Search

Many types of slow earthquakes have been discovered at subduction zones around the world. However, the physical process of these slow earthquakes is not well understood. To monitor offshore slow earthquakes, a marine seismic and geodetic experiment was conducted at the Hikurangi subduction margin from May 2014 to June 2015. During this experiment, a large slow slip event (Mw 6.8) occurred directly beneath the ocean bottom seismometer (OBS) network. In this study, S‐wave splitting and polarization analysis methods, which have been previously used on onshore data to investigate tremor and anisotropy, are applied to continuous OBS waveform data to identify tremors that are too small to detect by the envelope cross correlation method. Continuous tremor activity with stable polarization directions is detected at the end of the 2014 slow slip event and continued for about 2 weeks. The tremors are generated around a southwest bend in the slow slip contours and at the landward edge of a subducted seamount. Our findings corroborate a previous interpretation, based on burst‐type repeating earthquakes and intermittent tremor, that localized slow slip and tremor around the seamount was triggered by fluid migration following the large plate boundary slow slip event and indicate tremor occurred continuously rather than as isolated and sporadic individual events. Plain Language Summary: Slow earthquakes and tremor are characterized by slow fault rupture. Their generation mechanism has not been well understood. The slow earthquake itself does not cause damage but may inform forecasts of large earthquakes that could cause strong ground shaking or tsunamis. A large slow slip event occurred on the Hikurangi subduction plate boundary, offshore New Zealand, in 2014 directly beneath a temporary ocean bottom seismometer network. We applied new methods to this data to detect and characterize an earthquake tremor signal. The methods determine the polarization of the wave emanating from the tremor and the fast direction of the anisotropic subsurface structure through which the tremor wave propagates. These parameters enable us to detect small amplitude tremors and to determine their spatial and temporal distribution. We observed continuous tremor activity for about 2 weeks duration while the slow slip event was waning. This tremor activity occurred over a mapped subducted seamount and on the plate boundary, which likely experienced large stress changes due to the slow slip event. Our results are consistent with previous studies that used different methods on the same data, which indicate that the tremor activity in the vicinity of the seamount was triggered by fluid migration. Key Points: S‐wave splitting and polarization analyses of continuous offshore data detect small amplitude tremor activityTremor activity with stable polarization direction started near the end of the 2014 Gisborne slow slip event (SSE) and was continuous for about 2 weeksTremor activity occurred around the landward edge of the subducted seamount and was triggered by fluid migration following the SSE [ABSTRACT FROM AUTHOR]

Abstract: Shallow slow slip events have been well documented offshore Gisborne at the northern Hikurangi subduction margin, New Zealand, and are associated with tectonic tremor downdip of the slow slip patch and increases in local microseismicity. Tremor and seismicity on the shallow subduction interface are often poorly resolved due to their distance from land‐based seismic and geodetic networks. To address this shortcoming, the Hikurangi Ocean Bottom Investigation of Tremor and Slow Slip experiment deployed 24 absolute pressure gauges and 15 ocean bottom seismometers on the seafloor above the Gisborne slow slip patch to investigate the spatial and temporal extent of slow slip and associated tremor and earthquake activity. We present a detailed spatiotemporal analysis of the seismic signatures of various interplate slip processes associated with the September/October 2014 Gisborne slow slip event. Tectonic tremor begins toward the end and continues after the geodetically constrained slow slip event and is localized in the vicinity of two subducted seamounts within and updip of the slow slip patch. The subsequent, rather than synchronous occurrence of tremor suggests that tremor may be triggered by stress changes induced by slow slip. However, Coulomb failure stress change models based on the slow slip distribution fail to predict the location of tremor, suggesting that seamount subduction plays a dominant role in the stress state of the shallow megathrust. This and the observed interplay of seismic and aseismic interplate slip processes imply that stress changes from slow slip play a secondary role in the distribution of associated microseismicity. [ABSTRACT FROM AUTHOR]

We use a prestack depth migration reflection image and magnetic anomaly data across the northern Hikurangi subduction zone, New Zealand, to constrain plate boundary structure and geometry of a subducting seamount in a region of shallow slow slip and recent International Ocean Discovery Program drilling. Our 3‐D model reveals the subducting seamount as a SW‐NE striking, lozenge‐shaped ridge approximately 40 km long and 15 km wide, with relief up to 2.5 km. This seamount broadly correlates with a 20‐km‐wide gap separating two patches of large (>10 cm) slow slip and the locus of tectonic tremor associated with the September–October 2014 Gisborne slow slip event. Largest slow slip magnitudes occurred where the décollement is underlain by a 3.0‐km‐thick zone of highly reflective subducting sediments. Wave speeds within this zone are 7% lower than adjacent and overlying strata, supporting the view that high fluid pressures within subducting sediments may facilitate shallow slow slip along the north Hikurangi margin. Plain Language Summary: Using a suite of geophysical data from the northern Hikurangi margin, New Zealand, we determine the location and geometry of a subducting seamount on the subducting Pacific Plate and establish its spatial relationship with slow slip and tremor that occurred on the plate boundary in September–October 2014. We infer that slow slip appears to occur preferentially where there are sediments with high fluid pressure in pore fluids subducting adjacent to the seamount but is reduced above the seamount itself. This observation has implications for understanding what physical conditions contribute to spatial variation in frictional properties of the plate interface that may control fault slip behavior on large, plate boundary subduction thrusts. Key Points: Seismic images reveal Hikurangi margin accretionary wedge architecture and seismic velocity distributionMagnetic anomaly modeling shows seismic tremor focused on the landward flanks and downdip of subducting seamountsStructural heterogeneity of the plate interface may influence the distribution of slow slip and tremor [ABSTRACT FROM AUTHOR]

• Shear wave splitting used to measure crustal anisotropy during a slow slip event. • Fast azimuth variations near subducted seamount suggest structural control. • Temporal variations in Vp/Vs and delay time suggest some stress or fluid control. In September 2014, a five week long slow slip event (SSE) occurred near Gisborne at the northern Hikurangi subduction zone, New Zealand, and was recorded by offshore instruments deployed by the Hikurangi Ocean Bottom Investigation of Tremor and Slow Slip (HOBITSS) project. Up to 25 cm of slip occurred directly below the HOBITSS array. We calculate shear wave splitting (SWS) and V P / V S ratios for event-station pairs on HOBITSS ocean bottom seismometers and onshore GeoNet seismic stations to determine the relationship in time and space between slow slip and these seismic properties. Spatial averaging of SWS fast azimuths yields trench-perpendicular fast azimuths in some areas, suggesting that compressive stress from plate convergence closes microcracks and controls anisotropy in the upper-plate. Variations from the trench perpendicular directions are observed near a subducting seamount, with directions closely resembling fracture and fault patterns created by subducting seamounts previously observed in both laboratory and field experiments. Temporal variations in fast azimuths are observed at three stations, two of which are located above the seamount, suggesting measurable variations in stress orientations. During the SSE, median V P / V S measurements across all offshore stations increase from 1.817 to 1.894 and SWS delay times decrease from 0.178 s to 0.139 s (both changes are significant within 95% confidence intervals). Temporal variations in V P / V S and delay time are consistent with fluid pressurization below a permeability barrier and movement of fluids during the rupture of a slow-slip patch. [ABSTRACT FROM AUTHOR]