Stéphane Rondenay

Seismic imaging of subduction zone metamorphism

Combined analysis of high-resolution seismic images of the Alaska and Cascadia subduction zones reveals where metamorphic fl uids are released. Both images show the subducted oceanic crust as a dipping low-velocity layer with a clear termination depth. However, in Alaska the crust is thicker (15-20 km compared to 8 km) and terminates at greater depth (120 km com- pared to 40 km) than in Cascadia. Based on metamorphic reaction estimates and geodynamic models, we demonstrate that the termination depth corresponds to eclogitization of the crust triggered by dehydration of water-bearing minerals, and that the location of this reaction is dependent on the thermal structure of the subducted slab.

A sharp lithosphere–asthenosphere boundary imaged beneath eastern North America

Plate tectonic theory hinges on the concept of a relatively rigid lithosphere moving over a weaker asthenosphere, yet the nature of the lithosphere–asthenosphere boundary remains poorly understood. The gradient in seismic velocity that occurs at this boundary is central to constraining the physical and chemical properties that create differences in mechanical strength between the two layers. For example, if the lithosphere is simply a thermal boundary layer that is more rigid owing to colder temperatures, mantle flow models indicate that the velocity gradient at its base would occur over tens of kilometres. In contrast, if the asthenosphere is weak owing to volatile enrichment or the presence of partial melt, the lithosphere–asthenosphere boundary could occur over a much smaller depth range. Here we use converted seismic phases in eastern North America to image a very sharp seismic velocity gradient at the base of the lithosphere—a 3–11 per cent drop in shear-wave velocity over a depth range of 11 km or less at 90–110 km depth. Such a strong, sharp boundary cannot be reconciled with a purely thermal gradient, but could be explained by an asthenosphere that contains a few per cent partial melt or that is enriched in volatiles relative to the lithosphere.

P‐to‐S and S‐to‐P imaging of a sharp lithosphere‐asthenosphere boundary beneath eastern North America

[1] S-to-P (Sp) scattered energy independently confirms the existence of a seismic velocity discontinuity at the lithosphere-asthenosphere boundary that was previously imaged using P-to-S (Ps) scattered energy in eastern North America. Exploration of the different sensitivities of Ps and Sp scattered energy suggests that the phases contain independent yet complementary high-resolution information regarding velocity contrasts. Combined inversions of Ps and Sp energy have the potential to tightly constrain associated velocity gradients. In eastern North America, inversions of Sp and Ps data require a strong, 5–10% velocity contrast that is also sharp, occurring over less than 11 km at 87–105 km depth. Thermal gradients alone are insufficient to create such a sharp boundary, and therefore another mechanism is required. A boundary in composition, hydration, or a change in anisotropic signature could easily produce a sufficiently localized velocity gradient. Taken separately, the magnitudes of the effects of these mechanisms are too small to match our observed velocity gradients. However, our observations may be explained by a boundary in hydration coupled with a boundary in depletion and/or anisotropy. Alternatively, a small amount of melt in the asthenosphere could explain the velocity gradient. The tight constraints on velocity gradients achieved by combined modeling of Ps and Sp energy offer promise for defining the character of the lithosphere-asthenosphere boundary globally.

Imaging the source region of Cascadia tremor and intermediate-depth earthquakes

The subduction of hydrated oceanic crust releases volatiles that weaken the plate boundary interface, trigger earthquakes, and regulate transient phenomena such as episodic tremor and slip (ETS). It is not clear how dehydration can separately induce earthquakes within the subducting plate and ETS, partly because few data exist on their relationship to subduction zone structures. We present results of a seismic experiment in the Washington Cascades, United States, that images a region producing both earthquake types. Migration of scattered teleseis-mic waves provides images of low-velocity subducting crust at depths <40–45 km with sharp boundaries above and below it. The sharp upper boundary indicates a layer of weak sediment or an overpressured fault zone that terminates abruptly downdip at 40–45 km depth. Regular earthquakes are at the top of the mantle within the downgoing plate everywhere the plate is <95 km deep, but ETS only exists where the sharp upper boundary occurs. The ETS location supports models of slow slip that require near-lithostatic fluid pressure, whereas regular earthquakes nucleate closer to the origin of metamorphic dehydration. Very low shear stresses on the plate boundary may limit seismicity to ETS and similar phenomena.

Lithospheric assembly and modification of the SE Canadian Shield: Abitibi-Grenville teleseismic experiment

This paper presents the results of a joint Lithoprobe-Incorporated Research Institutions for Seismology (IRIS)/Program for Array Seismic Studies of the Continental Lithosphere (PASSCAL) teleseismic experiment that investigates portions of the Grenville and Superior Provinces of the Canadian Shield along the Quebec-Ontario border. Data from a 600-km-long, N-S array of 28 broadband seismographs deployed between May and October 1996 have been supplemented with additional recordings from an earlier 1994 deployment and from stations of the Canadian National Seismograph Network and the Southern Ontario Seismic Network. Relative delay times of P and S waves from 123 and 40 teleseismic events, respectively, have been inverted for velocity perturbations in the upper mantle and reveal a low-velocity, NW-SE striking corridor that crosses the southern portion of the line at latitude 46°N and lies between 50 and 300 km depth. Multievent S K S-splitting results yield an average delay time of 0.57±0.22 s and a direction of fast polarization of N93°E±18°, which is consistent with an earlier interpretation as being due to fossil strain fields related to the last major regional tectonic event. Subtle variations in splitting parameters over the low-velocity corridor may suggest an associated disruption in mantle fabric. Profiling of radial receiver functions reveals large and abrupt variations in Moho topography, specifically, a gradual thickening in crust from 40 to 45 km between latitudes 45°N and 46°N, which is followed by an abrupt thinning to 35 km at 46.6°N, some 65 km southeast of the Grenville Front. This structure is interpreted as a subduction suture extending the full length of the Front and punctuating a major pre-Grenvillian (Archean-Proterozoic) episode of lithospheric assembly in the southeast Canadian Shield. The low-velocity mantle corridor, by contrast, is better explained as the extension of the Monteregian-White Mountain-New England seamount hotspot track below the craton and is here postulated to represent interaction of the Great Meteor plume with zones of weakness within the craton developed during earlier rifting episodes.

Seismic investigation of the transition from continental to oceanic subduction along the western Hellenic Subduction Zone

National Science Foundation (U.S.) (project MEDUSA, funded by the NSF Continental Dynamics Program, grant EAR-0409373)

Multichannel Inversion of Scattered Teleseismic Body Waves: Practical Considerations and Applicability

We investigate the resolving power and applicability of a recently developed technique for multichannel inversion of scattered teleseismic body waves recorded at dense seismic arrays. The problem is posed for forward- and back-scattered wavefields generated at discontinuities in a 2D isotropic medium, with the backprojection operator cast as a generalized Radon transform (GRT). The approach allows for the treatment of incident plane waves from arbitrary backazimuths, and recovers estimates of material property perturbations about a smoothly varying reference model. An investigation of the main factors affecting resolution indicates that: (1) comprehensive source/station coverage is necessary to optimize geometrical resolution and recover accurate material property perturbations; (2) the range in dip resolution diminishes with increasing depth and is inversely proportional to array width (e.g., reaches [-45°,45°] at depths equivalent to ∼1/2 array width); (3) distortion of the image due to spatial aliasing is only significant at depths ≤2 x [station spacing]; and (4) unaccounted for departures from model assumptions (i.e., isotropy and 2D geometry) result in defocusing and mismapping of structure. Two applications to field data are presented. The first considers data from the Abitibi 1996 broadband array, in which stations were deployed at ∼20 km intervals. Imaging results show that this level of spatial sampling, which is characteristic of modem broadband arrays, is sufficient to adequately resolve structure below mid-crustal depths. For these data, we introduce a new preprocessing algorithm that uses eigenimage decomposition of seismic sections to suppress wavefield contamination by PcP and PP phases. The second application involves short period data from the Los Angeles Region Seismic Experiment and shows that images obtained from high frequency records are subject to significant contamination by scattered surface waves.

Pathway from subducting slab to surface for melt and fluids beneath Mount Rainier

Convergent margin volcanism originates with partial melting, primarily of the upper mantle, into which the subducting slab descends. Melting of this material can occur in one of two ways. The flow induced in the mantle by the slab can result in upwelling and melting through adiabatic decompression. Alternatively, fluids released from the descending slab through dehydration reactions can migrate into the hot mantle wedge, inducing melting by lowering the solidus temperature. The two mechanisms are not mutually exclusive. In either case, the buoyant melts make their way towards the surface to reside in the crust or to be extruded as lava. Here we use magnetotelluric data collected across the central state of Washington, USA, to image the complete pathway for the fluid–melt phase. By incorporating constraints from a collocated seismic study into the magnetotelluric inversion process, we obtain superior constraints on the fluids and melt in a subduction setting. Specifically, we are able to identify and connect fluid release at or near the top of the slab, migration of fluids into the overlying mantle wedge, melting in the wedge, and transport of the melt/fluid phase to a reservoir in the crust beneath Mt Rainier.

Geophysical Detection of Relict Metasomatism from an Archean (~3.5 Ga) Subduction Zone

Building Early Continents Cratons, the roots of Earths continents, have survived billions of years of accretion, volcanism, and plate motion. Due to this tumultuous history, existing evidence for how and when they formed is hard to find. C.-W. Chen et al. (p. 1089) use geophysical data collected below the Slave craton in Canada to show that subduction of lithospheric plates in the Archean may have been a major process that controlled its assembly. Spatially aligned seismic and conductive discontinuities over 100 kilometers below the surface are caused by minerals that formed from hot fluids generated as ancient crust melted at a subduction zone. Other old cratons on Earth show similar features, suggesting plate tectonics was operating at least 3.5 billion years ago. Seismic profiles of the Slave craton in Canada suggest that subduction is responsible for its formation. When plate tectonics started on Earth has been uncertain, and its role in the assembly of early continents is not well understood. By synthesizing coincident seismic and electrical profiles, we show that subduction processes formed the Archean Slave craton in Canada. The spatial overlap between a seismic discontinuity and a conductive anomaly at ~100 kilometers depth, in conjunction with the occurrence of mantle xenoliths rich in secondary minerals representative of a metasomatic front, supports cratonic assembly by subduction and accretion of lithospheric fragments. Although evidence of cratonic assembly is rarely preserved, these results suggest that plate tectonics was operating as early as Paleoarchean times, ~3.5 billion years ago (Ga).

Alaska Megathrust 2: Imaging the megathrust zone and Yakutat/Pacific plate interface in the Alaska subduction zone

We image the slab underneath a 450 km long transect of the Alaska subduction zone to investigate (1) the geometry and velocity structure of the downgoing plate and their relationship to slab seismicity and (2) the interplate coupled zone where the great 1964 earthquake (Mw 9.2) exhibited the largest amount of rupture. The joint teleseismic migration of two array data sets based on receiver functions (RFs) reveals a prominent, shallow-dipping low-velocity layer at ~25–30 km depth in southern Alaska. Modeling of RF amplitudes suggests the existence of a thin layer (Vs of ~2.1–2.6 km/s) that is ~20–40% slower than underlying oceanic crustal velocities, and is sandwiched between the subducted slab and the overriding plate. The observed megathrust layer (with Vp/Vs of 1.9–2.3) may be due to a thick sediment input from the trench in combination with elevated pore fluid pressure in the channel. Our image also includes an unusually thick low-velocity crust subducting with a ~20° dip down to 130 km depth at ~200 km inland beneath central Alaska. The unusual nature of this subducted segment results from the subduction of the Yakutat terrane crust. Our imaged western edge of the Yakutat terrane aligns with the western end of a geodetically locked patch with high slip deficit, and coincides with the boundary of aftershock events from the 1964 earthquake. It appears that this sharp change in the nature of the downgoing plate could control the slip distribution of great earthquakes on this plate interface.