用功率谱时间函数识别地震前兆异常
|
摘要
多年的前兆观测和地震预报实践表明,与前兆观测资料的噪声水平相比,中强地震的前兆信息多属弱信息,信息识别较为困难。用功率谱分析方法,将前兆观测时间函数分解为不同周期的谱密度随时间变化的多个时间函数。通过震例分析,探讨在强干扰背景下识别中强地震前兆信息的可能性。从识别中短期地震前兆这一基本目的出发,采用“数据窗”滑动采样的计算方法,满足不同周期对样本长度的要求,实现了频率域的计算结果在时间域具有一定密度。对首都圈地区10项地电阻率及地下水位资料进行了功率诺分析,资料最长的达24a。对小华北地区M_S>≥5.5地震,唐山老震区M_S≥4.5地震进行了震例分析。结果表明,30d、120d、240d周期的功率谱时间函数,能在原始观测资料中识别出接近噪声水平的中强地震前兆信息。以1989年10月19日大同M_S5.8地震为例,所计算的10项资料中有6项出现明显异常。“单峰”异常峰值以后两个月内发震的比例较高。用功率谱时间函数对前兆观测资料进行时空扫描,当“单峰”异常达到一定概率时,可作为中强地震的中短期前兆异常指标。
Many years' practices of precursor observation and earthquake prediction show that it is more difficult to discern the precursor information because most precursors before moderately strong earthquakes are weak compared with the noise level of observed data. This paper makes use of power spectrum analysis (PSA) to resolve the time function of precursor observation into many time functions with different periods and their spectrum's density changing with time. Taking some earthquakes as example, it makes a study on the possibility of discriminating earthquake precursors from their strongly disturbed background. In order to discern the short- and medium -term earthquake precursors, this paper adopts the method of data window slip sampling to meet the needs of different sample length for different periods. The data processing result of frequency domain has a density to some extent in time field. The PSA is applied to ten items of geoelectrical resistivity and groundwater level data in the Capital Circle. The longest one covers 24 years. It takes earthquakes with Ms≥5.5 in Minor North China and those with Ms≥4.5 in Tangshan earthquake region for example analyses. The result shows that the power spectrum time functions (PSTF) can discriminate precursors before moderately strong earthquakes, which close to noise level, from the original data. Taking earthquake with Ms≥5.8 in Dating on Oct. 19. 1989 as an example, there are distinct anomalies in the six of the ten items through data process. An earthquake takes place more likely within two months after the peak value of a " single peak" anomaly. This paper applies the PSTF to time and space scanning of the observed data, and finds when " single peak" anomaly reachs a certain possibility, it can be the precursor index of short - and medium -term of moderately strong earthquakes.
Many years' practices of precursor observation and earthquake prediction show that it is more difficult to discern the precursor information because most precursors before moderately strong earthquakes are weak compared with the noise level of observed data. This paper makes use of power spectrum analysis (PSA) to resolve the time function of precursor observation into many time functions with different periods and their spectrum's density changing with time. Taking some earthquakes as example, it makes a study on the possibility of discriminating earthquake precursors from their strongly disturbed background. In order to discern the short- and medium -term earthquake precursors, this paper adopts the method of data window slip sampling to meet the needs of different sample length for different periods. The data processing result of frequency domain has a density to some extent in time field. The PSA is applied to ten items of geoelectrical resistivity and groundwater level data in the Capital Circle. The longest one covers 24 years. It takes earthquakes with Ms≥5.5 in Minor North China and those with Ms≥4.5 in Tangshan earthquake region for example analyses. The result shows that the power spectrum time functions (PSTF) can discriminate precursors before moderately strong earthquakes, which close to noise level, from the original data. Taking earthquake with Ms≥5.8 in Dating on Oct. 19. 1989 as an example, there are distinct anomalies in the six of the ten items through data process. An earthquake takes place more likely within two months after the peak value of a " single peak" anomaly. This paper applies the PSTF to time and space scanning of the observed data, and finds when " single peak" anomaly reachs a certain possibility, it can be the precursor index of short - and medium -term of moderately strong earthquakes.
引文
1 国家地震局科技监测司编.PC-1500实用程序及通讯接口.北京:地震出版社,1987．
2 郑治真编.波谱分析基础.北京:地震出版社,1981．
3 冯康编.数值计算方法.北京:国防工业出版社,1978．
4 安鸿志,陈兆国,杜金观,等著.时间系列的分析与应用.北京:科学出版社,1983．
5 冯德益,楼世博,林命周,等编著.模糊数学方法与应用.北京:地震出版社,1983．
2 郑治真编.波谱分析基础.北京:地震出版社,1981．
3 冯康编.数值计算方法.北京:国防工业出版社,1978．
4 安鸿志,陈兆国,杜金观,等著.时间系列的分析与应用.北京:科学出版社,1983．
5 冯德益,楼世博,林命周,等编著.模糊数学方法与应用.北京:地震出版社,1983．
版权所有：© 2023 中国地质图书馆 中国地质调查局地学文献中心