Case study of a hypertemporal terrestrial lidar to monitor a macrotidal beach

Abstract

In order to protect the Belgian coast, knowledge on natural sand dynamics is essential. Monitoring sand dynamics is commonly done through sediment budget analysis, which relies on determining the volumes of sediment added or removed from the coastal system. These volumetrics require precise and accurate 3D data of the terrain on different time stamps. Earlier research states the potential of permanent long-range terrestrial laser scanning for continuous monitoring of coastal dynamics. For this paper, this methodology has been implemented at an ultradissipative macrotidal North Sea beach in Mariakerke (Ostend, Belgium). A Riegl VZ-2000 LiDAR, mounted on a 42 m high building, scanned the intertidal and dry beach in a test zone of ca. 200 m wide on an hourly basis over a time period of one year. It appeared that the laser scanner cannot be assumed to have a fixed zenith for each hourly scan. The scanner compensator measured a variable deviation of the Z-axis of more than 3.00 mrad. This results in a deviation of ca. 900 mm near the low water line. A robust calibration procedure was developed to correct the deviations of the Z-axis. ‘Robust’ meaning (1) independent of a truth set of reference points (2) independent of the model and (3) independent of outlier elimination strategy. Several strategies for calibration of the vertical axis of the scanner have been tested and compared. A TIN model of the terrestrial scan – calibrated on an airborne LiDAR truth of the unchanging seawall and groin – combined with iterative statistical outlier elimination (mean ± 2.5 × standard deviation) resulted in the best outcome.