GNGTS 2023 - Atti del 41° Convegno Nazionale

Session 2.1 GNGTS 2023 Earthquake data were downloaded from ISIDe Working Group database (Italian Seismological Instrumental and parametric database; http://iside.rm.ingv.it) , to obtain a subset of ‘‘significant events’’ with magnitudes (Mw) greater or equal to 1.0. The location of June 2010 to November 2013 earthquake epicenters is shown in Figure 1. The seismic dataset was declustered using the Matlab ® zmap routine, to reduce the total number of events to a manageable quantity of 241. Earthquake locations were then compared with the Gallicano Station location with respect to the Dobrovolsky radius D (Dobrovolsky, 1979), to define a useful distance where a physical interaction can connect the earthquake occurrence and the presence of CO 2 . Earthquakes were considered only when their hypocentre distances were  3 D. The earthquake database was digitized considering the presence of a seismic event as “1” and the absence as “0”, for each day. The resulting time series was characterized by 1259 days, where 39 “1” indicated the main shock events (EQ). The statistical elaboration of the CO 2 irregular component time series from Gallicano, retrieved during the same earthquake time interval, was realized by a Pearson type VII distribution (Pearson, 1916), which is finely peaked and symmetric. The goodness of the approximation was evaluated by the Chi-square test greater than 99%. So, a threshold over which the irregular component fluctuation had a 99% of probability being not accidental was calculated to define anomalies through the cumulative probability. It was retrieved by the Hypergeometric function (Johnson et al., 1995), with x + = 0.07444 and x - = -0.07374 around the average value x = 3.5 10 -4 . 23 CO 2 anomalies, or CO 2 events (ECO2 ), were selected from the irregular component and indicated by “1” in the series of daily anomalies. Fig. 2. Cross-correlation valuesR between CO 2 anomalies and earthquakes in the range Δ t from -20 to +20 days, three pronounced peaks occur at -11, -1, and 0 days, which means that CO 2 anomalies prevalently anticipate seismic activity. The correlation coefficient R was obtained by filling a histogram with the coincidence events of one earthquake with one CO 2 anomaly, ∑ {EQ;ECO2} (EQ × ECO 2 ) as in Fidani (2021), inside every time difference Δ t of one day, of the earthquake time and the CO 2 anomaly time. The digital time series can be correlated in the same way as the normal series and the Pearson correlation, looking for the so called “Matthews correlation” (Matthews, 1975). Figure 2 is the correlation plot between regional seismic events and Gallicano CO2 anomalies. The histogram spans over time differences