Compost amendment on agricultural soil is a current practice to compensate the loss of organic matter. As a consequence, dissolved organic carbon concentration in soil leachates can be increased and potentially modify the transport of other solutes. This study aims to characterize the processes controlling the mobility of dissolved organic matter (DOM) in deep soil layers and their potential impacts on the leaching of organic contaminants (pesticides and pharmaceutical compounds) potentially present in cultivated soils receiving organic waste composts. We sampled undisturbed soil cores in the illuviated horizon (60-90 cm depth) of an Albeluvisol. Percolation experiments were made in presence and absence of DOM with two different pesticides, isoproturon and epoxiconazole, and two pharmaceutical compounds, ibuprofen and sulfamethoxazole. Two types of DOM were extracted from two different soil surface horizons: one sampled in a plot receiving a co-compost of green wastes and sewage sludge applied once every 2 years since 1998 and one sampled in an unamended plot. Results show that DOM behaved as a highly reactive solute, which was continuously generated within the soil columns during flow and increased after flow interruption. DOM significantly increased the mobility of bromide and all pollutants, but the effects differed according the hydrophobic and the ionic character of the molecules. However, no clear effects of the origin of DOM on the mobility of the different contaminants were observed.

The retention of crop residues as mulch on the soil surface in conservation agriculture systems greatly influences the fate of pesticides, as most of the applied pesticide is intercepted by mulch before moving to the soil. This work was conducted in order to model the effect of maize decomposition on glyphosate degradation in mulch and soil. Labelled 14 C-glyphosate degradation was monitored for 49 days in three treatments with the same soils but with maize residues at different stages of decomposition (0, 20 and 49 days). Fresh residues of maize (0 days) exhibited an evolution of their biochemical fractions to a greater extent than decomposed residues. Glyphosate mineralization was faster in the 0-day treatment in mulch residues and in the soil layer below the mulch. However, a greater formation of non-extractable residues (NERs) was observed in mulch residues and soils in the 20-and 49-day treatments than in the 0-day treatment. Modelling maize mulch decomposition with the COP-soil model indicated that microbial activity was different in the three treatments and depended on the initial composition of maize residues. Glyphosate mineralization in mulch and soil can be simulated with an assumption of co-metabolism by coupling the modules of pesticide degradation and mulch carbon decomposition. Glyphosate and its metabolites, including soluble and adsorbed fractions, were simulated with the same adsorption coefficients for all treatments. The simulation of NER formation, however, suggested that more than one microbial population may be involved in the degradation process and could be added in the future development of the model.