Researchers have managed to piece together a machine learning algorithm that will help them be better at mapping the Earth’s underground structure. What does this mean? Right now, it can be a tricky process to map the Earth’s underground structure due to a “hum” that the Earth gives out. This can interfere with low-frequency waves that would be used to map under the Earth.
However, by turning to the use of AI, researchers can now use earthquakes to help estimate those low-frequency waves and take advantage of it to map underground structures.
According to the study’s co-author Laurent Demanet, professor of applied mathematics at MIT, “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. Now we’ve shown that deep learning offers a solution to be able to fill in these missing frequencies.”
As to why we would want to map the Earth’s underground structures, besides the obvious reasons such as enlightening ourselves, it can be used to discover resources that might be hidden, as well as maybe potential pockets of space where captured CO2 can be stored.