Implications of a statistical physics approach for earthquake hazard assessment and forecasting

There is accumulating evidence that distributed seismicity is a problem in statistical physics. Seismicity is taken to be a type example of self-organ ized criticality. This association has important implications regarding ear thquake hazard assessment and forecasting. A characteristic of a thermodyna mic system is that it exhibits a background noise that is self-organized. I n the case of a dilute gas, this self-organization is the Maxwell-Boltzmann distribution of molecular velocities. In seismicity, it is the Gutenberg-R ichter frequency-magnitude scaling; this scaling is fractal. Observations f avor the hypothesis that smaller earthquakes in moderate-sized regions occu r at rates that are only weakly dependent on time. Thus, the rate of occurr ence of smaller earthquakes can be extrapolated to assess the hazard of lar ger earthquakes in a region. We obtain the rate of occurrence of earthquake s with rn > 4 in 1 degrees x 1 degrees areas from the NEIC catalog. Using o nly this data we produce global maps of the seismic hazard. Observations al so favor the hypothesis that the stress level at which an earthquake occurs is a second-order critical point. As a critical point is approached, corre lations extend over increasingly larger distances. In terms of seismicity, the approach to a critical point is associated with an increase in the rate of occurrence of intermediate-sized earthquakes prior to a large earthquak e. This precursory activation has been shown to exhibit power-law scaling a nd to occur over a region about ten times larger than the rupture length of the large earthquake. Analyses of the spinoidal behavior associated with s econd-order critical points predict the power-law increase in seismic activ ity prior to a characteristic earthquake. This precursory activation provid es the basis for intermediate-range earthquake forecasting.