Adsorption distribution impact on preferential transport within horizontal flow constructed wetland (HFCW)

Predicting preferential water flow and solute transport in soils is in the interest of scientists and engineers in the fields of agricultural soil, forest hydrology, soil physics and wastewater treatment by constructed wetland. In artificial wetlands, any preferential pathway induces an insufficient residence time of pollutants in the soil, making an incomplete and unfinished biodegradation process, a wrong evaluation of the hydraulic residence time of the system which would hinder its management in a complete system with several entities of treatment and a non-homogeneous growth of the biofilm in the solid filter mass. This paper is a contribution in tracer experiment data analysis within a horizontal flow constructed wetland built in a storm water basin. A two-dimensional numerical model was used to simulate flow and reactive solute transport. The flow model was successfully calibrated in very dry soil conditions. The adsorption profiles used in the reactive transport modeling are those of five molecules: metolachlor, atrazine, deethylatrazine (DEA), deisopropylatrazine (DIA), and hydroxyatrazine (HA). We show that the adsorption distribution is an internal factor of soil which is responsible to the “preferential” pathway transport in a homogeneous gravel texture. The mean residence time of pollutants within the filter is strongly correlated with the average value of the adsorption coefficient. Moreover, we note a lack of significant impact of the heterogeneity of the medium on the statistical moments of breakthrough curve. It appears that a uni-modal breakthrough curve is not significant to the absence of preferential flow in the medium and at least a two-dimensional display can provide sufficient evidence on the presence or absence of preferential pathways. Finally, using a PLS regression, it is possible to perfectly discriminate the number of peaks of concentration and the asymmetry of the breakthrough curve.