Anisotropy Of Surface Waves Research Articles

Surface Wave Isotropic and Azimuthally Anisotropic Dispersion Across Alaska and the Alaska‐Aleutian Subduction Zone

AbstractComprehensive observations of surface wave anisotropy across Alaska and the Aleutian subduction zone would help to improve understanding of its tectonics, mantle dynamics, and earthquake risk. We produce such observations, using stations from the USArray Transportable Array, regional networks across Alaska, and the Alaska Amphibious Community Seismic Experiment in the Alaska‐Aleutian subduction zone both onshore and offshore. Our data include Rayleigh and Love wave phase dispersion from earthquakes (28–85 s) and ambient noise two‐ and three‐station interferometry (8–50 s). Compared with using two‐station interferometry alone, three‐station interferometry significantly improves the signal‐to‐noise ratio and approximately doubles the number of measurements retained. Average differences between both isotropic and anisotropic tomographic maps constructed from different methods lie within their uncertainties, which is justification for combining the measurements. The composite tomographic maps include Rayleigh wave isotropy and azimuthal anisotropy from 8 to 85 s both onshore and offshore, and onshore Love wave isotropy from 8 to 80 s. In the Alaska‐Aleutian subduction zone, Rayleigh wave fast directions vary from trench parallel to perpendicular and back to parallel with increasing periods, apparently reflecting the effect of the subducted Pacific Plate. The tomographic maps provide a basis for inferring the 3‐D anisotropic crustal and uppermost mantle structure across Alaska and the Aleutian subduction zone.

Effect of uneven noise source and/or station distribution on estimating the azimuth anisotropy of surface waves

Effect of uneven noise source and/or station distribution on estimating the azimuth anisotropy of surface waves

Subduction-controlled mantle flow and seismic anisotropy in South America

Seismic anisotropy records both the past and present deformation inside the solid Earth. In the mantle, seismic anisotropy is mainly attributed to the lattice preferred orientation (LPO) of mineral fabrics, caused by the shear deformation due to mantle flow. However, contributions from different tectonic processes remain debated, and a single geodynamic model that simultaneously explains the observed mantle structures and various seismic anisotropy measurements is still lacking. Here, we present a model for the Cenozoic subduction history in South America using a geodynamic simulation constrained by both past plate reconstructions and present mantle seismic structures. With a recently developed software package DRexS, we further predict azimuthal seismic anisotropy at different depths and generate synthetic shear wave splitting (SWS) measurements using the resulting mantle flow. Our results provide a good match to both depth-dependent surface wave anisotropy and various land-based SWS records. We find that the dominant control on seismic anisotropy in South America comes from subduction-induced mantle flow, where anisotropy below the subducting Nazca Plate aligns with plate-motion-induced Couette flow and that below the overriding South American Plate follows slab-induced Poiseuille flow. This large-scale mantle flow can be diverted by secondary slabs, such as that below the Antilles subduction zone. In contrast, the contribution to SWS from fossil continental anisotropy and from the effects due to mantle flow modulation by lithosphere thickness variation are minor. Upper-mantle fast seismic anomalies beneath the southern Atlantic margin should have close-to-neutral buoyancy in order to satisfy the observed seismic anisotropy.

Rayleigh-wave dispersion reveals crust-mantle decoupling beneath eastern Tibet.

The Tibetan Plateau results from the collision of the Indian and Eurasian Plates during the Cenozoic, which produced at least 2,000 km of convergence. Its tectonics is dominated by an eastward extrusion of crustal material that has been explained by models implying either a mechanical decoupling between the crust and the lithosphere, or lithospheric deformation. Discriminating between these end-member models requires constraints on crustal and lithospheric mantle deformations. Distribution of seismic anisotropy may be inferred from the mapping of azimuthal anisotropy of surface waves. Here, we use data from the CNSN to map Rayleigh-wave azimuthal anisotropy in the crust and lithospheric mantle beneath eastern Tibet. Beneath Tibet, the anisotropic patterns at periods sampling the crust support an eastward flow up to 100°E in longitude, and a southward bend between 100°E and 104°E. At longer periods, sampling the lithospheric mantle, the anisotropic structures are consistent with the absolute plate motion. By contrast, in the Sino-Korean and Yangtze cratons, the direction of fast propagation remains unchanged throughout the period range sampling the crust and lithospheric mantle. These observations suggest that the crust and lithospheric mantle are mechanically decoupled beneath eastern Tibet, and coupled beneath the Sino-Korean and Yangtze cratons.

Large-scale coherent anisotropy of upper mantle beneath the Italian peninsula comparing quasi-Love waves and SKS splitting

We document quantitatively observations of quasi-Love waves obtained at permanent (Italian National Seismic Network) and temporary seismic stations deployed in Italy between 2003 and 2006 (Retreat, CAT/SCAN projects). We analyzed large earthquakes with source parameters that favor quasi-Love wave generation within this time-span, including the Sumatra–Andaman earthquake of 12/26/04. The presence or the absence of the quasi-Love phase is compared to the smoothed anisotropic pattern defined by the numerous SKS splitting measurements obtained in peninsular Italy, and to the Italian upper mantle structure as defined by seismic tomography. The large-scale anisotropic features, responsible for shear-wave splitting and documented also by Pn and surface-wave anisotropy, generally display the correct geometry to explain the scattered quasi-Love waves. Quasi-Love observations do not demand a tilted-axis anisotropic geometry. We argue instead for anisotropy with laterally-variable horizontal symmetry axis in the upper mantle below the Italian peninsula.

Rayleigh and Love wave anisotropy in Southern California using seismic noise

We study surface wave anisotropy using three‐component frequency‐wave number analysis of 1 year (2012) of ambient seismic noise measured by the Southern California Seismic Network. Significant 2θ and 4θ Rayleigh wave anisotropy is observed over most of the frequency range 15 to 100 mHz (Millihertz). The wide Rayleigh wave illumination and large data volume allow for relatively high precision and sensitivity: the estimation variability above 35 mHz as well as the magnitude of the weakest significant detections is about 0.1%. The estimates are consistent with previous anisotropy studies of the region. We also show preliminary results for Love waves, but ambient Love wave illumination in Southern California may not be sufficient for the approach.

Time‐lapse analysis of ambient surface wave anisotropy: A three‐component array study above an underground gas storage

We perform a time‐lapse analysis of Rayleigh and Love wave anisotropy above an underground gas storage facility in the Paris Basin. The data were acquired with a three‐component seismic array deployed during several days in April and November 2010. Phase velocity and back azimuth of Rayleigh and Love waves are measured in the frequency range 0.2–1.1 Hz using a three‐component beamforming algorithm. In both snapshots, higher‐surface wave modes start dominating the signal above 0.4 Hz with a concurrent increase in back azimuth ranges. We fit anisotropy parameters to the array detections above 0.4 Hz using a bootstrap approach which also provides estimation uncertainty and enables significance testing. The isotropic phase velocity dispersion for Love and Rayleigh waves match for both snapshots. We also observe a stable fast direction of NNW‐SSE for Love and Rayleigh waves which is aligned with the preferred orientation of known shallow (<300 m) and deeper (∼1000 m) fault systems in the area, as well as the maximum horizontal stress orientation. At lower frequencies corresponding to deeper parts of the basin, the anisotropic parameters exhibit higher magnitude in the November data. This may perhaps be caused by the higher pore pressure changes in the gas reservoir in that depth range.

Azimuthal anisotropy beneath southern Africa from very broad-band surface-wave dispersion measurements

Seismic anisotropy within the lithosphere of cratons preserves an important record of their ancient assembly. In southern Africa, anisotropy across the Archean Kaapvaal Craton and Limpopo Belt has been detected previously by observations of SKS-wave splitting. Because SKS-splitting measurements lack vertical resolution, however, the depth distribution of anisotropy has remained uncertain. End-member interpretations invoked the dominance of either anisotropy in the lithosphere (due to the fabric formed by deformation in Archean or Palaeoproterozoic orogenies) or that in the asthenosphere (due to the fabric formed by the recent plate motion), each with significant geodynamic implications. To determine the distribution of anisotropy with depth, we measured phase velocities of seismic surface waves between stations of the Southern African Seismic Experiment. We applied two complementary measurement approaches, very broad-band cross-correlation and multimode waveform inversion. Robust, Rayleigh- and Love-wave dispersion curves were derived for four different subregions of the Archean southern Africa in a period range from 5 s to 250–400 s (Rayleigh) and 5 s to 100–250 s (Love), depending on the region. Rayleigh-wave anisotropy was determined in each region at periods from 5 s to 150–200 s, sampling from the upper crust down to the asthenosphere. The jackknife method was used to estimate uncertainties, and the F-test to verify the statistical significance of anisotropy. We detected strong anisotropy with a N–S fast-propagation azimuth in the upper crust of the Limpopo Belt. We attribute it to aligned cracks, formed by the regional, E–W extensional stress associated with the southward propagation of the East African Rift. Our results show that it is possible to estimate regional stress from short-period, surface wave anisotropy, measured in this study using broad-band array recordings of teleseismic surface waves. Rayleigh-wave anisotropy at 70–120 s periods shows that the fabric within the deep mantle lithosphere of the Limpopo Belt and northern Kaapvaal Craton is aligned parallel to the Archean–Palaeoproterozoic sutures at block boundaries. This confirms that the fabric within the lithosphere created by pervasive ancient deformation is preserved to this day. Suture-parallel fabric is absent, however, in the deep lithosphere of the western Kaapvaal Craton, suggesting that it was not reworked in the collision with the craton’s core, either due to its mechanical strength or because the deformation mechanism was different from those that operated in the north. Anisotropy at periods greater than 120–130 s shows fast directions parallel to the plate motion and indicates shear wave anisotropy in the asthenosphere. The depth distribution of anisotropy revealed by surface wave measurements comprises elements of both end-member models proposed previously: anisotropy in the asthenosphere shows fast-propagation directions parallel to the plate motion; anisotropy in the Limpopo and northern Kaapvaal lithosphere shows fast directions parallel to the Archean–Palaeoproterozoic sutures. The distribution of SKS-splitting orientations across southern Africa reflects anisotropic fabric both within the lithosphere (dominating the splitting beneath the Limpopo Belt and northern Kaapvaal Craton) and within the asthenosphere (dominating beneath the western Kaapvaal Craton).

Theoretical and numerical investigations of global and regional seismic wave propagation in weakly anisotropic earth models

Smith and Dahlen demonstrated that in a weakly anisotropic earth model the relative surface wave phase‐speed perturbation δc/c may be written in the form , where An and Bn are frequency‐dependent depth integrals and ζ denotes the ray azimuth. In this approximation, surface wave anisotropy is governed by 13 elastic parameters and the azimuthal dependence of the phase speed is represented by an even Fourier series in ζ involving degrees zero (five elastic parameters), two (six elastic parameters), and four (two elastic parameters). Jech and Ps̆enc̆ík demonstrated that in such a weakly anisotropic earth model the relative compressional‐wave phase‐speed perturbation may be expressed as δc/c= (2c2)−1B33, whereas the relative shear wave phase‐speed perturbations are given by δc/c= (4c2)−1{B11+B22±[(B11−B22)2+ 4B212]1/2}. We demonstrate that the coefficients B33, B11, B22, and B12 may be written in the generic form , where ζ denotes the local azimuth and i the local angle of incidence. For B11, B22 and B33 the coefficients an(i) and bn(i) are an even Fourier series of degrees zero, two and four in i, but for B12 they are an odd Fourier series of degrees one and three. Like surface waves, the azimuthal (ζ) dependence of body waves involves even degrees zero (five elastic parameters), two (six elastic parameters), and four (two elastic parameters), but, unlike surface waves, it also involves the odd degrees one (six elastic parameters) and three (two elastic parameters). Thus, weakly anisotropic body‐wave propagation involves all 21 independent elastic parameters. We use spectral‐element simulations of global and regional seismic wave propagation to assess the validity of these asymptotic body and surface wave results. The numerical simulations and asymptotic predictions are in good agreement for anisotropy at the 5 per cent level.

Spatiospectral localization of isostatic coherence anisotropy in Australia and its relation to seismic anisotropy: Implications for lithospheric deformation

We investigate the two‐dimensional (2‐D) nature of the coherence between Bouguer gravity anomalies and topography on the Australian continent. The coherence function or isostatic response is commonly assumed to be isotropic. However, the fossilized strain field recorded by gravity anomalies and their relation to topography is manifest in a degree of isostatic compensation or coherence which does depend on direction. We have developed a method that enables a robust and unbiased estimation of spatially, directionally, and wavelength‐dependent coherence functions between two 2‐D fields in a computationally efficient way. Our new multispectrogram method uses orthonormalized Hermite functions as data tapers, which are optimal for spectral localization of nonstationary, spatially dependent processes, and do not require solving an eigenvalue problem. We discuss the properties and advantages of this method with respect to other techniques. We identify regions on the continent marked by preferential directions of isostatic compensation in two wavelength regimes. With few exceptions, the short‐wavelength coherence anisotropy is nearly perpendicular to the major trends of the suture zones between stable continental domains, supporting the geological observation that such zones are mechanically weak. Mechanical anisotropy reflects lithospheric strain accumulation, and its presence must be related to the deformational processes affecting the lithosphere integrated over time. Three‐dimensional models of seismic anisotropy obtained from surface wave inversions provide an independent estimate of the lithospheric fossil strain field, and simple models have been proposed to relate seismic anisotropy to continental deformation. We compare our measurements of mechanical anisotropy with our own model of the azimuthally anisotropic seismic wave speed structure of the Australian lithosphere. The correlation of isostatic anisotropy with directions of fast wave propagation gleaned from the azimuthal anisotropy of surface waves decays with depth. This may support claims that above ∼200 km, internally coherent deformation of the entire lithosphere is responsible for the anisotropy present in surface wave speeds or split shear waves.

Mapping the lithosphere–asthenosphere boundary through changes in surface-wave anisotropy

The alignment of olivine crystals is considered as the dominant source of seismic anisotropy in the subcrustal lithosphere and asthenosphere. Different components of large-scale anisotropy can be traced in depth distributions of the radial and azimuthal anisotropy of surface waves. We propose a global model of the lithosphere–asthenosphere boundary (LAB) as a transition between a ‘frozen-in’ anisotropy in the lithosphere to anisotropy in the sublithospheric mantle related to the present-day flow. Due to different orientations of velocity maxima in the anisotropic subcrustal lithosphere and the asthenosphere, the velocity contrast related to the LAB can increase in particular directions. Because of their long wavelengths and horizontal propagation, surface waves suffer from poor lateral resolution. However, surface waves with various wavelengths allow us to map gross features of the LAB with a good vertical resolution. We estimate depths to the LAB to be between 200 and 250 km for the Precambrian shields and platforms, around 100 km for the Phanerozoic continental regions and 40–70 km beneath oceans from the world-wide depth distribution of the radial and azimuthal anisotropy of surface waves.

How to relate body wave and surface wave anisotropy?

Seismic anisotropy is one of the most efficient geological and geodynamical tools for understanding the dynamics of the Earth. Upper mantle anisotropy is evident in seismic data sets for the last 30 years primarily from surface wave dispersion curves and body wave SKS data. We demonstrate in this paper that surface wave and body wave derived anisotropy can be explained by the same anisotropic parameters (L, Gc, Gs) in the simplest case of a horizontal fast symmetry axis. One application of this method is to display the SKS delay time and the corresponding azimuth, which can be derived from surface wave global (with a lateral resolution of around 2000 km) and regional (lateral resolution of 350 km) anisotropy tomography. A global‐scale comparison is disappointing since there are only a few areas where both data types are correctly retrieved at the same spatial scale (same lateral resolution). The anisotropy is well resolved in oceanic basins from surface waves, whereas most measurements of SKS splitting are available below continents. However, there is a good agreement between surface wave and body wave anisotropy in regions where large‐scale processes (primarily tectonic) are taking place, such as the western United States and in central Asia. In particular, we show that, at high frequency (> 1Hz), the observed SKS delay time and azimuth depend on the order in which S waves propagate through the layers. A comparison between synthetic and observed SKS phases in central Asia highlights the varying sensitivity of SKS waves at depth. We can obtain a good correlation between synthetic and observed SKS by considering anisotropy only between 80 and 200 km depth in the tomographic model, rather than in the entire upper mantle. However, it is more difficult to explain the discrepancy between synthetic and observed SKS under the eastern United States, where a large‐scale coherent SKS anisotropy is present. Such results call for more sophisticated SKS waveform modeling and for enhanced lateral resolution in anisotropic tomographic models.

Where Can Seismic Anisotropy Be Detected in the Earth’s Mantle? In Boundary Layers...

—During the last 30 years, considerable evidence of seismic anisotropy has accumulated demonstrating that it is present at all scales, but not in all depth ranges. We detail which conditions are necessary to detect large-scale seismic anisotropy. Firstly, minerals must display a strong anisotropy at the microscopic scale, and/or the medium must be finely layered. Secondly, the relative orientations of symmetry axes in the different crystals must not counteract in destroying the intrinsic anisotropy of each mineral, and there must be efficient mechanisms of orientation of minerals and aggregates. Finally, the strain field must be coherent at large scale in order to preserve long wavelength anisotropy. Part of shallow anisotropy can be related to the past strain field (frozen-in anisotropy), however the deep anisotropy is due to the present strain field. All these conditions are fulfilled only in boundary layers of convective mantle.¶We review in this paper, the seismic data sets which provide insight into the location at depth of large-scale anisotropy from the D"-layer up to the lithosphere. In addition to the well-documented seismic anisotropy in the lithosphere and asthenosphere, there is new evidence of seismic anisotropy in the upper (400–660 km) and lower (660–900 km) transition zones and in the D"-layer. Nonetheless the bulk of the lower mantle seems close to isotropy. If we assume the hypothesis that seismic anisotropy is associated with boundary layers in convective systems, these observations strongly suggest that the transition zone is a boundary layer which makes the pasage of matter between the upper and the lower mantle difficult. However, this general statement does not rule out flow circulation between the upper and lower mantles. Finally, the geophysical, mineral physics and geological applications are briefly reviewed. An intercomparison between surface wave anisotropy and body-wave anisotropy data sets is presented. We discuss the scientific potential of seismic anisotropy and how it makes it possible to gain more insight into continental root, deformation and geodynamics processes.

Global anisotropy in the upper mantle inferred from the regionalization of phase velocities

A data set of 2600 paths for Rayleigh waves and 2170 paths for Love waves was selected in order to resolve azimuthal anisotropy of surface waves in the period range 70–250 s. The epicenters were chosen so that the paths provide uniform spatial and azimuthal coverage of Earth. A synthetic seismogram is computed for each path and compared to the observed seismogram in order to infer the phase velocity along each path. To process such a large data set, a new tomographic technique is designed. A good correlation is found with surface tectonics in the whole period range for Love waves and at shorter periods for Rayleigh waves. The directions of maximum velocities are also well correlated with plate tectonics motions for the 2ψ azimuthal term of Rayleigh waves. When the power spectrum of heterogeneities is computed, it is found that for Love waves the even orders are usually larger than the odd orders. For Rayleigh waves the degrees 2 and 6 dominate only at long periods, suggesting a deep origin for these components.

Azimuthally anisotropic model of the oceanic upper mantle

A mixed form of forward and backward modelings of azimuthally anisotropic oceanic upper mantle is attempted for a data set of surface-wave velocities of periods 30–100 s. One of the basic configurations of the anisotropy postulated in the LID and LVZ is axial symmetry of elastic properties with a horizontal axis parallel to the spreading direction as found in ultramafic rocks from ophiolite complexes of the Bay of Islands and Oman. Basic features of the resulting model are a thin lithosphere of ∼45 km, a decrease of S-wave velocities, and an increase of P-wave velocities and of partial melting/dislocation relaxation with depth in the LVZ. This single model explains the overall features of anisotropy observations in the Pacific ocean: P n velocity anisotropy, SH-SV anisotropy and azimuthal anisotropy of surface waves and ScS polarization anisotropy. This fact leads us to a conclusion that the ultramafic rocks from the ophiolite complexes can explain the seismic properties of the oceanic LID and LVZ.

A simple method for inverting the azimuthal anisotropy of surface waves

We investigate the problem of retrieving anisotropy as a function of depth in the mantle, from the observed azimuthal variations of Love and Rayleigh wave velocities. Following the approach of Smith and Dahlen, this azimuthal dependence is expressed in terms of a Fourier series of the azimuth θ. For the most general case of anisotropy (provided it is small enough), some simple linear combinations of the elastic tensor coefficients are shown to describe the total effect of anisotropy (both polarization anisotropy and azimuthal anisotropy) on the propagation of surface Waves. For the terms that do not depend on the azimuth the combinations are related to the elastic coefficients of a transversely isotropic mantle. For the azimuthal terms the relevant combinations are explicited. It is found that the partial derivatives of the azimuthal terms with respect to these combinations are easy to compute for they are proportional to the partial derivatives of a transversely isotropic model in the case of a plane‐layered model. In a first approximation the same property holds true for a spherical earth and we calculate from PREM all the partial derivatives needed for performing the inversion of the azimuthal anisotropy of surface waves in the period range 50–300 s. It is observed that very shallow anisotropy can be responsible for substantial azimuthal variations up to the longest periods. With this approach it is also easy to compute the azimuthal variations of surface wave velocities produced by any anisotropic model. When a Ci j elastic tensor is chosen for the upper mantle, azimuthal variations up to 2% are obtained for Rayleigh waves. The azimuthal variations of Love wave velocities are very small. The 2 θ term of the azimuthal variations of Rayleigh wave velocities is the dominant term. Its fast axis corresponds to the fast axis of P waves.

Azimuthal anisotropy of surface waves and the possible type of the seismic anisotropy due to preferred orientation of olivine in the uppermost mantle beneath the Pacific Ocean.

The group velocities of surface waves in the period range 22-89 s were measured to retrieve the anisotropy pattern of the surface wave velocities in the Eastern Pacific Ocean. WWSSN long period records at KIP, Hawaii, from ridge events in the Gulf of California, the East Pacific rise, and the Easter Island Cordillera were used to cover propagation directions N60°E-N140°E. These station and epicenters are in very favorable positions to minimize the effects of the lateral heterogeneity owing to ocean floor ages. The fast direction of the Rayleigh wave group velocity is normal to ridge, showing 3-5% azimuthal anisotropy. The Love wave group velocity shows little azimuthal anisotropy. A model of a possible type of preferred orientation of constituent olivine beneath the Pacific Ocean is proposed, based on the elastic constants and the petrofabrics of dunite from Hidaka, Hokkaido, Japan, and that of olivine nodules in the ophiolite complex from the Bay of Islands and Oman. In this model, a-axis is concentrated along a particular direction and b- and c-axes form a girdle. Elastic constants of the model are C11=2.44, C22=C33=2.05, C12=C13=0.72, C23=0.67, C44=0.69, and C55= C86=0.75 (1012 dyn·cm), where X1-axis, or the direction of the a-axis concentration, is normal to ridge and X3-axis is vertical. If we assume that the type of the preferred orientation is an age-independent property of suboceanic lithosphere, this model can explain the overall features of the large-scale anisotropy in the Pacific Ocean: the azimuthal anisotropy of surface wave velocities as described above; the SH-SV anisotropy of surface waves of about 0.15 km/s; the Pn, anisotropy of 7.9 km/s to 8.6 km/s; the high Sm, velocity of about 4.8 km/s; and the ScS polarization anisotropy