Determination of glacier surface displacements through multi-temporal UAV data

Abstract

The shelf and outlet glaciers fluctuation characteristics are classically performed by comparing multi-temporal mapping sources, geophysical and remote sensing methods. The unmanned aerial vehicles (UAVs) usage equipped with GNSS receivers and digital photographic cameras enables obtaining georeferenced information and images. This data can be successfully employed for characterizing the movement of the glacier. The author proposes a modified pseudo-parallax method, which uses the multi-temporal survey materials obtained from the UAVs to determine the spatial displacements of the glacier surface. Its efficiency is demonstrated on the movement pattern determination example of the Dålk Glacier (North Antarctica) in the period from January, 20 to February, 9, 2017. Through that we can determine the total displacement vector on a point by three coordinates, they are measured by pseudo-stereopair obtained from different-time orthophotomaps and DEM. Modern automated techniques for identifying the corresponding points on a pair of images allow obtaining a sufficiently dense measurement network with predictable accuracy and precision. The investigation made it possible to determine the glacier displacement in 10 days and the distribution nature of these offsets without ground-based geodetic measurements. The average drift was 6,28 m, with a scatter from 0,07 m to 10,8 m. A method for estimating the shifts without ground control measurements was proposed and the main sources of systematic and random errors in coordinates and points heights determination were described.