P- and S-wave detection and picking
Automatic identification of seismic phase arrivals and estimation of their arrival times using machine-learning models.
About the group
Research objective
We develop and validate methods for automated seismic data processing. We combine seismological workflows with modern machine-learning techniques to process large volumes of continuous recordings reliably and reproducibly.
Our main focus is the Ostrava region, with its distinctive combination of natural and human-induced seismicity. We also test and develop the methods using data from stations of the Czech Regional Seismic Network and beyond the region.
Research topics
What we work on
Seismic event location
Automatic location of seismic events detected in regional seismic data.
Event-type classification
Classification of mining-induced seismic events, surface explosions, and natural earthquakes based on the characteristics of seismic records.
Data quality and denoising
Detection of poor-quality segments, noise suppression, and preparation of data for subsequent automated analysis, including machine-learning-based methods.
Data and collaboration
Regional context
Our research uses seismic data from the Ostrava region and data from stations of the Czech Regional Seismic Network. We collaborate with the Institute of Geophysics of the Czech Academy of Sciences on seismic monitoring, data interpretation, and the development of automated processing methods.
People
Research team
Michael Skotnica
Research group leader at the Institute of Geonics CAS
michael.skotnica@ugn.cas.cz
Bohdan Rieznikov
bohdan.rieznikov@ugn.cas.cz
Jana Rušajová
jana.rusajova@ugn.cas.cz
Marek Pecha
marek.pecha@ugn.cas.cz
Vít Wandrol
vit.wandrol@ugn.cas.cz
Results
Activities
- Workshop: Neural Networks in Seismology
- Seismic event classification using machine learning approaches, Michael Skotnica, Marek Pecha, Jana Rušajová, Bohdan Rieznikov, Vít Wandrol, conference presentation at MedGU-25
- Machine learning-based seismic event classification at selected stations of the Czech Regional Seismic Network, M. Skotnica, M. Pecha, J. Pazdírková, J. Rušajová, B. Rieznikov, abstract in EGU 2026
- Machine learning approaches to seismic event classification in the Ostrava region, Marek Pecha, Michael Skotnica, Jana Rušajová, Bohdan Rieznikov, Vít Wandrol, Markéta Rösnerová, Jaromír Knejzlík, manuscript in preparation; preprint available on arXiv