Insights on permeable pavement hydraulic performance from large-scale laboratory experiments and physically based modelling

Among other Sustainable Urban Drainage Systems, Permeable Pavements (PPs) are one can be easily retrofitted in the urban environment. However, they suffer of clogging phenomena that reduces their efficiency over time. Laboratory experiments to assess the hydraulic performance of a newly constructed PP subjected to different rainfall intensities have been conducted using a large-scale laboratory model (2x6 m2 with 1.2\% slope). The surface of the upstream portion (1.7x2 m2) is impermeable to simulate runoff generation over impermeable surfaces, while the downstream portion (4.3x2 m2) is realized with PICP. The downstream vertical side of the PP is made of permeable bricks and two gutter channels are placed crosswise to separately collect runoff and subsurface discharge. The remaining sides, as well as the bottom of the model, are impermeable. The filter package below the PICP consists of three layers: 5 cm bedding (3-6 mm gravel), a 10 cm base layer (8-12 mm gravel) and a 30 cm sub-base layer (20/40 mm gravel), which is laid on top of a 40 cm layer of native sand (silty sand with d50=0.23 mm). A geotextile separates the bedding and base layers and a 4m long drainpipe (D=150 mm) was inserted in the sub-base layer. The facility is equipped with probes on both lateral sides: 6 tensiometers in the native sand, 4 water content reflectometers in the base and sub-base layers, and 3 piezometers to record water table evolution throughout the experiments and degree of saturation of the filter layer package. Runoff and subsurface discharge are separately conveyed to two tipping bucket rain gauges. A rainfall simulator is used to generate quite uniform rainfall distribution (80 - 150 mm/h intensity) for 15 minutes or 30 minutes. Moreover, an Integrated Surface-Subsurface Hydrological model (CatHy) has been used to model the permeable pavement, assess and support data collected from the laboratory experiments. Results from the laboratory experiments performed have proven the efficiency of a newly constructed permeable pavement to very intense rainfall events. The monitoring with spatially distributed sensors allowed to assess the evolution in time of the water table as well the recovery phase to pre-event conditions after the event. This is useful to assess the effect of repeated rainfall events at short distance in time. For each experiment performed, a rapid increase of subsurface discharges was recorded by the tipping bucket, whereas surface runoff occurred only for short and intense rainfall events (approximately 150 mm/h for 15 min). The system did not reach saturated conditions in any of the performed experiments due to the high permeability of the filter layer package. The monitoring with spatially distributed sensors also allowed to assess the heterogeneities of the physical processes (synthetic rainfall events, infiltration processes) as well as of the filter layer package. Future laboratory experiments simulating clogging phenomena will be performed and compared to the results obtained from the developed experiments up to now and of the ISSH model.