A rational approach to the assessment of aluminium solubility controls in freshwaters

The problem of assessing aluminium equilibrium solubility controls in freshwaters using thermodynamic techniques is placed in a statistical hypothesis testing framework. As an aid in interpreting field information, artificial data are generated which correspond to ideal equilibrium conditions, but which incorporate realistic analytical and temperature related uncertainties. These artificial data are then subjected to thermodynamic interpretations. The technique is useful in answering "what if" questions by showing the scatter to be expected in real data under true equilibrium conditions. Aluminium hydroxide is used as the example in this report. Reasons for discarding the conventional dependent axis plot (pAl3+ vs pH) in favour of a plot with independent or nearly independent axes are emphasized. An improved method is suggested allowing for temperature variations in the equilibrium constants when constructing stability diagrams. Also, the need for caution when regressing variables which both contain errors is noted.