THE DENSITY AND SHEAR VELOCITY CONTRAST AT THE INNER CORE BOUNDARY

A systematic search of short‐period GDSN seismograms from 1980 to 1984 at ranges from 20d̀ to 90d̀ identifies two probable PKiKP arrivals. PKiKP/PcP amplitude ratios for these phases are consistent with previous studies. However, more typically PKiKP is not observed, even when clear PcP arrivals are seen. We use these data to place upper bounds on PKiKP/PcP amplitude ratios for 100 event‐station pairs. These bounds indicate that most measurements of PKiKP amplitudes are biased toward large values and predict reflection coefficients at the inner core boundary (ICB) which are too high. Our upper limits on PKiKP amplitudes roughly constrain the density jump at the ICB to be less than 1.0 g cm−3 and the shear velocity at the top of the inner core to be greater than 2.5 km s−1, assuming a sharp discontinuity at the ICB. Upper bounds on PKiKP/P amplitude ratios at ranges between 70d̀ and 90d̀ are consistent with these results but are less reliable due to take‐off angle differences between P and PKiKP. Approximately 50 observed free oscillations of the Earth are sensitive to the structure of the inner core. Modern models derived from these and other mode data typically have a density jump at the ICB of 0.5–0.6 g cm−3. An experiment in which we varied the mean density of the inner core indicates that the mode frequencies are roughly linear functional of this parameter. The fit to the data is seriously degraded if the density jump is significantly different from 0.55 g cm−3. Many of the modes are also strongly sensitive to the shear velocity in the inner core, and forward modelling indicates that the average inner‐core shear velocity is probably 3.45 ± 0.1 km s−1. These results are compatible with the short‐period PKiKP amplitude bounds, indicating that there is no inconsistency between PKiKP and normal mode data regarding the density and shear velocity structure at the inner core boundary. Copyright © 1990, Wiley Blackwell. All rights reserved