Author: Austin John Elliott

Publisher:

Published: 2014

Total Pages:

ISBN-13: 9781321608465

Geometric complexities such as bends and stepovers along strike-slip faults impact the propagation of earthquake ruptures and can control the ultimate sizes of earthquakes. The ability of a rupture to propagate through a geometric complexity constitutes a fundamental predictor of seismic hazard, as the resulting length of a seismic fault rupture dictates the extent, intensity, and duration of damaging ground motion. Simulations of individual ruptures along a simple fault system indicate that bends of sufficient length or angle halt earthquake ruptures, yet simulations of rupture over multiple seismic cycles reveal that specific local geometry and the history of prior ruptures further modulate this behavior. Thus, assessing the proportion of ruptures that terminate at versus propagate through a geometric complexity requires specific geologic observations of fault geometry and seismic history. To investigate to what extent geometry alone controls rupture length, and validate the predictions of numerical models with observational data, I investigate the geomorphic record of multiple Quaternary earthquake cycles at the Aksay restraining double-bend on the Altyn Tagh fault (ATF) in western China. At the Aksay bend two overlapping subparallel strike-slip faults (the northern--NATF--and southern--SATF--Altyn Tagh faults) permit testing of model predictions for different fault bend angles. First I document the size and extent of the most recent earthquake (MRE) along the SATF, mapping 95 km of continuous fresh rupture as well as 70 measurements of small offsets that represent average coseismic slip of 5.6 m. Importantly, I constrain the eastward extent of this MRE and several before it at the most highly misoriented reach of the Aksay bend. Through Beryllium-10 exposure age dating of an undeformed Pleistocene alluvial deposit covering the fault, I demonstrate that no other Quaternary ruptures of the SATF have propagated farther through the bend than the MRE. Together with 270 km of fresh rupture previously mapped to the west, this minimum rupture length of 95 km, and average slip of 5.6 m, indicate a large magnitude M(w)>7.8) for this event. I measure Quaternary slip-rates at four locations spanning the bend on each of the two faults, in order to assess, using accumulated slip, how frequently and where prior ruptures have terminated within the bend. I present a new geomorphic interpretation of the controversial Huermo Bulak He slip rate site on the eastern NATF, at which prior studies reported contradictory slip rates based on conflicting mapping. The rate I determine of 6.3 (+2.1)/(-1.6) mm/yr−1 is substantially lower than some earlier estimates at this site, but agrees with rates determined here from both geodetic modeling and older offset geomorphic markers. At this site and the others I employ optically stimulated luminescence (OSL) burial-age dating of surface-capping loess deposits to interpret abandonment ages of geomorphic surfaces. Using cross-cutting relationships to interpret geomorphic history of deposition and incision at these sites, I relate these surface ages to offset piercing lines to obtain time-averaged slip rates. The resulting distribution of slip rates on each fault define opposing gradients on the west side of the Aksay bend, ranging from 6.3 (+2.1)/(-1.6) mm/yr−1 in the east to 2.1 ± 0.7 mm/yr−1 in the west on the NATF over a 150 km length of fault, but declining abruptly within 50 km on the SATF from 4.1 ± 0.4 mm/yr−1 in the west to effectively zero in the middle of the bend, with only a fraction of the fault-zone slip rate accommodated locally in the east (0.8 ± 0.3 mm/yr−1). This distribution of slip rates indicates that ruptures repeatedly stop at the bend on the SATF, but propagate through on the NATF. These slip gradients reveal persistence of a geometric barrier along the SATF through multiple earthquake cycles, and suggest the absence of a barrier on the NATF. These observed slip rates agree well with the synthetic slip rate distributions derived from numerical models of multiple rupture cycles along the Aksay bend fault system, validating the physics-based behavior in the models. These models, developed by collaborators in parallel with this observational study, provide the extents and distributions of individual earthquake ruptures that sum to produce the long-term slip rates, presenting the ensemble of possible ruptures that geology alone cannot distinguish. Together, the observational results presented here and the corresponding model results indicate that the vast majority of large ruptures halt along the most highly misoriented reach of the SATF, but that the less misoriented NATF remains favorable for occasional rupture. These results demonstrate that numerical modeling, tuned by field observations, may offer probabilistic estimates of the proportion of ruptures that violate expected barriers to propagation and thus generate larger, more damaging earthquakes.