Figure 8 . CRP images along the 35°N parallel determined for (a) PREM synthetics, (b) S40RTS synthetics, and (c) S40RTS synthetics after ray-theoretical corrections have been applied. The color scale is the same as Figure 4a.
The CRP image derived from S40RTS waveforms is more complex (Figure 8b). The 410-km and 660-km discontinuities deepen from east to west because S40RTS predicts that Ssds traveltimes through the upper mantle are shorter beneath the central and eastern US than beneath the western US and we use the PREM velocity structure to convert traveltimes to reflector depths. The velocity heterogeneity in S40RTS causes misalignments of Ssds signals and therefore fluctuations in the strength of the 410-km and the 660-km discontinuities from west to east by up to a factor of two. For example, the 660 appears as a relatively weak reflecting boundary between longitudes -120°E and -110°E, near the transition between the low-velocity upper mantle of the western US and the high-velocity upper mantle beneath the central US. In addition, spurious reflectors are particularly strong between -120°E and -100°E, where horizontal gradients in the uppermost mantle are strongest. It is difficult to identify how complex wave propagation produced the complexity in the CRP image, but the CRP image based on USArray waveforms is also most complex for the western US, and a tilted reflective structure in the upper mantle has been observed by SB19 in their data image, albeit with an eastward dip and a greater depth extent.
Figure 8c shows the CRP image based on the S40RTS synthetics after applying ray-theoretical traveltime corrections following the procedure outlined in section 4.3. The traveltime corrections do not remove, and may even amplify the CRP image artifacts for depths shallower than 100 km and deeper than 750 km. More significantly, the ray-theoretical calculations appear to overpredict the contribution of shear-velocity heterogeneity to the Ss410s-S and Ss660s-S traveltime differences. After traveltime corrections, the 410-km and 660-km discontinuities are projected shallower beneath the western US than the central US, opposite to the imaged depths of the 410-km and 660-km discontinuities prior to corrections.
The inaccuracy of ray theory in predicting the shear-wave traveltime perturbations is illustrated further in Figure 9. It shows the estimated depths of the 410-km discontinuity and the thickness of the MTZ based on the 1-D CRP method applied to synthetic waveforms computed for S40RTS. Supplementary Figure S3 shows that we obtain similar results for SEMUCB-WM1 and TX2015. The total variation in the depths of the 410-km discontinuity is about 15–20 km. As expected, the depth of the 410-km discontinuity (Figure 9a) mimics the shear-velocity variations in the upper mantle of S40RTS (Figure 7a) and the S-wave traveltime delay map shown in Figure 5. Variations in the thickness of the MTZ (Figure 9b) of about 10 km are small compared to the depth variations of the 410-km and 660-km discontinuities because shear velocity variations in the MTZ are much weaker than in the uppermost mantle.