Seismo-ionospheric coupling correlation analysis of earthquakes in Greece, using empirical mode decomposition

Abstract. Ionospheric variability as a result of earthquake events is a confirmed phenomenon as published in various seismo-ionospheric coupling studies. Generally, ionospheric variations resulting from earthquake activity are much weaker than disturbances generated by different sources, e.g. geomagnetic storms. However, geomagnetic storm disturbances exhibit more global behaviour, whereas seismo-ionospheric variations occur only locally in an area that is specified by the magnitude of the earthquake. Cross-correlation coefficient analysis is a technique proposed some years ago, and ensures cancelation of geomagnetic storm variations of the ionospheric plasma, provided that the measurements are taken from stations with similar behaviour in these phenomena. In this paper we will use the aforementioned technique for analyzing data from ionospheric stations in Rome and Athens, and apply it to a series of earthquakes in Greece. Considering the local behaviour of the seismo-ionospheric variations, we expect that the Athens station, which happens to be inside the area affected by the earthquake, will accurately capture the disturbances. Due to its distance from the activity, we also do not expect the Rome station measurements to be affected by the seismic events in Greece. In addition, due to the fact that ionospheric plasma parameters exhibit non-stationary and nonlinear behaviour, we propose a novel signal processing technique known as the Hilbert-Huang transform in order to denoise the data before we calculate the cross-correlation coefficient of the two signals. Results from our analysis are in accordance with previously-conducted studies covering the same topic, clearly demonstrating that there are ionospheric precursors 1 to 7 days prior to strong seismic events as well as 1 to 2 days following such events.