DI12A-01 INVITED
Normal-Mode Seismology at nHz Frequencies
Oscillations in the Earth's core are inferred from variations in the magnetic field over periods of several decades. The observed periods are consistent with expectations for a type of hydro-magnetic wave known as torsional oscillations. We view these oscillations as a set of very-low-frequency normal modes in which the internal magnetic field replaces elasticity and gravity as the primary restoring forces. By adapting the methods of normal-mode seismology, we construct estimates for the internal structure of the magnetic field, as well as a number of other parameters on which the waves depend (including the shape and viscosity of the inner core). Comparison of the recovered field structure with the predictions of numerical dynamo models provides valuable insights into the nature of convection and field generation in the core. We find strong evidence of columnar convection in the core, although the strength of the resulting field is not substantially different from the observable part of the field at the top of the core. We also use the normal modes to recover the source function for the oscillations. Much of the excitation appears to originate near the surface of a cylinder that is tangent to the equator of the inner core. Distinct events (core-quakes) rise above a background level of excitation, and are probably related to flow instabilities in the geodynamo. We speculate about the origin of the instabilities and describe how future progress will illuminate deep Earth processes
DI12A-02 INVITED
Mass flux across the 670 km discontinuity as constrained by the Earth's Argon-40 budget
Approximately 50% of the 40Ar that has been produced from radioactive decay of 40K now resides in the atmosphere and continental crust. If the 40Ar concentration of the upper mantle is representative of the whole mantle, ~ 30-50% of the 40Ar produced over Earth history would be missing. Processing the mantle through partial melting efficiently extracts 40Ar. Therefore, a large mantle reservoir that is largely isolated from the convective upper mantle is required to balance the Earth's 40Ar budget. Geophysical and geochemical observations, however, suggest significant mass exchange between the upper and lower mantle, implying that most of the Earth's mantle has been processed through partial melting at either mid-ocean ridges or at hotspots. This is the so-called Ar paradox. We suggest that convective mixing of noble gas depleted subducted slabs with the ambient mantle preserves high concentrations of volatiles in a mantle reservoir that has been extensively processed by partial melting. Efficient mixing of noble gas depleted slabs with the ambient mantle gradually dilutes the gas concentrations in the mixed mantle assemblage. The rate at which noble gases are lost from the mantle is proportional to the gas concentration of the mantle assemblage being melted. Hence, the monotonic decrease in noble gas concentrations in the mixed mantle assemblage results in a decrease in outgassing rate for 40Ar, even if the rate at which the reservoir is being processed through partial melting is held constant. We quantify the above process using geochemical reservoir modeling. We find that even for a net mass transport across the 670 km discontinuity that is on order 0.5-1.0 lower mantle mass, 40Ar concentrations in the lower mantle can be a factor of 10 higher than in the upper mantle, sufficient to satisfy the 40Ar budget. Overall, the 40Ar budget does not require hidden reservoirs or convective isolation for any length of time and can be satisfied within the framework of a more processed upper mantle and less processed lower mantle.
DI12A-03
Constraints from Earth's heat budget on the origins of deep mantle heterogeneity
To understand the origin and evolution of mantle heterogeneity, a multidisciplinary approach is required. A successful model must satisfy the constraints provided by seismology, which reveals the current state of the mantle, and geochemistry, which reflects the time-integrated history of the planet. One of the most important observations is the prominence of the degree-2 structure in the lower mantle. The large structures in the lower mantle beneath Africa and the Pacific contribute to the degree-2 signal, but their origin is not yet understood. A number of mantle models have been proposed containing multiple reservoirs in various configurations to produce a depleted source of mid-ocean ridge basalts along with the range of long-lived sources inferred from the geochemistry of ocean islands. One important constraint is provided by the Earth's total heat loss. Heat is produced by radioactive decay in the depleted mantle and continental crust; an additional radioactive source is required to make up the heat production of the bulk silicate earth known from as revealed by cosmochemistry and geodynamical models. This heat source may be located in the lower mantle, possibly associated with the superplume structures. This requires a tradeoff between the radiogenic heat productivity of the reservoir and its size: a small volume must contain a higher concentration of heat producing elements to balance the global heat budget. A very small reservoir at the base of the mantle can likely be ruled out because it would become extremely hot (resulting in velocities that are incompatible with seismic models) or buoyant (and therefore short-lived.) The dynamical implications of a dense layer can be assessed using the criteria that the layer must be stable through time, shows topography at its interface, has an effective density profile compatible with seismic models, and has an appropriate heat flux across the core- mantle boundary. A reservoir that is enriched in radiogenic elements compared to the depleted source of mid-ocean ridge basalts would be responsible for the peculiar geochemical signature in some ocean island basalts, but would be rarely sampled at the surface. The geochemical, seismological, and geodynamical evidence together suggest the presence of a hot, abyssal layer that stabilizes mantle plumes and produces the seismic signature of the superplumes.
DI12A-04
Thermal Coupling of the Inner Core to the Lower Mantle
Landmark studies by Adam Dziewonski and colleagues in 1977 and 1984 delineated the basic large-scale 3D seismic structure of the lower mantle, and showed how the long wavelength structure of the lower mantle is dynamically related to the long wavelength gravity field. These pioneering studies emphasized that the spectrum of lower mantle seismic heterogeneity is dominated by the low spherical harmonic degree terms, most notably l=2, an interpretation that has changed very little over three decades time. More recent seismic evidence indicates that the solid inner core also contains large-scale lateral heterogeneity. In particular there is evidence for an m=1 inner core hemispheric dichotomy in P-wave speed and also in P-wave attenuation that extends a few hundred kilometers beneath the inner core boundary. Models of the geodynamo indicate that this inner core hemispheric dichotomy may be dynamically linked to the lower mantle heterogeneity. Dynamo models driven by a heat flow pattern on the core-mantle boundary derived from the seismic heterogeneity of the lower mantle, the so-called "tomographic boundary condition", produce quasi-stationary, large-scale circulations in the liquid outer core that extend from the core-mantle boundary to the inner core boundary. This mantle-controlled circulation tends to concentrate magnetic field on the core-mantle boundary in a way that produces long-term deviations from geomagnetic axial symmetry. The same circulation produces lateral variations in the rate of light element release from the inner core boundary, with higher rates of light element release in the eastern hemisphere compared to the western hemisphere in some models. Provided that the time-averaged orientation of the inner core with respect to the lower mantle heterogeneity has not changed drastically over the past 100 Ma or so, it is possible that the lateral seismic heterogeneity at the top of the inner core is dynamically controlled by the long wavelength 3D structure of the lower mantle.
DI12A-05
Seismic Velocity Anomalies in the Outer Core: The Final Frontier
Variation in density along outer core geoid surfaces must be very small (of order one part in a billion) since the resulting fluid motions and buoyancy fluxes would otherwise be prohibitively large for any reasonable choice of outer core viscosity. In any situation where seismic velocity variations are proportional to density variations (a generalized Birch's "law") this means that the resulting seismic travel time variations in the outer core would be unobservable. The largest lateral variations in the outer core are thus likely to arise from the distortion of geoid surfaces caused by density anomalies in the mantle or inner core. However, these do not change on decadal timescales and would be very difficult to separate from the inner core or mantle variations that cause them. Nonetheless, a recent study (Dai and Song, GRL, vol. 35, L16311, doi:10.1029/2008GL034895) provides evidence for time-variable outer core seismic velocity at the level of ten parts per million. Assuming this is real, I argue that the best candidate explanation is that all or part of the outer core is a two-phase medium consisting of a small mass fraction of small (ten or 100 micron-sized) particles of exsolving silicate material suspended in the convecting liquid. The seismic velocity of this two phase medium can vary at the desired level should the size distribution of particles vary from place to place (and with time) as one would expect in a convecting system, even though the mean density of the medium is invariant at the level of a part per billion, as required by dynamical considerations (thus invalidating Birch's "law"). The seismic velocity variation depends on the ratio of diffusion times to seismic periods, where the diffusion times are thermal or compositional for the particles or the particle spacing. This idea is not new (cf. Stevenson, JGR, 1983) but gains increased impetus from recent work on the nature of core formation and the desirability of an additional energy source for driving the geodynamo, as would arise if of order 10km of mantle underplating occurred over all of geologic time. The amount of suspended material will be tiny at any one time, illustrating the remarkable sensitivity of seismic waves to the microstructure of the medium. Consequences of this picture include some dissipation (finite Q) in the outer core and a significant frequency dependence of this effect, but precise predictions are difficult because of uncertainties in particle kinetics and convective velocities. The two-phase region may also influence radial seismic velocity profiles, particularly in the layers immediately adjacent to the boundaries (e.g., the layer just below the core-mantle boundary), an effect that has been suggested in the literature on many occasions. Even so, this explanation for lateral variability remains marginal at best, suggesting that the claimed observation is either not real or that some other explanation still awaits discovery.
DI12A-06
Models of thermo-chemical convection: what ingredients are needed to explain probabilistic tomography?
A growing amount of seismological observations, including models of probabilistic tomography, indicate that strong lateral density anomalies, likely due to compositional anomalies, are present in the deep mantle. Thus, a successful model of mantle convection must be able to create large thermo-chemical structures in the lower mantle – in particular at its bottom – that can survive convection for a long period of time, and explain present day tomographic images. To determine the parameters that control the formation and survival of large thermo-chemical structures, we have conducted a systematic study in the model space of thermo- chemical convection using STAG3D, and tested the thermo-chemical distributions from these models against those from probabilistic tomography. We have identified five important ingredients for a successful model of thermo-chemical convection. (1) A reasonable buoyancy ratio, between 0.15 and 0.25 (corresponding to chemical density contrasts in the range 60-100 kg/m3). Larger density contrasts induce stable layering for long period of time, rather than pools with strong topography as suggested by seismic observations. (2) A moderate (in the range 0.1-10), chemical viscosity contrast. Small chemical viscosity contrasts induce rapid mixing, whereas large chemical viscosity contrasts lead again to stable layering. (3) A large (> 104) thermal viscosity contrast. Temperature-dependent viscosity creates and maintains pools of dense material with large topography at the bottom of the mantle. (4) A 660-km viscosity contrast around 30. (5) A Clapeyron slope of the phase transition at 660-km around 1.5-3.0 MPa/K. We have also tested models that include combinations of the previous ingredients. The power spectra of these models are in excellent agreement with those from probabilistic tomography except for the spectra of chemical anomalies in the layer located right below the 660-km boundary. This discrepancy might be related to the stacking of slabs around these depths. Because our treatment of the chemical field does not specifically account for the compositional differences between the descending slabs (MORB and harzburgite) and the regular mantle (pyrolite), our models cannot reproduce the chemical signal due to eventual stacked slabs. This will be fixed in future works, which will also include calculations in spherical geometry and the effects of the post-perovskite phase transition.
DI12A-07
S-wave velocity structure beneath the South Pacific Superswell derived from passive seismic experiments
Three-dimensional S-wave velocity structure beneath the South Pacific Superswell is obtained from joint broadband seismic experiments on the ocean floor and islands. As the noise level is significantly high in the horizontal components, we collected only approximately 800 relative times of long-period teleseismic SH- waves (T = 13 to 33 s) by using a waveform cross-correlation from 76 events occurring from Jan. 2003 to May 2005. We conducted relative time tomography with TOMOG3D2 (Zhao et al., 1992) to depths of 2000 km. The velocity perturbations are assigned to grid nodes with horizontal intervals of 5 and 6 degrees for latitudes and longitudes, respectively, and vertical interval of 400 km. The rms residual of 1.2 s is improved to 0.54 s after three time iterations. In the resultant image, we find an elongated large low velocity region beneath the Society to Pitcairn hotspots at 1200km depth. Such a low velocity region can be identified in the some global S-wave velocity models, for example SAW24B16 (Megnin and Romanowicz, 2002). At 800 km depth, two linear low velocity regions are located beneath Tuamotu and Austral islands. These S-wave low velocity regions in the lower mantle are qualitatively similar to those of P-wave by Tanaka et al. (2008). However, unlike the P-wave tomographic image, the S-wave low velocity region at 1200 km depth seems to be isolated from the deeper low velocity region as seen in the global models. Furthermore low velocity regions beneath major hotspots are unclear at 400 km depth. Only a relatively low velocity channel is found in the eastern neighbor of the Society hotspot between 400 and 800 km depth. These features are not consistent with the P-wave velocity structure, suggesting the depth variation of dlnVs/dlnVp.
DI12A-08
Migration and Radon Imaging of the Western Pacific Subduction Zones Using SdS Waves
Mantle reflectors are often used as mantle barometers of thermal and compositional variations near major subduction zones, but observations pertaining to their distributions and strengths still could suffer from limited geographical coverage (e.g., converted waves) and lateral resolutions (e.g., SS, PP precursors and ScS reverberations). In this study, we combine waveform migration with Radon transform, two high-resolution methods often utilized in exploration seismology, to explore the three-dimensional (3D) geometry and morphology of subducted oceanic lithosphere beneath the western Pacific island arcs. The availability of the large data set (covering 1980-2007) and the high-resolution techniques substantially decrease the size of the 'footprint' associated with typical cap-averaging schemes of past SS precursor studies. Our results provide a volumetric view of well-known seismic discontinuities and postulated reflectors from the surface down to ~1300 km depths. The migration stacks show convincing evidence of stagnant, sub-horizontal subducting slabs west of the Wadati-Benioff zone at the base of the upper mantle; the olivine phase boundary could be depressed by as much as 35 km along northeastern Manchuria. Furthermore, multiple stockpiles of the original/modified subducted oceanic lithosphere could be present in the deflected part of the slab. In comparison, the 410-km discontinuity has significantly less topography laterally, though modest and highly localized elevations of the associated phase boundary could be identified within the down-going slab. The existence, spatial extent and strength of transition zone signatures are supported by images obtained from a high-resolution Radon transform of the data (based on an adaptive averaging scheme). Analysis of other reflective signatures, such as a disjointed but unmistakable 520 km discontinuity and localized lower- mantle reflectors, provide further evidence of littered oceanic lithosphere within the mantle transition zone and, quite possibly, at shallow lower mantle (down to 1000 km) depths.