DI52A-01 INVITED
Mantle flow beneath subducting slabs and implications for mantle dynamics
The character of the mantle flow field in subduction zone regions and the dynamic interaction between downgoing slabs and the ambient mantle remain poorly understood, despite their importance for our understanding of subduction processes and of the geometry and dynamics of mantle convection. In particular, little attention has been paid to the pattern of mantle flow beneath subducting slabs. In order to identify subduction processes that make first-order contributions to the global pattern of sub-slab mantle flow in subduction regions, we have compiled shear wave splitting measurements from subduction zones worldwide from previously published studies. We have estimated average splitting parameters for the sub- wedge region and compared them to subduction parameters. We tested for relationships between sub-slab splitting parameters and indicators of tectonic processes such as trench migration velocity, convergence velocity and obliquity, age and spreading history of subducting lithosphere, slab dip, curvature, seismicity, thickness, and morphology, arc length, and subduction history. We used several different plate motion models to describe plate and trench motions and evaluate the differences among the models. Our comparisons support a model in which the mantle beneath subducting slabs is dominated by three- dimensional flow induced by trench migration. We explore several implications of our model for various aspects of mantle dynamics, including the choice of a suitable reference frame(s) for large-scale mantle flow, the nature of a likely decoupling zone between downgoing slabs and the sub-slab mantle, and constraints on mass transfer between the upper and lower mantle. We find, for example, that a relationship between sub- slab anisotropy and trench migration velocity is most easily discerned using a Pacific hotspot reference frame. Our model also suggests that downgoing slabs entrain little, if any, of the surrounding ambient mantle when (or if) they penetrate the transition zone and enter the lower mantle; this may place an upper limit on material flux between the upper and lower mantle reservoirs.
DI52A-02
Continuous subduction of oceanic crust into the deep mantle beneath central America
Recent tomographic images imply that subducted slabs may penetrate into the lower mantle in some regions. However the behavior of the subducted materials around and below the mantle transition zone remains poorly understood. In order to investigate the fate of the subducted slab beneath central America, we have analyzed broadband teleseismic data from intermediate- and deep-earthquakes in south America recorded at several Californian seismic networks. To suppress artifacts and obtain a high resolution image, we have applied seismic migration method called Slowness Back azimuth Weighted Migration (SBWM) which utilizes not only travel time but also slowness and back azimuth information in the wavefield. We have observed reflected/scattered waves from heterogeneities associated with subduction processes. The migrated seismic energy has then been evaluated using the jackknife algorithm to determine statistically significant seismic signals. The observed reflected seismic waves can be explained by the subducted former oceanic lithosphere (MORB) in the deep mantle, which provides independent evidence for slab penetration into the lower mantle and mass transportation across the mantle transition zone, at least in this region.
DI52A-03
Plate Boundary Forces at Subduction Zones: Effects of Plate Bending and Back-Arc Orogeny on Global Plate Motions
Deformation of both subducting and overriding at convergent plate boundaries tends to dissipate energy that would otherwise be used to drive plate motions. For subducting plates, the magnitude of the bending deformation is not known because of poor constraints on slab strength. For overriding plates, back-arc orogeny results from upper plate shortening and frictional stresses on the plate interface that also resist subduction. Because both processes tend to slow oceanic plates, the observed plate motions can be used to constrain their importance relative to other plate-driving forces. We estimated the resistance from plate bending and back-arc orogeny for 207 globally-distributed subduction zone transects. For bending, we used new measurements of the bending curvature determined from slab seismicity. For orogeny, we estimated the resisting force based on the orography of the overriding plate. The effective viscosity of the bending slab and the effective frictional resistance along the subduction zone interface were considered to be unknown parameters. Using a global mantle flow model to predict plate motions under the influence of these effects, we constrain the viscosity of the bending slab to be at most ~ 300 times more viscous than the upper mantle; stronger slabs are intolerably slowed by the bending deformation. Weaker slabs, however, cannot transmit a pull force sufficient to explain rapid trenchward plate motions unless slabs stretch faster than seismically observed rates of ~ 10-15 s-1. The constrained bending viscosity (~ 2 × 1023 Pa s) is larger than previous estimates because slabs bend more gently than previously thought, with an average radius of curvature of 390 km that permits subduction of strong slabs. We find that back-arc orogeny also significantly affects plate motions, especially for plates such as the Nazca plate where orogenic uplift dominates the back arc region.
DI52A-04
A Global Model of Mantle Convection that Incorporates Plate Bending Forces, Slab Pull, and Seismic Constraints on the Plate Stress.
We introduce a global mantle convection model employing mantle density anomalies inferred from seismic tomography to determine present day plate motions. Our approach addresses two aspects that are not usually considered in previous work. First, we include forces associated with the bending of subducting plates. The bending forces oppose the plate motion, and may be comparable in magnitude to other important forces at subduction zones, including slab pull. Second, our model incorporates data from the Global CMT Catalog. We use the focal mechanisms of earthquakes associated with subducting slabs to estimate the relative occurrence of compressional and tensional axes in the down-dip direction of subducting slabs. This information is used to infer the state of stress in the subducting slab, which we use to calculate slab pull forces. We investigate regional variations in slab pull by comparing plate motions derived using seismic constraints with those derived using slab pull forces based solely on the age of subducting plates. Furthermore, we constrain the rheology of subducted plates by comparing plate motions predicted with and without bending forces. Although our current model uses only radial variations in mantle viscosity, we include the capability of permitting lateral variations in viscosity by calculating buoyancy and plate-driven flows using Citcom
DI52A-05 INVITED
Estimates of the Basal-Strength Torques and Tractions That Drive the Plates From Below
We divide the torques on each surface plate into 3 parts: lithostatic-pressure, side-strength, and basal-
strength. (The strength tensor is the stress tensor minus lithostatic pressure.) We compute each part for 52
plates using a thin-shell finite-element model of the lithosphere with: topography, variable heat-flow, variable
crust and lithosphere thicknesses from seismic data, transient geotherms, nonlinear rheology, and weak
faults. We iterate solutions, adjusting boundary conditions to get correct plate velocities without the need for
any model of deep mantle flow. Uncertainty remains because side-strength torques and inferred basal-
strength torques depend on the effective friction of faults. Therefore, we compute a suite of models with
differing trench resistance and differing fault friction, and evaluate their misfits relative to: seafloor spreading
rates, geodetic velocities, intraplate stress directions, and azimuths of seismic anisotropy. The minimum
misfit occurs at effective fault friction of 0.1 and trench resistance 2x1012 N/m.
Unfortunately, computed values of mean basal-strength traction systematically increase for smaller plates.
We analyze error sources and find that the largest is unmodeled variation in effective friction of plate-
boundary faults. Discounting highly uncertain results, we find mean basal shear tractions of no more than 1
MPa for the 6 largest slabless plates: AF 0.2 MPa; AN 0.1 MPa; NA 0.6 MPa; EU 1.0 MPa; SA 1.0 MPa; SO
0.9 MPa.
The directions of basal shear traction on these plates are generally forward, meaning subparallel to absolute
velocity. Basal-strength torques on plates with subducting slabs represent the sum of net slab-pull and
distributed basal shear traction; if these torques are attributed to net slab-pull alone, net slab-pull is generally
toward the trench and of order 5x1012 N/m. However, our attempts to correlate these inferred net slab-
pull values with objective measures like trench depth or plate age have failed, raising doubts about the
validity of this distribution of basal strength.
Present plate motions on Earth appear to be driven primarily by deep mantle convection, rather than by
topography and associated lithostatic pressures. Our conceptual model is that dense slabs drive convective
rolls in the mesosphere, which provide forward/active driving force to many slow-moving (e.g., continental)
plates through basal shear tractions. Plumes contribute only then they lie on spreading boundaries, and
then primarily through their effects on topography and lithostatic pressures.
http://peterbird.name
DI52A-06 INVITED
Dynamically-consistent images of 3-D heterogeneity in Earth's present-day mantle
Images of mantle heterogeneity are most commonly in the form of seismic velocity since seismic waves are the most direct mantle probe. Although these images of present-day structure represent heterogeneity in the mantle, it is difficult to directly translate them to mantle flow for a variety of reasons. These difficulties include the inherent non-uniqueness of tomographic inversion and the uncertainties in the mineral physics parameters linking seismic velocity to density perturbations that are the driving force behind mantle flow. To overcome these obstacles, we have developed a tomographic imaging procedure involving simultaneous inversion of seismic shear-wave constraints and a suite of convection-related observations including the global free-air gravity field, tectonic plate divergences, dynamic surface topography and the excess ellipticity of the core-mantle boundary. The convection-related observations are interpreted via viscous-flow response functions and density perturbations are internally linked to velocity heterogeneity with mineral physics constraints. This joint inversion procedure has allowed us to directly test several hypotheses regarding the style of mantle flow as well as the sources of mantle heterogeneity. Detailed testing reveals that temperature variations likely dominate both shear-wave and density heterogeneity in the non-cratonic mantle. However, notable compositional anomalies are detected that are most evident within the imaged African superplume structures. Time-dependent flow calculations from the jointly-derived density models provide evidence that the (usually) minor compositional anomalies play an important dynamic role, particularly beneath the African plate. The present-day density models have also been used in dynamic flow calculations that predict anomalous flow patterns that coincide with known tectonic features including the New Madrid Seismic Zone, the Colorado Plateau, and several features within the African plate. Collectively, these observations lend strong support to the validity of jointly-derived images of mantle heterogeneity. This work performed under the auspices of the U.S. Department of Energy by Lawrence Livermore National Laboratory under Contract DE-AC52-07NA27344. LLNL-ABS-406713
DI52A-07
Transition zone structure under a stationary hot spot: Cape Verde
We report on a two-year seismic deployment in the Cape Verde Islands to study the upper mantle to determine its structure under a hot spot that is stationary in the hot spot reference frame. We find from analysis of time- and frequency-domain P-to-S receiver functions estimated from broadband seismic recordings that, within uncertainty, the time separation between the 410 and 660 km discontinuities is normal compared to radial earth models. The average of lags between P-to-S conversion from 660 km and from 410 km, measured at frequenies between 0.1 and 2 Hz, is 24.21±0.43 s (frequency domain) and 24.15±0.47 s (time domain), which is identical to radial earth model predictions: 24.03 s (AK135) and 24.12 s (iasp91). The mean 410 lag is 46.2±0.14 s, about 2.3 s longer than model prediction, and suggests that a slow upper mantle exists under Cape Verde, in accord with other studies. These results indicate that to exist, even stationary hot spots do not require vertical thermal anomalies from deep melting sources anchored in the lower mantle or at the core-mantle boundary.
DI52A-08
A Hotspot-Independent Absolute Plate Motion Model With Implications for Mantle Flow and Hotspot Motions
The motion of the lithosphere relative to the deeper mantle, i.e., absolute plate motion (APM), is a fundamental attribute of the Earth's inner dynamics, and so APM can provide a window into plate tectonic evolution, hotspot motion, and mantle deformation and flow. Although hotspot tracks have historically been used to estimate APM, more recent studies indicate that hotspot tracks provide unreliable measures of APM (not all plumes originate in the lower mantle and/or are fixed to each other). I report an APM model that is, for the first time, independent of hotspot motion, and so provides an unbiased reference frame for quantifying hotspot motions and associated mantle flow. The APM model is derived from the alignment of geodetic plate velocities with published polarisation-directions of split SKS waves underneath stable plates. The used splitting directions reflect lattice preferred orientation of minerals in the asthenosphere generated by the differential motion between the lithosphere and the lower mantle. The model fits data underneath some stable continental lithosphere (India, Africa, Canada, eastern Antarctica) about equally well as underneath oceans, suggesting that the presence of APM-induced asthenospheric flow is not limited to oceanic mantle. Elsewhere, anisotropy directions are likely affected by anomalous density-driven flow, effects of continental keels, and lithospheric structure. The results show that for most hotspots the predicted APM directions follow observed track azimuths very closely, which confirms that APM has first order control on hotspot trends. Small differences between observed hotspot migration vectors and predicted APM velocity vectors, together with information of the plume's depth of origin, can often be explained in terms of (sub-)asthenospheric flow, consistent with independent conclusions and models. For example, the predicted sub-asthenospheric motion of the Hawaiian plume (12.5 mm/yr to the south) is very close to that predicted from mantle flow models. Particularly consistent is the evidence for asthenospheric return-flow near oceanic ridges and for plate motions being self-driven rather than driven by sub-asthenospheric flow (at least at used hotspots locations).