Field scale variability of solute transport parameters and related soil properties

The spatial variability of transport parameters has to be taken into account for a reliable assessment of solute behaviour in natural field soils. Two field sites were studied by collecting 24 and 36 small undisturbed soil columns at an uniform grid of 15 m spacing. Displacement experiments were conducted in these columns with bromide traced water under unsaturated steady state transport conditions. Measured breakthrough curves (BTCs) were evaluated with the simple convective-dispersive equation (CDE). The solute mobility index (MI) calculated as the ratio of measured to fitted pore water velocity and the dispersion coefficient (D) were used to classify bromide breakthrough behaviour. Experimental BTCs were classified into two groups: type I curves expressed classical solute behaviour while type II curves were characterised by the occurrence of a bromide concentration maximum before 0.35 pore volumes of effluent (MI<0.35) resulting from preferential flow conditions. Six columns from site A and 8 from site B were identified as preferential. Frequency distributions of the transport parameters (MI and D) of both sites were either extremely skewed or bimodal. Log-transformation did not lead to a normal distribution in any case. Contour maps of bromide mass flux at certain time steps indicated the clustering of preferential flow regions at both sites. Differences in the extent of preferential flow between sites seemed to be governed by soil structure. Linear cross correlations among transport parameters and independently measured soil properties revealed relations between solute mobility and volumetric soil water content at time of sampling, texture and organic carbon content. The volumetric field soil water content, a simple measure characterising the soil hydraulic behaviour at the sampling location, was found to be a highly sensitive parameter with respect to solute mobility and preferential flow situations. Almost no relation was found between solute transport parameters and independently determined soil properties when non-preferential and preferential samples were considered separately in regression analyses. Future work should concentrate to relate integrated parameters such as the infiltration rate or the soil hydraulic functions to solute mobility under different flow situations.