An approach to measuring complex topographic change with airborne and drone-derived lidar of a large migrating river channel

Repeat lidar precisely records how Earth's surface topography changes over time. However, it is difficult to measure 3D change between lidar surveys on steep, curved surfaces due to inherent directionality in both lidar instrument geometry and processing methods. Here, we use two pairs of point clouds to measure lateral river channel erosion along the ~200 m-wide Trinity River in Texas between September 2023 and February 2024. One pair of point clouds was obtained with crewed airplanes, while the other was obtained with a drone-based system. All four datasets are classified using a custom workflow that improves ground point density and data coverage as compared to default classification methods. We find that the custom-classified drone-derived data has ~2–8x more points on steep surfaces (> 55°) than the airplane-derived data. We then compute 3D change volumes between surveys collected from the same platform. Points that changed are identified by calculating the horizontal M3C2 distances between subsequent surveys, then Poisson surface reconstruction is used to enclose volumes of the river bank surface that eroded. Results are compared to volumes calculated by mapping channel edge movement and differencing 2.5D rasters of the exposed bank. Each method estimates similar eroded volumes, but our reconstruction approach utilizes the full, detailed 3D geometry of the point clouds, resolving both individual channel bank failures and larger channel bend migration patterns. This approach could be applied to precisely measure mass fluxes on other steep, dynamic, and complex topographic surfaces, such as beach cliffs and glacier termini.