Seismicity and structure of Nazca Plate subduction zone in southern Peru

We define subducting plate geometries in the Nazca subduction zone by (re)locating intra-slab earthquakes in southern Peru (14-18 degrees S) and using previously published converted phase analysis results to clarify the slab geometry and inferred relationships to the seismicity. We also provide both P- and S-wave velocity models of the subducting Nazca Plate and mantle layer above the slab using double-difference tomography to understand upper-plate volcanism and subduction zone process. The double-difference constraints for determining the hypocenters and velocity model ensure high accuracy of the relative location of earthquakes with respect to velocity structure. The relocated seismicity shows a smooth contortion in the slab-dip transition zone for similar to 400 km between the shallow (30 degrees)-to-flat dipping interface to the northwest and the 30 degrees-dipping interface to the southeast. We find a significant slab-dip difference (up to 10 degrees) between our results and previous slab models along the profile region sampling the horizontal slab at a depth of similar to 85-95 km. Robust features in both P- and S-wave tomography inversions are both arc-normal and along-arc velocity variations. In the arc-normal direction, all profile results show that the slab velocities beneath the forearc (down to a depth of similar to 100 km) transition to higher velocities beneath the backarc (at similar to 100-140 km depth). In the along-arc direction, velocities of the subducting flat slab are shown to be depressed compared to those of the normal-dip slab. In addition, high shear-wave velocities and low Vp/Vs are detected in the mantle layer above the flat slab, indicating its cold and dry environment. Such differences in the velocity structures for the slab and mantle wedge between the two regions may indicate different hydration states, which greatly affects the upper-plate structure of southern Peru. (C) 2018 Elsevier B.V. All rights reserved.