Mantle convection and plate dynamics transfer and deform solid material on scales of hundreds to thousands of km. However, viscoplastic deformation of rocks arises from motions of defects at sub-crystal scale, such as vacancies or dislocations. In this study, results from numerical experiments of dislocation dynamics in olivine for temperatures and stresses relevant for both lithospheric and asthenospheric mantle (800–1700 K and 50–500 MPa; ) are used to derive three sigmoid parameterizations (erf, tanh, algebraic), which express stress evolution as a function of temperature and strain rate. The three parameterizations fit well the results of dislocation dynamics models and may be easily incorporated into geodynamical models. Here, they are used in an upper mantle thermo-mechanical model of subduction, in association with diffusion creep and pseudo-brittle flow laws. Simulations using different dislocation creep parameterizations exhibit distinct dynamics, from unrealistically fast-sinking slabs in the erf case to very slowly-sinking slabs in the algebraic case. These differences could not have been predicted a priori from comparison with experimentally determined mechanical data, since they principally arise from feedbacks between slab sinking velocity, temperature, drag, and buoyancy, which are controlled by the strain rate dependence of the effective asthenosphere viscosity. Comparison of model predictions to geophysical observations and to upper-mantle viscosity inferred from glacial isostatic adjustment shows that the tanh parameterization best fits both crystal-scale and Earth-scale constraints. However, the parameterization of diffusion creep is also important for subduction bulk dynamics since it sets the viscosity of slowly deforming domains in the convecting mantle. Within the range of uncertainties of experimental data and, most importantly, of the actual rheological parameters prevailing in the upper mantle (e.g. grain size, chemistry), viscosity enabling realistic mantle properties and plate dynamics may be reproduced by several combinations of parameterizations for different deformation mechanisms. Deriving mantle rheology cannot therefore rely solely on the extrapolation of semi-empirical flow laws. The present study shows that thermo-mechanical models of plate and mantle dynamics can be used to constrain the effective rheology of Earth's mantle in the presence of multiple deformation mechanisms.

How to cite: Garel, F., Thoraval, C., Tommasi, A., Demouchy, S., and Davies, D. R.: Using thermo-mechanical models to bridge scales between experimental rheology and large-scale observational constraints on mantle and plate dynamics, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-2815, https://doi.org/10.5194/egusphere-egu21-2815, 2021.

The knowledge of the density structure of the lithospheric mantle is critical to our comprehension of tectonic and magmatic events occurring within the lithosphere and crucial to model the evolution of complex geodynamic processes (e.g., subduction dynamics, mantle plume upwelling etc). Furthermore, a thorough understanding of the density evolution at mantle conditions is essential to interpret geophysical data such as seismic tomography (e.g., Afonso et al., 2008; Stixrude and Lithgow-Bertelloni, 2012).

The density of mantle peridotites is related to chemical composition, modal abundance and elastic properties of their constituent minerals, which in turn are controlled by pressure, temperature and bulk composition of the system. Accordingly, the elastic properties of mantle minerals combined with the thermal state of the lithosphere can predict how the physical properties (e.g., density, elastic moduli) of mantle peridotites vary with depth. To this aim, (i) we examined the existing literature data (compressibility, thermal expansion and elasticity) suitable to constrain the elastic properties of peridotite minerals and (ii) we addressed the density structure of two potential lithospheric mantle sections (fertile and depleted) across different thermal regimes from the perspective of the Equations of State (EoS) of their constituent minerals.

In a mantle characterized by a relatively cold geotherm (45 mWm^-2), the density of a depleted peridotitic system remains nearly constant up to about 4 GPa, while it moderately increases in a fertile system. In a mantle characterized by a relatively hot geotherm (60 mWm^-2), the density of both depleted and fertile systems decreases up to about 3 GPa, due to the more rapid raise of temperature compared to pressure, and then it increases downwards.

These preliminary results show that the thermal state of the lithosphere produces a first-order signature in its density structure, with few differences owing to different modes and crystal chemical compositions.

References

Afonso, J.C., Fernàndez, M., Ranalli, G., Griffin, W.L., Connolly, J.A.D., 2008. Integrated geophysical-petrological modeling of the lithosphere and sublithospheric upper mantle: Methodology and applications. Geochemistry, Geophys. Geosystems 9, Q05008.

Stixrude, L., Lithgow-Bertelloni, C., 2012. Geophysics of Chemical Heterogeneity in the Mantle. Annu. Rev. Earth Planet. Sci. 40, 569–595.

How to cite: Faccincani, L., Faccini, B., Casetta, F., Mazzucchelli, M., Nestola, F., and Coltorti, M.: Density structure of the lithospheric mantle: upscaling from minerals to peridotites, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-9565, https://doi.org/10.5194/egusphere-egu21-9565, 2021.

According to geochemical and geophysical observations, Earth's lower mantle appears to be strikingly heterogeneous in composition. An accurate interpretation of these findings is critical to constrain Earth's bulk composition and long-term evolution. To this end, two main models have gained traction, each reflecting a different style of chemical heterogeneity preservation: the 'marble cake' and 'plum pudding' mantle. In the former, heterogeneity is preserved in the form of narrow streaks of recycled oceanic lithosphere, stretched and stirred throughout the mantle by convection. In the latter, domains of intrinsically strong, primordial material (enriched in the lower-mantle mineral bridgmanite) may resist convective entrainment and survive as coherent blobs in the mid mantle. Microscopic scale processes certainly affect macroscopic properties of mantle materials and thus reverberate on large-scale mantle dynamics. A cross-disciplinary effort is therefore needed to constrain present-day Earth structure, yet countless variables remain to be explored. Among previous geodynamic studies, for instance, only few have attempted to address how the viscosity and density of recycled and primordial materials affect their mutual mixing and interaction in the mantle.

Here, we apply the finite-volume code STAGYY to model thermochemical convection of the mantle in a 2D spherical-annulus geometry. All models are initialized with a lower, primordial layer and an upper, pyrolitic layer (i.e., a mechanical mixture of basalt and harzburgite), as is motivated by magma-ocean solidification studies. We explore the effects of material properties on the style of mantle convection and heterogeneity preservation. These parameters include (i) the intrinsic strength of basalt (viscosity), (ii) the intrinsic density of basalt, and (iii) the intrinsic strength of the primordial material.

Our preliminary models predict a range of different mantle mixing styles. A 'marble cake'-like regime is observed for low-viscosity primordial material (~30 times weaker than the ambient mantle), with recycled oceanic lithosphere preserved as streaks and thermochemical piles accumulating near the core-mantle boundary. Conversely, 'plum pudding' primordial blobs are also preserved when the primordial material is relatively strong, in addition to the 'marble cake' heterogeneities mentioned above. Most notably, however, the rheology and the density anomaly of basalt affect the appearance of both recycled and primordial heterogeneities. In particular, they control the stability, size and geometry of thermochemical piles, the enhancement of basaltic streaks in the mantle transition zone, and they influence the style of primordial material preservation. These results indicate the important control that the physical properties of mantle constituents exert on the style of mantle convection and mixing over geologic time. Our numerical models offer fresh insights into these processes and may advance our understanding of the composition and structure of Earth's lower mantle.

How to cite: Desiderio, M., Gülcher, A. J. P., and Ballmer, M. D.: The interplay between recycled and primordial heterogeneities: predicting Earth's mantle dynamics via numerical modeling, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10251, https://doi.org/10.5194/egusphere-egu21-10251, 2021.

Signatures from hotspot lavas fed by mantle plumes suggest a heterogenous mantle source. Deep plumes sample the core-mantle boundary (CMB) region and this region is thought to host primordial and recycled crustal material, possibly in the form of thermochemical piles. The formation of these piles depends on the amount of oceanic crust subducted into the lower mantle and how much is entrained back toward the surface. However, it is unclear how and under which conditions the oceanic crust can segregate from subducted slabs to form these piles and eventually be entrained in ascending mantle plumes. It has been suggested that the bridgmanite to post-perovskite phase transition facilitates this segregation, as low viscosity post-perovskite allows for thinning and stretching of crustal material. This process is difficult to model numerically, since crustal material is often thinned to very small length scales. Thus, it usually cannot be resolved in global convection models, leading to over-estimates of entrainment and consequently impacting the predicted formation of basaltic piles.  Furthermore, the deformation of the crust as the slab descends into the lower mantle changes the initial surface crustal thickness and hence how likely the material is to form piles or become entrained. To address these uncertainties, we model a descending slab in the lower mantle to re-assess basalt entrainment and accumulation near the CMB. We use an adaptive mesh and tracers in order to track the deformation of the crust to achieve high resolution and also test different crustal thicknesses. These models provide insights into how material is added and removed from reservoirs in the lowermost mantle, and how these rates of material exchange have varied throughout Earth history.

How to cite: Chotalia, K., Dannberg, J., and Gassmoeller, R.: Crustal thickness and resolution controls on basalt entrainment in the lower mantle, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10836, https://doi.org/10.5194/egusphere-egu21-10836, 2021.

Long-lived, Mesozoic-Cenozoic subduction zones such as the Pacific slab under the Americas and the Tethyan slab under Eurasia consumed thousands of kms of lithosphere of which remnants are detected in today’s mantle by seismic tomography. Major differences, however, in subduction zone evolution occurred between these systems which include strong variations in subduction rate, slab morphological evolution, and trench motion, which all appear mostly to be accommodated in the upper 1000 km of the mantle (van der Meer et al. 2018). Furthermore, sinking rates of slabs below this zone tend to be similar for different subduction systems and an order of magnitude smaller than their plate/subduction velocities. Working from the premise that the mantle rheology that accommodated these subduction systems is basically similar, although still poorly constrained, we test the hypothesis that the contrasting evolution of these subduction systems is primarily tied in with the global plate tectonic forcing of subduction.

It is generally accepted that plate motion is primarily driven by slab pull with contributions from ridge push, rather than the drag of the underlying mantle. If correct, numerical subduction models should be able to obtain upper as well as lower mantle subduction velocities and sinking rates similar to those reconstructed from geological records. We are at the start of this investigation and will present the numerical model setup, modeling strategy, and preliminary results of a 2-D subduction modelling experiment. We implement a 2D-cylindrical model setup for solving the conservation of momentum, mass and energy with the open-source geodynamics code ASPECT (Kronbichler et al. 2012) using a nonlinear visco-plastic rheology and including the major phase changes. Our focus is on the possible role of the absolute motion of the subducting and overriding plates in concert with slab pull variation reconstructed from plate tectonic evolution models, while in both subduction cases the same (partly nonlinear) mantle rheological processes are required to accommodate slab morphology change and slab sinking. Kinematic modelling constraints are derived from global plate tectonic evolution models, while the tomographically inferred present-day stage provides the end-stage geometry of slabs.

van der Meer, D. G., Van Hinsbergen, D. J., & Spakman, W. (2018). Atlas of the underworld: Slab remnants in the mantle, their sinking history, and a new outlook on lower mantle viscosity. Tectonophysics, 723, 309-448.

Kronbichler, M., Heister, T., & Bangerth, W. (2012). High accuracy mantle convection simulation through modern numerical methods. Geophysical Journal International, 191(1), 12-29.

How to cite: van der Wiel, E., Thieulot, C., Spakman, W., and van Hinsbergen, D.: Dynamics of upper and lower mantle subduction and its effects on the amplitude and pattern of mantle convection., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-11502, https://doi.org/10.5194/egusphere-egu21-11502, 2021.

The rheology of the Earth’s lower mantle is poorly constrained due to a lack of knowledge of the rheological behaviour of its constituent minerals. In addition, the lower mantle does not consist of only a single, but of multiple mineral phases with differing deformation behaviour. The rheology of Earth’s lower mantle is thus not only controlled by the rheology of its individual constituents (bridgmanite and ferropericlase), but also by their interplay during deformation. This is particularly important when the viscosity contrast between the different minerals is large. Experimental studies have shown that ferropericlase may be significantly weaker than bridgmanite and may thus exert a strong control on lower mantle rheology.

Here, we thus explore the impact of phase morphology on the rheology of a ferropericlase-bridgmanite mixture using numerical models. We find that elongated ferropericlase structures within the bridgmanite matrix significantly lower the effective viscosity, even in cases where no interconnected network of weak ferropericlase layers has been formed. In addition to the weakening, elongated ferropericlase layers result in a strong viscous anisotropy. Both of these effects may have a strong impact on lower mantle dynamics, which makes is necessary to develop upscaling methods to include them in large-scale mantle convection models. We develop a numerical-statistial approach to link the statistical properties of a ferropericlase-bridgmanite mixture to its effective viscosity tensor. With this approach, both effects are captured by analytical approximations that have been derived to describe the evolution of the effective viscosity (and its anisotropy) of a two-phase medium with aligned elliptical inclusions, thus allowing to include these microscale processes in large-scale mantle convection models.

How to cite: Thielmann, M., Golabek, G., and Marquardt, H.: The importance of phase morphology for rheology of ferropericlase-bridgmanite mixtures, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-4384, https://doi.org/10.5194/egusphere-egu21-4384, 2021.

In situ measurement of solid-state deformation in a large volume press has historically required use of neutron and x-ray scattering facilities. The lack of widespread availability of these facilities has limited the abilities of researchers to measure in situ deformations on a regular basis. We have developed an assembly that utilizes a piezoelectric crystal within a typical large volume press assembly in a 6-axis press at pressures up to 5 GPa. The basic design of the assembly can be applied to multiple assembly sizes for a wide range of possible pressures. The piezoelectric crystal is a round disk, <1 mm in diameter, that is sputter coated with Au. Copper wires are placed through drilled holes in the side of the assembly, one connected to each side of the disk. The crystal generates a voltage across the two faces when a deviatoric stress is applied that is measured and plotted in real-time during the experiments. The voltage is then used to calculate strain and strain-rate in uniaxial compression. Using the known equation of state of the piezoelectric crystal, such as quartz or gallium orthophosphate, the stresses responsible for the strain can be calculated. Thus, we can measure the stress and strain regime of simple deformation within an assembly in situ in real-time during the deformation. We have measured strain-rates as low as 10^-7 s^-1 over a greater than 30-minute timescale. The total strain on the assembly can be measured by the total distance advanced by the press piston, which must be accommodated. Comparing the differences in strain accommodated by the piezoelectric crystal between separate experiments allows us to infer the strain accommodated by the sample under investigation.

Current limitations in measuring lower strain-rates are charge-leakage around the piezoelectric crystal causing a voltage drift during measurements and limitations in high-temperature experiments due to phase transitions during heating in the piezoelectric crystals to phases that are not piezoelectric. Future work will concentrate on finding a suitable, high-resistance material to place around the piezoelectric crystal to limit charge leakage and designing the assembly such that the piezoelectric crystal experiences lower temperature during heating than the sample to avoid phase transitions in the crystal.

How to cite: Dolinschi, J. and Frost, D.: Development of in situ measurement of solid-state deformation in a large anvil press utilizing a piezoelectric crystal, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-6928, https://doi.org/10.5194/egusphere-egu21-6928, 2021.

The mechanics of olivine deformation play a key role in large-scale, long-term planetary processes, such as the response of the lithosphere to tectonic loading or the response of the solid Earth to tidal forces, and in short-term processes, such as the evolution of roughness on oceanic fault surfaces or postseismic creep within the upper mantle. Many previous studies have emphasized the importance of grain-size effects in the deformation of olivine. However, most of our understanding of the role of grain boundaries in deformation of olivine is inferred from comparison of experiments on single crystals to experiments on polycrystalline samples.

To directly observe and quantify the mechanical properties of olivine grain boundaries, we use high-precision mechanical testing of synthetic forsterite bicrystals with well characterised interfaces. We conduct nanoindentation tests at room temperature on low-angle (13^o tilt about [100] on (015)) and high-angle (60^o tilt about [100] on (011)) grain boundaries. We observe that plasticity is easier to initiate if the grain boundary is within the volume tested. This observation agrees with the interpretation that certain grain-boundary configurations can act as sites for initiating microplasticity.

As part of continuing efforts, we are also conducting in-situ micropillar compression tests at high-temperature (above 600^o C) within similar bicrystals. In these experiments, the boundary is contained within the micropillar and oriented at 45^o to the loading direction to promote shear along the boundary. In these in-situ tests, our hypothesis is that the low-angle grain boundary displays a higher viscosity relative to the high-angle interface. Key advantages of performing in-situ experiments are the direct observation of grain-boundary migration or sliding, simplified kinematics of a single boundary segment, and  potentially changes in style of deformation with different grain-boundary character.

These small deformation volume experiments allow us to qualitatively explore the differences between the crystal interior and regions containing grain boundaries. Overall, the variation in strain and temperature in our small scale experiments allows the fundamental investigation of the response of well characterised forsterite grain boundaries to deformation. 

How to cite: Avadanii, D., Hansen, L., Darnbrough, E., Marquardt, K., Armstrong, D., and Wilkinson, A.: Micromechanical testing of olivine grain boundaries, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-8346, https://doi.org/10.5194/egusphere-egu21-8346, 2021.

      Localized deformation is a possible scenario that may explain the preservation of geochemical heterogeneity in the lower mantle. Recent experimental studies (e.g., Girard et al., 2016) showed that Fp (ferropericlase), which is a weak and volumetrically minor phase (~20 %), accommodates a large fraction of strain of its mixture with Br (bridgmanite), which is a stronger (approximately order of 2 - 3) and volumetrically major phase (~60-70 %). Localized deformation of the Fp phase within this two-phase mixture may provide an important insight to the long-standing question of the mechanical differentiation process between the weak boundary layer and the relatively unmixed volume in the lower mantle. Since the dominant deformation mechanism in the lower mantle is thought to be the diffusion creep, and the deformation state is mostly simple shear, it is important to understand how the deformation by diffusion creep occurs in a mixture of Fp-Br under the simple shear.

In the context of multiscale modeling methodology, we approach a grain’s length scale deformation where a single 2D elliptic Fp grain is embedded in the infinite Br matrix medium. These two phases are treated as linear viscous incompressible materials where deformation occurs by the fluxes of atoms (vacancies) caused by the applied boundary normal stress gradient. We conducted a theoretical analysis to investigate the nature of the diffusion-induced deformation of the Fp grain under the simple shear. We focused on the following issues: (i) when the two-phase mixture deforms under the far-field simple shear, what the local stress and strain rate fields within the Fp inclusion are, (ii) how the local stress gradients induce the diffusion fluxes of vacancies of the Fp grain, and (iii) the dependences of diffusion creep of the Fp to its shape (changing with the strain) and its viscosity contrast against the Br.

      To investigate the internal stress states, we used the Eshelby’s inhomogeneous inclusion theory translating its elastic formulations to the linear viscous ones using the Hoff analogy. This approach provides the stress, strain rate, and vorticity within a 2D elliptic Fp grain embedded in Br (3 orders of magnitude greater viscosity than Fp) matrix subjected to the far-field simple shear. From these mechanical states, the lattice diffusion within Fp grain and its influences on the rheology were found by using the Finite Element method solving the Fick’s laws of diffusion. This study shows that the diffusion creep rate increases as the ellipse elongates and rotates. As the Fp ellipse elongates (i.e., its aspect ratio increases), the local shear stress in the Fp increases, and the stress is somewhat concentrated near the small radius tips, which induces the strong diffusion fluxes due to the high normal stress gradients. These theoretical and numerical results support that the strain localization under diffusion creep regime can occur and be a possible mechanism that created the localized mantle flow particularly where the shear deformation is dominantly applied.

How to cite: Cho, H. and Karato, S.: Toward the Better Understanding of Shear Localization in the Lower Mantle Caused by the Strength Contrast between Bridgmanite and Ferropericlase: the Role of Stress/Strain Rate Heterogeneity in Diffusion Creep Regime, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-10321, https://doi.org/10.5194/egusphere-egu21-10321, 2021.

How strain localizes in the lower crust and upper mantle to accommodate transcurrent plate motions is not well understood. Here we focus on a suite of lower crustal and upper mantle xenoliths from the San Quintin Volcanic Field (SQVF) in Baja California, Mexico, located along transcurrent faults at the margin of the Pacific plate. Previous work has suggested that in addition to significant strain localization, the lower lithosphere below SQVF has experienced partial melting, possibly through shear heating. The presence of even minor amounts of melt could significantly affect the deformation mechanisms accommodating strain. While previous studies of SQVF have largely focused on deformation in the upper mantle, less is known about strain localization in the lower crust. We have analyzed the composition and microstructures of nine xenoliths using wavelength dispersive spectroscopy (WDS) and electron backscatter diffraction (EBSD) to elucidate the relationship between melt infiltration and deformation in the lower crust of this actively-deforming region.

We categorize the suite of SQVF xenoliths into two textural and chemical groups: Group 1, consisting of undeformed mafic cumulates, and Group 2, consisting of foliated ultramafic peridotites and mafic granulites. Symplectites and corona textures with olivine-orthopyroxene-clinopyroxene+spinel symplectite-plagioclase layering preserved in Group 2 samples are interpreted to have resulted from basaltic melt infiltration during deformation. The orientation of the shape preferred orientations (SPO) of spinel and orthopyroxene grains relative to foliation in Group 2 samples is consistent with experimental studies of crystallization during melt infiltration. Evidence for deformation is also preserved in the form of moderate crystallographic preferred orientations (CPO), present in plagioclase, orthopyroxene, and olivine. Oxide weight percentages, calculated using electron microprobe data and modal phase abundances from WDS maps, were used to construct pseudosections in order to estimate equilibrium temperatures and pressures. The range of pressures across samples suggest a changing degree of deformation and degree of rock-melt interaction with depth in the lower crust of Baja California.

How to cite: Murphy, C., Bernard, R., and Chin, E.: Interplay between melt infiltration and strain localization in the lower lithosphere of San Quintin, Baja California, EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-7947, https://doi.org/10.5194/egusphere-egu21-7947, 2021.

Rocks of the Earth's crust and mantle commonly consist of aggregates of different minerals with contrasting mechanical properties. During progressive, high temperature (ductile) deformation, these rocks tend to develop an extrinsic mechanical anisotropy related to the strain competition of the different minerals, the amount of accumulated bulk strain and the bulk strain geometry. Extrinsic anisotropy is thought to play an important role in a wide range of geodynamic processes up to the scale of mantle convection. However, the evolution of grain-scale and rock-scale associated with this anisotropy cannot be directly implemented in large-scale numerical simulations. For two-phase aggregates -a good rheological approximation of most Earth's rocks- we propose a methodology to indirectly approximate the extrinsic viscous anisotropy by a combination of (i) 3-D mechanical models of rock fabrics, and (ii) analytical effective medium theories. The resulting 3-D mechanical models, confirm that the weak least abundant phase induces substantial rock weakening by forming an inter-connected network of thin layers in the flow direction. 3-D models further suggest, however, that the lateral inter-connection of these weak layers is quite limited, and the maximum structural weakening is considerably less than previously estimated. Ont the other hand, presence of hard inclusions does not have a profound impact in the effective strength of the aggregate, with lineations developing only at relatively low compositional strength contrast. When rigid inclusions become clogged, however, the aggregate viscous resistance can increase over the theoretical upper bound. We show that the modelled grain-scale fabrics can be parameterised as a function of the bulk deformation and material phase properties and can be combined with analytical solutions to approximate the anisotropic viscous tensor. At last, the resulting parameterisation of the extrinsic viscous tensor is implemented in a bi-dimensional global mantle convection code. Preliminary results show that extrinsic is responsible for an increase of the upwelling speed of hot material from the lowermost mantle, different convective cell shapes, and deflection of mantle plumes at the uppermost mantle.

How to cite: de Montserrat Navarro, A., Faccenda, M., and Pennacchioni, G.: Extrinsic anisotropy of two-phase Newtonian aggregates: fabric characterisation and parametrisation, and application to global mantle convection., EGU General Assembly 2021, online, 19–30 Apr 2021, EGU21-12719, https://doi.org/10.5194/egusphere-egu21-12719, 2021.