Influence of the vegetative cover on the fate of trace metals in retention systems simulating roadside infiltration swales.

Large-scale outdoor mesocosms were designed and co-contaminated with metals (Cd, Pb, Zn) and organic compounds to better understand the complex functioning of urban roadside swale environments. Infiltration systems were planted with macrophytes (P. arundinaceae, J. effusus and I. pseudacorus) or grassed, and natural or spiked target metals were monitored over two years. In the non-spiked mesocosms, atmospheric metal inputs were slightly higher than outputs, leading to low metal accumulation in topsoils and to very low outflow water contamination (<0.7% of the initial metal stock). In the spiked infiltration systems that simulated point pollution through water inflow, transfer of the initial stock of metals to the deeper soil layers was quite low and outflow water contamination was very low (<0.6% of the initial stock). The main metal output from these systems occurred in the first days of their installation because of the high metal solubility in water and insufficient plant cover at that time. The infiltration systems stabilized after a few weeks, probably because of stronger sorption to soil aggregates, and because of plant root development. Mephytoextraction in plant roots was more efficient in mesocosms planted with P. arundinacea and grass. Metal phytoextraction in plant aerial parts was also better for grass and P. arundinacea, when considering metal standing stocks instead of their concentration in plants. J. effusus was a good metal accumulator, but its low aboveground biomass development was less favorable to metal removal through harvesting.