The lead author of the paper, John Rundle, works as the director for the Colorado Center for Chaos and Complexity, focussing on finding patterns in complex systems. As it says on their site, "These new paradigms are characterized by the growing conviction that many of the important properties of natural systems cannot be understood from studying the individual system components in isolation nor from a top-down, reductionist approach. Rather, these new approaches are distinguished from more traditional ones by the notion that these important properties are "emergent" attributes of the system as a whole."
What this means if you apply that idea to earthquakes, for example, is that you can't predict them by how often they tend to hit a given area on average, but rather on how earthquakes manifest themselves in relation to a larger system -- in this case, a large system of geologic faults. This is why those who study earthquakes were frustrated when, after building a really incredible earthquake monitoring center at Parkfield, California, the major quake they were expecting never happened.
Earthquakes manifest themselves by simultaneously releasing and increasing stress on segments of the earth's crust, so any good predictions have to model these stresses appropriately. Well, they took a look at earthquakes in South and Central California, ran a ton of data through a supercomputer, and came up with there predictions for the areas most at risk for magnitude 5 or greater quakes over the next 10 years.
Well, it turns out that since they released the report, they've successfully identified the location of four earthquakes, including the latest quake in Gilroy, California.
If their data is correct, it looks like there will be a lot of earthquakes over the next decade in Southern California... specifically around Bishop, Ridgecrest, Coalinga, Barstow, the area north of the 1992 Landers quake, the area NW of Palm Springs, Oceanside, Superstition Hills, and the Imperial Valley. The highest risk areas for magnitude 7 quakes are Coalinga, Landers, and the Imperial Valley. Northern California, however, appears to get off easy this decade.
Oh... and it looks likely that Parkfield will finally get that quake they've been waiting for. Maybe, with a better idea of the interrelatedness of how faults work, combined with monitoring for the most minute signals of an impending earthquake, they just might learn something about both the macro and the micro of what makes quakes tick.