The Minshan Uplift Zone (MUZ) is located at the eastern margin of the Tibetan Plateau, which is the junction of three tectonic terranes. The observed discrepancy between a high uplifting and low shortening rate over the MUZ is attributed to the intrusion of a viscous lower-crust. In the last 50 years, several significant earthquakes occurred at the boundaries of the MUZ, i.e. the Huya and Mingjiang fault. On August 8, 2017, the Jiuzhaigou earthquake (Mw 6.5) occurred on the northern extension of the Huya fault. We adopt a joint inversion of the Interferometric Synthetic Aperture Radar and teleseismic body-wave data to investigate the rupture process of this event. The obtained slip model is dominated by left-lateral strike-slips on a sub-vertical fault presenting significant shallow slip deficit. The rupture initiation is composed of both thrust and strike-slip mechanisms producing a none-double-couple solution. We also resolve a secondary fault branch forming an obtuse angle with the main fault plane at its northern end. These phenomena indicate that the northern Huya fault is a young (less mature) fault system. Focal mechanisms of the regional earthquakes demonstrate that the northern and southern Huya fault presents different combinations of strike-slip and reversed motion. We attribute such discrepancy to the lateral extension of the viscous lower-crust, which appears to extrude to the east beyond the northern Huya fault, in comparison with that confined under the MUZ near the southern Huya fault. This conceptual model is also supported by geomorphological and magneto-telluric observations.

Detection of weak seismic signal is important to investigate seismicity and/or aseismic behavior of a fault system, thus in the last decades, people adopted specially designed networks to detect such weak signals. High sensitivity networks generally require dense station coverage, long deployment period, special field environment or high funding cost, which are not suitable for immediate monitoring of aftershocks. To investigate aftershock activities of the 2017 Mw = 6.5, Jiuzhaigou earthquake, we designed and deployed an experimental seismic network (Sept - Dec 2017) near the source region. This network is composed of 9 refined sub-arrays, which is composed of 8-16 short period seismic stations with lateral coverage of several hundred meters. Such design utilizes the waveform coherency within each sub-array to enhance seismic signals and to suppress random and impulsive noises. We conduct a theoretical analysis of two signal enhancement algorithms, e.g. linearly stacking and Geometric Mean Envelope (GME) weighting, and found the GME algorithm could significantly suppress impulsive noise presented at a single station. We test the performance of two algorithms with short time average over long time average (STA/LTA) criterion and found the GME algorithm can significantly reduce the occurrence of earthquake over-counting. Such array geometry and the GME weighting algorithm serve as a promising network design and signal enhancement algorithm to improve the monitoring of micro-seismic signals.