Soft computing methodologies for spectral analysis in cyclostratigraphy

• Published in: Computers & geosciences Vol. 27; no. 5; pp. 535 - 548
• Main Authors: Tagliaferri, R, Pelosi, N, Ciaramella, A, Longo, G, Milano, M, Barone, F
• Format: Journal Article
• Language: English
• Published: Oxford Elsevier Ltd 01-06-2001; Elsevier Science
• Subjects: Cyclostratigraphy; Earth sciences; Earth, ocean, space; Exact sciences and technology; Genetic algorithm; Neural networks; Nonlinear principal component analysis; Stratigraphy; Time-series analysis; Southern Europe; Italy; Apennines; Southern Apennines; Cyclostratigraphy; Time-series analysis; Nonlinear principal component analysis; Neural networks; Genetic algorithm; time series analysis; Mesozoic; algorithms; Europe; neural networks; eccentricity; spectral analysis; Cretaceous; carbonates; obliquity of the ecliptic; paleoclimate; principal components analysis; cyclic processes; periodicity; Milankovitch theory; Earth orbit; secular variations; Phanerozoic; shallow-water environment; precession
• ISSN: 0098-3004; 1873-7803
• DOI: 10.1016/S0098-3004(00)00166-7

In this paper we present some soft computing methodologies for time-series analysis applied to cyclostratigraphy. An application to some stratigraphic signals to detect Earth orbital (Milankovic’) periodicities which are expected to be recorded in Cretaceous shallow water carbonate sequences outcropping in Southern Apennines (Italy), is described. The results obtained with classical spectral analysis techniques, based on the modified periodogram, are compared to the results of our methods based on neural nets and genetic algorithms. The aim of these cross comparisons is to find the most reliable, fast and accurate methodology to identify orbital periodicities in noisy and segmented stratigraphic signals.