Each year, 65 million m3 of sediment are removed by dredging of harbors, canals, rivers and French rivers. These materials are often contaminated with heavy metals, pesticides and other organic pollutants accumulated over time, and should be stored ashore. The re-use of contaminated sediments under various scenarii is supervised by technical guidelines and regulations. These guides offer an evaluation methodology and thresholds for protecting human health and the environment, using the total solid concentrations of hydrophobic organic compound (HOC). This work evaluates the long-term environmental behavior of contaminated sediments and precisely what are the parameters and mechanisms controlling the mobility of organic pollutants, to assess risk rather than hazard. Five marine, canal and storm basin sediments were monitored in the laboratory for 18 months, and various methods of measuring solid, colloidal and dissolved fractions were performed. The organic matter in sediments is heterogeneous and has a multitude of sorption sites. The presence of exogenous origin of soot and condensed organic matter (Black Carbon) and hydrocarbons is responsible for the high retention of HOC in the sediment (high partition coefficients Kd and very slow desorption kinetics). In the solid phase, PCBs are mainly adsorbed on condensed fraction of organic matter and on the hydrocarbon phase, and nearly 80 % of PAHs are retained on the condensed fraction of organic matter. The dissolved concentrations in the porewater were obtained by passive sampler (LDPE). The results obtained in the liquid phase in batch leaching and percolation column shows, particularly for PAHs, an important colloidal vectorization (> 80 % and up to > 99.9 % of the heaviest PAHs). A simple method for the detection of the colloidal fraction of COH was tested. Classical modeling of the dissolved phase by partition coefficients, even refined, may significantly under-estimate the environmental risk (factor up to 1000) because the mobility of COH through colloids or dissolved macromolecules is not taken into account in the conventional transmission and transport models. In the study of sediment stored on the ground, the issue of dissolved (colloïdal) organic matter should be measured by a percolation test in the laboratory, and modeling of the transfer should be based on this result.
