Can sub-slab low-velocity anomalies be an artifact caused by anisotropy? A case study from the Alboran slab area in the western Mediterranean

Although it is well understood that mantle flow in a subduction setting can develop seismic anisotropy, many tomographic studies still assume an isotropic mantle. In this study, we explore how unaccounted-for seismic anisotropy contributes to P-wave anomalies, focusing on apparent sub-slab low velocities. We use data from the westernmost Mediterranean, where the Alboran slab has subducted beneath the Gibraltar Arc. Our isotropic P-wave velocity model shows a low-velocity anomaly at depths of 75–225 km to the west of the high-velocity-anomaly that represents the Alboran slab. We explore whether this low-velocity anomaly could be an anisotropy-induced artifact as a significant shear-wave-splitting (SWS) along the arc is observed. We consider five different hypothetical anisotropy models representing different mantle configurations. The anisotropy models are based on 1) the SWS observations, 2) the depth distribution of the low-velocity anomaly and 3) geodynamic modeling of the development of entrained and toroidal mantle flow in generic subduction zones. Given the underdetermined nature of an anisotropic inversion, we include the anisotropy models as a priori constraints on the tomography. The tomographic results including anisotropy models have smaller travel time residuals and isotropic anomalies in the region of interest while fitting the SWS observation better than the isotropic model. We propose that unaccounted-for seismic anisotropy can produce artifacts in isotropic tomography of real datasets, and this is at least partially the case for the low-velocity anomaly below the Alboran slab. Further, we show that not taking this into account can lead to erroneous conclusions about mantle temperature and the presence of melt. From the insights that our anisotropy models provide, we also suggest that toroidal mantle flow is currently dominant below and around the Alboran slab.