Session C9 - Infiltration systems: treatment performance

Stormwater infiltration systems are increasingly implemented as alternatives to conventional piped drainage systems. However, the infiltration of stormwater runoff raises concerns over groundwater pollution. The research on stormwater pollutant transport towards the saturated zone in urban areas is relatively limited, and this study contributes to this area by investigating the quality of stormwater collected from two contrasting catchments on infiltration through three pilot infiltration basins. During three experimental runs, physicochemical parameters, selected metals and organic micropollutants (e.g. PAHs, organotin compounds, BPA, PFAS) were monitored in both the inflow and outflow. Zn and specific organic pollutants (e.g. BPA and PAH) showed reductions, whereas for Cu and organotin compounds we found increases in outflow concentrations relative to inflow. In particular, the organotin compounds MBT and MOT were reported at higher levels in the outflow, indicating leaching from soil media or construction materials within the infiltration basins. Further analyses of the soil media and potential material leaching will help clarify internal contaminant sources. The findings highlight that infiltration systems may both attenuate and introduce contaminants, with implications for the quality of shallow groundwater.

In urban areas, stormwater management projects are increasingly faced with constrained sites where the soil displays high anthropogenic backgrounds (backfills, brownfield sites). A methodology is needed to assess the potential remobilization of contaminants from such soil horizons, and thereby the suitability of implementing stormwater infiltration. The present approach combines (1) hydrodynamic modelling in the vadose zone to estimate the water depth reaching the contaminated horizon, and (2) leaching column experiments to assess the concentration dynamics for various contaminants. Modelling results suggest that capillary processes and evapotranspiration may contribute to mitigate exfiltration fluxes by up to 60% in “decentralized” systems. The experimental method was applied to five soils representing a gradient of contamination but excluding cases of severe pollution. Metals, metalloids and hydrocarbons typically exhibited decreasing concentrations at the column outlet, with dissolved-phase contamination levels of the same order of magnitude as the dissolved concentrations found in urban runoff. Flow interruption had a low to moderate impact on contaminant remobilization. The “steady-state” phase nevertheless represented a significant contribution to the total mass load exported from the column. Monitoring of effluent pH also revealed highly alkaline pH for one soil, thus pointing out a potential additional impact of stormwater infiltration that should be looked upon.

Bioswales play a major role in managing roadway runoff, yet their hydrologic performance varies widely. This study analyzes approximately 300 roadside bioswales in New Haven, Connecticut, the largest dataset of its kind to our knowledge, to quantify how watershed size, land-use setting, vegetation cover, and inlet sedimentation influence infiltration capacity. Watersheds were manually delineated using GIS-based flow interpretation, and double-ring infiltrometer tests were conducted at each inlet and outlet. Inlet infiltration was consistently lower, with a median initial rate of 3.5 cm/min, compared with 5.0 cm/min at outlets, reflecting sediment and pollutant accumulation from roadway runoff. Land-use differences were substantial: residential bioswales showed higher median infiltration (≈4.45 cm/min) than commercial systems (≈4.00 cm/min). Surprisingly, infiltration increased modestly with watershed size (ρ ≈ 0.12, p < 0.001), contrary to the expectation that larger watershed loads would increase clogging and reduce hydraulic conductivity. Vegetation had a positive but modest influence (ρ ≈ 0.10–0.11). These city-scale findings demonstrate that bioswale performance emerges from interacting influences of watershed loading, land-use pressures, and inlet sedimentation rather than design differences, offering guidance for future planning of green-infrastructure networks.

Stormwater (SW) infiltration is increasingly used to restore urban hydrology, but its effects on groundwater (GW) quality across different land uses remain poorly constrained. We investigated GW contamination in five aquifers in northern Stockholm that have been influenced by SW for several decades of urbanization. In total, 121 GW samples from 31 locations were analyzed, covering, natural lands, arable, mixed, and urban catchments. Sites were characterized by hydrogeological vulnerability and proximity to potential sources, including hazardous goods routes, suspected PFAS sites, and known contaminated sites. Sixty-six parameters were analyzed, including nutrients, major ions, trace metals, radon, 13 chlorinated VOCs, and 21 PFASs. Many trace metals, short-chain PFASs, and PFAS4 were frequently detected, whereas most longer-chain PFASs (C9−C13 PFSAs and C11−C13 PFCAs) and several VOCs were rarely or never quantified. Redundancy analysis showed that land use, aquifer identity, and proximity (<500 m) to suspected PFAS sources together explained 29% of the variability in GW concentrations. PFASs and trace metals were positively associated with urban areas, while natural, arable, and mixed catchments showed weaker or negative associations. The results indicate that SW infiltration can mobilize both conventional and emerging contaminants into GW, with impacts that are altered by land use, aquifer conditions, and nearby pollution sources.