摘要： •The z test is employed to identify seismo-ionospheric anomalies in the GPS TEC.•The ROC curves are used to examine seismo-ionospheric precursors in China.•Youden index finds the optimal cutoff value for determining anomaly bounds. Evidence of the seismo-ionospheric precursor (SIP) is reported by statistically investigating the relationship between the total electron content (TEC) in global ionosphere map (GIM) and 56M⩾6.0 earthquakes during 1998–2013 in China. A median-based method together with the z test is employed to examine the TEC variations 30days before and after the earthquake. It is found that the TEC significantly decreases 0600–1000 LT 1–6days before the earthquake, and anomalously increases in 3 time periods of 1300–1700 LT 12–15days; 0000–0500 LT 15–17days; and 0500–0900 LT 22–28days before the earthquake. The receiver operating characteristic (ROC) curve is then used to evaluate the efficiency of TEC for predicting M⩾6.0 earthquakes in China during a specified time period. Statistical results suggest that the SIP is the significant TEC reduction in the morning period of 0600–1000 LT. The SIP is further confirmed since the area under the ROC curve is positively associated with the earthquake magnitude. 出版者： Elsevier Ltd 出版日期： 2015-12-15 出處： Journal of Asian Earth Sciences, 2015-12, Vol.114, p.393-402 資源來源： Elsevier ScienceDirect Journals 版權： 2015 Elsevier Ltd 識別號： ISSN: 1367-9120 識別號： EISSN: 1878-5786 識別號： DOI: 10.1016/j.jseaes.2015.05.028
