Earthquake Early Warning Research Articles

Take Advantage of Earthquake Early Warning Systems

Take Advantage of Earthquake Early Warning Systems

Cumulative absolute velocity prediction for earthquake early warning with deep learning

AbstractA rapid and accurate estimate of earthquake damage is a key component in a successful earthquake early warning (EEW) system. The cumulative absolute velocity (CAV) is an important and widely used parameter to measure ground motion intensity, but it cannot be correctly estimated via the traditional approach with the limited information available in typical EEW systems. Therefore, current EEW systems cannot effectively use CAV to predict earthquake damage. Herein, a CAV prediction model (DLcav) based on convolutional neural networks was proposed for EEW systems. DLcav is an end‐to‐end solution to continuously predict CAV using arriving seismic waves of increasing length and supplemented with additional auxiliary information. The effectiveness of DLcav to predict CAV was tested based on Japanese ground motion records, and the generalization ability of DLcav was assessed using the ground motion records from Chile. The results demonstrate that DLcav can rapidly predict CAV with good accuracy, which will help better estimate earthquake damage in EEW systems.

Monitoring seismicity in the southern Sichuan Basin using a machine learning workflow

Monitoring seismicity in real time provides significant benefits for timely earthquake warning and analyses. In this study, we propose an automatic workflow based on machine learning (ML) to monitor seismicity in the southern Sichuan Basin of China. This workflow includes coherent event detection, phase picking, and earthquake location using three-component data from a seismic network. By combining PhaseNet, we develop an ML-based earthquake location model called PhaseLoc, to conduct real-time monitoring of the local seismicity. The approach allows us to use synthetic samples covering the entire study area to train PhaseLoc, addressing the problems of insufficient data samples, imbalanced data distribution, and unreliable labels when training with observed data. We apply the trained model to observed data recorded in the southern Sichuan Basin, China, between September 2018 and March 2019. The results show that the average differences in latitude, longitude, and depth are 5.7 ​km, 6.1 ​km, and 2 ​km, respectively, compared to the reference catalog. PhaseLoc combines all available phase information to make fast and reliable predictions, even if only a few phases are detected and picked. The proposed workflow may help real-time seismic monitoring in other regions as well.

Enhancing earthquake detection from distributed acoustic sensing data by coherency measure and moving-rank-reduction filtering

Distributed acoustic sensing (DAS) enables the recording of earthquake signals at an unprecedented low cost with high durability using fiber optic cables. It is often compromised by relatively lower data quality in single-channel measurements compared with traditional seismic receivers but compensated by high-density recordings from hundreds of channels per earthquake event. The multichannel nature of DAS data sets facilitates the applications of well-developed coherency-based denoising methods arising from reflection seismology for detecting more earthquakes. We first introduce a coherency measure for detecting earthquake signals from DAS data sets. Then, we apply a moving-rank-reduction (MRR) filter to enhance the DAS data quality so as to improve the earthquake detectability. The MRR filter is tailored from a rank-reduction filter that is widely used in processing multichannel reflection seismic data. We find that a simple band-pass or median filter is incapable of revealing weak signals generated from small-magnitude or far-away earthquake events, whereas the MRR filter significantly improves the signal-to-noise ratio that enables the detection of those weak signals. We apply the MRR method and the coherency measure on the San Andreas Fault Observatory at Depth DAS data sets for denoising and earthquake detection. As a result, our framework detects all 31 cataloged events, outperforming the previous detection of 25 events using the same data set.

Kalman Filter, ANN-MLP, LSTM and ACO Methods Showing Anomalous GPS-TEC Variations Concerning Turkey’s Powerful Earthquake (6 February 2023)

On 6 February 2023, at 1:17:34 UTC, a powerful Mw = 7.8 earthquake shook parts of Turkey and Syria. Investigating the behavior of different earthquake precursors around the time and location of this earthquake can facilitate the creation of an earthquake early warning system in the future. Total electron content (TEC) obtained from the measurements of GPS satellites is one of the ionospheric precursors, which in many cases has shown prominent anomalies before the occurrence of strong earthquakes. In this study, five classical and intelligent anomaly detection algorithms, including median, Kalman filter, artificial neural network (ANN)-multilayer perceptron (MLP), long short-term memory (LSTM), and ant colony optimization (ACO), have been used to detect seismo-anomalies in the time series of TEC changes in a period of about 4 months, from 1 November 2022 to 17 February 2023. All these algorithms show outstanding anomalies in the period of 10 days before the earthquake. The median method shows clear TEC anomalies in 1, 2 and, 3 days before the event. Since the behavior of the time series of a TEC parameter is complex and nonlinear, by implementing the Kalman filter method, pre-seismic anomalies were observed in 1, 2, 3, 5, and 10 days prior to the main shock. ANN as an intelligent-method-based machine learning also emphasizes the abnormal behavior of the TEC parameter in 1, 2, 3, 6, and 10 days before the earthquake. As a deep-learning-based predictor, LSTM indicates that the TEC value in the 10 days prior to the event has crossed the defined permissible limits. As an optimization algorithm, the ACO method shows behavior similar to Kalman filter and MLP algorithms by detecting anomalies 3, 7, and 10 days before the earthquake. In a previous paper, the author showed the findings of implementing a fuzzy inference system (FIS), indicating that the magnitude of the mentioned powerful earthquake could be predicted during about 9 to 1 day prior to the event. The results of this study also confirm the findings of another study. Therefore, considering that different lithosphere–atmosphere–ionosphere (LAI) precursors and different predictors show abnormal behavior in the time period before the occurrence of large earthquakes, the necessity of creating an earthquake early warning system based on intelligent monitoring of different precursors in earthquake-prone areas is emphasized.

Sensitivity limits for strain detection of hypothetical remote fluid-induced earthquakes (Mw ≥ 4): a case study in Taiwan

Capturing and quantifying the timing of remotely triggered earthquakes and understanding the physical processes responsible for this delay represent major challenges in earthquake forecasting. In this study, we propose a physical framework for the integration of borehole strainmeter observations for the investigation of remote triggering of moderate to large earthquakes (Mw ≥ 4) in Taiwan. Based on the time-delay computation between regional events and global earthquakes, we establish a selection of earthquakes showing fault zone properties (hydraulic diffusivity and nucleation length) that may be compatible with a magnitude-dependent fluid-induced nucleation process. Using theoretical fault zones parameters, we calculate the evolution of fluid pressure transiting along the nucleation region under the assumption of a one-dimensional, homogeneous poroelastic medium. Pore pressure levels reached before earthquake rupture are ranging from about 0.02 kPa to 3 kPa in the case of teleseismic wave-induced elastic pressure ranging from 0.15 kPa to 27.3 kPa. To compute the time-dependent evolution of deformation generated by a remote diffusing pressure front, we model the nucleation region using the analogue volcano source represented by a horizontal circular crack, and calculate synthetic dilatation at the strainmeter location from displacements using a finite-difference approach. In general, predictions are about two to four orders of magnitude smaller than observations (∼ 10–5 to 10–3 nϵ). Therefore, this suggests that detection of pore pressure-related deformation would have required change of volume in the nucleation region that is at least one order of magnitude larger than for the hypothetical cases considered here. The study represents the first attempt to analyze strain time-series for detecting pre-earthquake strain anomalies related to fluid-induced earthquakes and illustrates the challenge for detecting and characterizing intermediate-to far-field earthquake precursors caused by fluid flow in active regions.

Detecting the Unseen: Understanding the Mechanisms and Working Principles of Earthquake Sensors.

The application of movement-detection sensors is crucial for understanding surface movement and tectonic activities. The development of modern sensors has been instrumental in earthquake monitoring, prediction, early warning, emergency commanding and communication, search and rescue, and life detection. There are numerous sensors currently being utilized in earthquake engineering and science. It is essential to review their mechanisms and working principles thoroughly. Hence, we have attempted to review the development and application of these sensors by classifying them based on the timeline of earthquakes, the physical or chemical mechanisms of sensors, and the location of sensor platforms. In this study, we analyzed available sensor platforms that have been widely used in recent years, with satellites and UAVs being among the most used. The findings of our study will be useful for future earthquake response and relief efforts, as well as research aimed at reducing earthquake disaster risks.

Detecting strain with a fiber optic cable on the seafloor offshore Mount Etna, Southern Italy

Oceans cover more than 70 percent of the Earth's surface making it difficult and costly to deploy modern seismological instruments here. The rapidly expanding global network of submarine telecom cables offers tremendous possibilities for seismological monitoring using laser light. Recent pioneer studies have demonstrated earthquake detection using lasers in onland and submarine fiber optic cables. However, permanent strain at the seafloor has never before been measured directly as it happens. With this aim, we deployed a dedicated 6-km-long fiber optic strain cable, offshore Catania Sicily, in 2000 m water depth, and connected it to a 29-km long electro-optical cable for science use. We report here that deformation of the cable equivalent to a total elongation of 2.5 cm was observed over a 21-month period (from Oct. 2020 to Jul. 2022). Brillouin laser reflectometry observations over the first 10 months indicate significant strain (+25 to +40 microstrain) at two locations where the cable crosses an active strike-slip fault on the seafloor, with most of the change occurring between 19 and 21 Nov. 2020. The cause of the strain could be fault slip or seabottom currents. During the following 11 months, the strain amplitude increased to +45 to +55 microstrain, affecting a longer portion of the cable up to 500 m to either side of the first fault crossing. A sandbag experiment performed on the distal portion of the cable (3.2–6.0 km) starting Sept. 2021 demonstrates how the fiber optic cable deforms in response to an applied load and how the deformation signal partially dissipates over time due to the elastic properties of the cable. These preliminary results are highly encouraging for the use of BOTDR (Brillouin Optical Time Domain Reflectometry) laser reflectometry as a technique to detect strain at the seafloor in near real time and to monitor the structural health of submarine cables.

Shaking Table Design for Testing Earthquake Early Warning Systems

The unpredictability in time of seismic activities and the dependence of tectonic movements on a multitude of factors challenges specialists to identify the most accurate related methods to avoid catastrophes associated with hazards. Early warning systems are critical in reducing negative effects in the case of an earthquake with a magnitude above 5 MW. Their precision is all the better as they corroborate and transmit more information collected from the regional or on-site sensory nodes to a central unit that discloses events and estimates the epicentral location, earthquake magnitude, or ground shaking amplitude. The shaking table is the proper instrument for evaluating an early warning systems’ dynamic response and performance under specific vibration conditions. To this issue, the paper presents a laboratory single-axis shaking table with a small-scale, low-cost design and an accurate displacement control. Experiments based on a suite of 12 real earthquakes provided results with very small errors related to similar models, bearing out the designed shaking table is suitable for early earthquake warning system response testing for high magnitude earthquakes.

IPP: A Seismic P-Wave’s Detection Method Based on the Initial Power of the P-Wave

ABSTRACT The detection of P-wave is essential for the earthquake early warning system. How to detect the P-wave under noise circumstance accurately and rapidly is a challenge work. Based on the relative power of the seismic signal, a new parameter, the initial power of the P-wave (IPP), is proposed to detect the P-wave. By using the relative power of the seismic signal, the interference of the background noise is eliminated, and the initial power of the P-wave and sigma criterion are both applied to detect the P-wave. Through simulation analysis and example tests whose magnitudes are varied from Ms4.0 to Ms8.0, both the misjudgment and the missing rate of the IPP method are 0, and the time deviation within 0.2 s can reach 78.95%. Comparing with the STA/LTA, STP/LTP and PhaseNet methods, the IPP method proposed in this paper has strong robustness and high accuracy, which represents the initial power characteristic of P-wave and can effectively eliminate the interference of background noise.

En echelon faults reactivated by wastewater disposal near Musreau Lake, Alberta

SUMMARY We use machine-learning and cross-correlation techniques to enhance earthquake detectability by two magnitude units for the earthquake sequence near Musreau Lake, Alberta, which is induced by wastewater disposal. This deep catalogue reveals a series of en echelon ∼N–S oriented strike-slip faults that are favourably oriented for reactivation. These faults require only ∼0.6 MPa overpressure for triggering to occur. Earthquake activity occurs in bursts, or episodes; episodes restricted to the largest fault tend to have earthquakes starting near the southern end (distant from injectors) and progressing northwards (towards the injectors). While most events are concentrated along these ∼N–S oriented faults, we also delineate smaller faults. Together, these findings suggest pore pressure as the triggering mechanism, where a time-dependent increase in pore pressure likely caused these faults to progressively reawaken. Analysis of the ‘next record-breaking event’, a statistical model that forecasts the sequencing of earthquake magnitudes, suggests that the next largest event would be ML ∼4.3. The seismically illuminated length of the largest fault indicates potential magnitudes as large as Mw 5.3.

Bayesian Updating and Model Class Selection for Magnitude Estimation in Earthquake Early Warning: Application to Earthquake Sequences in Sichuan Region, China

ABSTRACT A key component of earthquake early warning system is the development of an accurate and robust predictive model relating the recorded waveform to seismicity and ground-shaking characteristics. This article presents a probabilistic Bayesian inference methodology to address relevant sources of uncertainty in the development of such predictive models. We investigate the calibration of models for the earthquake magnitude based on the maximum predominant period, considering the Sichuan region of Southwestern China as a case study. Established approaches for developing predictive models in this context adopt deterministic tools for some aspects of this development. They consider a linear regression calibrated typically through least squares optimization and frequently utilize the mean observations for each event averaging across the data available from different stations. The proposed Bayesian learning accommodates the following improvements: a model class selection is established, comparing across different candidate models to promote the most appropriate from accuracy and robustness perspectives; the full posterior distribution of the model parameters is identified, quantifying relevant uncertainties in their values; a heteroscedastic model is considered for the estimation error variance; and the observations are separately considered at the calibration stage. Each of these improvements ultimately addresses a different source of uncertainty impacting the predictive model development. We utilize transitional Markov chain Monte Carlo for obtaining samples from the posterior and for calculating the evidence to perform the model class selection. Different regression models are examined, and the Bayesian-based model identification is compared against the common least squares identification approach. Results show the value added by comparing across the different models and by considering a heteroskedastic variance model, offering insights into the advantages of Bayesian-based predictive models in earthquake early warning applications.

Earth quake early warning systems (EEWSs) and their applicability in modern research – a review

Sending alerts or warnings about ground tremors is known as an earthquake early warning system (EEWS). It is issued after the earthquake has already started to cause observable ground motion, setting it apart from the forecast. The interval between detection and the moment a user feels ground motion is known as the available warning time. Therefore, the potential warning window can be anywhere between seconds and minutes. Similarly to this, users must be able to act quickly because only seconds to minutes are available to collect, interpret, and transmit seismological data and warning signals. By reviewing previously published, high--quality literature, the review article offers clear insight into EEWSs and how they operate. This will support researchers in understanding the current technologies used to send signals to save lives from earthquakes at the appropriate time. The effects of earthquakes and how they have been tracked using EEWSs are also covered in the article.

Stable operation process of earthquake early warning system based on machine learning: trial test and management perspective

Earthquake Early Warning (EEW) is an alert system, based on seismic wave propagation theory, to reduce human casualties. EEW systems mainly utilize technologies through both network-based and on-site methods. The network-based method estimates the hypocenter and magnitude of an earthquake using data from multiple seismic stations, while the on-site method predicts the intensity measures from a single seismic station. Therefore, the on-site method reduces the lead time compared to the network-based method but is less accurate. To increase the accuracy of on-site EEW, our system was designed with a hybrid method, which included machine learning algorithms. At this time, machine learning was used to increase the accuracy of the initial P-wave identification rate. Additionally, a new approach using a nearby seismic station, called the 1+ α method, was proposed to reduce false alarms. In this study, an on-site EEW trial operation was performed to evaluate its performance. The warning cases for small and large events were reviewed and the possibility of stable alert decisions was confirmed.

Earthquake Magnitude With DAS: A Transferable Data‐Based Scaling Relation

AbstractDistributed Acoustic Sensing (DAS) is a promising technique to improve the rapid detection and characterization of earthquakes. Previous DAS studies mainly focus on the phase information but less on the amplitude information. In this study, we compile earthquake data from two DAS arrays in California, USA, and one submarine array in Sanriku, Japan. We develop a data‐driven method to obtain the first scaling relation between DAS amplitude and earthquake magnitude. Our results reveal that the earthquake amplitudes recorded by DAS in different regions follow a similar scaling relation. The scaling relation can provide a rapid earthquake magnitude estimation and effectively avoid uncertainties caused by the conversion to ground motions. Our results show that the scaling relation appears transferable to new regions with calibrations. The scaling relation highlights the great potential of DAS in earthquake source characterization and early warning.

Deep Long‐Period Earthquakes at Akutan Volcano From 2005 to 2017 Better Track Magma Influxes Compared to Volcano‐Tectonic Earthquakes

AbstractBoth volcano‐tectonic (VTs) and deep long‐period earthquakes (DLPs) have been documented at Akutan Volcano, Alaska and may reflect different active processes helpful for eruption forecasting. In this study, we perform high‐resolution earthquake detection, classification, and relocation using seismic data from 2005 to 2017 to investigate their relationship with underlying magmatic processes. We find that the 2,787 VTs and 787 DLPs are concentrated above and below the inferred magma reservoir respectively. They both are clustered as swarms and occur preferentially during inflation episodes with no spatial migrations. However, moment release rates of DLP swarms show a stronger correlation with inflation and their low‐frequency content is likely a source instead of a path effect. Therefore, we infer that DLPs are directly related to unsteady magma movement through a complex pathway. In comparison, repeating events are observed in VTs. Thus, we conclude that they represent fault rupture triggered by magma/fluid movement or larger earthquakes.

Social Science Contributions to Earthquake Warnings: Commemorating the Work of Dennis S. Mileti

The pioneering work of Dennis Mileti and those with whom he collaborated have established a sound basis in understanding how human beings behave when encountering hazard warnings. His work greatly expanded our understanding of the social environment into which warnings are issued and provided clear guidelines on how warnings should be delivered to optimize behavioral compliance; how warning message content can be construed to motivate protective actions; and why people may fail to comply or delay actions after receipt of a warning. This paper addresses Mileti’s work in earthquake warnings, first in the context of earthquake predictions and later, as optimism waned that accurate earthquake predictions were possible, as his insights were applied to newer approaches in the form of earthquake early warnings and operational earthquake forecasting. This paper will first briefly trace the historical development of approaches to warn of a potentially damaging earthquake because it is important to place Mileti’s work in the historical context in which it was carried out. It then assesses the manner in which the current methods of earthquake warning for short- and intermediate-term time frames via earthquake early warning and operational earthquake forecasting have benefited from Mileti’s very significant observations and insights. His work in evaluating and clarifying the social environment in which warnings are issued is an important and enduring element of his legacy, has significantly influenced how hazard warnings are now formulated, and has greatly shaped the work of those of us who carry on the study of hazard warning systems.

Application of deep learning for seismicity analysis in Ghana

We present the characterization of regional seismicity in Ghana by processing the Ghana Digital Seismic Network (GHDSN) data set recorded between September 2012 and April 2014, implementing deep learning (DL). Local earthquakes are detected in this dataset using EQTransformer, a DL model with a hierarchical attentive mechanism (HAM) for simultaneous earthquake detection and P- and S-phase picking. A Conservative Strategy (CS) is devised to detect the missing phases and to associate the detected phases to circumvent the false-negative issue of EQTrans-former processing low signal-to-noise ratio (SNR) seismograms. We performed a joint inversion by grid search in 1D velocity model space and simultaneous inversion for the hypocentral parameters, incorporating 559 detected arrival times (292 P and 267 S phases). The results obtained by velocity inversion contain thicknesses of 1, 13, 8, 13, and 10 km, from the surface to a depth of 45 km, with Vp = 5.9, 6.1, 6.3, 6.5, 6.9, and 7.2 km/s, respectively. The updated velocities for the first and last layers are 6% and 26% and the Vp/Vs (=1.70) is 3.03% higher than the previously reported values. A total number of 73 earthquakes with a local magnitude of 2.5 < Ml < 3.9 are located, comprising four main clusters of events, showing a high correlation with the mapped faults zones. The hypocentral depth distribution is mainly in the range of 7-15 km, confined to the upper crust in the region. No specific seismic activity in the eastern branch of Coastal Boundary Fault (CBF) and the continuation of Romanche Fracture Zone (RFZ) in the study period was observed, casting further doubt on the activity of this branch and the hypothesis of stress transfer from RFZ to southern Ghana. The results reinforce the intraplate nature of the tectonic activities in the region. Finally, an updated seismic catalog up to April 2022 is presented for Ghana by incorporating all reported catalogs and combining the newly detected events.

A Specific Earthquake Processing Workflow for Studying Long‐Lived, Explosive Volcanic Eruptions With Application to the 2008 Okmok Volcano, Alaska, Eruption

AbstractBy providing unrivaled resolution in both time and space, volcano seismicity helps to chronicle and interpret eruptions. Standard earthquake detection methods are often insufficient as the eruption itself produces continuous seismic waves that obscure earthquake signals. We address this problem by developing an earthquake processing workflow specific to a high‐noise volcanic environment and applying it to the explosive 2008 Okmok Volcano eruption. This process includes applying single‐channel template matching combined with machine‐learning and fingerprint‐based techniques to expand the existing earthquake catalog of the eruption. We detected an order of magnitude more earthquakes, then located, relocated, determined locally calibrated magnitudes, and classified the events in the enhanced catalog. This new high‐resolution earthquake catalog increases the number of observations by about a factor of 10 and enables the detailed spatiotemporal seismic analysis during a large eruption.

Application of MEMS Data to Fast Inversion of Rupture Process: Tests with Recordings from the IRREEW Network

AbstractThe China Earthquake Administration established a network for intensity rapid report and earthquake early warning (IRREEW) in 2016–2020, which consists of approximately 5000 conventional strong-motion and approximately 10,000 low-cost micro-electro-mechanical system (MEMS) seismometers. These seismographs, particularly MEMS stations, can provide a large amount of near-field waveform data suitable for rapid source inversion. Compared with conventional strong-motion data, MEMS recordings have rarely been used in the previous source inversions, because the MEMS technology is newly applied in earthquake monitoring, and the seismograph has a relatively lower signal-to-noise ratio and more severe baseline shifts. However, from waveform comparisons at collocated MEMS and strong-motion stations, we find that they are highly consistent with each other, particularly at frequencies above 0.04 Hz. To explore the application prospect of MEMS data to source inversion, we inverted both MEMS and strong-motion data for three strong earthquakes recorded by the IRREEW network during 2021–2022 to determine their rupture processes. In applications to the 2021 Mw 6.1 Yangbi earthquake, the 2022 Mw 6.6 Menyuan earthquake, and the 2022 Mw 6.6 Luding earthquake, the MEMS data equally well constrain the rupture model. The resulting source information, including the moment magnitude, rupture direction, and rupture dimension, are consistent with those obtained from the strong-motion inversions. Because the low-cost MEMS instruments can be deployed densely around seismically active regions, they can provide urgent waveform data for rapid determination of rupture process, which is crucial for simulation of strong ground motions, and assessments of earthquake and related disasters.