Methods for predicting dispersion coefficients in rivers

The majority of approaches for predicting pollutant transport in rivers rely on being able to estimate dispersion coefficients. This paper describes four methods for doing this and argues that in most cases practising engineers would rely on one of numerous predictive equations that have been proposed over the past 30 years or so. The paper reviews these equations and discusses some recent improvements to the theoretical representation of pertinent transport processes that implies that the latest predictors should be more reliable than older ones. Use of seven predictors for a simple test case shows considerable scatter in predictive values, indicating that predicting dispersion coefficients remains a difficult task. Interestingly, theoretically based predictors seem to give reasonably consistent, but possibly low, values while predictors that are heavily reliant on regression to observed values seem to overpredict. The paper will be of interest to practising engineers who use computer models to predict the fate of pollutants that are discharged to river systems.