Treatment Effect of Long-Term Subsurface-Flow Constructed Wetland on Mariculture Water and Analysis of Wetland Bacterial Community

The purification effect of organic matter and suspended particulate matter in the subsurface-flow wetland is shown in Figure 2. The average concentration of COD in the influent was 3.25 (±0.51) mg/L. According to China’s water standard GB 3097-1997 as a reference [15], the seawater belonged to the third category of water quality, which was not suitable for aquaculture without treatment (Table 1). The average concentration of COD in the effluent treated by the subsurface-flow wetland was 2.60 (±0.50) mg/L, and the water quality was upgraded to the second category. The average removal amount was 0.96 (±0.11) g/m²·d⁻¹, and the average removal rate was 20.29 (±2.88) %. In the subsurface-flow wetland system, the physical processes of precipitation, matrix adsorption, filtration, and microbial biodegradation were the main ways organic matter in seawater was removed [16]. As an important evaluation index of water quality monitoring, COD could reflect the organic pollutants in water. Because the influent COD content was low, the system treatment effect was not obvious.

The clogging problem has always been an important factor affecting the long-term stable operation of wetland systems [17]. Therefore, to prevent the reduction in its water conductivity and removal effect, the system had been cleaned and maintained regularly every year since its establishment and operation. Previous studies have suggested that the maximum TSS load of a wetland system should be 15 g/m²·d⁻¹ [18]. The removal amount of TSS was 12.27 (±2.74) g/m²·d⁻¹, the average removal rate was 49.33 (±12.75) %, and the average concentration of effluent was less than 10 mg/L. The removal amount was 14.91 and 15.21 g/m²·d⁻¹ in June and October, respectively.

Excessive nitrogen may affect the normal physiology and behavior of aquatic organisms [19]. It can be seen from Table 2 that the concentration of TN in the influent and effluent was 0.92–2.06 mg/L and 0.72–1.15 mg/L, respectively. The average removal amount was 0.78 (±0.43) g/m²·d⁻¹, and the average removal rate was 36.94%. The concentration of DIN (dissolved inorganic nitrogen) in the influent was 0.19–0.29 mg/L, and the effluent was 0.17–0.24 mg/L. The average removal amount and removal rate of DIN was 0.04 g/m²·d⁻¹ and 10.88%. The concentration of DON (dissolved organic nitrogen) in the influent and the effluent was 0.69–1.77 mg/L and 0.51–0.93 mg/L. The average removal amount was 0.74 g/m²·d⁻¹, and the average removal rate was 41.82%. The removal rate of DON in the subsurface-flow wetland characterizes the mineralization intensity of organic nitrogen, and the removal rate was positively correlated with temperature [20]. In April, the total nitrogen concentration of the influent and effluent of the system was 0.92 (±0.02) mg/L and 0.75 (±0.05) mg/L, respectively, and the removal rate was only 17.8%. Due to higher temperatures and more active microbial activity from June to October, the TN removal rate remained at a high level, with the highest removal rate at 44.08%. The main ways of nitrogen removal in the subsurface-flow wetland system were ammonia volatilization, denitrification, deposition, and adsorption [21,22]. Related studies have shown that subsurface-flow wetland systems are suitable for the removal of nitrogen pollutants in rivers and lakes, and could maintain a high TN removal rate even under long-term operation [23].

The form of nitrogen in the influent was mainly DON, which accounted for 75.68–89.26%, while DIN accounted for 10.74–24.32% (Figure 3). After subsurface-flow wetland treatment, the proportion of organic nitrogen decreased, the proportion of inorganic nitrogen increased, and DIN in the effluent was mainly NO₃⁻-N. The subsurface-flow wetland system had the most significant effect on the removal of DON, while for DIN, NH₄⁺-N decreased from 0.11 (±0.03) mg/L to 0.06 (±0.02) mg/L and NO₃⁻-N increased from 0.09 (±0.03) mg/L to 0.14 (±0.02) mg/L. It is generally believed that NO₂⁻-N has high biological toxicity to aquatic organisms. Elevated NO₂⁻-N can hinder biological growth and even induce death in aquatic organisms [24]. NO₂⁻-N decreased from 0.028 (±0.021) mg/L to 0.006 (±0.003) mg/L during the monitoring period, and the average removal rate reached 78.54% (Figure 4).

The subsurface-flow wetland system mainly achieved the removal of phytoplankton through precipitation, matrix adsorption, physical interception, flocculation precipitation, and microbial degradation [25]. The dominant species in influent and effluent were Microcystis (30,577.29 × 10³ ind./L), Synechocystis (2673.70 × 10³ ind./L); Phaeodactylum (24,030 × 10³ ind./L), Melosira (135.30 × 10³ ind./L), and Chlorophyta Dunaliella (148.69 × 10³ ind./L) (Table 3).

The effect of the subsurface-flow wetland system on algae removal is shown in Figure 5. The wetland system had a significant removal effect on Cyanophyta and Pyrrophyta, and the average removal rates were 92.90% and 97.85%, respectively. The density of Cyanophyta in the influent increased first and then decreased with the seasons, while the removal rate of Cyanophyta in the subsurface-flow wetland system was stable. The density of Cyanophyta in the influent reached its highest value in June, 129.60 × 10⁶ ind./L, and decreased to 3.64 × 10⁶ ind./L after treatment by the subsurface-flow wetland (Figure 5a). The density of Pyrrophyta in the influent was low. In October and December, during the treatment by the subsurface-flow wetland system, the densities of Pyrrophyta were 0.41 × 10⁶ ind./L and 0.66 × 10⁶ ind./L, which were reduced to 0.02×10⁶ ind./L and 0.003 × 10⁶ ind./L, respectively (Figure 5b). The density of Bacillariophyta in the influent increased first and then decreased with the change in seasons. The density of algae was the highest in June, with an average of 105.61 × 10⁶ ind./L. The density of Bacillariophyta decreased to 0.26 × 10⁶ ind./L after subsurface-flow wetland treatment. When the Bacillariophyta density in the influent was low, the Bacillariophyta density in the effluent increased. When the density of Bacillariophyta in the influent was high, the removal efficiency of Bacillariophyta in the subsurface-flow wetland system was stable, and the average removal rate was 97.24% from June to December (Figure 5c). The density of Chlorophyta in the influent was low. From April to June, the density of Chlorophyta in the influent was less than 0.10 × 10⁶ ind./L. After treatment in the subsurface-flow wetland, the density of Chlorophyta increased. After subsurface-flow wetland treatment, the density of Chlorophyta clearly decreased, and the average removal rate was 84.17% (Figure 5d).

Currently, chemical measures are usually used to regulate the planktonic state of the water, such as the addition of copper sulfate [26,27]. Although the treatment effect is good, there are chemical residues and cost problems. From the point of view of treatment effect, a constructed wetland is a better alternative scheme, which has a better removal effect on cyanobacteria and greatly reduces the operation cost. However, since the removal of harmful algae involved a series of biological and physicochemical processes, it is difficult to determine which pathway plays a major role in the system.