Prediction of the temperature dependence of electrical conductance for river waters

Abstract

The application of mixed electrolyte theories to the prediction of the electrical conductivity of 14 river-water samples covering a range of ionic strength and temperatures between 5 and 25°C has been evaluated using a new program, CONCAL. Both the Fuoss-Onsager, FO, and Lee-Wheaton, LW, theories produce better agreement between the measured and predicted conductivities for the majority of the samples than the simple product equation. For river waters with a charge imbalance of less than 10%, the prediction of the simple product, FO and LW equations are on average within 11.1, 3.4 and 3.7%, respectively, of the experimental values over the range of temperatures considered. The major error in the predictions of the FO and LW models is associated with the reliability of the chemical analytical data input to CONCAL. The accuracy of the predictions of the models is also discussed in terms of the magnitude and sign of the charge imbalance calculated for each of the samples from the analytical data. Five equations that are available to apply a temperature compensation to conductivity data to predict values at 25°C have been tested over a temperature range of 5–25°C. The best performance is achieved using the modified Walden's product. The relationship between the measured conductivity at 25°C, κ(25), and ionic strength, I, has been calculated to be I = 0.0021 + 0.0148 κ(25) with I in mmol dm −3. This observation is discussed by reference to the predicted relationship for simple binary mixtures of NaCl and Ca(HCO 3) 2.