Assessing the uncertainties of Radar and LiDAR derived digital elevation information over static and dynamic terrain

Abstract

The satellite era brought the advantages of obtaining fast, reliable, and repeated measurements over time on a global scale. Digital Elevation Information is a necessity for any scientific and engineering project that involves terrain, urban planning, and monitoring of natural hazards. The accuracy of Digital Elevation Information is influenced by systematic errors which may be caused by the instruments, the effects of climate on some sensors and regional dynamic processes. The repeatability of measurements collected by satellite sensors introduces time as a new dimension and a new source of uncertainty. In this study, the uncertainties of Synthetic Aperture Radar and Light Detection And Ranging derived products are assessed in static and dynamic terrain cases. A synthetic aperture radar derived digital elevation model is validated over an urban area using airborne LiDAR, for which the side-looking nature of synthetic aperture radar and the variations in wavelength depending on penetration depth are observed as the principal factors influencing uncertainty over built and vegetated environments. Sand dune migration is studied using satellite laser altimetry. By using the phase of the elevation profiles and applying a cross-correlation between overlapping profiles, we estimate migration vectors resulting in a database that can be correlated with the region’s wind patterns to locate the effects of extreme events such as sand or dust storms. Cloud cover, among other factors, interferes with the altimetry signal causing data losses. To fill these gaps, Projection onto Convex Sets is employed to interpolate between known elevation locations using the known properties of the elevation profiles in the region. Understanding these dynamic processes from satellite observations allows us to address long-term and short-term sand dune migration hazards and develop mitigation strategies.