Skip to content

Find the earthquakes hidden in your seismic data

Qseek detects and locates earthquakes in large seismic data sets. It stacks machine learning phase annotations along modeled travel times and focuses an adaptive octree on the seismic sources, in continuous archives and in real time.

Get started Configuration

  • Data and AI driven

    Machine learning pickers, trained on large seismic data sets, annotate the phase arrivals. Stacking them over the whole network makes the detection robust.

    How Qseek works

  • Automatic

    From continuous waveforms to located earthquakes with phase picks, magnitudes and features, without manual picking.

    Quick start

  • Extremely fast

    Built for large networks and years of data, with compiled stacking and phase annotation on the GPU.

    Benchmark

  • Geological settings

    Applied to tectonic swarms, volcano-tectonic unrest and induced seismicity at geothermal sites.

    Showcase

Features

  • Machine learning phase detection

    PhaseNet, EQTransformer, OBSTransformer and LFEDetect through SeisBench, on CPU or GPU.

    Image functions

  • Adaptive octree

    The search volume refines itself around the seismic sources, for fast searches and precise locations.

    Search volume

  • 1D and 3D velocity models

    Constant velocity, 1D layered models with fast marching or Pyrocko Cake, and 3D NonLinLoc models.

    Ray tracers

  • Station corrections

    Station-specific and source-specific travel time corrections, extracted from previous runs.

    Station corrections

  • Magnitudes

    Local magnitudes (ML) with regional attenuation models and moment magnitudes (Mw) from modeled peak amplitudes.

    Magnitudes

  • Real-time monitoring

    Stream waveforms from SeedLink servers and send detection alerts to Telegram.

    Real-time monitoring

  • Web UI

    Explore detections, magnitudes and stations in the browser, also for runs on remote machines.

    Explore results

  • Open formats

    Detections as JSON, CSV and Pyrocko markers, and export to HypoDD for double-difference relocation and to VELEST for velocity model inversion.

    Export detections

Cite Qseek

Please cite the Qseek paper when you use it in your work.

Isken, M., Niemz, P., Münchmeyer, J., Büyükakpınar, P., Heimann, S., Cesca, S., Vasyura-Bathke, H., & Dahm, T. (2025). Qseek: A data-driven Framework for Automated Earthquake Detection, Localization and Characterization. Seismica, 4(1). doi:10.26443/seismica.v4i1.1283

@article{isken2025qseek,
  title   = {Qseek: A data-driven Framework for Automated Earthquake
             Detection, Localization and Characterization},
  author  = {Isken, M. and Niemz, P. and M{\"u}nchmeyer, J. and
             B{\"u}y{\"u}kakp{\i}nar, P. and Heimann, S. and Cesca, S. and
             Vasyura-Bathke, H. and Dahm, T.},
  journal = {Seismica},
  year    = {2025},
  volume  = {4},
  number  = {1},
  doi     = {10.26443/seismica.v4i1.1283}
}

Supported by