Surface deformation associated with fractures near the 2019 Ridgecrest earthquake sequence

MEA-seqX: High-Resolution Profiling of Large-Scale Electrophysiological and Transcriptional Network Dynamics

Concepts of brain function imply congruence and mutual causal influence between molecular events and neuronal activity. Decoding entangled information from concurrent molecular and electrophysiological network events demands innovative methodology bridging scales and modalities. The MEA-seqX platform, integrating high-density microelectrode arrays, spatial transcriptomics, optical imaging, and advanced computational strategies, enables the simultaneous recording and analysis of molecular and electrical network activities at mesoscale spatial resolution. Applied to a mouse hippocampal model of experience-dependent plasticity, MEA-seqX unveils massively enhanced nested dynamics between transcription and function. Graph-theoretic analysis reveals an increase in densely connected bimodal hubs, marking the first observation of coordinated hippocampal circuitry dynamics at molecular and functional levels. This platform also identifies different cell types based on their distinct bimodal profiles. Machine-learning algorithms accurately predict network-wide electrophysiological activity features from spatial gene expression, demonstrating a previously inaccessible convergence across modalities, time, and scales.

Extensive off-fault damage around the 2023 Kahramanmaraş earthquake surface ruptures

Quantifying coseismic fault offsets for surface ruptures of major earthquakes is important for earthquake cycle and slip-rate studies, and thus for earthquake hazard assessments. However, measurements of such offsets generally underestimate fault slip due to inelastic deformation and secondary fault offsets, i.e., off-fault damage. Here, we use satellite synthetic aperture radar images to quantify off-fault damage in the two 2023 Kahramanmaraş (Türkiye) magnitude 7.8 and 7.6 earthquakes. We first derive three-dimensional coseismic surface displacements and show that on average ~35% of the coseismic slip is accommodated by off-fault damage within 5-7 km of the coseismic surface ruptures. Fault sections exhibiting geometrical complexities (e.g., bends and step-overs) experienced a higher level of off-fault damage than simpler fault sections. Our results highlight the importance of extending off-fault damage assessments to several km away from fault ruptures and indicate that fault offset measurements may underestimate slip-rate estimations by as much as a third.

Landscape changes caused by the 2024 Noto Peninsula earthquake in Japan

Landscapes are shaped by tectonic, climatic, and surface processes over geological timescales, but we rarely witness the events of marked landscape change. The moment magnitude 7.5 Noto Peninsula earthquake in central Japan was caused by a large thrust faulting, up to nearly 10 meters of slip, that expanded more than 150 kilometers along the fault zone. The deformation field reconstructed from satellite data and field surveys reveals up to 4.4 meters of uplift and associated coastal advance along the entire northern coast of the peninsula, meter-scale systematic movement of the mountain slopes consistent with slip on flexural faults, and activation of secondary inland faults, suggesting synchronized ruptures. The findings show excellent consistency between the coseismic deformation and geomorphic features and provide a vivid example of the role of a major earthquake in landscape formation.

Self-evolving artificial intelligence framework to better decipher short-term large earthquakes

Large earthquakes (EQs) occur at surprising loci and timing, and their descriptions remain a long-standing enigma. Finding answers by traditional approaches or recently emerging machine learning (ML)-driven approaches is formidably difficult due to data scarcity, interwoven multiple physics, and absent first principles. This paper develops a novel artificial intelligence (AI) framework that can transform raw observational EQ data into ML-friendly new features via basic physics and mathematics and that can self-evolve in a direction to better reproduce short-term large EQs. An advanced reinforcement learning (RL) architecture is placed at the highest level to achieve self-evolution. It incorporates transparent ML models to reproduce magnitude and spatial location of large EQs ([Formula: see text] 6.5) weeks before of the failure. Verifications with 40-year EQs in the western U.S. and comparisons against a popular EQ forecasting method are promising. This work will add a new dimension of AI technologies to large EQ research. The developed AI framework will help establish a new database of all EQs in terms of ML-friendly new features and continue to self-evolve in a direction of better reproducing large EQs.

Gauss curvature-based unique signatures of individual large earthquakes and its implications for customized data-driven prediction

Statistical descriptions of earthquakes offer important probabilistic information, and newly emerging technologies of high-precision observations and machine learning collectively advance our knowledge regarding complex earthquake behaviors. Still, there remains a formidable knowledge gap for predicting individual large earthquakes’ locations and magnitudes. Here, this study shows that the individual large earthquakes may have unique signatures that can be represented by new high-dimensional features—Gauss curvature-based coordinates. Particularly, the observed earthquake catalog data are transformed into a number of pseudo physics quantities (i.e., energy, power, vorticity, and Laplacian) which turn into smooth surface-like information via spatio-temporal convolution, giving rise to the new high-dimensional coordinates. Validations with 40-year earthquakes in the West U.S. region show that the new coordinates appear to hold uniqueness for individual large earthquakes ($$M_w \ge 7.0$$ M w ≥ 7.0 ), and the pseudo physics quantities help identify a customized data-driven prediction model. A Bayesian evolutionary algorithm in conjunction with flexible bases can identify a data-driven model, demonstrating its promising reproduction of individual large earthquake’s location and magnitude. Results imply that an individual large earthquake can be distinguished and remembered while its best-so-far model can be customized by machine learning. This study paves a new way to data-driven automated evolution of individual earthquake prediction.