Source patterns of Zn, Pb, Cr and Ni potentially toxic elements (PTEs) through a compositional discrimination analysis: A case study on the Campanian topsoil data

Publication date: 1 December 2018
Source:Geoderma, Volume 331
Author(s): Attila Petrik, Matar Thiombane, Stefano Albanese, Annamaria Lima, Benedetto De Vivo
One of the main objectives of environmental geochemistry is to reveal the source and spatial patterns of different inorganic and organic elements/compounds with special emphasize on potentially toxic elements. In the last couple of decades, environmental geochemists mainly focused on the identification and separation of geochemical anomalies from background/baseline values of potentially toxic heavy metals and the estimation of their ecological and human health risks. However, the main concern with previously published papers on this issue is that the majority of them simply neglected the compositional nature of geochemical data; hence results became spurious and biased and can be interpreted with reservations. Our goal is to identify, interpret and discriminate the source patterns of 4 potentially toxic elements in the Campania Region (Italy) emphasising on their ratios and spatial abundance using multivariate compositional data analysis. This study contributes to understand the compositional behaviour and relative proportion of 4 potentially toxic elements whose precise contamination sources have been unclear before. A workflow of compositional data analysis including a new discrimination index has been elaborated to identify their possible sources of contamination and enrichment. The investigated data set includes 18 elements derived from 3669 topsoil samples, collected at an average sampling density of 1 site per 3.2 km². First, robust biplots and factor analysis were performed to get an overview of elemental associations and reduce the dimensionality of the data set. They revealed that the 4 PTEs belong to different groups. The multivariate regression analysis using alr-transformed (additive logratio) data proved the strong linear relationship between Fe, Mn (independent) and each investigated PTE (dependent), but also unveiled deviation trends in case of Zn and Pb. Based on the multivariate regression result, a sequential binary partition was performed by means of 6 variables (the 4 PTEs, Fe, and Mn) to obtain balances. Balances are ilr-coordinates (isometric-logratio) which can be interpreted as ratios of specific groups of elements. They were used to generate interpolated maps by using multifractal method to see spatial patterns and proportions of elemental associations. A new index has been elaborated based on the bivariate regression of balances and their standardised residuals, which was particularly useful to identify and separate the sources of anthropogenic contamination and geogenic enrichment of respective elemental associations. The large urban and industrial areas (e.g. Naples, Salerno) along the coastline are mainly contaminated by Pb and Zn due to heavy traffic and alloy production. Some Cr and Ni contamination was discerned in the Sarno Basin where the Solofra industrial district is likely to be the principal source through releases from tannery industry. The large volcanic complexes (e.g. Mt. Somma-Vesuvius, Phlegraean Fields, Mt. Roccamonfina) are all characterised by geogenic enrichment of Zn and Pb. In contrast, Cr and Ni-geogenic enrichment is mainly related to the siliciclastic deposits.



https://ift.tt/2tQkyjE