New constraints on the thermochemical properties of Earth’s upper and mid-mantle from ScS reverberation data

Seismic topography models reveal that both upwelling plumes and downgoing slabs are deflected or stagnate at various depths in Earth’s mantle transition zone (MTZ) and mid-mantle (MM). Deflection within the MTZ is associated with the mineral physics phase changes at 410 and 660-km depth, however the cause of deflection in the MM remains debated. There are no candidate mineral transformations to explain the varied MM reflectors that have been detected [Waszek et al., 2018], instead indicating widespread compositional heterogeneities. Furthermore, our recent thermal model [Waszek et al., 2021] reveals a link between high temperatures in the MTZ and surface activity, indicating that some plumes are able to traverse this region unimpeded. Illuminating the detailed seismic structures of the upper and mid-mantle is key to determine the link between reflectors, temperature, composition, and dynamics.

Here, we present a new large global dataset of ScS reverberations, compiled using an automatic waveform identification code based on Convolutional Neural Networks [Garcia et al., 2021]. Mantle discontinuities and reflectors generate precursors to ScSn phases, and postcursors to sScSn. Here, we present a new method to correct for 3D mantle structure in which we remove the symmetry problem suffered by most of these phases. The data are stacked to reveal the small amplitude reverberation signals, and our correction method allows us to stack for five ScSn and sScSn phases simultaneously to obtain the highest possible data coverage. For the global MTZ discontinuities, we use “adaptive stacking”. Based on Voronoi tessellation, the method automatically adjusts for topography, noise, and data coverage. Regional-scale fixed bin parameterisations of varying sizes are used to search for the intermittent MM reflectors.

We incorporate our seismic observations with mineral physics modelling, inverting for a realistic range of potential temperatures and basalt-harzburgite mixtures to obtain the best-matching thermochemical model for the MTZ. We first compare our new ScS MTZ model with its counterpart generated from SS and PP precursors [Waszek et al., 2021], to benchmark observational differences between data types. We next investigate the link or lack thereof between our MTZ model and detections of MM signals, to place improved constraints on variations in properties with depth. The final step is interpretation of our observations and modelling in the context of geodynamical simulations of mantle convection. Our outputs will contribute to greater understanding of the complex relationship between MTZ discontinuities and MM reflectors, with implications for global mantle circulation, compositional layering beneath the MTZ, and even surface activity. 

 

 

References:

Waszek, Schmerr, Ballmer. 2018.

Garcia, Waszek, Tauzin, Schmerr. 2021.