2017/02/19 by Bertrand Rouet-Leduc, Bertrand Rouet‐Leduc, Claudia Hulbert +5 · 2 citations
Computer Science · Earth and Planetary Sciences · Physics and Astronomy · #Earthquake Detection and Analysis #Earthquake prediction #Fault (geology) #Noise (video) #Quake (natural phenomenon) #SIGNAL (programming language) #Seismology and Earthquake Studies #Signal processing #Supervised learning #earthquake and tectonic studies #physics.geo-ph
paper · pdf · doi:10.1002/2017gl074677
17 pages, 4 figures
arxiv created 2017/02/19 · openalex created_date 2017/06/23 · openalex publication_date 2017/08/31 · arxiv updated 2017/10/25 · openalex updated_date 2026/08/05
Forecasting fault failure is a fundamental but elusive goal in earthquake science. Here we show that by listening to the acoustic signal emitted by a laboratory fault, machine learning can predict the time remaining before it fails with great accuracy. These predictions are based solely on the instantaneous physical characteristics of the acoustical signal, and do not make use of its history. Surprisingly, machine learning identifies a signal emitted from the fault zone previously thought to be low-amplitude noise that enables failure forecasting throughout the laboratory quake cycle. We hypothesize that applying this approach to continuous seismic data may lead to significant advances in identifying currently unknown signals, in providing new insights into fault physics, and in placing bounds on fault failure times.