Comment on the article “Probability of radial anisotropy in the deep mantle” by Visser et al. (2008)

Informations

• Publié par: Shuel
• Nombre de lectures: 83
• Langue: English

Extrait

Earth and Planetary Science Letters 276 (2008) 223–225
Contents lists available at ScienceDirect
Earth and Planetary Science Letters
journal homepage: www.elsevier.com/locate/epsl
Discussion
Comment on the article “Probability of radial anisotropy in the deep mantle”
by Visser et al. (2008) EPSL 270:241–250
⁎Andréa Tommasi , David Mainprice
Geosciences Montpellier, CNRS & Université Montpellier 2, CC.060, Pl. Eugène Bataillon, Montpellier, F-34095, France
article info
Article history:
Accepted 17 July 2008
Available online 17 September 2008
Editor: R.D. van der Hilst
Based on the inversion of a large surface waves dataset, including Strong anisotropy at the rock scale is produced when the crystal
both fundamental and higher modes, Visser et al. (2008) constructed preferred orientations have strong single maxima concentrations of
new maps of the probability of surface wave polarization (or radial) the crystallographic directions corresponding to the highest and
anisotropy. These maps bring new constraints on the variation of lowest propagationvelocities or S-wavesplittingof thesingle crystal.
radial anisotropy with depth as a function of the geodynamic This is the case, for instance, for olivine under shallow mantle
environment. However, the interpretation of the seismic anisotropy conditions, where dislocation creep with dominant activation of the
distributionin terms of mantledeformation is oversimplified, sinceit (010)[100] slip system leads to a strong orientation of the [100] axis,
isessentiallybasedondataoftheevolutionofolivinecrystalpreferred which is the fastest P-wave propagation and the lowest S-wave
orientations and seismic anisotropy at high temperature, but low splitting direction, and of the [010] axis, which is the slowest P-wave
pressure conditions, neglectingrecentfindingsonthe deformation of propagation direction (Tommasi et al., 2000). In contrast, dominant
mantle minerals under high-pressure in-situ mantle conditions. activationof{hk0}[001]systemsinthedeepuppermantleresultsina
Three points, in particular, deserve a throughout discussion, since point concentration of the [001] axis that has intermediate values for
they highlight the limitations of using basic concepts that have been bothP-wavepropagationvelocitiesandS-wavesplittingandingirdle
established from the analysis of shallow mantle rocks and low distributionsof the[100]and[010]axesinaplanenormaltotheflow
pressure deformation experiments, as for instance: (i) lack of direction and hence in averaging between high and low values
anisotropy implies deformation by diffusion creep or (ii) fast (Mainpriceetal.,2005).Similarly,forwadsleyite,seismicvelocityand
polarization (or propagation for P or SV waves) directions mark anisotropydistributionsofapolycrystallineaggregatedonotcorrelate
systematically theflow direction,tointerpret seismicanisotropy data inasimplemannerwiththesingle-crystalproperties(Tommasietal.,
below depths of 200–250 km in the mantle. 2004). In the single crystal, P-wave propagation is fastest parallel to
the [010] axis and slowest parallel to [001]. The maximum S-wave
1. Absence of anisotropy does not necessarily imply polarization anisotropy is observed for propagation parallel to b110N,
diffusion creep the fast S-wave being polarized parallel to the [010] axis. In
wadsleyite-rich polycrystals, the slowest P-wave propagation is still
Estimation of seismic anisotropy using polycrystal plasticity parallel to the [001] maximum, but the fastest propagation direction
models based on recent experimental data on the deformation of and the polarization of the fast S-wave are not parallel to the [010]
olivine, wadsleyite, and perovskite highlights that development of a maximum,butatlowangletothe[100]maximum,i.e.,atlowangleto
crystal preferred orientation does not necessarily result in significant the shear direction. The dispersion of [010] axes in the polycrystal
seismic anisotropy, even if the single crystal has strongly anisotropic results in a weak seismic anisotropy for wadsleyite-rich rocks (N2%),
elastic constants (Tommasi et al., 2004; Mainprice et al., 2005, 2008). evenif thesingle-crystalisclearlyanisotropic(∼13%,Mainpriceetal.,
2000) and the rock has a well-developed wadsleyite CPO. Finally, in
the lower mantle, coupling of ab initio and polycrystal plasticity
models suggests that even if deformation by dislocation creep
produces clear perovskite CPO, the resulting seismic anisotropy is
DOI of original article: 10.1016/j.epsl.2008.03.041.
weak (Mainprice et al., 2008). The maximum polarization anisotropy
⁎ Corresponding author.
E-mail address: [e-mail] (A. Tommasi). is N2% at the low temperatures and pressures that prevail just below
0012-821X/$ – see front matter © 2008 Elsevier B.V. All rights reserved.
doi:10.1016/j.epsl.2008.07.061224 A. Tommasi, D. Mainprice / Earth and Planetary Science Letters 276 (2008) 223–225
the transition zone and it decreases with increasing temperature and velocity. In the light of these models, the higher probability of V-
pressure. NV in the transition zone below convergent or divergent plateSV SH
boundaries observed by Visser et al. (2008), Beghein and Trampert
2. Variations in sign of radial anisotropy (from V NV to (2004), and Panning and Romanowicz (2006) may imply dominantSH SV
V NV ) with increasing depth cannot be univocally interpreted verticalflowbeneath plate boundaries, whereas the global azimuthalSV SH
as a change in flow direction anisotropy patterns inferred from Love-wave overtone data by
Trampert and van Heijst (2002) are better explained by horizontal
This interpretation is based on the premise, derived from shearing,whichmaypredominateinmostofthetransitionzone,away
experiments at high temperature and moderate pressure and from plate boundaries. Similar models using lower transition zone,
extensive data on naturally deformed shallow upper mantle rocks, ringwoodite-rich compositions show that anisotropy below 520 km
that olivine deforms by dominant slip on {0kl}[100] slip systems. depth cannot be explained by strain-induced CPO of ringwoodite
Under these conditions, the deformation-induced CPO results in the (Carrez et al., 2006). Thus, if, as suggested by Visser et al. (2008)
fastest S-waves being polarized parallel to the flow direction. It also probability maps, seismic anisotropy is higher in the lower transition
leads to faster propagation of P and Rayleigh waves parallel to the zone, another mechanism, such as alternating layers with highly
flow direction. However, if the relation between finite strain axes contrasting seismic properties has be invoked explain the observa-
and olivine CPO changes, this relation is no longer valid. Recent tions. Such a layering may, for instance, be associated with stagnant
experimental data show that at the high pressures that prevail in the hydrated slabs in the transition zone as proposed by Ritsema et al.
deep upper mantle (Couvy et al., 2004; Raterron et al., 2007)orat (2004) and Mainprice et al. (2007).
high water contents (Jung and Karato, 2001) olivine deforms by
dominant glide on {hk0}[001] systems. There is still debate on the References
relative importance of these two parameters, but they are not
Beghein, C., Trampert, J., 2004. Probability density functions for radial anisotropy:independent, sincewatersolubilityin olivine increaseswith pressure
implicationsfortheupper1200kmofthemantle.EarthPlanet.Sci.Lett.217(1–2),
(Bolfan-Casanova et al., 2000). Moreover, both pressure and water 151–162.
havethe same effect, resulting in the sameolivine CPO evolution and Bolfan-Casanova, N., Keppler, H., Rubie, D.C., 2000. Water partitioning between
nominally anhydrous minerals in the MgO–SiO2–H2O system up to 24 GPa:hence in fast polarization directions normal to the flow direction.
implications for the distribution of water in the Earth's mantle. Earth Planet. Sci.
Polycrystal plasticity models based on these data show indeed that
Lett.182, 209–221.
the change in intensity and sign (V bV ) of anisotropy withSH SV Carrez, P., Cordier, P., Mainprice, D., Tommasi, A., 2006. Slip systems and plastic shear
anisotropy in Mg2SiO4 ringwoodite: insights from numerical modelling. Eur. J.increasing depth in the upper mantle observed in previous studies
Mineral.18,149–160.(Montagner and Kennett, 1996; Panning and Romanowicz, 2006;
Couvy, H., Frost, D., Heidelbach, F., Nyilas, K., Ungár, T., Mackwell, S., Cordier, P., 2004.
Zhou et al., 2006) and confirmed by Visser et al. (2008) is perfectly Shear deformation experiments of forsterite at 11 GPa – 1400 °C in the multianvil
apparatus. Eur. J. Mineral.16 (6), 877–889.explained as resulting from the change in olivine deformation with
Jung, H., Karato, S.-I., 2001. Water-induced fabric transitions on olivine. Science 293,
depth from dominant [100] slip to dominant [001] slip (Mainprice
1460–1463.
et al., 2005). Karato, S.-I., Dupas-Bruzek, C., Rubie, D.C., 1998. Plastic deformation of silicate spinel
under the transition zone conditions of the Earth's mantle. Nature 395, 266–269.
Li,L.,Long,H.,Raterron,P.,Weidner,D.,2006.Plasticflowofpyropeatmantlepressure3. Anisotropy in the transition zone cannot be interpreted based
and temperature. Am. Miner. 91, 517–525.
on olivine deformation data Mainprice, D., Barruol, G., Ben Ismaïl, W., 2000. The seismic anisotropy of the Earth's
mantle: from single crystal topolycrystal. In: Karato, S.-I., Forte, A.M., Liebermann,
R.C., Mast