Probing fault frictional properties during afterslip up- and down-dip of the 2017 Mw 7.3 Iran-Iraq earthquake

We use Interferometric Synthetic Aperture Radar (InSAR) data collected by the Sentinel-1 mission to study the co- and postseismic deformation of the 2017 Mw 7.3 earthquake that occurred in NW Zagros along the Iran-Iraq border. We derive the fault geometry of the mainshock in a Bayesian inversion framework, assuming a single rectangular slip patch. We find that the strike of the rupture that yields the highest value of PDF of posterior data fitting is 355 deg., which is at least 20-30° larger than the average strike of the structures in this region, but in good agreement with the focal mechanism solution ( 351 deg.) determined for the east-dipping nodal plane from seismic data. The amount of dextral slip is comparable to the thrust component, although the dip angle of the fault is found to be small ( 16 deg.). Most of the coseismic moment release is found to be at a depth range between 15 and 21 km, well beneath the boundary between the sedimentary cover and underlying basement. Data from all four tracks also reveal robust postseismic deformation during 8 month after the mainshock (from November 2017 to July 2018). Kinematic inversions show that the observed postseismic InSAR LOS displacements are well explained by oblique (thrust + dextral) afterslip both updip and downdip of the coseismic peak slip area. The dip angle of the shallow afterslip fault plane is found to be significantly smaller than that of the coseismic rupture, corresponding to a shallowly dipping detachment located near the base of the sediments. Aftershocks during the same time period exhibit a similar temporal evolution as the InSAR time series, with most of the aftershocks being located within and around the area of maximum surface deformation. We also observe prominent coseismic and postseismic surface deformation along several secondary fold-like structures. The postseismic deformation data are consistent with stress-driven afterslip models, assuming that the afterslip evolution is governed by rate-and-state friction. Under the assumption that the effective normal stress on the fault is hydrostatic and depth dependent, the average (a-b) is estimated to be 7*10^-3, consistent with previous laboratory results and in situ measurements from geodetic data.