SeisLM: a Foundation Model for Seismic Waveforms

TL;DR

SeisLM is a large-scale, self-supervised foundational model for seismic waveforms that improves various seismological tasks through pretraining on unlabeled data and fine-tuning.

Contribution

The paper introduces SeisLM, a novel seismic foundation model pretrained with contrastive learning, enabling versatile seismic analysis without task-specific labels.

Findings

SeisLM achieves high accuracy in event detection.

The model performs well in phase-picking and onset time regression.

SeisLM demonstrates effective classification of foreshocks and aftershocks.

Abstract

We introduce the Seismic Language Model (SeisLM), a foundational model designed to analyze seismic waveforms -- signals generated by Earth's vibrations such as the ones originating from earthquakes. SeisLM is pretrained on a large collection of open-source seismic datasets using a self-supervised contrastive loss, akin to BERT in language modeling. This approach allows the model to learn general seismic waveform patterns from unlabeled data without being tied to specific downstream tasks. When fine-tuned, SeisLM excels in seismological tasks like event detection, phase-picking, onset time regression, and foreshock-aftershock classification. The code has been made publicly available on https://github.com/liutianlin0121/seisLM.