Shoreline Extraction from Earth Observation data using the Isoradiometric Method: a QGIS scripting and Google Earth Engine workflow supporting landscape-evolution analyses

Sandy coasts are among the most dynamic and vulnerable environments, yet the development of robust and scalable methods for shoreline extraction from satellite imagery remains under discussion. Most current approaches delineate shorelines at the pixel scale, so their accuracy is intrinsically constrained by sensor resolution and pixel geometry. In this contribution, we present an isoradiometric sub-pixel shoreline extraction method, operationally implemented via QGIS scripting and Google Earth Engine (GEE), to support coastal morphodynamics and landscape-evolution analyses over multi-decadal timescales. We first develop and test the method on three Sicilian sandy beaches (Torre Salsa, Balestrate, and Vendicari), combining field radiometric measurements with multi-sensor satellite data (Landsat-8, Sentinel-2, and Planetscopes). Radiometric profiles acquired across the water–sediment interface are used to define spectral profiles and an iso-reflectance line at bottom-of-atmosphere associated with the shoreline, with particular emphasis on red-edge, near-infrared, and shortwaves bands. These iso-radiometric lines are then interpolated across the satellite scenes to obtain a continuous, sub-pixel shoreline. Independent NRTK GNSS surveys collected along the swash zone are used to validate the extracted shorelines and to quantify positional accuracy. Our results show that the isoradiometric approach on near infrared (NIR) bands can achieve shoreline position accuracy comparable to, and in several cases exceeding, those of more complex state-of-the-art methods, while remaining conceptually simple and cost-effective. The analysis of method performance across sensors highlights the key role of NIR bands and allows us to discuss the trade-offs between spatial and temporal resolution. We also systematically assess critical factors that can bias shoreline position, such as breaking waves, beached Posidonia oceanica, cloud and terrain shadows, and water-body radiometric effects, providing practical guidance to minimize their impact. In a second step, we iteratively apply the method to multi-decadal satellite image series, as well as seasonal datasets, to quantify long-term shoreline change and to distinguish phases of regression and transgression. The resulting time series are integrated into a simple quantitative landscape-evolution framework, enabling the estimation of differential erosion and accretion along the coast. Finally, we implement the full workflow in an open and reproducible way by combining GEE for large-scale image access and pre-processing with QGIS models and Python scripting for shoreline extraction and analysis. This integrated environment allows non-expert users and decision-makers to apply the isoradiometric method at a regional scale, thus offering a practical tool to support coastal erosion assessment and adaptation planning in the context of ongoing and projected climate-driven changes.