Deep learning of aftershock patterns following large earthquakes

Aftershocks are a response to changes in stress generated by large earthquakes and represent the most common observations of the triggering of earthquakes. The maximum magnitude of aftershocks and their temporal decay are well described by empirical laws (such as Bath's law¹ and Omori's law²), but explaining and forecasting the spatial distribution of aftershocks is more difficult. Coulomb failure stress change³ is perhaps the most widely used criterion to explain the spatial distributions of aftershocks^4-8, but its applicability has been disputed^9-11. Here we use a deep-learning approach to identify a static-stress-based criterion that forecasts aftershock locations without prior assumptions about fault orientation. We show that a neural network trained on more than 131,000 mainshock-aftershock pairs can predict the locations of aftershocks in an independent test dataset of more than 30,000 mainshock-aftershock pairs more accurately (area under curve of 0.849) than can classic Coulomb failure stress change (area under curve of 0.583). We find that the learned aftershock pattern is physically interpretable: the maximum change in shear stress, the von Mises yield criterion (a scaled version of the second invariant of the deviatoric stress-change tensor) and the sum of the absolute values of the independent components of the stress-change tensor each explain more than 98 per cent of the variance in the neural-network prediction. This machine-learning-driven insight provides improved forecasts of aftershock locations and identifies physical quantities that may control earthquake triggering during the most active part of the seismic cycle.

Prospects for Larger or More Frequent Earthquakes in the Los Angeles Metropolitan Region

Far too few moderate earthquakes have occurred within the Los Angeles, California, metropolitan region during the 200-year-long historic period to account for observed strain accumulation, indicating that the historic era represents either a lull between clusters of moderate earthquakes or part of a centuries-long interseismic period between much larger (moment magnitude, M(w), 7.2 to 7.6) events. Geologic slip rates and relations between moment magnitude, average coseismic slip, and rupture area show that either of these hypotheses is possible, but that the latter is the more plausible of the two. The average time between M(w) 7.2 to 7.6 earthquakes from a combination of six fault systems within the metropolitan area was estimated to be about 140 years.

Response of High-Rise and Base-Isolated Buildings to a Hypothetical Mw 7.0 Blind Thrust Earthquake

High-rise flexible-frame buildings are commonly considered to be resistant to shaking from the largest earthquakes. In addition, base isolation has become increasingly popular for critical buildings that should still function after an earthquake. How will these two types of buildings perform if a large earthquake occurs beneath a metropolitan area? To answer this question, we simulated the near-source ground motions of a M(w) 7.0 thrust earthquake and then mathematically modeled the response of a 20-story steel-frame building and a 3-story base-isolated building. The synthesized ground motions were characterized by large displacement pulses (up to 2 meters) and large ground velocities. These ground motions caused large deformation and possible collapse of the frame building, and they required exceptional measures in the design of the base-isolated building if it was to remain functional.