Constraints on the core-mantle boundary topography from P4KP-PcP differential travel times

P4KP-PcP differential travel times are examined to infer the core-mantle boundary (CMB) topography. A total of 362 P4KP-PcP times are measured with a measurement error of 0.5 s. The travel times are corrected for the Earth's hydrostatic ellipticity and mantle heterogeneity using a P-wave tomographic model. Spherical harmonic expansion up to degree 4 is adopted for model parameterization. The P-wave velocity heterogeneity is then estimated in the lowermost 150 km of the mantle to overcome the problem of underestimation of the velocity perturbation at the base of the mantle in the global 3-D P-wave mantle model. Subsequently, the CMB topography is inferred using the residues of the above processes. Since the odd-degree components of the CMB topography are insensitive to the P4KP-PcP times, only the components of degrees 2 and 4 are solved for. The resultant features indicate that the maximum amplitude of the CMB topography does not exceed +/- 2 km, with an uncertainty of less than 0.5 km. A numerical test confirms that the pattern of degree 4 is more reliable with less amplitude recovery. The obtained degree 4 pattern shows an amplitude of less than +/- 1 km and indicates the presence of depressions under the circum-Pacific, the central Pacific, and South Africa.