Numerical Investigation of the Stability of Toppling Rock Slopes Subjected to Glacier Retreat

Glacial retreat is often cited as a cause of rock slope instabilities in mountain regions. Until recently, glacial debuttressing was thought to be the main mechanism by which glaciers influence slope stability, however recent work has questioned the efficacy of this mechanism.  It appears that other mechanisms, including slope kinematics and hydro-mechanical interactions between the glacier and slope are important drivers of paraglacial rock slope instabilities.  In the present work, we use discontinuum numerical models to investigate the interaction between rock slope kinematics, slope/glacial hydrology and glacial retreat. 
We perform both a theoretical analysis using a simplified slope geometry, as well as a back-analysis of the Moosfluh Landslide.  For the theoretical analysis, we investigate the response of both toppling and sliding slopes to two factors: the weight of the ice, assumed to be applied as a ductile load acting normal to slope topography, and the variation of the slope water table, which is linked to the ice level and lowers as the glacier retreats. We then apply the insights from the theoretical analysis to investigate the Moosfluh Landslide.  This landslide, which is located at the left flank of the Great Aletsch Glacier Valley (Valais, Switzerland), at the present-day glacial terminus, underwent a dramatic acceleration in 2016 in response to glacier retreat.  The landslide was extensively monitored during this acceleration, and analysis of this data has revealed that the kinematics of movement changed from toppling to secondary sliding.  We simulate the behaviour of the Moosfluh Landslide by implementing a structural model determined from field mapping, and systematically lowering the ice level and slope water table, to simulate glacial retreat.
We find that the interaction between slope kinematics and glacial retreat leads to a complex slope response.  For sliding slopes, the stability of the slope is relatively insensitive to glacial ice loss.  For toppling slopes, the slope response is highly sensitive to ice loss, and the slope is the most unstable at a critical ice level, before ice has completely retreated.  For the Moosfluh instability, we are able to simulate the initial toppling kinematics of this landslide, as well as the transition to sliding triggered by the ice reaching a critical elevation.  Our analysis has important implications for understanding rock slope response to glacial retreat, and highlights the disparate behaviour of toppling and sliding slopes.