Estimating core-mantle boundary temperature from lateral variations in shear velocity and seismic attenuation

<p>The temperature at the Earth&#8217;s core-mantle boundary (CMB), <em>T</em>_CMB, is a key property for a better understanding of our planet dynamics and evolution, but remains poorly constrained. Because the CMB is a material boundary between silicate rocks and molten iron, <em>T</em>_CMB cannot be deduced directly from a single phase diagram. Estimates from mineral physics extrapolate either the solidus of mantle rocks to large pressure, or the ICB temperature to lower pressure, assuming that the outer core is adiabatic. Here, we propose a new approach to the determination of <em>T</em>_CMB based on the analysis of observed lateral variations in shear-wave seismic velocity, <em>V</em>_S, and attenuation, measured with the quality factor <em>Q</em>_S, in the lowermost mantle. Shear velocity is sensitive to the presence of post-perovskite (pPv), a high pressure phase of bridgmanite that is stable at lowermost mantle conditions. Because the stability field of pPv strongly depends on temperature, the presence of this phase and its impact on <em>V</em>_S provides a constraint on the horizontally averaged temperature in the lowermost mantle. On another hand, seismic attenuation is a thermally activated process, implying that its amplitude depends on temperature. At a given depth, and using an appropriate modeling of <em>Q</em>_S, its deviation from a global average can then give access to the average temperature at this depth. Our approach is based on the knowledge of several parameters, of which the temperature of the phase transition to pPv, <em>T</em>_pPv, appears to be the most sensitive. We then performed a preliminary application using models of <em>V</em>_S and <em>Q</em>_S obtained for beneath the Central America and the Northern Pacific, and found that for <em>T</em>_pPv = 3150 K, the CMB temperature should be in the range 3500-3800 K.</p>