The transformer earthquake alerting model: a new versatile approach to earthquake early warning

Abstract

Summary Earthquakes are major hazards to humans, buildings and infrastructure. Early warning methods aim to provide advance notice of incoming strong shaking to enable preventive action and mitigate seismic risk. Their usefulness depends on accuracy, the relation between true, missed and false alerts, and timeliness, the time between a warning and the arrival of strong shaking. Current approaches suffer from apparent aleatoric uncertainties due to simplified modelling or short warning times. Here we propose a novel early warning method, the deep-learning based transformer earthquake alerting model (TEAM), to mitigate these limitations. TEAM analyzes raw, strong motion waveforms of an arbitrary number of stations at arbitrary locations in real-time, making it easily adaptable to changing seismic networks and warning targets. We evaluate TEAM on two regions with high seismic hazard, Japan and Italy, that are complementary in their seismicity. On both datasets TEAM outperforms existing early warning methods considerably, offering accurate and timely warnings. Using domain adaptation, TEAM even provides reliable alerts for events larger than any in the training data, a property of highest importance as records from very large events are rare in many regions.

Neural Network-Based Strong Motion Prediction for On-Site Earthquake Early Warning

“…CNN has been applied to several seismic problems, such as earthquake detection [ 15 ], earthquake localization [ 16 ], earthquake arrival-time picking [ 17 , 18 ], and PGA prediction [ 19 , 20 ]. The above studies [ 15 , 16 , 17 , 18 , 19 , 20 ] show that CNN techniques can extract implicit features from the waveforms to obtain useful information. For PGA prediction, Jozinović et al [ 19 ] presented a CNN-based model to predict the intensity of ground shaking.…”

“…The scheme provides a reliable estimation of ground shaking 15 to 20 s after an earthquake occurs. Münchmeyer et al [ 20 ] developed a transformer earthquake alerting model to predict the PGA in target areas. The model issues warnings by predicting whether the output probability of the PGA exceeds a pre-defined probability or not.…”

Graph Neural Networks for Multivariate Time Series Regression with Application to Seismic Data

“…In addition, CNNs were developed for P-and S-wave arrival times picking [e.g., 37,39]. Others used multistation waveforms which were analysed for the estimation of the location of earthquakes [5] or earthquake early warning [40].…”

A study on the effect of input data length on deep learning based magnitude classifier

“…Furthermore, we investigate the effect of using different lengths of data after the first P-arrival (1s, 3s, 10s, 20s and 30s) on the performance of this classifier model. In each case the P-wave data is preceded by 2.8-3.0 seconds of pre-signal noise, so the model can learn the noise characteristics of the station [36]. The data labels 0, 1, and 2 are used to denote the classes noise, low-magnitude and high-magnitude, respectively.…”