秒采样精度下磁暴急始自动拾取方法研究

磁暴急始的自动识别拾取与磁暴报告自动产出是国家地磁台网数据产品产出中重要的内容。为实现在1s采样率下的磁暴急始高精度拾取,本文针对性地提出基于Walsh变换和Akaike信息准则(AIC)的起跳点检测算法——Walsh-AIC算法,将其应用于98个磁暴急始事件的拾取上,并采用由国际地磁指数服务(ISGI)发布的磁暴时刻为标准对应用效果进行评估。结果表明,该算法不仅能够将急始磁暴起跳时刻的位置更加明显地刻画出来,而且能够有效避免传统识别急始磁暴的AIC算法中存在的识别起跳时刻结果晚于实际起跳时刻的情况。同时,将Walsh-AIC算法与其他主流拾取方法进行对比,发现前者定位结果的平均偏差和标准差均明显小于其他算法。此外,Walsh-AIC算法结果的误差关于0点的对称性较好,表明该算法拾取磁暴急始起跳时刻的精度较高,适合于解决磁暴急始的自动拾取问题。

Research on Methods of Automatically Picking Sudden Commencement on 1-second Resolution Data

The Automatically picking sudden commencement (SC) and the generation of report is a pressing problem in data production in the Geomagnetic Network of China (GNC). A method based on Walsh transform and Akaike Information Criteria (AIC) is proposed in this paper which focuses on solving the problem of high-precision SC picking in 1-second resolution data,so-called Walsh-AIC method. A comparative study with other popular methods was conducted by using 98 horizontal components of magnetic data. The SC menu provided by International Service of Geomagnetic Indices (ISGI) is adopted as a standard. The picking precision of Walsh-AIC method is better than that of others through comparison including standard deviation and average deviation. Our method is suitable for automatic picking of SC.