摘要
对剪切-粘贴方法(cut and paste,CAP)和地震软件包(computer programs in seismology,CPS)两个较常用的震源机制格点搜索反演方法的加权策略进行分析,发现二者权重分别仅单独考虑了波形数据信噪比或振幅差异,且权重数值大小随震中距的变化趋势相互冲突。为了解决此数值矛盾并保留二者权重的优势,将两种加权策略进行联合,综合考虑波形信噪比和振幅差异对反演的影响。此外,考虑到震中距定权的粗糙性,提出了针对各波形本身的数据信息精化计算权重方案。为了检验联合定权的优越性,以2013-04-20的芦山地震为例,利用CPS程序分别用联合及单独定权进行多次反演,比较发现联合加权方法反演结果最优。最终的震源机制解为(走向211°±3°,倾角41°±1°,滑动角94°±2°),震源深度17km,与其他研究者的研究成果有很好的一致性,且与震源区的应力及地质构造情况均相互吻合,表明所提出的权重方法优化效果明显。
We discuss weighting in two grid searching methods, cut and paste (CAP) and computer programs in deismology (CPS), to determine source mechanisms from waveform inversion. We found their weightings differs: CAP weighting focuses on the influence from amplitute only; while CPS stresses the difference of SNR for all waveforms. The weighting from the two methods are incompati- ble, as CAP weighting increases when epicentral distance increases, while CPS weighting decreases. In order to take advantage of these two weighting methods, we combined them and improved the pre- cision of weighting by estimating the amplitute or relative error from the waveform itself. By applying grid search inversion on the 2013 LuShan earthquake data, we show the advantages of our weighing method in a comparison of three experiments based on CPS with different weightings: SNR weighting (CPS), amplitute weighting(CAP) and our improved weighting(CPS&CAP). The preferred source mechanism of LuShan earthquake from our study were strike 211±3°, dip 41±1°,and rake 94±2° and source depth was 17km, and consistent with the mechanisms other researchers have found for stress distribution and geological structure. The case affirms the advantages of our weighting.
出处
《武汉大学学报(信息科学版)》
EI
CSCD
北大核心
2017年第5期616-623,共8页
Geomatics and Information Science of Wuhan University
基金
国家自然科学基金(41374053)~~
关键词
震源机制
波形反演
芦山地震
反演权重
source mechanism
waveform inversion
Lushan Earthquake
weighting