Abstract: In order to quantitatively analyze the reliability of the composite fault plane solution of small earthquakes, the Bootstrap sampling technique is introduced into the grid search method, and the ideas and methods for calculating the confidence interval from the grid search method are proposed initially. There are two sample sets that can represent the characteristics of the composite fault plane solution. One sample set is the optional solutions obtained by the grid search method and the other is obtained by the Bootstrap sampling technique. Then, we calculate the confidence intervals of the two sample sets (P, B and T axis). The research results of tectonic stress field in southern Jiangxi Province are relatively few. In view of such situation, we use the focal mechanism solutions of small earthquakes to calculate and analyze the composite fault plane solution and the confidence interval. This study shows that the confidence interval of the principal stress axis can be obtained well by both of the sample methods. The reliability of the results and the confidence range of the principal stress axis can be better represented by the confidence intervals. The middle principal stress in southern Jiangxi Province is nearly vertical, and the maximum and minimum principal stresses are nearly horizontal. The direction of maximum principal stress is NWW-SEE and that of the minimum principal stress is NNE-SSW. And, the area is in a strike-slip stress regime. The results are consistent with previous studies, and the stress directions obtained by previous researchers are within the confidence interval calculated by this paper.

Key words: tectonic stress field, confidence interval, Bootstrap sampling, southern Jiangxi Province, composite fault plane solution

摘要： 为定量分析小震综合断层面解计算结果的可靠性，将Bootstrap抽样技术引入到格点尝试法的计算中，初步提出了对格点尝试法求解结果进行置信区间计算分析的思路和方法，即以2种方式来确定表示综合断层面解分布特征的样本集合：一种是以格点尝试法给出的可选解作为样本集合，另一种是引入Bootstrap抽样技术来获取的样本集合；然后借助分位数法对这2种样本集合（综合断层面解的P、B、T轴）的置信区间进行计算。针对赣南地区构造应力场研究成果相对较少这一状况，文中在回顾已有成果的基础上，利用近期的小震震源机制解数据，对赣南地区综合断层面解及其置信区间进行了计算分析。结果表明，2种样本方式均可以较好地获得主应力轴的置信区间；置信区间能够更好地定量展示格点尝试法计算结果的可靠性以及主应力轴的置信范围；赣南地区中间主应力轴近垂直（倾角95%的置信范围为45°～90°），最大和最小主应力轴近水平（倾角95%的置信范围分别为0°～40°、0°～20°），最大主应力轴方位为SEE向（方位95%的置信范围为110°～140°），最小主应力轴方位为NNE向（方位95%的置信范围为15°～45°），为走滑应力类型。该结果不仅与前人的计算分析结果相一致，并且前人给出的赣南地区应力方向均位于文中给出的置信区间内。

关键词: 构造应力场, 置信区间, Bootstrap抽样, 赣南地区, 综合断层面解