Part 6/10:
Buoyed by this success, researchers then tackled real-world faults, focusing on slow slip events—gentle, slow earthquakes that occur over days or weeks. These events, occurring beneath Vancouver Island, provided a natural analogue for laboratory signals. Seismic data from these slow slips were fed into the same machine learning models, which, once trained, successfully projected the timing of future slip events.
Remarkably, findings from both the laboratory and field data produced consistent results, suggesting that the same underlying principles could bridge the vast scale gap between controlled experiments and Earth's complex systems.