SEARCH FOR STRANGE ATTRACTORS IN MAGNETIC-FIELD DATA - AN EXAMPLE FROM A TERTIARY VOLCANIC AREA NEAR BONN, GERMANY

Citation
S. Leonardi et Hj. Kumpel, SEARCH FOR STRANGE ATTRACTORS IN MAGNETIC-FIELD DATA - AN EXAMPLE FROM A TERTIARY VOLCANIC AREA NEAR BONN, GERMANY, Journal of geodynamics, 21(4), 1996, pp. 329-338
Citations number
26
Categorie Soggetti
Geochemitry & Geophysics
Journal title
ISSN journal
02643707
Volume
21
Issue
4
Year of publication
1996
Pages
329 - 338
Database
ISI
SICI code
0264-3707(1996)21:4<329:SFSAIM>2.0.ZU;2-F
Abstract
Spatial variations of the magnetic field, measured at the Earth's surf ace, are mainly due to geological inhomogeneities within the crust. Th e present state of the crust, in turn, reflects its past dynamical pro cesses and evolution. We analyzed the magnetic field intensity along a few but densely sampled profiles in a Tertiary volcanic province near Bonn, Germany. The aim was to find out whether the variation of the m agnetic intensity can be related to a deterministic, although chaotic, generating system or to a stochastic one. In the former case, the geo dynamical history would depend on a few degrees of freedom, whereas in the latter case, the parameters to take into account were so many tha t a statistic approach appears to be more suitable. Since strange attr actors do not emerge from the analyzed data, we exclude the presence o f a chaos generating system. Autocorrelation functions and a rescaled- range analysis, on the other hand, reveal and quantify a certain degre e of correlation among successive data points and allow calculation of a range of spectral exponents in log-log diagrams, which are clearly distinct from those of white noise and Brownian motion. Only one of th e profiles exhibits similarity to uncorrelated or white noise. In this case, the signal variations do not result from the local geology but from a buried gas pipe-line, aligned subparallel to that profile. If s elf-organized crustal evolution generally produces signatures that can be described as correlated noise, within a limited bandwidth of spect ral exponents, erraneous or disturbed data could eventually be discrim inated from records containing purely natural data by applying analysi s tools of the dynamic system theory together with autocorrelation tes ts. Copyright (C) 1996 Elsevier Science Ltd