Search

Temperature is used to trace ocean density variations, and reveals internal waves and turbulent motions in the deep ocean, called ‘internal motions.’ Ambient temperature detected by geophysical differential pressure gauges (DPGs) may provide year-long, complementary observations. Here, we use data from four DPGs fixed on the ocean bottom and a high-resolution temperature sensor (T-sensor) 13 m above the seafloor as a square-kilometer array deployed offshore ~ 50 km east of Taiwan facing the open Pacific Ocean to examine the impact of temperature on DPG signals related to internal motions. The DPG signals correlate with T-sensor temperature variations between 0.002 and 0.1 mHz, but have time shifts partially caused by slow thermal conduction from the ambient seafloor to the DPG chamber and partially by internal motion propagation time across the array. Applying beamforming-frequency-wavenumber analysis and linear regression to the arrayed T-sensor and DPG data, we estimate the propagating slowness of the internal motions to be between 0.5 and 7.4 s m−1 from the northwest and northeast quadrants of the array. The thermal relaxation time of the DPGs is within 103–104 s. This work shows that a systematic scan of DPG data at frequencies

In recent years, Ocean Bottom Seismometers (OBSs) have become widely used to expand the coverage of seismic networks onto the ocean. This study takes advantage of offshore observations at the northern end of the Hikurangi margin and southwestern Okinawa Trough to study the tectonics in both regions.In the Hikurangi subduction zone, slow slip events (SSEs) have been observed, which are caused by the subduction of the Pacific Plate under New Zealand. The behaviour of SSEs and how they influence the physical properties of Earth materials are open to question. From 2014 to 2015, 15 OBSs were deployed offshore Gisborne, New Zealand on the Hikurangi margin. Ambient noise data from the OBSs are used to study velocity changes related to SSEs. Single station cross-component correlations and auto-correlations are computed, from which coda waves are used to monitor the velocity changes before, during and after the SSEs in 2014 and 2015 to analyse the slow earthquake behaviour and its relation to stress changes. Different rotation on horizontal components is tested by rotating horizontal components to N-E direction and parallel-perpendicular to the coastline. The dv/v computed by different components or rotation show different changes. The averaged dv/v displays a 0.1% velocity decrease during the SSE in October 2014.The southwestern Okinawa Trough tapers towards Taiwan. How the back-arc crust accommodates the narrowing processing remains to be understood. At various times between 2010 and 2017, 22 OBSs on a small scale (~0.2°×0.3°) were deployed in Southwestern Okinawa Trough offshore northeast Taiwan. Ambient noise recorded on vertical velocity and pressure sensors is used to retrieve Scholte waves for studying shear wave velocity structure. Phase velocities are forward-modeled according to a model proposed by Kuo et al. 2015 and shear strength and density results from ODP1202. Phase velocity dispersion curves are measured from cross-correlations and unwrapped according to the modeled phase velocities. The fundamental mode phase velocities averaged from different station pairs are 0.62 km/s at 3 s period and 1.56 km/s at 6.5 s period. A 3D inversion will be conducted for a shear wave velocity structure from the basin center to the edge.

Seismic attenuation accompanying the velocity structures demonstrates the variations of the physical and chemical properties of the earth. The t* measurement using the seismic body wave spectrum, however, typically encounters the trade-off of corner frequency, t*, and site effect. Ko et al, [2012] proposed the cluster event method (CEM) that reduced the model parameter numbers by grouping the spatial-closed enough events for those traveling to each station along the adjacent paths and sharing one t*. Yet, the site effects among different stations collected in the same cluster bring the challenges on fitting all spectrum. We adapt the cluster strategy to group multiple nearby events recorded by one station only. Moreover, the new iterative CEM algorithm includes both the spectrum and spectral ratio data which provide constraints on seismic moments and corner frequencies of each earthquake inside the cluster, respectively. The final t* and corner frequencies are determined again by including the side effects which are averaging from spectrum residuals in the initial CEM stage. We applied the iterative CEM for earthquakes recorded at dense deployed F-net and Hi-net by NIED in the Tohoku area, Japan. The multitaper spectrums are retrieved from direct P waves with coda wavetrains tapered. Combining the spectral ratio and spectrum data with proper weightings, our new approach increases the stability of t* measurements contributed from better constrains on the corner frequency estimations.

The intermediate-depth seismicity below the Hindu-Kush orogen is thought to mark the Indian-plate subduction with the bottom half of the slab currently breaking off. Unique features of this continental subduction are the near-vertical slab and the roughly stationary convergence boundary. How this subduction affects the mantle flow patterns remains to be understood. In this study we measured source-side shear wave splitting on the S waves from Hindu Kush intraslab events to sample the surrounding mantle. The observed fast polarization directions exhibit a circular pattern around the slab resembling that predicted for the toroidal flow driven by slab rollback. However, the rollback scenario is not favored because it hardly sustains in dynamic models without a considerable retreat of convergence boundary. We propose that the observed pattern is produced by the sub-vertical shear flow entrained by the steep descent of the slab and the ongoing breakoff. This scenario requires the existence of A-type or AG-type olivine fabrics with strong orthorhombic anisotropy in mid- to lower upper mantle, which is consistent with the global models of azimuthal and radial anisotropy. This interpretation circumvents the debate on the cause of trench-parallel anisotropy in some oceanic subduction zones where slab entrainment and rollback may coexist, and supports the notion that orthorhombic anisotropy of olivine may play an important role in shaping mantle anisotropy.

How the underlying mantle responds to the subduction of an oceanic plate remains controversial. Numerical models demonstrate the presence of 3D flow in the sub-slab mantle with a poloidal component induced by slab entrainment and a toroidal component driven by slab rollback. These flows produce plate–motion–parallel and trench-parallel fast directions of seismic anisotropy, respectively, allowing them to be distinguished via the anisotropy pattern. However, seismology studies have so far documented the dominance of either, but not both, type of flow in most subduction zones, evoking debates concerning the flow mechanisms. We employed source-side anisotropy technique to map the sub–slab mantle of the Cocos subduction zone. The observed anisotropy pattern can be best explained by a plate–motion–parallel anisotropy layer, roughly 130 km thick, underlain by a trench-oblique anisotropy regime. We interpret this phenomenon as an entrained poloidal flow transitioning to a rollback–induced toroidal flow modified to suit the regional tectonics. This finding reconciles the current debate over which flow prevails behind a subduction zone. The upper, entrained layer in our model extends below the lithosphere–asthenosphere boundary previously defined in the vicinity of the middle America trench, suggesting strong coupling along the lithosphere–asthenosphere boundary in this region.

The lowermost 100–300 km of the Earth's mantle commonly regarded as the thermal boundary layer (TBL) of mantle circulation is characterized by its complex physical properties. Beneath the Caribbean this so-called D″ layer features relatively high velocities and abrupt impedance increase at the top (designated as the D″ discontinuity). These seismic characteristics have been attributed to the accumulation of ancient subducted slab material and the phase transition in the major lower mantle mineral of pervoskite. Geodynamic models predict that the blanketing cold slabs may trap enough heat from core to be buoyantly destabilized, and eventually broken apart and entrained into the bottom of the convection cell. Here we explore the D″ structure with unprecedented resolution through modeling traveltimes, amplitudes, and waveform shapes from the USArray. We find an east-to-west asymmetrical undulation of the D″ discontinuity with a V-shaped depression of ∼70–160 km over a lateral distance of 600 km beneath northern South America. The shear velocity perturbations vary in the same trend showing the most pronounced reduction of ∼3–4% below the thinnest D″ layer in close proximity to an intermittently undetected discontinuity. The strong correlation between the D″ topography and velocity variations indicates the phase transition boundary has been perturbed or even disrupted by the large lateral temperature gradient of slab material which has been reheated from the core over extended periods of time.

T waves are conventionally defined as seismically generated acoustic energy propagating horizontally over long distances within the minimum sound speed layer in the ocean (SOFAR axis minimum). However, T waves have also been observed by ocean-bottom seismometers in ocean basins at depths greater than the SOFAR axis minimum. Previously, non-geometrical processes, such as local scattering at rough seafloor and water-sediment interface coupling, have been proposed as possible mechanisms for deep seafloor detection of T waves. Here, we employ a new T-wave modeling approach based on hydroacoustic ray theory to demonstrate that seismoacoustic energy can propagate to reach deep seafloor, previously considered as shadow zone of acoustic propagation. Our new hydroacoustic simulations explain well the observations of T waves on ocean-bottom seismometers at deep ocean basins east of Taiwan, and shed new light on the mechanism for deep ocean T-wave propagation.

Lower crustal deformation takes a turn Collisions creating mountain belts frequently involve a tectonic plate plunging into the mantle. Huang et al. connect the deformation of rock from the subducting plate to the surface topography in Taiwan (see the Perspective by Long). Subsurface deformation mapping required interpreting certain seismic wave velocities as they travel through the crust. The subsequent images of Taiwan's deep crust show two distinct layers of deformation. The bottom layer comprises the subducting slab, which is being pulled into the mantle. This mechanically couples with the upper layer of crust, compressing it into a mountain range. Science , this issue p. 720 ; see also p. 687

The elastic response of seafloor to ocean gravity wave loading, or seafloor compliance, provides a constraint on the elastic properties of the crust. We measured seafloor compliance at three ocean bottom seismometer (OBS) sites around Taiwan—two in the southwestern (SW) Okinawa Trough and one on the Ryukyu arc—and performed inversion for crustal structures beneath them. Models best fitting the data demonstrate a decrease in upper crustal shear velocity and an increase in the compressional/shear velocity ratio from the arc site to the trough sites with increasing amount of back-arc extension. This variation suggests that the upper continental crust is highly faulted and hydrated during rifting of the Eurasian lithosphere.

Vertical ocean‐bottom seismograph (OBS) data at frequencies below 0.05 Hz are contaminated by noise induced by infragravity waves. We constructed the transfer function between pressure and velocity data from OBSs deployed in Taiwan waters to remove the wave pressure‐induced noise from seismic recordings. Data were analyzed from five portable broadband OBSs deployed each for 10 months at water depths from 1740 to 4600 m and from a cabled, shallow‐buried seismograph (EOS1) installed on the seafloor at 300 m depth. Removing long‐period noise from these OBS data improves the identification of teleseismic phases such as P , S , SS , P diff , and PKIKP that are otherwise ambiguous or unidentifiable. For EOS1, infragravity‐wave signals completely mask the P and S waveforms in the 10–50 s period band suitable for centroid moment tensor (CMT) solutions for most of the local events. Application of the transfer functions to predict and remove wave deformation yielded clean prominent P and S waveforms at these periods and aided in the CMT determination for small events jointly with land stations. The relative amplitudes of the wavenumber‐normalized transfer function for some of the OBSs are mostly determined by the thickness of the sediment at the OBS site.

Robust determination of earthquake source parameters over a continuous depth range is central to inferring rupture mechanisms dominant at different depths. We employed a cluster-event method to constrain the source parameters as well as along-path attenuation for earthquakes over 0–150 km depths and 4 orders of seismic moments in the Japan subduction zone. We found that corner frequency and stress drop increase with depth, whereas the radiated energy scaled by seismic moment declines with depth slightly. As a result, the radiation efficiency exhibits a notable deficit for events deeper than 60 km. Together these suggest an increased energy dissipation during faulting in ductile deformation regime, consistent with shear heating instability as an important faulting mechanism for intermediate-depth earthquakes.

[1] The Okinawa trough in the Ryukyu subduction system is one of a few actively extending back arc basins within the continental lithosphere. Recent shear-wave splitting measurements show variable fast directions along the trough suggesting a complex three-dimensional mantle flow field. In this study we use numerical subduction models to demonstrate that the thickness variations of the continental lithosphere bounding the edge of a subduction zone can result in complicated mantle circulation and regional dynamics. The model results for the southern Ryukyu show a combination of two effects: Along the Okinawa trough, the increasing thickness of the overriding Eurasian plate toward Taiwan due to a gradually diminishing rifting induces pressure gradients that drive trench-parallel flow to the edge in the shallow portion of the mantle wedge, and the thick Eurasian lithosphere to the west effectively blocks and diverts this along-arc flow and limits the toroidal flow around the slab edge below it. The toroidal flow thus enters the mantle wedge at depths of more than about 100 km, opposite in direction with and largely below the along-arc flow. These combined geometry effects of the Eurasian lithosphere create an intricate three-dimensional flow structure at the southern edge of the Ryukyu subduction zone. Model predictions for lattice preferred orientations of olivine aggregates show rotation patterns that agree with the observed shear-wave splitting patterns. This three-dimensional scenario echoes the geochemical and seismological evidence worldwide that indicates complex, depth-varying mantle circulations in subduction systems.

Intraslab events recorded by ocean-bottom seismometers in the Okinawa trough provide an extended depiction of shear wave splitting in the southwest section of the Ryukyu subduction zone. At 100–200 km from the western edge of the subduction system, we observed trench-normal fast polarization direction in the back arc compatible with 2D slab- or rifting-driven corner flow. Towards the edge, the fast directions are sub-parallel to the trench in the arc—back arc region, and rotate to trench-normal within 50 km of the edge. Splitting constrained by land stations with paths mostly in the mantle wedge exhibits similar trench-normal fast directions in the subduction edge zone. Further inland, the dominant component of fast directions becomes roughly parallel to the Taiwan orogenic fabric. Splitting of P-to-S phases converted at the Moho of the Okinawa trough and of S phases from shallow events suggest that crustal anisotropy may affect the measured splitting, but the observed pattern reflects predominantly mantle anisotropy. The variation in splitting along the Okinawa trough cannot be explained by a B-type–A-type olivine fabric transition in the mantle wedge. It may indicate the presence of an along-arc flow in the mantle wedge towards the edge where it is blocked and deflected by the Eurasian lithosphere. This scenario bolsters previous studies suggesting a significant impact of Eurasian lithosphere on the dynamics of the Ryukyu subduction system.

The lateral edge of a subduction zone is usually depicted as an opening to the asthenosphere where invigorated dynamics and amplified magmatism take place. In this study we present evidence from seismic data suggesting the presence of a cold and dynamically sluggish edge environment at the southwest end of the Ryukyu subduction system. We measured attenuation, or 1/ Q , for P waves from subduction zone events at ∼100 km depths received by OBSs in the Okinawa trough and land stations in NE Taiwan. In the Okinawa trough 100–200 km from the edge, Q values are lower than 100. In the vicinity of the edge, Q values increase from 100 to over 1000 towards Taiwan. To reconcile arguments from geophysical and geochemical observations, we propose that the mantle wedge near the edge has high Q values due to low temperatures and probably low water content. These may result from coupling of the slab laterally with the thick Eurasian lithosphere, which inhibits back-arc rifting, retards subduction, and reduces the water supply to the mantle wedge. The SW Ryukyu subduction system represents a subduction-zone edge type distinct from more commonly documented free or warm edges.

SUMMARY A broad-band ocean-bottom seismometer (OBS) deployed ∼180 km east of Taiwan provides a first glimpse into the upper mantle beneath the westernmost section of the Philippine Sea or the Huatung basin (HB). We measured interstation phase velocities of Rayleigh waves between the OBS and stations on the eastern coast of Taiwan. The phase velocities show smooth variations from 3.8 to 3.9 km s −1 for periods of 25–40 s. In this short period range, phase velocities are comparable to those characterizing the 15–30 Ma Parece-Vela basin of the Philippine Sea. Modelling of the finite-frequency effect proves the validity of the measurement for the average HB. The shear-wave velocity models inverted from the 25 to 40 s dispersion show a velocity at lithospheric depths about 0.1 km s −1 lower than that of the west Philippine Sea, which agrees with the age effect derived from the Pacific pure-path model. Inversions incorporating the less reliable data above 40 s yield a shear velocity

Earthquakes from the Indonesian region recorded at distances between 78 and 104° from broad-band stations of the Kaapvaal network were used to investigate the P-wavespeed structure of the lowermost 700 km of the mantle below the central part of the Indian Ocean. Analysis of arrival times indicates a maximum in slowness gradients at a distance of 88°, and wavespeeds in the lowermost 700 km of the mantle that are lower than in the reference model IASP91. These results suggest that the lowermost mantle is relatively hot. Measurements of P-wave amplitudes also show a pronounced maximum at a distance close to 88°. Phase-weighted stacking of waveforms from three deep-focus earthquakes do not show any later arrivals that might identify a triplication in travel times associated with the presence of a discontinuity at the top of D′′. However, synthetic seismograms produced by both the WKBJ and reflectivity methods for models having a small discontinuity at the top of D′′ show the presence of an amplitude maximum, due to constructive interference between the closely spaced signals near the cross-over point of a triplication whose branches are not clearly resolvable visually. The observed amplitude maximum, however, is too large to be explained by the smooth model MIDIND01 fitted to the travel times. Stacking of synthetic seismograms constructed from models with small discontinuities at the top of D′′ gives similar results to those for the real seismograms. The observed amplitudes, travel times, results of phase-weighted stacking and synthetic seismograms, together suggest the presence of a small P-wave discontinuity of 0.4–0.6% that defines the top of D′′ 210 ± 20 km above the core–mantle boundary.

The hypothesis that thermal plumes contribute to the Cenozoic magmatism in East Africa is now widely accepted. The controversy lies on how many plumes exist and where they may be located. In this study we show numerical experiments of mantle convection models for a number of thermal plume models and discuss the implications for the melt generation in East Africa. We investigate how the plume(s), the Tanzania craton, and the African lithospheric structure may interplay to result in the magmatism distribution in East Africa since the Eocene. Our results demonstrate that the variable thickness of the lithosphere modulates melt generation. A single-plume model cannot reproduce the observations consistently. Double-plume models with plumes located at Afar and Kenya regions are viable with reasonable physical properties. The distribution of the plume material, however, is sensitive to the angle at which the Tanzania craton and regions of thick lithosphere approach the plume, as the African plate moves. Models that have present-day location of the second plume (Kenya plume) under the Eastern rift or the interior of the Tanzania craton can best match the basalt distribution. Our model results suggest that the basaltic eruptions associated with the Afar plume tap a relatively deep source of the plume body in general, whereas melting occurs at shallower depths for the Kenya plume except for the Eocene episode. The magmatism is derived from a more depleted mantle source in the low-Ti basalt province of northwestern Ethiopia. Our experiments indicate the thermal influence of the Afar plume but predict an absence of plume-derived melts, suggesting the melt generation within lithosphere triggered by thermal influence of Afar plume in this region. Our model results suggest that plume plays an active role on the initiation of the rifting process in East Africa.

Negative thermal anomalies beneath a mid-ocean ridge are dynamically isolated from the ambient upwelling and diverging flow field in the asthenosphere whose viscosity is on the order of 5×10 19 Pa s or less. This study examines on what condition a near-ridge cold anomaly ascends with the upwelling passive flow and spread off-axis. Three-dimensional numerical modeling demonstrates that, for a given magnitude of the cold anomaly, the viscosity of the asthenosphere, the spreading rate and the interference from continental rifting are the dominant controlling factors to the ascent/descent of the anomaly. To overcome the weight of such an anomaly and couple it with the upwelling, either the spreading rate or the asthenospheric viscosity has to be high. In a low viscosity asthenosphere, the cold anomaly also ascends during the early stage of continental rifting due to the enhanced upwelling induced by the thick continental lithosphere. The dynamic interaction between the cold anomaly and the ambient flow renders a transient nature of the subsidence of the seafloor, which may lead to exaggerated temperature variation estimated by using a conduction model alone. The scenarios examined are employed to place a constraint on dynamic models recently proposed for the Australian–Antarctic Discordance, in which the source of the negative anomaly is hypothesized to be deeply rooted in the upper mantle. With the asthenospheric viscosity less than 10 20 Pa s, the upwelling of the cooler material from great depths, which causes a significant topographic low at the Discordance, is made possible only by rifting of the Australian continent off the Gondwanaland.

T waves excited by earthquakes propagate along the SOFAR channel with low transmission loss, and therefore can be recorded on land-based seismic stations and hydrophones located thousands of kilometers away from earthquake epicenters. Early T-wave observations are mostly based on recordings by land-based stations due to the mechanics of the energy conversion of acoustic waves into seismic phases. Recently, T-wave signals have also been detected by ocean-bottom seismometers (OBSs) at deep ocean basin offshore eastern Taiwan, raising the question of how deep ocean environment affects the generation and propagation of T-waves. In this study, to understand how acoustic energy scatters and interacts with different seafloor topography, we apply the acoustic ray theory to simulate acoustic propagation in the presence of realistic seafloor topography and sound speed profile. Our simulations indicate that seafloor topography indeed affects the acoustic propagation pattern, part of which may reach deep ocean regions. We also simulate seismic energy of T-waves by stacking energy coming from a series of potential conversion points within a specific time-window. The stacked energy distribution expresses a pattern similar to the envelope function of T-waves, indicating that the long-lasting waveform may result from a series of seismic-acoustic conversion processes.

Summary Seismic traveltime tomography is commonly discretized by a truncated expansion of the pursued model in terms of chosen basis functions. Whether parametrization affects the actual resolving power of a given data set as well as the robustness of the resulting earth model has long been seriously debated. From the perspective of the model resolution, however, there is one important aspect of the parametrization issue of seismic tomography that has yet to be systematically explored, that is, the space–frequency localization of a chosen parametrization. In fact, the two most common parametrizations tend to enforce resolution in each of their own particular domains. Namely, parametrization in terms of spherical harmonics with global support tends to emphasize spectral resolution while sacrificing the spatial resolution, whereas the compactly supported pixels tend to behave in the opposite manner. Some of the significant discrepancies among tomographic models are very likely to be manifestations of this effect, when dealing with data sets with non-uniform sampling. With an example of the tomographic inversion for the lateral shear wave heterogeneity of the D″ layer using S–SKS traveltimes, we demonstrate an alternative parametrization in terms of the multiresolution representation of the pursued model function. Unlike previous attempts of multiscale inversion that invoke pixels with variable sizes, or overlay several layers of tessellation with different grid intervals, our formulation invokes biorthogonal generalized Harr wavelets on a sphere. We show that multiresolution representation can be constructed very easily from an existing block-based discretization. A natural scale hierarchy of the pursued model structure constrained by the resolving power of the given sampling is embedded within the solution obtained. It provides a natural regularization scheme based on the actual ray-path sampling and is thus free from a priori prejudices intrinsic to most regularization schemes. Unlike solutions obtained through spherical harmonics or spherical blocks that tend to collapse structures onto ray paths, our parametrization imposes regionally varying Nyquist limits, that is, robustly resolvable local wavelength bands within the obtained solution.

Differential travel times between S (or diffracted S) and SKS were measured to study the global distribution of shear wave velocity heterogeneity in the lowermost 250 km of the mantle (the D″ region). Commencing with ∼3000 S-SKS times with variable qualities, we minimize uneven path coverage by thinning redundantly sampled regions (e.g., the Fiji-Tonga to North America corridor) and collect additional data for sparsely sampled areas, especially in the Southern Hemisphere. About 1500 paths were retained, distributed to reconcile both high-density and homogeneous sampling. We compare (1) spherical harmonic and (2) equal-area block parameterizations of D″ shear velocity heterogeneity for identical minimum resolving lengths and mean model errors. We show that the two parameterizations result in indistinguishable patterns of heterogeneity and power distribution for resolution down to a block size of 4.5°×4.5° and maximum spherical harmonic degree of 40. We demonstrate synthetically that a high-degree (L) inversion followed by a lower-degree (Ls) spherical harmonic synthesis effectively circumvents model contamination from expansion truncation effects. Tomographic inversion of this data set yields a global distribution of D″ heterogeneity robust up to degree 12. In our preferred model (L = 40, Ls = 12, rms ∼ 1%), surface hotspots and estimated lower mantle plume roots are located in or at edges of low-velocity regions (δVs< −2%), for both the Atlantic-Indian low-velocity corridor as well as the low velocities beneath the central Pacific. Although the above parameterizations entail no intrinsic difference in resolvability, higher resolution is regionally achieved by removing the derived degree 12 model from the data then inverting the corrected residuals using 4.5°×4.5° blocks. This hybrid method prevents power loss at intermediate to long wavelengths from otherwise severe damping and recovers smaller-scale structure where constraints are better than the global average, such as beneath the Pacific and Eurasia. A synthetic recovery test yields resolvability maps that reflect both the path geometry and the quality of data, which help to identify robust features in both the degree 12 and the hybrid D″ models.

The amplitudes of diffracted SH (S diff ) normalized to SKS, together with the S diff –SKS times, were analyzed to constrain the structure of the D” region beneath Asia and the northernmost Indian Ocean. While the S diff –SKS residuals (δt; relative to the Preliminary Reference Earth model, or PREM) are consistently negative from 95° to 120°, the amplitude residuals of S diff /SKS (δ A ) show two trends of distance dependence, corresponding to distinct seismic structures in two adjacent zones in D”. In zone A, δ A increases significantly with distance, suggesting the presence of a negative velocity gradient in the base of the mantle. The travel time residuals independently require that the average velocity of zone A be faster than that of PREM. One-dimensional structures that reconcile both sets of constraints were sought through systematic forwarding modeling. Models with negative gradients that satisfy δt's match δ A 's to an acceptable degree only if a high-quality factor ( Q s ) is assumed. The preferred model for zone A has a 400–500 km thick negative gradient layer, with a ~4% velocity discontinuity at the top and Q s = 1000, an about three-fold increase from the PREM value. In zone B, the amplitude–distance curve is virtually flat, and a 200–300 km thick high-velocity layer with PREM-like gradient and Q s explains both observations well. To assess the role of mantle asphericity in δ A, we estimate the strength of focusing of the S waves into the Fresnel zone at the onset of diffraction in vertical cross-sections of 3-D tomographic models SAW12D and SKS12WM13. Both models predict stronger focusing in zone A than in zone B. The focusing effect is translated to a positive base-line shift in δ A, which, if applied to the model predictions, alleviates the need for an extremely high Q s in zone A. The simple 2-D experiment suggests that velocity gradient and the anelastic attenuation of the D” layer as well as the mantle heterogeneity all probably contribute to the decay characteristics and the level of amplitude of S diff . The slab subducted in the Mesozoic may be responsible for the structure depicted in this study.

A global map of shear velocity in the D″ layer results from the inversion of 340 differential travel times of diffracted S(SH) minus SKS(SV) (Sd-SKS), from long-period records of global seismic networks. The two-phase design reduces contamination from upper mantle heterogeneities and errors in location and origin time of the events. Additional corrections are made for (1) azimuthal anisotropy at stations where shear wave splitting parameters are available and for (2) travel time perturbations due to lower mantle asphericity, although both effects are minor compared with the observed residuals with respect to the preliminary reference Earth model (PREM) [Dziewonski and Anderson, 1981]. The corrected residuals, ranging from −16 to 18 s, are attributed to anomalies in D″ sampled by both phases. Taking these residuals as data and assuming a constant, 250-km-thick D″ layer, we invert for a lateral velocity variation model of D″ using spherical harmonics. In parameterizing D″ velocities, a high degree expansion (L=14) avoids aliasing, but only the reliably determined, low degree components (LI

To examine the hypothesis that the Australian-Antarctic Discordance overlays an anomalously cold upper mantle, we have measured Rayleigh wave group velocities along paths between the Southeast Indian Ridge and the Australian station NWAO. The group velocity peaks for events in the vicinity of 117°E, rather than for those within the discordance zone between 120 and 127°E. This variation is resolved at least for periods of 35–45 s. To model the cause, assumed velocity anomalies with simple geometry were tested using the surface wave Gaussian beam method in a forward sense. The preferred model consists of an elongated structure centered 300–500 km west of the discordance and stretching northward for at least 1000 km. The position of the anomalous structure is consistent with recent global tomographic models in this region.

SUMMARY The anomalous sea-floor topography and geoid height over the AustralianAntarctic Discordance (AAD) are examined and attributed to dynamic mantle flow in this study. Because the subsidences of sea-floor depth and geoid with agc exhibit forms of isostatic response of cooling from the surface, they are removed from the data to retain the dynamic components. In removing the cooling effect we allow for different subsidence rates for different spreading corridors. The reduction yields primarily north-south-trending lows crossing the AAD in both the residual topography and the residual geoid. Assuming that the mantle behaves viscously, the residual data were inverted to thermal perturbations in the assigned depth extents. The obtained thermal perturbations exhibit mainly eait-west, or axis-parallel variation and indicate that along the 3000 km long ridge, the mantle beneath the AAD is about 150 and 50°C colder than average for perturbation layers confined within 100 and 300km from the surface, respectively. As a result, the 3-D flow driven by the thermal anomaly is also predominantly east-west oriented with a major downwelling near the AAD and its aged mantle. To independently constrain these results, we derive a simple formula to estimate the mantle temperature beneath the ridge from the subsidence rates of topography and geoid for different corridors. This simple estimation yields a variation of about *200"C with the low near AAD and high at the segment to the east. The trend of variation is consistent with that from the dynamic modelling, and the magnitude is relatively high but favours the model in which the dynamic source is concentrated within the upper asthenosphere. The thermal anomaly could be, however, eliminated by lateral advection in the along-axis direction which is not considered in the inversion. The advection effect is estimated to cause a decay of the dynamic topography at a rate of the order of 0.005 km Myr-', making the downwelling die out very slowly with age. The actual 3-D anomalous flow in the south-east fndian Ridge area is probably still characterized more by a downwelling stretched across-axis as depicted by our dynamic modelling than in the form of a focused plume sinking at the AAD.

Taiwan is located near the edge of the continental Eurasia plate and the Philippine Sea plate, a quake-prone zone. Averagely speaking, out of 1000 sensible quakes detected per year, more than 50% of the epicenters are located in the surrounding waters. Therefore, there is a great need from our seismology community to deploy ocean bottom seismometer (OBS) to have high fidelity data, and widen the aperture of the observation as well. In this paper, we report the development of a “midweight” OBS called Yardbird which has one year deployment duration. We utilized the newly available ultra-low power MCU (micro control unit) and SD card to design a compact data logger. The response of 4.5-Hz geophones was extended to 3 s and used as the seismic sensors. With the compactness of the sensor module, it enables us to design a compact dual-axis motor-driven stage to level the vertical component to be less than 0.1 degree with respect to the gravity. Currently, the total power consumption of Yardbird OBS is less than 0.2 mW for three channels. A pilot experiment and was conducted in the south western Okinawa trough in 2010. During the three months experiment, an array of five Yardbirds captured and deep quakes from the subducting oceanic plate. The results prove that the performance of Yardbird OBS suffices the basic needs of seismic research.

[1] An improved inversion technique is needed to effectively separate the frequency dependence of the source from the intrinsic attenuation of the medium. We developed a cluster-event method (CEM) in which clusters of nearby events, instead of individual events, pair with stations to form the basis for measurements ofQ value and corner frequency (fc). We assume that the raypaths from one cluster to a station share an identical Q while each event in the same cluster is allowed for only one fcin the inversion process. This approach largely reduces the degrees of freedom to achieve a robust inversion. We use an optimization algorithm of simulated annealing to solve the nonlinear inverse problem. The CEM was applied to events at 70–150 km depths in the Japan subduction zone recorded by F-net. We show that the method proposed here leads to better constraints on both source parameters and attenuation. The resultantQ's in the mantle wedge increase from lower than 300 beneath the arc and back-arc to greater than 600 in the fore-arc region. Thefc's satisfy a self-similar scaling relationship with seismic moment ofM0 ∝ fc−3 with a best fit stress drop of 21.9 ± 6.9 MPa in Madariaga's form. This contrasts to the stress drop of 1.4 ± 1.1 MPa for a global data set composed of prior measurements for crustal events. The results of this study agree with results from previous studies, except with an upward deviation due to higher corner frequencies and stress drops.

Short-period stations ENT and NSK in northern Taiwan have frequently recorded split S waveforms from earthquakes in the underlying subduction zone. To determine splitting parameters together with objective, straightforward error estimations, we employ a waveform cross-correlation method. The error for each measurement is estimated by translating the lower limit of the 95% confidence interval of the cross-correlation coefficient into azimuth and time. Only measurements with 95% confidence regions excluding zero delay time are accepted as splitting data. We also use an energy minimization method to constrain the solutions with minor splitting. We interpret the splitting as caused by stress-aligned cracks in the upper crust. The average delay times are 0.03–0.04 s for ENT and 0.07 s for NSK. The fast directions at the two stations are predominantly N-S, which roughly match the stress trajectories predicted by a model simulating the arc-continent collision in the Taiwan area. However, at ENT there is a subset of data with fast directions oriented in roughly E-W. The nearly 90° rotation of the inferred crack orientation occurs in a horizontal distance of 10–15 km in the upper crust. Both the predicted heterogeneous stress regime and the local structural setting in northeastern Taiwan favor a lateral variation with a similar length scale.

Taiwan is located near the edge of the continental Eurasia plate and the Philippine Sea plate, a quake-prone zone. Averagely speaking, out of 1000 sensible quakes detected per year, more than 50% of the epicenters are located in the surrounding waters. Therefore, there is a great need from our seismology community to deploy ocean bottom seismometer (OBS) to have high fidelity data, and widen the aperture of the observation as well. In this paper, we report the development of a “midweight” OBS called Yardbird which has one year deployment duration. We utilized the newly available ultra-low power MCU (micro control unit) and SD card to design a compact data logger. The response of 4.5-Hz geophones was extended to 3 s, and the modified geophone was used as the seismic sensors. With the compactness of the sensor module, it enables us to design a compact dual-axis motor-driven stage to level the vertical component to be less than 0.1 degree with respect to the gravity. Currently, the total power consumption of Yardbird OBS is less than 0.2 mW for three channels. A pilot experiment was conducted in the south western Okinawa trough in 2010. During the three months experiment, an array of five Yardbirds captured tele-seismic event from Chile, and deep quakes from the subducting oceanic plate. The results prove that the performance of Yardbird OBS suffices the basic needs of seismic research.

[1] Melt generation in the arc-basin system of the subduction zones depends not only on fluid distribution and detailed temperature and pressure conditions, but also on melting mechanism and melting history. These factors are closely linked to the lithospheric extension, the proximity of the back-arc basin to the arc, and the migration of the spreading center. In this study the evolution of the back-arc basin is incorporated to explore the first-order features of the dynamics and melting processes in the Mariana subduction zone, using coupled thermomechanical numerical experiments. In the models continuous stretching thins the lithosphere, ultimately resulting in breakup of the lithosphere. The spreading center develops and migrates subsequently to form the new ocean basin in the back-arc region. Potential flux melting and decompression melting regimes are delineated based on the resultant thermal models, experimentally determined solidi, and phase diagrams. The results predict an overlap between the two regimes for several million years, suggesting that both melting mechanisms contribute to arc magma and back-arc basin basalts during the rifting phases and the earlier stages of spreading. The decompression melting regime progressively separates from the flux melting regime as the ridge migrates along. The mantle flow field driven by subduction and spreading, together with the ridge migration, generate intricate particle paths. This could lead to multiphase melt extraction for both the arc magma and the back-arc basin basalts and thus cause complex geochemical signatures. These processes may explain the first-order geochemical characteristics observed in most of the Mariana arc-basin system.

SUMMARY To test a prediction of the anisotropic properties of the upper mantle beneath the central North Atlantic, we measure splitting of the waveforms of SKS phases recorded at WWSSN stations in the Azores (PDA), and on Bermuda (BEC). Because the analysis of individual, analogue records does not provide sufficient resolution, we have developed a new technique for representing the confidence limits that permits a straightforward estimation of the average splitting beneath a station from a group of events. If events from a variety of different azimuths are recorded, the average splitting parameters can be tightly constrained. We found five records for PDA and eight for BEC that have good signal-to-noise ratio and a prominent SKS phase. Only one of the phases at PDA shows evidence of splitting; the average anisotropy must be small if it exists at all. At BEC, the average anisotropy is significant with a fast direction between NW-SE and N-S and a time delay between the two split components of about 1 s. Both of these observations are significantly smaller than predicted by a model of the upper mantle anisotropy based on SS-S differential traveltimes. The absence of systematic anisotropy at PDA may be attributable to a complex flow pattern in the upper mantle at this site near the Eurasian-African plate boundary in the vicinity of an active hotspot. At BEC, the observed direction is consistent with the prediction and both the direction and the magnitude of the average anisotropy are indicative of a sublithospheric source. To check the reliability of these results from analogue records, we have performed a similar analysis for station KIP on Oahu where digital records are available for comparison. The results from digital and analogue records are similar. At KIP, the splitting is more variable than expected for uniform anisotropy, with the fast directions for individual records ranging from about N26"E to about N83"E and with variable time delays. Similar variability is observed at all three stations. The average anisotropy at KIP is well-constrained, however, with a time delay of about 0.6 s and fast direction N75"E. The fast direction coincides with the fossil sea-floor spreading direction beneath the Hawaiian islands, suggesting that the dominant source of anisotropy beneath KIP is lithospheric.

[1] Three-dimensional tomographic images of the subducting Philippine Sea slab and the fore-arc mantle beneath NE Taiwan and the westernmost Ryukyu were generated in this study. More than 5600 events recorded simultaneously by seismic networks in Taiwan and Japan were relocated for the inversion for variations in VP, VS, and VP/VS. Analysis of the trade-off between the data variance reduction and model variance helps to determine the appropriate strictness for regularization to avoid either overinterpretatioin or underinterpretation of data. The regularization parameters were also chosen to ensure suppression of artificial VP/VS anomalies. The subducting slab is characterized by high VP, high VS, and intermediate to low VP/VS. Notable in the mantle wedge is the high VP/VS anomalies that abut the surface of the subducting slab at depths of 30–80 km. The previously identified positive VP/VS channel connecting the slab and the arc volcano interpreted to be a melt pathway is not reproduced in this study. We convert VS and VP/VS to perturbations of temperature and serpentinization of the mantle. The slab is cooled by 200°C–400°C relative to the mantle, in accord with the estimates from theoretical modeling of subduction. The serpentinization reaches ∼15%, or 2% water content, at 50 km depth in the fore-arc mantle. We interpret the peak serpentinization as hydrated by the water released from the basalt-eclogite metamorphic reaction in the oceanic crust of the subducting Philippine Sea Plate. The spatial limitation of the present seismic networks in this region with respect to subduction zone events hinders a full description of the pattern of melting in much of the mantle wedge. Resolution tests of the tomographic inversion provide a basis to guide our interpretation to better resolved regions.

Taiwan is located in a junction corner between the Philippine sea plate and Eurasian plate. Because of active convergence, numerous earthquakes have taken place in and around Taiwan. On average, there are about two earthquakes greater than magnitude 6 each year and over 70% of earthquakes occurred in the offshore area. Because of the subduction of Philippine Sea Plate beneath the western end of the Ryukyu Arc and northern Taiwan, both the tectonics and seismic activity are intensive. The 2004 Sumatra earthquake has induced giant tsunami attacking coastal countries of South Asia. In a similar geodynamic context, the Sumatra event has aroused the attention of Taiwan government. Specialists from Taiwan earth scientists and ocean engineers have quickly teamed up to discuss the potential and mitigation of natural hazards from the western end of the Ryukyu subduction zone. To construct a submarine cable observatory off eastern Taiwan (MACHO project) was proposed. MACHO means a sea goddess who protects people at sea. The purpose of MACHO project has several folds. Firstly, the extension of seismic stations on land to offshore area can increase the resolution of earthquake relocating. Secondly, the extension of seismic stations may obtain tens of second before the destructing seismic waves arrive on land or tens of minute before the arrival of giant tsunami, which is helpful for earthquake or tsunami warning. Thirdly, the seafloor scientific station can monitor the active volcanoes in the Okinawa Trough, which is directly adjacent to the Ilan plain in northeastern Taiwan. Fourthly, the seafloor observatory can be used to continuously study the Kurosho current, off eastern Taiwan. The MACHO project has been granted for the fiscal year of 2007. The MACHO project is expected to be fulfilled in 2009.

[1] The termination of the Ryukyu trench against Eurasia and the oblique subduction of the Philippine Sea plate create a subduction-collision complex offshore Taiwan, which has not previously been elucidated in detail. We combine traveltime data from the seismic networks in Taiwan and Japan to better illuminate how the subducting Ryukyu slab deforms in this subduction-collision zone. More than 5000 events recorded by both networks were relocated with the double-difference method using an optimal regional one-dimensional velocity model. The offshore seismicity indicates that the double seismic zone, with a gap of 15–20 km, exists in the subducting slab in the depth range of 40–80 km. Focal mechanisms suggest that the double seismic zone is caused by east-west compression resulting from oblique convergence. The improved hypocentral locations for the first time reveal folding of the slab into a horizontal curvature larger in magnitude than and opposite in sign to that of the Ryukyu trench in the depth range 50–100 km. The anomalous curvature, together with the focal mechanisms, suggests that the slab folds against the Eurasian lithosphere and that this deformation cannot be fully elastic. We model this deformation mode as the developing instability of a viscoelastic Maxwell layer embedded in a viscous medium. The characteristic wavelength of the instability, i.e., ∼250 km, is consistent with folding of a slab whose viscosity is 100 times higher than that of the surrounding mantle for an along-strike elastic membrane strain as small as 0.01, or more than 3 orders of magnitude higher if 5% elastic strain is allowed.

[1] The shear velocity structure in the lowermost 500 km of the mantle beneath the Caribbean and surrounding areas is determined by seismic tomography applied to a suite of Sd-SKS, ScS-S, (Scd + Sbc)-S, and ScS-(Scd + Sbc) differential times, where (Scd + Sbc) is a pair of overlapping triplication arrivals produced by shear wave interaction with an abrupt velocity increase at the top of the D″ region. The inclusion of the triplication arrivals in the inversion, a first for a deep mantle tomographic model, is possible because of the widespread presence of a D″ velocity discontinuity in the region. The improved ray path sampling provided by the triplication arrivals yields improved vertical resolution of velocity heterogeneity within and above the D″ region. The reference velocity model, taken from a prior study of waveforms in the region, has a 2.9% shear velocity discontinuity 250 km above the core-mantle boundary (CMB). Effects of aspherical structure in the mantle at shallower depths than the inversion volume are suppressed by applying corrections for several different long-wavelength shear velocity tomography models. Born-Frechet kernels are used to characterize how the finite frequency data sample the structure for all of the differential arrival time combinations; inversions are performed with and without the kernels. The use of three-dimensional kernels stabilizes the tomographic inversion relative to a ray theory parameterization, and a final model with 60- and 50-km correlation lengths in the lateral and radial dimensions, respectively, is retrieved. The resolution of the model is higher than that of prior inversions, with 3–4% velocity fluctuations being resolved within what is commonly described as a circum-Pacific ring of high velocities. A broad zone of relatively high shear velocity material extends throughout the lower mantle volume beneath the Gulf of Mexico, with several percent lower shear velocities being found beneath northern South America. Concentrated low-velocity regions extend through the D″ layer under the Caribbean, Colombia, and Ecuador, suggestive of small-scale plumes in the boundary layer and possible lateral variations of the D″ discontinuity. One scenario consistent with the imaged features involves subducted Farallon plate ponding at the base of the mantle and laterally displacing hot boundary layer material that piles up and destabilizes on its margins.

[1] The temperature gradient of the thermal boundary layer in the D″ region is one of the basic properties of the mantle that remains difficult to evaluate. We characterize this quantity for a region in the western Pacific using gradient-sensitive seismological probes. First, both the amplitudes and travel times of the diffracted P and S phases (Pdiff, Sdiff) were analyzed to constrain the regional-scale vertical gradients of VP and VS. Acceptable models were grid-searched using synthetic waveforms for trial models that employ only one velocity gradient zone in D″. The VS models were searched over the parameter space of the thickness (H) and the gradient. Solutions show negative deviation from Preliminary Reference Earth Model and a trade-off between gradient and H from −0.00088 s−1, for190 km to −0.00049 s−1, for 240 km with a typical error of ±0.00015 s−1. Rather than pinning down the best solution, we consider two sets of solutions for VS with H = 190 and 240 km. For these two H values, the gradients of VP were determined solely by amplitude to be −0.00032 s−1 for 190 km and 0.00015 s−1 for 240 km, shrouded in relatively large errors of 0.0002–0.0003 s−1. We derive the temperature gradient from the resolved velocity gradients using the seismological-thermodynamic equation of Doornbos et al. (1986) and arrive at 9.0 ± 5.8 K/km and 5.7 ± 4.1 K/km over 190 and 240 km, respectively. The difference in gradient between these two models is statistically significant at the 99% confidence level, and the odds that the apparent steeper gradient is caused by squeezing energy of diffracted waves into a thinner zone can also be rejected with high confidence. The combination of the two models is thus consistent with a nonlinear temperature profile that steepens toward the core-mantle boundary. Both models detect the superadiabatic gradient of temperature in the lowermost mantle.

Broadband records from ocean-bottom seismometers deployed in the Okinawa trough and the Huatung basin were analyzed to provide seafloor noise characteristics and the detection thresholds for teleseismic body and surface waves. Ambient noise levels on the horizontal components are 10-40 dB higher than on the vertical component, with the sensor seated on the surface of the sediment. On the vertical components, infragravity waves are 10-30 dB more energetic at the shallower Okinawa trough sites (≤2000 m depth) than at the deeper Huatung basin site (~4700 m). From 0.03 to 0.2 Hz, the Huatung basin noise levels are comparable to that of the broadband stations in Taiwan on a quiet day. The microseism peaks (~0.2-0.5 Hz) of OBSs reach or exceed the high noise model of continental stations. At regional distances M(subscript w) 6.5 is required for recording prominent Rayleigh waves if the source radiation is unfavorable, but 6.2 is sufficient for a favorable focal mechanism. Several tens to over one hundred high-fidelity P, P(subscript diff) and PKP waveforms have been recorded per year by OBSs at high corner frequency of 0.1 Hz with a minimum M(subscript w) 5.3-6.0. The number of recording drops to less than 5 per year at 1 Hz with M(subscript w)≥6.4 and distances less than 90°.