Update: A more detailed analysis of this earthquake can be found here
Japan was hit by a couple of large-ish earthquakes yesterday. According the USGS moment tensor solutions, the first magnitude 6.6 was caused by the rupture of a normal (extensional) reverse (high angle contractional – thanks Kim!) fault just off the west coast of Japan at a depth of about 8 km. About 12 hours later, a magnitude 6.8 occurred in the same region, but at a much greater depth of 360 km or so, which means that it has to have been within the cold, westward dipping slab of Pacific plate that is being subducted beneath the east coast of Japan (at that depth, normal mantle is much too hot and weak for brittle fracturing to occur).
Note that the quake that caused all the damage (some photos from BBC news) was actually the weaker of the two; the other one was so deep that the seismic energy had probably mainly dissipated by the time it got to the surface.
Perhaps the most worrying news was the report that the quake caused a fire at a large nuclear power plant, and a leak of water containing “a small amount of radioactive material” (whatever that means). This does not fill me with optimism for the future, because these quakes were nothing to do with the subduction megathrust off the east coast of Japan, which is capable of producing much larger earthquakes. Whatever your wider thoughts on the general safety/desirability of nuclear power, not building them next door to giant earthquake-generating faults is surely a bit of a no-brainer.
Thermochronic also has some comments up.