The coseismic surface ruptures and deformation field that we have measured from the 2015 M7.2 Murghob earthquake in Tajikistan reveal that it occurred along the least-recently ruptured portion of the left-lateral Karakul-Sarez fault (KSF). Analysis of the displacement field, detailed geomorphic mapping of the fault system, and seismological comparison of the 2015 event to the similar 1911 Sarez earthquake reveal a sequence of earthquake ruptures occurring along different reaches of this fault system. The apparent boundaries between these reaches coincide with thrust faults and suture zones, suggesting a structural control on the extent of modern earthquakes.

We derive the horizontal displacement field in the 2015 earthquake from Sentinel-1 interferograms and Landsat-8 pixel tracking. We invert for slip on a fault whose detailed geometry is tightly constrained by the near-field discontinuity in the Landsat-based displacement map, combined with geomorphic mapping on 1.5-meter resolution SPOT imagery. We produce a detailed map of surface ruptures visible in post-event high resolution imagery (and not visible in pre-event SPOT imagery), the extent of which agrees well with the distribution of shallow slip in the geodetic inversion. Notably, the northern third of the rupture appears to be blind, with ~1-2m of far-field slip but no surface rupture. This northward deepening of slip coincides with a steepening of the rake, which is evident in both the coseismic inversion as well as in the larger scale geomorphology of the region. To the northeast of the 2015 rupture, the KSF exhibits moletracks, small offsets, and general morphology that predate 2015 but suggest a relatively recent earthquake.

To investigate whether this northern reach could have been associated with the 1911 left-lateral M7.2 Sarez earthquake, we reevaluated the misfit of seismic waveforms of that earthquake to synthetic waveforms from a variety of geomorphically plausible sources in the region. We find that a source to the north of the 2015 event gives the poorest residual fit to the 1911 seismic observations, while sources to the southwest and west of the 2015 rupture fit the observed seismograms better. This results remains in agreement with earlier studies that place the 1911 epicenter westward of 2015. Thus we have evidence of rupture on three separate portions of this fault system, indicating a likely interacting sequence of earthquakes.

Based on our coseismic slip distribution and geomorphic interpretation of the fault system from high resolution imagery, we suggest that rupture extents along the Karakul-Sarez Fault are in part controlled by structural segmentation due to intersections with Mesozoic to Cenozoic suture zones within the Pamirs.