Characterisation and Analysis of Catastrophic Landslides and Related Processes using Digital Topographic Data

Abstract

This thesis represents a large body of work that seeks to describe, quantify, and simulate the behaviour of large rock slope failures (> 1 Mm³), in the form of landslides and rock avalanches, and their secondary processes, such as landslide-dammed lakes, utilizing remotely sensed data. Remotely sensed data includes aerial photography, high resolution satellite imagery from various platforms (e.g. LANDSAT, ASTER, EO-1, SPOT), and digital topographic elevation models of the Earth’s surface (e.g. SRTM-3, ASTER GDEM2, LiDAR). This thesis focused on regions in northwest North America (British Columbia, Yukon Territory, and Alaska), and on regions in the Himalaya and Pamirs Mountain chains (Tajikistan, Afghanistan, Pakistan, Tibet, and India). These study regions are each highly dynamic landscapes, where the occurrence of rock slope failures per area is higher than non-mountainous regions, and these events are aiding to the shape and profile of the landscapes and surfaces found today. This thesis focuses on: 1) the ability to accurately calculate geometrics (e.g. areas, volumes, runouts, debris depths) for large scale landslides and their associated landslide dammed lakes (e.g. areas, volumes, outbursts), utilizing data from remotely sensed sources; 2) the attempt to successfully simulate the observed dynamics for both landslide emplacement and their resulting debris deposits (DAN-W, DAN3D), and possible outburst flood scenarios (FLO2D); and, 3) attempt to quantify the kinetic and specific energy involved in rock avalanches, and how these energetics relate to fragmentation, as well as the lateral spreading and thinning of debris sheets. The river valleys of the northwest Himalayas (Pakistan and India) and the adjacent Pamirs Mountains of Afghanistan and Tajikistan contain in excess of two hundred known rockslide deposits of unknown age that have interrupted surface drainage and previously dammed major rivers in the region in recent and prehistoric time. Some prehistoric rockslide dams in the northwest Himalayas have impounded massive lakes with volumes in excess of 20 Gm³. The region contains: 1) the highest rockslide dam in the world (the 1911 Usoi rockslide, Tajikistan), which impounds the current largest rockslide-dammed lake (Lake Sarez) on Earth (est. volume 17 Gm³); 2) the largest documented outburst flood (6.5 Gm³) associated with a historical rockslide dam outburst (the 1841 Indus Flood, Pakistan); and, 3) the world’s most recent rockslide-dammed lake emergency, the 2010 Attabad rockslide dam on the Hunza River, in the Upper Indus basin, including the newly created Lake Gojal. By accurately quantifying the volume of an impoundment, and the downstream valley topography (DEM), floodwave scenarios can be created for various breaching situations, allowing for the delineation of downstream inundation areas, or the creation of hazard and risk scenarios. Two methods are used to attempt to quantify the volumes of landslide-dammed lakes: 1) a contour interpolation method, focusing on the creation of contours to represent lake levels in the DEM data; and, 2) a new technique using digitized shorelines and statistical methods to obtain lake elevations on specific dates. A new technique has also been developed to quantify the larger block fragmentation from rock avalanches in the glacial environment, and a credible grain-size curve for the largest blocks can be obtained, aiding in the creation of a more complete grain-size curve for a particular event. The combination of landslides and their associated landslide dammed lakes are an important geomorphic process to study, as these events have a direct relationship to the hazard and risk faced by local communities living and working in these regions. By understanding the emplacement and deposit dynamics of large landslides and/or the outburst flood scenarios from naturally impounded reservoirs, we can attempt to reduce the direct impacts these events have to local communities.