BeamKer: Connecting beamform methods and noise source inversion. (version 0.1, Python 3.7)

The following set of Python Jupyter notebooks are meant to illustrate some of the ideas in:

Bowden, D. C., Sager, K., Fichtner A., Chmiel M. (2020). “Connecting Beamforming and Kernel-based Source Inversion”, Geophysical Journal International, external pagehttps://doi.org/10.1093/gji/ggaa539
The paper describes both Beamforming and Matched Field Processing (MFP) algorithms for locating or characterizing seismic noise sources. From there, modifications lead to the full waveform, gradient-based Noise Source Inversion framework used by Sager et al (2020) and others. The Python notebooks here:
• Generate ambient noise raw data and correlations using the Generalized Interferometry package from Fichtner et al (2016). This uses 2D, analytic Green's functions for a homogeneous structure. It also directly recovers noise correlation functions without needing to simulate long, raw timeseries for convergence.
• Implement an example correlation-based Beamforming and MFP directly.
• Implement the first iteration of the Noise Source Inversion framework, again using GI.

All the necessary GI scripts are included here for a complete working example, but users are encouraged to also explore the standalone GI package for a more complete README and numerous examples specific to that package. This can be found also in the Research Software page.

These notebooks are NOT meant to reproduce every figure from the paper. They are also NOT meant to be a production-quality, black-box Beamforming or MFP code. They are meant to illustrate some of the key concepts in the paper, and to illustrate how some of the various algorithms are similar to eachother. Some of the concepts (namely, MFP and Noise Kernels) are only shown for a simple 3-station array, while others (namely, Beamforming and the connection to noise-correlation functions) are shown only for a more realistic 11-station array meant to resemble a subset of stations from Parkfield, CA. See the README.txt for more information.

The Jupyter notebooks here use Python version 3.7, with matplotlib, numpy, and scipy. Again, see the README for more information.

Other references:
Sager, K., Boehm, C., Ermert, L., Krischer, L., & Fichtner, A. (2020). Global-Scale Full-Waveform Ambient Noise Inversion. Journal of Geophysical Research: Solid Earth, 125, 1–17. external pagehttps://doi.org/10.1029/2019JB018644

Fichtner, A., Stehly, L., Ermert, L., & Boehm, C. (2016). Generalised interferometry - I. Theory for inter-station correlations. Geophysical Journal International, ggw420. external pagehttps://doi.org/10.1093/gji/ggw420

Download

BeamKer can be downloaded via the following link: DownloadBeamKer.zip (ZIP, 83.9 MB)