Multifractal Approach to Study the Earthquake Precursory Signatures Using the Ground-Based Observations

Abstract

In a series of papers by Smirnova and Hayakawa, fractal analyses of the ULF emissions in the frequency range of f = 0.001–0.3 Hz have been performed based on the geomagnetic data obtained in seismic active regions including Guam Island and Izu Peninsula. A certain dynamics of the spectrum slopes and the corresponding first-order fractal dimensions have been revealed in relation to the preparation phase of some strong earthquakes (EQs). In order to advance such a fractal approach which seems to be very promising for development of the EQ forecasting methods, we consider multifractal aspects in the analysis of geophysical data dynamics. First, as a case study, the seismicity distribution in the Kobe area of Japan is concerned in relation to the powerful Kobe EQ on 17 January 1995. Applying a multifractal approach, we have revealed that there is a gradual decrease in the certain high-order fractal dimensions which were obtained from the spectrum of singularities (multifractal spectrum) of the local seismicity distribution when the date of the Kobe EQ was approached. Many examples of the multifractal spectrum dynamics in relation to major EQs is introduced as a statistical study. It is then concluded that the multifractal analysis of geophysical data could be a promising tool for extraction of the precursory signatures of the extreme natural events including strong EQs.