Spatial-temporal organization of large earthquakes along continental fault systems

The India-Eurasia collision zone represents one of the largest regions of continental deformation on Earth. Strain is partitioned between crustal shortening, accommodated by thrust faults and folds, and lateral extrusion, taken up by large strike-slip faults. To date, the kinematics of major strike-slip faults are relatively well constrained by geodetic and geologic measurements. It is also known from instrumental records that these faults can sustain large earthquakes. Yet, how earthquakes accommodate deformation along these faults remains fairly unknown. Hence, the main objective of this thesis is to investigate the long-term organization of past earthquakes along intracontinental strike-slip fault systems. To address this objective, I adopt a paleoseismological approach to test two different strike-slip fault configurations across the Indio-Eurasia collision zone. The first case study focuses on the Altyn Tagh Fault (ATF), located on the northern margin of the Tibetan Plateau; this fault is the longest active strike-slip fault system in Asia and is characterized by a moderate to fast slip rate (~10 mm/yr). The second case study focuses on the Bulnay Fault System (BFS) in northwestern Mongolia, which accommodates the northernmost deformation related to the India-Asia collision and is characterized by slow-moderate slip rates (0.3-3 mm/yr). The scientific results and contributions of this thesis are presented across three main chapters.
In this thesis, we investigate the spatiotemporal organization of past earthquakes along the central-eastern ATF. We used trench excavations and offset measurements derived from high-resolution satellite imagery to characterize past earthquakes along the Aksay fault section. Then, we expand our analysis to a regional scale (multiple fault sections) by integrating our observations from the Aksay section with previously published investigations on neighboring fault sections. We found that while individual fault sections tend to rupture in a quasi-periodic manner, at the regional scale, earthquakes appear to cluster in time as a result of long-term fault interactions between the fault sections. Since reconstructing the organization of past earthquakes in a fault system requires integrating paleoseismic data across different spatial scales. We introduce a statistically based methodological framework to correlate and combine paleoseismic data from multiple trench sites into a unified paleoearthquake time series. Applying this approach to the ATF, we reduce event age uncertainties and provide new insights into the recurrence of large earthquakes. Notably, we found a regular decrease in the return times of past earthquakes during clustering periods, suggesting high probabilities of a large earthquake in the next decades along the central-eastern ATF. Finally, we focused on northwestern Mongolia, where we analyzed lake sediment records from Oygon, Kholboo, and Sangiyn Dalai lakes. Given the broad morphologies, shallow depths, and very low sedimentation rates of these lakes, no macroscopic event deposits were visually recognized. We used a multi-proxy approach and introduced the product-based semblance method to identify sections along the core with geochemical anomalies. From these analyses, we identified micrometer-scale event deposits related to strong earthquake shaking in continental interiors. We argue that the BFS releases strain by episodic independent ruptures along the Bulnay fault and synchronized major events involving the Bulnay and nearby Tsetserleg faults. The results presented in this thesis open new perspectives for understanding how large intracontinental strike-slip faults accommodate strain over long timescales. Building upon these findings, several research directions emerge, ranging from targeted field investigations and refined paleoseismic chronologies to better constrain numerical modeling simulations and seismic hazard assessment.