Magnitude and sign scaling in power-law correlated geoelectrical time series measured in southern Italy

Abstract. Recently, it has been proposed a new method to investigate the linear/nonlinear properties of a time series by decomposing it into two sub-series: magnitude and sign sub-series (Ashkenazy et al., 2003). The long-range correlation properties of the magnitude and sign sub-series relate to nonlinear and linear behavior of the original time series respectively. We apply this method to the hourly time variability of geoelectrical signals, measured from January 2001 to September 2002 by four stations installed in Basilicata region (southern Italy). Three stations (Giuliano, Marsico and Tito) are located in a seismic area, and one (Laterza) in a low-seismicity area. A site-dependence of the linear/nonlinear properties has been shown, revealing that different dynamics govern signals measured in seismic and low-seismicity areas.