Rapid source characterization of the Maule earthquake using Prompt Elasto-Gravity Signals

Abstract

The recently identified Prompt Elasto-Gravity Signals (PEGS), generated by large earthquakes, propagate at the speed of light and are sensitive to the earthquake magnitude and focal mechanism. These characteristics make PEGS potentially more accurate than P-wave based early warning algorithms (which produce saturated magnitude estimations) and faster than Global Navigation Satellite Systems (GNSS)-based systems. We use a deep learning model called PEGSNet, originally developed for application in Japan, to track the temporal evolution of the magnitude of the 2010 Mw 8.8 Maule earthquake. The model is a Convolutional Neural Network (CNN), trained on a database of synthetic PEGS -- simulated for an exhaustive set of possible earthquakes distributed along the Chilean subduction zone -- augmented with empirical noise. The approach is multi-station and leverages the information recorded on all the available stations to estimate as fast as possible the magnitude and location of an on-going earthquake. Our results indicate that PEGSNet could have estimated an  Mw > 8.7 earthquake after 100 seconds in the Maule case. Our synthetic tests using real data and noise recordings further support the instantaneous tracking of the source time function of the earthquake.