Category: Newsletter articles

Susan Schwartz (UC Santa Cruz), Anne Sheehan (University Colorado, Boulder), Rachel Abercrombie (Boston University)

In the last fifteen years, it has become evident that slow slip events (SSEs) are a common and important part of the subduction process. They produce millimeters to centimeters of surface displacement over days to years that can be measured by geodesy and are often accompanied by seismic tremor and earthquake swarms. Slow slip and tremor have been observed in subduction zones in Cascadia, Japan, Mexico, Alaska, Ecuador, northern Peru, Costa Rica and New Zealand.

The 2014-15 HOBITSS deployment of 24 ocean bottom pressure sensors and fifteen ocean bottom seismometers (OBSs) at the northern Hikurangi margin, New Zealand captured a M7.0 SSE. The vertical deformation data collected were used to image one of the best-resolved slow slip distributions to date, and indicated slip very close, if not all the way to the trench (Wallace et al., 2016). The Fall 2016 GeoPRISMS Newsletter reported on this experiment and how for the first time, ocean bottom pressure recorders successfully mapped a SSE displacement field (Wallace et al., 2016). The HOBITSS results were instrumental in demonstrating that Absolute Pressure Gauges are a valuable tool for seafloor geodesy. Seismologists from UC Santa Cruz, University of Colorado Boulder and Boston University are now using the seismic data collected during the same experiment to evaluate the spatiotemporal relationship between seismicity (both earthquakes and tremor) and the slow slip event and the role that seismic structure plays in controlling slip behavior. One of our primary goals is to determine if slow and fast interplate slip modes spatially overlap or are segregated.

An initial catalog of local earthquakes was constructed and relocated in a New Zealand-derived velocity model to produce a catalog of 2,619 earthquakes ranging in magnitude between 0.5 and 4.7. Locations indicate that Hikurangi seismicity is concentrated in two NE-SW bands, one offshore beneath the Hikurangi trough and outer forearc wedge, and one onshore beneath the eastern Raukumara Peninsula, with a gap in seismicity between the two beneath the inner forearc wedge. We do not find an increase in seismicity during the 2014 slow slip event, though seismicity is slightly higher in the month following the SSE. The majority of earthquakes are within the subducting slab rather than at the plate interface. The few events that locate close to the plate interface were assumed to be thrust events and used as templates in a waveform matching technique to identify similarly located earthquake swarms within the entire dataset.

Like the general seismicity increase in the month following the SSE, repeating families of interplate events (Fig. 1) also cluster in time at the end of the SSE. They are spatially concentrated within the slow slip patch and associated with a well-imaged subducted seamount (Bell et al., 2010).

Tectonic tremor was also identified toward the end and continuing after the slow slip event. Like the interplate earthquake families, tremor is also co-located with slow slip and localized in the vicinity of subducted seamounts (Fig. 2). The subsequent, rather than synchronous occurrence of tremor and interplate earthquakes and slow slip suggests that seamount subduction plays the dominant role in the stress state of the shallow megathrust. While northern Hikurangi seamounts appear to primarily subduct aseismically, their subduction may generate elevated pore-fluid pressures in accumulated underplated sediment packages and a complex, interconnected fracture network such that tremor and microseismicity occur as seismic components of seamount subduction during shallow slow slip. This study indicates that the location of subducted seamounts is strongly correlated with the distribution of SSE-associated tectonic tremor and repeating earthquakes. The seamounts appear to be responsible for slow slip, tremor, and microseismicity rupturing adjacent regions in a range of slip processes.

Ongoing work more fully utilizes the rich data set of local earthquakes and includes analysis of seismic attenuation using the body wave spectra of local earthquakes, local earthquake seismic velocity tomography, and earthquake source parameter analysis including focal mechanisms and seismic moment. Knowing the physical state of the subducting plate interface is important for the slow slip modeling, and our attenuation and velocity tomography models will be key to infer the physical properties and structure in the area where slow slip occurs. For example, recent work revealed large differences between SSE slip inversions that assume homogeneous elastic properties versus those that utilize a more realistic elastic structure (Williams and Wallace, 2015). ■

Rough seas off the shores of New Zealand during the HOBITSS deployment in 2014. Retrieved from the Fall 2016 issue of the GeoPRISMS Newsletter. Photo credit: Justin Ball

Figure 1. Repeating earthquake families identified near the plate interface color coded by time of occurrence of first earthquake in the family. Note the concentration of events where the seamount is subducting and their timing just after the 2014 slow slip event

Figure 2. Tectonic tremor associated with the 2014 Gisborne SSE color-coded by time. Tremor is primarily located on and adjacent to subducted seamounts Puke and S1 and mostly occurs at the end and after slow slip has ended. Triangles indicate the location of the land-based stations of the New Zealand National Seismic Network and HOBITSS OBS used to locate tremor.

MGL1801 Participants – Ryuta Aral (JAMSTEC), Stephen Ball (Univ. of Wisconsin, Madison), Nathan Bangs (UT Austin), Dan Barker (GNS Science), Joel Edwards (UC Santa Cruz), Melissa Gray (Imperial College London), Shuoshuo Han (UT Austin), Harold Leah (Cardiff Univ.), Tim Reston (Univ. of Birmingham), Hannah Tilley (Univ of Hawai’i), and Harold Tobin (Univ. of Wisconsin, Madison)

When the Pacific plate slips beneath the Australian plate along the northern Hikurangi margin, the subduction megathrust does not typically generate earthquakes as it often does in subduction zones. Instead, stress accumulated on the megathrust is released within patches every 2-4 years over a period of weeks in well documented slow slip events (SSEs) (i.e. Wallace and Bevean, 2010). While SSEs are not unique to the Hikurangi margin, SSEs often occur at depths of 30-40 km down dip of the seismogenic zone. This makes them difficult to access and thus difficult to examine the physical conditions that control whether the megathrust slips quickly in regular earthquakes or instead slips slowly. However, the Hikurangi margin has documented regular patterns of SSEs that extend updip along the megathrust to unusually shallow depths of ~2 km below the seafloor. This unusually shallow setting and the well documented distribution of slip makes these SSEs accessible with geophysical tools and even drilling.

From January 6th to February 9th 2018, a team of marine geologists and geophysicists from the US, UK, Japan, and New Zealand sailed on the R/V Langseth to acquire a new 3D seismic reflection data volume across the northern Hikurangi margin offshore of the North Island of New Zealand (Fig. 1). Previous seismic surveys have imaged the subsurface structures and offered hints into the unusual megathrust slip behavior along the Hikurangi margin. Bell et al. (2010) showed large seamounts on the subducting plate that can generate thrust faults within the upper plate, and entrain fluid rich sediments and carry them below the megathrust deep into the subduction zone. It is these impacts on the shallow subduction zone that are thought to generate conditions for high fluid content along the megathrust and fluid migration pathways from the megathrust through the upper plate. It is this fluid supply and flow system that is thought to lead to high fluid pressures and control effective stresses along the megathrust, which are also considered critical controls for slip behavior (Saffer and Wallace, 2015). However, it was also evident from earlier 2D seismic images that this complex setting required 3D data to correctly image the shallow megathrust and upper plate structures. Such high resolution 3D data can map out fluid content and faults to fully characterize this system.

The NZ3D experiment was designed to acquire 3D seismic images to map reflectivity and structures, and it provided an opportunity for a novel wide-angle seismic reflection and refraction component to measure seismic velocities in unprecedented detail and in 3D using full waveform inversion (FWI). The detailed seismic velocity data will reveal rock physical properties and will complement observations of reflectivity and structural geometry seen in 3D seismic images.

In most years this large ambitious geophysical experiment would by itself be a major achievement for any given site; however, the NZ3D project was designed to contribute to larger efforts on the New Zealand primary site that included: The NSF-funded SHIRE active source experiment (Nov–Dec 2017) to examine the crustal scale structure of the Hikurangi margin using ocean bottom seismometers, onland seismic receiver stations, and 2D seismic reflection imaging (p.22); IODP drilling to recover core samples, measure physical properties, and install observatories – Expeditions 372 (Nov 2017–Jan 2018) and 375 (Mar–Apr 2018) (p.16); and other related studies.

During the Langseth cruise we surveyed an area 14 x 60 km from the trench to the shelf across the Expedition 375 drilling transect (Fig. 1). Langseth fired one of two 3,300 in3 airgun arrays every 25 m in flip-flop mode and recorded returns on four 6-km-long, 468-channel seismic streamers spaced at 150 m. We made 62 passes through the survey area, fired 145,924 shots and recorded over 5Tbytes of seismic reflection data. With calm seas during most of the 35 days at sea, few equipment issues, and very few interruptions from protected species, we acquired a high-quality seismic data volume that will enable us to examine reflectivity of the megathrust down to more than 10 km in the area of SSEs and map the geometry of faults and stratigraphic horizons. In order to acquire the data needed for FWI, in December 2017, prior to NZ3D acquisition, the R/V Tangaroa deployed a hundred ocean bottom seismometers (OBSs) provided by JAMSTEC in a randomized grid with nominal 2 x 2 km spacing (Fig. 1). Shots for FWI were also recorded on stations deployed around the Gisborne area specifically for NZ3D and stations that had been deployed initially for SHIRE and remained for NZ3D (Fig. 1). A total of almost 300 onland stations recorded Langseth shots during NZ3D. We were also able to take advantage of the close line spacing during the 3D survey to increase the resolution of multibeam bathymetry and backscatter images across the margin. These data provide some of the best detail of the northern Hikurangi margin seafloor to date.

From here, we will spend the next few years processing the 3D volume (with emphasis on water column multiple removal) and OBS data sets to produce high-quality, detailed 3D images in depth, seismic velocity data, and interpret these results in the context of new results from the coordinated projects. Structures in 3D are already emerging from preliminary results (Fig. 1) and are only going to get better. There are lots of exciting results to come for studies of slow slip along the Hikurangi megathrust. ■

Scientists carefully examining a split core from the main fault zone at Site U1518

Figure 3. Seismic reflection profile crossing the HIkurangi drilling transect showing locations of IODP sites. Red lines highlight recently active thrust faults. Black lines highlight older, inactive faults. Interpretation is simplified from Pecher et al. (2018). See Figure 1 for position of trackline.

Figure 1. Index maps of the Hikurangi study area offshore New Zealand. Enlarged map shows locations of IODP drilling sites and contours of slip associated with recent slow-slip events identified by Wallace et al. (2016) using offshore geodetic data.

IODP personnel transferring a sedimentfilled core liner from the rig floor to the core-cutting area.

IODP personnel working on one of the CORK heads at the moon-pool level of JOIDES Resolution

JOIDES Resolution at the dock in Timura, New Zealand, prior to departure.

MGL1801 Participants – Ryuta Aral (JAMSTEC), Stephen Ball (Univ. of Wisconsin, Madison), Nathan Bangs (UT Austin), Dan Barker (GNS Science), Joel Edwards (UC Santa Cruz), Melissa Gray (Imperial College London), Shuoshuo Han (UT Austin), Harold Leah (Cardiff Univ.), Tim Reston (Univ. of Birmingham), Hannah Tilley (Univ of Hawai’i), and Harold Tobin (Univ. of Wisconsin, Madison)

When the Pacific plate slips beneath the Australian plate along the northern Hikurangi margin, the subduction megathrust does not typically generate earthquakes as it often does in subduction zones. Instead, stress accumulated on the megathrust is released within patches every 2-4 years over a period of weeks in well documented slow slip events (SSEs) (i.e. Wallace and Bevean, 2010). While SSEs are not unique to the Hikurangi margin, SSEs often occur at depths of 30-40 km down dip of the seismogenic zone. This makes them difficult to access and thus difficult to examine the physical conditions that control whether the megathrust slips quickly in regular earthquakes or instead slips slowly. However, the Hikurangi margin has documented regular patterns of SSEs that extend updip along the megathrust to unusually shallow depths of ~2 km below the seafloor. This unusually shallow setting and the well documented distribution of slip makes these SSEs accessible with geophysical tools and even drilling.

From January 6th to February 9th 2018, a team of marine geologists and geophysicists from the US, UK, Japan, and New Zealand sailed on the R/V Langseth to acquire a new 3D seismic reflection data volume across the northern Hikurangi margin offshore of the North Island of New Zealand (Fig. 1). Previous seismic surveys have imaged the subsurface structures and offered hints into the unusual megathrust slip behavior along the Hikurangi margin. Bell et al. (2010) showed large seamounts on the subducting plate that can generate thrust faults within the upper plate, and entrain fluid rich sediments and carry them below the megathrust deep into the subduction zone. It is these impacts on the shallow subduction zone that are thought to generate conditions for high fluid content along the megathrust and fluid migration pathways from the megathrust through the upper plate. It is this fluid supply and flow system that is thought to lead to high fluid pressures and control effective stresses along the megathrust, which are also considered critical controls for slip behavior (Saffer and Wallace, 2015). However, it was also evident from earlier 2D seismic images that this complex setting required 3D data to correctly image the shallow megathrust and upper plate structures. Such high resolution 3D data can map out fluid content and faults to fully characterize this system.

The NZ3D experiment was designed to acquire 3D seismic images to map reflectivity and structures, and it provided an opportunity for a novel wide-angle seismic reflection and refraction component to measure seismic velocities in unprecedented detail and in 3D using full waveform inversion (FWI). The detailed seismic velocity data will reveal rock physical properties and will complement observations of reflectivity and structural geometry seen in 3D seismic images.

In most years this large ambitious geophysical experiment would by itself be a major achievement for any given site; however, the NZ3D project was designed to contribute to larger efforts on the New Zealand primary site that included: The NSF-funded SHIRE active source experiment (Nov–Dec 2017) to examine the crustal scale structure of the Hikurangi margin using ocean bottom seismometers, onland seismic receiver stations, and 2D seismic reflection imaging (p.22); IODP drilling to recover core samples, measure physical properties, and install observatories – Expeditions 372 (Nov 2017–Jan 2018) and 375 (Mar–Apr 2018) (p.16); and other related studies.

During the Langseth cruise we surveyed an area 14 x 60 km from the trench to the shelf across the Expedition 375 drilling transect (Fig. 1). Langseth fired one of two 3,300 in3 airgun arrays every 25 m in flip-flop mode and recorded returns on four 6-km-long, 468-channel seismic streamers spaced at 150 m. We made 62 passes through the survey area, fired 145,924 shots and recorded over 5Tbytes of seismic reflection data. With calm seas during most of the 35 days at sea, few equipment issues, and very few interruptions from protected species, we acquired a high-quality seismic data volume that will enable us to examine reflectivity of the megathrust down to more than 10 km in the area of SSEs and map the geometry of faults and stratigraphic horizons. In order to acquire the data needed for FWI, in December 2017, prior to NZ3D acquisition, the R/V Tangaroa deployed a hundred ocean bottom seismometers (OBSs) provided by JAMSTEC in a randomized grid with nominal 2 x 2 km spacing (Fig. 1). Shots for FWI were also recorded on stations deployed around the Gisborne area specifically for NZ3D and stations that had been deployed initially for SHIRE and remained for NZ3D (Fig. 1). A total of almost 300 onland stations recorded Langseth shots during NZ3D. We were also able to take advantage of the close line spacing during the 3D survey to increase the resolution of multibeam bathymetry and backscatter images across the margin. These data provide some of the best detail of the northern Hikurangi margin seafloor to date.

From here, we will spend the next few years processing the 3D volume (with emphasis on water column multiple removal) and OBS data sets to produce high-quality, detailed 3D images in depth, seismic velocity data, and interpret these results in the context of new results from the coordinated projects. Structures in 3D are already emerging from preliminary results (Fig. 1) and are only going to get better. There are lots of exciting results to come for studies of slow slip along the Hikurangi megathrust. ■

The NZ3D science party, technical staff and crew of the R/V Langseth.

Figure 1. Frontal region of the NZ3D seismic data volume. Preliminary post-stack 3D time migration conducted during MGL1801. Gray shading shows multibeam bathymetry overlying the seismic image. Inset: Bathymetric map of the northern Hikurangi margin. The 3D seismic volume covers a 14 x 60 km area offshore Gisborne, NZ (black rectangle). Gray shading shows outline within the 3D area of the seafloor depicted in the data cube. Red dots: locations of 100 OBS deployed to record Langseth shots during NZ3D; White stars: IODP Expedition 372 and 375 drilling locations; Blue triangles: locations of onshore seismometers deployed for the SHIRE project and left in place to record Langseth shots during NZ3D; Yellow triangles: onshore seismometer locations for instruments deployed for NZ3D.

Jaime D. Barnes (UT Austin), Jeffrey Cullen (UT Austin), Shaun Barker (Univ. of Waikato), Samuele Agostini (Istituto di Geoscienze e Georisorse), Sarah Penniston-Dorland (Univ. of Maryland), John C. Lassiter (UT Austin), Andreas Klügel (Univ. of Bremen), Laura Wallace (UT Austin & GNS Science), Wolfgang Bach (Univ. of Bremen)

Much work has focused on defining the elemental concentration and isotopic composition of subduction zone inputs (e.g., sediments, altered oceanic crust, serpentinites), as well as outputs (e.g., volcanic gases and melt inclusions), in order to assess global cycling of volatiles. However, with most work focused on outputs from the volcanic front, poor constraints on the geochemistry of forearc outputs hamper global volatile flux calculations. In fact, most global flux calculations ignore contributions from the forearc because their outputs and the subducted sources contributing to the fore-arc outputs are poorly constrained (e.g., Barnes et al., 2018). The Hikurangi margin of New Zealand has a largely subaerial forearc which hosts numerous fore-arc seeps and springs (Fig. 1). This portion of the forearc is typically submerged beneath the ocean at most other subduction zones. Easy access to fore-arc springs along the margin allows for quantification of volatile flux and sources through the shallow portion of the subduction system (< 50 km).

In addition, there are dramatic along strike variations in subduction parameters along the length of the Hikurangi margin (Fig. 1). The northern portion of the margin has a thin layer of sediments (~ 1 km thick) and many seamounts on the subducting plate, a steep taper angle (7º to 10º) for the accretionary wedge, and shallow (< 15 km depth) aseismic creep. In marked contrast, the southern portion of the margin has a thick (3 to 6 km thick) package of sediments on the incoming plate, low taper angle (4º to 6º), and undergoes stick-slip behavior (e.g., Wallace et al., 2009). Interestingly, previous studies have documented an overall decrease in the Cl, B, Br, Na, and Sr concentrations in fore-arc spring waters from the north to the south (Giggenbach et al., 1995; Reyes et al., 2010). These observations raise numerous questions, such as: the amount of slab-derived fluid component to the springs; whether fluid sources vary along the length of the margin; and particularly whether any chemical variations in the spring fluids record dehydration metamorphic reactions that may be linked to changes in slip behavior along the margin. In order to address these questions, we have sampled fluids from sixteen cold and two thermal springs from along the Hikurangi margin and analyzed them for their cation and anion concentrations, as well as their B, Li, and Cl stable isotope compositions. Because Li, Cl, and B are highly fluid-mobile elements, their incompatibility limits modification by fluid-rock interaction making them excellent tracers of fluid source.

Data show that Cl, Br, I, Sr, B, Li, and Na concentrations are high in the forearc springs, consistent with previous studies. Most of these elements show a general decrease in concentration from north to south, a high in the central part of the margin, and limited variability through time. Despite the dramatic change in concentration along the margin, there is no corresponding trend in isotopic composition. Cl and B isotope compositions are remarkably consistent along the margin, suggesting fluids dominated by seawater and sedimentary pore fluids. Lithium isotope compositions are highly variable, suggesting fluids sourced from seawater and locally modified by interaction with host rock. High Br/Cl and I/Cl weight ratios also support a dominant seawater and pore fluid source.

The decrease in the absolute volatile concentrations along the margin is therefore not due to changes in subduction parameters (e.g., convergence rate, sediment subduction) altering the fluid source along the strike. In addition, the shift in seismic behavior along the margin is not linked to a change in fluid source within the forearc region. Instead, we hypothesize that the shift in volatile concentrations along the margin is controlled by fluid flux through the upper plate, due to increasing upper plate permeability from south to north. In the northern portion of the margin, the upper plate is undergoing extension, whereas in the southern portion, the upper plate is undergoing transpression (Fig. 1) (Wallace et al., 2004). The extension in the northern section of the margin could increase the permeability of the upper plate allowing for fluid loss along normal faults, and possibly lower fluid pressure within the forearc and near the interface. In contrast, the transpressional regime in the south could decrease the permeability of the upper plate, trapping fluids and increasing fluid pressure in the upper plate (Fagereng and Ellis, 2009). This model of changing permeability from north to south explains the decrease in volatile concentrations in spring fluids along strike. In the south, trapping of expelled seawater and pore fluids in the upper plate will allow the fluids to become diluted by meteoric groundwater, but their isotopic compositions will remain unchanged. Whereas in the north, the seawater and pore fluids will be able to pass through the upper plate with less dilution by groundwater. Seismic tomographic and attenuation data also suggest that more fluids are present in the northern and central portions of the upper plate of the Hikurangi subduction zone, compared to the south (Eberhart-Phillips et al., 2017; Eberhart-Phillips et al., 2005; Eberhart-Phillips et al., 2008). Springs with the highest concentrations of volatile elements are located in regions with some of the highest seismic attenuation (which can be interpreted as abundant inter-connected fluids). It is possible that the fluid pressure conditions in the upper plate may play an important role in seismic behavior along the Hikurangi margin. Higher fluid pressures in the south suppress the transition from brittle to viscous deformation, resulting in a deeper brittle-viscous transition and the occurrence of stick-slip behavior to greater depths (Fagereng and Ellis, 2009; Wallace et al., 2012). Greater structural permeability in the northern Hikurangi margin may allow fluids to bleed off, without building significant overpressures in the forearc — this could lead to a comparatively shallower brittle to viscous transition. This work highlights the role of the upper plate tectonics and permeability in controlling the flow of fluid through the forearc and the geochemical consequences on shallow outputs through the subduction system. ■

Mangapakeha. Photo credit: Jeff Cullen

Evan A. Solomon (University of Washington), Marta Torres (Oregon State University), and Robert Harris (Oregon State University)

Fluid generation, migration, and pore fluid pressure at subduction zones are hypothesized to exert a primary control on the generation of seismicity, low-frequency earthquakes, and slow slip events (SSEs) (e.g. Ranero et al., 2008; Obana and Kodaira, 2009; Saffer and Tobin, 2011; Saffer and Wallace, 2015). The SAFFRONZ (Slow-slip and fluid flow response offshore New Zealand) project addresses the GeoPRISMS Subduction Cycles and Deformation Initiative Science Plan by testing interrelationships among fluid production, fluid flow, and slow slip at the Hikurangi Margin. The recognition of dramatic changes in the along-strike distributions of SSEs and their recurrence intervals, interseismic coupling, inferred pore pressure, and other subduction-related parameters (Fig. 1) have resulted in a concerted international effort to acquire seismological, geodetic, other geophysical, and geomechanical data both onshore and offshore the Hikurangi margin. This effort includes recent scientific ocean drilling, logging, and the deployment of two subseafloor observatories during IODP Expeditions 372 and 375 (see page 16 of this issue), as well as a 3-D seismic reflection survey on the northern margin. (p. 14).

SAFFRONZ will complement and extend these efforts by providing:

A continuous two-year record of fluid flow rates and composition over the timeframe of the next expected SSE,

Information on the present background state and past locations of fluid flow and how they relate to inferred pore fluid overpressure along the plate boundary, and

Comparative geochemical and hydrologic data between the northern and southern sections of the margin.

The SAFFRONZ field program is scheduled for January 10 to February 14, 2019 on the R/V Revelle. Our field strategy employs a nested approach to constrain the margin-wide fluid flow distribution tied to estimated pore pressure evolution along the plate boundary from modeling studies and seismic attributes.

We are specifically targeting fault zones and off-fault locations between the deformation front and the shelf-break (Fig. 2). Site locations will be guided by pre-existing multi- and single-beam sonar data (including seafloor backscatter and water column indicators of gas seepage), 2D/3D seismic reflection data, and real-time water column multi-beam sonar surveys during the research expedition. Violin-bow heat flow measurements and piston coring will guide ROV Jason hydroacoustic surveys, Jason heat flow probe measurements, and the collection of push cores to further identify sites of active fluid discharge. Finally, the ROV surveys will guide the deployment of benthic fluid flow meters (Fig. 3) to generate a record of fluid flow rates and composition over a two-year period – the approximate recurrence interval for SSEs in this region. We anticipate deploying about sixteen benthic fluid flow meters, some co-located with seafloor bottom pressure recorders managed by GNS Science, during the 2019 field program.Although our focus is on the northern margin, the location of shallow SSEs and most research activity, we will also conduct ship and ROV operations and deploy a subset of fluid flow meters in the southern region of the margin.

Comparison of fluid flow pathways, fluid composition, fluid flow rates, and flow transients between the northern and southern areas will provide information on the differences in the nature of dewatering between the accretionary southern portion of the margin hosting deep SSEs and the dominantly non-accretionary northern portion with shallow SSEs. The along-strike comparison will also provide a control (reference) transect to compare regional (i.e. flow in response to SSEs in the north) to other hydrologic phenomena.

From both scientific and societal perspectives, results from this project will contribute to our understanding of fault slip behavior offshore New Zealand that have global implications for the postulated interdependence of temperature, pore pressure, fluid flow, and SSEs at subduction zones. The ability to integrate the SAFFRONZ experiment with concurrent bottom pressure recorder deployments, onshore cGPS data, IODP borehole monitoring experiments, 2D and 3D seismic reflection data, and other data to be collected along the margin in the next few years greatly enhance the outcomes of this project. The synthesis of the results from all the concurrent experiments being conducted at Hikurangi promises to be very exciting and the integration of the hydrologic data produced from this project will result in an unprecedented view of deformation and hydrological responses to slow slip at subduction zones. ■

Figure 1. Overview and tectonic setting of the Hikurangi margin (Barnes et al., 2010). A) New Zealand within the context of the SW Pacific. B) Bathymetry, topography, and active faulting of the onshore and offshore margin. White arrows show plate convergence rate and azimuth. Bold white dashed line shows position of 20 mm/yr slip deficit (Wallace et al., 2004). Magenta areas show region of observed slow slip events. Boxes show locations of our northern and southern survey areas shown in Figure 2. C) Perspective view of the Hikurangi margin (modified from Wallace and Beavan, 2010) illustrating the portions of the subduction interface that undergo stick-slip and aseismic slip in terms of the coupling coefficient. Coupling coefficients of one indicate locked areas and coefficients near zero indicate regions of steady aseismic slip. Green contours show areas of slip in slow slip events since 2002. Along strike variations in margin properties are summarized below the figure.

Figure 2. Seismic transects in the Top) northern field area along the IODP Expedition 375 drilling transect (Saffer et al., 2017), and Bottom) southern field area along seismic line PEG09-23. Locations shown as boxes in Figure 1. Arrows T1-T13 in the seismic reflection profiles are primary fault targets for surveys, sampling, and fluid flow meter deployments. Note that flow meters will also be deployed in off-fault locations to monitor local volumetric strain associated with SSEs.

Figure 3. A) The Mosquito benthic fluid flow meter (Solomon et al., 2008). The Mosquito uses OsmoSamplers and a tracer injection device to continuously measure fluid flow rates and sample for fluid composition at multiple depths beneath the seafloor. Picture shows Mosquito prior to tripping the release plate that pushes the sampling needles into the sediment and during recovery after sampling for one year at Hydrate Ridge, offshore Oregon. B) OSU violin heat flow probe that will be used to further constrain the thermal state of the forearc and to identify locations of fluid transport during the SAFFRONZ experiment. C) Schematic of the CAT meter and a CAT meter deployed in the vicinity of the Mosquito at Hydrate Ridge (Tryon et al., 2002; Brown et al., 2005). The CAT meter uses dilution of a chemical tracer to measure fluid flow through outlet tubing exiting the top of a collection chamber. An osmotic pump injects the tracer at a constant rate into the water stream as it moves through the outlet tubing. Both Mosquitos and CAT meters will be deployed during the SAFFRONZ experiment.

Noel Bartlow (University of Missouri), Laura Wallace (UT Austin & GNS Science), Ryan Yohler (University of Missouri), and Charles Williams (GNS Science)

The New Zealand and Cascadia subduction zones are two GeoPRISMS primary sites that have captured the interest of geophysicists from around the world. Both subduction zones feature significant geological hazards, including the potential for large earthquake ruptures and tsunamis. In New Zealand, the capital city of Wellington sits directly atop a large patch of the main subduction plate interface which is frictionally locked, with the potential to rupture in future earthquakes. In Cascadia, there is also potential for a large earthquake on the main subduction interface, which may impact cities such as Portland, OR and Seattle, WA.

Both New Zealand and Cascadia also host slow slip events (SSEs). Slow slip events consist of slip on the subduction plate interface – just as would occur in an earthquake – except the slip takes place more slowly than it would during an earthquake. Now that slow slip events are widely recognized at many subduction zones, it is critical to better understand their role in the accommodation of plate motion. Many questions also exist regarding implications of slow slip for subduction zone mechanics, and the relationship between slow slip events and seismic slip events on the plate boundary.

In New Zealand, slow slip events have been shown to trigger regular earthquakes up to magnitude 6 (e.g. Wallace et al., 2017). Additionally, some evidence exists from the 2011 Tohoku-Oki Mw 9.0 earthquake (e.g. Ito et al., 2013) and the 2014 Mw 8.1 Iquique earthquake (Ruiz et al., 2014) that slow slip events may be able to trigger damaging megathrust events. Most slow slip events do not trigger earthquakes and we currently cannot differentiate slow slip events that might trigger large earthquakes from those that will not. It is possible, however, that further study will lead to methods that allow slow slip events to be used as reliable earthquake precursors.

PIs Noel Bartlow (Univ. of Missouri) and Laura Wallace (UTIG), along with collaborator Charles Williams (GNS Science New Zealand) and graduate student Ryan Yohler (Univ. of Missouri) are studying slow slip events and frictional locking in New Zealand and Cascadia.

One goal of the project is to create self-consistent catalogs of slow slip events in both subduction zones that capture the time varying behavior of slow slip, including how these events grow and decay and move along the subduction plate interface. The data used for these models consists of land-based geodetic GPS time series, and in New Zealand, we also use vertical deformation of the seafloor recorded for one slow slip event using absolute pressure sensors (Wallace et al., 2016).

Previous time varying slow slip event modeling studies usually assume a uniform, elastic half-space (e.g. Bartlow et al., 2014). These new models utilize spatially-varying elastic properties within the earth based on seismic velocity models in both New Zealand and Cascadia, calculated using the PyLith finite element code. This leads to more accurate models of slip during slow slip events, and therefore, more accurate estimates of the amount of slip taken up in slow slip as opposed to being available for release in future earthquakes. Preliminary models for both New Zealand (Williams et al., 2017) and Cascadia (Bartlow et al., 2017) were shown at the American Geophysical Union 2017 Fall meeting. Additionally, a time-dependent model incorporating both onshore GPS and offshore pressure measurements for the 2014 Gisborne, New Zealand slow slip events was shown at the meeting by graduate student Ryan Yohler (Yohler et al., 2017). This model is shown in Figure 1. This slow slip event occurred near the locations of two historical 1947 earthquakes that caused damaging tsunami waves. This is the first time that seafloor geodetic data have been used in a time-dependent deformation model.

As part of this project, a team led by Laura Wallace, including Bartlow and other collaborators, have recently reported the occurrence of a large, shallow (<15 km), two-week slow slip event at the Northern Hikurangi margin triggered dynamically by passing seismic waves from the November 2016 magnitude 7.8 Kaikōura earthquake, over 600 km away (Fig. 2; Wallace et al., 2017). Long-duration (>1 year), deep (>25 km) slow slip was also triggered at the southern Hikurangi margin (Kapiti region), and afterslip occurred on the subduction interface beneath the northern South Island of New Zealand (Fig. 1).

Triggered slow slip at southern Hikurangi is more likely due to large static stress changes induced by the Kaikōura earthquake, given that area’s closer proximity to the earthquake (Wallace et al., 2018). Prior studies had already identified cases of slow slip events triggering earthquakes, and nearby earthquakes prematurely stopping ongoing slow slip events, but these studies are the first to show that dynamic and/or static stress changes from passing seismic waves may also trigger large-scale, widespread slow slip events. We are still discovering the wealth of possible complex interactions between slow slip events and earthquakes, and what they might mean for hazards. ■

Figure 1. (A) Map of New Zealand. Red box indicates area shown in panels (B) and (C). (B) Total slip on the subduction plate interface during the 2014 slow slip event near Gisborne, New Zealand, from both onshore GPS and offshore ocean bottom pressure data. (C) Snapshots of slip rate on the subduction plate interface during the 2014 slow slip event. The addition of the ocean bottom pressure data catches the onset of the slow slip several days earlier than based only on the onshore GPS data.

Figure 2. Total slow slip (colors) on the subduction plate interface over the year following the 2016 magnitude 7.8 Kaikōura earthquake (Wallace et al., in 2017, 2018). The green lines show the surface traces of faults that ruptured in the earthquake. The black dashed contours represent the subduction plate interface, with depths in kilometers below sea level.