Environmental Monitoring and Peat Assessment Using Multivariate Analysis of Regional-Scale Geochemical Data

Abstract

A compositional multivariate approach was used to analyse regional-scale soil geochemical data obtained as part of the Tellus Project generated by the Geological Survey of Northern Ireland. The multi-element total concentration data presented comprise X-ray fluorescence (XRF) analyses of 6862 rural soil samples collected at 20-cm depth on a non-aligned grid at one site per 2 km 2 . Censored data were imputed using published detection limits. Each soil sample site was assigned to the regional geology map, resulting in spatial data for one categorical variable and 35 continuous variables comprised of individual and amalgamated elements. This paper examines the extent to which soil geochemistry reflects the underlying geology or superficial deposits. Since the soil geochemistry is compositional, log-ratios were computed to adequately evaluate the data using multivariate statistical methods. Principal component analysis (PCA) and minimum/maximum autocorrelation factors (MAF) were used to carry out linear discriminant analysis (LDA) as a means to discover and validate processes related to the geologic assemblages coded as age bracket. Peat cover was introduced as an additional category to measure the ability to predict and monitor fragile ecosystems. Overall prediction accuracies for the age bracket categories were 68.4 % using PCA and 74.7 % using MAF. With inclusion of peat, the accuracy for LDA classification decreased to 65.0 and 69.9 %, respectively. The increase in misclassification due to the presence of peat may reflect degradation of peat-covered areas since the creation of superficial deposit classification.

Aitchison's Compositional Data Analysis 40 Years On: A Reappraisal

“…This case study, using a data set with relatively few components, is additional evidence that a simple approach to compositional data analysis is possible. Thus, other geochemical studies that have used the CLR transformation [62,64] would most probably have found almost identical results using a well-chosen ALR transformation -in particular, the study of the Tellus data in [64]. It is not claimed that this will always be the case, but sufficient successful applications have been found [48] to warrant considering the original approach of John Aitchison, using additive logratios, as well as alternative approaches that allow zero values, such as the power-transformed version of correspondence analysis.…”

“…The cluster of points on the left represents peat, that is, soils that are highly enriched in organic material, chalcophile elements (S, Se, Cd, Pb), and lithophile elements (Cl, Br, I). The relative enrichment of I also reflects coastal effects associated with proximity to sea water [62]. To demonstrate the near isometry of the ALR transformation with Al as the reference, Fig.…”

Occurrence, Geochemistry and Speciation of Elevated Arsenic Concentrations in a Fractured Bedrock Aquifer System

“…The study area is predominately made up of rocks of the Palaeozoic Southern Uplands-Down-Longford terrane (SUDLT), the Paleogene Slieve Gullion complex with younger intrusives and volcanic rocks to the north, west and east of the study area and Carboniferous sediments found to the south. The SUDLT, which extends across Scotland and Northern Ireland is dominated by Lower Palaeozoic marine sedimentary rocks (lithic arsenites and sandstones) which have undergone low-grade metamorphism (Steed and Morris 1986;Anderson 2004;Lusty et al 2012;McKinley et al 2017). The SUDLT bedrock is dominated by well-bedded Ordovician and Silurian turbidite sequences consisting of greywacke sandstone, siltstone and mudstone (Anderson 1 2004; Lusty et al 2012).…”

Compositional data analysis of regional geochemical data in the Lhasa area of Tibet, China

“…Subsequently, univariate statistical analysis (Filzmoser et al, 2009a), bivariate statistical analysis (Filzmoser et al, 2010;Reimann et al, 2017) and classical multivariate statistical analysis of compositional data (e.g. classification techniques, analysis of variance, linear discriminant analysis and spatial analysis, regression analysis, principal component analysis, factor analysis, clr-biplot and clustering of variables) have been proposed (Carranza, 2011(Carranza, , 2016Grunsky, 2010;Grunsky et al, 2014;McKinley et al, 2016aMcKinley et al, , 2016bMcKinley et al, , 2017Parent et al, 2014;Reimann et al, 2012;Thiombane et al, 2018;Tolosana-Delgado and McKinley, 2016;Wang et al, 2014). Moreover, the application of robust statistics (Carranza, 2016;Filzmoser et al, 2009b;Filzmoser et al, 2012b;Zuo, 2014), geostatistics (Grunsky et al, 2014;Pawlowsky-Glahn and Egozcue, 2016;Tolosana-Delgado and van den Boogaart, 2013;Tolosana-Delgado and van den Boogaart, 2014) and the use of fractal methods (Carranza, 2011) have all been applied in geochemical compositional data analysis .…”