Re:How long should we wait for an overdue earthquake on the ...
by cossatot on Friday July 22, 2016 @11:05AM
(#12143964)
With dice, each roll is independent of the others. With earthquakes, the probability of an earthquake occurring in any given year is (almost always) dependent on how long it has been since the last event. That is represented by the hazard function in the post (lambda t). I say 'almost always' because for a few distributions the hazard rate is constant with time.
An interesting thing about the hazard function with earthquakes is that it isn't always monotonically increasing; I've done this analysis on some other faults that show a very high hazard in the years to decades after an earthquake, and then the hazard drops down and doesn't recover until the far right tail of the distribution. This reflects temporal clustering of earthquakes, which is something we're gradually becoming aware of.
An interesting thing about the hazard function with earthquakes is that it isn't always monotonically increasing; I've done this analysis on some other faults that show a very high hazard in the years to decades after an earthquake, and then the hazard drops down and doesn't recover until the far right tail of the distribution. This reflects temporal clustering of earthquakes, which is something we're gradually becoming aware of.