Diffuse reflectance spectroscopy (DRS) parameters, such as redness index and hematite/goethite
ratios are popular in the study of loess-paleosol sequences. The suggested parameters,
however, have their limitations: only some parts of the whole reflectance spectrum
have been used, and the spectra of some important components may have been excluded
from the analysis.
A mathematical statistical method, hierarchical cluster analysis, was applied to the
loess-paleosol succession at Malá nad Hronom (Slovak Republic) to characterize the
entire reflectance spectrum and classify sediment groups. The sedimentary units were
classified into seven different cluster groups such as sediments, weak and well-developed
paleosols. The distribution of the spectral features in the reflectance spectra was
located using the key-wavelength, given by the Wilk’s lambda analysis. The key-minerals,
which play an important role in the classification, were determined by characteristic
extrema in the second derivative of reflectance spectra of the seven reference minerals
(hematite, goethite, muscovite, illite, montmorillonite, kaolinite and calcite).
Mineral groups influence together certain curve sections within the measured range,
therefore curve sections modified by individual minerals were to differentiate only
in a few cases. The mentioned mineral groups refer to characteristics of the formation
environment and to climate change during the Pleistocene.
Acknowledgement: The research was supported by NKFI K119366, NKFI K120620 and OTKA
100180. The paper was also supported by Balázs Bradák’s fellowship at the Research
Center for Inland Seas, Kobe University, Japan was supported by the Japan Society
for the Promotion of Science (JSPS). Agnes Novothny was funded by a post doctoral
project of the Hungarian Scientific Research Fund (OTKA PD-100315).