Laserfiche WebLink
1~ <br /> the 100 year exposure period occurred immediately after the <br />l <br />h <br />k <br /> ast eart <br />qua <br />e, or immediately before the next - the <br /> probability would be the same. <br />t There is ample evidence, however, which suggests that <br /> earthquakes that accompany surface faulting are correlated in <br />1 location, size and time (Mogi, 1968; Sykes, 1971; and Kelleher <br /> et a1., 1974). Physically, great earthquakes are caused by <br />' the gradual accumulation of strain (approximately 0.3 mm/yr <br /> for the Rio Crande Rift) interrupted intermittently by abrupt <br />' episodes of rapid energy release. It makes physical sense that <br />h <br />L <br /> t <br />e <br />onger a particular fault segment has been storing strain <br /> energy, the more Likely it is that the energy will be <br />' released within a given period of time. Such a view is especially <br /> appropriate for a situation such as the San Luis Valley where <br />1 seismic activity appears to be nonexistant between surface <br /> faulting events. <br />' An approach that accommodates temporal correlation of <br /> great earthquakes is the semi-Markhov method proposed by <br /> Patwardhan et al. (1980). Briefly, the idea is that a great <br /> earthquake having magnitude M <br />defines a state of the earth- <br /> i <br />. quake process (state i). Once we have entered state i (ie. an <br />' earthquake having magnitude Mi has just occurred), we choose <br /> a new state, state j, which defines the magnitude of the next <br />' earthquake (M <br />). Before moving on to state j, however, the <br /> j <br /> process holds in state i for some "holding time" t <br />which <br /> i <br />j <br /> is the likely time between earthquakes of magnitude M. and <br />' ~. <br /> M <br />respectively. Adjustments can be made to account for <br /> j <br />' the fact that we may come upon the process some time t:0 after <br />i <br />h <br />d <br />h <br />h <br />h <br />k <br /> t <br />as entere <br />t <br />e current state. In our case t <br />e eart <br />qua <br />e <br /> process on the San Luis fault segment has been in its current <br /> state for between 1,940 and 4,715, we don't know for ::ure. We <br /> also have only a fuzzy idea of precisely the state we are in <br /> (the magnitude of the Last earthquake). Patwardhan et a1. (1980) <br /> give the probability that the earthquake process, which has <br /> been in state i for t0 years, will make K1 transitions to <br /> <br />13 <br />