Seismic noise tomography

This project is dedicated to provide a Python framework for seismic noise tomography, based on ObsPy and numerical Python packages such as numpy and scipy.

Requirements

The code is developped and tested on Ubuntu (but should run on other platforms as well) with Python 2.7.

In addition to Python 2.7, you need to install the following packages:

• numpy >= 1.8.2
• scipy >= 0.13.3
• matplotlib >= 1.3.1
• ObsPy >= 0.9.2
• pyshp
• pyproj >= 1.8.9
• pyPdf

It is recommended to install these packages with pip install ... or with your favourite package manager, e.g., apt-get install ....

Optionally, you may want to install the Computer Programs in Seismology to be able to invert your dispersion maps for a 1-D shear velocity model, as these programs take care of the forward modelling.

How to start

You should start reading the example configuration file, tomo_Brazil.cnf, which contains global parameters and detailed instructions. You should then create your own configuration file (any name with cnf extension, *.cnf) with your own parameters, and place it in the same folder as the scripts. It is not advised to simply modify tomo_Brazil.cnf, as any update may revert your changes.

You may then use the scripts in the following order:

• crosscorrelation.py takes seismic waveforms as input in order to calculate and export cross-correlations between pairs of stations,

• dispersion_curves.py takes cross-correlations as input and applies a frequency-time analysis (FTAN) in order to extract and export group velocity dispersion curves,

• tomo_inversion_testparams.py takes dispersion curves as input and applies a tomographic inversion to produce dispersion maps; the inversion parameters are systematically varied within user-defined ranges,

• tomo_inversion_2pass.py takes dispersion curves as input and applies a two-pass tomographic inversion to produce dispersion maps: an overdamped inversion is performed in the first pass in order to detect and reject outliers from the second pass.

• 1d_models.py takes dispersion maps as input and invert them for a 1-D shear velocity model at selected locations, using a Markov chain Monte Carlo method to sample to posterior distribution of the model's parameters.

The scripts rely on the Python package pysismo, which must thus be located in a place included in your PATH (or PYTHONPATH) environment variable. The easiest choice is of course to place it in the same folder as the scripts.

How to update

The code is still experimental so you should regularly check for (and pull) updates. These will be backward-compatible, except if new parameters appear in the configuration file.

In other words, after any update, you should check whether new parameters were added to the example configuration file (tomo_Brazil.cnf) and insert them accordingly to your own configuration file.

References

The cross-correlation procedure of ambient noise between pairs of stations follows the steps advocated by Bensen et al. (2007). The measurement of dispersion curves is based on the frequency-time analysis (FTAN) with phase-matched filtering described in Levshin and Ritzwoller (2001) and Bensen et al. (2007). The tomographic inversion implements the linear inversion procedure with norm penalization and spatial smoothing of Barmin et al. (2001). The Markov chain Monte Carlo method is described by Mosegaard and Tarantola (1995), and the forward modelling is taken care of by the Computer Programs in Seimology (Herrmann, 2013).

• Barmin, M. P., Ritzwoller, M. H. and Levshin, A. L. (2001). A fast and reliable method for surface wave tomography. Pure Appl. Geophys., 158, p. 1351–1375. doi:10.1007/PL00001225 [journal] [pdf]

• Bensen, G. D. et al. (2007). Processing seismic ambient noise data to obtain reliable broad-band surface wave dispersion measurements. Geophys. J. Int., 169(3), p. 1239–1260. doi:10.1111/j.1365-246X.2007.03374.x [journal] [pdf]

• Herrmann, R. B., 2013. Computer Programs in Seismology: an evolving tool for instruction and research, Seismol. Res. Let., 84(6), p. 1081-1088 doi: 10.1785/0220110096 [pdf]

• Levshin, A. L. and Ritzwoller, M. H. (2001). Automated detection, extraction, and measurement of regional surface waves. Pure Appl. Geophys., 158, p. 1531–1545. doi:10.1007/PL00001233 [journal] [pdf]

• Mosegaard, K. and Tarantola, A. (1995) Monte Carlo sampling of solutions to inverse problems, J. Geophys. Res., 100(B7), p. 12431–12447 [journal] [pdf]

Publications

Please let me know of your published works making use of this project.

• Goutorbe, B., Coelho, L.O. and Drouet, S. (2015). Rayleigh wave group velocities at periods of 6–23 s across Brazil from ambient noise tomography. Geophys. J. Int., 203, 869–882. doi:10.1093/gji/ggv343 [journal] [pdf]