Search

The M 7.8 Kaikoura earthquake occurred in the northern South Island of New Zealand on 3 Nov., 2016, involving the rupture of >20 faults. To understand the complexity of the Kaikoura earthquake, details of the fault geometry, seismic velocity distribution, and stress field are necessary. We have undertaken seismic tomography along the c. 200 km length of the rupture zone. Data from both 51 temporary stations and 22 permanent (GeoNet) stations were collected from March 2011 to December 2018. The hypocenter of the Kaikoura earthquake and aftershocks near the Kekerengu fault locate along lineaments where seismic velocity changes laterally in the epicentral region. In the uppermost crust, lower velocities occur beneath the Emu Plain and Cape Campbell. A higher velocity region near Kaikoura may have acted as a barrier that prevented eastward rupture from the hypocenter and led to the complex fault distribution in this area. These complexities in the seismic velocity structure may relate to the multi-segment rupture character of the Kaikoura earthquake. Spatial correlations between rupture areas and high Vp/Vs suggest the involvement of overpressured fluid in the nucleation and propagation of rupture segments, which is also supported by the reactivation of unfavourably oriented strike-slip ruptures, many lying at c.70° to the regional maximum compressive stress trajectories. [Display omitted] • The aftershock alignments are consistent with most of the corresponding sub-faults. • High seismic velocity body may have acted as a barrier during the rupture process. • High Vp/Vs is present along the aftershock area. • Overpressured fluid likely promoted the multi-fault rupture of the Kaikoura earthquake. [ABSTRACT FROM AUTHOR]

Actively deforming crust and upper mantle around the obliquely convergent Pacific-Australia plate boundary in the northern South Island of New Zealand have been investigated by seismic tomography using data from a temporary c. 50 station network plus GEONET. The Alpine-Wairau Fault (principal component of the plate boundary at the surface) transects the study area, separating the Buller-Nelson (BN) province of active compressional inversion involving steep reverse faulting to the northwest, from the Marlborough fault system (MA) dominated by dextral strike-slip faulting to the southeast. In the course of this study, MA hosted the 2013 Seddon earthquake sequence (M 5.9, 6.6) and the 2016 M 7.8 Kaikoura earthquake. Active fault structures and present seismicity are associated with a heterogeneous distribution of low-velocity and Vp/Vs anomalies. Elsewhere, there is a general association between crustal seismicity and low-velocity zones along major fault structures within both the MA and BN seismotectonic provinces along which high Vp/Vs anomalies are locally conspicuous. Areas of active crustal seismicity are also generally characterized by high Vp/Vs , for instance the hypocentre and aftershocks of the 2016 M 7.8 Kaikoura earthquake overlie a low-velocity region with high Vp/Vs. The coincidence of anomalously low Vp and Vs and anomalously high Vp/Vs with zones of high electrical conductivity defined by a previous MT transect is consistent with the notion that upward migration of overpressured hydrothermal fluid from the subducting slab at depth leads to a heterogeneous distribution of overpressured fluid in and around the base of the crustal seismogenic zone, weakening the overlying crust and promoting seismic rupture along major fault systems. It seems possible that the Association of mid-crustal low-velocity zones with anomalously high Vp/ Vs may diagnose rupture preparation zones where frictional strength is being lowered by the build-up of fluid overpressure concurrent with accumulating shear stress, so that eventual fault failure is 'dual-driven'. • Seismic low-velocity zones beneath both the Buller-Nelson inversion province and the Marlbourough fault system • Hypocenter of the 2016 Kaikoura earthquake within low velocity zone • Seismic low-velocity can be interpreted as fluid overpressured area. [ABSTRACT FROM AUTHOR]

Relative to more mature fault zones, immature fault zones that have accumulated smaller total displacement are characterized by less efficient strain localization and more complicated earthquake ruptures. How differences in maturation are reflected in regional-scale upper-crustal fracturing is not well known. Recently, complicated earthquake ruptures associated with immature fault zones, such as the 2016 Kaikōura earthquake in New Zealand, have occurred in areas that are in regional proximity (<100 km away) to more mature faults. Here we examine whether inefficient strain localization in less mature fault zones is associated with a broader distribution and anomalously elevated concentration of fractures over distances of tens of kilometers. We use regional seismic arrival-time tomography in a broad area around the Kaikōura earthquake to investigate lateral variations in Vp and Vp/Vs. Focusing on the extensively faulted but compositionally uniform Torlesse-Pahau terrane in the Marlborough region where the earthquake occurred, we attribute lateral variations in Vp and Vp/Vs to differences in concentration of fluid-filled fractures. Using numerical models relating seismic velocities and fracturing, we solve for the lateral variation in concentration of ∼0.01 aspect ratio fluid-filled fractures. We find that areas near the Kaikōura rupture have >3% elevated fracture porosity compared to the adjacent area to the north. The elevated regional fracturing in the Kaikōura area is interpreted to result from more distributed deformation, and broader distribution of earthquakes, due to inefficient localization of strain from a regionally uniform strain rate field, highlighting the relationship between relative maturity of upper-crustal fault zones and lateral variability of regional upper-crustal properties. Plain language summary: When earthquakes occur in immature fault zones (areas that have accumulated small total displacement), they tend to rupture multiple poorly developed faults of diverse orientations. In contrast, more mature fault zones are associated with more developed, smoother faults and more localized earthquake activity. How these differences in maturation are reflected in the regional distribution of fractures is not well known. Here we use the arrival times of P and S waves from earthquakes to constrain seismic velocities (Vp and Vp/Vs) and use numerical models to relate seismic velocities to fracture concentration. We focus on the Marlborough region of the South Island, New Zealand, where immature fault zones recently ruptured during the 2016 Kaikōura earthquake but which also has more mature fault zones <100 km away. We find elevated fracture concentrations (3% higher fracture porosity), indicative of more distributed deformation along immature fault zones relative to more mature fault zones. In immature settings, fracturing is not as effectively localized along individual fault traces, leading to a broadly distributed fracture network. • We image upper-crustal Vp and Vp/Vs, and interpret their variations in terms of amount of fluid-filled fractures. • The Marlborough region, which experiences near-uniform shear, has variable fluid-filled fracture concentration. • Within the Marlborough region, areas of less mature faulting have more fluid-filled fractures than more mature areas. [ABSTRACT FROM AUTHOR]

To understand the stress controls on the occurrence of a multi-fault rupture, we estimated the crustal stress between April 2013 to December 2018, i.e., before and after the Mw7.8 Kaikōura earthquake that occurred in New Zealand on 13 November 2016. We used both the focal mechanism solutions from the temporary seismic networks and the GeoNet moment tensor solutions and selected the solutions that differed significantly from the mainshock fault planes and rakes. Then, we performed stress tensor inversions for the selected focal mechanism solutions. Using the stress tensor inversion results, we also calculated the slip tendency. Prior to the Kaikōura earthquake, the stress regime was the strike-slip type, and the maximum eigenvalue of the stress tensor (σ 1) was oriented WNW–ESE. The stress field orientation did not change significantly after the earthquake. This suggests that the stress change during the Kaikōura earthquake was too small to alter the stress orientations, implying that there may have been large differential stress prior to the Kaikōura earthquake. However, the average stress ratio in different clusters changed in two different patterns after the earthquake, suggesting possible changes in the magnitude of different components of the stress tensor, or of pore pressure in different regions. A high slip tendency was observed at the hypocentre, while a low slip tendency was observed at the northern end of the Kaikōura earthquake faults. This may suggest that the stress orientation and the stress ratio controlled the initiation and the end of the multi-fault rupture. These results corroborate previous fault propagation models. • Little or no coseismic change in stress orientation occurred due to the 2016 Kaikōura earthquake. • A large differential stress value was observed prior to the Kaikōura earthquake. • A high slip tendency prior to the earthquake was observed along the epicentral fault • A low slip tendency prior to the earthquake was observed at the northern end of the rupture zone. [ABSTRACT FROM AUTHOR]