The proof she's speaking of probably lies somewhere deep in physics or geophysics that I'm sure I don't have the education to understand at this point.

I know they've said before that all earthquakes large and small start the same - there is nothing different about the start of the quakes just the length of the rupture and thus magnitude of the quake - given there are 30 quakes a day (according to the story) - the question becomes which of those 30 are going to just keep fracturing - there's nothing different about a big one versus a little one.

Maybe the bigger ones are preceeded by minor fractures (and indeed they are most of the time here in socal) - I guess you have to think about it like breaking a pencil in very, very slow constant pressure - can you predict when and how much the pencil will break under pressure? Eventually it breaks all the way, but it starts fracturing before it breaks completely. Slow the pencil break down to about a 1 year time frame and you have it.

Maybe if the following was known and could be modeled: Rate of plate motion, all fault planes and directions (right-lateral, left lateral, thrust, normal, etc) in an area, gps series to give relative motions over an area (becoming more readily available), exact composition of the crust/soil in areas (available from core data for some areas) - strength of the crust/soil materials. But I was pondering all this and even if you had it all, the ground is so fractured with faults that its probably much easier to move on an existing fault so you'd need to know the resistance for each fault (and where they all were) rather than just rock strengths in the area for each fault and it becomes a nightmare of required data to get it right. And then just when you think you have it it breaks somewhere new.

Canie

• News story