From the analysis of local earthquake coda waves recorded by the Benioff short-period vertical seismograph at Riverside, California, we found a surprisingly large and systematic temporal variation in Coda Q-1 in the southern California during the 55-year period from 1933 to 1987. Comparison of the coda Q-1 variation with the b value obtained for earthquakes with M≥3 occured in the area within 180 km from Riverside (corresponding to the area for which Q-1 is estimated by the coda method), led to an extraordinary result that coda Q-1 correlates positively with b value with the correlation coefficient of 0.79. The significance of correlation cannot be rejected at the confidence level of 0.974. Since the data used for measuring Q-1 (namely, the coda part of seismograms) and the data used for measuring b (namely, the seismicity catalog) are totally independent, the observed correlation supports strongly the physical reality of temporal change in both coda Q-1 and b value. The correlation between coda Q-1 and b value has been reported for several other cases, but the correlation was positive in some cases and negative in others. Our current preferred model for explaining the observed correlation is the following. Aseismic creep is going on from time to time over various sizes of cracks in the ductile part and, in some cases, in the brittle part of lighosphere at and near the plate boundary. Creep activites tend to increase the crack density and coda Q-1 observed in the region. If the creek occurs over distibuted cracks with a certain predominant scale length, spatial stress concentration may enhance seismicity for earthquakes with a similar size. If this characteristic size is in the lower part of the magnitude range for which the b value is estimated, the enhanced seismicity will increase b. On the other hand, if it is in the upper part of the magnitude range, it will decrease b. Thus we can explain the observed strong correlation between coda Q-1 and b value for both positive and negative cases. The existence of such a characteristic crack size in a seismic region has been supported by various observations on earthquake scaling. ¿ American Geophysical Union 1989 |