Physics-Aware Machine Learning for Seismic and Volcanic Signal Interpretation

TL;DR

This paper reviews recent machine learning methods for seismic and volcanic signal analysis, emphasizing the importance of physical awareness, robustness to domain shifts, and interpretability for operational monitoring.

Contribution

It organizes and evaluates ML approaches in seismic and volcanic monitoring, highlighting the role of physical constraints, self-supervision, and transfer evaluation protocols.

Findings

Classical signal processing provides essential inductive bias.

Self-supervision reduces dependence on labeled data.

Evaluation protocols are crucial for assessing transferability.

Abstract

Modern seismic and volcanic monitoring is increasingly shaped by continuous, multi-sensor observations and by the need to extract actionable information from nonstationary, noisy wavefields. In this context, machine learning has moved from a research curiosity to a practical ingredient of processing chains for detection, phase picking, classification, denoising, and anomaly tracking. However, improved accuracy on a fixed dataset is not sufficient for operational use. Models must remain reliable under domain shift (new stations, changing noise, evolving volcanic activity), provide uncertainty that supports decision-making, and connect their outputs to physically meaningful constraints. This paper surveys and organizes recent ML approaches for seismic and volcanic signal analysis, highlighting where classical signal processing provides indispensable inductive bias, how self-supervision and generative modeling can reduce dependence on labels, and which evaluation protocols best reflect transfer across regions. We conclude with open challenges for robust, interpretable, and maintainable AI-assisted monitoring.