The rapid development in observational capability provides both opportunities and challenges in seismology. Especially, the increase of seismic stations enhances the monitoring capability, which reveals more details of earthquake phenomena, including earthquake nucleation, spatial-temporal migration and mechanisms. However, the growing seismicity has increased the workload of the analyst greatly in the last decade. Traditional manual processing methods cannot process the large volume seismic data in time and effectively, which calls for new advances for efficient processing. In the past five years, there has been great progress on machine learning (ML) applications in seismology. This is driven by several factors such as: the increasing size of shared seismic data sets, new ML architectures and open-source codes, and improvements in computational power. ML methods have shown its great potential in automation tasks, such as seismic detection and phase arrival picking, and are thus being widely adopted. However, ML seismology is still a rapidly developing field. For example, new solutions are needed to resolve the difficulty in the generalization of ML methods. We are expecting progresses on novel ML model applications, dataset constructions, and innovative ways to apply ML methods to solve seismological problems.
The aim of this Research Topic is to cover new progress in ML seismology, promote the method development and applications of ML in seismic data processing, seismic detection, seismic location, seismic classification, seismic inversion, seismic imaging, earthquake early warning, as well as advance the intelligentization and efficiency of seismic data automatic processing and analysis.
We welcome submissions on the related topics but are not limited to:
• Development of ML-based seismic detection, phase picking, phase association and seismic denoising methods;
• Development of generalized ML models for large distance (> 150 km);
• Discrimination of different types of seismic sources (e.g., earthquakes, explosions, collapses and landslides);
• Realtime intelligent processing system and its application at multiple scales;
• Application of ML in regional seismic networks and portable dense seismic arrays;
• Application of ML in seismic tomography;
• Application of ML in ground motion prediction;
• New interpretations and understandings through applying ML-based methods;
• Comparison of the performance of existing ML models and methods in seismology; and
• The construction of benchmark datasets for ML training, validation, test, and evaluation.
The rapid development in observational capability provides both opportunities and challenges in seismology. Especially, the increase of seismic stations enhances the monitoring capability, which reveals more details of earthquake phenomena, including earthquake nucleation, spatial-temporal migration and mechanisms. However, the growing seismicity has increased the workload of the analyst greatly in the last decade. Traditional manual processing methods cannot process the large volume seismic data in time and effectively, which calls for new advances for efficient processing. In the past five years, there has been great progress on machine learning (ML) applications in seismology. This is driven by several factors such as: the increasing size of shared seismic data sets, new ML architectures and open-source codes, and improvements in computational power. ML methods have shown its great potential in automation tasks, such as seismic detection and phase arrival picking, and are thus being widely adopted. However, ML seismology is still a rapidly developing field. For example, new solutions are needed to resolve the difficulty in the generalization of ML methods. We are expecting progresses on novel ML model applications, dataset constructions, and innovative ways to apply ML methods to solve seismological problems.
The aim of this Research Topic is to cover new progress in ML seismology, promote the method development and applications of ML in seismic data processing, seismic detection, seismic location, seismic classification, seismic inversion, seismic imaging, earthquake early warning, as well as advance the intelligentization and efficiency of seismic data automatic processing and analysis.
We welcome submissions on the related topics but are not limited to:
• Development of ML-based seismic detection, phase picking, phase association and seismic denoising methods;
• Development of generalized ML models for large distance (> 150 km);
• Discrimination of different types of seismic sources (e.g., earthquakes, explosions, collapses and landslides);
• Realtime intelligent processing system and its application at multiple scales;
• Application of ML in regional seismic networks and portable dense seismic arrays;
• Application of ML in seismic tomography;
• Application of ML in ground motion prediction;
• New interpretations and understandings through applying ML-based methods;
• Comparison of the performance of existing ML models and methods in seismology; and
• The construction of benchmark datasets for ML training, validation, test, and evaluation.
