U42A-01
Seismic characteristics of the crust in the Paraná Basin-south Brasília Belt surface limit, southeast Brazil.
The limit between Paraná Basin and Brasília Belt provinces is a suture zone formed due to the convergence of São Francisco and Paranapanema cratons/plates during Neoproterozoic time. The Paranapanema Craton is nowadays under the Paraná Basin. The Moho depth, mean P velocity (Vp) and mean Vp/Vs ratio of the crust along Paraná Basin-south Brasília Belt surface limit is herein presented. They were obtained by the conjugation of analysis of Santa Juliana deep refraction data and of receiver function results from NP4B and IGPB seismographic stations. South Brasília Belt shows an upper crust 29 km thick with Vp of 6.0 km/s and a lower crust 10 km thick with Vp of 6.95 km/s. The Moho is 39 km deep slightly inclined to west. Mean crustal Vp and Vp/Vs are respectively 6.25 km/s and 1.69. The Vp/Vs ratio and Moho depth are in agreement with four other receiver function results of the external zone of the belt. It matches with refraction results of external zone of northern Brasília Belt as well. The difference is the upper crust (first 30 km) velocity that is 6.0 km/s in the south. It is a remarkable low value that suggests differences in composition and/or physical properties between basement rocks from north to south of the belt. This low velocity is possibly related to the strong absorption suffered by seismic energy in the southern experiment. The crust rocks of southern Brasília Belt were stretched in the past (~ 0.9 Ma) forming the basement of the passive margin environment related to former Goiás Ocean. In contrast, northern crust was steeply broke, being characterized by block movements, and forming an abrupt transition between the former continental and ocean crust. The Paraná Basin NE border was poorly solved by refraction data. Seismic energy was strong absorbed and even Moho reflection (PmP) was not undoubtedly determined. The PmP of the west shot suggests a Vp of 6.6 km/s and a Moho depth of 43 km. These results are corroborated by receiver function results from NP4B and IGPB stations that converge to Vp/Vs of 1.72 and to Moho depth of 42.5 km, considering crustal Vp of 6.6 km/s. The mean Vp/Vs ratio of 1.72 is surprisingly low to Paraná Basin domain. It suggests a felsic composition to the crust of this region, although the expectation was finding high Vp/Vs values expressing the basaltic flows of Paraná Basin. Mean Vp/Vs of 1.72 suggests the presence of Brasília Belt/São Francisco Craton basement rocks under the Paraná Basin NE border.
U42A-02
The Lowermost Mantle Beneath Central America Imaged by Kirchhoff Migration of Scatterers and Reflectors
We use tens of thousands of seismograms from South and Central American earthquakes recorded by western North American seismic networks to image the lowermost mantle beneath Central America using a 3D Kirchhoff migration method. P wave studies of the deep mantle often rely on some form of stacking of many records in order to enhance the signal-to-noise ratio of weak phases generated by deep structure, such as reflections off of the D" discontinuity. These methods, however, often assume one-dimensional structure, which is at odds with the evidence for significant heterogeneity. Kirchhoff migration is a three-dimensional stacking method that allows interactions with structure outside of the source-receiver plane, thus illuminating a much larger volume. The D" discontinuity beneath Central America has been readily observed in S wave studies and may be the result of the shear wave velocity increase associated with the recently discovered perovskite to post-perovskite phase transition. This phase transition is expected to have weaker effects on P wave velocities than on S wave velocities and the sharpness of this transition is unknown. Using data at post-critical distances, we observe structures consistent with a P velocity discontinuity about 200 km above the core-mantle boundary (CMB). Observing this using short period data suggests that the boundary must be less than a few 10s of km thick, while observation with lower frequency broadband data exclude the possibility of it being a thin layer. Whether this discontinuity is co-located for both P and S waves is difficult to resolve. Both the broadband and the short period P wave data sets also reveal a sharp out-of-plane scatterer, which may be located close to the CMB. The short period data also indicate reflectivity about 400 km above the CMB, well above the D" discontinuity, and similar reflectivity is observed under the Central Pacific. This feature appears to be more consistent with a discontinuity than a scatterer, is hinted at in the broadband data set and is not observed with S waves. We also present results using P wave data from pre-critical distances. While reflection and scattering coefficients are much lower at these distances, the imaging volume is much larger. Results from these data hint at a complicated lower mantle with multiple features whose origins may lie anywhere between the core-mantle boundary and many hundred km above it.
U42A-03
Boundary Layer Dynamics and Sub-Adiabaticity in Convecting Planetary Mantles
A broad range of phenomena are influenced by the behavior of thermal boundary layers in planetary mantles including plume temperatures, lithospheric stresses, resistance to plate motions, and the temperature structure of the mantle as a whole. The textbook picture of the temperature profile in a convecting layer consists of two boundary layers separated by a well-mixed, adiabatic interior. The sum of the temperature drops across the upper and lower boundary layers is equal to super-adiabatic temperature drop across the entire layer. This picture does not accurately describe, however, the horizontally averaged temperature structure derived from numerical solutions of the equations of infinite Prandtl number, Boussinesq convection. The sum of the average temperature drops across the boundary layers in such models is always greater than the super-adiabatic drop across the whole layer, with the result that some portions of the interior are sub-adiabatic. The excess average temperature drop across each boundary layer is due to the arrival of material from the other boundary layer which has not equilibrated with the well-mixed interior. It is this material which transfers heat conductively across the boundary and thus controls the heat transport of the layer. Internal heating breaks the symmetry of the boundary layers (as does temperature dependence of viscosity), and it is the interaction between the two boundary layers that sets the equilibrium temperature drops. The scaling of the temperature drop across each boundary layer is controlled by two competing factors which depend on the Rayleigh number in different ways: the scale of boundary layer instabilities and the velocity of plumes (hot and cold). Furthermore, these scalings change as the system becomes time-dependent at moderate Rayleigh number. At very high Rayleigh number, beyond that of most planetary mantles, the plumes do equilibrate with the interior and the textbook picture applies. A scaling theory for the average temperature drop across the boundary layers will be presented and compared to numerical solutions for isoviscous and non-Newtonian rheologies, with and without internal heating.
U42A-04
Density Structures of Oceanic Slabs and Surrounding Mantle Around the 660 km Discontinuity: Implications for the Fate of Old and Young Slabs
We calculated density variations as a function of temperature around the 660 km deep seismic discontinuity in the Earth's mantle in different types of compositional units associated with a subducting slab and the ambient mantle. The calculations are based on computational thermodynamic approach of minimization of Gibbs free energy at a specified P-T condition, subject to the bulk compositional constraints of the system, that simultaneously yield stable mineral assemblage, mineral compositions and modal abundances. These results are converted to density profiles using appropriate data for physical properties and equations of state that are applicable to high P-T conditions. In addition, we also calculated thermal structures of several slabs, with the extremes being given by Tonga (140 Myr at trench, vertical velocity: 14 cm/yr; average dip: 60 degrees) and Peru (41 Myr, 4.4 cm/yr, 35 degrees). A slab was assumed to be lithologically stratified with a top basaltic crust, followed downwards by residual harzburgite and slightly depleted pyrolite. The surrounding mantle is taken to be undepleted pyrolite. Integration of the results of thermal and density calculations show that that all components of the Tonga slab are heavier than both ambient and thermally perturbed adjacent mantle, which has been cooled due to contact with the slab. Thus, old slabs with thermal minimum below 750 C, as in Tonga, should easily penetrate into the lower mantle unless resisted by slab roll back and/or a viscosity jump at the top of the lower mantle. In contrast, the harzburgite layer in warmer slabs, such as Peru, Marianas and Izu-Bonin, is slightly lighter than ambient mantle, causing near neutral or marginally negative net buoyancy of the slab. In this case, other factors, such as rollback and slab dip angle, may explain why some slabs in the northwest Pacific appear to penetrate into the lower mantle while others do not. In Peru-type warm slabs, buoyancy of the harzburgite layer may cause it to stagnate at the top of lower mantle, and thus deform the slabs. Finally, we find that regardless of temperature, the basaltic top part of a slab is always significantly heavier than surrounding mantle. Thus, should it become mechanically decoupled, it should sink into the lower mantle instead of peeling off from the slab to form a "perched eclogite layer", as suggested by some earlier studies.
U42A-05
Integrating Full Three-dimensional Mantle Circulation Models with Lithospheric Deformation Models
Our study provides insight into the lithosphere-mantle coupling problem through joint modeling of lithosphere dynamics and three dimensional mantle circulation. We use the global lithospheric stress field as well as plate motions in order to infer the nature of coupling between the lithosphere and the mantle. There are two types of stresses acting on the Earth's lithosphere, (1) internal buoyancy forces arising from lateral density variations within the lithosphere and (2) basal tractions associated with mantle convection that gets coupled to the base of the lithosphere. The relative contribution of these stresses vary laterally depending on the viscosity of the asthenosphere. Lateral viscosity variation arises through various factors, such as, presence of continental keels below Archaean shields, temperature differences in the oceans due to cooling of old oceanic lithosphere, as well as presence of weak, deforming plate boundary zones. We incorporate laterally variable viscosities within the asthenosphere in our mantle convection model and compute the global lithospheric stress field response to the traction distribution. The lithosphere response is computed using a thin-shell method with lateral viscosity variations, internal density buoyancy variations, and applied basal traction distributions computed from mantle circulation models. The modeled total deviatoric stress field is scored with deformation indicators from the Global Strain Rate Map (GSRM), which is constrained using over 5000 GPS velocity vectors. We also compare the predicted plate motions generated by the convection model with observed plate velocities. Based on such a combination of matching plate velocities and deformation indicators we are able to eliminate certain models and the joint criteria enable us to significantly narrow down the range of possible models that fit the observations. We also employ an inverse method and directly solve for the distribution of basal tractions that yields a best-fit lithospheric response that is scored with the GSRM. We define the deviatoric stresses associated with the global distribution of basal tractions using a spherical harmonic expansion out to degree and order 9, and solve for the coefficients of this series in a least-squares inversion. The best-fit solution suggests roughly equal contributions of lithospheric stress associated with GPE differences and basal tractions associated with mantle circulation. Moreover, the best-fit model shows traction distributions that are similar to our best-fit mantle convection model. In both the inverse and forward models, zones of mantle downwelling, most likely associated with the history of subduction, play a prominent role for North and South America, and central Asia.
U42A-06
Connecting Top to Bottom: The Relationship Between Tectonic Plates and Mantle Plumes
Some 15 years ago Nataf noted that thermal convection in purely temperature dependent fluids could not produce thermal upwellings of the type invoked by mantle plume theories of the day. Since then there have been bursts of investigations addressing the question of how active overturn of the lithosphere, which does not happen in purely temperature-dependent viscosity convection, might influence the dynamics and morphology of mantle upwellings. The first wave of these investigations made specific theoretical predictions and conjectures that went largely untested. Recent advance in numerical techniques and computer power have allowed for the exploration of convection simulations with very large viscosity contrasts. At the same time, advances in laboratory methods have allowed for fluid tank experiments that can mimic active plate subduction. These two expanded research tools have allowed us to revisit the connection between plates and plumes. As well as launching new investigations we have also made efforts to confirm or refute previous theoretical speculations regarding the influence of plate tectonics on the dynamics of the core-mantle boundary region. We will briefly review the history of the problem and present our recent numerical and laboratory results that explore the potential connections between tectonic plate, mantle plumes, and the structure of the core mantle boundary region.
U42A-07 INVITED
Convective Differentiation of the Earth's Mantle
The differentiation of the Earth is likely to be influenced by convective motions within the early mantle. Double- diffusive convection (d.d.c), driven by thermally and compositionally induced density differences is considered as a vital mechanism behind the dynamic differentiation of the early mantle.. We demonstrate that d.d.c can lead to layer formation on a planetary scale in the diffusive regime where composition stabilizes the system whil heat provides the destabilizing force. Choosing initial conditions in which a stable compositional gradient overlies a hot reservoir we mimic the situation of a planet in a phase after core formation. Differently from earlier studies we fixed the temperature rather than the heat flux at the lower boundary, resembling a more realistic condition for the core-mantle boundary. We have carried out extended series of numerical experiments, ranging from 2D calculations in constant viscosity fluids to fully 3D experiments in spherical geometry with strongly temperature dependent viscosity. The buoyancy ratio R and the Lewis number Le are the important dynamical parameters. In all scenarios we could identify a parameter regime where the non-layered initial structure developed into a state consisting of several, mostly two layers. Initially plumes from the bottom boundary homogenize a first layer which subsequently thickens. The bottom layer heats up and then convection is initiated in the top layer. This creates dynamically (i.e. without jump in the material behavior) a stack of separately convecting layers. The bottom layer is significantly thicker than the top layer. Strongly temperature dependent viscosity leads to a more complex evolution The formation of the bottom layer is followed by the generation of several layers on top. Finally the uppermost layer starts to convect. In general, the multilayer structure collapses into a two layer system. We employed a numerical technique, allowing for a diffusion free treatment of the compositional field. In each case a similar evolution has been observed. This indicates that a temporary formation of layered structures in planetary interiors is a typical phenomenon. Moreover, in this scenario, plate tectonics appears only in later stages of the evolution.