this auxiliary in situ measurements required to obtain reliable time series of satellite derived shorelines and bathymetry is significantly lower than the quantity required by traditional monitoring methods. The results are discussed, highlighting the gaps that need to be filled in the future so that satellite-derived products can be used in usual coastal change monitoring practices.