An application of regional time and magnitude predictable model for long-term earthquake prediction

A regional time and magnitude predictable model has been applied to estimate the recurrence intervals for large earthquakes in the vicinity of 8 October 2005 Kashmir Himalaya earthquake (25-40N and 65-85E), which includes India, Pakistan, Afghani- stan, Hindukush, Pamirs, Mangolia and Tien-Shan. This region has been divided into 17 seismogenic sources on the basis of certain seismotectonics and geomorphological criteria. A complete earthquake catalogue (historical and instrumental) of magnitude Ms C 5.5 during the period 1853-2005 has been used in the analysis. According to this model, the magnitude of preceding earthquake governs the time of occurrence and magnitude of future mainshock in the sequence. The interevent time between successive mainshocks with magnitude equal to or greater than a minimum magnitude threshold were considered and used for long-term earthquake prediction in each of seismogenic sources. The interevent times and magnitudes of mainshocks have been used to determine the following predictive relations: logTt = 0.05 Mmin ? 0.09 Mp - 0.01 log M0 ? 01.14; and Mf = 0.21 Mmin - 0.01 Mp ? 0.03 log M0 ? 7.21 where, Tt is the interevent time of successive mainshocks, Mmin is minimum magnitude threshold considered, Mp is magnitude of preceding mainshock, Mf is magnitude of following mainshock and M0 is the seismic moment released per year in each seismogenic source. It was found that the magnitude of following mainshock (Mf) does not depend on the interevent time (Tt), which indicates the ability to predict the time of occurrence of future mainshock. A negative correlation between magnitude of following mainshock (Mf) and preceding mainshock (Mp) indicates that the larger earthquake is followed by smaller one and vice versa. The above equations have been used for the seismic hazard assessment in the