2 CURRENT SEISMIC QUIESCENCES WITHIN 40 KM OF TOKYO

Two highly significant seismic quiescence anomalies that started in 19 93.0 exist currently within about 40km of Tokyo. They are located in c entral Saitama province at 35.99 degrees N/139.56 degrees E and in Tok yo Bay, south of Funabashi, at 35.59 degrees N/139.94 degrees E, The r adii of the anomalous volumes are approximately 15 km, centred at abou t 25 and 20 km depth, respectively. The seismicity rate during the qui escence was less than 10 per cent background rate in both cases. The s ignificance of these quiescences, measured by the standard deviate Z-t est, comparing the rate of the last three years to the backgound rate, is higher than for any other rate changes present in the catalogue (Z (max) = 7.5 and 7.2 respectively). However, episodes of highly signifi cant quiescences, which could have been interpreted as precursors but were not followed by main shocks, also exist in the data set. A third seismic quiescence was found to be in progress since 1993.3 +/- 0.5 in NW Saitama province at 36.18 degrees N/139.22 degrees E. It has a rad ius of 14 km and is less significant than the other two anomalies (Z(m ax) = 6.8). In a polygon that covers the areas of all three quiescence s defined above, the seismicity rate since 1994 is only 30 per cent of the previous background rate (i.e. 100 earthquakes that would normall y have occurred did not). This analysis is based on the earthquake cat alogue of the National Institute for Earth Science and Disaster Preven tion for the years 1980-1996.2 and M(min) = 1.0, as well as for the ye ars 1986.2-1996.2 with M(min) = 0.1. The anomalies are equally clear i n the clustered and in the full catalogues. To avoid contamination by explosions, only nighttime data (19:00 to 7:00 hours local time) were used. The frequency-magnitude distributions in the quiescent volumes a t central Saitama, NW Saitama and Funabashi have b-values of 1.3, 1.5 and 1.1, respectively. Based on the hypothesis that crustal volumes wi th high b-values may not be capable of M6 + earthquakes, these three q uiescent volumes may be ruled out as source volumes for main shocks of these magnitudes. Using the seismic quiescence hypothesis alone, it i s a reasonable interpretation to suggest that M = 6.5 +/- 0.5 earthqua kes are expected to terminate each of the three quiescences within 1 y ear +/- 1 year from now, with a 50 per cent probability each. The inte rpretation that all three quiet volumes could rupture together in a la rge earthquake seems less likely. However, based on the information th at the quiescences are located in volumes with b-values above the aver age b = 0.8, it can be argued that the three current quiescences are f alse alarms. The fact that these two lines of evidence allow opposing interpretations in these cases shows that neither hypothesis is suffic iently tested. Thus, the quiescence hypothesis is not advanced enough for issuing earthquake predictions. Nevertheless, it is important that scientific tests of prediction scenarios are formulated in real time, so we can learn how to modify our hypotheses to make them useful for prediction in the future.