Is stereo satellite imagery a reliable method to infer geomorphic dynamics in fluvial systems at high spatial resolution and precision?

Abstract

Remote sensing techniques (e.g., SfM photogrammetry, LiDAR) have been proved as a consistent method to reconstruct fluvial landscapes and processes at high temporal and spatial resolutions. Despite the advantages, these methods still presenting some limitations as, for example spatial covering, which usually is limited to <100 km length. In this context, stereo satellite imagery is presented as a valuable remote sensing technique which allow to reconstruct Earth’s relief at high spatial resolution covering extensive areas (>100 km) in one single satellite caption. Several studies have proved the accuracy and precision of this method in various geomorphological contexts, despite this, today there are very few works that study the reliability of these methods in fluvial environments with different degrees of complexity. The main objective of this study is to evaluate the use of stereo satellite imagery as a reliable method to reconstruct the topography of fluvial systems at high spatial resolution, precision, and accuracy. Additionally, we assess the reliability of this method to infer on geomorphic processes through the comparison of multi temporal high-resolution topography. To pursue these objectives two sets of stereo satellite imagery (i.e., DEIMOS and Pleiades) were compared with reference topographic datasets composed of SfM-photogrammetric (i.e., UAV platform) and GNSS-RTK surveys collected at the same time. These datasets were confronted in three different fluvial reaches located in the Marecchia River (Nortern Appennines). Study reaches presented contrasted characteristics in terms of fluvial pattern, active width, confinement, and channel-slope, which cover a wide range of fluvial types of mid-mountain rivers. Residuals between datasets were compared with different morphometric variables (e.g., slope, roughness) and channel characteristics (e.g., water depth).