A P wave velocity model of Earth's core

Abstract

Present Earth core models derived from the retrieval of global Earth structure are based on absolute travel times, mostly from the International Seismological Centre (ISC), and/or free‐oscillation eigenfrequencies. Many core phase data are left out of these constructions, e.g., PKP differential travel times, amplitude ratios, and waveforms. This study is an attempt to utilize this additional information to construct a model of core P wave velocity which is consistent with the different types of core phase data available. In conjunction with our waveform modeling we used 150 differential time measurements and 87 amplitude ratio measurements, which were the highest‐quality observations chosen from a large population of Global Digital Seismograph Network (GDSN) records. As a result of fitting these various data sets, a one‐dimensional P wave velocity model of the core, PREM2, is proposed. This model, modified from the Preliminary Reference Earth Model (PREM) (Dziewonski and Anderson, 1981), shows a better fit to the combined data set than any of the existing core models. Major features of the model include a sharp velocity discontinuity at the inner core boundary (ICB), with a large jump (0.78 km/s), and a low velocity gradient at the base of the fluid core. The velocity is nearly constant over the lower 100 km of the outer core. The model features a depth‐dependent Qα structure in the inner core such that a constant t* for the inner core fits the amplitude ratios and waveforms of short‐period waves moderately well. This means the top of the inner core is more attenuating than the deeper part of the inner core. In addition, the P velocity in the lowermost mantle is reduced from that of PREM as a baseline adjustment for the observed separations of the DF and AB branches of PKP at large distances.