4.1 Receiver Function Imaging
Our final common conversion point stack produces a NW-SE-trending, approximately trench-perpendicular profile that samples a ~50 km segment of the Alaska subduction forearc up to 80 km deep (Fig. 2). Both the stacked 1.2 Hz (Fig. 2a) and the stacked 2.4 Hz images (Fig. 2c) show a coherent, SE to NW dipping positive conversion at ~ 30-40 km depth consistent with the expected slab Moho depth from previous studies. For reference, we plotted earthquakes from the AACSE catalog (Ruppert et al., 2021a; Ruppert et al., 2021b) beneath the study area (57.40-58.0 N, 152.083-152.75 W) which are within one standard deviation of the mean hypocentral depth of 24.96 km on our CCP images (black dots in Fig. 2b and 2d). We also plotted the top of the slab depth from Hayes et al. (2018) and inferred the slab Moho depth assuming an 8-km thick oceanic crust (blue and red dashed lines in Fig. 2b and 2d). We do not observe a negative top-of-slab conversion above the positive slab Moho conversion.
We observe intermittent segments of shallow (above ~10 km depth) positive conversions across the length of the profile in our high frequency (2.4 Hz) stacked image (Fig. 2d). One such horizon at ~ 5 km depth extends from about ~8-12 km along the profile, and another beneath Kalsin Bay at ~7 km depth extends from 28-35 km along the profile. Since the depths of these early arrivals vary along the line, the arrival times of their multiples are also different. A mixture of these reverberations and other possible primary arrivals could explain the chaotic character of the traces between ~ 5 km and 35 km depths. Increasing the Gaussian value to 10 (~4.8 Hz) did not increase the quality of the final image (Fig. S4).