论文部分内容阅读
极值理论在地震中长期预报与烈度区划中已被广泛应用。1945年以来不少人先后刑Ⅰ型渐近分布和重型渐近分布对许多地区的大震危险作了估计。 本文用我国川、滇、西藏、新疆、甘、青、宁以及台湾等地区的地震资料,把偏度、峰度的估计值与Ⅰ、Ⅲ型渐近分布及正态分布的偏度、峰度进行比较后得出以下结论: (1)用极值理论研究地震规律时正态分布与Ⅲ型渐近分布应受到同样的注意。 (2)我国大陆、台湾(及其邻近海域)的定时段最大震级的分布,以正态分布拟合较好。 (3)川滇、西藏的定时段最大震级的分布以夏型渐近分布拟合较好。 (4)新疆地区使用正态分布或Ⅲ型渐近分布拟合都可以。 (5)Ⅰ型渐近分布对以上地区都不适用。 以上结论和采用绝对差或均方差为尺度进行衡址得到的结论是一致的。
The extreme value theory has been widely used in mid-long term earthquake prediction and intensity zoning. Since 1945, many people have followed the penalty Ⅰ type asymptotic distribution and heavy asymptotic distribution of large areas of earthquake risk has been estimated. In this paper, we use the seismic data of Sichuan, Yunnan, Tibet, Xinjiang, Gansu, Qinghai, Ning and Taiwan to compare the estimates of skewness and kurtosis with the skewness and peak values of type Ⅰ and Ⅲ asymptotic distribution and normal distribution After comparing the degrees, we can draw the following conclusions: (1) The normal distribution and type Ⅲ asymptotic distribution should be treated with the same attention when using the extreme value theory to study the seismic laws. (2) The distribution of the maximum magnitude of the time period in China, Taiwan (and its adjacent waters) is better fitted by the normal distribution. (3) The distribution of the maximum magnitude in the time series of Sichuan, Yunnan and Tibet is better fitted by the summer type asymptotic distribution. (4) Normal distribution or type Ⅲ asymptotic distribution fitting can be used in Xinjiang. (5) I-type asymptotic distribution does not apply to the above regions. The above conclusion is consistent with the conclusion obtained by using absolute difference or mean square error as the scale.