U41F-01
Does the Sharpness of the Gutenberg Discontinuity Require Melt?
A low-velocity zone (LVZ) underlying the fast seismic lid has been recognized since Gutenberg (1959). While,
strictly speaking, the asthenosphere and the LVZ are defined by different properties, their upper boundaries
are functionally equivalent in the oceanic setting. The Gutenberg (G) discontinuity marks the seismically
sharp upper boundary of the lid-low-velocity transition beneath oceans and is characterized by an abrupt
shear wave impedance decrease of roughly 9%. Explanations for the low seismic velocities include a
contrast in volatile content, structurally bound water, and the effects of temperature and grain size. Partial
melt, which once enjoyed the status of an obvious cause, has fallen from favor, primarily because of the
difficulty of retaining a connecting fluid phase in the LVZ. But recent results demonstrate large velocity
decrements for very small grain boundary melt fractions that could remain unconnected and trapped in the
LVZ (Faul and Jackson, 2007). In order to provide better constraints on the structure of the low-velocity zone,
we have examined the G discontinuity beneath a large portion of the central and western Pacific. Our focus
is on the sharpness of the G reflector and the impedance decrease across it. Using multiple ScS
reverberations we sampled the G discontinuity along a series of oceanic paths. We obtain impedance
decreases between 4.7 and 14.2%, averaging 9.1%, assuming a first-order discontinuity. Our data do not,
however, require the change in impedance to be sharp and can easily tolerate an extended transition of 10
to 15 km with little change in the estimated impedance contrast. However, extending the transition further
requires a greater net impedance decrease across the interval. This provides an important constraint on the
thickness of the transition interval. Preliminary results that consider acceptable net impedance contrasts
(obtained from surface wave and turning wave constraints on LVZ severity) impose a conservative upper limit
on the transition interval of 30 km. Furthermore, almost all of the velocity drop from the lid to the LVZ
happens across this transition, limiting the contribution of an extended, but slow, velocity reduction in the
upper half of the LVZ. The sharpness of this transition in older oceanic lithosphere is difficult to recreate
using grain size and thermal effects only. An abrupt change in hydration or the upward termination of partial
melt seem the most likely solutions. Growing evidence of a subdued, but similar transition near 80-km depth
under many continental regions (Thybo, 2006; Rychert and Shearer, 2007) may require a separate
explanation.
Faul, U. H., and I. Jackson (2007), Diffusion creep of dry, melt-free olivine, J. Geophys. Res., 112, B04204,
doi:10.1029/2006JB004586.
Rychert, C.A., and P. M. Shearer (2007), A Global Lithosphere-Asthenosphere Boundary?, Eos Trans. AGU,
88(52), Fall Meet. Suppl., Abstract V33D-02.
Thybo, H. (2006), The heterogeneous upper mantle low velocity zone, Tectonophysics, 416, 53-79.
U41F-02 INVITED
The Lithosphere-Asthenosphere Boundary: Clues From Joint Interpretation of Global Seismic Attenuation and Velocity Models
The zone of low viscosity beneath oceanic plates appears to also be characterized by anomalous seismic properties (low velocity, high attenuation, strong radial anisotropy) and high electrical conductivity. In recent years, the oceanic seismic low-velocity zone (LVZ) has been attributed by various authors to solid-state thermal mechanisms, to partial melt, and to elevated water content. Considering seismic attenuation (1/Q) in addition to velocity may help in determining the origin of the LVZ, since attenuation and velocity have different sensitivities to temperature, composition, partial melt, and water. We have developed a new global model of shear attenuation in the upper mantle, QRFSI12. The model is derived from large data sets of Rayleigh wave amplitudes in the period range 50--250 s, and we account for source, instrument, and focusing effects on the data in order to isolate the signal of attenuation. QRFSI12 is strongly anti-correlated with global velocity models throughout the upper mantle, and individual tectonic regions are each characterized by a distinct range of attenuation and velocity values at 100-km depth. Comparison with laboratory measurements of the temperature sensitivity of attenuation and velocity (Faul and Jackson, 2005) shows good first-order agreement between the seismological and experimental values, suggesting that temperature variations can explain much, but not all, of the observed global variability in velocity and attenuation. Specifically, the seismological velocity-attenuation relationship for oceanic regions agrees with the experimental values at depths > 100 km, but the seismic properties of cratonic regions deviate from the experimental values for depths < 250 km, suggesting that temperature alone is not a sufficient explanation for the observations in the shallow cratonic mantle. Globally, seismic properties shift into better agreement with the experimental (thermal) trends in the depth range 100--150 km and 200--250 km for oceans and cratons, respectively, which may indicate the base of a chemical boundary layer and the depth of the lithosphere-asthenosphere boundary in these regions.
U41F-03 INVITED
Evidence for a heterogeneous astenosphere from intra-transform and seamount lavas
The asthenosphere is a mechanically weak region in the shallow mantle (between 100 to 300 km) underneath the lithosphere. Its unique physical properties (location, depth, viscosity, seismic velocity, anisotropy, attenuation, electrical conductivity) have been attributed to either mineral properties at relevant temperatures and pressures or to the presence of melt and/or water. To understand the processes controlling the physical properties of the asthenosphere we rely on geochemical studies of primitive basalts from the Mid-Ocean Ridges (MORB). In this regard, establishing the composition (especially volatile content) of the mantle source of MORB is a fundamental step in our understanding of this mechanically weak region of the upper mantle. However, first it is important to determine to what extent the geochemical variations in axial MORB do represent a homogeneous mantle composition and variations in the physical conditions of magma generation and transport (i.e., depth and extent of melting and melt migration); or alternatively, they are inherited from mixing processes during the aggregation of melts originated from an heterogeneous mantle beneath the mid-ocean ridge. The composition of melts within a ridge segment can be obscured by along-axis transport of magma within the crust in dikes or long-lived magma chambers. To address these issues, seamount and intra-transform lavas provide a better opportunity to deconstruct the source heterogeneity beneath mid-ocean ridges than axial lavas. Although they share a common mantle source with axial MORB, they represent smaller melt volumes tapped locally from areas lacking steady-state magma chambers and along-axis transport. Therefore, lavas from intra-transform faults and seamounts represent pre-aggregated melts experiencing relatively less mixing and differentiation, and their compositions provide insight into the heterogeneity of the asthenosphere Basalts from Quebrada/Discovery/Gofar (QDG) fracture zone system and northern seamount on the East Pacific Rise (EPR) were analyzed for volatiles, major and trace elements. Compositionally they range from ultra-depleted to fairly enriched compositions, where the level of enrichment correlates well with both ratios of volatiles to similarly incompatible refractory elements and indicators of the presence of residual garnet in the source during melting (e.g., Sm/Yb). Overall, the chemical variation of these basalts is greater than that previously found in other fracture zones (such as Siqueiros and Garrett FZ) and is similar to the compositional range defined by northern EPR seamounts. The new data, combined with previously published data of intra-transform and seamount lavas from the EPR, suggest the presence of a heterogeneous mantle forming the asthenosphere: one depleted in incompatible trace element and dry, and other enriched and wet. This heterogeneity should be considered when interpreting the physical properties of the asthenosphere and when constructing geodynamic models.
U41F-04 INVITED
Characteristics of Asthenospheric Flow Constrained from Models of Mantle Circulation and Observations of Seismic Anisotropy
Although convection in the Earth's mantle ultimately drives the tectonic plates, the upper mantle does not necessarily move together with the surface plates. Instead, the relative motion between the plates and the convecting mantle interior is largely accommodated by shear deformation within the asthenosphere. This shear flow tends to align the a- (seismically fast) axis of olivine in the direction of shear, producing an anisotropic fabric that can be detected by shear-wave splitting and surface wave observations. Thus, observations of anisotropy provide a strong constraint on the pattern of relative motion between the surface plates and the underlying mantle, and further on the asthenospheric rheology that controls the shear deformation. To utilize this constraint, we develop global mantle flow models by forming linear combinations of three components: (1) the relative motions of the surface plates over a viscous mantle (plate-driven flow), (2) tomographically-inferred density heterogeneity in the viscous mantle beneath a rigid lithosphere (density- driven flow), and (3) the net westward rotation of the entire asthenosphere in a Pacific hotspot (HS3) reference frame. We constrain the relative importance of each component by predicting asthenospheric anisotropic fabric from the combined flow models and comparing the predicted fabrics to anisotropy inferred from both global surface wave (Debayle et al. [2005]) and SKS splitting observations. We find that oceanic anisotropy is well fit if the upper mantle viscosity is between 0.5 and 2 × 1021 Pa s, which allows density-driven flow to affect asthenospheric shear in the Atlantic and Indian basins and to a lesser extent in the Pacific basin where anisotropy is dominated by more rapid plate motions. These observations place an upper bound on the total asthenospheric shear associated with net lithosphere rotation. If the asthenosphere is 10x less viscous than the upper mantle, then net rotation of up to ~2.4 cm/yr (50% of HS3) is permitted. Smaller net rotation is allowed for lower asthenosphere viscosities, which focus more of the total net rotation shear into the asthenospheric layer relative to the upper mantle.
U41F-05
On the Resolution of Radial Viscosity Structure in Modeling Long-wavelength Post-glacial Rebound Data
Studies of post-glacial rebound (PGR) constrain the viscosity of the mantle in an absolute sense, and have also been used to constrain the radial variation in mantle viscosity. Such radial variations probably have important effects on many aspects of solid-earth dynamics, including the existence and style of plate tectonics, rotational dynamics, thermal convection, and mixing of chemical species by convection. Most studies of PGR, as well as other studies constraining viscosity structure (e.g., the geoid), agree that there is some increase in viscosity with depth through the upper mantle. However, there is marked disagreement regarding the amount of the viscosity contrast and the thickness of this "low viscosity zone" (LVZ), with estimates for the viscosity contrast ranging from a mere factor of 2 to several orders of magnitude. We have examined the fundamental modeling tradeoff between viscosity contrast (η*) and layer thickness (D), and show that if only long-wavelength constraints are available, PGR models are largely indistinguishable for constant values of η*/D3 characterizing an LVZ. Models employing both residual sea level (RSL) data from Hudson Bay and time-variations in gravity from the GRACE satellite illustrate this principle. Models with LVZ viscosity contrasts ranging from less than one to more than three orders of magnitude fit the data equally well, depending on the thickness of the LVZ. Other traditional data sets constraining PGR (e.g., Fennoscandian strand lines) appear no more capable of independently resolving this ambiguity, although some studies that incorporate both long-wavelength (RSL) type constraints and shorter wavelength constraints, such as tilting of continental margins, appear to favor a relatively thin, high viscosity contrast LVZ. New GPS data, especially from North America, may be able to resolve this question even further. However, at the present time PGR studies do not preclude, and perhaps favor, a relatively thin LVZ of viscosity contrast at least one and perhaps as much as three orders of magnitude beneath the lithosphere in some regions, especially the oceanic lithosphere. This conclusion has significant implications for understanding the phenomenon of plate tectonics and the evolution and dynamics of the mantle.
U41F-06
Poiseuille-Couette flow in Low-Viscosity Asthenosphere
Strong viscosity contrasts effect first order characteristics of mantle convection. One major viscosity contrast is associated with the asthenosphere, a region of low viscosity bounded above by a strong thermal compositional boundary layer. Its lower bound is currently not well constrained and likely involves pressure effects of olivine. It has been shown that a low-viscosity channel such as the asthenosphere can lead to large wavelength convection because dissipation is lowered as flow becomes channelized. We will present mixed heated 3D mantle convection simulations with a low viscosity asthenosphere that show two long wavelength convection regimes with different scalings in terms of surface velocity and surface heat flux. We will show that mantle flow in the lithosphere-asthenosphere region is a Poiseuille-Couette flow, suggesting that velocity amplitudes in the asthenosphere can exceed surface velocities. These simulations predict preferentially localized fabric development and seismic anisotropy, and they suggest that a strong pressure driven flow can exist in the asthenosphere independent of mantle plumes. The regime crossover depends on the relative thicknesses of compositional and thermal boundary layers. Additional simulations with temperature- and yield-stress dependent viscosity show consistent behavior and further suggest that the regime crossover is also associated with a change between regimes with relative large versus small energy dissipation at plate margins.
U41F-07
Implications of grain-size evolution on the seismic structure of the oceanic upper mantle
We construct a 1-D steady-state channel flow model for grain size evolution in the oceanic upper mantle using a composite diffusion-dislocation creep rheology. Following Austin & Evans [Geology, 2007], grain size evolution is calculated assuming that grain size is controlled by a competition between dynamic recrystallization and grain growth. Applying this grain size evolution model to the oceanic upper mantle we calculate grain size as a function of depth, seafloor age, and mantle water content. The resulting grain size structure is used to predict shear wave velocity (Vs) and seismic quality factor (Q). For a plate age of 60 Myr and a mantle water content of 1000 ppm, we find that grain size reaches a minimum of ~10 mm at ~130 km depth and then increases to ~20-30 mm at a depth of 400 km. This grain size structure produces a good fit to the low seismic shear wave velocity zone (LVZ) in oceanic upper mantle observed by surface wave studies assuming that the influence of hydrogen on anelastic behavior is similar to that observed for steady state creep. Further it predicts a viscosity of ~1019 Pa·s at 150 km depth and dislocation creep to be the dominant deformation mechanism throughout the oceanic upper mantle, consistent with geophysical observations. In contrast, for a dry upper mantle (50 ppm H2O) the predicted variation in grain size (~5 to ~10 mm from 150 to 400 km depth) is not sufficient to explain the LVZ in the absence of water and/or melt. These results indicate that a combination of grain size evolution and a hydrated upper mantle is a plausible explanation for the LVZ.
U41F-08 INVITED
Possible Explanations for a Direction-dependent Conductance of the Asthenosphere
Two techniques for detecting an electrically conductive asthenosphere are presented. In the low-resolution technique, global geomagnetic induction data utilizing the magnetic daily variation are compared with a laboratory-based conductivity-depth profile. Whilst good agreement for the transition zone is achieved, the geomagnetic result for the conductivity of upper mantle above 400km, 0.01 – 0.02 S/m, does not fit the laboratory result for dry olivine, 0.003 S/m. The fit is improved if the homogenous upper mantle is replaced by a less conductive one, interupted by a conductive layer at the base of the lithosphere. Long-period magnetotellurics (MT) utilzing periods above 20.000 s at which the electromagnetic field penetrates through that layer provides its conductance and its depth. With this technique and in particular the MT strike analysis, mantle conductivity under the continental lithosphere has been found to be anisotropic in several –though not all – target areas in Europe and Australia. It has been suggested that the electrical anisotropy is a consequence of the anisotropy of hydrogene diffusity in olivine cristals which are aligned due to lattice-preferred orientation. But, because the degree of alignment is not 100%, the diffusion anisotropy factor does not readily translate to an conductivty anisotropy. Therefore, the high conductivity anisotropy factors found in particular for central Europe can not be explained with hydrogene diffusity alone. Random resistor networks (RRNs), consisting of two different resistor types, have been employed for modelling electrically anisotropic media. Because the resistivities of the two resistor types have to differ by a factor of 100, RRNs are a realistic approach for crustal structures, where these resistor types represent the rock matrix and a high conductive phase. But in order to explain the anisotropy of the sub-lithospheric mantle with this concept, a conductive phase with a conductivity 100 times higher than the one of dry olivine is required. Is there a conduction mechanism at grain boundaries that provides a conductivity of 0.3S/m? . RRNs create the strongest anisotropy if the interface between the conductive and the resistive structure has a fractal geometry, and a possibly fractal structure is the topography of the lithosphere-asthenoshere boundary under continents. In contrast, the oceanic lithosphere is much simpler. A fractal resistive/ conductive interface which exists only under continents explains the absence of strong anisotropy in seafloor MT results.