Impact of regional mantle flow on subducting plate geometry and interplate stress: insights from physical modelling

Physical models of subduction investigate the impact of regional mantle flow on the structure of the subducted slab and deformation of the downgoing and overriding plates. The initial mantle flow direction beneath the overriding plate can be horizontal or vertical, depending on its location with respect to the asthenospheric flow field. Imposed mantle flow produces either over or underpressure on the lower surface of the slab depending on the initial mantle flow pattern (horizontal or vertical, respectively). Overpressure promotes shallow dip subduction while underpressure tends to steepen the slab. Horizontal mantle flow with rates of 1–10 cm yr⁻¹ provides sufficient overpressure on a dense subducting lithosphere to obtain a subduction angle of  ∼60°  , while the same lithospheric slab sinks vertically when no flow is imposed. Vertical drag force (due to downward mantle flow) exerted on a slab can result in steep subduction if the slab is neutrally buoyant but fails to produce steep subduction of buoyant oceanic lithosphere. The strain regime in the overriding plate due to the asthenospheric drag force depends largely on slab geometry. When the slab dip is steeper than the interplate zone, the drag force produces negative additional normal stress on the interplate zone and tensile horizontal stress in the overriding plate. When the slab dip is shallower than the interplate zone, an additional positive normal stress is produced on the interplate zone and the overriding plate experiences additional horizontal compressive stress. However, the impact of the mantle drag force on interplate pressure is small compared to the influence of the slab pull force since these stress variations can only be observed when the slab is dense and interplate pressure is low.     

Double seismic zones and stresses of intermediate depth earthquakes

Summary. Data from Japanese local seismograph networks suggest that the stresses in double seismic zones are in-plate compression for the upper zone and in-plate tension for the lower zone; the stresses do not necessarily appear to be down-dip. It may therefore be possible to identify other double seismic zones on the basis of data which indicate that events with differing orientations of in-plate stresses occur in a given segment of slab.
A global survey of published focal mechanisms for intermediate depth earthquakes suggests that the stress in the slab is controlled, at least in part, by the age of the slab and the rate of convergence. Old and slow slabs are under in-plate tensile stresses and the amount of in-plate compression in the slab increases with increasing convergence rate or decreasing slab age. Young and fast slabs are an exception to this trend; all such slabs are down-dip tensile. Since these slabs all subduct under continents, they may be bent by continental loading. Double seismic zones are not a feature common to all subduction zones and are only observed in slabs which are not dominated by tensile or compressive stresses.
Unbending of the lithosphere and upper mantle phase changes are unlikely to be the causes of the major features of double zones, although they may contribute to producing some of their characteristics. Sagging or thermal effects, possibly aided by asthenospheric relative motion, may produce the local deviatoric stresses that cause double zones.     

The dynamical origin of subduction zone topography

Summary. Subduction zones are expressed topographically by long linear oceanic trenches flanked by a low outer rise on the seaward side and an island arc on the landward side. This topographic structure is reflected in free air gravity anomalies, suggesting that much of the topography originates from dynamical forces applied at the base of the crust. We have successfully reproduced the general topographic features of subduction zones by supposing that the stresses generated by the bending of the viscous lower lithosphere as it subducts are transmitted through the thin elastic upper portion of the lithosphere. The trench is due to a zone of extensional flow (associated with low pressure) in the upper part of the viscous lithosphere.
The stresses in the subducting slab are computed using a finite element technique, assuming a Maxwell viscoelastic constitutive relation. Various dips (10 to 90°) were investigated, as well as depth dependent and non-Newtonian (power law, n = 3) viscosities. Observed subduction zone dimensions are well reproduced by these models. The effective viscosity required at mid-depth in the lithosphere is about 6 × 10²² P. This low value is probably due to the stress dependence of the effective viscosity. However, these models also show that the topography of the subduction zone depends primarily upon the geometry of the subducting slab (dip, radius of curvature of the bend) rather than upon its rheology. Shear stresses beneath the trench reach maxima of approximately 50 MPa. An interesting feature of some solutions is a dynamically supported bench or platform between the trench and island arc.     

The lithosphere–asthenosphere boundary beneath the western United States

S receiver functions from 67 broad-band seismic stations in the western United States clearly reveal the existence of a mantle discontinuity with velocity reduction downward, which we interpret as the lithosphere–asthenosphere boundary (LAB). The average depth of the LAB is ∼70 km. The boundary is relatively sharp with an overall sharpness of less than 20 km. The boundary is more prominent south of the Mendocino Triple Junction, where the Farallon Plate has completely subducted. This may indicate partial melts at the base of the lithosphere caused by the upwelling of the asthenospheric flow through the slab window. A double low velocity zone is observed at base of the lithosphere beneath southern Sierra Nevada, implying a second melting zone at a depth of ∼100 km, well correlated with previous studies of lithospheric delamination in the area.     

Lengths of intermediate and deep seismic zones and temperatures in downgoing slabs of lithosphere

Summary. If intermediate and deep earthquakes occur in the coldest portions of the downgoing slabs of lithosphere, then different lengths of seismic zones represent different temperatures in the slabs. As the slab descends through the aesthenosphere, it warms primarily by conduction of heat through its upper and lower surfaces. Isotherms are advected downwards to distances approximately proportional both to the rate of subduction and to the square of the thickness of the lithosphere. Consequently, the lengths of seismic zones should be approximately proportional to the product of the rates and the squares of these thicknesses. As these thicknesses are approximately proportional to the square root of the age of the lithosphere, the lengths ought therefore to be approximately proportional to the product of the convergence rates times the ages. Although there is considerable scatter, observed lengths are approximately proportional to such products, and are not simply related to the rate, the age or the thickness alone. The data crudely fit the relationship: length = rate × age/10. Using this relationship, we infer that the Philippine sea and Pacific plates move slowly, if at all, with respect to one another and that the Farallon plate may have been too young to be subducted to a great distance beneath western North America in the Palaeogene. Calculations of temperatures at the depths of the deepest events suggest that these cut-off temperatures increase from about 600 ± 100°C at 200 km to 830 ± 50°C at 650 km depth, but the cut-off potential temperature is approximately constant. Assuming that the strength is a thermally activated parameter, and using the activation energy for olivine, a crude estimate of activation volume is obtained from the dependence of the cut-off temperature on depth.     

Small-scale mantle flows and induced lithospheric stress near island arcs

Summary. This paper explores the middle ground between complex thermally-coupled viscous flow models and simple corner flow models of island arc environments. The calculation retains the density-driven nature of convection and relaxes the geometrical constraints of corner flow, yet still provides semianalytical solutions for velocity and stress. A novel aspect of the procedure is its allowance for a coupled elastic lithosphere on top of a Newtonian viscous mantle. Initially, simple box-like density drivers illustrate how vertical and horizontal forces are transmitted through the mantle and how the lithosphere responds by trench formation. The flexural strength of the lithosphere spatially broadens the surface topography and gravity anomalies relative to the functional form of the vertical flow stresses applied to the plate base. I find that drivers in the form of inclined subducting slabs cannot induce self-driven parallel flow; however, the necessary flow can be provided by supplying a basal drag of 1–5 MPa to the mantle from the oceanic lithosphere. These basal drag forces create regional lithospheric stress and they should be quantifiable through seismic observations of the neutral surface. The existence of a shallow elevated phase transition is suggested in two slab models of 300 km length where a maximum excess density of 0.2 g cm⁻³ was needed to generate an acceptable mantle flow. A North New Hebrides subduction model which satisfies flow requirements and reproduces general features of topography and gravity contains a high shear stress zone (75 MPa) around the upper slab surface to a depth of 150 km and a deviatoric tensional stress in the back arc to a depth of 70 km. The lithospheric stress state of this model suggests that slab detachment is possible through whole plate fracture.     

Geometry and brittle deformation of the subducting Nazca Plate, Central Chile and Argentina

We use data from the Chile Argentina Geophysical Experiment (CHARGE) broad-band seismic deployment to refine past observations of the geometry and deformation within the subducting slab in the South American subduction zone between 30°S and 36°S. This region contains a zone of flat slab subduction where the subducting Nazca Plate flattens at a depth of ∼100 km and extends ∼300 km eastward before continuing its descent into the mantle. We use a grid-search multiple-event earthquake relocation technique to relocate 1098 events within the subducting slab and generate contours of the Wadati-Benioff zone. These contours reflect slab geometries from previous studies of intermediate-depth seismicity in this region with some small but important deviations. Our hypocentres indicate that the shallowest portion of the flat slab is associated with the inferred location of the subducting Juan Fernández Ridge at 31°S and that the slab deepens both to the south and the north of this region. We have also determined first motion focal mechanisms for ∼180 of the slab earthquakes. The subhorizontal T -axis solutions for these events are almost entirely consistent with a slab pull interpretation, especially when compared to our newly inferred slab geometry. Deviations of T -axes from the direction of slab dip may be explained with a gap within the subducting slab below 150 km in the vicinity of the transition from flat to normal subducting geometry around 33°S.     

Gravity constraints on the dynamics of the crust–mantle system during Calabrian subduction

The thermomechanic evolution of the lithosphere–upper mantle system during Calabrian subduction is analysed using a 2-D finite element approach, in which the lithosphere is compositionally stratified into crust and mantle. Gravity and topography predictions are cross-checked with observed gravity and topography patterns of the Calabrian region. Modelling results indicate that the gravity pattern in the arc-trench region is shaped by the sinking of light material, belonging to both the overriding and subduction plates. The sinking of light crustal material, up to depths of the order of 100–150 km is the ultimate responsible for the peculiar gravity signature of subduction, characterized by a minimum of gravity anomaly located at the trench, bounded by two highs located on the overriding and subducting plates, with a variation in magnitude of the order of 200 mGal along a wavelength of 200 km, in agreement with the isostatically compensated component of gravity anomaly observed along a transect crossing the Calabrian Arc, from the Tyrrhenian to the Ionian Seas. The striking agreement between the geodetic retrieved profiles and the modelled ones in the trench region confirms the crucial role of compositional stratification of the lithosphere in the subduction process and the correctness of the kinematic hypotheses considered in our modelling, that the present-day configuration of crust–mantle system below the Calabrian arc results from trench's retreat at a rate of about 3 cm yr⁻¹, followed by gravitational sinking of the subducted slab in the last 5 Myr.     

Asthenospheric counterflow: a kinematic model

Summary. Present-day plate motions imply that about 240 km³ of oceanic lithosphere is created by sea-floor spreading and destroyed by subduction per year. A greater volume of asthenosphere will be dragged along by plate motions. Given the fluxes generated at plate boundaries, the horizontal direction and net rate of counterflow required to maintain mass balance is determined globally by a simple analytical model. Time-dependent calculations indicate that the motions are approximately valid in the hotspot reference frame over the past 5 Myr. Under most plates, the model return flow is opposite to the lithospheric motion in the hotspot frame. The counterflow dominates the resisting stresses to plate motion, so driving force models based on plate drag alone are not valid where the directions of plate motion and counterflow differ. The most marked departure of the two directions is under the North American plate. The model counterflow directions indicate that the sources of mantle hotspots are not located within the asthenosphere. Model flux balances demonstrate exchange of material between asthenospheric reservoirs located beneath different plates. Suggestions of southward asthenospheric motion under the North Atlantic, based on physical features around Iceland and strontium isotope geochemistry, are consistent with the direction of flow predicted by the model.     

Free-surface formulation of mantle convection—II. Implication for subduction-zone observables

Viscous and viscoelastic models for a subduction zone with a faulted lithosphere and internal buoyancy can self-consistently and simultaneously predict long-wavelength geoid highs over slabs, short-wavelength gravity lows over trenches, trench-forebulge morphology, and explain the high apparent strength of oceanic lithosphere in trench environments. The models use two different free-surface formulations of buoyancy-driven flows (see, for example, Part I): Lagrangian viscoelastic and pseudo-free-surface viscous formulations. The lower mantle must be stronger than the upper in order to obtain geoid highs at long wavelengths. Trenches are a simple consequence of the negative buoyancy of slabs and a large thrust fault, decoupling the overriding from underthrusting plates. The lower oceanic lithosphere must have a viscosity of less than to²⁴ Pa s in order to be consistent with the flexural wavelength of forebulges. Forebulges are dynamically maintained by viscous flow in the lower lithosphere and mantle, and give rise to apparently stiffer oceanic lithosphere at trenches. With purely viscous models using a pseudo-free-surface formulation, we find that viscous relaxation of oceanic lithosphere, in the presence of rapid trench rollback, leads to wider and shallower back-arc basins when compared to cases without viscous relaxation. Moreover, in agreement with earlier studies, the stresses necessary to generate forebulges are small (∼ 100 bars) compared to the unrealistically high stresses needed in classic thin elastic plate models.     

Anomalous high-frequency wave propagation from the Tonga–Kermadec seismic zone to New Zealand

Summary. Bulletins of the International Seismological Centre (ISC) show very large residuals, up to 15 s early, for arrivals from events in the Tonga–Kermadec subduction zone to the New Zealand network of seismometers. The very early arrivals are confined to events south of about 22°S, and shallower than about 350 km. The waveforms show two distinct phases: an early, emergent, first phase with energy in the high-frequency band 2–10 Hz, and a distinct second phase, containing lower frequency energy, arriving at about the time predicted by JB tables.
The residuals are attributed to propagation through the cold, subducted lithosphere, which has a seismic velocity 5 per cent faster, on average, than normal. Ray tracing shows that the ray paths lie very close to the slab for events south of 22°S, but pass well beneath the slab for events further north, corresponding to the change in residual pattern. This characteristic of the ray paths is due to the curved shape of the seismic zone, and in particular to the bend in the zone where the Louisville ridge intersects the trench at 25°S.
The residuals can only be explained if the high velocity anomaly extends to a depth of 450 km in the region of the gap in deep seismicity from 32 to 36°S. The very high-frequency character of the first phase requires the path from the bottom of the slab to the stations to be of high Q , and to transmit 2–10 Hz energy with little attenuation.
The absence of low-frequency energy in the first phase is due to the narrowness of the high-velocity slab, which transmits only short-wavelength waves. The second phase, which contains low frequencies, is identified as a P -wave travelling beneath the subducted slab in normal mantle. There is no need to invoke any special structures, such as low-velocity waveguides or reflectors, to explain any of the observations. The S -wave arrivals show similar effects.     

The Zihuatanejo, Mexico, earthquake of 1994 December 10 (M= 6.6): source characteristics and tectonic implications

An analysis of the Zihuatanejo, Mexico, earthquake of 1994 December 10 ( M = 6.6), based on teleseismic and near-source data, shows that it was a normal-faulting, intermediate-depth ( H = 50 ± 5 km) event. It was located about 30 km inland, within the subducted Cocos plate. The preferred fault plane has an azimuth of 130°, a dip of 79° and a rake of −86°. The rupture consisted of two subevents which were separated in time by about 2 s, with the second subevent occurring downdip of the first. The measured stress drop was relatively high, requiring a Δσ of about a kilobar to explain the high-frequency level of the near-source spectra. A rough estimate of the thickness of the seismogenic part of the oceanic lithosphere below Zihuatanejo, based on the depth and the rupture extent of this event, is 40 km.
This event and the Oaxaca earthquake of 1931 January 15 ( M = 7.8) are the two significant normal-faulting, intermediate-depth shocks whose epicentres are closest to the coast. Both of these earthquakes were preceded by several large to great shallow, low-angle thrust earthquakes, occurring updip. The observations in other subduction zones show just the opposite: normal-faulting events precede, not succeed, updip, thrust shocks. Indeed, the thrust events, soon after their occurrence, are expected to cause compression in the slab, thus inhibiting the occurrence of normal-faulting events. To explain the occurrence of the Zihuatanejo earthquake, we note that the Cocos plate, after an initial shallow-angle subduction, unbends and becomes subhorizontal. In the region of the unbending, the bottom of the slab is in horizontal extension. We speculate that the large updip seismic slip during shallow, low-angle thrust events increases the buckling of the slab, resulting in an incremental tensional stress at the bottom of the slab and causing normal-faulting earthquakes. This explanation may also hold for the 1931 Oaxaca event.     

Stress and temperature in subduction shear zones: Tonga and Mariana

Summary. Because there is secondary sea-floor spreading in the Tonga and Mariana subduction systems, the island arcs are separate plates. Horizontal forces on the two sides of the arc must balance, and the maximum force on the back-arc side can be calculated from a lithostatic ridge model. This, in combination with gravity data, allows calculation of the average shear stress in the top 100 km of the subduction shear zone. Stress in Tonga is 220±100 bar, and in the Mariana it is 165±75 bar. These low stresses are probably made possible by a fluid pore pressure almost equal to the least compressive stress.
Knowledge of stress allows approximate calculation of temperature in the shear zone by integration of a single differential equation. These temperatures are too low to activate most dehydration reactions in the subducted crust. As it approaches the volcanic line, this crust is at 150–350°C in Tonga and 150–300°C in the Mariana. Shear melting of the crust is ruled out, and conductive melting of the slab by contact with the asthenosphere meets with geochemical objections. Magmas in these systems are probably produced by partial melting of asthenosphere, triggered by a sudden release of water from the slab.     

State of stress in the Southern Tyrrhenian subduction zone from fault-plane solutions

In this paper we present revised locations and original focal mechanisms computed for intermediate and deep earthquakes that occurred within the Southern Tyrrhenian subduction zone between 1988 and 1994, in order to improve our knowledge of the state of stress for this compressional margin. In particular, we define the stress distribution within a large portion of the descending slab, between 40 and about 450 km depth. The seismicity distribution reveals a continuous 40–50 km thick slab that abruptly increases its dip from subhorizontal in the Ionian Sea to a constant 70° dip in the Tyrrhenian. We computed focal mechanisms for events with magnitudes ranging from 2.7 and 5.7, obtaining the distribution of P - and T -axes for many events for which centroid moment tensor (CMT) solutions are not available, thus enabling the sampling of a larger depth range compared to previous studies. We define three portions of the slab characterized by different distributions of P - and T -axes. A general down-dip compression is found between 165 and 370 km depth, whereas in the upper part of the slab (40–165 km depth) the fault-plane solutions are strongly heterogeneous. Below 370 km the P -axes of the few deep events located further to the north have a shallower dip and are not aligned with the 70° dipping slab, possibly suggesting that they belong to a separated piece of subducted lithosphere. There is a good correspondence between the depth range in which the P -axes plunge closer to the slab dip (∼ 70°) and the interval characterized by the highest seismic energy release (190–370 km).     

VOLCANIC ACTIVITY IN TAIWAN

I.MesozoicEraVolcanicAchvityDlepre-Tertlalyl11etal11olphicbase111entisOnlyexposedalongtheeastemflankoftheBackboneRangeoftl1eCentraIRangeinTaiwan(Figurel).Overlpngthebasen1e11t,tl1estratumofTertiarytoQuaten1alysysten11sexposed(Flgure2).Deposltlonofthepre-TertlaryrocksprobablybopninPen1llanperiodandextendedintoMesozoic.Theserockswerelatertectonizedandlnotal1lorPhosedinlateMesozoictil11earldtl1el11aJororogenic111ovelnentisnamedYenshanlllovelnent(t1ieNanaoorogenylnTaiwan).A:lthe111eta11…     

Bending stresses in subducted lithosphere

Summary. The flow of the mantle near subducted lithosphere applies a force (dynamic load) to the slab which has a normal component as well as a shear component. In addition, the greater density of the slab compared to the surrounding mantle results in a downward force (negative buoyancy) which also has a component normal to the slab. Because the slab is long compared to its thickness, the stresses in the slab due to the normal components of the loads are accurately estimated from the bending moment. We have used previous estimates of the loads to compute the bending moment in the slab as a function of distance down the slab. Our results indicate that the loads applied to the slab are large compared to its strength; bending moments due to normal loads will not be transmitted over the length of the slab. Instead, the slab will deflect so that the normal loads are balanced locally.     

The transition from subduction to continental collision: crustal structure in the North Canterbury region, New Zealand

The North Canterbury region marks the transition from Pacific plate subduction to continental collision in the South Island of New Zealand. Details of the seismicity, structure and tectonics of this region have been revealed by an 11-week microearthquake survey using 24 portable digital seismographs. Arrival time data from a well-recorded subset of microearthquakes have been combined with those from three explosions at the corners of the microearthquake network in a simultaneous inversion for both hypocentres and velocity structure. The velocity structure is consistent with the crust in North Canterbury being an extension of the converging Chatham Rise. The crust is about 27 km thick, and consists of an 11 km thick seismic upper crust and 7 km thick seismic lower crust, with the middle part of the crust being relatively aseismic. Seismic velocities are consistent with the upper and middle crust being composed of greywacke and schist respectively, while several lines of evidence suggest that the lower crust is the lower part of the old oceanic crust on which the overlying rocks were originally deposited.
The distribution of relocated earthquakes deeper than 15 km indicates that the seismic lower crust changes dip markedly near 43S. To the south-west it is subhorizontal, while to the north-east it dips north-west at about 10. Fault-plane solutions for these earthquakes also change near 43S. For events to the south, P -axes trend approximately normal to the plate boundary (reflecting continental collision), while for events to the north, T -axes are aligned down the dip of the subducted plate (reflecting slab pull). While lithospheric subduction is continuous across the transition, it is not clear whether the lower crust near 43S is flexed or torn.     

New insight into the crust and upper mantle structure under Alaska

To better understand the seismic structure of the subducting Pacific plate under Alaska, we determined the three-dimensional P-wave velocity structure to a depth of approximately 200 km beneath Alaska using 438,146 P-wave arrival times from 10,900 earthquakes. In this study an irregular grid parameterization was adopted to express the velocity structure under Alaska. The number of grid nodes increases from north to south in the study area so that the spacing between grid nodes is approximately the same in the longitude direction. Our results suggest that the subducting Pacific slab under Alaska can be divided into three different parts based on its geometry and velocity structure. The western part has features similar to those in other subduction zones. In the central part a thick low-velocity zone is imaged at the top of the subducting Pacific slab beneath north of the Kenai Peninsula, which is believed to be most likely the oceanic crust plus an overlying serpentinized zone and the coupled Yakutat terrane subducted with the Pacific slab. In the eastern part, significant high-velocity anomalies are visible to 60–90 km depth, suggesting that the Pacific slab has only subducted down to that depth.     

Implications of stress concentration on a strike-slip fault in an elastic plate subject to basal shear stress

Summary We consider a long strike-slip fault in a lithosphere modelled as an elastic slab. To the base of the slab a shear stress distribution is applied which simulates the viscous drag exerted by the asthenosphere. The resulant stress on the fault plane may directly fracture the lithosphere in its brittle upper portion; alternatively it may give rise at first to a stable aseismic sliding in the lower portion. In the latter case, stress concentration due to the deep aseismic slip is the relevant feature of the pre-seismic stress acting on the upper section of the lithosphere. The two cases are examined by use of dislocation theory and their observable effects compared. Different depths of the aseismic slip zone and the presence or absence of a uniform friction on the seismic fault are allowed for. If the model is applied to the San Andreas fault region, where a steady sliding condition actually seems to be present at shallow depth, it turns out that the slip amplitudes commonly associated with large earthquakes are consistent with average basal stress values which can be substantially lower than a few bars, a value often quoted as the steady state basal stress due to a velocity gradient in the upper asthenosphere.     

Slab low-velocity layer in the eastern Aleutian subduction zone

Local earthquakes in the vicinity of the Alaskan Peninsula's Shumagin Islands often produce arrivals between the main P and S arrivals not predicted by standard traveltime tables. Based on traveltime and polarization, these anomalous arrivals appear to be from P -to- S conversions at the surface of the subducted Pacific Plate beneath the recording stations. The P -to- S conversion occurs at the top of a low-velocity layer which extends to at least 150 km depth and is 8 ˜ 2 per cent slower than the overlying mantle. The slab is ˜ 7 per cent faster than the mantle. The low-velocity layer contains the foci of the earthquakes in the upper plane of the double seismic zone and confines PS ray paths to lie within it. These observations indicate that layered structures persist to positions well past the surface location of the volcanic front. Reactions forming high-pressure minerals do not yield slab-like velocities until beyond the point that subduction zone magma genesis occurs. If the subducted oceanic crust forms the layer, it is subducted essentially intact.