Paleoearthquake History of the Cherchen He Reach of the Central Altyn Tagh Fault Xinjiang, China.

Although the Altyn Tagh Fault (ATF), Xinjiang, China, is a first-order structure within the Indo-Asian collision zone, its Holocene paleoseismic and slip history remain contested. Studies of offset fluvial terraces and other landforms indicate slip rates of 9-27mm/yr over millennial timescales. However, little is known about the record of paleoearthquakes along the ATF. Defining a paleoearthquake record for the last few millennia complements slip rate studies spanning 5-10 kyr as well as fault-behavior studies. A classification of slip rate over different time scales is vital to understanding the role such structures play in the large-scale tectonics of the Indo-Asian collision. We mapped ~20km of this reach of the ATF (N 37.43°, E 86.40°). Geomorphic and fault trace mapping (1:3,800) revealed a continuous fault morphology. This reach of the fault is characterized by north-facing scarps ~0.5-2m high superimposed on ~50m high south-facing escarpments. The north- side up component of slip may be associated with ~10 right steps of 10-50m along the fault trace. Along with mapping, measurements of drainages offset between 5 and 30m provide information about slip along this reach for the last few earthquakes. The smallest of these offsets were consistently 5-8m. We infer they result from the most recent earthquake. Two 15m-long and ~2m deep trenches, sites 123 and 105, were excavated across narrow fault traces where fan deposits from south-draining watersheds abut the 0.5m-high fault scarps. These scarps dam the south-draining watersheds, which were offset by the fault and form depressions in which younger sediment was deposited. Trench 123 was logged at a scale of 1:10 and trench 105, 5 km to the west, was logged at a scale of 1:12.5. The stratigraphy in both trenches revealed interfingering alluvial deposits and eolian fine silts and loess. Offset units, unconformities, onlap and drape sequences, and fissures clearly define the fault traces and paleoearthquakes within the trenches. Both sites preserve evidence for 2-3 events. We also collected nine organic samples for radiocarbon dating at site 105 from individual stratigraphic layers within the trench. We expect that ongoing 14C analyses on these samples will bracket the ages of these events. Because of the consistency of 2-3 events at each of the paleoseismic sites, we conclude that their composite record reveals 2-3 late Holocene earthquakes. A clear result of this mapping is that the most recent ATF earthquake along this reach resulted in 5-8m of surface offset (consistent with Mw 7.2). These results have important implications for both slip-rate studies and studies of strike-slip fault behavior. For example, if 5m is representative of slip in each earthquake along this fault reach, earthquake recurrence intervals of 555 years are expected with a 9mm/yr slip rate, whereas 185 years is expected if the slip rate is 27mm/yr. Paleoseismic investigations such as this can distinguish between this factor of three difference in event recurrence. A recurrence interval will help also in classifying which model of fault-behavior is appropriate for the ATF.