Efficient Pollutant Removal and Low-Carbon Emission Mechanisms in Constructed Wetlands Synergistically Driven by Low COD/N Ratio and Coastal Location

Abstract

Quantifying the variation in wetland greenhouse gas fluxes across large spatial scales and accurately assessing source–sink effects is crucial. However, there remains a limited understanding of the combined impacts of influent COD/N ratios and geographical distribution conditions on pollutant removal and GHG emissions. In this study, five typical constructed wetlands from across the country were selected to evaluate GHG emissions, pollutant removal efficiencies, and the main influencing factors for each wetland. The results showed that temperature, ammonia nitrogen concentration, COD, COD/N ratio, and geographical location were the main regulators of GHG emissions, with complex interactions among the factors. Overall, GHG emissions were higher in the coastal region than in the inland region, highlighting the importance of geographic distribution conditions on wetland operation. In addition, wetlands with a COD/N of 3 showed the best overall performance in terms of pollutant removal and GHG emission reduction. Moreover, COD/N had an important effect on the emission fluxes of all three greenhouse gases, which was an important influencing factor on the emission fluxes of greenhouse gases from constructed wetlands. Wetlands with lower COD/N ratios, especially coastal wetlands, showed stronger performance in pollutant removal and GHG emission reduction. This study emphasizes the need to fully consider the potential influence of influent COD/N ratio on GHG emissions when designing constructed wetlands for municipal wastewater treatment, providing valuable insights for future wetland design and GHG abatement strategies.

1. Introduction

Constructed wetlands (CWs), as an ecologically engineered water treatment technology, are widely used in the field of wastewater treatment due to their high efficiency, low cost, and environmentally friendly characteristics. Its core function is to remove pollutants, including organic matter, nutrients such as nitrogen and phosphorus, and heavy metals through the synergistic action of physical, chemical, and biological processes [1,2]. Studies have shown that the pollutant removal efficiency of constructed wetlands is influenced by a variety of factors, such as plant species, substrate type, hydraulic retention time, and environmental conditions [3,4]. For example, wetland plants can effectively remove nitrogen and phosphorus through uptake and microbial synergism, whereas the choice of substrate significantly affects the adsorption and precipitation of heavy metals [5]. Although a large number of studies have explored the pollutant removal mechanisms of constructed wetlands, most of them focus on small-scale experiments or optimization under specific conditions and lack systematic assessment at large spatial scales.

Emission of greenhouse gases (GHGs) is an environmental issue that cannot be ignored during the operation of constructed wetlands. The anaerobic environment and microbial activities in constructed wetlands may lead to the release of CH₄ and N₂O, which have a much higher global warming potential (GWP) than CO₂ [6,7]. The emission of GHGs not only undermines the environmental benefits of constructed wetlands but also may exacerbate global climate change. Therefore, controlling GHG emissions in constructed wetlands is crucial to achieving their sustainable operation. It has been shown that GHG emissions can be reduced to a certain extent by optimizing wetland design parameters (e.g., plant configuration, water level control) and operating conditions (e.g., influent carbon and nitrogen ratios, hydraulic loading rates) [8,9]. However, most of the existing studies have focused on the effects of single factors on GHG emissions, and systematic analyses of the synergistic effects of multiple factors are lacking.

The influent chemical oxygen demand to nitrogen (COD/N) ratio is one of the key factors affecting pollutant removal and GHG discharge from constructed wetlands. Studies have shown that the variation in COD/N significantly affects the nitrogen removal efficiency and GHG emission flux [10,11]. The pollutant removal efficiency of constructed wetlands was higher when the COD/N was between 5:1 and 10:1, while the GHG emissions were relatively low [12]. Guo et al. further found that the addition of biochar could effectively reduce the GWP of wastewater and promote the pollutant removal under different COD/N conditions [13]. However, these studies were mostly based on laboratory or small-scale pilots and failed to fully reflect the effect of COD/N on the performance of constructed wetlands in practical engineering applications.

Although a large number of studies have been conducted to explore the mechanisms of pollutant removal and GHG emissions from constructed wetlands, there are still some key issues that have not yet been resolved. Firstly, most of the existing studies focus on small-scale experiments and lack quantitative analyses of GHG flux and pollutant removal effects of constructed wetlands at large spatial scales. Secondly, studies on the effect of COD/N on the performance of constructed wetlands mostly focus on single pollutants or GHGs and lack a systematic assessment of the synergistic effect of pollutant removal and GHG emissions. In addition, the existing studies failed to fully consider the combined effects of environmental factors (e.g., temperature, pH) and operating conditions (e.g., hydraulic retention time) on the performance of constructed wetlands. Therefore, this study aims to quantify the GHG emissions flux and pollutant removal efficiency of constructed wetlands through field sampling at a large spatial scale and to explore the key influencing factors so as to provide a scientific basis for the optimal design and sustainable operation of constructed wetlands.

2. Materials and Methods

2.1. Description of Experimental Wetlands

Five CWs crossing from North to South China were investigated in this study. CW-1 is located in Dongguan, Guangdong, China (23°09′ N 113°90′ E), covering an area of 255,000 square meters, and the wetland system treatment scale is 120,000 m³ d⁻¹. The main vegetation types in the wetland included ardonopsis, canna, windmill grass, reeds, etc., as well as some landscape plants. The flow of the wetland is surface flow wetland, which mainly bears tail water of Dongguan sewage treatment plant. CW-2 is located in Zhongning, Ningxia Hui Autonomous Region, China (36°97′ N 105°63′ E), covering an area of 100,000 square meters, with a wetland system treatment scale of 20,000 m³ d⁻¹. The wetland is planted with reed, calamus, and scallions. The flow type for this is undercurrent flow. The main inflow for this wetland is from the tail water of the Zhongning County sewage plant. CW-3 is located in Haining, Zhejiang, China (30°57′ N 120°71′ E), covering an area of 1,738,700 square meters. The treatment scale of the wetland system is 300,000 m³ d⁻¹, and the wetland vegetation is dominated by reeds. CW-4 is located in Anxin, Hebei, China (38°90′ N 115°79′ E), covers an area of 2.15 million square meters, and the wetland system treatment scale is 250,000 m³ d⁻¹. Reed and calamus are planted in the wetland. The flow type is surface flow wetland, and the main treatment of this wetland is the tail water from four sewage treatment plants in Baoding City. CW-5 is located in Changsha, Hunan, China (28°55′ N 113°34′ E), covering an area of 540 square meters. The wetland system scale of this wetland is 360 m³ d⁻¹. The wetland is planted with green bristlecone, and the flow type of the wetland is surface flow wetland, which mainly treats tail water from livestock and poultry breeding wastewater.

The individual wetlands were classified into three types based on their COD/N. CW-1 and CW-2 were classified as low carbon-to-nitrogen ratio wetlands, CW-3 and CW-4 were classified as medium carbon-to-nitrogen ratio wetlands, and CW-5 was classified as a high carbon-to-nitrogen ratio wetland.

2.2. Sample Collection and Analysis

2.2.1. Liquid and Soil Sample Collection

The influent and effluent (liquid sample) were collected using 100 mL plastic bottles, 3 replicates were collected, and water temperature and pH were measured on site and then quickly brought back to the laboratory for timely measurement. The sampling point was located at the drain outlet at the bottom of the storage tank, and the water sample was collected with a 100 mL polyethylene bottle, and the temperature and pH of the water body were measured on the spot. Then, the sample was taken back to the laboratory for timely determination or short-term refrigeration in the freezer (4 °C). After filtration, the concentration of ammonium nitrogen and nitrate nitrogen in water samples was determined by flow injection analyzer. COD in and out water was measured by the potassium dichromate method [14].

The soil from different CWs was sampled as below. Three 0.5 × 0.5 × 1 m³ quadways were set parallel to each sampling point in the serpentine layout, and 500 g of soil samples were taken. The same soil on different sides of each sample site was thoroughly mixed, and all soil collected was placed in sterile Ziplock bags and promptly brought back to the laboratory in ice boxes. The content of total carbon (TC) and total nitrogen (TN) was determined by elemental analyzer (FLASH EA1112 Series CNS). The pH of soil was measured by potentiometric method, and the moisture content of soil was measured by the drying method. The basic properties of wetland soils are shown in

Table 1. Basic properties of constructed wetland soil.

	Soil pH	Soil Moisture Content (%)	Soil Carbon Content (%)	Soil Nitrogen Content (%)
CW-1	6.71 ± 0.64	5.14 ± 1.14	0.69 ± 0.38	0.11 ± 0.06
CW-2	7.69 ± 0.19	25.84 ± 9.84	1.43 ± 0.31	0.07 ± 0.03
CW-3	7.28 ± 0.11	20.78 ± 6.04	0.89 ± 0.19	0.12 ± 0.03
CW-4	7.80 ± 0.09	8.19 ± 2.38	0.64 ± 0.01	0.05 ± 0.01
CW-5	6.32 ± 0.23	16.23 ± 2.17	0.28 ± 0.02	0.07 ± 0.01

, and water quality-related data are shown in

Table 2. The water quality in the influent from different constructed wetlands.

	Average Temperature (°C)	COD/N	Influent COD (mg/L)	Influent NH4+-N (mg/L)	Influent TN (mg/L)
CW-1	26.8	3	34.28	0.50	10.23
CW-2	20.3	3	26.97	0.16	8.64
CW-3	25.8	6	33.77	0.23	5.64
CW-4	28.2	8	30.62	0.04	3.74
CW-5	14.0	>40	202.79	2.30	4.65

.

2.2.2. GHG Collection

All measurements were taken between March and May 2022. Greenhouse gases are collected and analyzed by static box–gas chromatography [15]. The static box is made of PVC material. In order to adapt to the difference in plant growth and reach out to the different height in the CWs, the static box adopts the sectional combined design, which is flexibly designed and composed of a base and a connecting box. A connecting box is a square cube with a bottomless box, made of a chemically stable resistant material. The size of the bottomless cuboid box was 65 cm (length) × 65 cm (width) × 100 cm (height). The bottomless box was covered with a layer of white Styrofoam, further packaged with a yellow plastic film, by which it could prevent the box from the solar radiation and keep the temperature stable during the transfer. Two 12-volt fans for gas fully mixing were installed in the static box. An external connection was provided with a gas sample interface, a circuit interface for fan connection, and a thermometer interface. The gas sample was collected with a 50 mL needle syringe, which was from the sampling hole equipped with a valve situated at the top of the static box. A medical three-way valve was equipped with each needle tube to seal the collected gas in the tube. Carrying handles are mounted on the side of the connecting box. The base of the box, a bottomless square, was made of porous iron material, with a scale of 70 cm (length) × 70 cm (width) × 40 cm (height). A row of round holes was dispersed near the lower half of the base, which was for the root system of plants to pass through. The base was completely immersed in the matrix before the experiment began, and was always inserted in the substrate during the experiment. An electronic digital thermometer mounted on the inner wall was used to measure the temperature of the sampling box.

The procedure of gas sampling was as follows: let the fan on, and then gently collect the first sample immediately with a 50 mL syringe equipped with a three-way valve; standby for 10 min, and then collect the second sample. The first group of samples was completed. After waiting for 2 h, the sampling for the next group was initiated with the same procedure as above. The sampling time was daytime from 8:00 to 18:00. Once collecting gas was done, the gas was transferred into a LabCo headspace sampling bottle by a syringe (12 mL, 839 W Dalian Delin Gas Packaging Co., Ltd., Dalian, China). Bring the sample back to the laboratory immediately to analyze greenhouse gas compositions and concentrations using a gas chromatograph (Agilent 4890D, Agilent Co., Santa Clara, CA, USA).

2.2.3. Gas Detection and Analysis

1. CH₄ and CO₂: Within 48 h, samples were detected with an Agilent 4890 gas chromatograph equipped with an ion flame detector (FID). The operational temperature in the system was as follows: the column temperature was set at 55 °C, and the operating temperature for the detector was set at 200 °C. The CO₂ gas was firstly separated by 60–80 mesh Porapak filler (2 m in length and 2 mm in diameter), which was then reduced to CH₄ by nickel catalyst and detected by FID. The high-purity nitrogen was the carrier gas, with a flow rate of 30 mL/min.

2. N₂O: The N₂O in gas samples was analyzed by an Agilent 7890A gas chromatograph equipped with a 63Ni electron capture detector (ECD). The ECD operating temperature was 330 °C, and the temperature of the column chamber was set at 55 °C. A high-purity nitrogen, with a 99.99% purity, was used as the carrier gas at a 30 cm³ min⁻¹ flow rate. The peak retention time was about 3.5 min for N₂O. The lengths of the pre-column and the analytical column were 1 m and 3 m, respectively. The column had 2 mm of the inner diameter (Porapak Q) made of stainless steel.

2.2.4. The Calculation of Greenhouse Gas Flux

The gas flux was determined by a linear model of the change of gas concentration in the static chamber with time (30 min). The certainty coefficient R² of the regression line was ≥0.80, and the extreme deviation data that existed in a few cases were excluded. Therefore, based on static chamber size, ambient temperature, and atmospheric pressure, the mathematical model shown in Equation (1) is adopted to quantify the gas flux: in the process of gas sampling, each flux value needs at least four gradients of gas concentration to calculate [16]. The formula for calculating greenhouse gas flux is as follows:

$${\mathrm{F}}_{\mathrm{G}\mathrm{H}\mathrm{G}}={\displaystyle \frac{\mathrm{d}\mathrm{c}}{\mathrm{d}\mathrm{t}}}\times {\displaystyle \frac{\mathrm{M}}{{\mathrm{V}}_0}}\times {\displaystyle \frac{\mathrm{T}}{{\mathrm{T}}_0}}\times \mathrm{H}$$

(1)

where F is the greenhouse gas flux, a positive value represents the emission of greenhouse gases (CH₄, CO₂, N₂O) from water to the atmosphere, and a negative value represents absorption. dc/dt: rate of change in greenhouse gas concentration; this value is calculated by linear regression of time based on the concentration of greenhouse gases, which is realized by the formula slope in Excel. M: Molar mass of greenhouse gases. T: absolute temperature at sampling time; the temperature in the sampling box recorded at each use shall be converted into the thermodynamic temperature (Kelvin temperature) by adding 273 Celsius temperature in the calculation process. V₀: molar volume under standard conditions. T₀: absolute temperature under standard conditions; the standard condition here is 0 degrees Celsius, or 273 Kelvin, and H (cm) is the height of the static chamber above the water.

2.3. Global Warming Potential (GWP) in the Wetlands

The intensity of greenhouse gas sources and sinks can be evaluated by their total emissions or absorption of greenhouse gases, combined with GWP, and converted into carbon dioxide equivalent (CO₂-eq) to assess their contribution to global warming [17]. The following Formula (2) was used to calculate GWP in each wetland:

$$\mathrm{G}\mathrm{W}\mathrm{P}=25\times \mathrm{F}\left({\mathrm{C}\mathrm{H}}_4\right)+298\times \mathrm{F}\left({\mathrm{N}}_2\mathrm{O}\right)+\mathrm{F}\left({\mathrm{C}\mathrm{O}}_2\right)$$

(2)

where F (CH₄) and F (N₂O) were methane and nitrous oxide fluxes, respectively. And 25 and 298 were constant coefficients that normalized CH₄ and N₂O into CO₂-eq emissions over a 100-year time range.

2.4. Statistical Analysis

We present these results as a method of replication with standard errors. We used analysis of variance (ANOVA) to measure the statistically significant differences among samples at p < 0.05. We use Pearson correlation analysis to investigate the correlation between environmental parameters, water quality indicators, and greenhouse gas fluxes. In all calculations, the ones with p < 0.05 were considered as significant differences between samples.

3. Results

3.1. Pollutant Removal

3.1.1. COD Removal

The influent COD concentrations in all wetlands, except for CW-5, remained relatively low, not exceeding 33.70 mg L⁻¹ (Figure 1). In contrast, the COD concentration from the influent in CW-5 was significantly higher, with a value of 195.40 mg L⁻¹ Among the wetlands, CW-1 (COD/N = 3) and CW-4 (COD/N = 8) achieved the highest COD removal efficiencies, with removal rates of 70% and 27.56%, respectively (Figure 1). Conversely, CW-2 (COD/N = 3) exhibited the lowest COD removal efficiency, with a removal rate of 12.50%.

With respect to plant types, reed bamboo from CW-1 exhibited the highest COD removal efficiency, followed by reed and green foxtail algae at other wetlands. Geographically, coastal wetlands from CW-1, CW-3, and CW-4 showed higher COD removal efficiencies than inland wetlands, e.g., CW-2 and CW-5. Additionally, the flow type of wetland showed an impact on COD removal performance, with subsurface flow wetlands (mainly based on CW-1) exhibiting substantially higher removal rates than surface flow wetlands (CW-2 and CW-5).

3.1.2. TN Removal

The total nitrogen (TN) concentration in the influent of each wetland remained consistent with a minor reduction throughout the experiment (Figure 1b). For example, even the highest TN removal efficiencies were seen in CW-2 (COD/N = 3) and CW-4 (COD/N = 8) with removal rates of 32.31% and 28.57%, respectively. However, the removal efficiencies of TN were less than 10% in CW-1 (COD/N = 3) and CW-5 (COD/N > 40) with the lowest TN removal efficiencies.

Reed from CW-2 showed the highest TN removal efficiency compared to the other plants employed in the remaining CWs. The high-latitude regions, e.g., CW-2 and CW-4, exhibited higher TN removal rates compared to the low-latitude regions (CW-1, CW-3, and CW-5). In terms of wetland water flow, the TN removal rate of subsurface flow wetlands (CW-2 and CW-3) was generally higher than that of surface flow wetlands (CW-4 and CW-5).

3.1.3. NH₄⁺-N Removal

The ammonia nitrogen (NH₄⁺-N) concentration in the influent for all wetlands, except for CW-5, remained relatively low, ranging from 0.0389 to 0.227 mg L⁻¹. In contrast, CW-5 exhibited a significantly higher influent concentration of 2.2939 mg L⁻¹. In all wetlands, CW-1 (COD/N = 3) achieved the highest NH₄⁺-N removal efficiency, with a removal rate of 91.23%. However, the lowest removal efficiency was observed in CW-5 (COD/N > 40), with a removal rate of only 15.02%. The NH₄⁺-N removal efficiencies in the rest wetlands were comparative.

Regarding plant types, reeds demonstrated the highest NH₄⁺-N removal efficiency in CW-1. In geography, the coastal areas (CW-1 and CW-3) exhibited higher NH₄⁺-N removal rates than the inland regions (CW-2 and CW-5). The significantly higher NH₄⁺-N removal rates were observed from subsurface flow wetlands (CW-1, CW-2, and CW-3) compared to surface flow wetlands.

3.2. CH₄ Flux

The CH₄ emission flux exhibited distinct temporal variations across CWs (Figure 2a). Two peaks were present for CW-1, the highest one at 10 am. One trough followed by one peak was observed for CW-2, and the peak was situated at 10 am, with a significant rise at the end. There was only one blunt and big peak in CW-3, showing between 2 pm and 4 pm. While one peak was situated at the middle of two troughs for both CW-4 and CW-5. Specifically, there was a great rise at the end (18:00) in CW-5. However, the peak points for the respective were at 2 pm and noon. We calculated the average CH₄ emission flux (Figure 2b). The average CH₄ emission flux followed the order: CW-5 > CW-3 > CW-4 > CW-1 > CW-2. Notably, CW-3 and CW-5 exhibited a significantly higher CH₄ flux than the other wetlands, with CW-5 having the highest average flux at 922.47 μg m⁻² h⁻¹ (Figure 2b).

The mean CH₄ flux was negative in CW-1 and CW-2 wetlands (COD/N = 3), indicating that the function of both wetlands was CH₄ sinks. Conversely, the average CH₄ fluxes were positive in CW-3 (COD/N = 5), CW-4 (COD/N = 8), and CW-5 (COD/N > 40), suggesting that they contributed to CH₄ sources. CW-5 showed the biggest CH₄ source with the highest CH₄ flux among them. Additionally, it also demonstrated the increase in COD/N may cause a transition from CH₄ sink to CH₄ source in the wetlands.

Geographically, CH₄ fluxes were generally higher in coastal areas than in inland regions. Regarding vegetation cover, CH₄ flux decreased in the following order: green foxtail algae > reeds > reed bamboo. In terms of wetland type, surface flow wetlands exhibited higher CH₄ fluxes than subsurface flow wetlands.

3.3. N₂O Flux

Figure 3a exhibits distinct temporal variations in the N₂O emission flux across different wetlands. There were two peaks in CW-1, respectively, situated at 12:00 and 16:00. In contrast, there was a single peak for the other CWs; however, the peak time was dissimilar. For example, the peak time for CW-2 and CW-4 was the same, both at 14:00; that for CW-3 was at 10:00. The peak time in CW-5 was delayed, and it was visible after 4:00 p.m. and near 6:00 p.m. The average N₂O emission flux in the different CWs followed the order: CW-1 > CW-3 > CW-2 > CW-5 > CW-4 (Figure 3b). Notably, CW-4 showed a substantially lower N₂O emissions than the other wetlands, with an average flux of 5.13 μg m⁻² h⁻¹. In contrast, the average N₂O emission flux for the other wetlands ranged from 14.39 to 50.54 μg m⁻² h⁻¹.

The mean N₂O flux across all wetlands was positive, indicating that all wetlands contributed to N₂O sources, while the lowest N₂O source was from CW-4. Geographically, N₂O fluxes were generally higher in coastal areas than in inland regions. Regarding vegetation cover, the highest N₂O flux was observed in wetlands dominated by reed and bamboo. In terms of wetland type, subsurface flow wetlands exhibited a generally higher N₂O removal efficiency compared to surface flow wetlands.

3.4. CO₂ Flux

The CO₂ emission curves show that the variation trends of CO₂ in different wetlands were dissimilar. One peak and one trough were present in all CWs except for CW-5. The peak time was at 10:00 for CW-1; however, there was a peak delay for CW-2, CW-3, and CW-4, and their peak times were in the afternoon. Specifically, the peak time for CW-3 was the same as that for CW-4, both at 14:00. In CW-5, there was one peak situated in the middle of two troughs, and the peak time was 12:00. Additionally, a significant rise was observed in CW-5 at the end (18:00). The average CO₂ flux of CW-5 was 56.75 mg m⁻² h⁻¹, which was significantly higher than that of other wetlands (Figure 4b). The CO₂ flux of CW-1 was significantly lower than that of other wetlands, and the average CO₂ flux of CW-1 was −2.29 mg m⁻² h⁻¹. The average CO₂ emission flux in other wetlands ranged from 11.17 mg m⁻² h⁻¹ to 23.68 mg m⁻² h⁻¹. The average CO₂ flux of CW-1 (COD/N = 3) was negative, indicating the function of CW-1 as the sink of CO₂. The average CO₂ flux of the other constructed wetlands, i.e., CW-2 (COD/N = 3), CW-3 (COD/N = 5), CW-4 (COD/N = 8), and CW-5 (COD/N > 40), was positive, suggesting that they contributed to the source of CO₂. The strongest CO₂ source was from CW-5 with the highest CO₂ emission flux. Additionally, it suggested that the increase in the ratio of COD/N lead to the transition of COD flux from sink to source.

Geographically, CO₂ fluxes are generally lower in coastal areas compared to inland regions. Regarding vegetation cover, bamboo-dominated wetlands exhibited the highest CO₂ flux. In terms of wetland types, subsurface flow wetlands typically showed a higher CO₂ removal rate than surface flow wetlands.

3.5. Global Warming Potential (GWP) from Different Wetlands

The ratio of GWP for each CW among all CWs showed the largest proportion, up to 59.51%, in CW-5 (Figure 5a). The second largest was from CW-2 with 16.89%. Both CW-1 and CW-4 exhibited a less than 10% among all CWs, in which CW-1 only stood for 0.95%. The average GWPs for all CWs were over zero, suggesting all CWs result in GHG sources (Figure 5b). The lowest GWP was 1.510 mg CO₂-eq m⁻² h⁻¹ from CW-1, while the highest GWP was 94.874 mg CO₂-eq m⁻² h⁻¹ from CW-5. The contribution of each GHG on the GWP capacity for CWs is shown in Figure 5c. From the perspective of source sink, CW-1 and CW-2 have two kinds of contributions, but the other wetlands are all source contributions. From the point of view of gas, CH₄ accounted for a relatively small contribution in each wetland, except for CW-5, which accounted for a relatively large contribution of methane. N₂O can exist as the second largest contribution gas in the contribution ratio of all wetlands, except CW-1 wetland, where N₂O contribution is the largest. Except for CW-1, other wetlands accounted for the largest proportion of CO₂ contribution as source and the second largest proportion of CO₂ contribution in CW-1, and the contribution type was sink.

3.6. The Correlation of GHG with the Environmental Factors in Constructed Wetlands

In this experiment, temperature was negatively correlated with CH₄ flux (p < 0.05) and CH₄ emission flux was significantly correlated with NH₄⁺-N, COD, and COD/N (p < 0.05) (Figure 6). It has been shown that high COD/N ratios lead to rapid oxygen depletion and a decrease in redox potential, which promotes CH₄ production under sufficient carbon source conditions [18], and therefore the magnitude of COD/N can have a significant effect on CH₄ fluxes. In this study, temperature was negatively correlated (p < 0.05) with N₂O flux (Figure 6), which may be caused by the fact that denitrification takes place in a wide range of temperatures, and too high or too low temperatures inhibit its activity, whereas the influent COD concentration and NH₄⁺-N were positively correlated (p < 0.05) with N₂O flux (p < 0.05), and the influent COD concentration and NH₄⁺-N were positively correlated (p < 0.05); therefore, the magnitude of COD/N will have some effect on N₂O flux. In this study, temperature was negatively correlated with CO₂ flux, and CO₂ flux was correlated with influent NH₄⁺-N concentration, which was significantly and positively correlated with COD concentration (p < 0.05), and influent COD concentration and NH₄⁺-N were positively correlated (p < 0.05) (Figure 4b); thus, the magnitude of COD/N would have some effects on CO₂ flux.

4. Discussion

4.1. Pollutant Removal Efficiency

CWs consume organic matter through plant photosynthesis, respiration, and aerobic and anaerobic biochemical reactions of microorganisms. Most organic matter in sewage is used by heterotrophic microorganisms for their own growth and development and converted into CO₂ and H₂O [18]. Studies have shown that although wetlands are effective and reliable in removing COD, their performance in removing inorganic nutrients from domestic wastewater is often limited and unsatisfactory. Zhao et al. [11] believed that the removal efficiency of COD in the wetland system was higher when COD/N was 5:1, while Chen et al. believed that COD/N of 10 was the best COD/N for COD removal [19]. In our study, the COD removal rate of common reed wetland was not high, while the COD removal efficiency of the CW-1 foxtail algae wetland was up to 70%. These results indicated that actual plant species played an important role in the removal of organic matter differences.

The main ways of denitrification in constructed wetlands were nitrification and denitrification of microorganisms, adsorption and filtration of substrates, nitrogen fixation of plants, and ammonia volatilization, which were also through physical, chemical, and biological synergies [1]. Nitrification and denitrification were the main denitrification mechanisms of continuous wastewater treatment, and carbon source was the key factor of denitrification. The initial COD/N ratio was one of the most critical parameters in the nitrification process because it directly promoted the growth competition between autotrophic and heterotrophic microbial populations. Under different influent COD/N, different wastewater had different removal effects on pollutants.

Among the wetlands studied, those located in coastal areas exhibited better removal efficiencies for COD and NH₄⁺-N compared to wetlands in inland areas. This suggested that air humidity may play a significant role in the removal efficiency of COD and NH₄⁺-N. The constructed wetland with a reed bamboo vegetation cover demonstrated relatively higher removal efficiencies for COD and NH₄⁺-N, although its NH₄⁺-N removal efficiency was lower than that of the other two vegetation types. Analysis of wetland types revealed that subsurface flow wetlands generally exhibited better removal efficiencies for all three pollutants compared to surface flow wetlands. Therefore, when designing and constructing constructed wetlands, subsurface flow wetlands should be prioritized if the removal of these three pollutants is a key objective.

Some studies believe that the TN removal rate in wetland systems increases with the increase in COD/N. Zhu et al. found that when the COD/N was 5, the removal efficiency of TN was the highest, and the removal efficiency increased with the increase in the COD/N [4]. Fan et al. showed that when COD/N was 10, the maximum nitrogen removal rate could reach 90% [8]. This may be due to the abundance of organic matter and nutrients in the water, which provided good nutrient conditions for the growth and enrichment of anaerobic denitrifying bacteria. However, some studies have shown that a lower COD/N ratio can obtain a better pollutant removal effect. The research results of Zhao et al. showed that the removal efficiency of TN was the highest when the COD/N was 2.5–5 [11]. Zhou et al. studied the nitrogen removal response of domestic wastewater with different COD/N. The nitrogen removal efficiency was optimal when the influent COD/N was 1–3, and the nitrogen removal rate decreased when the influent COD/N increased [20]. Microbial processes were the driving force behind nutrient removal in treating wetlands, and plant-mediated microbial processes were the main pathway of nutrient removal in these systems [21]. Plants may also provide carbon to these microbes, which allowed the microbes to survive and remove nutrients more successfully. However, the degradation of excess organic matter will consume more dissolved oxygen in constructed wetlands, thus inhibiting the activity of nitrifying microorganisms. In our study, the optimal COD/N for pollutant removal is 3. When the COD/N ratio of the running water in the wetland systems is close to this value, they all achieve the best results of nutrient removal.

4.2. Effect of Influent COD/N Ratio on GHG Emissions

4.2.1. CH₄ Flux

In general, the final CH₄ flux was determined by the balance of CH₄ production and consumption by CH₄-producing bacteria and CH₄-vegetative bacteria, which was often influenced by organic carbon availability, O₂ availability, and electron acceptors. The main factors of CH₄ production included temperature [22,23], influent COD, ammonia nitrogen, wetland soil moisture content [15], wetland water level, salinity, etc. The higher organic carbon content in soil promoted the microbial activity and further increased the CH₄ flux.

In the optimal temperature range of methanogenic bacteria activity, the increase in temperature was conducive to the microbial decomposition of organic matter more effectively, accelerating the decomposition of organic matter and finally providing substrate for the production of CH₄, and also improving the activity of methanogenic bacteria [24]. There was no uniform conclusion on the relationship between CH₄ flux and temperature in the current research results. Some studies suggested that CH₄ emission flux was closely related to temperature and showed a positive correlation. Previous studies had also suggested that the higher CH₄ flux may be due to the rapid oxygen consumption as the COD/N ratio increases, resulting in a lower REDOX potential. Under conditions of large amounts of available carbon, CH₄ production may be more frequent, resulting in higher CH₄ fluxes [25].

Nutrient loading increased the lake’s native primary production and promoted oxygen consumption and anaerobic decomposition in the sediment, resulting in increased methane release from the lake to the atmosphere [26]. The methane emission rate of wetlands receiving HUSB reactor wastewater was 12 times higher than that of wetlands receiving main sedimentation wastewater, indicating that the higher the COD/N ratio, the greater the methane emission. The presence of plants also reduced methane flux. Chen et al. found that C. indica had a better effect on pollutant removal and a lower methane flux (21.88 ± 2.51 mg m⁻² h⁻¹) in their study on the effect of plant species on greenhouse gas emission and pollution removal in constructed wetlands [27,28]. In the study of Guo et al., the average CH₄ emission flux when COD/N = 3 was significantly lower than that when COD/N = 6 and 9 [13]. However, the average CH₄ flux under each COD/N gradient in their study was higher than that in our experiment, which was only 0.18 mg m⁻² h⁻¹, while theirs was greater than 165 μg m⁻² h⁻¹. Our results also indicated that lower COD/N was associated with lower CH₄ emissions.

The CH₄ emission flux in coastal wetlands was higher than that in inland wetlands, with no significant variation observed along the north–south axis. However, wetlands located in coastal areas generally exhibited higher CH₄ emission fluxes compared to those in inland areas. This suggested that wetlands with higher COD/N ratios tended to have higher CH₄ emission fluxes than wetlands with lower COD/N ratios. Furthermore, differences in COD/N and geographical distribution can influence the variation in CH₄ emission fluxes.

4.2.2. N₂O Flux

It was generally believed that the nitrogen removal in the constructed wetland system was mainly the result of the combined action of the nitrification and denitrification processes with the participation of microorganisms. N₂O was produced in both nitrification and denitrification processes and was generally regarded as an intermediate product and an incomplete product of nitrification and denitrification [29]. There were many factors affecting the N₂O release in wetlands. Pollutant load, carbon-to-nitrogen ratio, REDOX potential, and pH can all affect the composition, quantity, and biochemical reaction process of microorganisms in the system, thus affecting the N₂O release [30,31,32]. During spring and summer, when plant growth and metabolism were vigorous, plant synthesis and respiration were strong, and there were obvious aerobic and anaerobic environmental conditions in the wetland, thus affecting the intensity of nitrification and denitrification in the system. The wetland had a high influent COD, which also promoted N₂O emissions since there was sufficient carbon input to the system to support nitrogen removal. In addition, in an anaerobic environment, the increase in organic matter in wetland and the discharge of local high-carbon and nitrogen sewage can also lead to the increase in N₂O emission. Our experimental results also confirmed this point, showing a positive correlation between influent COD concentration, NH₄⁺-N and N₂O flux (p < 0.05) (Figure 6).

As for the influence of the carbon–nitrogen ratio, Yan et al. showed that when the COD/N was 5, the wetland treatment effect was good and the nitrogen oxide emission was low [10]. In our study, the lowest emission flux of N₂O was at a high carbon–nitrogen ratio, that is, COD/N is 8, which was consistent with the previous research results of Li et al. [33]. This may be due to the fact that NO₂⁻-N accumulation inhibited the growth and activity of nitrous oxide reductase, inducing denitrification or high COD/N ratios in ammonia-oxidizing bacteria, and hypoxia limited nitrification and complete denitrification [34,35]. In our study, the emission of N₂O was lower at a higher carbon-to-nitrogen ratio, while the emission of N₂O was higher at a lower carbon-to-nitrogen ratio. In addition, it was generally believed that higher ambient temperature led to a higher metabolic rate of microorganisms that can promote nitrification and denitrification in wetland system and accelerate N₂O emission. During the period of gas sample collection, the ambient temperature was higher, the microbial activity was stronger, and the emission of N₂O flux was also higher. However, our results showed a negative correlation between temperature and N₂O flux (p < 0.05) (Figure 6). This may be due to the fact that denitrification can be carried out over a wide range of temperatures and be enhanced with increasing temperature, but it can be inhibited by either too high or too low temperatures.

The overall N₂O emission flux in coastal wetlands was higher than in inland wetlands, and the N₂O emission flux in wetlands located in the southeastern region exceeded that in the northwestern region. This suggested that varying land–sea conditions, along with differences in COD/N ratios, played significant roles in influencing N₂O emissions from different wetlands.

4.2.3. CO₂ Flux

Although CO₂ was a major greenhouse gas, few studies had looked at CO₂ emissions from clean water. Constructed wetlands can fix atmospheric CO₂ through plant photosynthesis and enter the atmosphere in the form of CO₂ through plant respiration and microbial oxidation and decomposition. The CO₂ emission flux measured in our study mainly referred to the CO₂ emission flux caused by microbial respiration in water and soil and plant root respiration. Under aeration, plants provided substrates for soil microorganisms through root secretions, which stimulated soil microbial activity and promoted microbial respiration. At the same time, plants transported oxygen to plant roots, forming an aerobic environment, improving the decomposition rate of soil organic carbon, and accelerating the release of carbon-based CO₂ and CH₄ gases [36]. Most of the vegetation in our experimental wetland was foxtail algae, reed, etc., which had relatively developed aerenchyma and had conditions for full release of CO₂.

Temperature was the main influencing factor of the plant growth process, which directly determined the vegetation type and coverage rate in the region. It also affected plant photosynthesis by affecting the enzyme reaction of the dark reaction, which made it another important influencing factor of CO₂ emission flux in wetlands [37]. The results showed that, within a certain temperature range, with the increase in soil temperature, the activities of microorganisms and enzymes in soil decreased, the mineralization of organic matter in soil was weakened, the respiration of roots was weakened, and the CO₂ emission flux of wetland showed a downward trend with the increase in temperature. Our results showed a negative correlation between temperature and CO₂ flux (Figure 6). It was generally believed that wet conditions in wetlands were conducive to the growth of microorganisms, and more CO₂ will be produced through anaerobic respiration, which may be because aerobic respiration was more effective than anaerobic respiration for CO₂ emission. Our experiment period was the plant growth period, and higher plant biomass and CO₂ flux may be the function of increased bacterial activity in growing crops, which was the function of plant secretions to obtain more unstable carbon. The reason why both CH₄ and CO₂ were negative may be that the amount of CO₂ fixed by plant photosynthesis was higher than that produced by plant and microbial respiration, resulting in negative net CO₂ exchange. The study of Chen et al. confirmed that the CO₂ flux increased with the increase in influent COD/N ratio, which was also confirmed by our experimental results [19]. In addition, it was also found in this study that CO₂ flux was correlated with influent NH₄⁺-N concentration and significantly positively correlated with COD concentration (p < 0.05) (Figure 6).

The CO₂ emission flux in coastal wetlands was lower than in inland wetlands, and the CO₂ emissions in wetlands located in the southeastern region were lower than those in the northwestern region. This suggested that wetlands with higher COD/N ratios tended to have higher CO₂ emissions than those with lower COD/N ratios. Additionally, it indicated that land–sea factors may play an important role in CO₂ emissions across different wetlands.

4.2.4. GWP

It was important to note that high greenhouse gas emission fluxes did not represent a high greenhouse effect. The magnitude of GWP can be used to reflect the capacity of the greenhouse effect. The fluxes of three greenhouse gases were measured in this study, so it was necessary to assess the total greenhouse gases in different greenhouses to elucidate the impact of greenhouse gas emissions on greenhouse mitigation under different influent COD/N conditions. CH₄ and CO₂ emissions were the main factors affecting the warming potential. In the past, many studies ignored N₂O emissions from wetlands, but because of the high GWP value of N₂O, the impact of N₂O on global warming was quite significant. In this study, although the N₂O emission flux was lower than the other two greenhouse gases, GWP (298) N₂O was the highest. A large amount of nitrogen-containing nutrients flowing into the wetland system may lead to a large amount of N₂O produced by the system and a higher carbon source, thus enhancing the greenhouse effect. Yang et al. suggested that nitrogen content and environmental conditions suitable for the production of nitrous oxide by microorganisms may be the main factors for the transformation of the wetland system from a carbon sink to a carbon source [38]. In Chen et al.’s study, the reduction in greenhouse gases from biochar was mainly manifested in the reduction in nitrous oxide flux, while methane flux was promoted and carbon dioxide flux was not affected [39]. Studies had been conducted to mitigate the effects on GWPs by adding biochar, biochar matrix, tidal flow, and intermittent aeration modes [40], walnut shell, manganese (Mn), ore and activated alumina [18,25]. It was worth noting that in our study, CH₄ fluxes were positively correlated with CO₂ and N₂O fluxes (p < 0.05) Figure 6), while CO₂ and N₂O fluxes were only correlated with CH₄ fluxes. This will enlighten us to consider collaborative emissions reductions when controlling greenhouse gas emissions in the future.

The influent COD/N ratio was 3, which was the best choice to achieve relatively high pollutant removal efficiency and low greenhouse gas emissions at the same time [11,19]. In addition, through the introduction of oxygen by artificial aeration, the selected vegetation, especially the catypuff, can obtain the lower GWP value, the maximum nutrient removal rate [28]. It had also been suggested that VSSFS were an effective option for mitigation of greenhouse gas emissions from the wastewater sector, especially in developing countries such as China, where VSSFS should be widely established for decentralized wastewater treatment, which may also reduce greenhouse gas emissions by 8–107 million tons of CO₂ per year compared to centralized schemes [41].

The accuracy of our study may still be limited by some inherent uncertainties. General conclusions about wetland greenhouse gas fluxes may be inherently uncertain due to site-specific environmental factors. In our study, the use of chamber-based methods for collecting and measuring gas fluxes may have limitations or disadvantages in large spatial scale experiments. Therefore, other techniques, such as continuous eddy covariance [42], the NDIR technique [43], and high-frequency soil oxygen sensors for predicting wetland greenhouse gas emissions [44], should be used to compare and complement the data from the chamber method in larger spatial scale studies. High heterogeneity in the data set, including changes in sampling time, experiment duration, and vegetation type, further increases uncertainty in greenhouse gas estimates. Therefore, our results needed to be validated with additional data from standardized, continuous field observations of wetlands distributed over larger spatio-temporal ranges and scales to improve quantification of spatio-temporal variations in greenhouse gas fluxes and accurate assessment of long-term source–sink effects. In addition, plant productivity was an important factor that needs to be taken into account in follow-up studies [21,45]. The results of our analysis will contribute to understanding the key drivers and mechanisms of greenhouse gas fluxes between wetlands that treat urban tailwater. However, more detailed studies of the spatio-temporal coupling and the key drivers controlling greenhouse gas fluxes are needed.

The GWP values of wetlands in inland areas are higher than those in coastal areas, with western regions exhibiting higher values than eastern regions. This suggests that wetlands with higher COD/N ratios tend to have higher GWP values compared to those with lower COD/N ratios. Additionally, it indicates that geographical location has a significant impact on the magnitude of GWP values.

5. Conclusions

This study investigated the effects of varying influent COD/N ratios and sea–land distribution conditions on pollutant removal and GHG emissions in constructed wetlands. The results indicate that temperature, ammonia nitrogen, COD, COD/N ratio, and geographical distribution are the primary factors influencing GHG fluxes. Temperature exhibited a negative correlation with GHG emissions, while ammonia nitrogen, COD, and COD/N were positively correlated. Further analysis revealed that COD/N significantly influenced the emission fluxes of all three GHGs, making it a key determinant of GHG emissions in constructed wetlands. GHG fluxes were significantly higher in coastal areas than in inland regions. CH₄ was positively correlated with both CO₂ and N₂O, whereas no significant correlation was observed between CO₂ and N₂O. This lack of correlation suggests the need for further research on potential synergistic effects in GHG emission reduction or control strategies.

Constructed wetland technology did not have a significant impact on GHG emissions. Among the wetlands studied, those with a COD/N ratio of 3 demonstrated the highest overall efficiency in pollutant removal and GHG mitigation. Additionally, wetlands with lower COD/N ratios and those located in coastal areas were more effective in both pollutant removal and GHG reduction. These findings highlight the importance of considering influent COD/N ratios in the design of constructed wetlands for municipal wastewater treatment to optimize pollutant removal while minimizing GHG emissions.