Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Continuous chatter of the Cascadia subduction zone revealed by machine learning


Tectonic faults slip in various manners, which range from ordinary earthquakes to slow slip events to aseismic fault creep. Slow slip and associated tremor are common to many subduction zones, and occur down-dip from the neighbouring locked zone where megaquakes take place. In the clearest cases, such as Cascadia, identified tremor occurs in discrete bursts, primarily during the slow slip event. Here we show that the Cascadia subduction zone is apparently continuously broadcasting a low-amplitude, tremor-like signal that precisely informs of the fault displacement rate throughout the slow slip cycle. Using a method based on machine learning previously developed in the laboratory, we analysed large amounts of raw seismic data from Vancouver Island to separate this signal from the background seismic noise. We posit that this provides indirect real-time access to fault physics on the down-dip portion of the megathrust, and thus may prove useful in determining if and how a slow slip may couple to or evolve into a major earthquake.

This is a preview of subscription content, access via your institution

Relevant articles

Open Access articles citing this article.

Access options

Buy article

Get time limited or full article access on ReadCube.


All prices are NET prices.

Fig. 1: Map and schematic of the region analysed: Vancouver Island and the subduction zone.
Fig. 2: Estimating the GPS displacement rate from the continuous seismic data.
Fig. 3: Strong correlations between the statistics of the seismic data and the GPS displacement rate.

Data availability

All the data used are publicly available and can be found online. The seismic data are from the CNSN27 (, and the GPS data are from the Western Canada Deformation Array operated by the Geological Survey of Canada, preprocessed by the United States Geological Survey28 (, NA-fixed trended data). The work flow described in Methods uses open source software (python and python packages including scikit-learn37 and obspy36). We are currently unable to make the python script associated with this paper available, but we aim to make it available in the near future. Please contact for details.


  1. Obara, K. Nonvolcanic deep tremor associated with subduction in southwest Japan. Science 296, 1679–1681 (2002).

    Article  Google Scholar 

  2. Rogers, G. & Dragert, H. Episodic tremor and slip on the Cascadia subduction zone: the chatter of silent slip. Science 300, 1942–1943 (2003).

    Article  Google Scholar 

  3. Perfettini, H. et al. Seismic and aseismic slip on the Central Peru megathrust. Nature 465, 78–81 (2010).

    Article  Google Scholar 

  4. Wech, A. G., Creager, K. C. & Melbourne, T. I. Seismic and geodetic constraints on Cascadia slow slip. J. Geophys. Res. Solid Earth 114, B10316 (2009).

    Article  Google Scholar 

  5. Bartlow, N. M., Miyazaki, S., Bradley, A. M. & Segall, P. Space–time correlation of slip and tremor during the 2009 Cascadia slow slip event. Geophys. Res. Lett. 38, L18309 (2011).

    Article  Google Scholar 

  6. Kao, H. et al. Spatial–temporal patterns of seismic tremors in northern Cascadia. J. Geophys. Res. Solid Earth 111, B03309 (2006).

    Article  Google Scholar 

  7. Kao, H., Shan, S.-J., Dragert, H. & Rogers, G. Northern Cascadia episodic tremor and slip: a decade of tremor observations from 1997 to 2007. J. Geophys. Res. Solid Earth 114, B00A12 (2009).

    Article  Google Scholar 

  8. Wech, A. G. & Creager, K. C. Automated detection and location of Cascadia tremor. Geophys. Res. Lett. 35, L20302 (2008).

    Article  Google Scholar 

  9. Wang, K., Hu, Y. & He, J. Deformation cycles of subduction earthquakes in a viscoelastic earth. Nature 484, 327–332 (2012).

    Article  Google Scholar 

  10. Goldfinger, C., Nelson, C. H., Johnson, J. E. & Party, T. S. S. Holocene earthquake records from the Cascadia subduction zone and northern San Andreas fault based on precise dating of offshore turbidites. Ann. Rev. Earth Planet. Sci. 31, 555–577 (2003).

    Article  Google Scholar 

  11. Wech, A. G. & Bartlow, N. M. Slip rate and tremor genesis in Cascadia. Geophys. Res. Lett. 41, 392–398 (2014).

    Article  Google Scholar 

  12. Frank, W. B. Slow slip hidden in the noise: the intermittence of tectonic release. Geophys. Res. Lett. 43, 125–10,133 (2016).

    Article  Google Scholar 

  13. Hawthorne, C. J. & Rubin, A. M. Short-time scale correlation between slow slip and tremor in Cascadia. J. Geophys. Res. Solid Earth 118, 1316–1329 (2013).

    Article  Google Scholar 

  14. Brace, W. F. & Byerlee, J. D. Stick–slip as a mechanism for earthquakes. Science 153, 990–992 (1966).

    Article  Google Scholar 

  15. Scholz, C. H. The Mechanics of Earthquakes and Faulting (Cambridge Univ. Press, Cambridge, 2002).

    Book  Google Scholar 

  16. Obara, K. & Kato, A. Connecting slow earthquakes to huge earthquakes. Science 353, 253–257 (2016).

    Article  Google Scholar 

  17. Satake, K., Shimazaki, K., Tsuji, Y. & Ueda, K. Time and size of a giant earthquake in Cascadia inferred from Japanese tsunami records of January 1700. Nature 379, 246–249 (1996).

    Article  Google Scholar 

  18. Aguiar, A. C., Melbourne, T. I. & Scrivner, C. W. Moment release rate of Cascadia tremor constrained by GPS. J. Geophys. Res. Solid Earth 114, B00A05 (2009).

    Article  Google Scholar 

  19. Ito, Y. et al. Episodic slow slip events in the Japan subduction zone before the 2011 Tohoku-Oki earthquake. Tectonophysics 600, 14–26 (2013).

    Article  Google Scholar 

  20. Hasegawa, A. & Yoshida, K. Preceding seismic activity and slow slip events in the source area of the 2011 Mw 9.0 Tohoku-Oki earthquake: a review. Geosci. Lett. 2, 6 (2015).

    Article  Google Scholar 

  21. Kato, A. et al. Propagation of slow slip leading up to the 2011 Mw 9.0 Tohoku-Oki Earthquake. Science 335, 705–708 (2012).

    Article  Google Scholar 

  22. Radiguet, M. et al. Triggering of the 2014 Mw 7.3 Papanoa earthquake by a slow slip event in Guerrero, Mexico. Nat. Geosci. 9, 829–834 (2016).

    Article  Google Scholar 

  23. Ruiz, S. et al. Intense foreshocks and a slow slip event preceded the 2014 Iquique Mw 8.1 earthquake. Science 345, 1165–1169 (2014).

    Article  Google Scholar 

  24. Kostoglodov, V. et al. The 2006 slow slip event and nonvolcanic tremor in the Mexican subduction zone. Geophys. Res. Lett. 37, L24301 (2010).

    Article  Google Scholar 

  25. Rouet-Leduc, B. et al. Estimating fault friction from seismic signals in the laboratory. Geophys. Res. Lett. 45, 1321–1329 (2018).

    Article  Google Scholar 

  26. Hulbert, C. et al. Laboratory earthquake prediction illuminates connections between the spectrum of fault slip modes. Nat. Geosci. (2018).

  27. IRIS DMC FDSNWS Dataselect Web Service (IRIS, Geological Survey of Canada, accessed 1 September 2017);

  28. Murray, J. R. & Svarc, J. Global positioning system data collection, processing, and analysis conducted by the US Geological Survey Earthquake Hazards Program. Seismol. Res. Lett. 88, 916–925 (2017).

    Article  Google Scholar 

  29. Gripp, A. E. & Gordon, R. G. Young tracks of hotspots and current plate velocities. Geophys. J. Int, 150, 321–361 (2002).

    Article  Google Scholar 

  30. Xue, M. & Allen, R. M. The fate of the Juan de Fuca plate: implications for a Yellowstone plume head. Earth Planet. Sci. Lett. 264, 266–276 (2007).

    Article  Google Scholar 

  31. Zhang, J. et al. Cascadia tremor spectra: low corner frequencies and earthquake-like high-frequency falloff. Geochem. Geophys. Geosyst. 12, Q10007 (2011).

    Google Scholar 

  32. Bensen, G. et al. Processing seismic ambient noise data to obtain reliable broad-band surface wave dispersion measurements. Geophys. J. Int. 169, 1239–1260 (2007).

    Article  Google Scholar 

  33. Breiman, L. Random forests. Machine Learning 45, 5–32 (2001).

    Article  Google Scholar 

  34. Holtkamp, S. & Brudzinski, M. Determination of slow slip episodes and strain accumulation along the Cascadia margin. J. Geophys. Res. Solid Earth 115, B00A17 (2007).

    Google Scholar 

  35. Rouet-Leduc, B. et al. Machine learning predicts laboratory earthquakes. Geophys. Res. Lett. 44, 9276–9282 (2017).

    Article  Google Scholar 

  36. Beyreuther, M. et al. Obspy: a python toolbox for seismology. Seismol. Res. Lett. 81, 530–533 (2010).

    Article  Google Scholar 

  37. Pedregosa, F. et al. Scikit-learn: machine learning in python. J. Machine Learning Res. 12, 2825–2830 (2011).

    Google Scholar 

  38. Gregorutti, B., Michel, B. & Saint-Pierre, P. Correlation and variable importance in random forests. Stat. Comput. 27, 659–678 (2017).

    Article  Google Scholar 

  39. Breiman, L., Friedman, J. H., Olshen, R. A. & Stone, C. J. Classification and Regression Trees (CRC Press, New York, 1999).

    Google Scholar 

  40. Jones, D. R., Schonlau, M. & Welch, W. J. Efficient global optimization of expensive black-box functions. J. Global Optim. 13, 455–492 (1998).

    Article  Google Scholar 

Download references


This work was funded by Institutional Support (LDRD) at Los Alamos and the DOE Office of Basic Research, Geoscience Program. We are grateful to H. Dragert, H. Kao, T. Melbourne, J. Gomberg, D. Trugman, R. Guyer, A. Delorey, J. Murray, I. McBrearty and C. Lee for fruitful comments and discussions. We thank W. Frank for his extensive review of our work and his suggestions. We thank T. Cote, X. Jin and M. Kolaj from the CNSN for their data and help.

Author information

Authors and Affiliations



All the authors devised the original study. B.R.L. and C.H. conducted the machine-learning work and all the authors contributed to writing the manuscript.

Corresponding author

Correspondence to Bertrand Rouet-Leduc.

Ethics declarations

Competing interests

The authors declare no competing interests.

Additional information

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Further information on machine learning and GPS displacement and Supplementary Figures 1–6.

Rights and permissions

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Rouet-Leduc, B., Hulbert, C. & Johnson, P.A. Continuous chatter of the Cascadia subduction zone revealed by machine learning. Nature Geosci 12, 75–79 (2019).

Download citation

  • Received:

  • Accepted:

  • Published:

  • Issue Date:

  • DOI:

This article is cited by


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing