Longitudinal dispersion in rivers: The persistence of skewness in observed data

For some selected field studies of longitudinal dispersion the variances of observed concentration distributions increase linearly with time, as predicted by solutions to the one-dimensional Fickian-type diffusion equation. However, observed values of the skewness coefficient are almost constant, so deviations from the theoretical values become greater with increasing time. The usual assumption that concentration distributions converge to the Gaussian solution of the one-dimensional diffusion equation is not supported by the empirical observations. One explanation for the persistence of the skewness is the existence of dead zones which temporarily trap portions of the dispersant. The coefficients of skewness predicted by a dead zone model fit the observed values more closely, but this model is, like the Fickian model, characterized by a decay in the skewness, which is not exhibited by the observed distributions.