| Abstract: | Statistics of ultimate strain are improved by adding new data to the previous ones. The critical value for horizontal strain seems somewhat larger than that for vertical strain, although parameters of a Weibull distribution, which is customarily used for quality-control research and which fits in very well with the present statistics, are calculated for the whole set of data making no distinction between the two subsets because of their scantiness.On the basis of the parameters thus determined and strain rates obtained from geodetic data, probabilities of earthquake occurrence in a few regions in Japan and the U.S. are estimated. Probability of having an earthquake in an area southwest of Tokyo, where we had the 1923 earthquake (magnitude 7.9), at this time amounts to 20%, a value almost the same as that obtained in the previous papers. The probability will reach some 50 and 90% by 2000 and 2050, respectively. In the North Izu district, where an earthquake of magnitude 7.0 occurred in 1930, a shearing crustal motion is going on to an extent for which we have a probability for an earthquake recurring there in these 40 years amounting to 40%. By the end of this century, it will become as high as 85%.Similar estimates of such cumulative probabilities are made for the San Francisco and Fort Tejon regions, where great earthquakes occurred respectively in 1906 and 1857, yielding values of 30 and 80% at present. These probabilities are tentative because of possible errors in evaluating geodetic measurements and uncertainty of the ultimate crustal strain assigned to the San Andreas fault. |
