A strong seismic velocity contrast in the shallow mantle across the Zagros collision zone (Iran)

Surface wave dispersion measurements are interpreted jointly with the inversion of teleseismic P-wave traveltime residuals along a dense 620-km long temporary seismic profile across the Zagros to investigate its upper-mantle structure. The S-wave model determined from Rayleigh wave dispersion in the Zagros fold and thrust belt has high velocities from 4.5 ± 0.2 km s−1 below the Moho to 4.9 ± 0.25 km s−1 at 200 km depth, which are comparable to a shield-like structure. Beneath the suture region from the Main Zagros Thrust (MZT) to the Urumieh-Dokhtar volcanic arc, S-wave velocities are lower than beneath the Zagros in the top 50 km of the upper mantle, with a minimum of 4.4 ± 0.2 km s−1 at 80 km depth. From 150 km and deeper, S velocities are as high as beneath the Zagros. We suggest that part of the velocity difference at shallow depth is due to higher mantle temperatures and/or higher fluid content beneath the northern half of the profile, but that velocities are too high to support the hypothesis of mantle lid delamination under the suture zone. Teleseismic P traveltime relative residuals display a long-wavelength variation along the transect, with a difference of 1.1 s between negative residuals in the Zagros Simple Folded Belt and positive residuals in Central Iran. This difference backprojects into a 6–7 per cent lateral variation of P-wave velocity in the shallow upper mantle, with higher VP beneath Zagros and lower VP beneath Central Iran. The main short wavelength variation of the residual is located in the suture region, with late P arrivals in the region of the MZT and early arrivals in the Sanandaj-Sirjan zone (SSZ). Using synthetic models of VP perturbations, we show that the high velocities of the Arabian platform have to extend laterally at least to the SSZ to fit the observed P delays. This model also predicts Rayleigh wave phase velocities, which are within the error bars of the observed dispersion. It supports the model of crustal-scale overthrusting at the MZT.