An experimental cell was built over the Saint-Tite-des-Caps landfill in order to evaluate the efficiency of a cover with capillary barrier effect (CCBE). The latter is composed of a layer of deinking by-products (DBP) installed over a capillary barrier, which is in turn composed of a layer of coarse sand over a layer of gravel. In order to monitor the performance of this system, lysimeters, tensiometers, water content probes, settlement plates and weather station were installed. Design parameters were selected in such a way that the diversion length (DL) would be attained right before the length of the CCBE (30 m). As for the design of the cell itself, the design of the lysimeters also depended on the unsaturated properties of the materials employed, in particular their hydraulic conductivity functions. The lysimeters were designed in order to remain functional at least during the study period. The water retention curves (WRC) of the materials composing the CCBE were determined in the laboratory. These results allowed obtaining the hydraulic conductivity functions (k-fct), which is the principal element, as far as optimisation of the design of the CCBE and of the lysimeters are concerned. Using the k-fcts, it is possible to evaluate the theoretical DL. Numerical simulations showed that the DL can be significantly altered by slight changes in the parameters used to describe the WRC and the k-fcts. For each material, the WRC was determined using various experimental methods; and they all gave slightly different parameters. This caused the significant variation in the DL. The DL determined based on field data allowed for the selection of the most representative WRC. Analysis of field data, made it possible to show the change in saturated hydraulic conductivity, ksat, of DBP decreased by nearly one order of magnitude, which affected its WRC and k-fct. Field monitoring also shows that the hydraulic barrier remains saturated and that the volume of water reaching the sand/gravel capillary barrier increased progressively, as one moves downslope. Data from the four tensiometers, placed along the sand/gravel interface, and the measured infiltration rates in the 3 lysimeters installed in the gravel layer show that the DL varied between 24 m and 29 m during the 2006 monitoring period. During some short periods in 2007, the DL was greater than the 30 m of the cell. The infiltration rate into the gravel layer at the toe of the slope was continuously lower than 3,0 x 10-9 m/s, i.e. lower than the more stringent standards of low permeability cover design (Règlement québécois sur l'enfouissement et l'incinération des matières résiduelles ). In order to quantify the amount of infiltrating water trough DBP and sand layers, several simulations were performed using by Visual HELP software. Infiltrations rates obtained thereof were used as boundary conditions in steady state simulations of the response of the capillary barrier system. Software SEEP/W and Hydrus were used in these simulations and allowed to check the exactness of the initial design hypotheses and assumptions. As the measured DL is somehow similar to the simulated DL, it was concluded that the design methodology was appropriate. The simulations results also show that much greater DL can be attained using such an alternative cover system, depending on the hydraulic properties of the materials.