James A. Conder

Asymmetric mantle dynamics in the MELT region of the East Pacific Rise

Abstract The mantle electromagnetic and tomography (MELT) experiment found a surprising degree of asymmetry in the mantle beneath the fast-spreading, southern East Pacific Rise (MELT Seismic Team, Science 280 (1998) 1215–1218; Forsyth et al., Science 280 (1998) 1235–1238; Toomey et al., Science 280 (1998) 1224–1227; Wolfe and Solomon, Science 280 (1998) 1230–1232; Scheirer et al., Science 280 (1998) 1221–1224; Evans et al., Science 286 (1999) 752–756). Pressure-release melting of the upwelling mantle produces magma that migrates to the surface to form a layer of new crust at the spreading center about 6 km thick (Canales et al., Science 280 (1998) 1218–1221). Seismic and electromagnetic measurements demonstrated that the distribution of this melt in the mantle is asymmetric (Forsyth et al., Science 280 (1998) 1235–1238; Toomey et al., Science 280 (1998) 1224–1227; Evans et al., Science 286 (1999) 752–756) at depths of several tens of kilometers, melt is more abundant beneath the Pacific plate to the west of the axis than beneath the Nazca plate to the east. MELT investigators attributed the asymmetry in melt and geophysical properties to several possible factors: asymmetric flow passively driven by coupling to the faster moving Pacific plate; interactions between the spreading center and hotspots of the south Pacific; an off-axis center of dynamic upwelling; and/or anomalous melting of an embedded compositional heterogeneity (MELT Seismic Team, Science 280 (1998) 1215–1218; Forsyth et al., Science 280 (1998) 1235–1238; Toomey et al., Science 280 (1998) 1224–1227; Wolfe and Solomon, Science 280 (1998) 1230–1232; Evans et al., Science 286 (1999) 752–756). Here we demonstrate that passive flow driven by asymmetric plate motion alone is not a sufficient explanation of the anomalies. Asthenospheric flow from hotspots in the Pacific superswell region back to the migrating ridge axis in conjunction with the asymmetric plate motion can create many of the observed anomalies.

Seismic structure beneath the Tonga arc and Lau back-arc basin determined from joint Vp, Vp/Vs tomography

The Tonga arc and associated Lau basin exhibit many geologically important processes that link subduction and mantle flow with plate separation and crustal production. We create seismic tomograms of the Tonga-Lau region by jointly inverting for Vp and Vp/Vs structure using data from the LABATTS ocean bottom seismograph experiment and several island deployments to better constrain dynamic processes in the mantle wedge. Jointly using P and S data can help distinguish between the various mechanisms responsible for seismic velocity anomalies such as temperature and the presence of melt and/or volatiles. Because high attenuation in the wedge limits the S wave data set, we focus on 2-D inversions beneath the linear OBS array where resolution is best and also parameterize the solution in terms of the Vp/Vs ratio. As expected, the subducting slab has fast Vp and Vs and a low Vp/Vs ratio, consistent with the cold downgoing plate. The Central Lau Spreading Center (CLSC) exhibits stronger anomalies in Vp/Vs than in Vp, with the anomalies larger than would be predicted purely by temperature variations. The CLSC anomaly extends >100 km to the west of the axis, suggesting a broad region of melt production driven by passive upwelling from plate separation rather than active upwelling mechanisms. The anomaly is asymmetric about the axis, suggesting that slab-induced corner flow possibly influences mantle dynamics several hundred kilometers away from the arc. There is a strong anomaly beneath the volcanic arc that gradually deepens as it trends toward the back arc, likely outlining a hydrated region of melt production that feeds the volcanic front. Hydration possibly continues throughout the wedge to at least 400 km depth. The Lau ridge exhibits a thicker lithosphere relative to the rest of the Basin, while the Fiji platform likely has a thinner lithosphere than the Lau Ridge from more recent extension. There is also a reasonable likelihood of a small degree of partial melt in the uppermost mantle beneath the platform.

Anomalous seafloor spreading of the Southeast Indian Ridge near the Amsterdam-St. Paul Plateau

The Amsterdam-St. Paul Plateau is bisected by the intermediate-rate spreading Southeast Indian Ridge, and numerous geophysical and tectonic anomalies arise from the interactions of the Amsterdam-St. Paul hotspot and the spreading center. The plate boundary geometry on the hotspot platform evolves rapidly (on timescales <1 Myr), off-axis volcanism is abundant, the seafloor does not deepen away from the axis, and transform faults do not have fracture zone extensions. Away from the hotspot platform the ridge-transform geometry is typical of mid-ocean ridges globally. In contrast, the Amsterdam-St. Paul Plateau spreading segments are shorter, they often overlap each other significantly, and the intervening discontinuities are smaller, more ephemeral, and more migratory. Abyssal hills are smaller and less uniform on the hotspot platform than on neighboring spreading segments. From gravity and isostasy analysis the average thickness of the platform crust is ∼10 km, approximately 50% thicker than that of typical oceanic crust. Most of the isostatic compensation of the hotspot plateau occurs at the Moho or within the lower crust, and the effective elastic thickness of the plateau lithosphere is ∼1.6 km, less than half that of adjacent spreading segments. Away from the platform some transform faults contain intratransform spreading centers; on the platform the two transform faults have valleys which may be depocenters for abundant axial or off-axis volcanism and mass wasting. Although not wellconstrained by magnetics coverage, the Amsterdam-St. Paul hotspot appears to have been “captured” by the Southeast Indian Ridge, enhancing crustal production at the ridge since about 3.5 Ma. Prior to this time the hotspot formed a line of smaller, isolated volcanoes on older Australian plate. The underlying cause for the present-day crustal accretion anomalies is the effect of melt generation from separate sources of mantle upwelling (due to plate spreading and the hotspot) which has a consequent effect of weakening the lithosphere.

Investigation of microearthquake activity following an intraplate teleseismic swarm on the west flank of the Southern East Pacific Rise

Between February 1991 and May 1992, 33 intraplate earthquakes having body wave magnitudes between 4.3 and 6.0 were located on the west flank of the Southern East Pacific Rise by the International Seismological Center. Seven months after the last teleseismic event, we deployed four ocean bottom seismometers at the site of the teleseismic swarm. One hundred and ninety-two microearthquakes were located using P and S travel times of events recorded by three or more instruments during the 16-day deployment. Most of the microearthquakes were in a band about 30 km long and 6 km wide between and parallel to seamount chains. In addition, several events were distributed along a line perpendicular to the main seismicity band and parallel to the ridge axis. The focal depths of the microearthquakes range from 1 to 15 km, and most are between 5 and 12 km, similar to the depth range of the teleseismic events [Hung and Forsyth, 1996]. The composite P wave polarities indicate that the microearthquakes had a variety of focal mechanisms. We developed a new grid-search, inversion technique that utilizes the P wave polarities and the empirically corrected ratios of P and S wave amplitudes to find the focal mechanisms of individual events. Within the acceptable travel time and amplitude misfits, focal solutions are fairly stable. Normal faulting is found in the ridge-parallel seismicity line. The thrust and strike-slip faulting in the main seismicity band is distinctly different from the exclusively normal faulting mechanisms of the teleseismic events. There is no apparent depth dependence of fault types. None of the existing models of the sources of stress (ridge push, thermoelastic stresses, loading by local topographic features, caldera collapse, and north-south extension of the Pacific Plate) provides a satisfactory explanation for both the teleseismic swarm and microearthquakes. We propose a new tectonic scenario. In this scenario, the lithosphere is prestressed by the cooling of the plate. Magma rising from the deeper mantle induces normal faulting ahead of the dike tips in the lower lithosphere, which is already under extensional, thermal stress, producing the larger, teleseismically detected events. Once the dikes propagate into the lithosphere, the region surrounding the dikes behind the tips is compressed by the overpressure of magma. Depending on the geometry of the dikes, the local orientations of the minimum principal stress, and the local weaknesses in the lithosphere, thrust or strike-slip faulting (the microearthquakes) may occur.

Seafloor spreading on the Amsterdam‐St. Paul hotspot plateau

The Amsterdam-St. Paul (ASP) platform on the intermediate rate Southeast Indian Ridge (SEIR) is the only oceanic hotspot plateau outside the Atlantic Ocean containing an active, mid-ocean ridge spreading axis. Because the ASP hotspot is small and remotely located, it has been relatively unstudied, and the ridge axis location in many places near the ASP plateau was previously unknown or ambiguous. We mapped the SEIR out to 1 Ma crust (Jaramillo anomaly) both on and near the ASP platform. We located the spreading center to within a few kilometers, based on side-scan sonar reflectivity. Recent off-platform magnetic anomalies and lineated abyssal hill topography are consistent with a simple spreading history. Off-platform full spreading rates increase from ∼63 km/Myr on segment H to the north of the platform to ∼65.5 km/Myr on segment K to the south. In contrast, inversions of seafloor magnetization based on uniform and variable thickness magnetic source layers reflect a complex on-platform tectonic history with ridge jumps, off-axis volcanism, and propagating rifts. On one section of the ASP plateau the spreading location has stabilized and is beginning to rift the plateau apart, generating symmetric magnetic anomalies and lineated topography for the last several hundred thousand years. The larger, more stable, spreading segments of the ASP platform are aligned with major volcanic edifices, suggesting that along-axis magma flow away from plume-fed centers is an important influence on spreading geometry. Many complex tectonic features observed on the ASP plateau, such as ridge jumps, en echelon, oblique spreading centers, and transforms oblique to the spreading direction, are comparable to features observed on Iceland. The similarities suggest that moderate crustal thickening at an intermediate rate spreading center may have similar effects to pronounced thickening at a slow rate spreading center.

Seismic evidence of effects of water on melt transport in the Lau back-arc mantle

Processes of melt generation and transport beneath back-arc spreading centres are controlled by two endmember mechanisms: decompression melting similar to that at mid-ocean ridges and flux melting resembling that beneath arcs. The Lau Basin, with an abundance of spreading ridges at different distances from the subduction zone, provides an opportunity to distinguish the effects of these two different melting processes on magma production and crust formation. Here we present constraints on the three-dimensional distribution of partial melt inferred from seismic velocities obtained from Rayleigh wave tomography using land and ocean-bottom seismographs. Low seismic velocities beneath the Central Lau Spreading Centre and the northern Eastern Lau Spreading Centre extend deeper and westwards into the back-arc, suggesting that these spreading centres are fed by melting along upwelling zones from the west, and helping to explain geochemical differences with the Valu Fa Ridge to the south, which has no distinct deep low-seismic-velocity anomalies. A region of low S-wave velocity, interpreted as resulting from high melt content, is imaged in the mantle wedge beneath the Central Lau Spreading Centre and the northeastern Lau Basin, even where no active spreading centre currently exists. This low-seismic-velocity anomaly becomes weaker with distance southward along the Eastern Lau Spreading Centre and the Valu Fa Ridge, in contrast to the inferred increase in magmatic productivity. We propose that the anomaly variations result from changes in the efficiency of melt extraction, with the decrease in melt to the south correlating with increased fractional melting and higher water content in the magma. Water released from the slab may greatly reduce the melt viscosity or increase grain size, or both, thereby facilitating melt transport.

Seafloor spreading on the Southeast Indian Ridge over the last one million years: a test of the Capricorn plate hypothesis

Abstract Plate motions in the Indian Ocean are inconsistent with a rigid Indo-Australian plate. An equatorial, diffuse boundary dividing the plate into separate Indian and Australian plates significantly improves the fit of kinematic plate models to the spreading rates, transform azimuths, and earthquake slip vectors on the spreading center boundaries. An additional boundary, further dividing the Australian plate into Australian and Capricorn plates has been proposed to account for much of the remaining inconsistency and the pattern of intraplate earthquakes [J.-Y. Royer, R.G. Gordon, Science 277 (1997) 1268–1274]. The proposed boundary is ∼2000 km wide where it intersects the Southeast Indian Ridge. Several recent geophysical cruises to the Southeast Indian Ridge, including a cruise within the proposed boundary, provide many new data for investigating the validity of the Capricorn plate model. These new observations strongly support the hypothesis that the Capricorn plate exists. Statistical tests of the data from the Southeast Indian Ridge alone are not sufficient to confirm it, but motion about the Rodriguez Triple Junction (RTJ) suggests some non-rigidity in the Antarctica–Australia–Somalia circuit. Inferred deformation with enforced closure about the RTJ leads to an estimate of plate motion consistent with the Capricorn plate model. However, the diffuse Capricorn–Australia boundary does not extend south of the St. Paul Fracture Zone, 800 km narrower than the previously proposed boundary.

Do the 1998 Antarctic Plate earthquake and its aftershocks delineate a plate boundary

Plate motion data along the eastern Southeast Indian Ridge (SEIR) are not fit by a rotation pole consistent with data along the western SEIR. DeMets et al. (1988) looked for, but did not find, evidence of a microplate north of the SEIR to explain the discrepancy. We test the hypothesis that the large (Mw=8.2), strike-slip, 1998 Antarctic earthquake lies on the boundary of a microplate (here termed “Balleny”) south of the SEIR that explains the plate motion discrepancy. A small circle about an Antarctic-Balleny Euler pole, determined by plate closure, fits the east-west band of aftershocks, consistent with a strike-slip boundary. However, the Antarctic-Balleny pole predicts right-lateral slip, while the seismicity exhibits left-lateral, strike-slip motion. The opposite sense of slip not only rules out the Balleny plate hypothesis, but requires deformation elsewhere to accommodate both the plate motion differences along the SEIR and the strain accumulated from the Antarctic event.

Mantle Structure and Flow Patterns Beneath Active Back‐Arc Basins Inferred from Passive Seismic and Electromagnetic Methods

Passive seismic and electromagnetic (EM) imaging methods can provide strong constraints on mantle processes associated with active back-arc spreading systems. Seismic velocity and attenuation structures are strongly sensitive to temperature, melt content, melt-pore geometry, and water content. In contrast, EM imaging is sensitive to the presence of fluids and melt when there is high connectivity and is relatively insensitive to temperature. Both seismic and EM methods can detect anisotropy, which may be caused by lattice-preferred orientation (LPO) or oriented melt or fluid pockets. EM studies of the Mariana Trough show a low-conductivity region in the shallow mantle at depths of less than 70 to 150 km, indicating dry conditions resulting from basaltic melt extraction in the uppermost mantle. Regional seismic waveform studies of four active back-arc basins suggest that the Lau back arc is characterized by the slowest upper mantle velocity, and the Mariana Trough shows the fastest velocity, with primary differences occurring between 40 and 100 km depth. These findings are consistent with petrological evidence that suggests higher mantle temperatures for the Lau system. Seismic tomographic images show large low-velocity and high-attenuation anomalies beneath the Lau Basin, indicating that a broad zone of magma production feeds the back-arc spreading center. Slow velocities extend to at least 250 km depth, and these deep anomalies may result from the release of volatiles transported to depth by hydrous minerals in the slab. Shear wave splitting data from the Lau Basin show southward along-strike mantle flow, in agreement with geochemical data. The existence of along-arc flow patterns in many subduction zones suggests that viscous coupling does not exert a strong control on the mantle flow pattern in back-arc regions.