Combining Deep Learning With Physics Based Features in Explosion‐Earthquake Discrimination

This paper combines the power of deep-learning with generalizability physics-based features, to present an advanced method for seismic discrimination between earthquakes and explosions. The proposed contains two branches: a deep learning branch operating directly on waveforms or spectrograms, second parametric features. These features are high-frequency P/S amplitude ratios difference local magnitude (ML) coda duration (MC). combination achieves better generalization performance when applied new regions than models that developed solely learning. We also examined which parts waveform data dominate decisions (i.e., via Grad-CAM). Such visualization provides window into black-box nature machine-learning offers insight how derived use make decisions.