Slip rate gradients along parallel strands of the eastern Altyn Tagh fault confirm modeled rupture behavior at a transpressional bend

Abstract:

Characterized in sufficient spatial proximity, Holocene fault slip rates may be used to test numerical model predictions of recurrent rupture behavior, including whether fault geometry produces a persistent control on rupture length. We document slip rate gradients along two subparallel strands of the Altyn Tagh fault (ATF) in western China as they pass through the 50-km-long, 15-20° transpressional double-bend at Aksay (~93°E). In the middle of the bend the southern fault (SATF) attains a strike 20° clockwise from regional fault strike, a bend angle that prior model results predict will inhibit throughgoing propagation of rupture. The northern fault (NATF) reaches only 15° clockwise from regional strike, at which angle model results indicate rupture may continue unimpeded. The sum of NATF and SATF slip rates at a given longitude gives the full sinistral slip rate of the ATF. Our mapping, optically stimulated luminescence (OSL), Be-10, and C-14 geochronology, and paleoseismic results show that ruptures have not propagated through the bend on the SATF during at least the past 5-6 earthquake cycles, including the most recent surface rupture, which we have mapped and measured in detail. We map an abrupt west-to-east decline in mid-Holocene slip rate on the SATF from 5 ± 2 mm/yr west of the bend to 0 mm/yr in the middle of the bend. As the SATF straightens out again to the east, slip has largely been transferred to the NATF; a >10.7 ± 0.8 ka alluvial fan on the east side of the bend is offset merely 9.1 ± 1.2 m by the SATF, giving a Holocene slip rate of 0.9 ± 0.3 mm/yr and representing possibly only 2-3 Holocene earthquakes. On the NATF, we document a more gradual decline in slip rate through the bend, from 6.5 ± 2.5 mm/yr in the east to 2-3 mm/yr on the west. Fault morphology and the absence of an abrupt drop in slip rate on the NATF stand in stark contrast to the SATF and indicate that earthquakes on the NATF may span the bend. Our measured gradients in slip rate match the distribution of cumulative slip predicted by numerical dynamic-rupture models of this fault system. Correspondence of our field observations with the results of numerical models indicates that physics-based rupture simulations with realistic fault geometry may provide reliable predictions of plausible rupture extents and recurrence intervals.