The convergence zone between the Pamir and Tien Shan Mountains apparently accommodates between 10 and 50% of the relative motion between India and Eurasia at 073°E latitude. Much of it is spatially focused as fault slip along narrow thrust belts. We examined the Holocence history of deformation along the Main Pamir Thrust (MPT) in the Alai Valley region with two paleoseismic investigations in the area north of Peak Lenin (Komansu\textemdash 39° 28.3' N, 072 ° 37.5' E) and 60 km west at Shivie Su (39° 28.5' N, 071 ° 55.2' E). No strike-slip offset is evident along these ENE-striking fault zones.
At Shivie Su, we exposed medium to coarse-grained fluvial gravels with stage I-II CaCO3 soil development. Both the deposit and terrace are vertically offset 2 m along a 28° S-dipping thrust fault. The footwall is capped by a tan-brown matrix-supported gravel unit that merges to the north with retransported loess that covers the terrace and thickens to 60 cm toward the south where it is overridden by the footwall. We infer that two earthquakes ruptured the surface at this location since the formation of the terrace with 4 m total dip slip.
At Komansu, we exposed a major S-dipping thrust fault and N-facing monocline (main trench) and a minor N-dipping backthrust (south trench). The 5-m high scarp is developed on a terrace surface that is younger than 6 ka. Three wedges of poorly sorted sand and silt with both fine pebbles and medium to coarse gravels are evident in the main trench. These deposits have clear relative age relationships with the southernmost and lowest being oldest. The wedges thicken to the south and the older two are truncated to the south. All three merge northward with the retransported loess cover of the terrace. We infer that each of these deposits represents colluvial deposition from the erosion of the surface rupture of a major thrust earthquake. The earliest event is indicated by a wedge that is truncated by a 30 degree S-dipping fault. The fault dip is horizontal above the second wedge. The youngest wedge is not faulted. The thrust fault in the last 2 m of extent up dip actually dips 10 degrees to the north and indicates that in the last earthquake the rupture displaced the hanging wall downhill over the pre-existing scarp. The decrease in dip forms a monocline in the hanging wall. Offset of a distinctive gravel in the last earthquake was 1.6 m. The shape and size of the three wedges suggests that the events had similar dip slip.
While no events larger than M6 were recorded instrumentally, these data indicate that least 2-3 earthquakes of similar magnitude have ruptured the surface along the MPT in the Holocene. If these events and a slip rate of > 6 mm/yr are representative, a recurrence interval for large earthquakes along the MPT of less than 300 years can be inferred.