Active accommodation of plate convergence in Southern Iran: Earthquake locations, triggered aseismic slip, and regional strain rates

Abstract

We present a catalog of interferometric synthetic aperture radar (InSAR) constraints on deformation that occurred during earthquake sequences in southern Iran between 1992 and 2011, and explore the implications on the accommodation of large‐scale continental convergence between Saudi Arabia and Eurasia within the Zagros Mountains. The Zagros Mountains, a salt‐laden fold‐and‐thrust belt involving ~10 km of sedimentary rocks overlying Precambrian basement rocks, have formed as a result of ongoing continental collision since 10–20 Ma that is currently occurring at a rate of ~3 cm/yr. We first demonstrate that there is a biased misfit in earthquake locations in global catalogs that likely results from neglect of 3‐D velocity structure. Previous work involving two M ~ 6 earthquakes with well‐recorded aftershocks has shown that the deformation observed with InSAR may represent triggered slip on faults much shallower than the primary earthquake, which likely occurred within the basement rocks (>10 km depth). We explore the hypothesis that most of the deformation observed with InSAR spanning earthquake sequences is also due to shallow, triggered slip above a deeper earthquake, effectively doubling the moment release for each event. We quantify the effects that this extra moment release would have on the discrepancy between seismically and geodetically constrained moment rates in the region, finding that even with the extra triggered fault slip, significant aseismic deformation during the interseismic period is necessary to fully explain the convergence between Eurasia and Saudi Arabia.

Fault creep rates of the Chaman fault (Afghanistan and Pakistan) inferred from InSAR

“…Each synthetic data set was then inverted by using the same procedure described above, resulting in 50 populations of model estimates for a single across‐fault profile (Figure ). One‐sigma error bounds were defined as the 16th and 84th percentiles of the resulting population of models [e.g., Devlin et al , ; Barnhart et al , ]. Through this approach, reported error bounds quantify variability introduced by the presence of spatially coherent noise in the InSAR time series.…”

Seismotectonics of the Zagros (Iran) From Orogen‐Wide, Calibrated Earthquake Relocations

“…As the trend of the range changes from E‐W to NW‐SE in the central and northwestern Zagros, GPS velocities decrease to ∼4 mm/year relative to central Iran, and convergence obliquity is partitioned into shortening on range‐parallel thrusts and dextral shear along the NW‐SE trending Main Recent Fault (Authemayou et al, ; Talebian & Jackson, ; Figures b and ). Seismic deformation rates calculated from ∼100 years of seismicity account for around half of the geodetic (GPS‐derived) shortening rate in the northwestern Zagros and for less than one third of the geodetic rate in the southeastern Zagros (Barnhart et al, ; Jackson & McKenzie, ; Masson et al, ; Palano et al, ). The shortfall is likely accounted for by a mixture of folding, aseismic fault creep (e.g., Barnhart et al, ), and ductile shortening of the basement (Allen et al, ; Nissen et al, ).…”

“…Seismic deformation rates calculated from ∼100 years of seismicity account for around half of the geodetic (GPS‐derived) shortening rate in the northwestern Zagros and for less than one third of the geodetic rate in the southeastern Zagros (Barnhart et al, ; Jackson & McKenzie, ; Masson et al, ; Palano et al, ). The shortfall is likely accounted for by a mixture of folding, aseismic fault creep (e.g., Barnhart et al, ), and ductile shortening of the basement (Allen et al, ; Nissen et al, ).…”

“…In addition, there has never been a permanent seismic network in the Zagros and the earthquake catalog epicenters used by Berberian () were determined in single‐event analyses using arrival times at distant Global Digital Seismic Network stations. We now know that in the Zagros, epicenters determined in this way—such as those listed in the widely used GCMT, ISC, and NEIC catalogs—may be mislocated by as much as 50–60 km, as revealed by coseismic surface deformation mapped with InSAR (Barnhart et al, ). These mislocations reflect both scatter and bias introduced by seismic velocity perturbations in regions surrounding the Zagros which are not accounted for in standard whole Earth velocity models.…”

Geodetic Constraints on the 2014 M 6.0 South Napa Earthquake

“…To derive a source model and corresponding uncertainties for the South Napa earthquake, we begin by jointly inverting the GPS and downsampled InSAR observations for the location (longitude, latitude, and depth), orientation (strike, dip, and rake), and dimensions (along-strike length and down-dip width) of a single-fault patch with homogeneous slip embedded in a homogeneous elastic half-space using the neighborhood algorithm (Okada, 1992;Sambridge, 1999). We do not use seismic event location information (i.e., hypocenter) to precondition the inversion as earthquakes in other locations globally may be mislocated by tens of kilometers (e.g., Elliott et al, 2010;Devlin et al, 2012;Barnhart et al, 2013). Event moment tensors variably indicate the ruptured fault dipped steeply to the west or east (http://earthquake.usgs.gov/earthquakes/ eventpage/nc72282711#scientific_moment-tensor; last accessed December 2014), so we allow the neighborhood algorithm to sample both dip domains, searching over a broad range of strike, dip, and rake within each region of model space.…”