Enhanced Nitrogen Removal from Aquaculture Wastewater Using Biochar-Amended Bioretention Systems

Abstract

Aquaculture wastewater is characterized by large discharge volumes and variable nitrogen concentrations, posing challenges for stable and efficient treatment. This study investigated biochar-amended bioretention systems (BBSs) under varying temperatures (8.0–26.0 °C), influent TN levels, and operation modes (intermittent and continuous flow). In intermittent runs, the 20% biochar system (BBS20) achieved 72.4% TN removal at low influent TN (9.55 mg/L) and 80.4% at high TN (29.96 mg/L), significantly outperforming the control (CBS). In continuous runs, BBS20 reduced effluent TN to 1.75 mg/L within 72 h, yielding higher average HRT, HLR, and ELR than CBS. Mechanistic analyses showed that biochar addition enhanced extracellular polymeric substance (EPS) secretion, stimulated electron transport system activity (ETSA), and increased the relative abundance of denitrifying genera and functional genes (e.g., nirS, narG). These synergistic effects optimized nitrification–denitrification coupling, particularly under low-temperature conditions. The findings demonstrate that biochar amendment is a practical and effective strategy for improving nitrogen removal from aquaculture wastewater.

1. Introduction

The rapid expansion of global aquaculture, driven by increasing food demands, has resulted in substantial wastewater generation, with annual production volumes ranging from 22,640 to 133,120 m³ across major regions, including the USA, Europe, Asia, and the Middle East [1,2]. Aquaculture wastewater is characterized by two key challenges: first, exceptionally high discharge volumes, and second, significant fluctuations in pollutant concentrations due to diverse production systems [1,3,4,5]. Notably, the concentration of nitrogenous compounds, such as ammonia (0.63–289.10 mg/L) and nitrate (0.38–101.30 mg/L), varies widely [6]. These substantial variations in pollutant loads complicate the development of efficient treatment strategies.

Various biotechnological approaches, such as constructed wetlands and floating islands, have been applied to treat aquaculture wastewater [7]. Despite their use, achieving consistent nitrogen removal efficiency remains a challenge, with systems often underperforming, particularly under low-temperature conditions [8,9]. Thus, there is an urgent need to develop innovative and more efficient treatment technologies to enhance nitrogen removal performance [1].

Bioretention systems, widely recognized for their effectiveness in mitigating urban stormwater runoff pollution, demonstrate considerable potential for application in agricultural non-point source pollution control [10,11,12]. These systems offer two distinct advantages: cost-effectiveness and efficient nitrogen removal capabilities. Their typical configuration consists of multiple functional layers, including planting soil layers, filtration layers, and gravel drainage layers, which contribute to their operational simplicity [12,13]. While previous research has primarily focused on stormwater runoff treatment, recent investigations have expanded to examine the efficacy of bioretention systems in treating other wastewater types with similar characteristics to stormwater. Notably, existing studies have confirmed the effectiveness of bioretention systems in controlling nitrogen pollution from nursery runoff and dairy farm runoff [14,15].

Biochar is widely adopted in bioretention systems due to its multifunctional roles in nitrogen removal. Its high porosity and ion-exchange capacity facilitate nitrogen adsorption and create favorable micro-anoxic niches for denitrification [16]. In addition, redox-active functional groups (e.g., –OH, C=O) enhance extracellular electron transfer, accelerating microbial nitrate reduction [17]. Biochar also reshapes microbial ecology by enriching nitrogen-cycling taxa and stimulating EPS secretion, which further supports denitrification processes [18,19]. Collectively, these effects—physical adsorption, electron transfer facilitation, and microbial community modulation—synergistically improve the performance of biochar-amended systems, as demonstrated in previous studies on stormwater and agricultural runoff.

However, limited studies have systematically examined the effects of biochar amendment on aquaculture wastewater treatment using bioretention systems, and the underlying nitrogen removal mechanisms remain insufficiently understood [20]. To bridge this knowledge gap, we constructed three bioretention systems with varying biochar addition ratios (0%, 10%, and 20% by volume) and operated them under controlled conditions simulating typical aquaculture wastewater characteristics, including seasonal temperature variations and different influent nitrogen concentrations [1,6]. The objectives were: (1) to evaluate nitrogen removal performance of biochar-amended bioretention systems under aquaculture-specific conditions, and (2) to elucidate the mechanistic roles of biochar in enhancing system function through EPS production, ETSA, and microbial community analysis. This work provides the first systematic demonstration of biochar-enhanced bioretention for aquaculture wastewater, offering both mechanistic insights and practical guidance for scalable applications.

2. Construction and Operation of Bioretention Systems

2.1. Bioretention Systems Construction and Operation

(1) Three bioretention columns, each with a height of 1.0 m and an inner diameter of 0.2 m, were constructed at Southwest University. The columns were vertically stratified into four distinct layers, installed from bottom to top as follows: a 10 cm drainage layer composed of crushed stones, a 20 cm filtration layer of river sand, a 30 cm media layer consisting of a 7:3 (v/v) mixture of soil and sand, and a 20 cm water storage layer. At the base of each column, a perforated pipe with an inner diameter of 20 mm was installed to facilitate controlled drainage. The drainage outlet was elevated 30 cm above the bottom of the drainage layer, thereby creating a saturated zone that enhances denitrification processes. To standardize plant variables across the systems, each column was planted with Iris Pseudacorus, selected for its robust root length and density. Plants were carefully weighed on an electronic scale to ensure biomass consistency, with no more than a 1.0 g difference among the systems. In this study, CBS was not amended with biochar, while the other two columns were supplemented with straw biochar at volumes of 20% (BBS20) and 10% (BBS10), respectively (Figure 1).

Following the completion of bioretention system construction, pond water was utilized for irrigation over a 60-day period to ensure full development of the systems. The study was designed with two operational modes: intermittent flow and continuous flow. Intermittent flow experiments were conducted over 210 days from October 2022 to May 2023, encompassing three concentration levels—high (October–November 2022, 6 trials, 8.0–13.0 °C), medium (November 2022, 3 trials, 17.0–21.0 °C), and low (December 2022–May 2023, 10 trials, 8.0–13.0 °C). A total of 19 sequential batch experiments were performed, with a one-day interval between trials. Each concentration gradient was replicated three times. During each experiment, 5 L of simulated influent water was applied via sprinkler irrigation, and all effluent was collected for water quality analysis.

Continuous flow experiments were conducted under three distinct temperature conditions: 8.0–12.80 °C (December 2022, 9 trials), 16.50–17.30 °C (March 2023, 9 trials), and 24.50–26.30 °C (May 2023, 9 trials). Each continuous flow experiment lasted 72 h and was repeated three times. Water samples from both inflow and outflow were collected at intervals of 1, 2, 4, 7, 12, 24, 36, 48, and 72 h. In total, 27 continuous flow experiments were carried out.

(2) Hydraulic characterization

Intermittent (batch) mode: Each cycle consisted of a one-time dosing of 5 L influent (V_b) into the column (internal diameter 0.20 m, surface area A = 0.0314 m²). Effluent appeared within 1–2 min and largely drained out within approximately 2 h; however, the system was operated on a once-per-day cycle. Thus, the operational cycle length of 24 h was defined as the batch contact time, which represents the apparent HRT. The areal hydraulic loading rate (HLR) and areal nitrogen loading rate (ELR) were calculated as:

$$HRT={\displaystyle \frac{V_{\mathrm{b}}}{A}}$$

(1)

$$ELR={\displaystyle \frac{\left(C_{\mathrm{T}\mathrm{N}}\times V_{\mathrm{b}}\right)}{A}}$$

(2)

where C_TN is the influent TN concentration (mg/L), V_b is the batch volume (L), and A is the surface area (m²). For influent TN concentrations of 9.55, 20.23, and 29.96 mg/L, the ELRs were 1.52, 3.22, and 4.77 g N·m⁻²·d⁻¹, respectively (Table S1).

Continuous-flow mode: Influent was continuously supplied through a faucet at a low but unmetered rate. Effluent flow rates (Q_out(t), L·h⁻¹) were recorded at 1, 2, 4, 7, 12, 24, 36, 48, and 72 h. Under quasi-steady conditions without ponding, Q_in(t) ≈ Q_out(t) was assumed. The apparent HRT, HLR, and ELR were calculated using the following equations:

$$HRT\left(t\right)={\displaystyle \frac{V_{\mathrm{P}\mathrm{o}\mathrm{r}\mathrm{e}}}{Qout\left(t\right)}}$$

(3)

$$HLR\left(t\right)={\displaystyle \frac{Q\mathrm{o}\mathrm{u}\mathrm{t}\left(t\right)}{A}}$$

(4)

$$ELR\left(t\right)={\displaystyle \frac{\left(C_{\mathrm{T}\mathrm{N}}\times Q\mathrm{o}\mathrm{u}\mathrm{t}\left(t\right)\right)}{A}}$$

(5)

where V_pore = 4.08 L represents the pore water volume. Time-averaged values over 0–72 h and cumulative throughput are summarized in Table S2, while the time-resolved effluent flow dynamics are provided in Figure S1.

2.2. Materials

The biochar used in this study was provided by Lize Environmental Technology Co., Ltd. (Zhengzhou, China) and produced from straw via pyrolysis at 500 °C. Crushed stone (8–23 mm), river sand (0.2–4 mm), and planting soil were sourced from Southwest University (Chongqing, China). Among the four fillers, biochar had the lowest bulk density (0.17 g/cm³) and highest void ratio (88.1%), while river sand had the highest density (1.49 g/cm³) and lowest void ratio (7.62%). Biochar showed moderate porosity (44%), comparable to planting soil (46%) and higher than river sand (34%), but lower than gravel (62%). In terms of pH, all materials were alkaline, ranging from 7.58 (soil) to 10.12 (biochar), highlighting biochar’s porous, low-density, and highly alkaline characteristics favorable for microbial activity.

The selected TN levels (9.55, 20.23, and 29.96 mg/L) represent the typical low, medium, and high ranges reported for pond and recirculating aquaculture system effluents [1,21,22], Corresponding COD concentrations were adjusted to reflect characteristic C/N ratios of aquaculture wastewater, ensuring that the simulated influent closely mimicked real discharge conditions. Three simulated influent types were prepared using fish manure, KNO₃, NaNO₂, KH₂PO₄, (NH₄)₂SO₄ (Sinopharm Chemical Reagent Co., Ltd., Shanghai, China), and glucose to replicate pond and RAS wastewater characteristics. The low-concentration influent contained 9.55 mg/L TN, 6.49 mg/L NO₃⁻–N, 0.68 mg/L NH₄⁺–N, 0.12 mg/L NO₂⁻–N, 1.98 mg/L TP, and 94.0 mg/L COD. Medium concentration included 20.23 mg/L TN, 14.67 mg/L NO₃⁻–N, 2.21 mg/L NH₄⁺–N, 0.43 mg/L NO₂⁻–N, 4.57 mg/L TP, and 184.67 mg/L COD. High concentration reached 29.96 mg/L TN, 19.77 mg/L NO₃⁻–N, 2.30 mg/L NH₄⁺–N, 0.59 mg/L NO₂⁻–N, 6.12 mg/L TP, and 290.0 mg/L COD.

2.3. Test Method for Water Quality

The effluent from bioretention system was collected in polyethylene plastic bottles. Samples were filtered through 0.45 μm cellulose acetate membrane and analyzed as follows: NO₃⁻–N was determined by ultraviolet spectrophotometry (HJ/T 346-2007) [23] NH₄⁻–N by Nessler’s reagent method (HJ 535-2009) [24], NO₂⁻–N by spectrophotometry (GB 7493-87) [25], and TN by alkaline potassium persulfate digestion (HJ 636-2012) [26]. All measurements were performed using the same TU-1901 double-beam UV–Vis spectrophotometer (Beijing Purkinje General Instrument Co., Ltd., Beijing, China).

2.4. Extraction and Detection of Extracellular Polymeric Substance

EPS was extracted following the method described by [27]. Soil samples were collected at depths of 15 cm and 45 cm using the equilateral triangle sampling method. After homogenizing equal volumes of the collected soil, EPS extraction was performed using the cation exchange resin method. The detailed extraction procedure is provided in the Supplementary File.

Polysaccharides (PS) were quantified using the phenol-sulfuric acid method. Proteins (PN) were measured using a BCA assay kit (Ranjco Technology Co., Ltd., Beijing, China), with bovine serum albumin as the standard. Absorbance was recorded at 562 nm, and concentrations were calculated from a standard curve. For fluorescence analysis of EPS, an F-7000 fluorescence spectrophotometer (Hitachi, Tokyo, Japan) was employed. The EPS extraction solution was filtered through a 0.45 μm membrane. Fluorescence measurements were conducted with excitation wavelengths (Ex) ranging from 200 to 550 nm and emission wavelengths (Em) set within the same range, both with intervals of 5 nm and a scanning speed of 12,000 nm/min.

2.5. Quantification of Electron Transport Systems Activities

The measurement method for activities of ETSA was referenced from [28,29]. The ETSA was based on the reduction of 2-(p-iodophenyl)-3-(p-nitrophenyl)-5-phenyl tetrazole chloride (INT) to formazan (INF). A total of 10 g media soil sample was taken and washed twice with 4 mL of phosphate buffer (PBS, 100 mM, pH 7.4), and then resuspended in PBS solution to obtain a suspension sample. Next, 1 mg of nicotinamide adenine dinucleotide (NADH) and 1 mL of INT were added to 5 mL of the above extraction solution, and incubated in darkness at 30 °C for 30 min. After incubation, 1 mL of formaldehyde was added to terminate the reaction, and the supernatant was discarded after centrifugation at 10,000 rpm for 5 min. Then, 5 mL of methanol was added to extract the formazan. Finally, the formation of formazan was measured at a wavelength of 490 nm, and ETSA was calculated using the following Equation (6).

$$\mathrm{E}\mathrm{T}\mathrm{S}\mathrm{A}(\mathrm{\mu }\mathrm{g}{\mathrm{O}}_2·{\left(\mathrm{g}·\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{t}\mathrm{e}\mathrm{i}\mathrm{n}·\mathrm{m}\mathrm{i}\mathrm{n}\right)}^{-1})={\displaystyle \frac{{\mathrm{A}\mathrm{B}\mathrm{S}}_{490}}{15.9}}{\displaystyle \frac{{\mathrm{V}}_1}{{\mathrm{V}}_0\mathrm{t}}}{\displaystyle \frac{32}{2}}{\displaystyle \frac{1}{\mathrm{m}}}$$

(6)

In Equation (6), ABS₄₉₀ denotes the absorbance of formazan at 490 nm, 15.9 is the molar absorption coefficient of INT-formazan (L·mmol⁻¹·cm⁻¹), V₁ is the volume of methanol for formazan extraction (mL), V₀ is the incubated suspension volume (mL), t is the incubation time (min), 32/2 represents the stoichiometric conversion factor (32 g O₂ per mol electron equivalent divided by two electrons per INT), and m is the protein concentration of the sample (mg/mL).

2.6. Metagenomic Analysis

Soil samples were collected from depths of 15 cm and 45 cm, mixed in equal volumes, and stored at −80 °C for subsequent DNA extraction. Briefly, genomic DNA was extracted and its integrity verified via 1% agarose gel electrophoresis (Beijing Solarbio Science & Technology Co., Ltd., Beijing, China). The extracted DNA was then sheared to an average fragment size of approximately 400 bp using a Covaris M220 instrument (Covaris, Woburn, MA, USA). Following fragmentation, PCR amplification and sequencing were both performed by Shanghai Majorbio Bio-pharm Technology Co., Ltd. (Shanghai, China).

Open reading frame (ORF) prediction was performed on the contigs obtained from the assembly results using Prodigal v2.6.3 (https://github.com/hyattpd/Prodigal accessed on 15 July 2025). Genes with nucleotide lengths equal to or greater than 100 bp were selected and translated into amino acid sequences. CD-HIT v4.8.1 was used to cluster all predicted gene sequences from samples, and the longest gene in each cluster was chosen as representative sequence to create non-redundant gene set. SOAPaligner v2.21 was used to align the high-quality reads from each sample to non-redundant gene set, and abundance of each gene in respective sample was determined.

2.7. Statistical Analysis

Statistical analyses were conducted using IBM SPSS Statistics 24. All experiments were performed in triplicate, and results are presented as mean ± standard deviation (SD). Differences in pollutant removal performance among systems were evaluated by one-way analysis of variance (ANOVA), with p < 0.05 considered statistically significant. When significant effects were detected, Tukey’s HSD test was applied for pairwise comparisons. ANOVA results are reported with F-values and degrees of freedom, and significant differences in figures are indicated as p < 0.05 (*) and p < 0.01 (**). In addition, Pearson’s correlation analysis was used to assess linear relationships between key variables.

3. Results and Discussion

3.1. Biochar Addition Enhanced Nitrogen Removal Under Intermittent Flow Conditions

Each 5 L dosing corresponded to an areal hydraulic loading (HLR) of 159 mm·batch⁻¹, with TN-ELRs of 1.52, 3.22, and 4.77 g N·m⁻²·batch⁻¹ under low (9.55 mg/L/), medium (20.23 mg/L), and high (29.96 mg/L/) influent TN, respectively (Table S1). The operational cycle length of 24 h was considered the batch contact time (apparent HRT). Under the low TN condition, CBS, BBS10, and BBS20 achieved TN removal efficiencies of 36.5%, 46.4%, and 72.4%, respectively, with similar trends for NO₃⁻–N and NH₄⁺–N. These results demonstrate that biochar amendment enhanced nitrogen removal performance across increasing ELR levels.

At low temperatures (8.0–13.0 °C, Figure 2), biochar addition markedly improved nitrogen removal performance. Under low-concentration influent, CBS achieved only 36.46 ± 1.81% TN removal, whereas BBS10 and BBS20 reached 46.32 ± 1.02% and 72.40 ± 0.97%, respectively. Similar trends were observed for NO₃⁻–N and NH₄⁺–N, with BBS20 showing removal efficiencies of 67.93 ± 0.60% and 84.23 ± 5.00%, both significantly higher than those of CBS and BBS10 (p < 0.05). For NO₂⁻–N, BBS20 exhibited 51.52 ± 2.62% removal, comparable to CBS (59.85 ± 6.20%) but still substantially better than BBS10 (25.51 ± 7.59%). These results indicate that biochar enhances nitrification and denitrification under cold conditions and helps stabilize system performance, although differences among treatments may vary by nitrogen species.

Under high-concentration influent, biochar continued to improve system performance (Figure 3). BBS20 achieved a TN removal rate of 80.40 ± 0.24%, significantly higher than BBS10 (70.12 ± 0.40%) and CBS (65.09 ± 0.62%) (p < 0.01). Similarly, BBS20 showed superior NO₃⁻–N and NH₄⁺–N removal rates (78.21 ± 0.26% and 74.08 ± 2.10%) compared to CBS (61.25 ± 0.13% and 55.56 ± 1.96%). Although the improvement in NO₂⁻–N removal was less pronounced under high loading, BBS20 still maintained a relatively high NO₂⁻–N removal rate (64.16 ± 3.62%), suggesting its robustness under stress conditions.

When the ambient temperature increased to 17.0–21.0 °C (Figure 3), microbial metabolism accelerated, boosting the activity of nitrifying and denitrifying bacteria. Notably, even under these favorable thermal conditions, biochar continued to exhibit a clear enhancement effect. BBS20 reached a TN removal rate of 78.81 ± 0.27%, significantly outperforming BBS10 (69.17 ± 0.36%) and CBS (58.67 ± 0.57%) (p < 0.01). Similarly, NO₃⁻–N removal improved from 58.75 ± 1.43% (CBS) to 76.50 ± 0.46% (BBS20). NH₄⁺–N and NO₂⁻–N removal also increased to 82.20 ± 0.41% and 77.33 ± 1.53%, respectively, further demonstrating the consistent benefits of biochar addition.

Notably, higher influent concentrations of TN and NO₃⁻–N correlated with elevated removal rates [30,31], whereas increased NH₄⁺–N concentrations inversely reduced removal efficiency. This may be due to the abundance of unused feed and excreta in the culture effluent, leading to higher levels of organic nitrogen [1,21,32]. In turn, ammoniation reaction could transfer organic nitrogen to ammonia nitrogen, causing the growth of NH₄⁺–N concentration and a slight reduction during its removal. Some studies have demonstrated that biochar-amended bioretention systems exhibit enhanced nitrogen removal efficiency under high-load conditions [14], primarily attributed to the synergistic interplay between adsorption and biodegradation mechanisms [33,34]. Biochar first removes nitrogen from water through physical adsorption, and more biochar addition may lead to greater nitrogen removal. Permeable structure of biochar naturally augmented the attachment sites for microorganisms, redox actives and conductivity [35,36,37,38,39]. With respect to biodegradation, biochar can also provide habitats for microorganisms, promote electron transfer rate and enhance associated enzyme activities at lower temperatures, further improving the denitrification capacity of the bioretention systems [40]. That may be the reason why the higher biochar addition, the superior nitrogen removal at different temperatures and influent loads.

3.2. Biochar Addition Improved Nitrogen Removal Under Continuous Operation Mode

During 72 h of continuous operation, the average effluent flow rates were 2.59, 1.69, and 2.04 L·h⁻¹ for CBS, BBS10, and BBS20, corresponding to mean apparent HRTs of 2.08, 2.72, and 2.39 h, respectively. The associated mean HLRs were approximately 1980, 1294, and 1556 mm·d⁻¹, while the TN-ELRs reached 21.2, 13.8, and 16.6 g N·m⁻²·d⁻¹ (Table S2). These hydraulic metrics indicate that biochar amendment enhanced the systems’ capacity to tolerate higher nitrogen loading while maintaining effective retention times, underscoring its potential scalability for aquaculture wastewater treatment. Despite temperature fluctuations, the biochar-amended systems—particularly BBS20—consistently achieved higher TN, NO₃⁻–N, and NH₄⁺–N removal than CBS, confirming the robustness of biochar under varying hydraulic conditions.

At low temperature (8.0–12.80 °C), after 72 h of operation, TN removal rates were 47.20% for CBS, 48.30% for BBS10, and 50.20% for BBS20. BBS20 maintained better performance under low-temperature constraints, especially in nitrate removal. The NO₃⁻–N effluent concentration in CBS was 1.99 mg/L, while BBS20 reached 1.37 mg/L, an increase of 11.70 percentage points. Additionally, NH₄⁺–N and NO₂⁻–N concentrations in BBS20 remained relatively stable between 24 and 48 h, indicating better buffering and regulation capacity. When temperature rose to 16.50–17.30 °C, all systems’ overall performance improved notably. BBS20 exhibited the most pronounced enhancement, with TN decreasing from 9.19 mg/L to 1.75 ± 0.12 mg/L at 72 h, achieving a removal rate of 80.9%, significantly higher than CBS (58.67%) and BBS10 (69.17%). NO₃⁻–N in BBS20 effluent dropped to just 0.02 ± 0.01 mg/L (99.7% removal), compared to 85.1% in CBS. As temperature increased further to 24.50–26.30 °C, nitrogen removal was most effective across all systems. TN effluent concentrations at 72 h were 0.85 mg/L (CBS, 91.70% removal), 0.33 mg/L (BBS10, 96.80%), and 0.26 mg/L (BBS20, 97.50%) (Figure 4).

This performance improvement suggested biochar amendment may promote the proliferation of ammonia-oxidizing bacteria, as previously documented in relevant microbial studies [29]. Biochar exhibits a substantial specific surface area coupled with abundant negatively charged functional groups (-COO- and -C₆H₄O-), which significantly enhances its ammonium adsorption capacity [39,41]. This modification promotes the development of microenvironments conducive to both autotrophic and heterotrophic denitrifying bacterial communities [40,42]. Furthermore, dissolved organic matter liberated through biochar dissolution could serve as an effective carbon substrate for denitrification processes [43].

3.3. EPS Content Analysis

The EPS content in all three bioretention systems increased with biochar supplementation and temperature elevation (Figure S2). Biochar amendment significantly enhanced microbial EPS production, with the BBS20 system exhibiting particularly pronounced effects. Compared to CBS, BBS20 showed higher PN and PS levels by 382.92 μg/g and 80.33 μg/g, respectively, under optimal thermal conditions (24.50–26.30 °C; p < 0.05). These EPS components play critical biochemical roles in denitrification efficiency through their redox mediator constituents, particularly C-type cytochromes and flavins. These molecules facilitate electron transfer between microbial donors and nitrate acceptors [44,45,46]. Enhanced EPS production improved electron mass transfer efficiency, a rate-limiting factor in denitrification kinetics [47]. Biochar’s stimulatory effects on microbial metabolism manifest through both EPS biosynthesis promotion and redox activity potentiation [48,49,50].

Temperature-dependent microbial activity further modulated EPS dynamics, with psychrophilic conditions (8.0–12.80 °C) yielding average PN and PS concentrations of 1044.16 μg/g and 201.36 μg/g, respectively. Thermal elevation to mesophilic ranges (24.50–26.30 °C) induced significant increases to 1965.68 μg/g PN and 427.48 μg/g PS, reflecting enhanced microbial metabolic activity and physiological regulation.

3.4. Mechanisms of Nitrogen Removal by Bioretention Systems

3.4.1. Analysis of Electron Transport Activity

The electron transport capacity of biochar plays a crucial role in facilitating extracellular electron transfer by microorganisms throughout the microbial nitrate reduction process [17]. Notably, the BBS20 system with 20.0% biochar addition demonstrated significantly higher ETSA compared to CBS (Figure S2) (p < 0.05). This enhancement can be attributed to two primary mechanisms: First, key EPS components such as humic acids and cytochrome C function as electron shuttles, facilitating efficient intercellular electron transfer. Previous studies demonstrate that EPS-mediated redox processes can increase denitrification efficiency by 19.0–34.0% through the release of redox-active mediators [45]. Second, microbial metabolism of carbon substrates generates electrons that are subsequently channeled through the electron transport system to drive denitrifying enzymes, a process fundamentally dependent on ETSA for effective nitrate removal [36,45]. Crucially, the redox-active components within EPS exhibit strong positive correlations with ETSA levels [51,52].

Temperature elevation further amplified ETSA performance (Figure S3). When ambient temperature increased from 8.0 to 12.80 °C to 24.50–26.30 °C, ETSA values rose by 0.12 μg O₂. (g. protein.min)⁻¹, 0.16 μg O₂. (g. protein.min)⁻¹ and 0.18 μg O₂. (g. protein.min)⁻¹ in CBS, BBS10, and BBS20 systems, respectively. This thermal enhancement likely stems from improved microbial metabolic activity under warmer conditions, which stimulates increased EPS secretion. The temperature-dependent accumulation of redox mediators in EPS enables more efficient electron shuttling, thereby amplifying ETSA.

3.4.2. Microbial Community Dynamics Analysis

Microbial α-diversity was further evaluated using Shannon and Simpson indices (Table S3). The results showed that CBS exhibited the highest diversity (Shannon = 6.49, Simpson = 0.0094), whereas biochar-amended systems BBS10 and BBS20 displayed slightly reduced diversity, suggesting selective enrichment of nitrogen-cycling taxa. In contrast, temperature elevation (T2, T3) increased microbial diversity relative to low temperature (T1), indicating thermal stimulation of community richness and stability.

Biochar supplementation significantly enhanced aerobic nitrification, accompanied by a clear shift in microbial community composition (Figure 5a,b). Compared with CBS, BBS10 and BBS20 exhibited increased relative abundances of nitrifying and denitrifying genera. Specifically, BBS10 showed higher abundances of Nakamurella (7.70%), Cellulomonas (4.30%), and Georgenia (2.90%), while BBS20 recorded increases of 7.10%, 4.60%, and 3.10%, respectively. Nakamurella, a member of Actinobacteria, was notably enriched in biochar-amended systems, suggesting potential adaptation to long-term aquaculture wastewater conditions [53,54]. Both Cellulomonas and Georgenia contribute to lignocellulosic degradation and humic substance production, which can facilitate extracellular electron transfer and promote denitrification [55,56]. These results indicate that biochar addition favors the enrichment of functional genera closely involved in nitrogen cycling, consistent with the improved nitrogen removal efficiency observed.

The relative abundance of dominant microbial groups also displayed temperature-dependent enrichment patterns (Figure 5c,d). At the phylum level, Proteobacteria (47.79%), Actinobacteria (40.95%), Bacteroidota (2.87%), and Firmicutes (1.05%) were predominant under mesophilic conditions (24.50–26.30 °C). Proteobacteria and Firmicutes showed marked thermal responsiveness, with relative abundances 25.96% and 0.47% higher, respectively, compared to psychrophilic conditions (8.0–12.80 °C). This enhanced representation of Proteobacteria is consistent with their well-documented role in nitrification and denitrification pathways [42]. Additionally, Planctomycetota, Actinobacteria, and Bacteroidota were also identified as important phyla contributing to nitrogen cycling in wastewater ecosystems [29,57]. At the genus level, thermal stratification analysis revealed progressive enrichment of Hydrogenophaga and Thauera with rising temperatures, both well-known denitrifiers that increased from baseline abundances of 1.10% and 0.46% to 2.60% and 1.10%, respectively. Moreover, members of the class Alphaproteobacteria also increased in relative abundance across temperature gradients, and taxa within this group are recognized contributors to nitrite reduction processes [58].

3.4.3. Analysis of Key Genes in Nitrogen Metabolism

This study demonstrates that elevated temperature and biochar addition synergistically enhanced the relative abundance of key functional genes involved in the nitrogen cycle, thereby restructuring the coupling between nitrification and denitrification pathways and improving overall nitrogen transformation efficiency (Figure 6). Temperature primarily promoted the stepwise reduction of nitrate and its intermediates, while biochar addition more strongly enriched genes associated with the middle stages of denitrification, optimizing the balance between carbon supply and electron flow.

In the denitrification pathway, increasing temperature (from T1 to T3) substantially increased the relative abundance of several essential reductase genes. At the initial stage, membrane-bound nitrate reductase (narG) and periplasmic nitrate reductase (napA) jointly catalyze the reduction of nitrate (NO₃⁻) to nitrite (NO₂⁻) [59,60], and their relative abundances increased by 60.40% and 143.10%, respectively. During the intermediate step, nitrite reductase (nirS), which converts NO₂⁻ to nitric oxide (NO) [61], increased by 275.20%, indicating that it was the most temperature-responsive gene and highlighting its importance as a known rate-limiting step in denitrification. Subsequently, norB and nosZ, responsible for reducing NO to nitrous oxide (N₂O) [62], and N₂O to N₂ [63], increased by 103.60% and 89.70%, respectively. These results indicate that elevated temperature facilitated a more complete denitrification cascade, which is beneficial for reducing N₂O accumulation and mitigating greenhouse gas emissions. In the nitrification pathway, ammonia monooxygenase (amoA) increased by 471.30% [64], suggesting an enhanced initiation of ammonia oxidation and greater provision of nitrite for subsequent denitrification, even though no major change was observed for hydroxylamine oxidoreductase (hao).

Biochar addition, in contrast, primarily affected intermediate denitrification steps. With increasing biochar proportion (CBS → BBS10 → BBS20), the relative abundance of nirS increased by 17.20%, reinforcing the reduction of NO₂⁻ to NO. In comparison, napA decreased by 23.70% while narG remained relatively stable, suggesting a system preference for the narG-mediated pathway under carbon-enriched conditions due to its stronger coupling with intracellular electron transport. Moreover, the abundance of NRT2, a high-affinity nitrate transporter [65], increased by 21.30%, implying enhanced nitrate uptake capacity and improved substrate supply to the denitrification pathway.

In summary, temperature elevation increased the relative abundances of narG, napA, nirS, norB, and nosZ, promoting a complete denitrification sequence from nitrate to nitrogen gas, while also enhancing amoA abundance and the supply of oxidized nitrogen intermediates from nitrification. Biochar addition primarily strengthened the nirS-mediated step and favored narG-linked nitrate reduction under carbon-rich conditions. Together, these factors enhanced the integrity and efficiency of nitrogen removal in bioretention systems, with important ecological implications for reducing N₂O emissions and engineering significance for improving stability and scalability in aquaculture wastewater treatment.

3.5. Mechanistic Reconstruction of Nitrogen Removal

This study integrated EPS characteristics, ETSA, and key denitrification enzyme expression levels to elucidate the synergistic pathways through which temperature elevation and biochar addition restructured nitrogen removal mechanisms in bioretention systems (Figure 7). The results revealed that enhanced EPS production, accelerated electron transfer, and activation of functional enzymes jointly contributed to improved nitrogen transformation efficiency.

First, biochar significantly promoted EPS secretion. Under optimal temperature conditions (24.50–26.30 °C), PN and PS contents in the BBS20 system reached 382.92 μg/g and 80.33 μg/g, respectively—significantly higher than in the CBS (p < 0.05). The redox-active components within EPS, such as humic acids and cytochrome c, act as electron shuttles that enhance inter-microbial electron transfer and thereby accelerate denitrification.

Second, ETSA—an indicator of electron transport intensity—increased significantly with temperature. ETSA values in CBS, BBS10, and BBS20 rose by 0.12, 0.16, and 0.18 μg O₂. (g. protein.min)⁻¹, respectively, indicating that biochar combined with elevated temperature effectively overcame electron donor limitations in denitrification, improving both electron availability and nitrogen removal rates.

Third, temperature and biochar jointly upregulated the expression of multiple key nitrogen-cycle enzymes, restructuring the nitrification–denitrification coupling pathways. Elevated temperature significantly enhanced the expression of narG, napA, nirS, norB, and nosZ, with nirS (nitrite reductase) showing a dramatic 275.20% increase—highlighting it as a major regulatory target in the rate-limiting step. Additionally, amoA expression increased by 471.30%, indicating substantial activation of the NH₄⁺–N oxidation step under warming conditions. Biochar primarily strengthened the nirS-mediated NO₂⁻–N reduction (BBS20 showed a 17.20% increase over CBS) and favored the narG-dominated intracellular NO₃⁻ reduction pathway. Furthermore, the 21.30% increased relative abundance of the high-affinity nitrate transporter NRT2 suggested improved NO₃⁻ uptake capacity alongside enhanced electron supply.

Engineering implications: From an engineering perspective, the calculated hydraulic parameters provide critical insights for scaling biochar-amended bioretention systems (BBS) to field conditions. The batch-mode TN loadings of 1.52–4.77 g N·m⁻²·d⁻¹ and continuous-mode loadings of ~13–21 g N·m⁻²·d⁻¹ fall within the typical range of aquaculture wastewater discharge, suggesting practical feasibility. The apparent HRT of ~2 h in continuous runs indicates that effective nitrogen removal can be achieved without long residence times, supporting compact reactor design. However, real-world applications must account for potential clogging, long-term biochar stability, and operational sustainability under fluctuating influent loads. These considerations highlight the need for pilot-scale studies before full-scale implementation.

4. Conclusions

(1) Bioretention systems are suitable for nitrogen removal in aquaculture wastewater treatment. Biochar amendment effectively enhances the denitrification efficiency of bioretention systems, particularly overcoming the technical limitation of inadequate nitrogen removal during winter low-temperature periods.

(2) Biochar improved EPS secretion and ETSA, enhancing electron transfer and denitrification efficiency, especially at low temperatures.

(3) Biochar enriched key denitrifying bacteria and optimized nitrification–denitrification pathways, resulting in higher nitrogen removal efficiency.