search

cossatot / lanf_earthquake_likelihood

Project on calculating the likelihood of observing a significant low-angle normal fault earthquake over some instrumental time period.
3 stars 9 forks source link

Weight recurrence interval sampling by recurrence interval duration #2

Closed cossatot closed 10 years ago

cossatot commented 10 years ago

Basically, over some 'long' period of time (mutliple seismic cycles), the majority the interseismic strain accumulation time will be winding up for a big event. The smaller events will mostly follow the big ones, and all of their small recurrence intervals shouldn't add up to more than the big events.

I am not sure that this comes across in the PDFs of 'recurrence intervals'. There has to be some way to weight the recurrence intervals in the PDF:

By golly, I think I have it: Take some 'long' earthquake series (hundreds or thousands of events pulled from the frequency-magnitude distribution) and calculate the statistics on what is the time to next event, which is related to what fraction of the total time is in accumulation for earthquakes of each magnitude (or whatever, I know this is a bad way of phrasing it).

This might be essentially the same as taking the F(M) samples and multiplying it by the recurrence interval...

cossatot commented 10 years ago

Maybe the way to do this is by making a CDF of total time vs. magnitude...