Modelling inorganic aluminium with WHAM in environmental monitoring

Due to the varying toxicity of different Al species, information about Al concentration and speciation is important when assessing water quality. Modelling Al speciation can support operational monitoring programmes where Al speciation is not measured directly. Modelling also makes it possible to retroactively speciate older samples where laboratory fractionation was not undertaken. Organic-rich waters are a particular challenge for both laboratory analysis and models. This paper presents the modelling of Al speciation in Swedish surface waters using the Windermere Humic Acid Model (WHAM). The model was calibrated with data from operational monitoring, the Swedish national survey of lakes and rivers, and covers a broad spectrum of physical and chemical conditions. Calibration was undertaken by varying the amount of DOC active in binding Al. A sensitivity analysis identified the minimum parameters required as model input variables primarily to be total Al, organic C, pH, F⁻, and secondly Fe, Ca and Mg. The observed and modelled Ali had no significant differences (Spearman rank, p < 0.01), however, lake samples modelled better than rivers. Samples were placed in the correct toxicological category in 89–95% of the cases. The importance of the size of the calibration data set was assessed, and reducing the calibration data set resulted in poorer correlations, but had little impact on the toxicological placement. Overall, the modelling gave satisfactory results from samples covering a broad spectrum of physical and chemical conditions. This indicates the potential value of WHAM as a tool in operational monitoring of surface waters.