jats:titleSUMMARY</jats:title>jats:pLarge thrust faults accommodate the convergence between India and Tibet along the southern margin of the Himalaya and have a history of producing great earthquakes that cause widespread damage. Along most parts of the Himalaya, there is geomorphological evidence that these thrusts can rupture to the surface in Mw &gt;8 earthquakes. However, in the Himalayan state of Jammu & Kashmir (NW India), the thrust faults are blind and large-scale folding is the only expression of active deformation at the surface, making it difficult to assess the seismic hazard in this region. In this paper, we use field, satellite, and seismological observations to determine the fault geometry in Jammu & Kashmir. We then estimate the ground motions from potential earthquakes in the region using models of the seismic wavefield that would be generated if the thrust fault beneath Jammu & Kashmir were to rupture. We find that earthquakes that rupture the buried, shallow part of the locked Main Himalayan Thrust could generate peak ground velocities that are &gt;3 times larger than earthquakes of the same magnitude on its deeper portions. We also model the ground motions that would result from the thrust fault geometries representative of different parts of the Himalayan arc. These simulations show that even seemingly minor variations in the shallow fault geometry can lead to large differences in the expected ground motions, highlighting the importance of accurately determining the shallow geometry of thrust faults along the margins of mountain ranges for estimating seismic hazard.</jats:p>