Reliable Real‐Time Seismic Signal/Noise Discrimination With Machine Learning

Abstract

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 nonlinear classifiers with variable architecture depths, including fully connected, convolutional and recurrent neural networks, and a model that combines a generative adversarial network with a random forest. We train all classifiers on the same data set, which includes 374 k local earthquake records (M3.0-9.1) and 946 k 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 3-s-long waveform snippets, the convolutional neural network and the generative adversarial network with a random forest 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.Plain Language Summary Seismic stations record not only earthquake signals but also a wide variety of nuisance signals. Some of these nuisance signals are impulsive and can initially look very similar to real earthquake signals. This is a problem for earthquake early warning (EEW) algorithms, which sometimes misinterpret such signals as being real earthquake signals, and which may then send out false alerts. For each registered impulsive signal, EEW systems need to decide (or, classify) in real-time whether or not the signal stems from an actual earthquake. State-of-the-art machine learning (ML) classifiers have been shown to strongly outperform more standard linear classifiers in a wide range of classification problems. Here we analyze the performance of a variety of different ML classifiers to identify which type of classifier leads to the most reliable signal/noise discrimination in an EEW context. We find that we can successfully train complex deep learning classifiers that can discriminate between nuisance and earthquake signals very reliably (accuracy of 99.5%). Less complex ML classifiers also outperform a linear classifier, but with significantly higher error rates. The deep ML classifiers may allow EEW systems to almost entirely avoid false and missed signal detections...

How Often Can Earthquake Early Warning Systems Alert Sites With High‐Intensity Ground Motion?

“…This retrospective study was conducted with triggered waveform data, that is, we have implicitly assumed that all waveforms were accurately detected and associated to the correct event. In practice, phase detection and association are difficult and error-prone tasks, especially during intense aftershock sequences (Cochran 10.1029/2019JB017718 et al, 2018Hoshiba et al, 2011;Meier et al, 2019). Here the PLUM and FinDer algorithms have a key advantage over standard point-source algorithms in that they do not strongly rely on phase associations.…”

Real-time determination of earthquake focal mechanism via deep learning

“…We process all the synthetic data in the same way and use them to train the network. After the FMNet is well trained, in case that one real earthquake is identi ed with the existing algorithms of automatic detection and phase picking [7][8][9][10][11] , we rst remove their instrument responses and then perform the bandpass ltering, arrival-time alignments, and amplitude normalizations on the data prior to feeding them to the FMNet.…”

“…Besides, the successful application of AI in determining source focal mechanism shows that advanced AI technologies can handle much more complex data patterns other than those in detection, picking, location, etc. [7][8][9][10][11] , and hence it may signi cantly provoke further studies in resolving the complexity of earthquake process.…”

“…When a destructive earthquake occurs, real-time reporting of the earthquake parameters is of crucial importance for immediate destruction assessment and emergency evacuations. Recent efforts have been re ned towards applying AI technologies to estimate the source parameters because of its full automation, high e ciency, and human-like capability [4][5][6] , which has been remarkably demonstrated in numerous seismic processing tasks such as earthquake detection 7,8 , seismic phase picking [9][10][11] , magnitude estimation 12 , and other tasks [13][14][15][16][17] . Besides reporting the three basic parameters of an earthquake (i.e., origin time, location, and magnitude), it is also exceedingly important to derive the source focal mechanism in time to better understand various aspects of the earthquake.…”

Local earthquakes detection: A benchmark dataset of 3-component seismograms built on a global scale

“…Ross et al, 2018). This has important implications, e.g., for the improvement of modern earthquake early-warning system techniques and therefore for the mitigation of risk (Meier et al, 2019). ML applications, however, always require a large number of samples to induce these models to generalize well, i.e.…”