Southern California Seismic Network Research Articles

A study on small magnitude seismic phase identification using 1D deep residual neural network

Reliable seismic phase identification is often challenging especially in the circumstances of low-magnitude events or poor signal-to-noise ratio. With improved seismometers and better global coverage, a sharp increase in the volume of recorded seismic data has been achieved. This makes handling seismic data rather daunting by using traditional approaches and therefore fuels the need for more robust and reliable methods. In this study, we develop 1D deep Residual Neural Network (ResNet), for tackling the problem of seismic signal detection and phase identification. This method is trained and tested on the dataset recorded by the Southern California Seismic Network. Results demonstrate that the proposed method can achieve robust performance for the detection of seismic signals and the identification of seismic phases. Compared to previously proposed deep learning methods, the introduced framework achieves around 4% improvement in earthquake detection and a slightly better performance in seismic phase identification on the dataset recorded by Southern California Earthquake Data Center. The model generalizability is also tested further on the STanford EArthquake Dataset. In addition, the experimental result on the same subset of the STanford EArthquake Dataset, when masked by different noise levels, demonstrates the model’s robustness in identifying the seismic phases of small magnitude.

Spatiotemporal Graph Convolutional Networks for Earthquake Source Characterization.

Accurate earthquake location and magnitude estimation play critical roles in seismology. Recent deep learning frameworks have produced encouraging results on various seismological tasks (e.g., earthquake detection, phase picking, seismic classification, and earthquake early warning). Many existing machine learning earthquake location methods utilize waveform information from a single station. However, multiple stations contain more complete information for earthquake source characterization. Inspired by recent successes in applying graph neural networks (GNNs) in graph-structured data, we develop a Spatiotemporal Graph Neural Network (STGNN) for estimating earthquake locations and magnitudes. Our graph neural network leverages geographical and waveform information from multiple stations to construct graphs automatically and dynamically by adaptive message passing based on graphs' edges. Using a recent graph neural network and a fully convolutional neural network as baselines, we apply STGNN to earthquakes recorded by the Southern California Seismic Network from 2000 to 2019 and earthquakes collected in Oklahoma from 2014 to 2015. STGNN yields more accurate earthquake locations than those obtained by the baseline models and performs comparably in terms of depth and magnitude prediction, though the ability to predict depth and magnitude remains weak for all tested models. Our work demonstrates the potential of using GNNs and multiple stations for better automatic estimation of earthquake epicenters.

Relative Focal Mechanism Inversion and Its Application to Ridgecrest Sequence

AbstractEarthquake focal mechanisms are important for characterizing the subsurface faulting geometry and evaluating stress distributions. Existing approaches usually strive to determine the absolute focal mechanisms and may be subject to large uncertainties due to incomprehensive knowledge of the velocity model, particularly for moderate-to-small earthquakes. Alternatively, difficulties that arise from the velocity model can be largely mitigated by inverting the relative data variations in a series of earthquakes, because effects from the velocity model are systematic among all events in the vicinity. In this study, we propose a novel relative focal mechanism inversion (RFMI) method to invert the second-order variations in a series of focal mechanisms utilizing a well-constrained primary event. We test the RFMI method on both synthetic data and 251 real earthquakes (M ≥3) in the 2019 Ridgecrest sequence. The synthetic test results show that the RFMI method is robust and insusceptible to location errors (<2 km) and systematic velocity errors (5%). The real data application results demonstrate improved consistency among the inverted focal mechanisms, resulting in better characterization of the fault orientations than the Southern California Seismic Network (SCSN) focal mechanism catalog. The retrieved earthquake depths are also well correlated with the depths of the Mw 6.4 and 7.1 mainshocks. Waveform cross-correlation analysis verifies the reliability of the results. Furthermore, dynamic stress monitoring is enabled with decent resolution. The proposed RFMI method paves a new path toward achieving a rich number of reliable earthquake focal mechanisms, which will benefit the investigation of the earthquake process.

Variability in the Natural Frequencies of a Nine-Story Concrete Building from Seconds to Decades

Abstract Since 2001, the Southern California Seismic Network has archived continuous waveform data from strong-motion station CI.MIK in Caltech Hall (formerly Millikan Library)—a nine-story reinforced concrete building. Simple spectral analysis of this 20 yr record reveals that the building’s fundamental frequencies have increased gradually by 5.1% (east–west) and 2.3% (north–south), with larger long-term variability up to 9.7% (east–west) and 4.4% (north–south). This finding is unexpected, as previous analysis of forced vibration tests and strong-motion records has shown that between 1968 and 2003 the fundamental frequencies decreased by 22% (east–west) and 12% (north–south), largely attributed to minor structural damage and soil–structure system changes from major earthquakes. Today, the building’s apparent structural stiffness is comparable to what it was in 1986, before the Whittier Narrows earthquake. Using data from earthquakes and forced vibrations, we also document the building’s nonlinear dynamic elasticity, which is characterized by a rapid softening (decrease in apparent frequencies) at the onset of strong motion, followed by a slower, log–linear recovery trend over the scale of minutes. This nonlinear behavior does not appear to have changed with time.

Temporal changes of seismicity in Salton Sea Geothermal Field due to distant earthquakes and geothermal productions

SUMMARY The Salton Sea Geothermal Field (SSGF) is one of the most seismically active and geothermally productive fields in California. Here we present a detailed analysis of short-term seismicity change in SSGF from 2008 to 2013 during and right following large distant earthquakes, as well as long-term seismicity change due to geothermal productions. We first apply a GPU-based waveform matched-filter technique (WMFT) to the continuous data recorded by the Calenergy Borehole (EN) Network and detect more than 70 000 new micro-earthquakes than listed in the standard Southern California Seismic Network catalogue. We then analyse the seismicity rate changes in the SSGF associated with transient stress fluctuations triggered by regional and large teleseismic earthquakes from 1999 to 2019. We find triggered seismicity in the SSGF following seven regional M > 5.5 earthquakes. In comparison, most teleseismic earthquakes with M > 8.0 did not trigger significant seismicity rate change in the SSGF, likely indicating a frequency dependence in remote dynamic triggering. We further characterize the correlation between the long-term seismicity rate and geothermal production rates, and the temporal and spatial distribution of Guttenberg–Richter b-values inside and outside the SSGF with the newly detected catalogue. The long-term seismicity shows that events with M > 1.5 are likely correlated with net production rates, while smaller events do not show any correlation. The b-values inside the SSGF are higher than those outside the SSGF, and the locations of dynamically triggered events are close to locations with high b-values.

Eikonal surface wave tomography with smoothing splines—application to Southern California

SUMMARY The densification of both permanent and temporary seismic networks has raised new interest in surface wave eikonal tomography from which phase velocity maps can be obtained without resolving a tomographic inverse problem. However, eikonal tomography requires to reconstruct traveltime surfaces from a discrete number of measurements obtained at the station locations, which can be challenging. We present a new method to reconstruct these traveltime surfaces with smoothing splines discretized in a regular 2-D Cartesian grid. We impose Neumann boundary conditions so that the phase gradients on the edges of the grid are equal to the apparent slownesses of the average plane wave along the normal direction measured by beamforming. Using the eikonal equation, phase velocity maps are then derived from the norm of the gradient of the interpolated traveltime maps. The method is applied to Rayleigh waves recorded by the Southern California Seismic Network to derive phase velocity surfaces. Robust, stable and finely resolved phase velocity maps at 25 and 33 s period are obtained after averaging the phase velocity maps derived from the analysis of a selection of recent large (Mw ≥ 6.5) teleseismic events. The phase velocity map at 25 s mainly constrains the thickness of the Southern California crust, with results that are in excellent agreement with previous tomographic studies.

Development of Qube: A Low-Cost Internet of Things Device for On-Site and Regional Earthquake Warning

Abstract Earthquakes are a major global risk. The current earthquake early warning systems based on public seismic stations face challenges such as high cost, low density, high latency, no alert zone, and difficulty in predicting ground motions at the location of the user. This article pursues an alternative consumer-based approach. An Internet of Things consumer device, called a “Qube,” was built for a cost below $100 and is about the size of a Rubik’s cube. The Qube successfully detected earthquakes and issued earthquake warnings through sounding the onboard alarm for on-site warning and sending text messages to local subscribers for regional warning. The Qube is highly sensitive. During nine months of testing from September 2020 to May 2021, it detected all earthquakes over M 3.0 magnitude around Los Angeles, as well as nearby earthquakes down to M 2.3. The Qube uses a geophone for ground-motion velocity sensing and captures earthquake waveforms consistent with a nearby broadband seismometer in the Southern California Seismic Network. By analyzing data of the earthquakes detected by the Qube, an empirical logarithmic formula that is used to estimate local earthquake magnitude based on detected ground-motion amplitude in digital counts was developed. Although the Qube’s response in digital counts to ground-motion velocity in μm/s has not been determined, the empirical formula between Qube’s output and local earthquake magnitude suggests the Qube’s consistency in ground-motion measurement. The Qube has Wi-Fi connectivity and is controllable via a smartphone or computer. The combination of low cost, high sensitivity, and integrated alarm function of the Qube is intended to enable a consumer-based approach with the potential for mass adoption and use in dense networks, creating new opportunities for seismic network, earthquake warning, and educational applications.

An Earthquake Detection and Location Architecture for Continuous Seismograms: Phase Picking, Association, Location, and Matched Filter (PALM)

Abstract We developed an earthquake detection and location architecture for continuous seismograms that incorporates phase picking, phase association, location, and matched-filter techniques (PALM). The PALM architecture incorporates two modules: (1) PAL, the initial detection following picking, association, and location processes, and (2) match, expand, shift, and stack (MESS), a matched-filter detector that augments the template catalog. The effectiveness of PALM is demonstrated in building an early aftershock catalog for the 2019 Ridgecrest, California, earthquake. By comparing with Southern California Seismic Network arrival times, we show that the PAL picker combines the strengths of short-term average/long-term average and the kurtosis picker, realizing robust phase detection and precise picking. Our final MESS catalog is compared with two other matched-filter catalogs by Ross, Idini, et al. (2019) and Shelly (2020). We find that PALM directly recovers unbiased and detailed features in seismicity from continuous seismograms, which can be efficiently implemented to scan continuous waveforms without the need for visual inspection.

Determination of Near Surface Shear‐Wave Velocities in the Central Los Angeles Basin With Dense Arrays

AbstractIn this study, we investigate the shallow shear wave velocity structure of the Los Angeles Basin in southern California, using ambient noise correlations between 5 dense arrays and 21 broadband stations from the Southern California Seismic Network (SCSN). We observe clear fundamental mode and first overtone Rayleigh waves in the frequency band 0.25–2.0 Hz, and obtain group velocity maps through tomography. We further derive a 3D shear wave velocity model, covering a large portion of the central LA Basin for the depths shallower than 3 km. We found that the small scale shallow velocity structure heterogeneities are better resolved compared with the SCEC Community velocity models. Our model captures the presence of the Newport‐Inglewood fault by a NW–SE trending high velocity belt. Our model provides more accurate constraints on local ground motion predictions with detailed mapping of structural heterogeneities.

Monitoring Human Activity at a Very Local Scale with Ground-Motion Records: The Early Stage of COVID-19 Pandemic in California, U.S.A., New York City, U.S.A., and Mexicali, Mexico

Abstract In this article, we analyze the change in anthropogenic seismic noise level within a frequency range of 4–14 Hz, through a survey of seismic stations in California, United States, New York City, United States, and Mexicali, Baja California, Mexico from early December 2019 to late April 2020. Our analysis shows that some stations recorded a drop in anthropogenic seismic noise during the COVID-19 pandemic, and the timing of the anthropogenic noise decrease typically correlates with the timing of a strict curtailment of personal and economic activity issued by the local government. In other locations, the drop in the anthropogenic seismic noise appears not to follow the lockdown timing perfectly. During our analysis, we observed that many stations did not record a drop during the early stage of COVID-19 pandemic. Of the 19 stations of the Southern California Seismic Network that were surveyed, we found that only five show a similar extent of drop in anthropogenic seismic noise comparable to the Christmas holiday break in 2019. This suggests that the human activity that caused seismic noise did not significantly reduce during the COVID-19 pandemic near most surveyed stations in southern California. A further analysis implies that the primary seismic noise source in southern California might be traffic, and the continuation of industrial traffic, such as cargo transportation, during the COVID-19 pandemic may be the reason why many stations did not record a noise drop. Our results show that the anthropogenic seismic noise recorded by seismic stations is capable of indicating human activity, and that this metric is, particularly, powerful in measuring how localized communities initially responded to the COVID-19 pandemic.

Calibrating Spectral Decomposition of Local Earthquakes Using Borehole Seismic Records—Results for the 1992 Big Bear Aftershocks in Southern California

AbstractEmpirical approaches are often used to correct earthquake spectra for path effects in order to estimate corner frequencies and stress drops, but suffer from tradeoffs among source parameters and the shape of the empirical correction or Green's function. This hampers absolute stress‐drop estimates and comparisons between different studies or regions. Borehole records are less affected by site and attenuation effects than surface seismometers and thus provide more direct constraints on source spectra. We analyze 1,787 earthquakes of the 1992 to 1993 Big Bear aftershock sequence in southern California, which were recorded by both the Southern California Seismic Network (SCSN) and the Cajon Pass borehole seismometer at 2.5‐km depth. Spectral decomposition of P‐waves recorded by the SCSN yields source terms that can be fit equally well by a variety of models, owing to tradeoffs between the model parameters and the empirical correction term. Earthquake spectra recorded by the Cajon Pass borehole are consistent with a P‐wave Q of about 1,000. For 25 Big Bear earthquakes in common, we find the correction spectrum that brings the event terms from spectral decomposition and the Q‐corrected borehole spectra into best‐fitting agreement. This provides absolute calibration for the SCSN spectral decomposition results, which can then be used to process all the Big Bear earthquakes and estimate earthquake source properties without the parameter tradeoffs normally associated with the EGF approach. Our results highlight the limits of purely empirical approaches to spectral analysis for earthquake source properties and the importance of determining crustal attenuation models.

Comparison and Analysis of Network Construction Methods for Seismicity Based on Complex Networks

The approach of the complex network has well described seismic complex systems. In this paper, this is the first time three classical network construction methods for seismicity are compared. By using the same dataset from the Southern California Seismic Network, three networks are constructed. They all present the scale-free, small-world properties, a strength-degree correlation, and an assortative mixing feature. However, they show some differences in the hierarchical clustering feature. On observing the evolution results, three measures show a similar correlation with seismicity dynamics, but one measure shows a different result. These results show that different network construction methods will present some similarities and differences in network properties. This situation needs to be considered, especially when discussing a predictive indicator of seismicity.

Latency of Waveform Data Delivery from the Southern California Seismic Network during the 2019 Ridgecrest Earthquake Sequence and Its Effect on ShakeAlert

Abstract The occurrence of the 4–6 July 2019 Mw 6.4 and Mw 7.1 Ridgecrest earthquake sequence provided the first full-scale test of the network and telemetry readiness of the Southern California Seismic Network (SCSN), to support the ShakeAlert earthquake early warning (EEW) system in California. ShakeAlert is a U.S. Geological Survey (USGS)-led collaboration to detect earthquakes and, when possible, to alert the public before the arrival of the strongest shaking. The SCSN performed well in its regional monitoring role for both the 4 July Mw 6.4 and the 6 July Mw 7.1 earthquakes. In the EEW role, it provided timely delivery of 5 s of P-wave data to ShakeAlert, which issued its first alert 6.9 s after origin time. Data delivery at peak data volumes for many stations exhibited some latency, and, as a consequence, some data arrived too late for analysis by one of the EEW algorithms. We find that the average link bandwidth for each station was sufficient, because all waveform data were delivered automatically to the archive, but link capacity for many stations was insufficient for peak demand. We describe the performance of the data telemetry for the sequence, including cellular, radio, hybrid, and backhaul systems. Cellular-based telemetry systems maintained low latency throughout strong shaking and after, but some stations, even at great distances, experienced subsequent brief increases in latency. Performance of radio links depended mostly on the signal strength of the link, with short-distance direct shots to high-bandwidth backhaul systems showing no latency impact, whereas stations on some long distance or marginal quality links suffered latencies of tens or hundreds of seconds. Improvements are being implemented to move telemetry links onto USGS and partner high-bandwidth microwave systems, and to reduce dependency on less robust long-distance radio shots.

Application of a convolutional neural network for seismic phase picking of mining-induced seismicity

Monitoring mining-induced seismicity (MIS) can help engineers understand the rock mass response to resource extraction. With a thorough understanding of ongoing geomechanical processes, engineers can operate mines, especially those mines with the propensity for rock-bursting, more safely and efficiently. Unfortunately, processing MIS data usually requires significant effort from human analysts, which can result in substantial costs and time commitments. The problem is exacerbated for operations that produce copious amounts of MIS, such as mines with high-stress and/or extraction ratios. Recently, deep learning methods have shown the ability to significantly improve the quality of automated arrival-time picking on earthquake data recorded by regional seismic networks. However, relatively little has been published on applying these techniques to MIS. In this study, we compare the performance of a convolutional neural network (CNN) originally trained to pick arrival times on the Southern California Seismic Network (SCSN) to that of human analysts on coal-mine-related MIS. We perform comparisons on several coal-related MIS data sets recorded at various network scales, sampling rates and mines. We find that the Southern-California-trained CNN does not perform well on any of our data sets without retraining. However, applying the concept of transfer learning, we retrain the SCSN model with relatively little MIS data after which the CNN performs nearly as well as a human analyst. When retrained with data from a single analyst, the analyst-CNN pick time residual variance is lower than the variance observed between human analysts. We also compare the retrained CNN to a simpler, optimized picking algorithm, which falls short of the CNN's performance. We conclude that CNNs can achieve a significant improvement in automated phase picking although some data set-specific training will usually be required. Moreover, initializing training with weights found from other, even very different, data sets can greatly reduce the amount of training data required to achieve a given performance threshold.

2019 Ridgecrest Earthquake Reveals Areas of Los Angeles That Amplify Shaking of High-Rises

AbstractThe populace of Los Angeles, California, was startled by shaking from the M 7.1 earthquake that struck the city of Ridgecrest located 200 km to the north on 6 July 2019. Although the earthquake did not cause damage in Los Angeles, the experience in high-rise buildings was frightening in contrast to the shaking felt in short buildings. Observations from 560 ground-level accelerometers reveal large variations in shaking in the Los Angeles basin that occurred for more than 2 min. The observations come from the spatially dense Community Seismic Network (CSN), combined with the sparser Southern California Seismic Network and California Strong Motion Instrumentation Program networks. Site amplification factors for periods of 1, 3, 6, and 8 s are computed as the ratio of each station’s response spectral values combined for the two horizontal directions, relative to the average of three bedrock sites. Spatially coherent behavior in site amplification emerges for periods ≥3 s, and the maximum calculated site amplifications are the largest, by factors of 7, 10, and 8, respectively, for 3, 6, and 8 s periods. The dense CSN observations show that the long-period amplification is clearly, but only partially, correlated with the depth to basement. Sites with the largest amplifications for the long periods (≥3 s) are not close to the deepest portion of the basin. At 6 and 8 s periods, the maximum amplifications occur in the western part of the Los Angeles basin and in the south-central San Fernando Valley sedimentary basin. The observations suggest that the excitation of a hypothetical high-rise located in an area characterized by the largest site amplifications could be four times larger than in a downtown Los Angeles location.

Tracking earthquake sequences in real time: application of Seismicity-Scanning based on Navigated Automatic Phase-picking (S-SNAP) to the 2019 Ridgecrest, California sequence

SUMMARYRecent improvements in seismic data processing techniques have enhanced our ability to detail the evolution of major earthquake sequences in space and time. One such advance is new scanning algorithms that allow large volumes of waveform data to be analysed automatically, removing human biases and inefficiencies that inhibit standardized monitoring. The Seismicity-Scanning based on Navigated Automatic Phase-picking (S-SNAP) workflow has previously been shown to be capable of producing high-quality earthquake catalogues for injection-induced seismicity monitoring. In this study, we modify the original S-SNAP workflow to enable it to delineate the spatiotemporal distribution of major earthquake sequences in real time. We apply it to the 2019 Ridgecrest, southern California earthquake sequence, which culminated in an Mw 6.4 foreshock on July 4 and an Mw 7.1 main shock on July 6 and generated tens of thousands of smaller earthquakes. Our catalogue—which spans the period 2019 June 1 to July 16—details the spatiotemporal evolution of the sequence, including early foreshocks on July 1 and accelerating foreshocks on July 4, a seismicity gap before the main shock around its epicentre, seismicity on discrete structures within a broad fault zone and triggered earthquakes outside the main fault zone. We estimate the accuracy and false detection rate of the S-SNAP catalogue based on the reviewed catalogue reported by Southern California Seismic Network (SCSN) and our own visual inspection. We demonstrate the advantages of S-SNAP over a generalized automatic earthquake monitoring software, Seiscomp3, and a customized real-time earthquake information system for southern California, TriNet. In comparison, the S-SNAP catalogue contains five times more events than the Seiscomp3 catalogue and 1.4–2.2 times as many events per hour as the TriNet catalogue at most times. In addition, S-SNAP is more likely to solve phase association ambiguities correctly and provide a catalogue with consistent quality through time. S-SNAP would be beneficial to both routine network operations and the earthquake review process.

Optimally Oriented Remote Triggering in the Coso Geothermal Region

AbstractDefining the nature of faulting, the stress needed to trigger earthquakes, and the stress state of faults remains fundamental to understanding the earthquake cycle. Studying remote dynamic triggering allows us to probe faults to systemically address the fundamental physical mechanisms of faulting. Using 13 years of data (2004–2016) from the EarthScope USArray Transportable Array and the Southern California Seismic Network, we search for remotely triggered seismicity in an extended region encompassing the Coso Geothermal Field (CGF+), California. We first apply a short term to long‐term average ratio detector to high‐pass (5 Hz) filtered waveforms spanning ±5 hr encompassing 211 M ≥ 7 global earthquakes. We visually inspect these waveforms to identify uncatalogued local earthquakes. We use the augmented local earthquake catalog to investigate remote earthquake triggering in the CGF+ region. Of the 211 remote mainshocks, we find 32 (15%) produce a statistically significant increase in seismicity following the P wave arrival in CGF+. An additional nine mainshocks had local earthquakes coincident with the surface waves passage but lacked a significant rate increase. Of the 41 (19%) triggering mainshocks, 28 and 13 exhibit instantaneous and delayed triggering, respectively. We find no correlation between triggering and mainshock depth, peak dynamic stress, nor mainshock focal mechanism type. Instead, results suggest the CGF+ may be optimally oriented for remote triggering from mainshocks in the West Pacific and South America and propose the reason for this is that the transient dynamic stresses align favorably with the local stress field (most compressive horizontal stress, SHmax) to promote triggering.

Evaluation of Ground-Motion Models for USGS Seismic Hazard Models Using Near-Source Instrumental Ground-Motion Recordings of the Ridgecrest, California, Earthquake Sequence

ABSTRACTWe evaluated the performance of 12 ground-motion models (GMMs) for earthquakes in the tectonically active shallow crustal region of southern California using instrumental ground-motion observations from the 2019 Ridgecrest, California, earthquake sequence (Mw 4.0–7.1). The sequence was well recorded by the Southern California Seismic Network and rapid response portable aftershock monitoring stations. Ground-motion recordings of this size and proximity are rare, valuable, and independent of GMM development, allowing us to evaluate the predictive powers of GMMs. We first compute total residuals and compare the probability density functions, means, and standard deviations of the observed and predicted ground motions. Next we use the total residuals as inputs to the probabilistic scoring method (log-likelihood [LLH]). The LLH method provides a single score that can be used to weight GMMs in the U.S. Geological Survey (USGS) National Seismic Hazard Model (NSHM) logic trees. We also explore GMM performance for a range of earthquake magnitudes, wave propagation distances, and site characteristics. We find that the Next Generation Attenuation West-2 (NGAW2) active crust GMMs perform well for the 2019 Ridgecrest, California, earthquake sequence and thus validate their use in the 2018 USGS NSHM. However, significant ground-motion residual scatter remains unmodeled by NGAW2 GMMs due to complexities such as local site amplification and source directivity. Results from this study will inform logic-tree weights for updates to the USGS National NSHM. Results from this study support the use of nonergodic GMMs that can account for regional attenuation and site variations to minimize epistemic uncertainty in USGS NSHMs.

Weighted Template-Matching Algorithm (WTMA) for Improved Foreshock Detection of the 2019 Ridgecrest Earthquake Sequence

ABSTRACT The template-matching algorithm (TMA) has become an important tool for detecting small and/or unconventional earthquakes, and newly detected seismic events have improved our understanding of earthquake physics, regional tectonics, and geological hazards. Standard TMA-based detection methods do not take into account the fact that the template waveforms themselves are contaminated with noise. In this study, we propose a weighted template-matching algorithm (WTMA), in which the normalized cross-correlation time series are weighted according to the signal-to-noise ratios of the corresponding template waveforms before they are being stacked for event picking. We present results from an extensive collection of numerical experiments to demonstrate that WTMA is capable of improving the detection rate of small to tiny (e.g., ML<0.0) earthquakes while reducing the computational cost associated with the standard TMA-based detection methods. The application of our WTMA to the continuous waveform recordings before the Mw 6.4 earthquake in the 2019 Ridgecrest sequence allowed us to discover ∼200 foreshocks, a larger number than the ∼150 foreshocks detected using the standard TMA, and significantly larger than the 17 foreshocks documented in the Southern California Seismic Network (SCSN) catalog. Relocated hypocenters of the foreshocks suggest a complex faulting pattern that is consistent with the hierarchical interlocked orthogonal faulting that characterized the main Ridgecrest sequence. Magnitudes of the foreshocks obtained through careful waveform amplitude calibration provided a more robust estimate of the magnitude–frequency distribution, which reduced the minimum magnitude threshold and increased the b-value, as compared with those obtained from the SCSN catalog.

Detailed Investigation of the Foreshock Sequence of the 2010 Mw7.2 El Mayor‐Cucapah Earthquake

AbstractForeshocks can provide valuable information about possible nucleation process of a mainshock. However, their physical mechanisms are still under debate. In this study, we present a comprehensive analysis of the earthquake sequence preceding the 2010 Mw7.2 El Mayor‐Cucapah mainshock, including waveform detection of missing smaller events, relative relocation, and source parameter analysis. Based on a template matching method, we find a tenfold increase in the number of earthquakes than reported in the Southern California Seismic Network catalog. The entire sequence exhibits nearly continuous episodes of foreshocks that can be loosely separated into two active clusters. Relocated foreshocks show several seismicity streaks at depth, with a consistently active cluster at depths between 14 and 16 km where the mainshock was nucleated. Stress drop measurements from a spectral ratio approach based on empirical Green's functions show a range between 3.8 and 41.7 MPa with a median of 13.0 MPa and no clear temporal variations. The relocation results, together with the source patches estimated from earthquake corner frequencies, revealed a migration front toward the mainshock hypocenter within last 8 hr and a chain of active burst immediately 6 min prior to the mainshock. Our results support combined effects of aseismic slip and cascading failure on the evolution of foreshocks.