MarsQuakeNet: An Extended Marsquake Catalogue from Deep Learning

MarsQuakeNet is an extended catalogue of marsquakes generated using deep learning techniques. The dataset includes a list of all detections and sol-wise predictions. It was created by Nikolaj Dahmen at ETH Zurich and published as supporting material for a 2022 paper in JGR Planets.

Use Cases

• Training seismic event detection models based on the extended marsquake catalogue
• Analyzing temporal patterns of Martian seismic activity based on sol-wise predictions
• Benchmarking deep learning techniques for planetary seismology using the provided code and detections

Strengths

• Dataset is associated with a peer-reviewed publication (JGR Planets, 2022JE007503)
• Includes both a static catalogue and dynamic code for reproduction from a GitHub repository
• Created by a researcher at a leading institution (ETH Zurich)

Limitations

• Column-level documentation is absent; field semantics must be inferred after download
• Row count is unknown, which may limit suitability assessment
• Description metadata is limited; actual data quality requires manual inspection after download

Provenance

Source

ETH Zurich

Collection Method

Generated using deep learning techniques applied to seismic data from the InSight mission.

Geography

Mars