Datasets:
QuakeFlow DAS
This repository contains Distributed Acoustic Sensing (DAS) datasets for PhaseNet-DAS. The DAS event format is explained here: Seismic Event Format for DAS.
Datasets
Arcata
The "arcata" dataset is from the Arcata, California, Distributed Acoustic Sensing (DAS) experiment: 2022 M6.4 Ferndale Aftershock Sequence and the Spring 2022 Arcata to Eureka California, DAS experiment.
- Period: December 22, 2022 – December 31, 2024
- Equipment: Luna QuantX DAS interrogator installed at the Arcata Police Station, connected to fiber-optic telecommunications infrastructure owned by Vero Communications running between Arcata and Eureka, California
- Institutions: U.S. Geological Survey (USGS), Cal Poly Humboldt University, Luna Inc.
- Files: 2,470 event files, named by timestamp (
YYYYMMDDTHHMMSSZ.h5)
This dataset was collected by Jeffrey J. McGuire, Andrew J. Barbour, Connie Stewart, Victor Yartsev, Martin Karrenbach, Mark Hemphill-Haley, Robert C. McPherson, Theresa Sawi, and Clara E. Yoon. Please inform the authors if you utilize this dataset in your research.
Related publication:
McGuire, J.J., Barbour, A.J., Stewart, C., Yartsev, V., Karrenbach, M., Hemphill-Haley, M., McPherson, R.C., Stockdale, K., Yoon, C.E., and Sawi, T., 2025, The GorDAS Distributed Acoustic Sensing Experiment Above the Cascadia Locked Zone and Subducted Gorda Slab. Seismological Research Letters, 96(4): 2489–2503. doi:10.1785/0220240415
Ridgecrest
The "ridgecrest_north" dataset is extracted from The SCEDC Earthquake Data AWS Public Dataset.
- Files: 751 event files, named by event ID (
ci########.h5)
This dataset is collected by Prof. Zhongwen Zhan (zwzhan@caltech.edu). Please inform the authors if you utilize this dataset in your research.
Data Format
Each .h5 file follows the DAS Seismic Event Format and contains:
data: 2D float32 array with shape(nch, nt)— DAS waveform data in microstrain/s- Attributes on the
datadataset:
| Attribute | Type | Description |
|---|---|---|
event_id |
str | Earthquake identifier |
event_time |
str | Origin time |
begin_time |
str | Window start time |
end_time |
str | Window end time |
latitude |
float | Epicenter latitude |
longitude |
float | Epicenter longitude |
depth_km |
float | Depth below surface (km) |
magnitude |
float | Seismic magnitude |
magnitude_type |
str | Magnitude scale |
dt_s |
float | Temporal sampling interval (s) |
dx_m |
float | Spatial channel spacing (m) |
unit |
str | Data measurement unit |
Note: Some files (especially in the arcata subset) may have missing event attributes (e.g., location, magnitude).
Usage
Requirements
h5pynumpyhuggingface_hub
Iterate over events
example.py provides DASDataset (a PyTorch Dataset) and helper functions. Files are downloaded on first access and cached locally:
from example import DASDataset
dataset = DASDataset("ridgecrest_north")
event = dataset[0]
print(event["data"].shape, event["event_id"], event["magnitude"])
Use with PyTorch DataLoader
from example import DASDataset
from torch.utils.data import DataLoader
dataset = DASDataset("arcata", max_events=100)
dataloader = DataLoader(dataset, batch_size=1, shuffle=True, num_workers=0)
for batch in dataloader:
print(batch["data"].shape)
break
Download a single file
import h5py
from huggingface_hub import hf_hub_download
filepath = hf_hub_download("AI4EPS/quakeflow_das", "arcata/data/20221223T043657Z.h5", repo_type="dataset")
with h5py.File(filepath, "r") as f:
data = f["data"][:] # shape: (nch, nt)
print(f"Waveform shape: {data.shape}")
for key, val in f["data"].attrs.items():
print(f"{key}: {val}")
Citation
If you use this dataset, please cite:
@article{zhu2023phasenet_das,
title={Seismic arrival-time picking on distributed acoustic sensing data using semi-supervised learning},
author={Zhu, Weiqiang and Biondi, Ettore and Li, Jiaxuan and Yin, Jiuxun and Ross, Zachary E. and Zhan, Zhongwen},
journal={Nature Communications},
volume={14},
pages={8192},
year={2023},
doi={10.1038/s41467-023-43355-3}
}
@misc{mcguire2024arcata_das,
title={Arcata, California, Distributed Acoustic Sensing (DAS) experiment: 2022 M6.4 Ferndale Aftershock Sequence (ver. 3.0, February 2026)},
author={McGuire, J.J. and Barbour, A.J. and Stewart, C. and Yartsev, V. and Karrenbach, M. and Hemphill-Haley, M. and McPherson, R.C. and Sawi, T. and Yoon, C.E.},
year={2024},
publisher={U.S. Geological Survey},
doi={10.5066/P1V7CKGA}
}
@article{mcguire2025gordas,
title={The GorDAS Distributed Acoustic Sensing Experiment Above the Cascadia Locked Zone and Subducted Gorda Slab},
author={McGuire, J.J. and Barbour, A.J. and Stewart, C. and Yartsev, V. and Karrenbach, M. and Hemphill-Haley, M. and McPherson, R.C. and Stockdale, K. and Yoon, C.E. and Sawi, T.},
journal={Seismological Research Letters},
volume={96},
number={4},
pages={2489--2503},
year={2025},
doi={10.1785/0220240415}
}
@article{gou2025leveraging,
title={Leveraging submarine DAS arrays for offshore earthquake early warning: A case study in Monterey Bay, California},
author={Gou, Yuancong and Allen, Richard M and Zhu, Weiqiang and Taira, Taka'aki and Chen, Li-Wei},
journal={Bulletin of the Seismological Society of America},
volume={115},
number={2},
pages={516--532},
year={2025},
publisher={Seismological Society of America}
}
- Downloads last month
- 3,903