Contribution of UAS to the monitoring at the Larchberg-Galgenwald landslide (Austria)

In April 2001, a rock fall occurred at the landslide Larchberg-Galgenwald, Austria. The movement of this landslide has been under observation for decades, but the methods used so far have several disadvantages. This article aims to show that unmanned aerial systems (UAS) provide a reasonable alternative in terms for monitoring the landslide. At the landslide, we focused on the most active area to (i) record changes of the landslide's surface, (ii) assess the accuracy of the resulting data, (iii) assess whether the number of ground control points (GCPs) could be reduced in subsequent UAS surveys, and (iv) conclude whether UAS-based recording is favourable compared to common terrestrial methods. UAS imagery, acquired under challenging site conditions in November 2015 and May 2016, were processed to digital elevation models and orthophotos with a ground sampling distance of 0.04m, which should display expected changes of up to similar to 0.17m (relating to previously reported movements of up to 30cm a(-1)) for the same period of time. Using DEM differencing, we calculated a vertical difference that was in the range of a few centimetres in most cases with only a few localized areas of much larger change. The latter are mainly related to single objects moving such as rock blocks or tree trunks. Apart from that, no general surface elevation lowering (a threshold of 0.27m was used to distinguish significant from insignificant changes) could be detected. Based on the resulting orthophotos, we calculated vectors of horizontal displacement, which confirmed that the changes are related to single objects and revealed that in upcoming flight campaigns the number of GCPs can be generously reduced with an acceptable loss of positional accuracy. Therefore, the suitability of UAS for expanding the monitoring approaches used at Larchberg-Galgenwald landslide is described.