ASGP (2024), vol. 94: 273–286

A MULTIVARIATE APPROACH TO PEBBLE MORPHOMETRY: A CASE STUDY OF THE WITÓW SECTION (SOUTHERN POLAND)

Szymon MOL , Krzysztof NINARD (*), Agata KUŹMA , Anastasiia DERYK & Piotr ŁAPCIK

Institute of Geological Sciences, Jagiellonian University, Gronostajowa 3a, 30-387 Kraków, Poland; e-mails: szymon.mol@student.uj.edu.pl, krzysztof.ninard@uj.edu.pl, aga.kuzma@student.uj.edu.pl, anastasiia3977.deryk@student.uj.edu.pl, piotr.lapcik@uj.edu.pl
*) Corresponding author

Mol, S., Ninard, K., Kuźma, A., Deryk, A. & Łapcik, P., 2024. A multivariate approach to pebble morphometry: a case study of Witów section (southern Poland). Annales Societatis Geologorum Poloniae, 94: 273–286.

Abstract: Gravel-dominated Neogene – Early Pleistocene braided river deposits of the Witów Series occur in the Carpathian Foredeep, about 20 km north of the front of the Polish Outer Carpathians, between the Szreniawa and Vistula rivers. For the first time, these deposits were subjected to numerical analyses, based on the morphometry and mass of almost 1,500 pebbles sampled at 10 cm intervals along the 4.4-m-high section in the Witów Quarry. In contrast to the traditional approach to pebble morphometry, multivariate statistics was utilised. This enables the examination of various aspects of the dataset holistically and simultaneously. A multivariate method, called principal component analysis (PCA), is widely used in the life sciences, but the employment of PCA for pebble morphometry has not yet been described. Here, PCA was applied to reveal the interrelations between pebble size, mass, lithological composition and stratigraphic height. Most notably, some differences in the distribution of morphometric features between different pebble lithotypes are displayed. Even though the morphometric features and petrological composition of pebbles remain similar in the section as a whole, overall upward-decreasing trends of stream-bed velocity proxies were recognized with the aid of PCA results and were validated, using standard bivariate correlation methods. This approach to the multivariate analysis of large quantitative and qualitative datasets should be considered as a possible part of the integrated sedimentological research of coarsegrained deposits. The consistency between results among the multiple indicators studied reduces the uncertainty of the sedimentological interpretations, presented in this work.

Manuscript received 23 February 2024, accepted 4 July 2024