The determination of background pollution levels is fundamental for the interpretation of the results obtained from environmental biomonitoring. In this paper we propose a new probabilistic method, based on a Gaussian mixture model, for determining the distribution of regional background levels of different pollutants. The distribution of the reference level is used to categorize the observations as "background" or "above-background" and spatial statistical techniques are then applied to determine the probability of the background level being exceeded. To exemplify its use, we applied the method to concentrations of five potentially toxic elements (Cd, Cu, Hg, Pb and Zn) measured in the moss Pseudoscleropodium purum. The proposed method was applied to data resulting from sampling at ca. 150 sampling stations in a regular grid (15 × 15 km) in Galicia (NW Spain). Sampling was carried out in June in 2000 and 2002, and in March and September in 2004, 2006, 2008 and 2014. The proposed method yielded consistent results for all of the different sampling surveys, and the pollution levels were found to be closely related to the sources of pollution identified in the study region. In short, although not an optimal solution, the proposed method seems to be suitable and realistic for the qualitative assessment of regional pollution.
Keywords: Atmospheric pollution; Biomonitoring; Bryophytes; Heavy metals; Mercury; Normal mixture model.
Copyright © 2022 The Authors. Published by Elsevier B.V. All rights reserved.