Identification Of Seismic Events Research Articles

Ensemble Learning Improves the Efficiency of Microseismic Signal Classification in Landslide Seismic Monitoring.

A deep-seated landslide could release numerous microseismic signals from creep-slip movement, which includes a rock-soil slip from the slope surface and a rock-soil shear rupture in the subsurface. Machine learning can effectively enhance the classification of microseismic signals in landslide seismic monitoring and interpret the mechanical processes of landslide motion. In this paper, eight sets of triaxial seismic sensors were deployed inside the deep-seated landslide, Jiuxianping, China, and a large number of microseismic signals related to the slope movement were obtained through 1-year-long continuous monitoring. All the data were passed through the seismic event identification mode, the ratio of the long-time average and short-time average. We selected 11 days of data, manually classified 4131 data into eight categories, and created a microseismic event database. Classical machine learning algorithms and ensemble learning algorithms were tested in this paper. In order to evaluate the seismic event classification performance of each algorithmic model, we evaluated the proposed algorithms through the dimensions of the accuracy, precision, and recall of each model. The validation results demonstrated that the best performing decision tree algorithm among the classical machine learning algorithms had an accuracy of 88.75%, while the ensemble algorithms, including random forest, Gradient Boosting Trees, Extreme Gradient Boosting, and Light Gradient Boosting Machine, had an accuracy range from 93.5% to 94.2% and also achieved better results in the combined evaluation of the precision, recall, and F1 score. The specific classification tests for each microseismic event category showed the same results. The results suggested that the ensemble learning algorithms show better results compared to the classical machine learning algorithms.

Identification of natural and technogenic seismic events by energy characteristics

One of the key problems of seismic monitoring is the identification of earthquakes and signals from technogenic sources detected by a network of seismic stations. In peacetime, technogenic events are mainly associated with industrial mining developments, however, with the beginning of russia's full-scale aggression against sovereign Ukraine, thousands of seismic signals from explosions as a result of missile, aircraft, artillery strikes were registered by the seismological network of the Main Center of Special Monitoring of the State Space Agency of Ukraine. This significantly complicates the process of assessing seismicity and makes the question of determining the nature of registered events extremely relevant. Based on the analysis of seismic signals, the relationships between energy classes (K), magnitudes (mb), maximum amplitudes of longitudinal volumetric phases , and yields (Y) of explosions in TNT equivalent in Kyiv, Zhytomyr, Vinnytsia, Khmelnytsky, Chernihiv regions. Energy characteristics can be used to identify the nature of seismic events, and the results of the analysis of the ratios , , make it possible to yield estimate of explosions in TNT equivalent and determine the probable types of ammunition based on the received data. The energy from the signal source in the case of an explosive event can be determined additionally by infrasound data, the presence of an acoustic wave serves as an additional criterion for identifying the event. At the same time, energy characteristics make it possible to identify natural sources, an example of which is the tectonic earthquake of May 26, 2023 in the Poltava region.One of the key problems of seismic monitoring is the identification of earthquakes and signals from technogenic sources detected by a network of seismic stations. In peacetime, technogenic events are mainly associated with industrial mining developments, however, with the beginning of russia's full-scale aggression against sovereign Ukraine, thousands of seismic signals from explosions as a result of missile, aircraft, artillery strikes were registered by the seismological network of the Main Center of Special Monitoring of the State Space Agency of Ukraine. This significantly complicates the process of assessing seismicity and makes the question of determining the nature of registered events extremely relevant. Based on the analysis of seismic signals, the relationships between energy classes (K), magnitudes (mb), maximum amplitudes of longitudinal volumetric phases , and yields (Y) of explosions in TNT equivalent in Kyiv, Zhytomyr, Vinnytsia, Khmelnytsky, Chernihiv regions. Energy characteristics can be used to identify the nature of seismic events, and the results of the analysis of the ratios , , make it possible to yield estimate of explosions in TNT equivalent and determine the probable types of ammunition based on the received data. The energy from the signal source in the case of an explosive event can be determined additionally by infrasound data, the presence of an acoustic wave serves as an additional criterion for identifying the event. At the same time, energy characteristics make it possible to identify natural sources, an example of which is the tectonic earthquake of May 26, 2023 in the Poltava region.

Seismic Event Identification Based on a Generative Adversarial Network and Support Vector Machine

Identifying appropriate seismic events is the primary precondition for conducting meaningful analysis in seismological research. The successful creation of a method to automatically identify earthquakes from large amounts of data has become increasingly vital, especially with the construction of seismic stations, the collection of extensive seismic data, and the development of earthquake early warning (EEW) systems. To accurately identify seismic events, a combined model based on a generative adversarial network (GAN) and a support vector machine (SVM) is proposed to distinguish between earthquakes and microtremors. We first use 52,537 strong ground motion records from Japan to train a GAN and extract the characteristics of P waves and then use an SVM to discriminate seismic events in the testing set, thereby transforming the complex seismic event identification into a simpler binary classification of earthquakes and microtremors. The results illustrate that the combined model can achieve accuracies of 99.74% for P waves and 99.93% for microtremors, which represents an increase in accuracy of 14.13% compared with the traditional short-term averaging/long-term averaging (STA/LTA) method. Additionally, 98% of the local seismic events in the Great East Japan earthquake were identified. Therefore, the combined model has a wide range of applications in EEW and earthquake monitoring.

Spatial and temporal multiplet analysis for identification of dominant fluid migration path at The Geysers geothermal field, California

Multiplet analysis is based on the identification of seismic events with very similar waveforms which are used then to enhance seismological analysis e.g. by precise relocation of sources. In underground fluid injection conditions, it is a tool frequently used for imaging of subsurface fracture system. We identify over 150 repeatedly activated seismic sources within seismicity cluster induced by fluid injection in NW part of The Geysers geothermal field (California). Majority of multiple events (ME) occur along N–S oriented planar structure which we interpret as a fault plane. Remaining ME are distributed along structures interpreted as fractures, forming together a system of interconnected cracks enabling fluid migration. Temporal analysis reveals that during periods of relatively low fluid injection the proportion of ME to non-multiple events is higher than during periods of high injection. Moreover, ME which occur within the fault differ in activity rate and source properties from ME designating the fractures and non-multiple events. In this study we utilize observed differences between ME occurring within various structures and non-multiple events to describe hydraulic conditions within the reservoir. We show that spatial and temporal analysis of multiplets can be used for identification and characterization of dominant fluid migration paths.

A Comodulation Analysis of Atmospheric Energy Injection Into the Ground Motion at InSight, Mars

AbstractSeismic observations involve signals that can be easily masked by noise injection. For the NASA Mars lander InSight, the atmosphere is a significant noise contributor, impeding the identification of seismic events for two‐thirds of a Martian day. While the noise is below that seen at even the quietest sites on Earth, the amplitude of seismic signals on Mars is also considerably lower, requiring an understanding and quantification of environmental injection at unprecedented levels. Mars’ ground and atmosphere are a continuously coupled seismic system, and although atmospheric functions are of distinct origins, the superposition of these noise contributions is poorly understood, making separation a challenging task. We present a novel method for partitioning the observed signal into seismic and environmental contributions. Atmospheric pressure and wind fluctuations are shown to exhibit temporal cross‐frequency coupling across multiple bands, injecting noise that is neither random nor coherent. We investigate this through comodulation, quantifying the synchrony of the seismic motion, wind and pressure signals. By working in the time‐frequency domain, we discriminate between the different origins of underlying processes and determine the site's environmental sensitivity. Our method aims to create a virtual vault at InSight's landing site on Mars, shielding the seismometers with effective postprocessing in lieu of a physical vault. This allows us to describe the environmental and seismic signals over a sequence of sols, to quantify the wind and pressure injection and estimate the seismic content of possible marsquakes with a signal‐to‐noise ratio that can be quantified in terms of environmental independence. Finally, we exploit the relationship between the comodulated signals to identify their sources.

Multi-view diffusion maps

In this paper, we address the challenging task of achieving multi-view dimensionality reduction. The goal is to effectively use the availability of multiple views for extracting a coherent low-dimensional representation of the data. The proposed method exploits the intrinsic relation within each view, as well as the mutual relations between views. The multi-view dimensionality reduction is achieved by defining a cross-view model in which an implied random walk process is restrained to hop between objects in the different views. The method is robust to scaling and insensitive to small structural changes in the data. We define new diffusion distances and analyze the spectra of the proposed kernel. We show that the proposed framework is useful for various machine learning applications such as clustering, classification, and manifold learning. Finally, by fusing multi-sensor seismic data we present a method for automatic identification of seismic events.

Identification of Seismic Events on and near the North Korean Test Site after the Underground Nuclear Test Explosion of 3 September 2017

Identification of Seismic Events on and near the North Korean Test Site after the Underground Nuclear Test Explosion of 3 September 2017

Identification of Local Seismic Events Using the Mikhnevo Seismic Array

A large number of events with sources in the immediate vicinity of an array are usually detected during seismological observations with seismic arrays. These events should be detected and correctly interpreted during processing of seismic array records in order to avoid clogging up the event catalog. This problem can be solved by classifying records of local events by genetic features and creating a databank with the most representative samples. The present paper considers local events recorded using a unique scientific setup, the Mikhnevo small aperture seismic array. Epicenters of local seismic events are located less than 5 km from the center of the array. Seismic responses of acoustic shock waves are also examined. Seismic events caused by anthropogenic sources are identified and classified using cluster, cross-correlation, and wavelet analysis. Events accompanied only by the arrival of surface waves, as well as events represented by body, surface, and acoustic waves, are identified. Shock wave events are classified as a separate category. A small group of supposedly natural weak events is also found. As a result, a databank of waveforms of local seismic events for the Mikhnevo seismic array is established. In the future, this will make it possible to automate their identification when investigating the seismicity of the East European Platform.

Identification of similar seismic events using a phase-only correlation technique

Identification and clustering of similar acoustic emission (AE) events are important to determine source locations precisely and to evaluate subsurface cracks. An AE cluster analysis method using phase-only correlation (POC) is proposed to identify and hierarchically classify similar AE waveforms. The POC of time-varying spectral representations is used to evaluate the similarity between two waveform images in the time–frequency domain, and cluster analysis is used to classify the waveforms into groups according to a distance measure. The method is applied to waveforms from local earthquakes in Japan to assess its ability to identify similar waveforms perturbed by white noise.

Use of artificial neural networks for classification of noisy seismic signals

Automatic identification of noisy seismic events is still a problem. The process involves the analysis of complex relationships between data from different sources. Moreover, there are disturbing factors such as poor signal-to-noise ratio, the presence of accidental bursts of man-made noise, and changes in the amplitude and phase of the waves as they travel through the medium. The amount of observed data increases rapidly, so it is imperative to develop suitable and effective methods for processing and analyzing the influx of big data. Artificial neural networks (ANNs) show promise as a disruptive technology that will accelerate and improve analysis of seismic signals. ANNs are easy to apply, and the results often outperform alternative methods. This paper gives an overview of the highs and lows of neural networks, discusses the possibility of routine processing of seismic signals using ANNs, and presents examples of some interesting applications. It is hoped that researchers who read the article will actively use this technique, because ANNs could become more robust and flexible for solving complex problems that currently cannot be solved by the standard approach.

Phase-only correlation of time-varying spectral representations of microseismic data for identification of similar seismic events

Identification of similar seismic events is important for precise estimation of source locations and for evaluation of subsurface structure. Phase-only correlation is well known as a real-time image-matching method for fingerprint identification. I applied the phase-only correlation in a geophysical context to identify similar waveforms among microseismic events. The waveforms were first transformed into time-varying spectral representations to express frequency content in the time-frequency domain. The phase-only correlation function is calculated between two time-varying spectral representations and similarity is evaluated using the peak value of the phase-only correlation function. This method was applied to arbitrarily selected waveforms from aftershocks of an earthquake in Japan to assess its ability to identify similar waveforms perturbed by white noise. The detection of similarity of the proposed algorithm was compared to the similarity as detected by a 2D crosscorrelation function of the time-varying spectral representation and a 1D crosscorrelation of the raw waveform. This showed that the phase-only correlation function exhibits a sharp peak that quantifies similarity and dissimilarity over a wide range of signal-to-noise ratio (S/N) and remained unaffected by the length of the time window used to estimate time-varying spectral representations. Phase-only correlation may also have applications in other geophysical analyses and interpretations that are based on waveform and seismic image data.

Detection and Identification of Low-magnitude Seismic Events near Bala, Central Turkey, Using Array-based Waveform Correlation

Research Article| January 01, 2011 Detection and Identification of Low-magnitude Seismic Events near Bala, Central Turkey, Using Array-based Waveform Correlation Korhan U. Semin; Korhan U. Semin Search for other works by this author on: GSW Google Scholar Nurcan M. Ozel; Nurcan M. Ozel Search for other works by this author on: GSW Google Scholar Ocal Necmioglu Ocal Necmioglu Search for other works by this author on: GSW Google Scholar Author and Article Information Korhan U. Semin Nurcan M. Ozel Ocal Necmioglu Publisher: Seismological Society of America First Online: 09 Mar 2017 Online ISSN: 1938-2057 Print ISSN: 0895-0695 © 2011 by the Seismological Society of America Seismological Research Letters (2011) 82 (1): 97–103. https://doi.org/10.1785/gssrl.82.1.97 Article history First Online: 09 Mar 2017 Cite View This Citation Add to Citation Manager Share Icon Share Facebook Twitter LinkedIn MailTo Tools Icon Tools Get Permissions Search Site Citation Korhan U. Semin, Nurcan M. Ozel, Ocal Necmioglu; Detection and Identification of Low-magnitude Seismic Events near Bala, Central Turkey, Using Array-based Waveform Correlation. Seismological Research Letters 2011;; 82 (1): 97–103. doi: https://doi.org/10.1785/gssrl.82.1.97 Download citation file: Ris (Zotero) Refmanager EasyBib Bookends Mendeley Papers EndNote RefWorks BibTex toolbar search Search Dropdown Menu toolbar search search input Search input auto suggest filter your search All ContentBy SocietySeismological Research Letters Search Advanced Search Belbasi Seismic Station is composed of two seismic sub-arrays. One is a medium-period seismic array with a radius of about 45 km located in Ankara, and the second is a short-period seismic array with a radius of about 3 km located in Keskin. The distance between the two arrays is approximately 80 km. Both arrays consist of seven elements. Each array has a broadband seismometer located at the middle element of the circular array. Short-period vertical seismometers (Geotech 23900) are installed at depth 30 m from the surface while medium and broadband instruments (Geotech KS54000) are installed at depth 60... You do not have access to this content, please speak to your institutional administrator if you feel you should have access.

Earthquake arrival identification in a record with technogenic noise

Analysis of the problem of seismic event identification at an increased level of noise, by the data of a single sensor or local group of sensors, is made. This problem is the key to development of an ultra-short-term warning system of an earthquake that has occurred. Catastrophic damage and human deaths are usually caused by the sudden character of earthquake occurrence. If data processing could be done rapidly (i.e., within 4–5 s), such an earthquake warning would be of value for decreasing human deaths and economic damage. The aim of the present study is to investigate the selection of information features of signals in order to reduce the feature dimensionality and to incorporate the particular features of self-contained systems. Special emphasis is placed on the possibility of neural network techniques for the analysis of class separability in the attribute space. Our system will be helpful for reducing the impact of a disastrous earthquake. The main problems are highlighted and the methods for their solution are discussed; tests are carried out for a test problem, consisting in detection of small earthquakes in a noise-contaminated record. It is intended to develop algorithms and programs for identification of earthquake phase arrivals at an increased level of industrial and other noise. Further work will consist in designing instrumental solutions of the system.

Seismic event identification

AbstractEarthquakes and explosions generate seismic waveforms that have different characteristics. However, the challenge of confidently differentiating between these two signatures is complex, and requires the integration of physical and statistical techniques. This article reviews the methods for constructing discrimination features from diverse physical observations. These discrimination features are appropriate for many statistical classification frameworks. Under the null hypothesis an event is an explosion, we discuss strategies for constructing P‐values which can be interpreted as standardized discrimination features. We develop standardized discriminants for both teleseismic (simple propagation path in the mantle) and regional (complicated propagation path in the crust) events, following the trend toward characterizing increasingly smaller single‐point explosions. Copyright © 2010 John Wiley & Sons, Inc.This article is categorized under: Applications of Computational Statistics > Computational Physics and Computational Geophysics

A polarization filter for surface wave analysis: statistical approach

Polarization properties of seismic surface waves have been used for the identification and interpretation of small seismic events. In the present method of polarization filtering, generally elliptically polarizing noise is discriminated against specially polarizing surface waves. Statistically random nature of noise is taken into account in the determination of filter weights.

Instrumental and macroseismic data on processes in the focal zone of the Krivoy Rog earthquake on December 25, 2007.

In this paper the results of the studies of kinematical and dynamic characteristics of seismic waves and source for tangible Krivorozhsky earthquake on 25.12.2007 are presented. The refined data about the depth of the earthquake hypocenter and shaking intensity distribution are quoted. For the first time for the earthquake of the territory of the Ukrainian Shield focus parameters are restored, namely: source mechanism, dynamic characteristics, seismic moment tensor components, parameters of tearing process. It is shown, that deformation of medium in focal zone was shearing, the sliding of rupture bands was rough, the tearing process developed as directed from the bottom upward to the northwestern direction.. Identification of seismic events for the last 15 years around Krivorozhsky earthquake has been carried out. The problems of improvement the earthquake representative estimate for the eastern part of Ukraine have been formulated.

The nature of seismic events recorded in the west of the Ukrainian Shield in 2005 - 2007.

For recognition of the nature of seismic events with epicentra located on the East-European platform within the limits of the western part of the Ukrainian Shield wave forms of seismic signals have been studied and the changes of spectral composition with time. The main distinctions have been revealed of weakly intense tectonic earthquakes and explosions produced in the open pits which are manifested in special features of distribution of spectrum density of transversal and surface waves depending on the filtration frequency according to spectrograms. In case of elaboration of criteria for identification of seismic events the records of earthquakes and explosions at the seismic stations of the Main Center of Special Control of NSA Ukraine, equipped with modern digital seismologic logistics have been used.

Evaluation of seismic events detection algorithms

Identification of seismic events from continuously recorded seismic data in real-time through a Digital Seismic Data Recording system is a difficult task. Despite the vast amount of research in this field, the signal processing and event parameters discrimination algorithms have not yet fully come of age. Presently, we have a wide spectrum of trigger algorithms, ranging from a very simple amplitude threshold type to the sophisticated ones based on pattern recognition approaches. Some of the other approaches use adaptive technique and neural network methods. Researchers are continuously making efforts for the development of algorithms using various techniques, which produce minimum false trigger. Some approaches have been reported which are accurate for detecting first phase of events and take minimum possible computational time. In this paper several approaches for detecting event signals in background noise are presented and their precision evaluation is discussed.

Spectral time method of identification of seismic events at distances of 15°–40°

The spectral time method is applied to the identification of underground nuclear explosions (UNEs) and earthquakes from P wave records obtained at stations at epicentral distances of 15°–40°. It is necessary to utilize data at such great distances because the density of stations of the International Monitoring System is small in several areas of the Earth. At such distances, it is difficult to detect S and surface LR waves and, therefore, the efficiency of the identification using the diagnostic parameters (discriminants) S/P and LR/P is drastically reduced. As a result, the role of spectral and spectral time methods in the identification of events from P waves becomes much more significant. An advantage of the spectral time method (STM) is that it is applicable in the region of interest, provided that data from one or two reference underground chemical explosions and a representative sample of earthquake records obtained at network stations of the region are available. Moreover, in the majority of cases, the STM identifies earthquakes that could not be recognized with the use of spectral diagnostic parameters. We used records of UNEs and earthquakes in Central Asia received from the archives of the Center of the Incorporated Research Institution for Seismology (IRIS) in Washington, the National Data Center in Dubna (Russian Federation), and the International Data Center (IDC) in Vienna. The effectiveness of earthquake record selection using the given method amounted to about 80%. Supposedly, the application of this method of identification can increase the percentage of filtration (screening) of regional events having a natural origin.

Automatic classification and a-posteriori analysis of seismic event identification at Soufrière Hills volcano, Montserrat

Seismic energy radiation at Soufrière Hills volcano, Montserrat, is made up by various types of transient signals, which are distinguished by the Montserrat Volcano Observatory (MVO) in different classes with respect to their characteristics and/or origin. There are five fundamental classes, i.e., Volcano-Tectonic Events, Regional Events, Long-Period Events, Hybrid Events, and Rockfalls. Identification and classification of these transients, which have been hitherto carried out manually by various staff members, yield important information for the assessment of the state of the volcano system. In the frame of the MULTIMO project, we proposed the application of Artificial Neural Networks (ANN) for the classification of these kind of data in order to handle large data sets, and to achieve reproducible results, emulating the expert's analysis. Using the manual routine classification as a-priori information, we obtained a fair performance of such an automatic processing, with 70% of the automatic classifications being consistent with the original ones. From an analysis of the misclassified events, however, we found that for most of them the original a-priori classification was incorrect. In this study, we first revised manually the original a-priori classification. Based on a data set of 6000 events, we carried out a reanalysis of the seismic traces recorded at different seismic stations. Then, using this new information, we trained and tested the ANN, obtaining a successful classification in ca. 80% of records. Particularly, the automatic classification was excellent in the identification of Rockfalls and Volcano-Tectonic Events. Among the misfits, we observe the erroneous attribution of Long-Period and Hybrid Events to Rockfalls. This may be partly explained by the fact that signals addressed to as Rockfalls contain frequently contributions of various sources. Overall, the failure in the classification between some types of transients highlights the problem of the concurrent activation and/or unclear separation of distinctive sources from which the signals stem. We conclude that the automatic classification with ANN is a powerful tool for handling large data masses as well as for the a-posterior analysis of the consistency of the classification problem.