Item

Abstract

We apply a variety of statistical and machine learning techniques to elucidate the key 1-D and 3-D features of Earth’s lowermost mantle as imaged by seismic tomography. Multiple observables are required to break the trade-off between temperature, composition and mineral phase changes. Using bulk and shear wave speeds we can distinguish lateral and vertical variations in silica content and place a broad constraint on the temperature. Including density allows us to further resolve variations in iron content and place a tighter constraint on the temperature. Globally, the base of the mantle appears to be enriched in SiO2 compared to the overlying mantle. We demonstrate that the large low shear velocity provinces (LLSVPs) are more likely to be enriched in post-perovskite than their surroundings, and that dense piles at the base of the LLSVPs are likely enriched in both iron and silica.