Abstract: On 17 June 2019, an earthquake with M_S6.0 occurred near Changning, Yibin, Sichuan Province. In order to detect the anomalous changes of Total Electron Content (TEC) seismic ionospheric, Autoregressive Integrated Moving Average Model (ARIMA) and prophet are used to describe the TEC time series and detect TEC anomalies variations before and after the Changning earthquake. In this study, geomagnetic indices (i.e. K_p, Dst, A_p and F10.7) are used to distinguish pre-earthquake anomalies from the other anomalies related to the geomagnetic and solar activities. The result of ARIMA shows that apart from the significant ionospheric anomalies in TEC time series detected at 5:45 UTC, 13 days before the earthquake, all the other ionospheric anomalies occurred 2~6 days after the event. While using Prophet method, multiple pre-earthquake anomalies are detected at 7:45 UTC before 16 days, 9:30 UTC before 13 days, 8:30 UTC before 5 days, and 09:30 UTC before 3 days. A comprehensive analysis indicates that the Prophet method is more likely to manifest anomalous states than traditional methods such as ARIMA. Meanwhile, due to its promising ability of nonlinearly learning, Prophet method is worthy of attention as a robust predictive tool for seismic ionospheric anomaly detection.

Key words: Prophet, ARIMA, Ionospheric anomalies, TEC, Earthquake