Dynamic modelling of nitrification in vertical flow constructed wetlands

Constructed wetlands (CWs) are attractive system for wastewater treatment for small communities, but interactions between the removal processes are often badly understood. Modelling of CW systems is a way to unravel them and improve their design. A modelling study with the HYDRUS/CW2D model was carried out on a French first-stage vertical flow constructed wetland (VFCW) at Evieu (Ain). Characterizing the hydraulic properties of the filtering material is a prerequisite. However hydraulic characterization is difficult due to varying proportion of organic matter in the filter. I present a combined laboratory and in-situ approach to achieve it. The laboratory approach gives prior estimates that are later refined by in-situ measurements and inverse modelling. Thanks to a large experimental dataset, it led to a reliable hydraulic parameter set allowing accurate simulation of water content within the VFCW. Prior to biological modelling, i developed and validated an original method to measure nitrification rates at the different depths of the VFCW and consequently at different organic matter contents. It is specifically adapted for partially saturated matrix and is based on respirometric principles. The obtained maximum growth rate coefficients for the autotrophic biomass are critical inputs for the VFCW plant modelling. Once the hydraulic calibration carried out and some of the biological parameters assessed, we simulate reactive transport. Particular attention was devoted to the nitrogen fate. During a feeding period, influent fractionation was determined, influent and effluent pollutant concentrations continuously monitored as well as in-situ oxygen concentrations. These measurements have been used to fit the biological model by adjusting kinetic parameters highlighting the need to include adsorption of ammonium. Good agreement between model and experimental data is obtained based on 2-4 h average concentrations, while instantaneous measurements are badly fitted.(AGRO 3) -- UCL, 201