Systèmes extensifs de gestion et de traitement des eaux urbaines de temps de pluie

Urban runoff management has become a priority for many years, regarding the issues of flooding and water course pollution that jeopardize the objectives of the European Framework Directive 2000/60/CE. As an extensive technique, subsurface flow constructed wetlands have proved to be reliable solutions for the treatment of sewage, and have been used for more than 20 years. Their adaptation to other types of functions or effluents has even been implemented over the last ten years (tertiary treatment, dilute effluents, sludge drying reed beds), and the management of urban runoff is part of this extension process. Despite this system combines both storage and treatment capacities, (Uhl and Dittmer, 2005) highlight dysfunctions as reed death during long rest periods or performance losses. These authors also underline the needs in optimizing design rules and tools.Then, the SEGTEUP project combines a large-scale pilot monitoring and hydrodynamic modeling, aiming at providing some basis for the implementation of a reliable design tool, applicable over the entire French territory. Then, the pilot study aimed at testing different materials and filter configurations, while a mechanistic approach was implemented to highlight the main flow influence parameters.Anyway, filtration of suspended solids appears as the major pollutant removal process, allowing the retention of at least 70% of SS and associated pollutants whatever the material considered. Nevertheless, dissolved organic matter adsorption remains negligible, while ammonium elimination is poor for the coarsest material. In the opposite, adsorbent zeolite present high ammonium retention abilities and ensures constantly low discharge levels. The use of this material is particularly recommended in case of specific issues concerning high pollutant loads or low available surface area. Furthermore, micropollutants (metals and PAHs) mainly bound with particles, are highly removed by filtration of SS. In our study, dissolved micropollutants concentrations were too low to highlight robust conclusions. However, further research are needed to better understand metals retention and release, and PAHs degradation processes, since these mechanisms are strongly dependent of pH and redox conditions and dynamics. Finally, the study of hydrodynamics shows that the throttle value influences adsorption and degradation reactions through variations in hydraulic retention time, particularly during saturated phases. On the other hand, the feeding structure is of special importance since a homogenous distribution of influent at the filter surface is needed to avoid hydraulic shortcuts and corresponding dead volumes. Finally, a HYDRUS 1D model was implemented around urban runoff issues, but cannot be used as a predictive design tool. Indeed, the throttle value, represented by a virtual layer with low conductivity, must be re-calibrated for each simulated event. Then, a 2D adaptation of the model is required, as well as the integration in the HYDRUS code of appropriate singular head losses as boundary conditions. In the end, the developed simplified model of water flow and pollutant degradation should be coupled with a sewer system model, aiming at creating a flexible and adaptable design tool.