Predicting A Global Perovskite And Post-Perovskite Phase Boundary

Daoyuan Sun,Don Helmberger,Xiaodong Song,Stephen P. Grand
DOI: https://doi.org/10.1029/174GM12
2007-01-01
Abstract:A lower mantle S-wave triplication with a S-cd branch occurring between S and ScS has been recognized for many years with a variety of explanations. It is particularly strong when sampling regions beneath the circum-Pacific lower mantle fast velocity belt seen in global tomographic models where it has been modeled with a 2-3% jump in S-velocity. General properties are (1) it tends to arrive earlier beneath the fastest anomalies and (2) its rapid changes in strength and timing relative to S. The recently discovered Perovskite (Pv) to Post-perovskite (PPv) phase transition [Murakami et al., 2004] is expected to show this change in timing assuming a positive Clapeyron slope (gamma) between 3 to 9 MPa/K; however, the predicted velocity jump is about half of the above 1 D modeling results. Here we model the phase boundary height by mapping S-wave tomography into temperature assuming uniform chemistry, or a Mono-Phase-Transition (MPT), and a more complex mapping procedure involving possible changes in Chemistry (CPT). A few adjustable parameters involving reference phase boundary height and velocity jump are determined from comparing synthetic seismogram predictions with densely sampled observations. Particularly strong Sed data are explained by focusing effects caused by small zones of enhanced velocities (slab buckling) seen in some tomographic models. These sharp features in S-wave tomography are confirmed by the behavior of differential PKP (AB-DF) which shows 4 s changes over a distance of 300 km correlating well with the corresponding S-wave models beneath Central America. Thus adding 3D propagational effects caused by these structures to the Pv to PPv velocity jump predicted from mineral physics appears to generate results compatible with S-cd waveform observations.
What problem does this paper attempt to address?