Skip to main content


Laboratory experiments undertaken within a compound channel allowed investigation of thetransverse mixing of a solute tracer during both in-bank and over-bank flows. Attention was focused on the effect of a cross-stream (transverse) depth variation from main channel to flood plain on the transverse variation of the local transverse mixing coefficient. The temporal variation of the longitudinal and transverse velocities and their spatial variation within the cross-section have been measured using two dimensional Laser Doppler Anemometry. The variation of fluorescent solute concentrations in the longitudinal and transverse directions has been quantified in both the near- and mid-field mixing zones. The generalised change of moment method is employed to estimate transverse variations of the transverse mixing coefficient. This shows that assuming a transverse variation of the transverse mixing coefficient of the form of the spatial distribution of eddy viscosities determined from turbulence measurements, reproduces the observed mixing, but that increased values are required to predict the mixing on the floodplain close to the main channel.


K.J. Spence - Sheffield Hallam University, Sheffield S1 1WB, UK

I. Guymer - Professor of Civil Engineering, School of Engineering, University of Warwick, Coventry CV4 7AL, UK.


KEY WORDS: Flood hydraulics, over-bank flow, solute, fluorescent tracer, transverse mixing, eddy viscosity, generalised change of moment



The basis for this paper was a series of laboratory experiments with the aim to investigate the transverse mixing of a solute tracer during in-bank and over-bank flows. Following on from previous work, models considering turbulence and secondary flows were used to form a two dimensional model concerning transverse mixing. A transverse mixing coefficient was obtained where depth, longitudinal velocity and local transverse mixing coefficients were constant. The non-dimensional transverse mixing coefficient measured in the main channel regions during in-bank and over-bank flows on the floodplain away from the influence of the main channel were similar to reports from other papers. Overall, a simple distribution of local transverse mixing coefficients has been proposed for future models as an approximation and has been shown to work well.

This paper will be of use to engineers who are investigating mixing within a compound channel (transversely), and anyone planning to use models to predict mixing. However, good background work is required in order to be able to apply the model to a problem and conclusions should be considered to be additions to current theory.