Illuminating Structural Variability of the East Aleutian Plate Interface From Local and Teleseismic P-to-S Converted Waves

The upper few kilometers of subducting plates govern subduction zones. Direct seismic imaging shows a variety of features there depending on the imaging method used. We compare P-to-S mode conversions from the subducting plate interface from teleseismic earthquakes to those from local earthquakes. Seismic signals are examined from 33 seismometers in the Alaska Peninsula region, including several permanent and temporary stations. The timing of the P-to-S conversions from local arrivals (local PS; 1–20 Hz signals) convert off an interface on average 8 km shallower than the equivalent teleseismic mode conversion, (teleseismic Ps; 0.05–1 Hz signals). The consistent positive teleseismic conversion is interpreted as the oceanic Moho, placing the local-PS conversions near the slab surface. Using 2D full waveform modeling, we find that models with a thin low-velocity layer at the top of the subducting plate containing elongated scatterers are required to reconcile observations at both frequencies. Scattering at high frequencies can be explained by anastomosing shear zones juxtaposing highly variable lithologies along foliation that parallels the plate surface. Local PS timing shows that a low-velocity layer exists across most of the array with velocities increasing as depth increases, in a manner resembling progressive metamorphism; the base of this layer generates the most prominent teleseismic Ps conversion. What constitutes this layer changes with depth; the layer is likely the metamorphosing oceanic crust at >40 km depth but shallower is better explained as a shear zone comprised of fluids, hydrous minerals, and/or metasediment.