Dispersion coefficients analysis by ADCP measurements, tracer test and numerical simulations

Abstract

The longitudinal and transverse dispersion coefficients of the two-dimensional advection–dispersion model were calculated using velocity and concentration data obtained from tracer tests in a meandering river. The velocity-based dispersion coefficients calculated by applying the theoretical equations to the velocity data were compared with the dispersion coefficients calculated by applying numerical simulation to the measured concentration curves. The results showed that the cross-sectional average values of the dimensionless longitudinal dispersion coefficient KL/ u∗H H calculated from the velocity profiles were of the order of 6.80 (range 1.40–11.0). This value is very close to Elder’s result (5.93) and to the best fit estimated from the concentration curves through numerical simulation. The average dimensionless transversal coefficient calculated by velocity method ranged 0.012–0.66, with reach average of 0.20 while the calibration value estimated by numerical simulation was 0.15. Although at the bend apex, the local velocity method values are up to seven times higher than the average, demonstrating the relevance of secondary currents in the mixing process. On average, the velocity method showed that between the apex of the bend and the straight reach, the transverse dispersion varied up to 4 times while the change in longitudinal dispersion was up to 1.5 times and in the opposite sense.