What lies beneath : investigating the distribution and dynamics of landslides in lakes, and fjords of New Zealand

Abstract

Subaerial and subaqueous landslides are a significant hazard in intramountainous lake and fjord landscapes across the globe. Landslides into, and within, the lakes and fjords can generate tsunamis that present a risk to shoreline communities and infrastructure. Understanding the landslide and associated hazard cascade (e.g. landslide tsunamis) is reliant on both identification of deposits to assess their magnitude-frequency relationship, and detailed analysis of the deposits, which can provide information on the landslide’s emplacement dynamics. This thesis aims to better understand the distribution and dynamics of landslides within lakes and fjords in the seismically active, high-mountain landscape of the South Island, New Zealand, to improve the understanding of the hazard. Despite the similar physiographic setting to many global counterparts (e.g. Norway, European Alps, western Canada, Patagonia), there have only been a few studies to quantify the hazard from landslides into and within lakes and fjords of New Zealand, compared to the subaerial landslide hazard. Previous research suggests that the lakes and fjords of New Zealand may provide a well-preserved archive of previously unknown landslide deposits. A new literature-based inventory of lakes and fjords with known landslides is presented within this thesis. In total, 15 lakes and 5 fjords were identified as having documented landslide events. The lakes and fjords contain a variety of landslide types; including subaerial originating (n=7), deltaic landslides (n=4), and mass-transport deposits, assumed to have originated from the subaqueous slopes (n=7). Previous bathymetric data investigations in Lake Tekapo and Milford Sound/Piopiotahi reveal deposits from multiple sources. Earthquakes are the main trigger, with 9 cases related to known earthquakes. Earthquakes are assumed to be the trigger in 5 further lakes and fjords. In Milford Sound/Piopiotahi, morphological and morphometric analysis, and numerical modelling by DAN3D of previously identified deposits from high-resolution bathymetric data indicate that the landslides exhibit a variety of morphologies, morphometries, and generally low apparent mobility. Factors such as transitioning from air to water, abrupt slope angle changes from the steep fjord walls to flat basin wall, and lack of lateral confinement resulting in freely spreading deposits appear to contribute to the apparent low mobility. Furthermore, the best fit numerical simulations required high basal friction angles, and/or low turbulence parameters to retard the landslide runout The modelling results also found a Voellmy rheology to be effective in simulating the subaqueous portion of landslide runout, as varying the velocity-dependent turbulence coefficient parameter could represent the resistance to flow ii by the water, in contrast to the frictional rheology where resistance is governed only by basal friction angle. In Lake Wānaka, newly collected bathymetric and sub-bottom data reveal comparatively few deposits on the lakebed (n=17), in contrast to the large number of subaerial deposits mapped and identified within the terrestrial catchment in this thesis (n=202). Limited subsurface penetration precludes identifying more deposits within the lake infill. Sub-bottom data do reveal that some of the visible deposits within the lake are draped by lake sediments, obscuring their true thicknesses. Therefore, burial of deposits may be obscuring the total number of deposits within the lake basin. The research presented within this thesis builds a proposed conceptual model for landslide distributions in New Zealand lakes and fjords. Landslides may occur more frequently in areas of frequent and strong seismicity. In contrast, regions with low frequency seismicity will likely trigger fewer landslides into, and within the subaqueous basins less often. Variations in sedimentation rates within the lakes and fjords control how long deposits will be visible on the basin floors for. Furthermore, the newly identified deposits, and quantitative insights into the landslide emplacement dynamics provide information that can be taken forward to assess tsunami hazard, which may present the greatest risk to shoreline communities and infrastructure around New Zealand lakes and fjords.