Humic substances in acid surface waters; modelling aluminium binding, contribution to ionic charge-balance, and control of pH

Abstract

A discrete-site model of ion-binding by humic substances (HS), incorporating a description of electrostatic effects, is evaluated with analytical data for surface waters of acid pH (3.5–6.5). After optimization of the model by adjustment of the binding-site content of the HS, the root-mean-square deviation (RMSD) between measured and calculated concentrations of organically-complexed monomeric aluminium—[Alm-org]—is 1 μM for a range of measured values of 0.1–9.0 μM (108 samples from 12 different locations). The optimization indicates that the dissolved organic matter of natural waters is only about 50% as ‘active’ (in the sense of ion-binding) as isolated HS. The model, optimized for Al-binding, also accounts for the contribution of HS to ionic balance; for 139 samples (from 8 locations) with dissolved organic carbon concentrations in the range 4.6–43.0 mg 1−1, and using measured pH as input for the computations, the mean calculated ratio of cations to anions was 1.03, with a standard deviation of 0.11. A similar result was obtained with an optimized version of the model of B. G. Oliver, E. M. Thurman and R. L. Malcolm (Geochim. cosmochim. Acta47, 2031–2035, 1983). For the same 139 samples, pH values were also calculated, using total measured anion concentration as inputs. The RMSD in pH was 0.35 for all samples, but only 0.18 for the 56 samples of pH ⩽ 4.5. Statistical analyses indicate that inadequacies in model assumptions, including the estimation of concentrations of HS in water samples, account for about one-third of the discrepancy between measured and calculated [Alm-org]; the remaining two-thirds is explained by errors in input data and measured [Alm-org]. In the case of pH prediction, no model inadequacy is apparent, because of the high sensitivity of the calculations to errors in input data.