Inner core boundary topography explored with reflected and diffracted P waves

The existence of topography of the inner core boundary (ICB) can affect the amplitude, phase, and coda of body waves incident on the inner core. By applying pseudospectral and boundary element methods to synthesize compressional waves interacting with the ICB, these effects are predicted and compared with waveform observations in pre-critical, critical, post-critical, and diffraction ranges of the PKiKP wave reflected from the ICB. These data sample overlapping regions of the inner core beneath the circum-Pacific belt and the Eurasian, North American, and Australian continents, but exclude large areas beneath the Pacific and Indian Oceans and the poles. In the pre-critical range, PKiKP waveforms require an upper bound of 2 km at 1-20 km wavelength for any ICB topography. Higher topography sharply reduces PKiKP amplitude and produces time-extended coda not observed in PKiKP waveforms. The existence of topography of this scale smooths over minima and zeros in the pre-critical ICB reflection coefficient predicted from standard earth models. In the range surrounding critical incidence (108-130°), this upper bound of topography does not strongly affect the amplitude and waveform behavior of PKIKP + PKiKP at 1.5 Hz, which is relatively insensitive to 10-20 km wavelength topography height approaching 5 km. These data, however, have a strong overlap in the regions of the ICB sampled by pre-critical PKiKP that require a 2 km upper bound to topography height. In the diffracted range (>152°), topography as high as 5 km attenuates the peak amplitudes of PKIKP and PKPCdiff by similar amounts, leaving the PKPCdiff/PKIKP amplitude ratio unchanged from that predicted by a smooth ICB. The observed decay of PKPCdiff into the inner core shadow and the PKIKP-PKPCdiff differential travel time are consistent with a flattening of the outer core P velocity gradient near the ICB and iron enrichment at the bottom of the outer core.