BUILDING A LATEST QUATERNARY SLIP HISTORY FOR THE CENTRAL ALTYN TAGH FAULT: IMPLICATIONS FOR DEEP-TIME PALEOSEISMOLOGY

The active, left-slip Altyn Tagh fault (ATF) defines the northwestern margin of the Tibetan plateau and is arguably the most important single structure accommodating Indo-Asian convergence north of the Himalayas because of its total offset (>450 km), plate-boundary length (>1500 km), and longevity (>30 Ma). Because the ATF plays a central role in competing models of deformation for this archetypal continental collision, its latest Quaternary slip rate has been disputed for over two decades. To resolve this debate, we have determined the post-16 ka slip history of the central ATF between 86.5° and 88.5°E longitude using morphochronology (measurements of dated and displaced landforms). We started by evaluating the uncertainties in slip rates introduced by epistemic uncertainties in morphochronologic data, such as uncertainty in the magnitude of lateral erosion of a displaced fluvial terrace riser or its age. We then obtained new morphochronologic data from 9 risers at 5 slip-rate sites: Kelutelage, Tuzidun, Yukuang, Keke Qiapu, and Yuemake. To analyze this volume of morphochronologic data, we developed a Monte Carlo approach for determining slip histories from suites of morphochronologic data that allows us to determine both precise average slip rates and evaluate the extent to which such rates have varied over time. Application of this model to all robust morphochronologic data for the central ATF yields a preliminary average rate of 9.1 ± 1.1 mm/yr from 16.6 ± 3.9 ka to present. This approach also reveals an apparent pulse of accelerated strain release in the mid Holocene that is 3 times greater than the average rate, which we interpret to represent a cluster of 2-6, >Mw 7.2 earthquakes during an 800 year period. This is the first possible earthquake-cluster detected using morphochronologic techniques. Our method for quantifying the extent to which fault slip varies over time is creating new opportunities in deep-time paleoseismology because the records of fault-slip preserved by faulted landforms are generally longer than those typically accessible via trenching studies. In particular, this approach is essential for testing the morphing block model of continental tectonics, in which frequent reorganization of the active fault network leads to secular variation in slip along faults at timescales of 10-100 kyr.