Description
E-poster session with display of each e-poster on an assigned touchscreen
At the International Data Centre (IDC) real-time data goes through a three-step process to determine if there is an event. These three steps in seismic/hydro/infrasound (SHI) operations are station processing, network processing and interactive review. Inside each of these steps are algorithms that conduct various tasks (e.g. detection, categorization, location). Many of these algorithms,...
Sandia National Laboratories is developing the Geophysical Monitoring System (GMS) for modernization of the United States National Data Centre waveform processing system, now focused on development of interactive analysis capabilities (IAN). The United States provides open source releases of GMS software to support International Data Centre (IDC) Re-engineering. The latest GMS release includes...
The identification and precise location of earthquakes are essential for understanding seismicity and mitigating associated risks. This research focuses on utilizing state-of-the-art deep learning methods, specifically PhaseNet, to enhance the accuracy of seismic phase picking and event localization. By analyzing 19 years of continuous seismic data from a gas storage field in Iran, our study...
Acoustic, hydroacoustic and seismic signals that are most often of practical interest are those whose nature is impulsive. The appearance of such forms of impulses is a consequence of the sudden release of energy at a location in a relatively short time. Estimation of the time of arrival of these signals, on several spatially distributed sensors, enables locating the place where these...
We present an end-to-end LLM engineering platform for fine-tuning, evaluation and registration of custom models and adapters. Our platform, built on top of open-source tools, provides a comprehensive suite of components for data processing, fine-tuning, evaluation, and deployment of LLMs. Key features include pipeline orchestration for batch-oriented workflows, model training and fine-tuning,...
Seismic observation at Syowa Station (69.0°S, 39.6°E; SYO), Antarctica started since 1959 associated with the International Geophysical Year (IGY; 1957-1958) campaign. Since the establishment of the INTELSAT telecommunication link, digital waveform data have been transmitted to the National Institute of Polar Research (NIPR) for the utilization of phase identification more clearly....
This study evaluates seismic activity in the Americas region over the past five years using data from the Seismic Event Bulletin (SEB) of the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO). By applying the Gutenberg-Richter law, key seismic parameters (a and b) are calculated to estimate recurrence intervals and maximum expected magnitudes for localized subregions. Using magnitude...
In this study on the use of IMS station data, the seismic event of October 12, 2024, at 17:43 UTC, in the northwest of Costa Rica will be analyzed, using nearby seismic stations and including non-IMS stations from the local OVSICORI-UNA network. Additionally, the results of the analysis of the main infrasound events detected by the portable array station I69CR will be presented, along with...
The use of Machine Learning (ML) methods in seismology has gained significant attention in recent years, driven by the availability of large, high-quality datasets. While ML is applied to various seismological tasks, it is most commonly used for seismic signal detection, phase picking, and classification. Recent studies show that deep learning models such as EQTransformer, PhaseNet, and...
Given the increasing volume of data from seismic networks, manually analyzing and identifying earthquake phases is becoming unfeasible. This has led to the adopting of automated methods, particularly deep learning models, for accurate and efficient phase identification. This study evaluated the performance of four popular deep learning models (PhaseNet, EQTransformer, GPD, and BasicPhaseAE)...
Data recorded for monitoring a global nuclear-test-ban treaty are a mixture of useful signals, ambient and measurement noise. Suppressing these noises and increasing the signal to noise ratio is a significant task. The paper describes a way to suppress the mentioned noises, so that by applying the proposed method, the signal to noise ratio is increased by more than 20 dB. This result is...
Low magnitude monitoring of explosions is an expanding field of interest, but due to low SNR and station coverage it is difficult to identify them. We hope to explore these limitations by using simple features from regional events and easy to implement machine learning algorithms to classify earthquakes vs explosions. Using the CNRST bulletin, we collected 4542 regional waveforms for both...
Continuous monitoring of environmental and cultural noise levels is critical in selecting and maintaining seismic station sites, especially in regions undergoing rapid urbanisation. Cultural noise is of particular concern for stations near densely populated or industrialising areas, where human activities generate vibrations that interfere with detecting and analysing seismic signals. This...
The ability of automatic data processing at the CTBTO results in identifying and estimating parameters for phases detected by the IMS stations. These processes contribute to provide automatic event locations generating the Standard Event List3(SEL3). During interactive analysis, the SEL3 event solutions are refined by modifying, or re-estimating phase attributes. These actions include: (1)...
The effective detection threshold of the CTBTO seismological network is a key tool for prioritizing repairs of primary and auxiliary seismic stations. However, many seismic networks implement it manually or semi-manually, using programs that delay the calculation and visualization process and are prone to errors.
This study aims to develop an algorithm to automate the calculation of the...
In order to be able to detect and characterize small magnitude events, even those with long propagation distances, seismic arrays are perfectly adapted tools with their high detection capabilities. We have studied the possibility to improve the detection and the localization of local and regional seimic events by using data from only one array.
International Monitoring System seismic arrays...
The International Data Centre (IDC) of the Comprehensive Nuclear Test Ban Treaty Organization (CTBTO) processes and analyses data from the International Monitoring System (IMS). This effort culminates in the daily production of the Reviewed Event Bulletin (REB), recognized as one of the most comprehensive global seismic bulletins.
This study compares the IDC REB bulletins with those...
We compared the locations of earthquakes with magnitudes greater than 3.0 that occurred around the Korean Peninsula from January 1, 2010 to April 11, 2024 between the KIGAM (Korea Institute of Geoscience and Mineral Resources) catalog and the CTBTO (Comprehensive Nuclear-Test-Ban Treaty Organization) IDC (International Data Center) REB (Review Event Bulletin) earthquake catalog. We initially...
The CTBTO Link to the database of the International Seismological Centre (ISC) is a service provided by arrangement with IDC/CTBTO. The Link provides PTS and National Data Centres (NDC) with dedicated access to long-term definitive global datasets maintained by the ISC. Functionality includes specially designed graphical interfaces, database queries and non-IMS waveform requests. This service...
For the accurate determination of local earthquake hypocenters, it is necessary to have reliable and good quality P- and S-wave readings. Data quality control is an important step to correct or filter anomalies that can occur during phase picking. The Chatelain method (Chatelain, 1978) is first applied to facilitate VP/VS ratio calculation. This method is based on calculating the differences...
We present our work on training and the application of deep learning algorithms for the automated phase picking of body waves on the the IMS network. We train new IMS data based seismic phase pickers from both EQT and PhaseNet architectures. Phase picking is a necessary step before event localization and characterization and deep learning based models have been proven to perform well at this...
Accurate modelling of infrasound transmission losses is essential for evaluating the performance of the International Monitoring System (IMS) infrasound network. The parabolic equation method provides accurate loss modeling but is computationally expensive for operational monitoring applications. To address this, a previous study trained a Convolutional Neural Network on regionally simulated...
On the night of 22 November 2022, at 22:22 UTC, a fireball was observed in the sky over Dellys, a town in north-eastern Algeria. According to eyewitness reports, the celestial body was seen travelling from the south-west to the north-east. Two recently installed infrasound stations recorded the signal from this trajectory, enabling the determination of some of the meteor's characteristics. The...
Time-space variations of infrasound source locations for three years, 2019-2021, were studied by using a combination of two local arrays in the Lützow-Holm Bay (LHB), Antarctica. The local arrays deployed at two coastal outcrops detected temporal variations in signal frequency content as well as propagating directions during these years. A large number of infrasound sources were detected with...
The production of high-quality event bulletins relies on high-accuracy and high-precision hypocentral locations and low completeness magnitude. The Terceira Rift provides an ideal setting to study rifting processes due to its seismicity and volcanism, driven by slow transtensional deformation between the Nubian and Eurasian plates. This study leverages data from the UPFLOW project, including...
The performance of earthquake detection and localization within the seismic network of Thailand was analyzed, encompassing the Thai Meteorological Department (TMD) network, which serves as the main authority for monitoring seismic activity in Thailand and adjacent regions, along with the seismic network of the Department of Mineral Resources (DMR) and the CMAR array of the CTBTO. The minimum...
Accurately distinguishing between nuclear explosions and earthquakes by manual discrimination or automatic discrimination is crucial in the field of seismic signal analysis. The technique of manual discriminating takes a lot of time, and it can occasionally become inaccurate. Thus, a high-accuracy automatic discrimination method is required. An inaccurate assessment of a region's inherent...
Infrasound sensing is essential for the global detection and precise geolocation of bolide events. However, discrepancies often arise between observed back-azimuths — the arrival direction of infrasound signals — and theoretical predictions based on the bolide's peak brightness location, especially for shallow angle entries. Shallow entry angles result in complex, extended acoustic signals...
The rapid detection, location and classification of seismic events, particularly offshore, has become increasingly important in a world where critical infrastructure on the seabed has been the subject of several damaging incidents. Given the intense political focus on the events, it is equally important to provide data to deescalate a situation if events have natural causes as it is to help...
The accuracy and completeness of the International Data Centre (IDC) seismic bulletins, such as the Late Event Bulletin (LEB) and Reviewed Event Bulletin (REB), are essential for global seismic monitoring of the Comprehensive Nuclear-Test-Ban Treaty Organization (CTBTO). This study assesses the contribution of scanning processes for missed events during the interactive analysis and their...
Infrasound monitoring usually requires ground truth information from other sources in order to classify a detection. Here, we present an ensemble model that combines a Random Forest trained on simple features derived from the Progressive Multi-Channel Correlation (PMCC) technique with a Neural Network trained on spectrograms calculated from the waveforms to overcome the necessity of ground...
Sandia National Laboratories is developing the Geophysical Monitoring System (GMS) for modernization of the United States National Data Center waveform processing system, now focused on development of interactive analysis capabilities (IAN). The United States provides open source releases of GMS software to support International Data Centre (IDC) re-engineering. Sandia has recently integrated...
The generation of synthetic seismograms through simulation is a fundamental tool of seismology required to run quantitative hypothesis tests. A variety of approaches have been developed throughout the seismological community and each has their own specific user interface based on their implementation. This causes a challenge to researchers who will need to learn new interfaces with each new...
Sandia National Laboratories is developing the Geophysical Monitoring System (GMS) for modernization of the United States National Data Center waveform processing system. The GMS development effort is now focused on development of interactive analysis capabilities (IAN) to replace the ageing Analyst Review Station. IAN now includes capabilities to filter, rotate, and beam station data; measure...
Geothermal energy exploration has been going on in Kenya for the last five decades for green energy exploitation. Between the month of April and August 2024, twenty seismometers were installed throughout the Menengai geothermal field/Caldera and its’ surrounding at a radius of 15Km using. The stations were installed with IGU-BD3C-5 Smart solo seismometer, a low-frequency, three-component...
Volcanic eruptions generate infrasound, which consists of low-frequency acoustic waves below 20 Hz. This phenomenon is valuable for monitoring volcanic activity, particularly in regions where local sensor networks are impractical. The expansion of infrasound stations of the International Monitoring System (IMS), along with advancements in infrasound data analysis, has improved the detection of...
Temporal variations of the noise conditions constrain the ability to detect and identify signals of interest at infrasound stations. Station-dependent factors that contribute to the noise include wind and turbulence. A coherent source of ambient noise at the global infrasound station network of the International Monitoring System are microbaroms from the oceans, which vary seasonally such that...
Discrimination between explosions and earthquakes is a major challenge in the field of seismology. This task is important not only to meet the expectations of the Comprehensive Nuclear Test Ban Treaty (CTBT) but also to refine seismic bulletins used in regional seismicity research, seismotectonic analysis, and seismic hazard assessment.
Seven models using different machine learning...
Seismic signal classification plays a critical role in monitoring and understanding seismic activity by attributing each detected event to its source, such as earthquakes, quarry blasts, volcanic events, or nuclear explosions. Traditional methods, such as cross-correlation-based approaches, offer the advantage of not requiring large databases or explicit feature extraction. However, these...
We assess the impact of the Event Definition Criteria (EDC) on the IDC bulletins from the perspective of introduced changes in IDC processing and growing number of IMS stations during the last decade.
Two main changes in IDC processing contributed to the increase SEL3 events reviewed by analysts – infrasound processing since 2016 and introduction of NET-VISA as additional associator to...
The Neighbourhood Algorithm is a grid search method that optimizes a user-supplied objective function over a computational domain using Voronoi cell tesselation. The algorithm is a method for solving geophysical inverse problems with the additional benefit of not requiring the estimation of travel-time derivative information (Sambridge, 1999)
In this application a misfit function for...
Routine analysis of infrasound data began in early 2010. Since then, analysts at the International Data Centre have reviewed more than 117,700 automatically built SEL3 infrasound events and included 62,700 events in the Late Event Bulletin (LEB). Of these, 34,800 events met the Reviewed Event Bulletin (REB) event criteria. Analysing infrasound data presents several challenges. These include...
On February 6, 2023, at 01:17 UTC and 10:24 UTC, two devastating earthquakes with magnitudes of Mw 7.7 and Mw 7.6 struck, with epicenters located in Pazarcık (Kahramanmaraş) and Elbistan (Kahramanmaraş), respectively. These earthquake couples are one of the most destructive earthquakes occurred in recent history affecting 11 provinces in Turkey’s Southeast region and causing more than 53000...
To study impulse sound of shell explosion position in a far field over 5km even for nuclear explosion’s position. One new way was designed to perform clear spot image of source position with accurate result, that is a reverse beam-forming of burst pulse sound, Meanwhile, time-delay variance is used to estimate source’s coordinates and sound’s average velocity suggested the sound’s propagation...
Discriminating tectonic events from artificial ones poses a significant challenge, particularly when both sources are small and geographically close. This issue is common in regions with quarries for rock blasting and tectonic seismic activity. This study focuses on distinguishing tectonic events from quarry explosions by integrating seismic and infrasound data collected from stations in Sete...
The Punggye-ri nuclear test site in North Korea, historically not seismically active, has likely experienced significant stress changes from the six nuclear tests, particularly the largest in 2017. These stress changes have facilitated the triggering of seismicity in the surrounding area long after the tests. Using multi-channel correlation detectors (Gibbons, 2012) at the IMS seismic arrays...
Recorded seismic data are generally contaminated by noise from different sources, which masks the signals of interest. We implemented a noise suppression approach based on the mathematical morphology theorem. The method involves compound operations of dilation and erosion using structuring elements of varying lengths and decomposes an input noisy waveform into several time functions with...
Capacity building effort of PTS to provide NDCes with software to independently reproduce IDC SHI catalogue has led to development and provisioning of the system NDC –in –a-Box (NIAB).
We present the results of work of NDC-in-a-Box (NIAB) containing the NET-VISA associator integrated with SeisComp (SC). Our configuration of SC also enables usage IDC DFX detector and Netvisa associator for...
We present NPLoc, a machine learning model designed to accurately locate earthquakes within a permanent seismic network. This model predicts earthquake origin time, hypocenter, magnitude and its associated uncertainties as rapidly as possible. This approach uses temporal patterns extracted from earthquakes within a seismic network and utilizes the Histogram-Based Gradient Boosting method. To...
This research explores the application of the YOLOv8 object detection framework for real-time earthquake detection using spectrogram images derived from seismic data. The proposed method is a promising candidate for real-time event detection, with the potential to improve detection accuracy and reduce the risk of human error. By leveraging the strengths of YOLOv8, known for its speed and...
Monitoring of phenomena in the atmosphere and at shallow depth of burial using infrasonic signals is often performed via automated detection, localization and characterization with refinement using interactive tools for an identified event of interest. Ongoing infrasound research and development at Los Alamos National Laboratory includes development and evaluation of various signal analysis...
In the field of seismic monitoring, the three critical tasks of phase picking, association and location are interconnected and tightly coupled. Current seismic monitoring methods typically tackle phase picking, association and location separately, and most existing phase picking methods focus on single-station waveform data processing. Graph Neural Networks (GNNs) are deep learning frameworks...
In this presentation, we demonstrate how to use the IMS FDSN web service to download and analyze seismic data. Specifically, we show how to integrate IMS waveforms and phase-picking data with locally collected datasets, and how to process this information to determine event locations and magnitudes.
For our analysis, we utilize the seismic analysis software package SEISAN. SEISAN (see...
With the projected increase in world nuclear capacity comes the hurdle of spent nuclear fuel. Pyroprocessing is one method to process irradiated fuel by using high temperatures and electrochemical steps to separate radioactive components of interest. High temperatures, high radiation levels, and equipment confined to a heavily shielded hot cell are a few of the challenges introduced regarding...
The NET-VISA software package features a physics-based probabilistic model combined with a heuristic inference algorithm to identify the most likely set of seismic events corresponding to a series of detections by a global seismic network. It has been enhanced to detect events occurring in three mediums—rock, air, and water—and supports seismic, hydro-acoustic, and infrasound sensors.
The...
The Standard Screened Event Bulletin (SSEB) is a sub-sequent automatic product of the Standard Screened Bulletin (SEB) produced by the International Data Centre (IDC). The SEB is a post-location processing of the Reviewed Event Bulletin (REB) that includes the event characteristics that are used to screen out seismoacoustic events generated from natural or non-nuclear man-made phenomena. The...
Deep learning approaches are effective for earthquake detection and phase picking. However, challenges in existing research include using non-uniform datasets, such as training on limited event distances and excluding noise samples. Models trained exclusively on local events often fail to perform well on teleseismic signals, while models trained only on signals struggle with robustness in the...
Recently, involving the capabilities of artificial intelligence (AI) to solve problems in exploration seismology attracted increasing interest. One example is the use of recurrent neural networks to estimate seismic activity. These networks can process time-series data, allowing to model dynamic processes through time-dependencies. Clustering to segment seismic data helps identify different...
The Idaho National Laboratory (INL), located in Eastern Idaho, in the United States, has a history of operating nuclear test reactors and is currently designated as a Reactor Innovation Center. The seismic monitoring network began as a single seismic station in 1973 and has grown to include over 100 instruments for the purpose of monitoring geophysical phenomena. As the network has evolved...
Limestone mining activity close to site KAPI (CTBTO) has been clearly visible through the land clearing from satellite imagery. The BMKG Region IV Makassar team has visited the site to analyze noise from heavy equipment activities that operating around the site, consisting of a bucket excavator on 25 August 2023 and a breaker excavator on 23 January 2024. Noise from the activities of these two...
To support National Data Centers (NDCs) in their ability to process IMS data, the Provisional Technical Secretariat provides the NDC-in-a-box software package. Many NDCs do real-time processing of continuous data from IMS waveform technology stations, with a particular focus on seismic monitoring data. This data can be used for verification purposes, as well as for civil and scientific...
The Swedish National (SNSN) currently operates 80 broadband seismic stations. In addition SNSN receives realtime data from about 120 stations located in Norway, Finland, Denmark, Germany, Poland, the Baltic States, and Russia. SNSN processes the waveform data of this virtual network using the SeisComp and Earthworm systems in parallel. In order to screen out spurious events we generate a...
Traditional moment tensor inversions are a common tool used to characterize events of interest for nonproliferation monitoring. Many inversions assume a known source time function and solve for the moment tensor of a seismic source. However, this requires a source time function to be assumed, which could result in inaccurate results if, for example, an explosion source time function is used...
The increasing volume and complexity of seismic data require advanced techniques for efficient signal classification, particularly in monitoring compliance with the Comprehensive Nuclear-Test-Ban Treaty (CTBT). This study introduces a robust approach using Convolutional Neural Networks (CNNs) to automate seismic signal classification, significantly improving both accuracy and classification...
Seismic wave detection and phase picking are the initial steps in most seismological studies. The increasing seismic data necessitates the development of capable auto-detection and precise auto-phase-picking algorithms. Deep learning approaches have played a crucial role in these tasks in recent years. EQTransformer and PhaseNet are among the most important models introduced for detection and...
At the International Data Centre (IDC), data received from the network goes through a three-step process (station processing, network processing and interactive review) to determine if a combination of detections can be built into an event. One of the major steps in determining if an event can be built or not, is the phase classification of the detected signals. For acoustic data, phases are...
Stockwell transform, commonly known as S Transform (ST, Stockwell et al., 1996) is an extension of the continuous wavelet transform (CWT) and involves an inverse frequency-dependence of the localizing Gaussian window as well as a modulating phase factor, which result in better frequency resolution than CWT or short-time Fourier transform (STFT). We are leveraging that advantage by implementing...
