The Hayward fault in California's San Francisco Bay area produces large earthquakes, with the last occurring in 1868. We examine how physics‐based dynamic rupture modeling can be used to numerically simulate large earthquakes on not only the Hayward fault, but also its connected companions to the north and south, the Rodgers Creek and Calaveras faults. Equipped with a wealth of images of this fault system, including those of its 3D geology and 3D geometry, in addition to inferences about its interseismic creep‐rate pattern and rock‐friction behavior, we use a finite‐element computer code to perform 3D dynamic earthquake rupture simulations. We find that the rock properties affect the locations and amount of slip produced in our simulated large earthquakes. Crucial factors that control rupture behavior in our modeling are the earthquake nucleation locations, the fault geometry, and the data that reveal where the fault system is creeping or locked. Our findings suggest that large Rodgers Creek‐Hayward‐Calaveras‐Northern Calaveras (RC‐H‐C‐NC) fault‐system earthquakes may result from dynamic rupture that starts in a locked part of the fault system, but is then stopped by the creeping parts, leading to high‐magnitude‐6 earthquakes; or, from dynamic rupture that starts in a locked part of the fault system, then cascades through some of the creeping parts, leading to magnitude‐7 earthquakes.

©2019. American Geophysical Union. All Rights Reserved. The 14 November 2016 Mw7.8 Kaikōura earthquake in the northern South Island, New Zealand, involved highly complex, multifault rupture. We combine data from a temporary network and the permanent national seismograph network to repick and relocate ~2,700 aftershocks of M≥3 that occurred between 14 November 2016 and 13 May 2017. Automatic phase-picking is carried out using REST, a newly developed hybrid method whose pick quality is assessed by comparing automatic picks for a subset of 138 events with analysts' picks. Aftershock hypocenters computed from high-quality REST picks and a 3-D velocity model cluster almost exclusively in the shallow crust of the upper plate and reveal linkages at depth between surface-rupturing fault segments. Only eight aftershocks are relocated on a deeper structure positioned between patches of geodetically detected afterslip. This indicates that afterslip has not triggered significant earthquake activity on the subduction interface during the period of aftershock activity analyzed.

We utilise seismic data from the central section of the Alpine Fault to locate earthquakes and image crustal structure in three dimensions. Tomography results from c. 6500 sources reveal the fault as either a southeast-dipping low-velocity zone or a marked velocity contrast in different parts of the study region. Where our model is best resolved, we interpret the Alpine Fault to be listric in nature, dipping steeply in the upper crust (50-60 degrees) and flattening to 25-30 degrees in the lower crust. The base of the seismogenic zone shallows from c. 15km beneath the footwall and Alpine Fault to c. 6km beneath the Southern Alps Main Divide, and then deepens to c. 15km by c. 10km further southeast. The shallow brittle-ductile transition overlies a broad low-velocity zone, which together likely result from the presence of fluids and elevated temperatures brought about by enhanced exhumation rate in this section of the Alpine Fault.

P>We map the b-value in the subduction zone of the Wellington region, New Zealand, using a high quality earthquake catalogue relocated with a 3-D seismic velocity model, consisting of 50 314 events that occurred between 1990 and 2005. In order to investigate heterogeneity in the crust of the overlying plate and in the upper plane of the Wadati-Benioff Zone (WBZ), we analyse a series of cross-sections perpendicular to the strike of the subduction zone. We calculate the b-values selecting events with magnitude of completeness >= 2.4 and depth < 65 km and projecting the seismicity within 20 km on each side of the cross-sectional planes. We observe areas of high b-value (similar to 1.7) near the plate interface and regions of low b-value anomalies are detected both in the WBZ in the northwest region below 40 km depth and in the overlying plate in the northern South Island at 10 km depth. The anomalies are statistically significant based on Utsu's p-test and the bootstrap method and are not data processing method or parameter dependent.We compare the b-value distribution with previously determined 3-D distributions of V-p, V-p/V-s and Q(p) from seismic tomography. This comparison suggests that material inhomogeneity, caused by fluid filled cracks resulting from dehydration of the subducted slab and subducted sediments, is the predominant cause of b-value variation in the shallow part of this subduction zone. Our observations are consistent with a previously proposed conceptual model that fluid distribution in the shallow part of this subduction zone is controlled by the permeability of geological terranes in the overlying plate.

Geological observations require that episodic slip on the Alpine fault averages to a long-term displacement rate of 2-3 cm/yr. Patterns of seismicity and geodetic strain suggest the fault is locked above a depth of 6-12 km and will probably fail during an earthquake. High pore-fluid pressures in the deeper fault zone are inferred from low seismic P-wave velocity and high electrical conductivity in central South Island, and may limit the seismogenic zone east of the Alpine fault to depths as shallow as 6 km. A simplified dynamic rupture model suggests an episode of aseismic slip at depth may not inhibit later propagation of a fully developed earthquake rupture. Although it is difficult to resolve surface displacement during an ancient earthquake from displacements that occurred in the months and years that immediately surround the event, sufficient data exist to evaluate the extent of the last three Alpine fault ruptures: the 1717 AD event is inferred to have ruptured a 300-500 km length of fault; the 1620 AD event ruptured 200-300 km; and the 1430 AD event ruptured 350-600 km. The geologically estimated moment magnitudes are 7.9 +/- 0.3, 7.6 +/- 0.3, and 7.9 +/- 0.4, respectively. We conclude that large earthquakes (M-w>7) on the Alpine fault will almost certainly occur in future, and it is realistic to expect some great earthquakes (M-w >= 8).

The central part of South Island of New Zealand is a product of the transpressive continental collision of the Pacific and Australian plates during the past 5 million years, prior to which the plate boundary was largely transcurrent for over 10 My. Subduction occurs at the north (west dipping) and south (east dipping) of South Island. The deformation is largely accommodated by the ramping up of the Pacific plate over the Australian plate and near-symmetric mantle shortening. The initial asymmetric crustal deformation may be the result of an initial difference in lithospheric strength or an inherited suture resulting from earlier plate motions. Delamination of the Pacific plate occurs resulting in the uplift and exposure of mid-crustal rocks at the plate boundary fault (Alpine fault) to form a foreland mountain chain. In addition, an asymmetric crustal root (additional 8 - 17 km) is formed, with an underlying mantle downwarp. The crustal root, which thickens southwards, comprises the delaminated lower crust and a thickened overlying middle crust. Lower crust is variable in thickness along the orogen, which may arise from convergence in, and lower lithosphere extrusion along, the orogen. Low velocity zones in the crust occur adjacent to the plate boundary (Alpine fault) in the Australian and Pacific plates, where they are attributed to fracturing of the upper crust as a result of flexural bending for the Australian plate and to high pressure fluids in the crust derived from prograde metamorphism of the crustal rocks for the Pacific plate.

We examine the expected elastic and inelastic strain accumulation and release across the transpressive Alpine Fault of New Zealand using a finite element model, which incorporates a frictional fault and material behaviour in the crust based on geophysical measurements and rock mechanics experiments. A zone of localized ductile creeping is predicted beneath the part of the fault that slips seismically. Localized creeping results from (a) thermal weakening along the fault ramp and (b) enhanced stresses due to interaction with the shallow, brittle fault above. The creeping zone controls the shorter-wavelength feature seen in the GPS-determined surface velocities. Using the model we show that: (1) seismic faulting loads the creeping shear zone elastically, transferring stress to the mid-crust, and enhancing creep localization in the lower crust; (2) seismic loading and additional thermal weakening, due to long-term advection and exhumation of rocks along the Alpine Fault, can explain the shallow depth and localization of creeping below the Alpine Fault several hundred years after an earthquake and (3) model surface velocities are in good agreement with present-day velocities determined from campaign GPS measurements. Model strain rates show a transient response after faulting that lasts for the first ca. 20 per cent of the interseismic period (ca. 100 yr), with a gradual broadening in the surface velocity signal and almost constant strain rates thereafter until the next faulting event occurs.

We show that the strain rate pattern measured at the surface of an oblique collision zone, the New Zealand Southern Alps, can be used to impose bounds on the width of present-day deformation in underlying mantle lithosphere. Geological and geophysical evidence supporting localized shearing along the downdip extension of the Alpine Fault, is used to constrain the upper slipping patch. Using finite-element models, we investigate the effect of different boundary conditions in the lower lithosphere on the predicted interseismic velocity profile at the surface. For a homogeneous elastic rheology, distributed lower lithosphere shear over widths greater than about 100 km causes the fit with interseismic velocities derived from Global Positioning System (GPS) measurements to significantly degrade. In comparison, either localized shear along a slipping patch or distributed shear over a width less than 100 km can fit the data. The strike-parallel velocity component is most diagnostic between the differing models. Finite-element models incorporating variations in elastic strength and non-linear viscous creep in the crust and mantle lithosphere, while indicating that some decoupling can occur between crust and mantle lithosphere, also support the conclusion that mantle lithosphere cannot be deforming over a width greater than about 100 km. In contrast, shear-wave splitting studies have been interpreted to show a wide, 200-400-km deforming zone beneath the Southern Alps. The discrepancy highlights the different timescales measured by GPS geodesy compared to estimates of finite strain that has accumulated over millions of years.

Recent dense deployments of portable digital seismographs have provided excellent control on earthquakes beneath the central North Island of New Zealand. Here we use a subset of the best-recorded earthquakes in an inversion for the 3-D Vp and Vp/Vs structure. The data set includes 39 123 P observations and 18 331 S observations from 1239 earthquakes and nine explosions. The subducted plate is imaged as a high Vp, low Vp/Vs feature. Vp within the mantle of the subducted slab is almost always > 8.5 km s(-1), which requires the ca. 120 Myr slab to be unusually cold. The low Vp/Vs within the subducted plate closely parallels the lower plane of the dipping seismic zone. It most likely indicates fluid resulting from dehydration of serpentine in the slab mantle, and the earthquakes themselves are likely to be promoted by dehydration embrittlement. We identify a region with Vp < 8.0 km s(-1) which coincides with the upper plane of the dipping seismic zone and extends to ca. 65 km depth with the subducted Hikurangi Plateau, which is about 17 km thick prior to subduction. The mantle wedge is generally imaged as a low Vp, high Vp/Vs feature. However, there are significant changes evident in the wedge along the strike of the subduction zone. The region where Vp is lowest (7.4 km s(-1)) and Vp/Vs is highest (1.87) occurs at 65 km depth, immediately west of the Taupo caldera. This region is best interpreted as a significant volume of partial melt, produced by the reaction of fluid released by dehydration of the subducted plate with the convecting mantle wedge. The region with lowest Vp, while paralleling the underlying dipping seismic zone, is located about 30 km from the upper surface of the zone. Material with Vp > 8.0 km s(-1) directly above the dipping seismic zone can be interpreted as sinking, entrained with the motion of the subducted slab and forming a viscous blanket that insulates the slab from the high-temperature mantle wedge. Material in the overlying low Vp region can be interpreted as rising within a return flow within the wedge. The volcanic front appears to be controlled by where this dipping low Vp region meets the base of the crust. The thickness of the backarc crust also shows significant variation along strike. In the central Taupo Volcanic Zone (TVZ) the crust is ca. 35 km thick, while southwest of Mt Ruapehu the crust thickens by ca. 10 km. There is no significant low Vp zone in the mantle wedge in this southwestern region, suggesting that this thicker crust has choked off mantle return flow. The seismic tomography results, when combined with constraints on mantle flow from previous shear-wave splitting results, provide a plausible model for both the distribution of volcanism in the central North Island, and the exceptional magmatic productivity of the central TVZ.

Two types of seismic velocity heterogeneity are often observed in a given region: pronounced patterns of anisotropy shown by local earthquake shear wave splitting and seismic velocity structure imaged by local earthquake traveltimes. We seek to combine these types of observations by including anisotropy in 3-D velocity inversion. The fast polarization directions from local earthquake shear wave splitting are used to define an initial anisotropy model. The ray paths are estimated for each of the shear wave splitting observations and the 3-D initial model is determined via an averaging process and a coherence analysis. The initial magnitude of anisotropy at a node is dependent on the number of adjacent ray paths and the consistency of the shear wave splitting anisotropy measurements. The traveltime inversion is parametrized with an isotropic component and two azimuthal anisotropy parameters for each node. The ray paths are defined with approximate 3-D ray tracing, and for each point along a ray path, the velocity is defined as a function of the ray path azimuth at that point. A blended approach is used to damp the perturbations to the magnitude of anisotropy and/or perturbations to the fast direction.Tests with synthetic traveltime data show that the anisotropic velocity inversion method is reliable. Application to the Marlborough region shows variations in anisotropy in the brittle crust, ductile lower crust, mantle and subducted slab. The patterns are consistent with the shear wave splitting observations, but show some additional features. The maximum anisotropy is approximately 12 per cent oriented northeast, between the Awatere and Wairau faults. This high anisotropy may reflect deformation and extensive fracturing in a region of high total shear strain. In the region of interaction of the ductile lower crust with the shallow subducting slab, moderate anisotropy with east-west orientation is imaged. This is consistent with ductile deformation roughly parallel to the plate velocity.

The Wellington region of New Zealand overlies a strongly coupled portion of the shallow subduction interface between the Pacific and Australian plates. To better understand the active deformation in the area, we perform a double-difference (DD) relocation of 6825 local earthquakes that occurred between 1990 and 2001. Using both cross-correlation (CC) and bispectrum (BS) methods, we calculate high-quality waveform-based differential times (WBDTs) for event pairs. We manage to more than triple the S-wave differential time measurements after reasonably estimating many of the unpicked S arrivals. After relocation, the image of the double seismic zone beneath this region is greatly sharpened. We find several northeast-striking linear seismic features near Lake Wairarapa on the North Island, which may result from slip along normal faults within the subducting Pacific Plate. We further search for repeating events that might define creeping patches along the interplate thrust. Using a 20-s data window for CC, we find 287 event clusters with very high waveform similarities. Most of these clusters contain just a doublet that occurred within a short time period. The failure to find repeating earthquakes along the plate boundary in this 12-yr time period is consistent with the hypothesized strong coupling between the Pacific and Australian plates under the Wellington region.

The Mw 6.1 Thompson Sound earthquake occurred on I November 2000, with an epicentre near the Fiordland, New Zealand, coastline (-45.112degrees, 166.952degrees). Aftershocks, recorded on temporary seismographs as well as the National Seismograph Network, define a 12.5 x 12.5 km planar zone, taken as the mainshock fault, with a strike of 175degrees and dip of 65degreesW, ranging in depth from c. 12 to 24 km. This is in accord with the mainshock focal mechanism determined by a body-wave inversion that indicates a rake of c. 590, that is, mainly thrust motion with a component of left-lateral strike-slip. This event follows a series of moderate to large earthquakes in the Doubtful Sound region: Te Anau, 1988, Mw 6.7, depth 60 km; Doubtful Sound, 1989, Mw 6.4, depth 24 km; Secretary Island, 1993, Mw 6.8, depth 22 km. The Secretary Island event was a thrust event near, or on, the subduction interface, with a dip of 30-40degreesSE, and our interpretation is that both the Doubtful and Thompson Sound events were oblique thrusts (with a left-lateral component) above the interface and shoreward of the Secretary Island earthquake. The Coulomb failure stress induced by all three large events prior to the Thompson Sound event would have loaded closer to failure faults such as that at Thompson Sound. Together, the four events induced a pattern of Coulomb failure stress on the nearby Alpine Fault that varies in sign with depth. Overall, little can be said about potential triggering or bringing forward/retarding of a large Alpine Fault event. Strong motion recordings for the Thompson Sound event are few, but peak accelerations are in accord with existing attenuation relationships.

Contraction, strike slip, and extension displacements along the Hikurangi margin northeast of the North Island of New Zealand coincide with large lateral gradients in material properties. We use a finite‐difference code utilizing elastic and elastic‐plastic rheologies to build large‐scale, three‐dimensional numerical models which investigate the influence of material properties on velocity partitioning within oblique subduction zones. Rheological variation in the oblique models is constrained by seismic velocity and attenuation information available for the Hikurangi margin. We compare the effect of weakly versus strongly coupled subduction interfaces on the development of extension and the partitioning of velocity components for orthogonal and oblique convergence and include the effect of ponded sediments beneath the Raukumara Peninsula. Extension and velocity partitioning occur if the subduction interface is weak, but neither develops if the subduction interface is strong. The simple mechanical model incorporating rheological variation based on seismic observations produces kinematics that closely match those published from the Hikurangi margin. These include extension within the Taupo Volcanic Zone, uplift over ponded sediments, and dextral contraction to the south.

It is rare to find earthquakes with depths greater than 30 km in continent-continent collision zones because the mantle lithosphere is usually too hot to enable brittle failure.However, a handful of small, intermediate-depth earthquakes (30-97 km) have been recorded in the continental collision region in central South Island, New Zealand.The earthquakes are not associated with subduction but all lie within or on the margins of thickened crust or uppermost mantle seismic highvelocity anomalies.The largest of the earthquakes has M L 4.0 corresponding to a rupture radius of between 100 and 800 m, providing bounds on the upper limit to the rupture length over which brittle failure is taking place in the deep brittle-plastic transition zone.The earthquake sources may be controlled by large shear strain gradients associated with viscous deformation processes in addition to depressed geotherms.

The material properties of the shallow subduction zone are imaged with data from a passive seismic deployment in the Raukumara Peninsula. Attenuation images are obtained using t* values measured from 2848 spectra of P wave arrivals. The t* values are inverted for frequency‐independent three‐dimensional (3‐D) Qp, using a 3‐D velocity model. Linked nodes in a preliminary 2‐D inversion are used to obtain the most reliable initial model for 3‐D inversions. The results show the benefits of obtaining 3‐D Q in addition to 3‐D Vp and Vp/Vs. Q is related to localized permeability and heterogeneity, and thus 3‐D Q is an important tool for interpreting active tectonic regions. Qp in the shallow overlying plate generally corresponds to the mapped geology. Two high Qp features (>600) represent subducted seamounts or Cretaceous volcanics within the overlying plate. The gradient above the subducting plate is the strongest feature in the Qp model, with the slab having high Qp (600–900). Below 20‐km depth and above the plate interface, there is a zone of 50% decrease in Qp and high Vp/Vs about 10 km thick which represents actively subducting sediment and has distributed microearthquake activity associated with active underplating.

Uppermost mantle seismic structure below the Southern Alps in South Island, New Zealand, is investigated by teleseismic P wave travel time residual inversion. The three‐dimensional tomographic images show a near‐vertical, high‐velocity (2–4%) structure in the uppermost mantle that directly underlies thickened crust along the NNE‐SSW axis of the Southern Alps. The center of the high‐velocity anomaly lies to the east of the Alpine fault which bounds Pacific and Australian plate rocks. The oblique collision of these plates resulted in the uplift of the Southern Alps during the past 5–7 m.y. Also, a high‐velocity anomaly (3–5%) corresponding to the Hikurangi subduction zone lies to the northeast of the Southern Alps anomaly, and low‐velocity anomalies (−3%) underlying parts of northwestern and southern South Island may be signatures of late Tertiary extension and volcanism. The data consist of teleseismic arrival times from the New Zealand National Seismograph Network and arrival times recorded during the 1995–1996 Southern Alps Passive Seismic Experiment. Crustal heterogeneity was accounted for by back projecting the rays through an independently obtained three‐dimensional crustal velocity and Moho depth model. The Southern Alps uppermost mantle velocity anomalies are most simply explained by lithospheric thickening below the center of convergence accompanied by thinning and asthenospheric upwelling adjacent to the region of convergence.

The Alpine fault is the Pacific‐Australian plate boundary in the South Island of New Zealand. This study analyzes 195 earthquakes recorded during the 6 month duration of the Southern Alps Passive Seismic Experiment (SAPSE) in 1995/1996 and two ML 5.0 earthquakes and aftershocks in 1997, which occurred close to the central part of the Alpine fault. Precise earthquake locations are derived by simultaneous inversion for hypocenter parameters, a one‐dimensional velocity model, and station corrections. Together with focal mechanisms calculated using a first motion and amplitude ratio method, these results provide a picture of the seismotectonics in the central South Island over a 6 month period. Moment tensor inversions of three earthquakes provide an independent means of comparison to the focal mechanisms derived using the amplitude/first motion method. To validate our observations over time, we compare the SAPSE seismicity with the seismicity recorded by the New Zealand National Seismic Network (NZNSN) and a local network at Lake Pukaki east of the Southern Alps (6 months versus 8 years). Our study indicates that the Alpine fault releases elastic strain seismically from the surface down to 10–12 km depth between Milford Sound in the south and the Hope fault in the north. The seismicity rate of the Alpine fault is low but comparable to locked sections of the San Andreas fault, with large earthquakes expected. Seismicity decreases north of Bruce Bay at the Alpine fault and within a triangular region along the Alpine fault located between the Hope and Porters Pass fault zones. We interpret this as the result of deformation distributed on the Alpine fault and the Hope and Porters Pass fault zones. The base of the seismogenic zone is fairly uniform at 12 km ± 2 km over large parts of the South Island. The high Alps region has a shallower base of the seismogenic zone, indicating localized elevated temperatures east of the Alpine fault. Most of the study region deforms under a uniform stress field with a maximum principal horizontal shortening direction of 110°–120°, similar to geodetic observations and plate motions. This confirms that the region is not undergoing strain partitioning. The earthquake data show that the deformation away from the Alpine fault is distributed on mainly NNE trending thrust faults and strike‐slip transfer faults with a maximum seismogenic depth of 12 km.