DASFormer: self-supervised pretraining for earthquake monitoring

Abstract Earthquake monitoring is a fundamental task to unravel the underlying physics of earthquakes and mitigate associated hazards for public safety. Distributed acoustic sensing, or DAS, which transforms pre-existing telecommunication cables into ultra-dense seismic networks, offers a cost-effec...

Full description

Saved in:
Bibliographic Details
Main Authors: Qianggang Ding, Zhichao Shen, Weiqiang Zhu, Bang Liu
Format: Article
Language:English
Published: Springer 2025-07-01
Series:Visual Intelligence
Subjects:
Online Access:https://doi.org/10.1007/s44267-025-00085-y
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849388828099346432
author Qianggang Ding
Zhichao Shen
Weiqiang Zhu
Bang Liu
author_facet Qianggang Ding
Zhichao Shen
Weiqiang Zhu
Bang Liu
author_sort Qianggang Ding
collection DOAJ
description Abstract Earthquake monitoring is a fundamental task to unravel the underlying physics of earthquakes and mitigate associated hazards for public safety. Distributed acoustic sensing, or DAS, which transforms pre-existing telecommunication cables into ultra-dense seismic networks, offers a cost-effective and scalable solution for next-generation earthquake monitoring. However, current approaches for earthquake monitoring like PhaseNet and PhaseNet-2 primarily rely on supervised learning, while manually labeled DAS data is quite limited and it is difficult to obtain more annotated datasets. In this paper, we present DASFormer, a novel self-supervised pretraining technique on DAS data with a coarse-to-fine framework that models spatial-temporal signal correlation. We treat earthquake monitoring as an anomaly detection task and demonstrate DASFormer can be directly utilized as a seismic phase detector. Experimental results demonstrate that DASFormer is effective in terms of several evaluation metrics and outperforms state-of-the-art time-series forecasting, anomaly detection, and foundation models on the unsupervised seismic detection task. We also demonstrate the potential of fine-tuning DASFormer to downstream tasks through case studies.
format Article
id doaj-art-d26e50fe38fb4e8095c8349fda7b8be7
institution Kabale University
issn 2097-3330
2731-9008
language English
publishDate 2025-07-01
publisher Springer
record_format Article
series Visual Intelligence
spelling doaj-art-d26e50fe38fb4e8095c8349fda7b8be72025-08-20T03:42:09ZengSpringerVisual Intelligence2097-33302731-90082025-07-013112010.1007/s44267-025-00085-yDASFormer: self-supervised pretraining for earthquake monitoringQianggang Ding0Zhichao Shen1Weiqiang Zhu2Bang Liu3Mila - Quebec AI Institute, University of MontrealWoods Hole Oceanographic InstitutionUC Berkeley Seismological LaboratoryMila - Quebec AI Institute, University of MontrealAbstract Earthquake monitoring is a fundamental task to unravel the underlying physics of earthquakes and mitigate associated hazards for public safety. Distributed acoustic sensing, or DAS, which transforms pre-existing telecommunication cables into ultra-dense seismic networks, offers a cost-effective and scalable solution for next-generation earthquake monitoring. However, current approaches for earthquake monitoring like PhaseNet and PhaseNet-2 primarily rely on supervised learning, while manually labeled DAS data is quite limited and it is difficult to obtain more annotated datasets. In this paper, we present DASFormer, a novel self-supervised pretraining technique on DAS data with a coarse-to-fine framework that models spatial-temporal signal correlation. We treat earthquake monitoring as an anomaly detection task and demonstrate DASFormer can be directly utilized as a seismic phase detector. Experimental results demonstrate that DASFormer is effective in terms of several evaluation metrics and outperforms state-of-the-art time-series forecasting, anomaly detection, and foundation models on the unsupervised seismic detection task. We also demonstrate the potential of fine-tuning DASFormer to downstream tasks through case studies.https://doi.org/10.1007/s44267-025-00085-yEarthquake monitoringImage imputationTime series forecastingSelf-supervised learning
spellingShingle Qianggang Ding
Zhichao Shen
Weiqiang Zhu
Bang Liu
DASFormer: self-supervised pretraining for earthquake monitoring
Visual Intelligence
Earthquake monitoring
Image imputation
Time series forecasting
Self-supervised learning
title DASFormer: self-supervised pretraining for earthquake monitoring
title_full DASFormer: self-supervised pretraining for earthquake monitoring
title_fullStr DASFormer: self-supervised pretraining for earthquake monitoring
title_full_unstemmed DASFormer: self-supervised pretraining for earthquake monitoring
title_short DASFormer: self-supervised pretraining for earthquake monitoring
title_sort dasformer self supervised pretraining for earthquake monitoring
topic Earthquake monitoring
Image imputation
Time series forecasting
Self-supervised learning
url https://doi.org/10.1007/s44267-025-00085-y
work_keys_str_mv AT qianggangding dasformerselfsupervisedpretrainingforearthquakemonitoring
AT zhichaoshen dasformerselfsupervisedpretrainingforearthquakemonitoring
AT weiqiangzhu dasformerselfsupervisedpretrainingforearthquakemonitoring
AT bangliu dasformerselfsupervisedpretrainingforearthquakemonitoring