Exploring topsoil geochemistry from the CoDA (Compositional Data Analysis) perspective: The multi-element data archive of the Campania Region (Southern Italy)

Abstract

Soil geochemistry is often investigated by considering a large number of variables, including major, minor and trace elements. Some of the variables are usually highly correlated due to coherent geochemical behaviour, but the effect of anthropic factors tends to increase data variability, sometimes obscuring natural relationships governing their distributions. In this framework it may be difficult to identify geochemical features linked to natural phenomena as well as to separate geogenic anomaly from the anthropogenic ones. Consequently the identification of background/baseline values may be seriously compromised. However, knowledge about these reference terms is fundamental to manage and protect natural resources on different scales. Moreover, adequate estimations of background/baseline values are possible only if a sufficient number of chemical analyses are stored in complex repositories. In this contribution the multi-element data archive of the Campania Region (Southern Italy) was explored from the CoDA (Compositional Data Analysis) multivariate perspective to characterise its structure. The archive contains abundance data of Al, As, B, Ba, Ca, Co, Cr, Cu, Fe, K, La, Mg, Mn, Mo, Na, Ni, P, Pb, Sr, Th, Ti, V and Zn (mg/kg) determined in 3535 new topsoils as well as information on coordinates, geology and land cover. Under CoDA the proportionality features of abundance data are fully taken into account enhancing their relative multivariate behaviour in the correct sample space.

Results indicate that the structure of the whole matrix appears to be constituted by a core that geographically is mainly given by topsoils developed on volcanic materials and several outlier compositions whose origin is different. Anomalous compositions can originate from the robust barycentre all around when the following conditions are present: 1) high Na K volcanic products, 2) limestones and dolostones with their terrigenous component, 3) flysch deposits or 4) fertiliser contribution.

The (1 x D) robust barycentre of the whole dataset together with the variation array of the core represents the most frequent (1 x D) multi-element vector as well as the proportionality relationships among its components. It might be considered a compositional baseline.