Search

On 14 November 2016, the Northern part of New Zealand's South Island was struck by the Mw7.8 Kaikōura earthquake, an event of unprecedented geometric complexity rupturing at least 21 segments of the contractional North Canterbury and of the Marlborough fault systems. Geological, geodetic, tsunami and seismic data reveal puzzling features as well as observational difficulties of the event leading to competing views on its key features, such as the role played by the Hikurangi subduction interface, the exact rupture sequence, and the higher than expected local tsunami. We present a dynamic rupture model of the 2016 Mw 7.8 Kaikōura earthquake to unravel the event's riddles in a physics-based manner and provide insights on the mechanical viability of competing hypotheses proposed to explain them. The event is modeled using the open-source software SeisSol (www.seissol.org), based on a arbitrary high-order accurate DERivative Discontinuous Galerkin method (ADER-DG). We verify our model with strong ground-motions, high rate GPS, long-period teleseismic and tsunami data. Our model reproduces key characteristics of the event and constraints puzzling features inferred from high-quality observations including a large gap separating surface rupture traces, the possibility of significant slip on the subduction interface, the non-rupture of the Hope fault, and slow apparent rupture speed. We show that a rupture cascade sequentially breaking the Papatea and Kekerengu faults is dynamically consistent with regional stress estimates and a crustal fault network geometry inferred from seismic and geodetic data. A coseismically triggered tsunami caused by the rare occurrence of an earthquake starting on-land and crossing to two faults off-shore matches local tide gauge data.We find, that such complex rupture cascade links the crustal fault system when operating at low apparent friction thanks to the combined effects of overpressurized fluids, low dynamic friction and stress concentrations induced by deep fault creep. Understanding the physical conditions that allow rupture cascades on complex fault systems advances our ability to quantify earthquake hazard, especially to evaluate the possibility of extreme events on real fault networks. [ABSTRACT FROM AUTHOR]

The 2004 Mw 9.1-9.3 Great Sumatra-Andaman Earthquake is one of the strongest and most devastating earthquakes in recent history. It generated a powerful tsunami, causing more than 220,000 fatalities. Despite its huge spatial extent, the analysis of the earthquake is impeded by a lack of near-source observations. Static ground-deformation data, polluted by significant after-slip in the near-source region, and teleseismic data do not fully constrain the source, leading to significant differences in its characterization. Specifically, the tsunamigenic aspects of the earthquake, such as the large but mostly horizontal seafloor displacements and potential splay fault activation, remain puzzling.In this study, we investigate the earthquake using high-resolution 3D dynamic rupture simulations. We use the open-source software SeisSol (www.seissol.org) which is based on an arbitrary high-order accurate DERivative Discontinuous Galerkin method (ADER-DG). We account for complex megathrust-splay fault geometries, high resolution topography and bathymetry, 3D subsurface structure and off-fault plasticity. We resolve the full frictional sliding process as well as the seismic wave field in the to-date longest (500 s) and largest (1500 km) dynamic rupture earthquake scenario.We discuss macroscopic earthquake source characteristics inferred from a smoothly initialized model, in which we prescribe regional, smoothly varying fault pre-stresses constrained by earthquake focal mechanisms, interacting with frictional fault conditions changing only depth-dependently. The interplay of complex fault geometry and simple pre-stress state yields good agreement of ground-deformations in the near field and very long-period teleseismic data. We are specifically interested in the effects of off-fault plasticity: overall reducing rupture speed but increasing or decreasing fault slip depending on regional stress orientation in conjunction with the fault geometry. We then refine the regional stress state based on static inversion and mechanical viability arguments to fully capture dynamic features such as lower slip release in the Southern megathrust and slow rupture velocity at around 8° latitude. Our models are verified by generating synthetic geodetic, teleseismic and tsunami records, employing off-line coupling of SeisSol to established global seismic wave propagation and tsunami simulation tools. [ABSTRACT FROM AUTHOR]