The influence of pore water velocity on transport of sorptive and non-sorptive chemicals through an unsaturated sand

Abstract

A set of laboratory column experiments were conducted to study the transport of two relatively non-sorptive (Br, NO 3) and three sorptive (PO 4, simazine, linuron) chemicals at three water flux conditions ( q =6.9, 3.1, 0.62 cmh −1) through an unsaturated sand. Special attention was paid to the evaluation of the effects of pore-water velocity on transport of the chemicals as well as on their adsorption-desorption characteristics. The breakthrough curves (BTC) for a non-sorptive chemical measured at three water fluxes were more similar when the pore water velocity ( q/ θ) was computed using mobile water content ( θ m) than total water content (θ). The BTCs for Br and NO 3 were almost identical in shape at any given water flux. There was evidence of some attentuation of NO 3, and its extent was larger at lower fluxes. The BTCs for the adsorption process, observed at different depths under three different flux conditions could be scaled approximately using only two similarities, i.e. pore volume for ‘mobile’ water phase and Brenner number, VX/D ∗, where V is the average pore-water velocity, X is the depth and D ∗ is the hydrodynamic dispersion coefficient. Establishment of the similarity requires the condition that the retardation factor ( R) be represented approximately as a function only of concentration. This suggests that in the assumed equilibrium isotherm, S = G( θ) F( C). The assumption that G( θ) is a linear function of θ is satisfactory for practical purposes. Here, S is the specific sorbed chemical (weight per unit dry soil weight), θ the volumetric water content and C the concentration. A comparison between the modified BTCs of phosphate and simazine for adsorption and desorption processes leads to the conclusion that both processes are more or less reversible. Based on the similarity, this requires a severe condition that R must be constant, i.e. d F/d C is constant. A similar analysis could not be applied to linuron data, as they exhibited a two-step concentration increment in the BTCs, and it is suggested that this may be caused by the presence of organic matter in the sand. The retardation factors for the three fluxes were found to be only approximately constant, with an average of 1.5 for phosphate and 2 for simazine. Strictly, R was dependent on V in that R increased with deceasing flux. Furthermore, this dependence was larger for chemicals with a higher sorption characteristic (i.e. the dependence of simazine was greater than that of PO 4). Calculated BTCs for PO 4-P and simazine using these average values of R showed a good agreement with observation and the dispersion coefficients were directly proportional to pore water velocity. It is concluded that transport of some sorptive solutes can be described adequately on the assumption that R is constant.