4.1 Testing ray-theoretical traveltime corrections
Figure 8 shows the CRP images along the 35°N parallel in the central region of the model domain determined for the PREM and S40RTS models. Supplementary Figure S2 shows similar CRP images for SEMUCB-WM1 and TX2015. The CRP image for PREM in Figure 8a is the ideal case as the assumed velocity structure of the mantle is identical to the structure used to calculate traveltimes and ray paths. Artifacts are entirely due to the implementation of the CRP mapping procedure, the limited frequency band of the waveforms, and wave interference. PREM’s velocity discontinuities at 220, 400, and 670 km depth are resolved about 10 km shallower in the mantle because the crust is not included in the waveform computations. Since the waveforms are computed for periods longer than 10 s and since shear wave speed increases with depth, reflectors at larger depths are more stretched than at shallower depths. The imaged 660-km discontinuity is therefore only about 60% stronger than the imaged 410-km discontinuity even though the impedance contrast at the 660 is a factor of two stronger than at the 410-km discontinuity. The CRP image derived from PREM waveforms is mostly free of artificial layering between 150 km and 750 km depth. The side lobes of the 660 near -65°E are artifacts near the boundaries of the station grid. Layering near 100 km depth, which is especially strong near the center of the CRP image, shows that the Ssds reverberation is not an ideal wave type for imaging the uppermost mantle. The arch-shaped structure below 750 km depth is likely the projection of shallow SS precursors misinterpreted as Ssds reflections deep in the transition zone as discussed in section 2.1.