Mapping inland water bathymetry with Ground Penetrating Radar (GPR) on board Unmanned Aerial Systems (UASs)

Bathymetry of inland water bodies is essential for river maintenance and flood risk management. Traditionally, in shallow water bodies, bathymetry is retrieved by operators wading through the water body with Real Time Kinematic (RTK) Global Navigation Satellite System (GNSS), whilst in deeper waters, it is retrieved with sonar instruments on manned or unmanned boats. In the past, researchers have documented the use of Ground Penetrating Radar (GPR) on boats (i.e. water-coupled GPR) for monitoring the bathymetry of frozen and non -frozen water bodies. Furthermore, GPR has been used on helicopters for monitoring ice and snow thickness. However, deployment of GPR on board Unmanned Aerial Systems (UASs) in non-frozen inland water bodies with electric conductivity higher than 100 mu S/cm (as is common in most inland waterbodies in non-polar regions) is unexplored. In this paper, we document the possibility to use drone-borne and water-coupled GPR in several cross-sections located in three different waterbodies (1 lake and 2 rivers) in Denmark. These waterbodies had different bed sediment materials and vegetation conditions, an electric conductivity varying from 200 to 340 mu S/ cm and depths up to 2.5 m. Drone-borne GPR showed accuracy similar to water-coupled GPR when compared to RTK GNSS ground-truth measurements, with a Mean Absolute Error (MAE) of approx. 8 cm. The only limitations of drone-borne GPR were i) more restrictive minimum depth requirement (typically 0.8-1.1 m for drone-borne GPR, while 0.3-0.4 m for water-coupled GPR) ii) requirement to fly the GPR antenna at altitudes of approx. 0.5 m above the water surface to avoid high spreading losses and strong surface clutter events hiding the signal. Finally, GPR measurements were benchmarked against traditional sonar measurements, showing that GPR measurements significantly outperform sonar measurements in waterbodies with medium or high density of aquatic vegetation.