Earthquake Forecasting as a System-Science Problem

The increasing exposure of society to natural hazards has made the forecasting of extreme events a pressing scientific concern. No aspect of this problem has been more vexing than earthquake prediction. The century-long failure to identify observable precursory signals diagnostic of impending events has led to an alternative approach, pioneered by earthquake engineers, in which a variety of constraints on earthquake location, magnitude, and long-term frequency are synthesized into probabilistic seismic hazard models, such as those produced by the USGS National Seismic Hazard Mapping Project. This presentation will describe how recent progress in earthquake system science is improving hazard and risk forecasting. These system-level problems can be partitioned according to causal sequences described in terms of conditional probabilities. For example, the exceedance probabilities of shaking intensities at geographically distributed sites conditional on a particular fault rupture (a ground motion prediction model or GMPM) can be combined with the probabilities of different ruptures (an earthquake rupture forecast or ERF) to create a seismic hazard map. Deterministic simulations of ground motions from very large suites (millions) of ruptures, now feasible through high-performance computational platforms such as SCEC's CyberShake, are allowing seismologists to replace empirical GMPMs with physics-based models that more accurately represent wave propagation through heterogeneous geologic structures, such as sedimentary basins that amplify seismic shaking. A notable advance is the development of ERFs conditioned on preceding seismic activity, such as the Uniform California Earthquake Rupture Forecasts produced by the Working Groups on California Earthquake Probabilities. These time-dependent probability models account for the stress-renewal processes of elastic rebound, and they are beginning to capture aftershock triggering. However, they have not fully reconciled the long-term phase modulation of stress renewal with the short-term clustering observed in aftershocks and extended seismic sequences. For this and other reasons, the stochastic ERFs are likely to be replaced by deterministic earthquake simulators that explicitly model rupture nucleation and stress evolution within fault systems. Several outstanding issues will be highlighted, such as the assimilation of historical seismicity data into simulator-based forecasting and the problem of forecasting rupture directivity, which can strongly influence ground motions.

Mackenzie Room 300