3.2 Ray-theoretical corrections
Since the tomographically predicted S-wave traveltime variation of 5–6 seconds across the USArray is a significant fraction of the recorded traveltime variation, we suspect that shear-velocity variations in the upper mantle influence the CRP imaging and our estimate of the depth of the 410-km discontinuity. To quantify this, we determine the CRP image for “corrected” Ssds-S difference times. From the measured Ssds-S difference time, we subtract the predicted difference time anomaly (positive or negative) by shifting the waveform around the theoretical Ssds arrival time. We predict the Ssds-S difference time by ray tracing through tomographic model S40RTS for the mantle and CRUST1.0 for the crust. The Ssds-S difference time predictions are similar for SEMUCB-WM1 and TX2015, as expected from Figure 5.
Figure 6a shows the CRP image along the 40°N parallel after traveltime corrections. The character of the corrected (Figure 6) and uncorrected (Figure 4) CRP images are the same, but the 410-km and 660-km discontinuities are flatter boundaries across the USArray. This is especially clear for the region between -100°E and -80°E where the 410-km and 660-km discontinuities are relatively simple. Figure 6b emphasizes that the depth variation of the 410-km discontinuity is much smaller when the CRP image is based on tomographically corrected Ssds-S difference times. The 410-km discontinuity is not deeper in the west than in the east. The thickness of the MTZ in the corrected and uncorrected images are similar because shear-velocity variations are relatively weak in the MTZ compared to the uppermost mantle. The histograms shown in Figures 6c illustrate that the depth variation of the 410-km discontinuity is about a factor of two smaller when traveltime corrections have been applied to the data and that the corrections do not change the range in MTZ thickness values. The traveltime corrections change the mean depth of the 410-km and 660-km discontinuities by about 10 km, which is similar to the change obtained by Shearer & Buehler (2019) using ray-theoretical corrections computed for a regional 3-D velocity model.
The ray-theoretically corrected images may not reflect the actual depth variations of the 410-km discontinuity. Since S40RTS and any other tomographic model does not perfectly explain the recorded S-wave traveltime variation (see Figure 3), it is unlikely that the traveltime corrections have completely removed the effects of the shear-velocity structure on the CRP image. Further, we show in the next section that ray-theoretical predictions of long-period Ssds-S traveltimes are inaccurate and that corrections can project as spurious signals in the CRP images.