Investigating the source of anomalous PKP travel-times on south-Sandwich to Alaska paths

Inner core anisotropy was proposed thirty years ago to explain differences in travel times of inner core phases (PKIKP) on polar and equatorial paths (Morelli et al., 1986). Over time, models of inner core anisotropy have become very complex, with evidence for depth dependence, hemispherical variations, and other localised features. Some models propose the strength of anisotropy to be in excess of 4% in the western hemisphere of the inner core. This is difficult to reconcile with predictions from mineral physics and dynamical models of inner core growth. The strong anisotropy is confined to anomalous paths between earthquakes in the south Sandwich Islands and stations in Alaska. In contrast, the strength of inner core anisotropy obtained from measurements of PKPPKP travel times on polar paths does not exceed 1-2% (Bréger et al., 2000; Frost and Romanowicz, 2017). We re-examine the trends of PKIKP travel times on polar paths, in order to reconcile the different measurements and to determine whether discrepancies can be explained by structure in the mantle, the outer core, or localized strong anisotropy in the inner core. For this, we combine existing and new measurements, taking advantage of recent deployments of broadband arrays in Alaska and Antarctica.