Session D3 - Pollution in SCMs: characterisation

Urban desealing raises questions about the fate of contaminants in stormwater runoff. Particulate and dissolved phases are relatively well documented, but colloidal phases remain poorly studied in urban environments. The ROCOCO project (role of colloids and particles in the dynamics of contaminants in urban stormwater runoff) aims to characterize the distribution of contaminants within different fractions. A centrifugal ultrafiltration methodology allowed the separation of four fractions (>0.45 µm, 0.45 µm-30 kDa, 30-3 kDa, <3 kDa) from samples collected at two Observatoire de Terrain en Hydrologie Urbaine (OTHU) sites. Results showed that contaminants are predominantly associated with the particulate phase, and revealed variable distribution of metals (Al, Cr, Cu, Fe, Zn and Ti) between dissolved and colloidal phases of the <0.45 µm fraction. Pesticides and pharmaceuticals were mainly found in the dissolved phase. These data can provide information for the design of infiltration systems and contribute to the assessment of contaminant migration risks to groundwater.

Stormwater quantity and quality within 3 co-located swales with different vegetative conditions (grassed, vegetated wetland, and unvegetated wetland) were monitored along a stretch of US highway 23 near Upper Sandusky, Ohio. Wetland conditions occurring within roadside swales are often unintended; therefore, periodic maintenance may be done to remove hydrophytic vegetation and accumulated sediment from the swale, resulting in unvegetated (i.e., “dipped out”) wetland conditions. Between July 2023 to July 2024, ISCO 6712 automated samplers collected flow-paced runoff samples and runoff hydrographs used to calculate annual pollutant loading rates (kg/ha/yr) and watershed runoff depths (watershed-mm) for each treatment. A total of 760 mm of rainfall occurred during the monitoring period. The unvegetated wetland condition conveyed the greatest annual sediment (583 kg/ha/yr), total phosphorus (TP; 0.96 kg/ha/yr), total nitrogen (4.52 kg/ha/yr), and runoff depth (283 watershed-mm) of all vegetation treatments. This is likely driven by the exposure of runoff to bare soils and reduced channel roughness following wetland swale dredging. In contrast, the vegetated wetland condition conveyed a lower runoff depth (13.25 watershed-mm) than the unvegetated wetland swale along with the lowest sediment, TP, and dissolved phosphorus loading rates amongst all treatments. Unintended wetland conditions within roadside swales serve to reduce runoff volume as well as hyperaccumulate sediment and nutrients from runoff. Wetland roadside swales provide a cost-effective means to reduce pollutant loads.

Simultaneous partial Nitrification, Anammox, and Denitrification (SNAD) is an emerging process for nitrogen removal in stormwater bioretention systems. However, its performance remains difficult to predict due to a limited understanding of coupled hydraulic-biogeochemical interactions. This study presents a process-based, two-zone mechanistic model that explicitly couples plug-flow hydraulic dynamics from the MPiRe water quantity model with sequential nitrogen transformations: ammonification, nitrification, Anammox, and denitrification. The model was calibrated using the bootstrap MCMC method on laboratory-scale bioretention columns data. The model accurately reproduces observed effluent concentrations of NH4−N and NO3−N. Results revealed that NO2−N accumulation represents a critical bottleneck in the SNAD process, indicating unbalanced two-step nitrification rates. Moreover, it was also found that the COD plays a major role in the denitrification process and impacts the Anammox pathway. Model predictions demonstrate that optimising saturated zone volume and carbon availability can enhance complete denitrification pathways. The validated modelling framework enables biofilter design optimisation for enhanced nitrogen removal performance, helping to understand the underlying mechanism.

Stormwater runoff is emerging as a significant contributor of microplastics (MP) to downstream waterways. To begin to answer management questions on this emerging contaminant, we sought to quantify the extent to which existing biofilter (a.k.a. bioretention) stormwater control measures reduce MP, and to evaluate whether engineered media specifications are adequate to capture the range of MP sizes occurring in runoff. We conducted a field monitoring study to measure MP event mean concentrations (EMCs) and size distributions (20 µm to >500 µm) in influent and effluent and quantify accumulation in engineered media with respect to the media’s pore and particle size distributions. Across 18 storm events collected from7 biofilters, median MP EMC reduction was 72%, EMC reduction ranged from 68–100% for different size fractions. Physical filtration (i.e., straining) was identified as the dominant removal mechanism for MPs in biofilters, supported by the close correspondence between the removal of MPs and all particles remaining after extraction. Media with a larger proportion of pores < 20 m accumulated more MP. The strong alignment in occurrence and removal across a wide concentration range indicates that all particle counts may serve as a practical surrogate for estimating MP occurrence and BMP treatment in urban runoff.