Attempts to understand the physics of earthquakes over the past decade generally have focused on applying methods and theories developed based upon phase transitions, materials science, and percolation theory to a variety of numerical simulations of extended fault networks. These analyses attempt to model and explain many of the important features of the fault system, such as the scaling properties of the system, the variety of periodic, quasi-periodic, and nonperiodic behaviors, and the space-time clustering of events. These results strongly support the view that seismic activity is highly correlated across many space and time scales within large volumes of the earth's crust. Here we describe the analysis of the various tectonic regimes using regional seismicity data using methods applicable to the dynamics of this driven, near mean-field system. In particular, we show that the statistics of fault systems are ergodic over the appropriate spatial and temporal scales, and this property can be used to define the extent and nature of these correlated regions.