Most quakes take place along "active margins" where plate subduction is taking place - Japan or the northern west coast. The southern to mid west coast is odd in that it has a transverse fault (plates moving past each other instead of one plate getting subducted under another).
However, on passive margins (no subduction) like the US east coast there are old faults dating back from when it was an active margin. These faults are mostly inactive but... sometimes the activate often to due to deroofing/sediment unloading (as the rock above them erodes the compressive forces reduce), sometimes due to isostatic rebound (from the retreat of ice sheets for example) or even drilling or occasionally mining activity. In this case I'm guessing (will have to look at the data as it comes out in the weeks and months ahead) that and old fault reactivated pretty close to the margin of the piedmont and coastal plain. My money is on erosonial forces finally allowing a fault to reactivate. It may be a reverse fault - in which case the accumulation of sediment on the coastal plain may have had a role.