Evaluation of Pre-Earthquake Anomalies of Borehole Strain Network by Using Receiver Operating Characteristic Curve

In order to monitor temporal and spatial crustal activities associated with earthquakes, ground- satellite-based monitoring systems have been installed in China since the 1990s. recent years, correlation between strain anomalies local major earthquakes has verified. this study, we further evaluate possibility of containing earthquake precursors by using Receiver Operating Characteristic (ROC) prediction. First, network were extracted borehole data recorded Western during 2010–2017. Then, proposed a new prediction strategy characterized number an anomaly window, Nano, length alarm Talm. We assumed that clusters indicate probability increase impending earthquake, consequently, window would be duration which possible occur. The Area Under ROC Curve (AUC) true predicted rate, tpr, false fpr, is measured efficiency strategies. found optimal short-term forecasts was established setting greater than 7 within 14 days at one day. results show performs significantly better when there are frequent enhanced prior larger surrounding region. detection indicates possibly contain precursory information highlights potential for forecasting.