Large Earthquakes Monitoring using High-Rate Global Navigation Satellite System Data through a Deep Learning approach

The High-rate Global Navigation Satellite System (HR-GNSS) instruments are devices that can detect seismic wave arrivals and measure ground displacement generated by an earthquake with high precision. By integrating HR-GNSS data with other sensors and models, we can improve the accuracy of earthquake assessments and provide valuable information for early warning and disaster preparedness. Our focus lies in developing deep-learning models leveraging HR-GNSS waveform data. These models significantly empower our capacity to detect and estimate the magnitude of large earthquakes. Yet, the rapid analysis of HR-GNSS data using deep learning algorithms remains a current challenge. To overcome this challenge, it is crucial to have access to large and high-quality datasets. Since the presence of noise in GNSS recordings particularly impacts data quality, especially for earthquakes measuring below magnitude 7, our training of Deep Learning (DL) models primarily relies on the data available from the largest earthquakes. This comes with a trade-off as these events provide a limited dataset because they occur less frequently, making the data poorly representative for model training. To overcome this limitation, we have used both synthetic earthquake signals combined with synthetic and real noise for model training, validation, and testing. Our investigation explores how diverse factors, such as noise, earthquake magnitude, station density, distance from the epicenter, and duration of the signal, affect the performance of our models. We aim to generalize the detection methodology and magnitude estimation for real-time monitoring of large earthquakes across diverse tectonic regions. The DL models proposed in this work will be integrated as complementary algorithms to the open-source Python package SAIPy.