The applicability of two computer techniques to separate the remanent magnetization (RM) components acquired in laboratory fields

Abstract

Thermal remanent magnetization (TRM) and anhysteretic remanent magnetization (ARM) components were imposed on natural rock samples. The artificial laboratory components had different directions and the blocking temperature and/or coercivity spectra were overlapping. Two methods, principal component analysis (PCA) by Kirschvink and analytical modelling of demagnetization data (by Stupavsky and Symons, S&S) were used to resolve these components. The PCA technique calculated lines fitted to the demagnetization path with ASD = 10° (angular standard deviation), and the S&S method used four types of intensity decay curves for calculated components. Both methods (PCA and S&S) resolved perfectly the one-component case. The two- or three-component case results strongly depended on spectra overlapping, and on the angles between component directions and magnetic minerals in samples. Principal component analysis gave more reliable results for separated spectra of TRM and thermally cleaned samples, whereas the S&S technique was more efficient for the case of strong spectra overlapping of ARM components and the alternative current field (AF) demagnetization method. Remarkable anisotropy of RM was observed which influences the results for the haematite-bearing samples.