SOLUTE TRAVEL TIME DISTRIBUTIONS IN SOIL: A FIELD STUDY

Abstract

Flow patterns change as water and solute move through different soil horizons. This study was conducted to determine how solute travel times change with depth in a layered soil and to model experimental results with convection-dispersion and transfer-function models. Solutions of KBr were applied to the surface of a prewetted Silawa-variant loamy fine sand (fine, mixed, thermic Udic Paleustalf) and intercepted on tunnel ceilings at 0.3-, 0.9-, and 1.2-m depths. Samples were analyzed for volume, Br concentration, and arrival time. Breakthrough curves were constructed as probability density distributions and fitted with a lognormal transfer function and a convection-dispersion model. Pore water velocity varied with depth and was greatest at the 0.9-m depth in the layer with the coarsest structure and the finest texture. Travel time distributions at 0.9 and 1.2 m that were predicted from model calibration at 0.3 m did not fit measured values. Incorporating a measured retardation factor less than 1.0, an acceleration factor, greatly improved the model results, whereby the lognormal transfer function model clearly created the most accurate predicted travel time distributions at 0.9 and 1.2 m depths. This showed that surface measurements alone may not be enough to predict contaminant transport to deeper parts of the soil and that knowledge of the subsoil should be incorporated into model parameters.