In Earthquake Early Warning (EEW), every sufficiently impulsive signal is potentially the first evidence for an unfolding large earthquake. More often than not, however, impulsive signals are mere nuisance signals. One of the most fundamental - and difficult - tasks in EEW is to rapidly and reliably discriminate real local earthquake signals from all other signals. This discrimination is necessarily based on very little information, typically a few seconds worth of seismic waveforms from a small number of stations. As a result, current EEW systems struggle to avoid discrimination errors, and suffer from false and missed alerts. In this study we show how modern machine learning classifiers can strongly improve real-time signal/noise discrimination. We develop and compare a series of non-linear classifiers with variable architecture depths, including fully connected, convolutional (CNN) and recurrent neural networks, and a model that combines a generative adversarial network with a random forest (GAN+RF). We train all classifiers on the same data set, which includes 374k local earthquake records (M3.0-9.1) and 946k impulsive noise signals. We find that all classifiers outperform existing simple linear classifiers, and that complex models trained directly on the raw signals yield the greatest degree of improvement. Using 3s long waveform snippets, the CNN and the GAN+RF classifiers both reach 99.5% precision and 99.3% recall on an independent validation data set. Most misclassifications stem from impulsive teleseismic records, and from incorrectly labeled records in the data set. Our results suggest that machine learning classifiers can strongly improve the reliability and speed of EEW alerts.
In this paper, we present an analysis of seismic spectra that were calculated from all broadband channels (BH?) made available through IRIS, NIED F-net and Orfeus servers covering the past five years and beyond. A general characterization of the data is given in terms of spectral histograms and data-availability plots. We show that the spectral information can easily be categorized in time and regions. Spectral histograms indicate that seismic stations exist in Africa, Australia and Antarctica that measure spectra significantly below the global low-noise models above 1 Hz. We investigate world…
Seismologists often need to gather information about the subsurface structure of a location to determine if it is fit to be drilled for oil. However, there may be electrical noise in seismic data which is often removed by disregarding certain portions of the data with the use of a notch filter. Instead, we use a convolutional encoder decoder network to remove such noise by training the network to take the noisy shot record as input and remove the noise from the shot record as output. In this way, we retain important information about the data collected while still removing coherent noise in se…
Noises are common events in seismic reflection data that have very striking features in seismograms, affecting seismic data processing and interpretation. Noise attenuation is an essential phase in seismic processing data, usually resulting in seismic interpretation improvement that enhances the signal to noise ratio. Groundroll presence is the major fashion of significant noise in the land seismic survey. It is a type of coherent noise present in seismograms that appears as linear events, in most cases overlapping the reflections and probably making it challenging to recognize. There are seve…
During inter-eruption periods, magma pressurization yields subtle changes of the elastic properties of volcanic edifices. We use the reproducibility properties of the ambient seismic noise recorded on the Piton de la Fournaise volcano to measure relative seismic velocity variations of less than 0.1 % with a temporal resolution of one day. Our results show that five studied volcanic eruptions were preceded by clearly detectable seismic velocity decreases within the zone of magma injection. These precursors reflect the edifice dilatation induced by magma pressurization and can be useful indicato…
We consider the problem of seismic velocity change estimation using ambient noise recordings. Motivated by [23] we study how the velocity change estimation is affected by seasonal fluctuations in the noise sources. More precisely, we consider a numerical model and introduce spatio-temporal seasonal fluctuations in the noise sources. We show that indeed, as pointed out in [23], the stretching method is affected by these fluctuations and produces misleading apparent velocity variations which reduce dramatically the signal to noise ratio of the method. We also show that these apparent velocity va…
This study presents the concept of spectral modulation and a time-frequency analysis, applied to broadband local ($5<Δ<200$ km) seismic data from quarry blasts and micro-earthquakes, kindly supplied by the Institute for Petroleum Research and Geophysics (IPRG), Holon, Israel, from the so-called GIF-array. The aim of this research is verification of ripple-firing by recognition of scalloping trends in amplitude-spectra and hence discriminating quarry blasts from other events. Most quarry blasts are ripple-fired, in northern Israel open pit blasts; consequently the event discrimination method ba…
Simultaneous source seismic acquisition is an efficient method of seismic surveying that can considerably reduce the cost of high density seismic acquisition. The method results in overlapping records, or interference, that must be removed prior to subsequent processing. Deblending methods typically rely on the incoherence of the blending noise relative to the underlying signal. There are many common situations where these assumptions break down, for instance, when the underlying signal contains noise or erratic amplitudes, or when shooting times are not sufficiently random. We present a robus…
Lawrence Livermore National Laboratory（劳伦斯利弗莫尔国家实验室）；University of New Mexico（新墨西哥大学）；Southern University of Science and Technology（南方科技大学）；University of Utah（犹他大学）
This paper combines the power of deep-learning with the generalizability of physics-based features, to present an advanced method for seismic discrimination between earthquakes and explosions. The proposed method contains two branches: a deep learning branch operating directly on seismic waveforms or spectrograms, and a second branch operating on physics-based parametric features. These features are high-frequency P/S amplitude ratios and the difference between local magnitude (ML) and coda duration magnitude (MC). The combination achieves better generalization performance when applied to new …
Mazzarella and Scafetta (2016) showed that the seismic activity recorded at the Bunker-East (BKE) Vesuvian station from 1999 to 2014 suggests a higher nocturnal seismic activity. However, this station is located at about 50 m from the main road to the volcano's crater and since 2009 its seismograms also record a significant diurnal cultural noise due mostly to tourist tours to Mt. Vesuvius. Herein, we investigate whether the different seismic frequency between day and night times could be an artifact of the peculiar cultural noise that affects this station mostly from 9:00 am to 5:00 pm from s…
When monitoring geophysical parameters, data from segments that are contaminated by noise may have to be abandoned. This is the case, for example, in the geoelectrical field measurements at some sites in Japan, where high noise -due mainly to leakage currents from DC driven trains- prevails almost during 70\% of the 24 hour operational time. We show that even in such a case, the identification of seismic electric signals (SES), which are long-range correlated signals, may be possible, if the remaining noise free data are analyzed in natural time along with detrended fluctuation analysis (DFA).
It is a well-established principle that cross-correlating seismic observations at different receiver locations can yield estimates of band-limited inter-receiver Green's functions. This principle, known as seismic interferometry, is a powerful technique that can transform noise into signals which allow us to remotely image and interrogate subsurface Earth structures. In practice it is often necessary and even desirable to rely on noise already present in the environment. Theory that underpins many applications of ambient noise interferometry makes an assumption that the noise sources are uncor…
Seismic noise with frequencies above 1 Hz is often called cultural noise and is generally correlated quite well with human activities. Recently, cities in mainland China and Italy imposed lockdown restrictions in response to COVID-19, which gave us an unprecedented opportunity to study the relationship between seismic noise above 1 Hz and human activities. Using seismic records from stations in China and Italy, we show that seismic noise above 1 Hz was primarily generated by the local transportation systems. The lockdown of the cities and the imposition of travel restrictions led to a ~4-12 dB…
The Roles of Low-Noise Stations, Arrays and Ocean-Bottom Seismometers in Monitoring UK Offshore Seismicity associated with Subsurface Storage of Carbon Dioxide
School of Geosciences, University of Edinburgh（爱丁堡大学地球科学学院）
Reliable earthquake detection and seismic phase classification 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 is witnessed. This makes the handling of the seismic data rather daunting based on traditional approaches and therefore fuels the need for a more robust and reliable method. In this study, we investigate two deep learningbased models, termed 1D ResidualNeuralNetwork (ResNet) and multi-branch ResNet, for tackling the …
Seismic noise poses challenges for gravitational wave detection. Effective vibration isolation and methods to subtract unsheildable Newtonian Noise are examples. Seismic arrays offer one way to deal with these issues assuming seismic coherency. In this paper we find that wind induced seismic noise is incoherent and will dramatically reduce the projected low frequency sensitivity of future gravitational wave detectors. To quantify this, we measure the coherence length of wind induced seismic noise from 0.06--20~Hz in three distinct locations: close to a building, among tall trees and in shrubs.…
Extracting Empirical Green's Functions (EGFs) by cross-correlating ambient seismic noise or coda, considered as an efficient approach of retrieving new response inside the media of two receivers, has been developed well in the past two decades. However, EGFs emerge with high Signal-to-Noise Ratios (SNRs), which needs us to stack a large number of cross-correlations. Here, we retrieve EGFs with high SNRs from ambient seismic noise released by hurricane Dorian moving along the eastern coast of the United States during 3 days in Sept 2019. We systematically analyze the energy sources of seismic n…
In this paper we report a study of the seismic noise measured in the underground Baksan Neutrino Observatory. The main spectral feature below 1 $Hz$ is the oceanic microseism, while for greater frequencies the measured horizontal and vertical accelerations approaches the Peterson low noise model. Using two synchronized seismometers we also studied the coherence in the microseismic band ($0.1-0.5Hz$) between three underground stations located at $0.3km$, $1.4km$ and $3.7km$ from the tunnel entrance. Finally, on the base of our measurements we evaluate the Newtonian noise limit at this site usin…
Noise suppression is an essential step in any seismic processing workflow. A portion of this noise, particularly in land datasets, presents itself as random noise. In recent years, neural networks have been successfully used to denoise seismic data in a supervised fashion. However, supervised learning always comes with the often unachievable requirement of having noisy-clean data pairs for training. Using blind-spot networks, we redefine the denoising task as a self-supervised procedure where the network uses the surrounding noisy samples to estimate the noise-free value of a central sample. B…
Recovery of low-magnitude seismic events before the July 29, 2025, Kamchatka megathrust earthquake using waveform cross-correlation enhanced by the addition of stochastic noise
Denoising and filtering are widely used in routine seismic-data-processing to improve the signal-to-noise ratio (SNR) of recorded signals and by doing so to improve subsequent analyses. In this paper we develop a new denoising/decomposition method, DeepDenoiser, based on a deep neural network. This network is able to learn simultaneously a sparse representation of data in the time-frequency domain and a non-linear function that maps this representation into masks that decompose input data into a signal of interest and noise (defined as any non-seismic signal). We show that DeepDenoiser achieve…
As new techniques exploiting the Earth's ambient seismic noise field are developed and applied, such as for the observation of temporal changes in seismic velocity structure, it is crucial to quantify the precision with which wave-type measurements can be made. This work uses array data at the Homestake mine in Lead, South Dakota and an array at Sweetwater, Texas to consider two aspects that control this precision: the types of seismic wave contributing to the ambient noise field at microseism frequencies and the effect of array geometry. Both are quantified using measurements of wavefield coh…
Incorporating wave physical priors into diffusion models: A novel approach to seismic resolution enhancement
King Abdullah University of Science and Technology（阿卜杜拉国王科技大学）；School of Petroleum, China University of Petroleum-Beijing at Karamay（中国石油大学（北京）克拉玛依校区石油学院）；School of Geography and Tourism, Zhengzhou Normal University（郑州师范学院地理与旅游学院）
University of Texas at Austin（德克萨斯大学奥斯汀分校）；BP plc.（英国石油公司）；Lawrence Livermore National Laboratory（劳伦斯利弗莫尔国家实验室）
Estimating porosity models via seismic data is challenging due to the signal noise and insufficient resolution of seismic data. Although impedance inversion is often used by combining with well logs, several hurdles remain to retrieve sub-seismic scale porosity. As an alternative, we propose a machine learning-based workflow to convert seismic data to porosity models. A ResUNet++ based workflow is designed to take three seismic data in different frequencies (i.e., decomposed seismic data) and estimate their corresponding porosity model. The workflow is successfully demonstrated in the 3D chann…
A seismic array has been deployed at the Sanford Underground Research Facility in the former Homestake mine, South Dakota, to study the underground seismic environment. This includes exploring the advantages of constructing a third-generation gravitational-wave detector underground. A major noise source for these detectors would be Newtonian noise, which is induced by fluctuations in the local gravitational field. The hope is that a combination of a low-noise seismic environment and coherent noise subtraction using seismometers in the vicinity of the detector could suppress the Newtonian noise…
An array of seismometers is being developed at the Sanford Underground Laboratory, the former Homestake mine, in South Dakota to study the properties of underground seismic fields and Newtonian noise, and to investigate the possible advantages of constructing a third-generation gravitational-wave detector underground. Seismic data were analyzed to characterize seismic noise and disturbances. External databases were used to identify sources of seismic waves: ocean-wave data to identify sources of oceanic microseisms, and surface wind-speed data to investigate correlations with seismic motion as…
