When gathering data for seismic records, seismologists usually place sensor arrays as far as possible from railways to avoid unwanted noise. In the past decade, however, researchers have begun to explore the possibility of using seismic energy generated by rumbling freight trains to image Earth’s crust, where changes in the transmitted wave’s velocity could indicate geological boundaries, faults, and mineral deposits.
The train-rumble technique struggles to image shallow parts of the crust, however, where seismic noise is dominated by waves that travel along the surface rather than those that travel through the layers of interest. Laura Pinzon-Rincon, François Lavoué, and their colleagues at the Grenoble Alpes University in France have overcome that limitation by developing a method for pinpointing the high-frequency humming signal from freight trains. Using the new method, the researchers mapped a mineral-rich formation located near a railway. The technique provides a low-cost method for shallow crust imaging, especially in regions that lack other reliable sources of high-frequency seismic noise. The research was presented in two talks at the Fall Meeting of the American Geophysical Union, which was held virtually in December.
When the weight of moving freight trains pushes down on the ground, they generate tremors equivalent to earthquakes of up to magnitude 2 that can be detected up to 100 km away. To separate those tremors from other sources of noise, researchers first had to identify both the mechanism behind train-generated seismic noise and its characteristic spectral signature. Based on observations in previous studies of equidistant peaks in spectrograms of passing freight trains traveling at different speeds, Lavoué proposed that seismic waves are emitted by regularly spaced wheels that excite harmonic vibrations in stationary horizontal railway ties. The resulting frequency depends on the train’s velocity, length, and wheel spacing. Most of the energy transmitted as a pressure wave comes from the high-frequency harmonics, rather than the fundamental frequencies of 1 to 3 Hz.
Equipped with those models, Pinzon-Rincon and her colleagues developed a method for extracting high-energy signals from an array of 1000 seismometers placed within 8 to 15 km of a railway and located near a platinum-rich mining site in Ontario, Canada. From computational models and 30 days of recordings, they picked out the high-frequency pressure waves generated by the trains and are using them to obtain three-dimensional images of the seismic velocity distribution in the environment. Seismic velocity depends mainly on the properties of the local rock. Interpreting the 3D image in terms of geology and lithology resulted in a map of a hundreds-of-meters-thick gabbro formation that’s known to host a mineral intrusion.
Given the extensive railroad networks in Asia, Europe, and North America, quantifiable seismic signals are likely to be widely available. A railway-based network provides an opportunity for seeking mineral deposits and subsurface water, observing the structural integrity of piers and tunnels, and monitoring active fault zones. (F. Lavoué et al., Seis. Res. Lett. 92, 287, 2021; L. Pinzon-Rincon et al., Seis. Res. Lett., in press.)