Influence of rubber particle inputs on nitrogen removal efficiency of bioretention systems

Urban stormwater runoff is a significant route for MPs to reach terrestrial and aquatic environments (Horton et al. 2017). Previous studies have found that tire particles are the most prevalent MPs in urban surface runoff, accounting for 42–62% of total MPs (Siegfried et al. 2017; Werbowski et al. 2021). This indicates that rainwater runoff should be treated before being released into nearby aquatic environments to lessen the ecotoxicological effects of MPs and contaminants carried by their diffusion in environmental media. Bioretention systems are an effective method of controlling and utilizing urban rainwater that can efficiently reduce peak runoff discharge and surface source pollution (Osman et al. 2019). Studies have shown that MPs can be physically retained in bioretention systems and removed at a rate of 84–100% (Smyth et al. 2021; Lange et al. 2022). Lange et al. (2022) discovered that MPs are primarily trapped in the upper layer of bioretention systems, with polypropylene, ethylene-vinyl acetate, polystyrene, and ethylene-propylene-diene-monomer (EPDM) rubber being the most common forms detected. However, it is unknown whether the MPs retained in bioretention systems affect their performance.

Once MPs are introduced into the natural environment, they can have direct or indirect impacts on ecosystem function (Wazne et al. 2023). After entering the soil, MPs eventually break down soil pore structure and soil aggregates, altering its chemical and physical characteristics, including its ability to retain water and other hydraulic qualities (Wang et al. 2022b; Sepehrnia et al. 2024). Plant root activity as well as water and nutrient transport can both be impacted by changes to soil's physical and chemical characteristics (Rillig et al. 2019). Plant growth is further impacted by MPs because they alter the synthesis of photosynthetic pigments and the activity of antioxidant enzymes (Rillig et al. 2019). MPs can also have an impact on the structure and richness of natural communities. MPs in marine environments can have a severe impact on the growth of native microzooplankton, with one experiment finding the biomass of the highest concentration MP treatment group being reduced by 96.59% compared to the control group after 72 h of exposure (Geng et al. 2021). In soil conditions, polylactic acid MPs increased and decreased the complexity of bacterial and fungal community structures, respectively (Liu et al. 2023). In addition, MPs adsorb a variety of environmental contaminants, including heavy metals and organic compounds; as a result, they pose a risk to the environment both as pollutants and through their interaction with other pollutants, causing combined toxicity (Li et al. 2018). From this, it would be expected that MPs would harm the fillers, plants, and microorganisms in bioretention systems, disrupting the initial material cycle and reducing pollutant removal.

Nitrogen is a contributing element to water eutrophication and a primary pollutant in stormwater runoff from bioretention systems (Li & Davis 2014; Osman et al. 2019; Shi et al. 2023; Zhou et al. 2023). Bioretention systems remove nitrogen from stormwater runoff primarily by filler adsorption, plant absorption, microbial mineralization, nitrification, and denitrification (Li & Davis 2014). However, it is unknown how MP deposition affects nitrogen removal in bioretention systems. Given the biotoxicity of MPs and the presence of MPs and nitrogen in stormwater runoff, it is vital to investigate the nitrogen migration and transformation process in bioretention systems when exposed to MP accumulation.

This study focused on the impact of EPDM input on the nitrogen removal performance of bioretention systems, determining the influence of varying EPDM concentrations on nitrogen removal efficiency (0, 50, 100, and 400 mg/L). In addition, the impact of EPDM input on nitrogen removal efficacy under various rainfall scenarios and its effect on the microbial community at different concentrations was determined. The influence of rubber MPs on the nitrogen cycle in the bioretention system was investigated by analyzing the nitrogen effluent concentration and microbial community composition under different EPDM concentrations, aiming to determine the influencing mechanism of long-term MP accumulation in a bioretention system for stormwater runoff.

The bioretention column used in this experiment was a high-density polyethylene cylindrical box (660 mm high, 100 mm inner diameter). The effluent sampling point of the bioretention system was set at a column height of 3 cm, and sampling points were set at 20, 30, 40, and 50 cm to facilitate the collection of filler samples. The system was composed of planting (6 cm), filler (50 cm), and gravel layers (10 cm) from top to bottom (Figure 1). The filler layer was a fully mixed cinder (1–3 mm) containing loess (<2 mm) (v/v = 5:5) (Xu et al. 2024). The planting layer was Ophiopogon japonicus, which is prevalent in northwest China. The gravel layer was composed of fine sand (0.2–0.4 cm), medium gravel (0.4–0.8 cm), and coarse gravel (0.8–1.2 cm) (v/v/v = 3:3:4). After the bioretention system was constructed, it was domesticated for 1 month, during which simulated rainwater was pumped into the system. Indicators (ammonia–nitrogen (⁠⁠), nitrite–nitrogen (⁠⁠), nitrate–nitrogen (⁠⁠), and total nitrogen (TN)) were regularly monitored to determine system stability. After the system stabilized, rubber particle injection experiments were carried out.

To determine the effects of interference by different concentrations of rubber particles, four groups of bioretention systems containing 0, 50, 100, and 400 mg/L (corresponding to E0, E50, E100, and E400, respectively) EPDM were set up for short- and long-duration rainfall simulation tests (Rodland et al. 2022). The highest concentration of EPDM (400 mg/L) was chosen to mimic the effects of long-term accumulation of rubber particles in the natural environment. Research has indicated that the majority of MPs in runoff rainfall are concentrated in the 20–100 μm range, significantly exceeding other particle sizes (>100 μm) (Jarlskog et al. 2020). For this experiment, EPDM particles with a size of 75–100 μm were used.

Two distinct rainfall situations were used. The first rainfall condition was primarily designed to investigate the influence of EPDM on the nitrogen removal efficiency of a bioretention system under short-duration rainfall, lasting 27 days with a total of 15 rainfall events. The bioretention system endured dry periods of 2, 4, 8, and 2 days in succession, with rainfall events lasting 120 min occurring 3 days after each drought period ended. The second experiment investigated the impact of EPDM on the nitrogen removal performance of the bioretention system with long-duration rainfall. In this stage, the system was exposed to successive dry periods of 2, 4, and 8 days, followed by a 24-h rainfall event.

The experiment results could be negatively impacted by significant variations in runoff and contaminants produced by various rainstorm events. To prevent this, an artificial rainwater simulation was used; Table 1 displays the pollutant configuration of the simulated runoff (Shi et al. 2023).

Table 1

Simulated runoff pollutant configurations

Pollutant .	Concentration (mg/L) .	Medicament .
	8	NH4Cl
	8	KNO3
	2	KH2PO4
COD	200	C6H12O6

Pollutant .	Concentration (mg/L) .	Medicament .
	8	NH4Cl
	8	KNO3
	2	KH2PO4
COD	200	C6H12O6

COD, chemical oxygen demand.

View Large

In an actual rainfall setting, when rainfall time increases, rubber particles on the road surface are gradually swept clean by rainwater. Rubber particles were only added during the first 2 h of long-duration rainfall to imitate this process.

Four groups of bioretention system effluent were collected in the first stage to examine water quality. To minimize measurement errors, two 100 mL effluent samples were collected from each bioretention system after each rainfall event. In the second stage, the effluent from each bioretention system was collected at 30 min and 1, 2, 4, 6, 8, 10, 12, 16, 20, and 24 h after rainfall events. During one rainfall event, 11 effluent water samples were taken from each group of bioretention systems. All samples were collected in glass containers and kept at 4 °C until analysis. Nessler's reagent photometry, N-(1-naphthyl)-ethylenediamine photometry, ultraviolet spectrophotometry, and potassium persulfate oxidation were used to analyze the concentrations of ⁠, ⁠, ⁠, and TN in each sample, respectively, in accordance with National Environmental Protection Agency (2022).

Filler samples from the bioretention system were taken after the experiment to examine the composition of the microbial community. Bioretention system filler (2.00 ± 0.01 g) was extracted from the filler sampling ports located at 20, 30, 40, and 50 cm in height, in that order. To create a mixed sample that accurately represented the microbial dispersion inside the filler at the corresponding concentration, the four samples of varying heights were mixed uniformly (Shi et al. 2023). Samples of mixed fillers from the different bioretention systems were kept at −20 °C. The mixed soil samples were sent to a genetic sequencing company for high-throughput sequencing using the Illumina platform. The sequencing process followed basic high-throughput sequencing steps; the DNA was sequenced after DNA extraction and PCR amplification on a Miseq high-throughput sequencer. For operational taxonomic unit (OTU) cluster analysis, related DNA sequences were clustered into the same OTU for further analysis. Each OTU was compared to known species to determine the microbial species present in the soil samples.

removal in bioretention systems relates to drought duration (Figure 2(a)). The concentration of effluent reduced after 2 and 4 days of drought but increased after 8 days. During the dry period, nitrifying bacteria can convert NH₄⁺−N adsorbed by the filler into NO₂⁻N and NO₃⁻N, reducing NH₄⁺−N accumulation in the filler and restoring NH₄⁺−N adsorption (Hatt et al. 2007). However, when drought lasts longer, runoff flows quickly through the filler, reducing the hydraulic retention time (Chen et al. 2022). The short contact period between stormwater runoff and the filler reduces the adsorption of ⁠, resulting in an increase in concentration in the effluent.

Under EPDM stress, the effluent concentration increased by 1.33–1.49 mg/L (P < 0.01) compared to E0, but the effluent concentration decreased by 1.31–2.08 mg/L (P < 0.05) (Figure 2(b) and 2(c)). This suggests that the addition of EPDM facilitates NO₃⁻N removal but also leads to the accumulation of ⁠. High concentrations in the effluent are caused by the anaerobic environment created by MPs input, which inhibits the activity of nitrite-oxidizing bacteria and makes it difficult for to be further oxidized to (Li et al. 2020; Zhou et al. 2023). Denitrification is the primary mechanism in bioretention systems for removing (Li & Davis 2014). The constant input of EPDM results in a large accumulation of rubber particles adhering to the filler medium, which can be used as an organic substrate for microorganisms to consume oxygen. This may not only encourage the growth of relevant functional bacteria but may also create a microenvironment conducive to nitrogen transformation (Huang et al. 2021). At the same time, the inner surface of the MPs provides an extra anaerobic habitat, promoting the growth and activity of denitrifying bacteria (Li et al. 2020).

EPDM stress resulted in reduced average effluent concentrations compared to E0 for three rainfall events following 2, 4, and 8 days of drought. This shows that EPDM inputs may mitigate the harmful consequences of drought on bioretention systems. This is consistent with Wang et al. (2022a) who found that some MPs boost soil water retention capacity, reducing the impact of drought stress on crops. The initial rainfall event at the end of the dry period had the greatest effluent concentration while following rainfall events had a progressively decreasing influence on effluent concentration. This could be because the activity of denitrifying bacteria eventually resumed, increasing denitrification.

The average TN concentration in the simulated runoff during the test phase was 17.85 ± 0.84 mg/L (Figure 2(d)). Under EPDM stress, the TN effluent concentration decreased by 0.59–1.52 mg/L in comparison to E0. The concentration trends in effluent and TN in Figure 2(c) and 2(d) are essentially the same, likely because the primary source of TN in the effluent was ⁠. While adding EPDM somewhat increased the amount of that accumulated, it also helped remove and ⁠, as well as TN. Following 2, 4, and 8 days of drought under EPDM stress, the average TN effluent concentrations after the three rainfalls were 0.41–1.67, 0.20–1.72, and 0.57–1.51 mg/L lower than those in the E0, respectively; this further showed that EPDM input lessened the detrimental effects of drought on the bioretention system.

The effluent concentration was higher at 30 min in the long-duration rainfall phase than at the later sampling times (Figure 5). This could be because air replaces the pore water in the filler during the dry period, promoting nitrification (Hatt et al. 2007; Rahman et al. 2020). The adsorbed by the filler is continually transformed into ⁠, which is washed out in the following rainfall event, resulting in high effluent concentrations and even leaching in the first rainstorm after drought. The efficacy of elimination improved steadily as rainfall duration increased. As the biomass and activity of most microbes rise when precipitation increases, this could be the cause of the overall decrease in effluent concentration (Nielsen & Ball 2015). The recovery rate of removal in each stage following the drought period can be calculated using a standard of 4 mg/L (⁠ removal rate greater than 50%). For reducing the concentration in the effluent below this level under various drought conditions, E400 took the least amount of time and recovered the fastest (2–4 h), followed by E50 and E100 (4–5 h), while E0 recovered the slowest (6–14 h). This could be because the higher concentrations of EPDM input blocked a portion of the channels in the filler, more quickly creating an anaerobic environment which is beneficial to denitrification (Li et al. 2020; Zhou et al. 2023). As a result, the addition of EPDM improved the recovery of removal capacity following drought under long-term rainfall. Throughout long-term rainfall, as the EPDM input concentration increased, the effluent concentration of the bioretention system steadily declined. Subsequent microbiological examination revealed that EPDM application increased the percentage of denitrifying bacteria, with the largest percentage occurring at the stage of the highest EPDM infusion (400 mg/L).

The Goods value represents the probability that a new species will not be detected in a sample, with a greater value indicating a lesser possibility. The Goods values in this experiment were all greater than 0.997, indicating that the sequencing coverage represented the actual population. The Chao1 index is commonly used to assess microbial community richness. E0 had a higher Chao1 index than E50, E100, and E400, showing that EPDM input reduced the number of microorganisms in the bioretention system (Table 2). Both the Shannon and Simpson indexes indicate microbial community diversity; higher Shannon and lower Simpson indexes suggest greater diversity. These two sets of data showed that the addition of EPDM reduced the microbial diversity of the bioretention system.

At the genus level, Dechloromonas, Mycobacterium, Pseudarthrobacter, Ramlibacter, Zoogloea, Nocardia, and Rhizobacter were the main genera in the bioretention system (Figure 8(b)). EPDM addition considerably boosted the relative abundance of Dechloromonas (5.54–8.69%), making it the dominant genus in E100 and E400. Dechloromonas is a heterotrophic denitrifying bacterium, and its relative abundance can increase denitrification capacity. Mycobacterium is a heterotrophic nitrifying and aerobic denitrifying bacterium that can participate in both processes in aerobic environments (Moura et al. 2018). The relative abundance of Mycobacterium, the dominant genus in E50, increased initially before subsequently decreasing as EPDM concentration increased. This could be because the high EPDM input clogs the pores of the bioretention system, preventing oxygen transport into the filler. The increased relative abundance of Mycobacterium in E50 may explain why this system removed more efficiently than the other systems. Other denitrifiers, including Zoogloea (0.53–1.63%), Ramlibacter (048–0.84%), and Aeromonas (0–2.3%), were similarly enhanced by the addition of EPDM (Lin et al. 2019; Wang et al. 2022a; Hu et al. 2023), whereas the relative abundance of Pseudarthrobacter and Nocardia decreased. The dominant genus of E0 was Pseudarthrobacter (2.99%), and the addition of EPDM reduced its relative abundance by 1.22–1.88%. According to Dhakal et al. (2019), Nocardia is a common aerobic bacteria that can take part in denitrification in aerobic environments. With an increase in EPDM input concentration, the relative abundance of Nocardia declined progressively; its relative abundance in E50, E100, and E400 fell by 1.99, 1.15, and 0.73%, respectively, in comparison to E0. This could be because EPDM introduces an anaerobic environment into the bioretention system, reducing the relative number of aerobic bacteria (Zhou et al. 2023). Conversely, some anaerobic denitrifying bacteria may flourish in this environment, increasing the capacity of the system to denitrify. Denitrifying bacteria with relative abundances more than 1% were found in E50, E100, and E400 at total relative abundances of 9.62, 10.11, and 12.63%, respectively, significantly higher than those seen in E0 (2.37%).

Two distinct rainfall conditions were investigated to study the impact of rubber EPDM particles on bioretention systems. Rubber particles were shown to increase the adsorption capacity of filler, produce an anaerobic environment, and give microorganisms a substrate to consume, all of which improved the ability of the bioretention system to remove nitrogen. In addition, rubber particles lessened the detrimental impacts of drought on the bioretention system and accelerated the recovery of nitrate removal capacity. Rubber particles also changed the structural composition of the microbial community and decreased its diversity. Following rubber particle input, denitrifiers such as Dechloromonas, Ramlibacter, and Zoogloea increased in relative abundance, whereas Pseudarthrobacter and Nocardia decreased. Overall, proper accumulation of captured rubber particles in the bioretention system improved nitrogen removal and drought adaptability. However, as the accumulation of MPs frequently influences the hydrological consequences of the system, the effects of the input of MPs on the performance of the bioretention system should be more comprehensively examined in future research.

This work was supported by the National Natural Science Foundation of China (No. 52070152) and the Qinghai Province Central Government Guides Local Science and Technology Development Fund Projects (No. 2022ZY039). The authors also appreciate the assistance provided by Fang Song at the Instrument Analysis Center of Xi'an University of Architecture and Technology (China).

All relevant data are included in the paper or its Supplementary Information.

The authors declare there is no conflict.