TO identify resources such as oil reserves, geothermal sources and, more recently, reservoirs where excess carbon dioxide could potentially be sequestered, scientists have evolved methods to map the structures within the earth’s crust. They do so by tracking seismic waves that are produced naturally by earthquakes or artificially using explosives or underwater air guns. The manner in which these waves reflect and scatter through the earth gives scientists an idea of the type of structures that lie beneath the surface.
There is a narrow range of seismic waves—those that occur at low frequencies of around 1 hertz—that could give scientists the clearest picture of underground structures spanning wide distances. But these waves are often drowned out by the earth’s noisy seismic hum and are, therefore, difficult to pick up with current detectors. Now researchers at the Massachusetts Institute of Technology (MIT) have come up with a technique based on machine learning in a neural network system to fill this gap.
In a paper appearing in the journal “Geophysics”, the scientists have described a method in which they trained a neural network on hundreds of different simulated earthquakes. When the researchers presented the trained network with only the high-frequency seismic waves produced from a new simulated earthquake, the neural network was able to imitate the physics of wave propagation and estimate accurately the quake’s missing low-frequency waves.
The new method could allow researchers to synthesise artificially the low-frequency waves that are hidden in seismic data, which can then be used to map more accurately the earth’s internal structures. “The ultimate dream is to be able to map the whole subsurface, and be able to say, for instance, ‘this is exactly what it looks like underneath Iceland, so now you know where to explore for geothermal sources,’” said Laurent Demanet, applied mathematician and one of the paper’s authors, in an MIT press release. “Now we’ve shown that deep learning offers a solution to be able to fill in these missing frequencies.” Demanet’s co-author is lead author Hongyu Sun, a graduate student at MIT.
A neural network is a set of algorithms modelled loosely after the neural workings of the human brain. The algorithms are designed to recognise patterns in data that are fed into the network, and to cluster these data into categories, or labels. Hongyu Sun and Demanet adapted a neural network for signal processing, specifically, to recognise patterns in seismic data.
They have reasoned in their paper that if a neural network was fed enough examples of earthquakes, and the ways in which the resulting high- and low-frequency seismic waves travel through a particular composition of the earth, the network should be able to “mine the hidden correlations among different frequency components” and extrapolate any missing frequencies if the network were only given an earthquake’s partial seismic profile.
As they improve the neural network’s predictions, the team hopes to be able to use the method to extrapolate low-frequency signals from actual seismic data, which can then be plugged into seismic models to map more accurately the geological structures below the earth’s surface. The low frequencies, in particular, are a key ingredient for solving the big puzzle of finding the correct physical model.