Using sparse data to predict lab quakes

A machine-learning approach developed for sparse data reliably predicts fault slip in laboratory earthquakes and could be key to predicting fault slip and potentially earthquakes in the field. The research by a Los Alamos National Laboratory team builds on their previous success using data-driven approaches that worked for slow-slip events in earth but came up short on large-scale stick-slip faults that generate relatively little data—but big quakes.

“The very long timescale between major earthquakes limits the data sets, since major faults may slip only once in 50 to 100 years or longer, meaning seismologists have had little opportunity to collect the vast amounts of observational data needed for machine learning,” said Paul Johnson, a geophysicist at Los Alamos and a co-author on a new paper, “Predicting Fault Slip via Transfer Learning,” in Nature Communications.

To compensate for limited data, Johnson said, the team trained a convolutional neural network on the output of numerical simulations of laboratory quakes as well as on a small set of data from lab experiments. Then they were able to predict fault slips in the remaining unseen lab data.

Read more…