Observations of seismic anisotropy are usually used as a proxy for lattice-preferred orientation (LPO) of anisotropic minerals in the Earth's mantle. In this way, seismic anisotropy observed in tomographic models provides important constraints on the geometry of mantle deformation associated with thermal convection and plate tectonics. However, in addition to LPO, small-scale heterogeneities that cannot be resolved by long-period seismic waves may also produce anisotropy. The observed (i.e. apparent) anisotropy is then a combination of an intrinsic and an extrinsic component. Assuming the Earth's mantle exhibits petrological inhomogeneities at all scales, tomographic models built from long-period seismic waves may thus display extrinsic anisotropy. In this paper, we investigate the relation between the amplitude of seismic heterogeneities and the level of induced S-wave radial anisotropy as seen by long-period seismic waves. We generate some simple 1-D and 2-D isotropic models that exhibit a power spectrum of heterogeneities as what is expected for the Earth's mantle, that is, varying as 1/k, with k the wavenumber of these heterogeneities. The 1-D toy models correspond to simple layered media. In the 2-D case, our models depict marble-cake patterns in which an anomaly in shear wave velocity has been advected within convective cells. The long-wavelength equivalents of these models are computed using upscaling relations that link properties of a rapidly varying elastic medium to properties of the effective, that is, apparent, medium as seen by long-period waves. The resulting homogenized media exhibit extrinsic anisotropy and represent what would be observed in tomography. In the 1-D case, we analytically show that the level of anisotropy increases with the square of the amplitude of heterogeneities. This relation is numerically verified for both 1-D and 2-D media. In addition, we predict that 10 per cent of chemical heterogeneities in 2-D marble-cake models can induce more than 3.9 per cent of extrinsic radial S-wave anisotropy. We thus predict that a non-negligible part of the observed anisotropy in tomographic models may be the result of unmapped small-scale heterogeneities in the mantle, mainly in the form of fine layering, and that caution should be taken when interpreting observed anisotropy in terms of LPO and mantle deformation. This effect may be particularly strong in the lithosphere where chemical heterogeneities are assumed to be the strongest.
<p>Massive surface wave datasets constrain upper mantle seismic heterogeneities with horizontal wavelengths larger than 1000 km, allowing us to investigate the large-scale properties and alignment of olivine crystals in the lithosphere and asthenosphere. The azimuthal anisotropy projected onto the direction of present plate motion shows a very specific relation with the plate velocity. Plate-scale present-day deformation is remarkably well and uniformly recorded beneath plates moving faster than &#8764;4 cm/yr. Recent geodynamic models suggest that cold sinking instabilities tilted in the direction opposite to plate motion below fast plates could produce a pattern of large-scale azimuthal anisotropy consistent with our observations. Beneath slower plates, plate-motion aligned anisotropy is only observed locally, which suggests that the lithospheric motion does not control mantle flow below these plates.</p><p>Radial anisotropy extends deeper beneath continents than beneath oceans, but we find no such difference for azimuthal anisotropy, suggesting that beneath most continents, the alignment of olivine crystal is preferentially horizontal and azimuthally random at large scale. As most continents are located on slow moving plates, this supports the idea that azimuthal anisotropy aligns at large scale with the present plate motion only for plates moving faster than &#8764;4 cm/yr.</p><p>The same inversion also provides 3D models of seismic velocity and attenuation. The simultaneous interpretation of global 3D shear attenuation and velocity models has a great potential to decipher the effect of temperature, melt and composition on seismic observables. We will discuss our findings from the simultaneous interpretation of our latest models.</p>
Cette these de geodynamique comporte deux parties. Dans chacune de ces parties, des processus de deformation des materiaux manteliques sont mis en jeu. Les donnees gravimetriques, geoide ou anomalies de Bouguer apparaissent comme des observations fondamentales pour contraindre ou guider nos modeles. La premiere partie de cette these presente une analyse spectrale de la topographie et des anomalies de gravite de la province du Basin and Range dans l'Ouest Americain. Des periodicites de differentes longueurs d'onde sont mises en evidence suivant des directions que nous relions a celles des extensions que cette region a subies au Miocene et a l'Actuel. Nous comparons ces observations a des modeles d'instabilite d'etirement. Nous proposons l'existence d'un boudinage a l'echelle lithospherique pour interpreter nos observations dans cette region. La deuxieme partie analyse les relations a grandes longueurs d'onde entre le geoide, les heterogeneites manteliques de densite et la tectonique des plaques. Nous montrons que la prise en compte de la dynamique interne de notre Terre fluide est absolument necessaire pour comprendre ces relations. Nos modeles considerent l'existence de variations laterales des proprietes mecaniques de la Terre. Ces modeles permettent de simuler la presence des plaques tectoniques rigides a la surface d'un manteau convectif homogene.
The effect of lateral variations in the sub-lithospheric viscosity, i.e. in the lithosphere-asthenosphere coupling, is incorporated in dynamic earth models driven by plate velocities. If the coupling is stronger under fast-moving continents, it is shown that a geoid low will develop in their wake and a high in front of them. Thus the well-known low in the Indian ocean could be explained, at least in part, in terms of induced upper-mantle dynamics. Similarly, if the thickness of oceanic plates increases with age the models show that trenches should be associated with marked geoid highs, whereas ridges could correspond to much weaker geoid lows. This also seems to agree with some features of the observed geoid at very long wavelengths. The mathematical framework of such dynamic earth models is developed extensively here. These models are characterized by lateral viscosity variations inside an outer shell and a purely radial viscosity structure at greater depth. Their internal flow patterns are only driven by imposed surface velocities, not by internal loads.
