Long-term seismicity is an effective tool to infer fault properties at depth, but the catalog construction is challenging because of the large data volume. We propose a new deep learning-based workflow that follows a “Train-Detect-Pick” procedure, which solves the generalization problem in AI pickers. We apply the new workflow on the preseismic phase (2008-2019) of Ridgecrest-Coso region. Results show that the new workflow realizes efficient and stable detection, and well substitutes matched filter. Our new catalog helps characterize the preseismic fault behavior: (1) the Ridgecrest area has a distributed deformation, and the 2019-ruptured segment has a persistent asperity; (2) the central Garlock fault is unfavorable for rupture propagation, because of its discontinuous geometry and low coupling ratio; (3) the Coso geothermal field generates intense and shallow seismicity, which has a high b-value that does not correlate with seismicity rate and industrial production, thus suggest a low stress level.

Characterizing fault behaviors prior to large earthquakes through long-term seismicity is crucial for seismic hazard assessment, yet constructing high-resolution catalogs over extended periods poses significant challenges. This study introduces LoSAR, a novel deep learning-driven workflow that enhances phase picking by Localizing a Self-Attention Recurrent neural network with local data, addressing the generalization problem common in data-driven approaches. We apply LoSAR to two distinct regions that are both featured by recent large earthquakes: (1) preseismic period of the Ridgecrest-Coso region (2008-2019), and (2) pre-postseismic period of the East Anatolian Fault Zone (EAFZ, 2020-2023/04). Through detailed comparisons, we demonstrate that LoSAR offers slightly higher detection completeness than the QTM matched filter catalog, while boosts an over 100 times faster processing and a superior temporal stability, avoiding low-magnitude gaps during background periods. Against PhaseNet and GaMMA, two established phase picker and associator, LoSAR proves more scalable and generalizable, achieving roughly 2.5 times more event detections in the EAFZ case, along with a ~7 times higher phase association rate. By leveraging the two enhanced catalogs and b-value analysis, we gain insights into the preseismic fault behaviors: (1) The Ridgecrest faults are characterized by sparse and distributed seismicity across a band of ~20 km, revealing multiple orthogonal preexisting faults; coupled with a low b-value that signifies this area as a persistent asperity; (2) The Erkenek-Pütürge segment of EAFZ exhibits complex fault geometry that forms a persistent rupture barrier, which consists of a hidden conjugate fault system that presents as a ~10-km wide fault zone.

The 2021 Mw 6.1 Yangbi earthquake in southwest China is preceded by three major foreshocks: 05/18 Mw 4.3, 05/19 Mw 4.6, and 05/21 Mw 5.2. It provides a valuable chance to revisit two end-member models describing earthquake interaction: cascade-up and pre-slip model. We first determine the associated fault structure with relocated aftershocks and focal mechanisms obtained from multi-point-source inversion. We find that the mainshock and two smaller foreshocks occur on an unmapped near-vertical fault, and the largest foreshock occurs on a mapped stepover fault that dips to NE. Secondly, for each major foreshock, we estimate and delineate their rupture area based on aftershocks and spectral ratio analysis. Based on the rupture model, we finally calculate the evolution of Coulomb stress, with which to interpret the causality of each major event. Results show that the Yangbi sequence can be explained by the cascade triggering mechanism, while we also find evidence for aseismic slip that contributes to the triggering process: the first foreshock is preceded by a short-term localized cluster, and the aftershock zone of the second foreshock extents through time. The nucleation of mainshock is probably contributed by multiple major foreshocks through both seismic and aseismic processes. This detailed seismological characterization of Yangbi sequence lend supports for a deeper understanding on the foreshock mechanism: (1) the controlling mechanisms are not limited to cascade-up & pre-slip, multiple mechanisms can operate together; and (2) aseismic slip does not always provide more predictability on the mainshock.

The Xiaojiang Fault (XJF) Zone locates in the southeastern of Tibetan Plateau and defines the boundary between the South China and Sichuan-Yunnan blocks. Historical large earthquakes were hosted on the XJF, though its seismic hazard in the near future is under debate. In this study, we utilize portable broad-band seismic network to unravel the microseismic activities along XJF, and to further investigate the fault structures and their properties. Adopting PALM, a newly developed earthquake detection algorithm, we obtained ~13,000 relocated events. The micro-seismicity reveals widespread off-fault structures showing conjugate geometry, while the major faults present low seismicity. The fault branches conjugate to the main-fault present intensive microseismicity, which hosts repeating events and presents high b-value. Regional GPS stations reflect slips are mostly concentrated along the XJF, while the slip rate on off-fault branches correlates with seismic activities on these structures. Combining with other recent seismological and magnetotellurics evidences, we suggest a low strength on these off-fault structures, which may partially release tectonic stress loading and serve as a barrier for future big earthquakes. On the XJF, the microseismic events are clustered on the fault junctions with low b-value. A special set of clusters between 25°N to 25.5°N show an along-strike variation of depth from 10 to 25-km, imaging the boundary between creeping and locked fault portions. We revisit the seismic hazard problem of XJF, and conclude that XJF is at the late stage of inter-seismic period.