Yixiao Sheng

The San Jacinto fault is one of the most active faults in Southern California and a major cause of seismic hazard. Estimating its capability to generate large earthquakes requires a detailed understanding of its mechanical properties and of their temporal changes at seismogenic depths, which remain difficult to characterize with standard seismic data and methods. In this work, we use 2.5 years of seismic noise recorded by a dense array of 300 nodes deployed at the Piñon Flat Observatory (PFO) from April 2022 to October 2024 to monitor the San Jacinto fault zone with tremor-like signals generated by freight trains travelling in the nearby Coachella valley. Recent studies have shown that trains are powerful and repeatable sources of high-frequency (> 1 Hz) body waves that can be used for monitoring seismic velocity variations at seismogenic depths (5 - 10 km) by seismic interferometry, but they have only exploited P waves so far. Here we show that cross-correlating train tremors recorded on both sides of the San Jacinto fault zone enables us to retrieve both P and S waves, and therefore to measure relative changes of both VP and VS with time, from which we can derive relative changes of an effective VP/VS ratio in the sampled volume. While the 2.5-year-long observation period does not include significant earthquakes in the target area, our results show fluctuations of seismic velocities and of the derived VP/VS ratio, with long-term trends as well as more rapid changes. The quantitative interpretation of these variations remains to be specified, but they are likely related to changes in stress, porosity, or fluid pore pressure at depth. Combining these results with other observables (e.g., detailed catalogs of micro-seismicity) will provide valuable information on the dynamic mechanical behaviour of the San Jacinto fault, potentially yielding insights into the evolution of cracks and fluids in the fault zone.

Large earthquakes involve processes that occur primarily at seismogenic depths of 5-10 km or more and that are thus difficult to observe in detail with traditional seismic networks. The FaultScan project establishes the first long-term, dense nodal array experiment by acquiring continuous data at 300 stations along the San Jacinto Fault (SJF, South California) at the Piñon Flat Observatory for 2.5 years (April 2022-Nov. 2024).We report on the experiment logistics, data quality and describe the characteristics of the High-Frequency (HF, >1 Hz) back-ground noise with a focus on incoherent wind generated noise, train traffic tremors and far distant coherent urban noise. We illustrate how these HF coherent noise sources recorded mostly as body-waves can be used to track velocity changes across the SJF by applying seismic interferometry between the array and nearby permanent seismic stations.We further describe how much slant-stacking increases the level of signal to noise ratio for the detection of small earthquakes along the San Jacinto Fault and show how newly detected events improve our description of foreshock/aftershock patterns. Our search for tectonic tremors along the San Jacinto Fault turns up empty but we observe tremor-like signals, mostly T-phases coming exclusively from the Tonga subduction and one intriguing T-phase like sequence originating from the October 2023, offshore Japan volcanic crisis.

Industrial facilities can act as persistent, high-energy seismic noise sources and are increasingly exploited for passive time-lapse monitoring. In geothermal settings, vibrations generated by power plants provide spatially stable sources that are well suited for long-term interferometric analyses. However, temporal variations in operational conditions may lead to changes in source spectral content, potentially biasing time-lapse measurements derived from ambient noise cross-correlation. Assessing and mitigating the effects of such source nonstationarity is therefore essential for reliable monitoring.We investigate continuous seismic noise recorded in 2008 at the Salton Sea geothermal field (southern California) and apply a Variational Autoencoder (VAE) to characterize temporal variability in source spectra. The VAE is trained on frequency spectra from relatively stable periods and subsequently used to identify time intervals exhibiting anomalous time–frequency behavior, interpreted as changes in the industrial noise source. We then compute noise cross-correlation functions and corresponding travel-time variations (dt) for both normal and anomalous periods.Our analysis reveals systematic differences in dt behavior associated with source spectral changes, including abrupt offsets at transitions between normal and anomalous intervals and increased high-frequency fluctuations during anomalous periods. These effects occur despite stable source locations, demonstrating that spectral variability alone can significantly contaminate time-lapse measurements.To reduce these biases, we construct separate correlation reference functions for distinct source regimes. This adaptive strategy suppresses spurious dt fluctuations during anomalous intervals and yields more physically interpretable travel-time variations. The results highlight the importance of explicit source characterization in passive seismic monitoring and demonstrate how machine learning–based approaches can enhance the robustness of time-lapse interferometry in geothermal fields and other environments dominated by industrial noise sources.

Shallow subsurface velocity variations provide key constraints for understanding the migration of soil water and associate hydrological processes. In this study, we use a distributed acoustic sensing (DAS) unit to record high-frequency ambient noise and apply ballistic wave seismic monitoring (Mi et al., 2025) to retrieve velocity variations within the upper 10 m of the shallow subsurface in Hefei, Anhui Province, China, over a two-month period. During non-precipitation periods, low-frequency (< 22 Hz) phase velocities exhibit a negative correlation with temperature, whereas high-frequency (>22 Hz) phase velocities show a positive correlation. Following precipitation events, phase velocities decrease significantly across all frequencies. Based on a reference 1-D shear-wave velocity model, we further invert the time-dependent phase velocity perturbations to obtain depth-dependent shear-wave velocity variations (Haney and Tsai, 2017). Integration with borehole observations reveals contrasting responses between shallow (5 m) layers to evaporation, infiltration, and loading: diurnal temperature variations regulate soil moisture and thereby control velocity changes during dry periods, while rainfall-induced infiltration becomes the dominant factor during precipitation. Our results demonstrate the effectiveness of DAS-based time-lapse velocity monitoring for characterizing shallow soil water cycling and highlight its potential for high spatiotemporal resolution hydro-geophysical monitoring of the near surface.RaeferencesHaney, M.M. & Tsai, V.C., 2017. Perturbational and nonperturbational inversion of Rayleigh-wave velocities, Geophysics, 82, F15–F28.Mi, B., Xia, J. & Li, J., 2025. On the measurement of relative phase velocity changes for ballistic wave seismic monitoring, Geophysical Journal International, 234, 1-9.

We utilize train tremors as P‐wave seismic sources to investigate velocity‐strain sensitivity near the San Jacinto Fault Zone. A dense nodal array deployed at the Piñon Flat Observatory is used to detect and identify repeating train energy emitted from a railway in the Coachella valley. We construct P‐wave correlation functions across the fault zone and estimate the spatially averaged dt / t versus strain sensitivity to be 6.25 × 10 ⁴ . Through numerical simulations, we explore how the sensitivity decays exponentially with depth. The optimal solution reveals a subsurface sensitivity of 1.2 × 10 ⁵ and a depth decay rate of 0.05 km ⁻¹ . This sensitivity aligns with previous findings but is toward the higher end, likely due to the fractured fault‐zone rocks. The depth decay rate, previously unreported, is notably smaller than assumed in empirical models. This raises the necessity of further investigations of this parameter, which is crucial to study stress and velocity variations at seismogenic depth.

SUMMARY: Human activities generate seismic waves, which have been traditionally analysed by seismic interferometry. However, this processing technique often neglects the seismic source distribution and is prone to introducing biases in interpretations. This study proposes an alternative processing technique known as seismic stereometry to analyse seismic signals generated by cars. The proposed algorithm exploits the highly coherent wavefields recorded by dense seismic arrays and directly measures the seismic wave traveltime differences between sensors. By doing so, it intrinsically considers the seismic source distribution. Traveltime differences can be used to invert the positions of vehicles and the near-surface velocity structures. We demonstrate the analysis using a dense seismic array deployed atop a southern San Andreas Fault segment and invert the shallow S-wave velocity using seismic waves generated by only one car. The velocity inversion result reveals a clear contrast across the fault, consistent with previous studies. The developed technique, combined with the distributed acoustic sensing technology, has great potential for applications in the urban environment.

Seismic velocities in rocks are highly sensitive to changes in permanent deformation and fluid content. The temporal variation of seismic velocity during the preparation phase of earthquakes has been well documented in laboratories but rarely observed in nature. It has been recently found that some anthropogenic, high‐frequency (>1 Hz) seismic sources are powerful enough to generate body waves that travel down to a few kilometers and can be used to monitor fault zones at seismogenic depth. Anthropogenic seismic sources typically have fixed spatial distribution and provide new perspectives for velocity monitoring. In this work, we propose a systematic workflow to seek such powerful seismic sources in a rapid and straightforward manner. We tackle the problem from a statistical point of view, considering that persistent, powerful seismic sources yield highly coherent correlation functions (CFs) between pairs of seismic sensors. The algorithm is tested in California and Japan. Multiple sites close to fault zones show high‐frequency CFs stable for an extended period of time. These findings have great potential for monitoring fault zones, including the San Jacinto Fault and the Ridgecrest area in Southern California, Napa in Northern California, and faults in central Japan. However, extra steps, such as beamforming or polarization analysis, are required to determine the dominant seismic sources and study the source characteristics, which are crucial to interpreting the velocity monitoring results. Train tremors identified by the present approach have been successfully used for seismic velocity monitoring of the San Jacinto Fault in previous studies.

Microseismic noise has been used for seismic velocity monitoring. However, such signals are dominated by low‐frequency surface waves that are not ideal for detecting changes associated with small tectonic processes. Here we show that it is possible to extract stable, high‐frequency body waves using seismic tremors generated by freight trains. Such body waves allow us to focus on small velocity perturbations in the crust with high spatial resolution. We report on 10 years of seismic velocity temporal changes at the San Jacinto Fault. We observe and map a two‐month‐long episode of velocity changes with complex spatial distribution and interpret the velocity perturbation as produced by a previously undocumented slow‐slip event. We verify the hypothesis through numerical simulations and locate this event along a fault segment believed to be locked. Such a slow‐slip event stresses its surroundings and may trigger a major earthquake on a fault section approaching failure.

Anthropogenic seismic signals attract more and more attention in recent years. Freight trains, among different seismic sources, are of particular interest for seismic velocity monitoring due to several advantages. Trains are persistent, powerful sources that generate seismic tremors equivalent to Mw 2 earthquakes and detectable up to 100 km distance; trains move along fixed trajectories, allowing us to properly account for the source distribution and its coupling between the structure; trains generate high-frequency body-wave energy enabling us to focus on the changes at depth with high spatial resolution. We propose a fault zone monitoring framework through a case study in southern California. Freight trains running through the Coachella Valley are used to monitor changes associated with the San Jacinto Fault. The general steps include identifying sources and constructing a train catalog, extracting body waves through seismic interferometry, measuring travel-time perturbation, and mapping seismic velocity change. We analyze the seismic data from 2010 to 2020 and discover an episode of velocity changes in early 2014, manifested on all station pairs considered. The velocity perturbation shows a complicated spatial pattern, with some station pairs exhibiting positive changes, and others negative changes. We interpret that this velocity perturbation results from an aseismic slip near the edge of the Anza seismic gap and further validate this idea using numerical simulations. We use the Coulomb software to simulate volumetric strain for velocity perturbation and full-waveform modeling to simulate correlation functions for estimating travel-time change. The proposed framework has great potentials to be applied in other settings, from wastewater injection to CO2 sequestration, using freight trains or other type of anthropogenic sources.