Moss interspecies comparisons in trace element concentrations

Abstract

Within the framework of a European-scaled moss survey, various moss species were sampled throughout The Netherlands [NL], Germany [D], and Switserland [CH], and used in moss interspecies comparisons of elemental concentrations. Moss species considered were Pleurozium schreberi [NL,D,CH], Brachythecium rutabulum [NL], Hypnum cupressiforme [D,CH], Hylocomium splendens [D,CH], and Scleropodium purum [D]. Element analysis was carried out directly (The Netherlands, Instrumental Neutron Activation Analysis: As, Br, Ce, Cr, Cs, Fe, La, Mn, Na, Rb, Sc, Se, Sm, Th, Ti, V and Zn), or after sample digestion (The Netherlands, ICP-MS: Pb, Germany, AAS/ICP-AES: Cd, Cr, Cu, Fe, Ni, Pb, Ti, V, and Zn; Switzerland, ICP-AES/ICP-MS: Co, V, and Zn).Local variations (=within sampling sites) in element concentrations were estimated, based upon in-site multiple sampling and analysis of Pleurozium schreberi species in The Netherlands. Element concentrations in moss species were compared in linear correlations, both in unweighted and weighted fits, with weighing factors based on the local variation data. Weighted fits were shown to generally improve the calibration characteristics, as tracked by X (2) calculations.The calibration data suggest the presence of previously unnoticed outliers in element concentrations. The absence of further information, however, may prescribe the use of all data in comparison procedures. These results indicate that interlaboratory analysis of replicate samples and the use of dedicated certified reference materials may help solving problems in the analysis of the sample series.For several of the considered interspecies comparisons, weighted calibrations could be based on significant correlations (P=0.05). Actual use, however, will remain arbitrarily decided upon, and may be based on decisions as to what to accept with respect to the levels of uncertainty in the calibration parameters. Furthermore, the use of calibrations in extrapolation modes is greatly restricted by the necessary reservations in geographically larger-scaled applications.