New global constraints on transition-zone topography from normal-mode tomography

Lateral variations in the depths of the transition-zone discontinuities are generally attributed to variations in temperature, causing local changes in the depth of the dominant phase transition. At moderate temperatures the dominant phase transitions are those of olivine, characterized by a positive Clapeyron slope (dP/dT) at 410 km depth and a negative Clapeyron slope at 660 km depth. An anticorrelation between topography on the 410 and 660-km discontinuities is therefore expected in the absence of variations in chemical composition, as an increase in temperature would lower the 410-km discontinuity and elevate the 660-km discontinuity. Simultaneously, this temperature increase would result in a decrease in seismic velocity and density of the mantle material. Comparing models of transition-zone topography, seismic velocity and density therefore gives valuable insight into the nature of transition-zone discontinuities. Existing global models of transition-zone topography have been created using SS and PP precursor measurements, which need to be corrected for mantle velocity structure using an independent velocity model before the discontinuity depths can be calculated. Here, we present new global models of transition-zone topography and whole-mantle S-wave velocity, P-wave velocity and density that have been simultaneously inferred from a different type of seismic data: Earth’s normal modes. Normal modes are whole-Earth oscillations induced by large earthquakes (M_w≥7.5). We use our models, which can be readily compared to one another, to analyze the nature of the transition-zone discontinuities. We also discuss the trade-offs between the different model parameters and the model uncertainties, the latter of which is additional information provided by the Hamiltonian Monte Carlo method used for our inversion. Finally, we compare our models to transition-zone topography obtained from SS precursor data.