D1.4 Ranking of pollutants on their risk of accumulation and the risk of break-through to the groundwater

1 D1.4 Ranking of pollutants on their risk of accumulation and the risk of break-through to the groundwater June 2025 Anders R. Johnsen, Ulla E. Bollmann and Trine Henriksen (GEUS – Geological Survey of Denmark and Greenland) Jan H. Christensen and Thomas M.M. Karlsson (University of Copenhagen) Data driven implementation of hybrid nature-based solutions for preventing and managing diffuse pollution from urban water runoff Ref. Ares(2025)4705957 - 12/06/2025 2 D1.4 Ranking of pollutants on their risk of accumulation and their risk of breakthrough to the groundwater Work Package WP1 Deliverable lead GEUS Author(s) Anders R. Johnsen, Ulla E. Bollmann and Trine Henriksen (Geological Survey of Denmark and Greenland, GEUS) Jan H. Christensen and Thomas M.M. Karlsson (University of Copenhagen, UCPH) Contributors: Jorge R. Hernandez (University of Cantabria) Simone Lippi (Acque SpA) Contact [email protected] Grant Agreement number 101060638 Start date of the project / Duration 1 September 2022 / 42 months Type of deliverable (R, DEM, DEC, other) R - Research Dissemination level (PU, SEN) PU - Public Project website www.d4runoff.eu Document history Version Date Authors (organisation) 0.01 24.10.2024 GEUS, report draft 0.02 29.10.2024 GEUS + UCPH, report revision 1.00 01.11.2024 VCS, report submitted 1.01 11.06.2025 GEUS, revision according to reviewers’ comments 3 R=Document, report; DEM=Demonstrator, pilot, prototype; DEC=website, patent fillings, videos, etc.; OTHER=other PU=Public, SEN=Sensitive, limited under the conditions of the GA ACKNOWLEDGEMENTS This project has received funding from the European Union’s Horizon Europe research and innovation programme under grant agreement No 101060638. COPYRIGHT STATEMENT The work described in this document has been conducted within the D4RUNOFF project. This document reflects only the D4RUNOFF Consortium views, and the European Union is not responsible for any use that may be made of the information it contains. This document and its content are the property of the D4RUNOFF Consortium. All rights relevant to this document are determined by the applicable laws. Access to this document does not grant any right or license on the document or its contents. This document or its contents are not to be used or treated in any manner inconsistent with the rights or interests of the D4RUNOFF Consortium or the Partners detriment and are not to be disclosed externally without prior written consent from the D4RUNOFF Partners. Each D4RUNOFF Partner may use this document in conformity with the D4RUNOFF Consortium Grant Agreement provisions. 4 Executive Summary The overall aim was to identify relevant stormwater pollutants, study the fate of stormwater pollutants in nature-based solutions for handling stormwater, and finally rank the pollutants based on their risk of accumulation in soil and the risk of breakthrough to the groundwater. Thirty-eight urban stormwater sources were sampled and analysed in Deliverable 1.3 - Inventory of pollutants in urban storm water across Europe. The pollutants in the inventory were initially ranked based on their detection frequency and pollutants present at >25% of sampled stormwater stations were used for ranking. By setting a threshold at 25%, we evaluated only pollutants with a high probability of presence in urban stormwater and avoided compounds that derived solely from wastewater stations. Model pollutants were selected from the draft inventory produced in Deliverable 1.3. We chose compounds that were found in high frequency and originated from the city environment such as triazines, compounds associated with vulcanization and the wear of car tires, and corrosion inhibitors. Five stations with nature-based solutions (NBS) for stormwater handling at the demonstration sites were selected for fate studies. The stations represent different strategies for using nature-based solutions for handling urban surface runoff and have highly variable soil types from coarse gravel below a permeable pavement in a parking area, filter sand in two rainbeds, compact clay in a detention basin and highly organic sediment in a retention basin. The degradation potential was tested by spiking NBS soil with a mixture of the model pollutants in the lab and follow the decrease in concentrations over time. The rain beds drain in less than a day, suggesting that the retention time in the active layer is only a few days at best for mobile pollutants. The rather long half-lives of the model pollutants and the low numbers of microbial degraders show that the model pollutants may leach from the rain beds to less bioactive subsurface layers, if the pollutants are mobile. The same applies to the permeable pavement in the parking bays where retention times in the geotextile and the gravel layers may be shorter. The situation may be different in detention basins where retention times may be weeks or months. A desorption study was carried out to determine which pollutants are released from the topsoil/sediment at the five stations. The desorption study gave information on which accumulated compounds, or even which concentrations for target compounds, are mobile and may leach. At the same time, it gave an indication of the bioavailability of these compounds. To measure which pollutants move towards the groundwater from the naturebased solutions, suction cups were installed approximately one meter below the ground or one meter below the sediment surface at the five NBS stations, and water was collected and analysed for pollutants. Finally, we ranked stormwater pollutants according to their risk of leaching from NBS towards the groundwater and/or accumulating in NBS topsoil. We were limited by the very limited information available for many of the compounds. The source stormwater data was combined with RP-LC-HRMS data from the suction cups from the NBS stations. The compounds were then sorted according to their presence in suction cup water. The rest of the compounds were sorted according to logD (pH 7.0), i.e., according to their tendency to 5 phase-partition into an organic phase. We recommend that nature-based solutions for local handling of urban stormwater should be constructed in such a way that the common, mobile stormwater pollutants are retained or degraded before water is infiltrated. Concentrations of trace elements and polycyclic aromatic hydrocarbons (PAHs) in soil were determined for three model NBS stations. In addition to the immobile micropollutants, PAHs and the trace elements Sn, Zn, B, Pb, Cu, Cr and V may also accumulate in topsoil based on their enrichment in topsoil compared to deeper layers. The PAHs probably have low leaching potentials due to a combination of high sorption to organic matter (high logP) and very large microbial PAH degrader populations in the topsoil. The tested trace elements probably also do not leach as they generally showed little or no depletion in topsoil compared to deeper layers. 6 1 Table of Contents 1 Table of Contents ......................................................................................................... 6 2 INTRODUCTION ............................................................................................................ 9 3 INITIAL RANKING OF ORGANIC MICRO-POLLUTANTS BASED ON DETECTION FREQUENCY IN SOURCE STORMWATER ........................................................................ 9 4 NATURE-BASED SOLUTIONS SELECTED FOR FATE STUDIES .............................13 4.1 Hørdumsgade rain bed, Odense (Denmark). ........................................................13 4.2 Kallerupvej rain bed, Odense (Denmark). .............................................................15 4.3 Car park, Santander (Spain). ................................................................................17 4.4 Centro Pontedera detention basin, Pontedera (Italy). ............................................19 4.5 Søparken Syd retention basin, Odense (Denmark) ...............................................21 5 FATE-STUDIES – DEGRADATION OF ORGANIC MICROPOLLUTANTS ..................23 5.1 Model pollutants ....................................................................................................23 5.2 Degradation potential ............................................................................................24 5.3 Quantification of microbial degrader populations ...................................................27 6 FATE-STUDIES – WHICH ORGANIC MICRO-POLLUTANTS LEACH FROM THE CASE STUDY NBS? ...........................................................................................................28 6.1 Desorption ............................................................................................................28 6.2 What leaches from the NBS stations? ...................................................................29 6.3 Results: desorption and leaching ..........................................................................33 7 FATE-STUDIES – RANKING OF POLLUTANTS ACCUMULATED IN THE NBS ........35 8 RANKING OF POLLUTANTS .......................................................................................38 8.1 Method ..................................................................................................................38 8.2 Results and discussion .........................................................................................39 9 CONCLUSION ..............................................................................................................42 10 REFERENCES ..............................................................................................................44 11 Annex A. Concentrations of trace elements and PAHs in topand bottom soil at Hørdumsgade, Kallerupvej and Centro Pontedera ..........................................................45 7 List of Tables Table 1. Organic target pollutants in the stormwater inventory detected at more than 25% of the source stormwater stations. Pollutants are sorted according to detection frequency. Some pollutants were detected by both methods in different frequencies. ............................10 Table 2. Organic suspect micropollutants in the stormwater inventory detected at more than 25% of the source stormwater stations. Pollutants are sorted according to detection frequency. Confidence levels: RP-LC, confidence level from 0-100 in RP-LC-HRMS; HILIC, confidence level from 1-5 according to Schymanski et al. (2014). Some pollutants were detected by both methods in different frequencies. ...............................................................11 Table 3. Model compounds selected for fate studies. ...........................................................23 Table 4. Degradation of model compounds in microcosms. Degradation rates are presented as half-lives in days (T1/2) for simple first-order fitting. 2-ABT: 2-aminobenzothiophene. 2BZSA: benzothiazole-2-sulfonic acid. 2-HBT: 2-hydroxybenzothiazole. 5-MBT: 5-methyl-1Hbenzotriazole. 6-PPDQ: 6PPD-quinone. 1,3-DPG: 1,3-diphenylguanidine. HMM: hexa(methoxymethyl)melamine. MLM: melamine. TERB: terbutryn......................................26 Table 5. Most probable number (MPN) of degrader microorganisms that can grow on selected model compounds as single source of energy and energy. ....................................28 Table 6. Target analyses (RP-LC-HRMS) of water from the desorption study and from suction cups installed in-situ. The median concentration is the median of the stations where the compound was detected. ................................................................................................33 Table 7. Suspect screening (RP-LC-HRMS) of water from the desorption study and from suction cups jnstalled in-situ. ................................................................................................35 Table 8. Enrichment factors for trace elements found in concentrations >LOQ in both the top layer and the bottom layer. Enrichment factors were calculated as the top-to-bottom concentration ratio for three replicate holes at each station. Trace elements are sorted according to decreasing mean enrichment factor. ................................................................36 Table 9. Enrichment factors for PAHs found in concentrations >LOQ in both the top layer and the bottom layer. Enrichment factors were calculated as the top-to-bottom concentration ratio. PAHs are sorted according to decreasing mean enrichment factor. .............................37 Table 10 Ranking of stormwater pollutants according to the risks of leaching from NBS (decreasing from top to bottom) and accumulation in NBS topsoil (increasing risk from top to bottom). A total of 12 suction cup samples represents leaching from four traffic impacted NBS stations. Red: compounds with high risk of leaching based on detections in NBS suction cups. Black: compounds ranked according to logD at pH 7,0, the ranking of these compounds is uncertain. Blue: QUATs/BACs with very low leaching risk. Pollutants used for degradation experiments and/or MPN-counting of specific degrader microorganisms are highlighted in bold. ...............................................................................................................40 Table 11. Mean concentrations of trace elements in soil samples (n=3) from the NBS-station Hørdumsgade. .....................................................................................................................45 Table 12. Mean concentrations of trace elements in soil samples (n=3) from the NBS-station Kallerupvej ...........................................................................................................................45 Table 13. Mean concentrations of trace elements in soil samples (n=3) from the NBS-station Centro Pontedera .................................................................................................................46 Table 14. Concentrations of PAHs in pooled soil samples (n=3) from the NBS-stations Hørdumsgade, Kallerupvej, and Centro Pontedera. .............................................................47 8 List of Figures Figure 1. Hørdumsgade sampling location in Odense. .........................................................13 Figure 2. Rainbed at Hørdumsgade......................................................................................14 Figure 3. Drainage area of the Hørdumsgade rain bed. ........................................................14 Figure 4. Schematic cross-section of the Hørdumsgade rain bed. ........................................15 Figure 5. Kallerupvej sampling location in Odense. ..............................................................15 Figure 6. Rainbed and drained area at Kallerupvej. ..............................................................16 Figure 7. Drainage area of the Kallerupvej rain bed ..............................................................16 Figure 8. Schematic cross-section of the Kallerupvej rain bed. .............................................17 Figure 9. car park sampling location in Santander. ...............................................................17 Figure 10. Permeable parking bays. .....................................................................................18 Figure 11. Schematic cross-section and drainage area of the permeable parking bay..........18 Figure 12. Car park sampling location in Pontedera. ............................................................19 Figure 13. Centro Pontedera detention basin. ......................................................................19 Figure 14. Ditch along road that drains to the Centro Pontedera detention basin. ................20 Figure 15. Drainage area of the Centro Pontedera detention basin. .....................................20 Figure 16. Søparken Syd sampling location in Odense. .......................................................21 Figure 17. Søparken Syd retention basin..............................................................................21 Figure 18. Drainage area of Søparken Syd retention basin. .................................................22 Figure 19. Søparken Syd retention basin, stormwater outlet to the basin. ............................22 Figure 20. Examples of degradation curves and fitting of simple first-order functions. ..........26 Figure 21. Stainless steel suction cup fitted to stainless steel tubing. ...................................29 Figure 22. Schematic cross section of NBS with installed suction cup. .................................30 Figure 23. Suction cup installed in Kallerupvej rain bed. .......................................................30 Figure 24. Suction cup installed in Hørdumsgade rain bed. ..................................................31 Figure 25. Suction cup installed in Søparken retention basin................................................31 Figure 26. Suction cups installed in Centro Pontedera detention basin. ...............................32 Figure 27. Collector pipe (drainpipe) in Santander permeable parking bay. ..........................32 Figure 28. Criteria used for ranking of pollutants. .................................................................39 9 2 INTRODUCTION This deliverable (D1.4 “Ranking of pollutants on their risk of accumulation and the risk of breakthrough to the groundwater”) represents work carried out in the Horizon Europe project “Data driven implementation of hybrid nature-based solutions for preventing and managing diffuse pollution from urban water runoff – D4RUNOFF”, work package 1 “Novel detection methods for urban runoffs pollutants characterization”, tasks 1.3 “Screening for CECs, pathogenic indicators and microbial resistance in urban runoff” and 1.4 “Determining the fate of pollutants in NBS“. The overall aim was to identify relevant stormwater pollutants, study the fate of stormwater pollutants in nature-based solutions for handling urban stormwater, and finally rank the pollutants based on their risk of accumulation in soil and their risk of breakthrough to the groundwater. The work was organized in three sections: 1. Initial ranking of organic micropollutants. The pollutants were ranked based on their detection frequency in a stormwater inventory produced in Deliverable 1.3 to identify the most common organic stormwater pollutants across Europe and select relevant model pollutants. 2. Selection and characterization of five nature-based solutions (NBS-stations) for handling urban stormwater. These stations were used for determining the fate of selected pollutants. Suction cups were installed at the stations below the surface to determine which compounds leach towards the groundwater. Soil was collected to determine mobility of compounds present in the soil (desorption experiment), carry out degradation experiments with model pollutants, and determine accumulation of selected pollutant classes (PAHs and trace elements). 3. Final ranking of organic stormwater pollutants. 3 INITIAL RANKING OF ORGANIC MICROPOLLUTANTS BASED ON DETECTION FREQUENCY IN SOURCE STORMWATER Thirty-eight urban stormwater sources were sampled and analysed in Deliverable 1.3 - Inventory of pollutants in urban storm water across Europe. In Table 1 and Table 2, we have summarized the results for organic micropollutants detected in source stormwater at ≥25% of the sampled stations. The pollutants were ranked according to their detection frequency. By setting a detection frequency threshold at 25%, we evaluated only pollutants with a high probability of presence in urban stormwater and avoided compounds that derive solely from wastewater stations (WWTP bypass, sewer overflow, etc.) that comprised 5 of the 38 sampled stations plus several stations where surface runoff seems to be unintentionally mixed with sewage. 16 Figure 6. Rainbed and drained area at Kallerupvej. Figure 7. Drainage area of the Kallerupvej rainbed 17 Figure 8. Schematic cross-section of the Kallerupvej rainbed. 4.3 Car park, Santander (Spain). Coordinates: 43.473209, -3.797802. Station in Deliverable 1.3 stormwater inventory: Sa02 Parking bay. The sampling was done in three identical parking bays, The parking bays were constructed with interlocking concrete blocks as surface layer, geotextile, a base layer of 4-8 mm of limestone chips, and in the bottom at 50 cm depth a waterproof liner and a collector pipe to collect infiltrated stormwater for analyses. Figure 9. car park sampling location in Santander. 18 Figure 10. Permeable parking bays. Figure 11. Schematic cross-section and drainage area of the permeable parking bay. 19 4.4 Centro Pontedera detention basin, Pontedera (Italy). Coordinates: 43.654270, 10.626498. Station in Deliverable 1.3 stormwater inventory: Po05. Road runoff drains to a detention basin via ditches that run along both sides of the road. The basin has two separate subbasins. Water enters the main basin from the ditches via two drainpipes. The main basin drains to a smaller subbasin that drains to sewers when full. The nearby parking lots and mall roof drain directly to the sewers and not to the basin. The basin is constructed as a simple depression. The soil is heavy clay. Figure 12. Centro Pontedera sampling location in Pontedera. Figure 13. Centro Pontedera detention basin. 20 Figure 14. Ditch along road that drains to the Centro Pontedera detention basin. Figure 15. Drainage area of the Centro Pontedera detention basin. * 21 4.5 Søparken Syd retention basin, Odense (Denmark) Coordinates: 55.350630, 10.3630977. Station in Deliverable 1.3 stormwater inventory: Od03. Constructed retention basin receiving runoff from a residential area. Water is mostly evaporated from the basin but runoff from basin to Odense River happens during heavy rain events. The basin was probably constructed during the 70’ties and has a PE membrane, geotextile and a clay liner. Figure 16. Søparken Syd sampling location in Odense. Figure 17. Søparken Syd retention basin. 22 Figure 18. Drainage area of Søparken Syd retention basin. Figure 19. Søparken Syd retention basin, stormwater outlet to the basin. *South inlet 23 5 FATE-STUDIES – DEGRADATION OF ORGANIC MICROPOLLUTANTS 5.1 Model pollutants Model pollutants were selected from the draft inventory produced in Deliverable 1.3. The draft inventory had data on the composition of source stormwater from only 16 stations. We chose nine compounds that were found at a high fraction of the stations and originated from the city environment. We deliberately excluded pharmaceuticals that mostly originate from stations contaminated by wastewater. The model compounds encompassed triazines, compounds associated with vulcanization and the wear of car tires, and corrosion inhibitors. Table 3 lists the occurrence of the model compounds based on all 38 inventory stations. Some of the compounds were detected with two different platforms and methods (RP-LCHRMS and HILIC-HRMS) that have different detection levels in the inventory which may partly be due to different limits of detection of the two methods. The methods furthermore differ with respect to sample pre-treatment where a major difference is that the stormwater in the RP-LC-HRMS was filtered before analysis, whereas the storm water for HILIC-HRMS was evaporated by lyophilization. This means that only the soluble fraction was detected by RP-LC-HRMS, whereas HILIC-HRMS detected the combined solubleand particle-bound fractions. Table 3. Model compounds selected for fate studies. Compound Trivial name CAS RN (detection frequency in inventory) StructureE Origin Melamine 108-78-1 (34%A, 66%C) Degradation product from the car tire compound hexa(methoxymethyl)melamine. Many other probable sources in the urban environment. Terbutryn 886-50-0 (63% A) Biocide, preservative added to mortar plaster and paint forl film protection. Hexa(methoxymethyl)melamine (HMMM) 3089-11-0 (74% A) Rubber cross-linker. Released from car tire particles (Foscari et al., 2024; Seiwert et al, 2020). 24 6PPD-quinone 2754428-18-5 (74% A) Rubber, 6PPD is an antidegradant in rubbers. 6PPD-Q is a degradation product from 6PPD. Released from car tire particles (Foscari et al., 2024). 1,3-diphenylguanidine 102-06-7 (76% A,95%C) Rubber, vulcanization accelerator. From car tire particles (Foscari et al., 2024; Johannessen et al., 2022; Seiwert et al, 2020). Benzothiazole-2-sulfonic acid 941-57-1 (92% A,87%C) Released from car tire particles (Foscari et al., 2024; Seiwert et al, 2020). Probable oxidation product from the rubber vulcanization accelerator 2mercaptobenzothiazole. 5-methyl-1H-benzotriazole 136-85-6 (63% A,53%C) Released from car tire particles (Johannessen et al., 2022). Also corrosion inhibitor. 2-aminobenzothiazole 136-95-8 (32%B, 92%C) Released from car tire particles (Foscari et al., 2024; Johannessen et al., 2022; Seiwert et al., 2020). Also corrosion inhibitor? 2-hydroxybenzothiazole (Benzothiazolone) 934-34-9 Released from car tire particles (Foscari et al., 2024; Johannessen et al., 2022; Seiwert et al, 2020). Also corrosion inhibitor? A RP-LC-HRMS target analysis. B RP-LC-HRMS suspect screening. C HILIC-LC-HRMS level 1 (with standard). D HILIC-LC-HRMS level 2 suspect screening. E Molecular structure from PubChem. 5.2 Degradation potential The degradation potential for the nine model pollutants was tested by spiking surface samples with a mixture of the pollutants in semi-realistic concentrations and the decrease in concentrations was followed over time. The soils were composite samples as follows: At Hørdumsgade, soil was collected from the upper 0-10 cm as 15 subsamples distributed throughout the rain bed. At Kallerupvej, soil was collected from the upper 0-10 cm as 15 subsamples distributed in the first half of the rain bed. Søparken sediment was collected with a 1-mm sieve as approximately 0-10 cm upper sediment as five subsamples 2-10 meters from the southen stormwater source. The sediment was drained in the sieve but was still very liquid. At Centro Pontedera, topsoil was collected from the upper 0-5 cm as 20 subsamples at 0-6 m from the inlet to detention basin, the soil was clayey and liquid. At Santander car park, gravel was collected from 3 permeable parking bays. Geotextile was 25 collected from the same 3 permeable parking bays but only geotextile from the parking bay with the most fine-grained material was used as the two other contained mostly geotextile and coarse gravel. The soil and sediment samples were homogenized in their plastic bags and sieved through 4 mm. The water content was determined gravimetrically after 24 hours of drying at 105 °C. 3 g of soil (wet weight, Hørdumsgade and Kallerupvej) or 2.5 ml of sediment slurry (Søparken and Centro Pontedera) was added to 12-ml centrifuge tubes. For the parking bay samples, 3x200 g of gravel (200 g from each parking lot, stones > 6 mm sorted out) was shaken with 100 ml of MilliQ water, and 2.5 ml of this slurry was added per tube. Geotextile pieces were cut out of the geotextile with the highest content of fine particles, and 3 g (wet weight) were added per tube. An aqueous spike solution with a mixture of the nine model pollutants (each 0.25 µg/mL) was prepared in Milli-Q water from concentrated stocks (1000 µg/mL in methanol). 0.5 ml of aqueous spike solution was added to all tubes and mixed into the soil/sediment with a small sterile spatula. The final concentration was approximately 50 ug/kg in the tubes. Controls were prepared with methanol instead of the concentrated stocks. The tubes were incubated aerobically in a water-saturated atmosphere in closed plastic boxes at 20 °C. At increasing time intervals, a set of tubes was closed with caps with PTFE liners and frozen in a tilted position until analysis. The model pollutants were semi-quantified by LC-MS and the decrease in concentration was expressed as C/C0. Simple first-order kinetics were fitted to the degradation curves and half-lives calculated, some examples are shown below. Many curves followed first-order kinetics, but often with some underestimation of initial degradation and overestimation of later degradation. The degradation was in some cases very slow and appeared almost linear (e.g. Figure 20 terbutryn). Some stations showed a high background concentration of some pollutants in the control samples (e.g. Figure 20, 1,3-benzothiazole-2-sulfonic acid) and a bi-phasic degradation where the added pollutant degraded much faster than the same pollutant in the original soil. This may for instance be caused by the pollutants being present in car tire fragments with low bioavailability in the soil or caused by aging phenomena. When this was seen, we only used the initial part of the curve for curve fitting, which to some degree underestimated degradation of the added fraction. It may be argued that we could instead have fitted a double-exponential-decay-in-parallel function, but many of these plots were more noisy than shown in Figure 20 and double-decay could not be reliably estimated. In some cases, the background concentrations were high and the data so noisy that we could not fit a curve, this was often the case for melamine (e.g. Figure 20, melamine) and to a lesser extent for 1,3-benzothiazole-2-sulfonic acid. The model pollutants were degraded fastest in one or both parking bay samples. This shows that the microbial communities were highly adapted for degrading a range of urban pollutants at this station, probably because of high in-situ pre-exposure. This was also seen in the MPN counts of specific degraders where especially the car park geotextile had counts for most compounds. The rain beds drain in less than a day, suggesting that the retention time in the active layer is only a few days at best for mobile pollutants. The rather long half-lives show that all model pollutants may leach from the rain beds to less bioactive subsurface layers, if the pollutants are mobile. The same applies to the permeable pavement in the parking bays where 32 Centro Pontedera. Three suction cups were installed on March 19th, 2024, one, four and seven meters from the ditch outlet pipes. The suction cup tips were 90 cm below the surface. The suction cups were sampled only one time. The soil was heavy clay with very little infiltration and very little water was collected during the 48 hours evacuation of the flasks. Water from one suction cup was slightly contaminated with sediment and was discarded. Water from the the two other suction cups was pooled to give a total of only 70 mL that was analysed only by RP-LC-HRMS. Figure 26. Suction cups installed in Centro Pontedera detention basin. Car park. Water was collected from three car park bays on April 23rd, 2024. The parking bays were constructed with the same materials. The water (500 mL) was collected from shallow drainpipes. The parking bays were excavated after collection of water. Figure 27. Collector pipe (drainpipe) in Santander permeable parking bay. 33 6.3 Results: desorption and leaching The results from topsoil desorption in the laboratory and suction cups placed subsurface in situ at the nature-based stations (leaching test) are presented in the two tables below. In the desorption study, we used both geotextile and gravel from the permeable parking bays. In the tables, these results have been combined (highest observed concentration) so that each station is represented equally. 5-methyl-1H-benzotriazole was the only compound detected in the gravel desorption water but not in the geotextile desorption water. RP-LC-HRMS was prioritized in the desorption study so that all water was used for this analysis to achieve low detection limits. There is consequently no HILIC-HRMS data from this experiment. The suction cup data consists of four samples from each of the Hørdumsgade, Kallerupvej and Carpark stations plus one combined sample from the Centro Pontedera station. There is no suction cup data from the Søparken station as it turned out, rather surprisingly, that the basin was constructed with an impermeable liner. Table 6. Target analyses (RP-LC-HRMS) of water from the desorption study and from suction cups installed in-situ. The median concentration is the median of the stations where the compound was detected. Desorption Leaching (suction cups) Compound Origin Station s, total Stations, detection Median conc. (ppb) Samples, total Samples, detection Median conc. (ppb) 1,3-Diphenylguanidine Industrial 5 3 0.228 12 12 0.025 HMMM Industrial 5 4 0.099 12 10 0.702 Benzothiazole-2-sulfonic acid Industrial 5 0 -- 12 10 0.605 Melamine Industrial 5 1 0.831 12 9 0.467 FBSA PFAS 5 3 0.001 12 9 0.001 5-methyl-1H-benzotriazole Industrial 5 4 0.018 12 7 0.019 DEET Pesticide 5 0 -- 12 6 0.006 PFBS PFAS 5 2 0.004 12 6 0.001 Dibenzylamine Industrial 5 0 -- 12 5 0.015 Terbuthylazine-desethyl-2hydroxy Pesticide 5 0 -- 12 5 0.133 Terbuthylazine-2-hydroxy Pesticide 5 2 0.005 12 4 0.002 4-Nitrophenol Industrial 5 0 -- 12 4 0.038 1H-benzotriazole Industrial 5 5 0.029 12 3 0.547 Caffeine Natural 5 0 -- 12 3 0.015 Tris(2-butoxyethyl) phosphate Industrial 5 1 0.007 12 3 0.005 34 PFOS PFAS 5 2 0.008 12 3 0.002 6PPD-Quinone Industrial 5 3 0.007 12 2 0.009 Lidocaine Pharmaceutical 5 3 0.001 12 2 0.001 Metformin Pharmaceutical 5 2 0.005 12 2 0.003 DNOC Pesticide 5 0 -- 12 2 0.005 Salicylic acid Natural 5 0 -- 12 2 0.695 6-(tert-Butylamino)-1,3,5triazine-2,4(1H,3H)-dione Pesticide/biocide 5 1 0.006 12 1 0.045 Benzyldimethylamine Industrial 5 1 0.143 12 1 0.018 Carbamazepine Pharmaceutical 5 0 -- 12 1 0.002 Diuron Pesticide 5 1 0.008 12 1 0.015 Isoproturon Pesticide/biocide 5 1 0.002 12 1 0.002 Propiconazole Pesticide/biocide 5 1 0.007 12 1 0.014 Tebuconazole Pesticide/biocide 5 1 0.006 12 1 0.006 Terbuthylazine metabolite SYN 545666 Pesticide/biocide 5 0 -- 12 1 0.008 Terbutryn Biocide 5 1 0.007 12 1 0.015 Terbutryn-desethyl Biocide 5 1 0.003 12 1 0.011 2-Mercaptobenzothiazole Industrial 5 1 0.018 12 1 0.013 Bis(2-ethylhexyl) phosphate Industrial 5 0 -- 12 1 0.227 Bisphenol S Industrial 5 0 -- 12 1 0.038 PFBA PFAS 5 2 0.003 12 1 0.013 PFHpA PFAS 5 2 0.008 12 1 0.007 PFHxA PFAS 5 2 0.013 12 1 0.015 PFHxS PFAS 5 0 -- 12 1 0.001 PFOA PFAS 5 4 0.011 12 0 -- Indole-3-acetic acid Natural 5 3 0.818 12 0 -- PFPeA PFAS 5 2 0.069 12 0 -- Metoprolol Pharmaceutical 5 1 0.003 12 0 -- O-Desmethylvenlafaxine Pharmaceutical 5 1 0.104 12 0 -- Prosulfocarb Pesticide 5 1 0.014 12 0 -- 4-Hydroxybenzoic acid Natural/Industrial 5 1 3.265 12 0 -- PFDA PFAS 5 1 0.009 12 0 -- PFNA PFAS 5 1 0.007 12 0 -- 35 Table 7. Suspect screening (RP-LC-HRMS) of water from the desorption study and from suction cups jnstalled in-situ. Desorption Leaching (suction cups) Compound (ID points) Origin Stations, total Stations, detection Samples, total Samples, detection Dicyclohexylurea (60) Industrial Pharmaceutical? 5 0 12 11 8-Hydroxyquinoline (25) Industrial? Pharmaceutical? 5 4 12 9 N,N-Dimethyldecylamine oxide (50) Biocide, Industrial? 5 1 12 7 Dinitrocresol (60) ? 5 0 12 5 4-Methylbenzotriazole (60) Industrial 5 3 12 3 Diethyl-phthalate (60) Industrial 5 0 12 3 Tramadol (45) Pharmaceutical 5 1 12 2 2-Aminobenzothiazole (45) Industrial 5 0 12 2 Atrazine-desethyl (15) Pesticide 5 0 12 2 Triethylene glycol bis(2ethylhexanoate) (60) Industrial 5 0 12 1 Icaridin (60) Pesticide 5 0 12 1 Phloretin (50) Natural, Personal care 5 0 12 1 Isoxaben (20) Pesticide 5 0 12 1 Phlorizin(40) Natural, food, (personal care products?) 5 0 12 1 Triethylcitrate (60) Food additive, industrial 5 2 12 0 Aminocarb (100) Pesticide 5 1 12 0 Tetraglyme (60) Industrial 5 1 12 0 Simeconazole (25) Pesticide 5 1 12 0 Bis(2-ethylhexyl) phthalate (20) Industrial 5 1 12 0 Methyl-1-naphthaleneacetate (20) ? 5 1 12 0 7 FATE-STUDIES – RANKING OF POLLUTANTS ACCUMULATED IN THE NBS Trace elements and polycyclic aromatic hydrocarbons (PAHs) are two contaminant classes that generally show a high risk of accumulating in soil. We furthermore had well-established methods for quantification of these pollutants in soil so that we could determine 36 concentrations and enrichment of trace elements and PAHs at the NBS stations Hørdumsgade, Kallerupvej and Centro Pontedera. Here, we could determine accumulation of pollutants and their enrichment factors by comparing the concentrations in the topsoil and the same soil type from deeper layers. This was not possible at Søparken and the Santander parking bays as these stations had thin layers of sediment and gravel. All samples were composite samples composed of 10-15 subsamples. At Hørdumsgade, the filter soil in the rain bed had a thickness of 80 cm. Samples were taken 1.0; 1.5 and 2.0 meters from the stormwater inlet at the center of the rain bed. The bottom samples were taken by digging to 60 cm below terrain followed by subsampling of the 60-70 cm layer to avoid contamination from the original soil below. At Kallerupvej, the filter soil in the rain bed had a thickness of only 40 cm. Samples were taken 1.0; 1.5 and 2.0 meters from the stormwater inlet at the center of the rainbed. The bottom samples were taken by digging to 40 cm below terrain until reaching the bottom of the filter soil. The bottom 3 cm of the filter soil was sampled horizontally around the bottom of the hole. At Centro Pontedera, samples of the original soil (top and bottom) were taken 1,5 meters downstream of each suction cup. The bottom samples were taken by digging to 60 cm below terrain followed by subsampling of the 60-70 cm layer. The samples were transferred to clean red-cap bottles with PTFEliners in the lids and frozen the same day as sampling (Hørdumsgade and Kallerupvej) or immediately cooled to 5°C and frozen the day after sampling (Centro Pontedera). Trace elements. We determined the concentration of acid-extractable trace elements. In the lab subsamples of 3-6 g were extracted with 25 mL 3.5% nitric acid, shaken for 8 hours, then diluted 4 times and analyzed on ICP-MS with internal calibration. Most samples were reanalyzed at further 100 times dilution because of instrument overload. The trace element concentrations are shown in Annex A. Enrichment factors were calculated as the top-tobottom concentration ratio. Enrichment factors were quite variable between different holes at the same NBS, which is not surprising given that the soil profiles appeared heterogeneous. There were also considerable differences between stations due to different soils, different exposures, etc. The mean enrichment factor across all samples, however, seems robust (Table 8). None of the trace elements were depleted in the topsoil, i.e., none of the mean enrichment factors were <1.0. When ranking the trace elements according to mean enrichment factor, especially Sn is a metal that we should expect to accumulate in topsoil of NBS handling surface runoff, and to lesser extent also Zn, B, Pb, Cu, Cr, and V. Table 8. Enrichment factors for trace elements found in concentrations >LOQ in both the top layer and the bottom layer. Enrichment factors were calculated as the top-to-bottom concentration ratio for three replicate holes at each station. Trace elements are sorted according to decreasing mean enrichment factor. Element Enrichment factor Hørdum Enrichment factor Kallerup Enrichment factor Centro Pontedera Enrichment factor Mean Enrichment factor Standard deviation Sn 40.6 18.5 <LOQ 29.5 15.6 Zn 2.8 8.7 1.8 4.4 3.7 B 2.1 2.0 3.5 2.5 0.8 Pb 0.9 5.1 1.4 2.5 2.3 Cu 1.7 3.7 1.3 2.2 1.3 Cr 1.0 3.3 2.0 2.1 1.2 V 0.8 2.6 2.6 2.0 1.0 37 Al 0.5 3.2 0.9 1.5 1.5 Mo 0.5 1.8 2.0 1.5 0.8 Se 0.6 2.2 1.1 1.3 0.8 Ni 0.8 2.1 0.8 1.2 0.7 Co 0.6 2.0 0.9 1.2 0.7 Ba 0.8 1.7 1.1 1.2 0.5 Mn 0.6 1.6 1.0 1.1 0.5 Cd 0.7 1.8 0.7 1.1 0.6 Fe 0.8 1.2 1.1 1.0 0.2 As 0.2 1.6 1.2 1.0 0.7 PAHs. Forty PAHs were quantified by GC-MS for pooled soil samples from the two depths at the three NBS stations. The PAH concentrations are shown in Annex A. Enrichment factors are shown in Table 9 for PAHs found in concentrations >LOQ in both the top layer and the bottom layer. Fluorene, anthracene, perylene and dibenzo[a,h]anthracene were also found in the top layer at Kallerupvej and Hørdumsgade, but the concentrations were below the LOQ in the bottom layer, we therefore could not calculate enrichment factors for these PAHs though they had clearly accumulated in the top soil. It was rather surprising that twoand three-ring PAHs showed high accumulation factors as the light PAHs are more biodegradable and more mobile than the heavier PAHs. This probably reflects the high numbers of microbial degraders of twoand three-ring PAHs ( Table 5) at these stations so that much of the light PAHs may be degraded before reaching the bottom layer. The results also reflect that the three NBS stations have been built rather recently, and therefore have had a short time to accumulate pollutants in the topsoil. Higher accumulation factors would probably be found where NBS have been in operation for longer time, and presumably, the heavier PAHs would accumulate more than the light, biodegradable PAHs. Table 9. Enrichment factors for PAHs found in concentrations >LOQ in both the top layer and the bottom layer. Enrichment factors were calculated as the top-to-bottom concentration ratio. PAHs are sorted according to decreasing mean enrichment factor. PAH Kallerupvej Enrichment factor Hørdumsgade Enrichment factor Centro Pontedera Enrichment factor Mean Enrichment factor Acenaphthylene 28.9 7.9 -- 18.4 Naphthalene 17.2 17.1 7.3 13.9 Phenanthrene 15.4 10.2 13.0 12.9 Benzo[a]pyrene 16.3 4.7 -- 10.5 Benzo[g,h,i]perylene 9.8 -- -- 9.8 Chrysene 11.3 8.2 -- 9.8 Indeno[1,2,3-c,d]pyrene 14.0 5.2 -- 9.6 Pyrene 11.9 7.2 -- 9.5 Benzo[e]pyrene 10.9 7.4 -- 9.1 Benzo[k]fluoranthene 11.3 6.7 -- 9.0 Fluoranthene 9.6 6.2 -- 7.9 Benzo[b]fluoranthene 7.3 5.2 -- 6.2 Benzo[a]anthracene -- 4.4 -- 4.4 2-Methylnaphthalene 2.7 3.9 1.6 2.7 38 8 RANKING OF POLLUTANTS In this section, we intend to rank stormwater pollutants according to their risk of leaching from NBS towards the groundwater and/or potential for accumulation in NBS topsoil. The risk of leaching and accumulation can be ranked in many ways, but in our case, we were limited by the very limited information available for many of the compounds. Ranking was therefore based on the detection of compounds in the source stormwater inventory, presence in infiltrated water at the five model NBS stations, degradation in the laboratory experiment, results found in the literature, and estimated sorption potential (logD) when no other data was available. 8.1 Method The criteria for ranking of pollutants were applied in prioritized order (Figure 28). We first selected source stormwater pollutants that were present at ≥25% of the stormwater stations in the source stormwater inventory (Table 1). By setting the threshold at 25%, we chose only pollutants with a high probability of presence in urban stormwater and at the same time avoided compounds that derive solely from wastewater stations. Three compounds were judged environmentally non-problematic and removed from the list, these are 3-methylindole (skatole) that originates from feces, triethylcitrate that is a food additive (E1505), and indole3-acetic acid that is a plant hormone present in all plants. The resulting list of source stormwater pollutants was combined with leaching data from the four sampled NBS stations (suction cup data). The suction cup samples represent leaching from traffic-impacted NBS fate-stations. The pollutants were then sorted according to their presence in suction cup samples for pollutants present in ≥3 of the analysed samples. The higher the detection frequency in suction cups, the higher the risk of leaching. Three positive samples was chosen as the limit because compounds detected in at least three samples are either detected repeatedly at one station or detected at two or more stations. RP-LC-HRMS was prioritized for suction cup samples with low volumes, so HILIC-HRMS was only applied to suction cup samples from Kallerupvej and Hørdumsgade. HILIC-HRMS data is not reported as two stations is too few for any general interpretations, but the HILIC data was used for evaluating a few specific compounds (QUATs/BACs). The remaining compounds were sorted according to their logD. LogD is the distribution coefficient between an organic phase and a water phase at a specified pH. Basically, we used logD as a proxy for sorption to soil organic carbon (logKoc). The higher the logD, the lower the risk of leaching and the higher the risk of accumulation. We could not find experimental logD data for any of the compounds, logD was therefore estimated in silico for each compound at pH 7.0 by using Chemicalize (chemicalize.com). The sorption of quaternary ammonium compounds (QUATs/BACs) is poorly described by their logD since they have a positive charge and thus bind to negatively charged soil particles, e.g. clay minerals, as well as to the soil organic matter. All QUATs/BACs were therefore assigned a very low probability of leaching, and a corresponding, high potential for accumulation. 39 Figure 28. Criteria used for ranking of pollutants. 8.2 Results and discussion The ranking is presented in Table 10 where compounds are sorted according to decreasing leaching potential so that compounds at the top are estimated to have the highest leaching potential and compounds at the bottom the smallest leaching potential. The risk of accumulation is the opposite but estimated with less certainty than the leaching potential. Ranking of compounds marked in red has the highest certainty as the ranking is based on their presence in water leached from the model NBS stations. The ranking therefore integrates the effects of both in-vivo sorption and degradation with respect to leaching potential. Basically, pollutants leached to suction cups were present in the source stormwater at that NBS-station, escaped degradation, and sorption was insufficient for retaining the pollutant. Five of the top pollutant candidates (marked in bold) also have degradation data from the in vitro fate studies (Table 4 and Table 5), which confirm slow degradation compared to expected residence times. The potential for accumulation is less well described for many ”red” compounds as degradation potential has only been assessed for five compounds. Also, absence in suction cup samples cannot be used to assign low leaching/high accumulation potential as the pollutant may not be in the source stormwater at those specific stations and rain events, or the pollutant may be degraded in the topsoil. Present at ≥25% of source stormwater stations in inventory? Compound not relevant, not ranked Identified with ID score ≥50 (RP-LC) or ≥3 (HILIC)? Found in ≥3 suction cup samples? Compound ID too uncertain, not ranked Binds to soil minerals (BACs and QUATs)? Yes Yes Yes No No Yes No No High leaching risk. Probably low accumulation risk. Ranked according to suction cup detection No fate data. Ranked according to logD. Ranking uncertain. Low leaching risk. Probably high accumulation risk. Criterion Ranking 40 Ranking of “black” compounds is very uncertain as it is based solely on predicted logD, i.e., the tendency to sorb to soil organic matter. First, it should be noted that logD is pHdependent for many compounds as lipophilicity/hydrophilicity changes for weak acids and bases depending on pH. At a given pH, these pollutants may be either ionized, partially ionized, or unionized. The pH of soils in NBS varies depending on source rock and the degree of illuviation. We chose a pH of 7.0 as a compromise between illuviated sandy topsoil of low pH and arid or calcareous soils of high pH. Second, degradation is a key parameter as both leaching and accumulation can be reduced by degradation. Degradation in NBS soil was tested for three of the “black” compounds and was slow (Table 4), but other compounds may have faster degradation and thus less tendency to accumulate and leach than indicated by their logD. Compounds embedded in particles are especially difficult to rank from logD. This is because the particles may accumulate in topsoil even when the compounds’ logD is moderate, which applies to rubber compounds present in car tire fragments and probably also to compounds present in façade paint and possibly also façade plaster. The sorption of quaternary ammonium compounds (QUATs/BACs) is poorly described by their logD, since they have a positive charge and thus bind to negatively charged soil minerals as well as to the soil organic matter. This has been quantified by Kahn et al. (2017) who conclude: “Due to their positive charge, BACs are expected to strongly adsorb on negatively charged soil particles. Furthermore, the long hydrophobic carbon chains facilitate their adsorption on soil organic matter…….Column tests conducted using soils amended with lime, stabilised biosolids and artificial rain water at a flow rate of 0.2 mL min−1 indicate very low leaching of BACs.” All QUATs/BACs were therefore assigned a very low probability of leaching which is in accordance with our results showing BAC C12 (benzododecinium chloride) and BAC C14 (myristalkoniumchloride) in the source stormwater at Kallerupvej and Hørdumsgade, but not in the suction cup samples from these stations when screened with HILIC-HRMS (data not shown). The low leaching risk of the “blue” QUATs/BACs therefore has high certainty, but the corresponding high accumulation estimate is less certain as degradation will likely take place in the topsoil. Table 10 Ranking of stormwater pollutants according to the risks of leaching from NBS (decreasing from top to bottom) and accumulation in NBS topsoil (increasing risk from top to bottom). A total of 12 suction cup samples represents leaching from four traffic impacted NBS stations. Red: compounds with high risk of leaching based on detections in NBS suction cups. Black: compounds ranked according to logD at pH 7,0, the ranking of these compounds is uncertain. Blue: QUATs/BACs with very low leaching risk. Pollutants used for degradation experiments and/or MPN-counting of specific degrader microorganisms are highlighted in bold. Compound LogD (pH7) Confidence level RP-LC-HRMS Inventory % HILIC -HRMS Inventory % Detection, suction cup samplesn Suction cup max. conc. (ppb) 1,3-Diphenylguanidine Target RP-LC, 1 HILIC 76 95 12 0.025 Dicyclohexylurea 60 RP-LC 89 11 Benzothiazole-2-sulfonic acid Target, 2 HILIC 92 87 10 0.605 41 Hexamethoxymethylmelamine (HMMM) Target RP-LC 74 10 0.702 Melamine Target RP-LC, 1 HILIC 34 66 9 0.467 5-Methyl-1H-benzotriazole Target, 1 HILIC 63 53 7 0.019 N,N-Dimethyldecylamine oxide 60 RP-LC 2 HILIC 42 74 7 Diethyltoluamide (DEET) Target, RP-LC 89 6 0.006 Dibenzylamine Target RP-LC 87 5 0.015 Dinitrocresol 60 RP-LC 95 5 Terbuthylazine-desethyl-2-hydroxy Target RP-LC 42 5 0.133 4-Nitrophenol Target RP-LC 53 4 0.038 Terbuthylazine-2-hydroxy Target RP-LC 1 HILIC 50 87 4 0.002 1H-benzotriazole Target RP-LC 55 3 0.547 4-Methylbenzotriazole 60 RP-LC 42 3 Caffeine Target RP-LC 66 3 0.015 Tris(2-butoxyethyl) phosphate Target RP-LC, 3 HILIC 68 92 3 0.005 Metformin -5,69 Target RP-LC 53 2 0.003 Guanylurea -3,95 Target RP-LC 32 <LOQ Propamocarb -1,50 3 HILIC 66 Acesulfame -1,49 Target RP-LC 34 <LOQ Salicylic acid -1,47 Target RP-LC 45 2 0.695 Tetraethylene glycol -1,35 3 HILIC 61 4-Hydroxybenzoic acid -1,24 Target RP-LC 71 0 <LOQ Theophyline -0,82 60 RP-LC 26 <LOD Metoprolol -0,47 3 HILIC 74 Nicotine -0,42 3 HILIC 87 2,4-Dinitrophenol -0,30 60 RP-LC 63 <LOD Tetraglyme -0,06 60 RP-LC 37 0 <LOD Benzyldimethylamine 0,02 Target RP-LC 29 1 0.018 Tramadol 0,10 45 RP-LC, 1 HILIC 32 66 2 <LOD Cotinine (nicotine metabolite) 0,21 50 RP-LC, 3 HILIC 47 95 <LOD 4,6-Dinitro-o-cresol (DNOC) 0,34 Target RP-LC 92 2 0.005 O-Desmethylvenlafaxine 0,59 Target RP-LC, 3 HILIC 47 63 0 <LOQ Dextromethorphan 0,73 3 HILIC 47 Paracetamol 0,90 Target RP-LC 47 <LOQ 48 1.3.6-Trimethylchrysene 1.086 <DL <DL <DL <DL <DL <DL