Search

Abstract: Observations indicate that earthquake faults occur in topologically complex, multi-scale networks driven by plate tectonic forces. We present realistic numerical simulations, involving data-mining, pattern recognition, theoretical analyses and ensemble forecasting techniques, to understand how the observable space–time earthquake patterns are related to the fundamentally inaccessible and unobservable dynamics. Numerical simulations can also help us to understand how the different scales involved in earthquake physics interact and influence the resulting dynamics. Our simulations indicate that elastic interactions (stress transfer) combined with the nonlinearity in the frictional failure threshold law lead to the self-organization of the statistical dynamics, producing 1) statistical distributions for magnitudes and frequencies of earthquakes that have characteristics similar to those possessed by the Gutenberg–Richter magnitude–frequency distributions observed in nature; and 2) clear examples of stress transfer among fault activity described by stress shadows, in which an earthquake on one group of faults reduces the Coulomb failure stress on other faults, thereby delaying activity on those faults. In this paper, we describe the current state of modeling and simulation efforts for Virtual California, a model for all the major active strike slip faults in California. Noting that the Working Group on California Earthquake Probabilities (WGCEP) uses statistical distributions to produce earthquake forecast probabilities, we demonstrate that Virtual California provides a powerful tool for testing the applicability and reliability of the WGCEP statistical methods. Furthermore, we show how the simulations can be used to develop statistical earthquake forecasting techniques that are complementary to the methods used by the WGCEP, but improve upon those methods in a number of important ways. In doing so, we distinguish between the “official” forecasts of the WGCEP, and the “research-quality” forecasts that we discuss here. Finally, we provide a brief discussion of future problems and issues related to the development of ensemble earthquake hazard estimation and forecasting techniques. [Copyright &y& Elsevier]

In weather forecasting, current and past observational data are routinely assimilated into numerical simulations to produce ensemble forecasts of future events in a process termed “model steering”. Here we describe a similar approach that is motivated by analyses of previous forecasts of the Working Group on California Earthquake Probabilities (WGCEP). Our approach is adapted to the problem of earthquake forecasting using topologically realistic numerical simulations for the strike-slip fault system in California. By systematically comparing simulation data to observed paleoseismic data, a series of spatial probability density functions (PDFs) can be computed that describe the probable locations of future large earthquakes. We develop this approach and show examples of PDFs associated with magnitude M > 6.5 and M > 7.0 earthquakes in California. [ABSTRACT FROM AUTHOR]

Abstract: Topologically realistic simulations of earthquake faults systems have been constructed to understand the physics of interacting earthquake fault systems. We focus on one of these models, a simulation called Virtual California, that represents a model for the strike-slip fault system in California. In weather forecasting, current and past observations are routinely extrapolated forward to forecast future weather. The question addressed in this paper is whether a similar application of numerical simulations can be used in earthquake forecasting. Present simulation models are discussed and their ability to successfully generate earthquake recurrence statistics is demonstrated. An important question relates to how paleoseismic data can be used to constrain simulations, and whether these constrained simulations provide improved forecasts of future earthquakes. Here, we show first results from a consideration of these issues using a method of “datascoring”. The data are divided into “training intervals” and “testing intervals”. In the training intervals, the time history of paleoseismic data are used to evaluate space–time windows of simulations. Earthquakes following high-scoring space–time windows in the simulations are then used as a basis for developing waiting time statistics and used to forecast data in the testing intervals. In our present method, we focus on the problem of determining the timing of future earthquakes having magnitude m >7. Our preliminary conclusion is that the amount of paleoseismic data currently available does not as yet improve the waiting time statistics to a level significantly beyond a random (temporal) predictor. However, this conclusion is based on a set of studies that are not extensive, so further investigations may well reveal important new avenues. In particular, it may be that the true value of this approach lies in defining the probable spatial locations of future earthquakes, rather than their timing. [Copyright &y& Elsevier]