A very slow basal layer underlying large-scale low-velocity anomalies in the lower mantle beneath the Pacific: evidence from core phases

A multi-phase analysis using long-period World Wide Standardized Seismograph Network and Canadian Network data has been conducted using core-phases for deep focus events from the southwest Pacific. These include SKS, S2KS, SV_diff, and SP_dKS. The last phase emerges from SKS near 106° and is associated with a P-wave diffracting along the bottom of the mantle. Patterns in S2KS - SKS differential travel times (T_{S2KS - SKS}) correlate with those in SP_dKS - SKS (T_{SPdKS - SKS}). T_{S2KS - SKS} values strongly depend on variations in V_S structure in the lower third of the mantle, whereas T_{SPdKS - SKS} values mainly depend on V_P structure and variations in a thin zone (100 km or less) at the very base of the mantle. Anomalously large T_{S2KS - SKS} and T_{SPdKS - SKS} values (relative to the Preliminary Reference Earth Model (PREM)) are present for Fiji-Tonga and Kermadec events (recorded in North and South America), along with anomalously large SV_diff amplitudes well into the core's shadow. More northerly paths beneath the Pacific to North America for Indonesian and Solomon events display both PREM-like and anomalous times. A model compatible with the observations is presented, and contains a thin very-low-velocity layer at the base of the mantle that underlies the large volumetric lower-mantle low-velocity regions in the southwest Pacific. A low-velocity layer of 20-100 km thickness with reductions of up to 5-10% (relative to PREM) can reproduce T_{SPdKS - SKS} as well as SV_diff amplitudes. Large-scale (more than 1000 km) lower-mantle V_S heterogeneity (2-4%) can explain long-wavelength trends in T_{S2KS - SKS}. The exact thickness and velocity reduction in the basal layer is uncertain, owing to difficulties in resolving whether anomalous structure occurs on the source- and/or receiver-side of wavepaths (at the CMB).