Title: Radiated Energy and State of Stress during Earthquake Rupture

We determined the energy release in earthquakes using the data obtained from TriNet in southern California. We investigated the effects of the depth, the mechanism and the propagation path by examining the results for events with different mechanisms and depths. We found that the path and site effects dominate so much that these effects are not obvious. By far the most dominant is the path-site effects. We determined the station site corrections for the integral of the square of ground motion velocity. The amplification factors are very large: a factor of 10 is common, and for some stations, it is as large as 30. Source directivity can produce a strong azimuthal variation of energy radiation. A good example is the 1992 Landers earthquake in which strong directivity was observed over the frequency band of energy spectrum. We developed a numerical method to correct for this effect. We found that the energy estimate obtained for the Landers earthquake listed in K93 is overestimated by a factor of 2.6. With these corrections, we determined the energy released by earthquakes which occurred in southern California for the period of 1995 to the present. Also, we updated the results for the larger earthquakes (1991 Sierra Madre earthquakes (M=5.8), 1992 Joshua Tree earthquake (M=6.4), 1992 Landers earthquake (M=7.3), 1992 Big Bear earthquake (M=6.4), and the 1994 Northridge earthquake (M=6.7)) applying the newly determined station corrections and directivity effects. The results show that the energy to seismic moment ratio, , for large earthquakes is 10 to 100 times larger than for small earthquakes. This can be interpreted in term of reduction of friction when the slip exceeds a threshold value, about 10 to 50 cm, for large earthquakes. ̃ e