Research papersUnsaturated hydraulic behaviour of a permeable pavement: Laboratory investigation and numerical analysis by using the HYDRUS-2D model

- Water flow through the pavement consisted of 3 layers is measured and simulated.
- Experimental measurements of hydraulic properties of the construction materials are carried out.
- To investigate the hydraulic parameters effect, the Morris sensitivity analysis is carried out.
- The model is calibrated by using the Levenberg-Marquardt Algorithm included in HYDRUS-2D.
- The model is validated on an independent set of data with good results.

An adequate hydrological description of water flow in permeable pavement systems relies heavily on the knowledge of the unsaturated hydraulic properties of the construction materials. Although several modeling tools and many laboratory methods already exist in the literature to determine the hydraulic properties of soils, the importance of an accurate materials hydraulic description of the permeable pavement system, is increasingly recognized in the fields of urban hydrology. Thus, the aim of this study is to propose techniques/procedures on how to interpret water flow through the construction system using the HYDRUS model. The overall analysis includes experimental and mathematical procedures for model calibration and validation to assess the suitability of the HYDRUS-2D model to interpret the hydraulic behaviour of a lab-scale permeable pavement system. The system consists of three porous materials: a wear layer of porous concrete blocks, a bedding layers of fine gravel, and a sub-base layer of coarse gravel. The water regime in this system, i.e. outflow at the bottom and water contents in the middle of the bedding layer, was monitored during ten irrigation events of various durations and intensities. The hydraulic properties of porous concrete blocks and fine gravel described by the van Genuchten functions were measured using the clay tank and the multistep outflow experiments, respectively. Coarse gravel properties were set at literature values. In addition, some of the parameters (Ks of the concrete blocks layer, and α, n and Ks of the bedding layer) were optimized with the HYDRUS-2D model from water fluxes and soil water contents measured during irrigation events. The measured and modeled hydrographs were compared using the Nash-Sutcliffe efficiency (NSE) index (varied between 0.95 and 0.99) while the coefficient of determination R2 was used to assess the measured water content versus the modelled water content in the bedding layer (R2=0.81÷0.87). The parameters were validated using the remaining sets of measurements resulting in NSE values greater than 0.90 (0.91÷0.99) and R2 between 0.63 and 0.91. Results have confirmed the applicability of HYDRUS-2D to describe correctly the hydraulic behaviour of the lab-scale system.