Understanding Debris-Flow Motion through Detailed Analysis of Timelapse Point Clouds Collected by High-Frequency 3D LiDAR Scanners: Illgraben, Switzerland

Abstract

Debris flows are extremely rapid, flow-like landslides with large impact forces and long runout distances, and they are one of the most dangerous types of mass movements in mountainous regions. Managing the risk posed by this type of landslide is particularly important for alpine countries like Switzerland, where debris flows have caused major damage and led to numerous fatalities. More detailed field-scale measurements of natural debris flows are required to better understand the fundamental mechanisms governing debris-flow motion and, ultimately, to reduce the associated risks in the future. This Master’s thesis aims to contribute to a better process understanding by analyzing a debris-flow event using timelapse point clouds from a high-resolution, high-frequency 3D LiDAR sensor (Ouster OS1-64). Such a sensor has been installed at the Illgraben catchment (Valais, Switzerland), one of the most active debris-flow catchments in the Alps. The author developed and applied both manual and automated algorithms to derive hazard-related debris-flow parameters at high spatial and temporal resolution over a 25m long channel segment upstream of a check dam. The following parameters were quantified: i) front velocity, ii) surface velocity, iii) cross-sectional area, iv) discharge and v) event volume. A decrease in the front velocity towards the check dam was observed and appeared to be related to a change in the channel slope and width. Moreover, the surface velocities measured directly behind the front were (on average 1.75×) faster than the front velocity, which likely led to the formation of the bouldery front observed in this event. It was further found that different features – including large, rolling boulders and woody debris – traveled at systematically different velocities during the second surge of the event. This observation was likely caused by different objects sampling the velocity profile at different depths and allowed for the vertical velocity profile of the debris flow to be inferred over the event. The discharge was estimated for three different channel sections upstream of the check dam based on automatically derived surface velocity and cross-sectional area values. In these discharge estimates, a systematic decrease towards the check dam could be observed. This observation, which was called the “discharge paradox”, could be explained by spatial and temporal variations in the velocity profile and the channel bed geometry. The ratio between discharge estimates at different channel sections allowed for an assessment of the contribution of these two effects. It is assumed that during the first surge of the event, potential changes in the channel bed geometry were the main cause for the observed discrepancy. In contrast, during the second surge, mainly spatial changes (along the channel) in the velocity profile caused the inconsistent discharge estimates. Eventually, the volume of the debris flow was derived as the sum of the discharge over the event (13800m3 to 16300m3; ca. ±3000m3, i.e. ca. ±20 %). The LiDAR data analyzed in this project is unique because it allows a truly 3D, high-resolution investigation of moving debris flows at sub-second intervals. The developed methods will be applied to LiDAR data from additional monitoring stations and events at the Illgraben, allowing for further inference into the internal dynamics of debris flows. Eventually, this might enhance the understanding of the fundamental debris-flow mechanisms, help to optimize numerical as well as empirical modeling approaches, improve hazard mitigation in general and reduce the risk posed by flow-like landslides in the future.