Overriding plate, mantle wedge, slab, and subslab contributions to seismic anisotropy beneath the northern Central Andean Plateau

Abstract

The Central Andean Plateau, the second‐highest plateau on Earth, overlies the subduction of the Nazca Plate beneath the central portion of South America. The origin of the high topography remains poorly understood, and this puzzle is intimately tied to unanswered questions about processes in the upper mantle, including possible removal of the overriding plate lithosphere and interaction with the flow field that results from the driving forces associated with subduction. Observations of seismic anisotropy can provide important constraints on mantle flow geometry in subduction systems. The interpretation of seismic anisotropy measurements in subduction settings can be challenging, however, because different parts of the subduction system may contribute, including the overriding plate, the mantle wedge above the slab, the slab itself, and the deep upper mantle beneath the slab. Here we present measurements of shear wave splitting for core phases (SKS, SKKS, PKS, and sSKS), local S, and source‐side teleseismic S phases that sample the upper mantle beneath southern Peru and northern Bolivia, relying mostly on data from the CAUGHT experiment. We find evidence for seismic anisotropy within most portions of the subduction system, although the overriding plate itself likely makes only a small contribution to the observed delay times. Average fast orientations generally trend roughly trench‐parallel to trench‐oblique, contradicting predictions from the simplest two‐dimensional flow models and olivine fabric scenarios. Our measurements suggest complex, layered anisotropy beneath the northern portion of the Central Andean Plateau, with significant departures from a two‐dimensional mantle flow regime.

Effects of change in slab geometry on the mantle flow and slab fabric in Southern Peru

“…The ambient conditions in the mantle wedge beneath the active volcanic arc with elevated temperatures and high water content are favorable for the occurrence of C‐type olivine [ Karato et al , ], which would be the best candidate to explain the observed pattern. This scenario would be consistent with trench‐parallel flow in the back‐arc mantle wedge further to the south recently inferred from the local splitting [ Long et al , ].…”

“…The model for upper mantle anisotropy obtained in this study provides constraints complementary to those provided by shear wave splitting, which has previously been studied in this region by Eakin et al [, , ] and Long et al []. Here we briefly discuss the comparison between the two types of studies, illustrated in Figure .…”

“…The model for upper mantle anisotropy obtained in this study provides constraints complementary to those provided by shear wave splitting, which has previously been studied in this region by Eakin et al [2014Eakin et al [ , 2015Eakin et al [ , 2016 and Long et al [2016].…”

“…Comparison between shear wave splitting and Rayleigh wave azimuthal anisotropy. Shear wave splitting measurements are from Eakin et al [], Eakin et al [], Eakin et al [], and Long et al []. (a) Local shear wave splitting (green bars) and 40 s. The measurements are plotted at the event location.…”

“…For comparison, we overlaid our anisotropic phase velocity map at 40 s period, with peak sensitivity at~50 km depth (Figure 2). Figure 8b shows SKS splitting measurements from Eakin et al [2015] and Long et al [2016], along with our 91 s period map, with peak sensitivity at~140 km depth (Figure 2). Sourceside measurements from Eakin et al [2016], which mainly reflect anisotropy in the slab and subslab mantle, are shown in Figure 8c, along with the 100 s phase velocity map; peak sensitivity at this period is~145 km.…”