余震序列性质判定单参数判据的统计研究

Statistic study on the criterion index for classification of aftershock sequences

收集整理了1970～2004年中国大陆地区293次记录相对完备的M0≥5．0地震序列，其中孤立型地震68次、主余型地震174次、多震型地震52次。选择h值、b值、能量熵、震级差等9个表征序列特征的参数，分震后不同时段进行计算。在0．05的置信水平下，对每一参数及规定的主震后时段，分孤立型地震、主余型地震及多震型地震三类统计其总体均值及置信区间，以此作为序列类型划分的判据。在此基础上对序列参数的变化特征及其序列分类能力进行讨论。总的来看，所选参数大多具有一定的序列分类能力，其中部分参数数值分布范围与主震震级有关，主震不同则序列分类标准有差异，如能量熵、时间熵、序列主震与最大或次大余震的震级差等参数；部分参数数值分布范围随时间变化，震后不同时段有不同的序列分类判据，如归一化频次、h值等参数；参数的有效时段存在差异，并且部分参数始终无序列分类能力，如b值、归一化能量、平均震级等。对有一定序列分类能力的参数，给出具体的适用条件及与时间或序列主震震级相关的判别指标。平均来看，能量熵、主震与统计时段内最大地震的震级差等参数具有相对较高的识别准确率。

294 earthquake sequences with magnitude greater than or equal to 5.0 occurred in China＇s mainland from 1970 to 2003 have been collected. Among of them, 68 are isolated earthquakes, 174 are mainshock-aftershock sequences and 52 belong to the multiple mainshock type. 9 parameters, such as h value, b value, energy entropy and so on, of these sequences have been calculated in different periods after the mainshock. With the given confidence level, a=0.05, average value of each parameter and its confidence limitation have been calculated for each sequence types, which are taken as the criterion index for classification of aftershock sequences. Based on the data, the changing characteristics of parameters and their capability for classification of aftershock sequences have been discussed. Generally, most selected parameters express the capability for classification of aftershock sequences to a certain extent. Among them, the numerical range of some parameters （such as energy entropy, time entropy, difference of the magnitude and so on） is concerned with the magnitude of mainshock and the classification rule is different when the mainshock are distributed in different magnitude range. Some parameters （such as normalized frequency, h value and so on） change with the time and there is different classification rule in different period after the mainshock. The time range, in which the parameter is available for classification of aftershock sequences, is different for different parameters, and some parameters seldom have the ability to classify the aftershock sequences, such as b value, normalized energy, average magnitude, and so on. For parameters which have a certain capability for classification of aftershock sequences, the suitable condition for daily use and the real judgment index which concerning with the lapse time from the mainshock and magnitude of mainshock have been given. Roughly and averagely, several parameters, such as energy entropy, magnitude difference between mainshock and the largest aftershock and so on, have a high identification rate for classification of aftershock sequences.