四川地区中短期前兆识别、指标体系、预报方法及效能评价研究

Studies of Precursor Recognition,Indexes,Methods and Their Efficient Evaluation for Medium Short Term Earthquake Prediction in Sichuan

本子专题在以往几次攻关研究的基础上，以１年时间尺度为目标，对于地震学、地形变、地下流体和地电、地磁等地球物理学科，应用四川地区２０多年或尽可能长时间的观测资料，在监测区作全时空扫描，对有震、无震背景开展系统研究，对前兆指标和判定方法的预报效能进行科学评价，研究各种指标的相互关系，力求形成有区域特色的有效的中短期地震前兆指标体系和预报方法。我们曾经借助Ｒ－ｔ图，结合震中分布图圈定弱震空区，用Ｄ－ｔ图划空段，用带内外频度比确定背景性条带和弱震密集条带，规定过明确的判别指标和预报规则，并用四川１９７０年７月～１９８７年的观测资料，讨论过预报效能，得到了肯定的结果。增添１９８８～１９９７年的观测资料，继续开展研究，得到大体相当或稍有提高的Ｒ值。说明这几种地震活动图象的预报效能比较稳定。关于△Ｄ０（△Ｄ１）、Ｃ（ｎ）、βｎ、ｂ、η和∑（ｔ）值等地震活动性指标，对于四川地区５个地震带，增添资料前后预报评分Ｒ的差值呈以△Ｒ＝－００１为轴的对称分布，５３％在［－００６，＋００２］区间内。也可以说，这些指标有大体稳定的预报效能。从物理意义和统计相关分析两个方面研究了若干地震学指标之间的相关性，主要结论是：地震时间分形Ｄ?

Based on the previous results of researches,for the purposes of the earthquake prediction with the time scale of a year,using all the observation data from seismology,crustal deformation,groundwater,terrestrial electricity,earths magnetism and so on,whose length is about 20 years or longer period in Sichuan ,the efficacy on precursor indexes and recognition methods are valued scientifically,and the relationships among several prediction indexes are studied systematically.Some precursor indexes and prediction methods with regional characteristics for the medium short term prediction of strong earthquake are put forward. The methods for determining weak earthguake gap by using both of R-t plot and epicenter distribution map,and for determining seismic gap along major seismic zones by using D-t plot ,and for determing background seismicity belts and microseismic active belts by using the earthquake number rate between inner and out belt have been studied before.The clear discriminatory indexes and predictive rules are put forward,and are tested with the observation data from 1970 to 1987 in Sichuan.More observation data from 1988 to 1997 are added for testing their prediction effectiveness once more.The results of calculation with new set of data from 1970 to 1997 show that the R value are equal to or greater than that based on the data from 1970 to 1987.So the prediction efficacy of indexes mentioned above is stable relatively. The seismicity indexes from 1970 to 1990 such as △D 0,△D 1,C(n),β n,b,(t) and their prediction grading R are calculated for five seismic zones in Sichuan.The difference ΔR between two sets displays a symmetrical distribution with the axies of symmetry-0 01.The 53% ΔR values are in the range from -0 06 to +0 02.In other words,these prediction indexes are stable roughly. From both physical senses and statistical correlative analysis,the correlative relationships among several seismological indexes are studied carefully and the results show that the three indexes,temporal fractal of earthquakes D 0 , D 1 and b,are independent.The D 0 and D 1 are two kinds of fractal dimensions with different physical senses.When the number of sample is small,the calculating results for D 0 and D 1 are same almost.From the physical nature of calculation,the C(n) is different for D 0 and b value.But when the number of samples in calculation is small and the earthquake magnitude is divided with a big span,the C(n) ,b, D 0 ( D 1 ) can pass the correlative test.As for the indexes M f and b,the relationship between them is not linear correlative.This conclusion can be gotten not only from the deduction analysis but also from the statistical correlative test for five seismic zones in Sichuan.The relationship between index M f and empirical probability P of earthquake ( M≥M t) is obviously correlative. Three kinds of seismic quiet conceptions are concluded.First kind is a quiet seismic precursor before a strong earthquake.Second one is a quiet interval between two strong earthquakes.Last one is a quiet period in the earthquake active procedure .Their mechanism and effects in earthquake prediction are analyzed. By using synthetic probability and pattern recognition methods,the seismic activity levels per year in Sichuan are discussed as well.The R values from 0 7 to 0 8 are obviously greater than the critical grading R 0 for the random earthquake prediction with the convidence level of 95% and nature probability of earthquake. The research thought for earthquake precursors and the complex of precursor recognition are discussed. Some necessary steps for the recognizing a precursor anoraml from a single curve of seismic observation data are put forward.The interference removal methods in several links are also deduced.Several typical interference curves are displayed here which are tested many times previously in practical situation.Four basic criterions for synthetically judging seismic precursors are offered.So based on s