摘要：
本文介绍了地电阻率时间序列数据的常用处理方法,即消除年变化法、无量纲法、相对均方差法和差分法共4类8种方法,包括“九五”以来对原方法的改进和新发展的方法,讨论了各方法的原理、数据处理过程、异常分辨能力、异常指标和异常物理机制及其在数据处理中存在的不足.结果显示:①一般来说,用有效的数据处理方法才能分析、识别出原始数据曲线上的“弱变化”异常;②文中8种方法的原理简明,异常物理机制清晰或较清晰,各方法的异常识别指标分别为明确、基本明确和定性的,定性的异常在震情研判中仅有参考意义;③消除年变化法和无量纲法通常用于识别地震中期、短临异常,而相对均方差法和差分法通常用于识别短临异常;④经数据处理得到的异常与原始曲线的“弱变化”异常相协调;⑤时间序列数据出现的异常并不等同与地震孕育、发生过程有直接联系的前兆异常,出现数据异常的台站附近不一定会发生显著地震.%This paper introduces the frequently used methods to analyze and process georesistivity observation data for the purpose of earthquake prediction,which include the improvements to some previous methods and the new methods since the ninth Five-Year Plan.In all,there are the eight methods that can be classified into four categories,that is,method of eliminating annual variation,dimensionless method,relative root-mean-square-error (RMSE) method and difference method.Some specific problems about the eight methods are discussed in detail,such as method principle,procedure of data processing,approaches of anomaly analyses,resolution capability for anomaly,indices for identifying anomaly,physical mechanism of anomaly as well as their shortcomings.The results show that:① in general,effective data processing methods ought to be used to analyze and identify the "weak change" anomaly on the raw data curve;② the principle of the eight methods are simple and clear,the physical mechanism on georesistivity anomalies based on the methods are clear or more clear,and the indices for identifying anomaly of each method are definite,relatively definite or qualitative,respectively.The qualitative anomalies are used only for reference in actual earthquake analysis;③ the method of eliminating annual variation and dimensionless method is usually used to identify the anomalies in the medium-term and imminent periods before an earthquake,and the latter two kinds of methods are usually used to identify imminent anomalies;④ the raw observation data from a station is true,so the anomaly got through a data processing method should be concordant with the "weak change" anomaly on a raw curve;⑤ the anomaly arising in the time series of observation data,called as "data anomaly",does not equate with the precursory anomaly directly related to the preparation and occurrence processes of earthquake,and it is not expected that a noticeable earthquake will be certain to occur once any data anomaly appears on the time series of data.
