[ckissick]

I used to work at the USGS on an earthquake prediction project in Parkfield, CA. Earthquakes can theoretically be predicted before they happen. There's evidence that there is a "strain event" that precedes the rupture by hours or days.

My project used anomalous water level changes in deep water wells to indicate a strain event. Unfortunately, the correlation between a strain event and a subsequent earthquake appears to sporadic. It can be done, but it's not reliable.

As for a quake in Japan triggering a quake in California, I doubt it. The Hayward fault is overdue, and if it goes next year, it won't be because of what happened in Japan. Faults do affect one-another, and one earthquake has caused another, but only where the faults are within much less than 100 miles of each other.