Modelling of pollutant mixing in the St. Lawrence River

Many of Canada's large rivers, such as the connecting channels of the Great Lakes, receive contaminant loads from industrial, municipal, and tributary sources. These contaminants experience two general types of mixing, outlet-dominated mixing processes (near field) and river-dominated mixing processes (far field). This note is concerned with the numerical modelling of the far field processes by a fully elliptic form of the two-dimensional depth-averaged river model proposed by Rodi et al. Many of the popular hydrodynamics codes experience numerical diffusion which far exceeds the real turbulent diffusion; a new method known as the semi-implicit skewed upwind method is introduced to minimize the numerical dispersion for rivers with highly nonlinear alignment. The new model is verified by comparison with the field data obtained from a dye study in the St. Lawrence River. Key words: pollutant transport, river mixing, far field model, numerical diffusion.