Figure 2 . (a) Epicenters of earthquakes (stars) used in this study. The dashed circles have a common center of [40°N, -95°E] and radii of 30, 60, 90, 120, and 150 degrees. (b) Red dots show reflecting points of Ss410s at the 410-km discontinuity for the 59,517 seismograms in our data set.
A record section of the sum of these waveforms brings out Ss410s and Ss660s as the strongest mantle reflections (Figure 3). The Ss410s and Ss660s have mean amplitudes of about 0.05 and are recorded without interference with ScS and SS at distances larger than 60° and 75°, respectively. The SS precursors S410S and S660S are weaker than Ss410s and Ss660s at distances smaller than 110° (e.g., Shearer, 1991). Although we cannot rule out that signals labeled ‘A’ and ‘B’ in Figure 3a are side lobes due to the applied Butterworth filter, signal ‘A’ may be a SS-precursor reflection at a depth of about 125 km and signal ‘B’ may be a Ssds reflection from the lithosphere-asthenosphere boundary (LAB), a boundary that has also been studied with P-wave and S-wave receiver functions (Rychert et al. 2007; Abt et al. 2010; Hopper & Fischer, 2018) and multiple S-wave reflection (Liu & Shearer, 2021).
Ssds reflections from the uppermost lower mantle below the 660 arrive more than 250 s after S and do not interfere with SS and the S410S precursor at distances larger than 95° in region ‘C’ of Figure 3a. However, it is difficult to differentiate reflections below the 660 from shallower SS precursors because S waves are attenuated by diffraction around the core and the slowness resolution is poor. The high-amplitude signal about 330 s after S waves has a slowness of roughly 1.0 sec/degree which is smaller than the slowness of any SS precursor. Its traveltime is similar to that of the phase Ss410s410s (i.e., the shear-wave reverberation with two up-and-down shear-wave segments between the surface and the 410-km discontinuity) and the phase PSs660s (i.e., the PS phase with an additional top-side reflection off the 660-km discontinuity). However, it is unlikely that these phases can be recorded with high amplitudes on transverse component records.