Initiation and mechanism of rock slope failures triggered by the 2016 Mw 7.8 Kaikōura earthquake

Abstract

<p>The 2016 Mw 7.8 Kaikōura earthquake on New Zealand’s South Island triggered c. 30,000 landslides. Around 70% of landslides occurred in Torlesse greywacke rock mass, which is characterised by closely spaced but low-persistence joints. Most failures in this rock mass were relatively shallow rock avalanches which do not appear to follow traditional failure mechanism models. Here, we use detailed site characterisation and dynamic numerical modelling to better understand landslide hazard and risk from Torlesse greywacke slopes. Using multi-method site characterisation including 3D pixel tracking in pre- and post-earthquake aerial imagery, geomorphic mapping, rock mass characterisation, geophysical ground investigations and a geotechnical borehole, we developed engineering geological ground models for individual sites. We then used these to develop a conceptual framework of failure mechanism in Torlesse greywacke and propose a ‘joint-step-path’ failure mechanism in which rupture surface propagation occurs along pre-existing, but low-persistence joints through multiple degrees of kinematic freedom. Torlesse greywacke failures typically evolve in three main landslide failure stages – incipient, transitional and rock avalanching. Hazard can increase for the same slope when it transitions from the incipient failure stage to sliding and/or avalanching. To quantify the transition between failure stages, we analysed coseismic displacement and strain for six landslides. As many displacement based coseismic landslide susceptibility models require some threshold, above which the slope is assumed to transition into a landslide, this information could potentially serve as a useful tool. For slopes at the incipient and transitional stage, 1D maximum total strain appears to be closely correlated with source slope angle. Based on these results, we develop the ‘transitional slope strain index’ (TSSI) that combines 1D maximum total strain with source slope angle. The TSSI relates to the likelihood of a slope transitioning into a more mobile, and therefore more hazardous, rock avalanche at a given level of earthquake shaking. Dynamic numerical back-analysis of the initiation of two landslides in Torlesse greywacke supports our empirical hypotheses that landslide susceptibility in this rock mass is strongly influenced by slope angle and rock mass strength. Coseismic failure initiation is, furthermore, strongly dependent on ground motion input. The geometry of failures can be reproduced using a random Voronoi joint network and adopting residual joint strength parameters, which further lends weight to the ‘joint-step-path’ failure mechanism hypothesis.</p>

Evolution of an earthquake-induced landslide complex in the South Island of New Zealand: How fault damage zones and seismicity contribute to slope failures

“…9). This interpretation is supported by dynamic numerical modeling conducted by Singeisen (2023). Modeling results showed that these defects facilitate flexural toppling and are required to replicate (1) the pattern of coseismic landslide displacements observed following the 2016 earthquake and (2) the type and depth of slope deformation observed in geotechnical and field investigations.…”

“…Amplification of ground motion in the damaged and weathered rock mass during seismic events (i.e., amplification caused by impedance contrasts in the fault damage zone and preexisting rock mass damage; Martino et al, 2006) likely additionally led to slope deformation and sliding along defects and weak zones. Dynamic numerical modeling by Singeisen (2023) showed topographic amplification at the slope crest at Half Moon Bay, with amplification increasing further when considering material strength effects and discontinuities. Based on the intense weathering and fracturing of the rock mass observed at Half Moon Bay (Fig.…”