Communications of the ACM
Applying Machine Learning Tools to Earthquake Data Offers New Insights
Columbia University researchers using machine learning algorithms have found a connection between earthquakes experienced at The Geysers geothermal field in California and water injected into the hot rocks underground.
Credit: cloudfront.net
Machine learning algorithms can identify different types of earthquakes from three years of seismic recordings at The Geysers in California, according to Columbia University researchers.
The team found the repeating patterns of earthquakes at The Geysers one of the world's oldest and largest geothermal fields, appear to match the seasonal rise and fall of water-injection flows into the hot rocks below, suggesting a connection to the mechanical processes that cause rocks to slip or crack, triggering an earthquake.
The researchers amassed a catalog of 46,000 earthquake recordings, each represented as energy waves in a seismogram. They then measured changes in the waves' frequency through time, and applied machine learning tools traditionally used for extracting patterns from music and human speech to identify spectral "fingerprints" in the quakes.
The machine learning program helped the researchers find the connection to the fluctuating amounts of water injected below ground during the energy-extraction process.
From Lamont-Doherty Earth Observatory
View Full Article
Abstracts Copyright © 2018 Information Inc., Bethesda, Maryland, USA
No entries found