It is now recognized that the earthquake hazard systems evolve naturally to the self-organized critical (SOC) state, in which the system is very sensitive to any external perturbations, and when the large avalanches (strong earthquakes) are probable events. Since the principal feature of the SOC state is fractal organization of the output parameters both in space (scale-invariant structure) and in time (flicker noise or 1/ f noise) we can use fractal methods to investigate the evolutionary processes in the earthquake hazard system at different stages of the catastrophic event preparation. Here we apply fractal methods for extraction of earthquake precursory signatures from ULF ( f=0.001–1 Hz) geomagnetic data obtained in seismic active regions before strong earthquake events. We focus our attention on the massive Guam earthquake of August 8, 1993 ( M=8, depth=60 km), and on a swarm of Japanese earthquakes of June–August 2000 occurred near Izu Peninsula ( M>6, depth ≈ 10 km). We analyze scaling (fractal) characteristics of ULF geomagnetic fields registered in the area less than 100 km from the earthquake epicenters, study their dynamics as approaching the earthquake dates, and compare the results obtained in both regions. Three methods have been applied to calculate scaling parameters (spectral exponents and fractal dimensions of the ULF geomagnetic time series): FFT procedure, Burlaga–Klein approach and Higuchi method. It is found that fractal characteristics of the ULF emissions manifest specific precursory behavior with some common and individual peculiarities in both seismic active areas. As the common feature, we have revealed the same increase of the ULF time series fractal dimensions (and the corresponding decrease of spectral exponents) before the both events. As the distinctive peculiarity, we have found different character of such dynamics: gradual increase (decrease) in the case of the Guam earthquakes and relatively sharp alteration in the case of the Izu earthquake swarm. For the case of Japanese earthquakes, it turned out to be possible to reveal the most effective frequency range (around f=0.01 Hz), in which precursory behavior of fractal characteristics is more pronounced and manifested earlier than in the other frequency ranges. We give a physical interpretation of the peculiarities revealed taking into account possible specifics of the SOC processes at different depths of the earthquake focuses.
Read full abstract