GNGTS 2022 - Atti del 40° Convegno Nazionale

GNGTS 2022 Sessione 2.1 207 PROBABILITY, CORRELATION, AND TRANSFER ENTROPY BETWEEN WESTERN PACIFIC EARTHQUAKES AND NOAA ELECTRON BURSTS C. Fidani 1,2 , A. De Santis 1 , L. Perrone 1 1 Istituto Nazionale di Geofisica e Vulcanologia, Roma, Italy 2 Central Italy Electromagnetic Network, Fermo, Italy A correlation between strong earthquakes (EQs) and high energy electron bursts (EBs) has been recently presented in several publications. The peak of correlation appeared when considering shallow EQs with hypocentre depth of less than 200 km, located in a large portion of the Earth’s crust in the Western Pacific, with a minimum magnitude of 6. It evidenced that the 30-100 keV electron detections by the NOAA-15 satellite anticipate by1.5-3.5 h the EQ events. Using this result for EQ forecasting requires unambiguously definition of EBs to verify a precursor phenomenology (Wyss, 1997), which was obtained only considering EBs with L-shell in a restricted interval (Fidani, 2021; see also Figure 1). Here, a contribution is presented which links the analytical transfer entropy, TE (Schreiber, 2000; Bossomaier et al. , 2009) for digital variables to conditional probabilities and correlations. This approach refers to the information theory (e.g. Jaynes, 1957; Cover and Thomas, 2005), and has been recently utilized in the lithosphere to ionosphere system (Fidani, 2022). Thus, mutual information and transfer entropy have been evaluated in the context of precipitating EBs from the inner Van Allen belt, which preceded strong EQs. Fig. 1. L-shell intervals relative only to EBs which were considered to unequivocally discriminate the EBs to use for a possible forecasting experiment as proposed in Fidani, 2020 and 2021. Considering EQ and EB as two sets of digital events, in the case of dependency between events each one contains information on the other. To quantify the amount of information that the EB event has on EQ, the mutual information (Cover & Thomas, 2005) is: I(EQ;EB) = ΣΣ {0,1} P(x,y) log 2 {P(x,y)/[P(x)P(y)]}, (1) where P(x), and P(y) are defined as frequencies of EQ and EB, respectively, while P(x,y) is the joint probability. The logarithm base is 2 as digital events are considered. Making more explicit the mutual information: