Seasonal Distribution of Nutrient Salts and Microbial Communities in the Pearl River Delta

Abstract

The transformations of iron (Fe), phosphorus (P), and sulfide (S) have been previously investigated in many areas, but quantifying the effects of the seasons on nutrient transformations and bacterial community distributions is a major issue that requires urgent attention in areas with serious anthropogenic disturbance. The authors used the diffusive gradients in thin films (DGTs) technique and 16S rRNA gene sequencing to determine the spatial heterogeneity in the nutrient distribution and bacterial community structure in the overlying water and sediment in the Pearl River Delta (PRD). Sampling campaigns were conducted in summer and winter. The results show that the nutrient salts exhibited greater differences in time than in space and there were higher water pollution levels in winter than in summer. During summer, the abundant non-point source pollution from the rainfall input provided a rich substrate for the bacteria in the water, leading to a strong competitiveness of the PAOs and nitrifying bacteria. Meanwhile, a high temperature was favorable for the exchange of elements at the SWI, with a greater release of P, Fe, and N, while, with the low temperatures and high DO and nutrient salts seen in winter, the SOB and denitrifying bacteria were active, which correctly indicated the high concentration of SO₄²⁻ and NH₄⁺-N in the water. The microbial diversity and abundance were also affected by the season, with a higher richness and diversity of the microbial communities in summer than in winter, and the high salinity and nutrient salt concentration had a significant inhibitory effect on the microorganisms. A Mantel test revealed that the spatiotemporal distribution patterns of the dominant bacteria were closely related to the TOC and DO levels and played an important role in the P, Fe, S, and N cycle. These observations are important for understanding the nutrient salt transformation and diffusion in the Pearl River Delta.

1. Introduction

River deltas are important links between terrestrial and marine ecosystems; they are characterized by dense water networks, strong tidal action, high population densities, and intensive development [1]. The Pearl River Delta (PRD) in southern China is the fourth largest river network area in the world. Due to climate change and anthropogenic pollution, the concentrations of nitrogen (N) and phosphorus (P) in the coastal waters have intensified, resulting in environmental pressure on the Pearl River [2,3]. These contaminants have a negative impact on the biotic components of the water and sediment, resulting in a eutrophication state and algal bloom events [4].

Nutrient salts in overlying water can accumulate in sediments through adsorption, flocculation, and sedimentation, making sediments an important reservoir for many pollutants [5]. In the absence of external inputs and effective control of N and P discharges, endogenous pollutants in sediments may be released in a continuous stream due to diffusion, bioturbation, and other effects [6,7]. Natural factors such as temperature, rainfall, wind, and waves, as well as human activities such as ship navigation, may affect the physicochemical properties of water and sediments and further control nutrient release, especially in river deltas [8]. The concentration of nutrients in water bodies is negatively correlated with the dissolved oxygen content of the water. When the content of N, P, and other nutrients in the water body is too high, under anoxia-driven reducing conditions, sulfate (S) reduction produces H₂S and dissimilatory nitrate reduction to ammonium, accompanied by the conversion of Fe (III) to Fe (II), generating FeS, which releases adsorbed phosphorus into the water, which is why Fe is widely recognized as an important variable controlling the release of P [9]. At the same time, changes in the concentrations of N and P in the reaction further affect the concentration of nutrients in the water body, indicating that there is a strong coupling relationship between P, Fe, N, and S [10]. Therefore, the study of P, Fe, S, and N distribution characteristics and diffusive fluxes in sediments is crucial to understanding the effects of seasonal distribution on nutrient distribution.

Microorganisms are an important component of river ecosystems, and their community composition and function often change with the accumulation of nutrient salts and chemical pollution caused by human activities [11,12,13]. Furthermore, microbes play important roles in energy conversion and material cycling, thereby maintaining the balance of river ecosystems [14,15]. Organisms such as Sulfate-Oxidizing Bacteria (SOB), Sulfate-Reducing Bacteria (SRB), and phosphate-accumulating organisms (PAOs) play an important role in the sulfur cycle and biological phosphorus removal system [16]. Iron-Oxidizing Bacteria (IOB) and Iron-Reducing Bacteria (IRB) are able to control the release and cycling of Fe. It has been found that there is competition between IRB and SRB in a limited number of substrates, and that S²⁻ produced by SO₄²⁻ reduction has an inhibitory effect on IRB, thus further controlling Fe reduction [17]. This process influences the redox conditions of water under aerobic and anoxic conditions and also influences the activities of PAOs and nitrifying and denitrifying bacteria, suggesting interactions between these three components [18,19]. Therefore, exploring the microbial pathways of coupled P, Fe, N, and S cycling in the sediment–water interface (SWI) from a microbial perspective is key to elucidating potential endogenous release mechanisms in the water and sediment.

In this study, the authors selected 14 sites in the PRD, located in upstream freshwater areas and downstream high-salinity areas (Figure 1). As summer and winter had completely opposite environmental conditions based on long-term water quality monitoring, such as DO (dissolved oxygen), water level, and temperature, ultimately leading to differences in the bacterial community, the two sampling seasons were selected to represent a large time span. The diffusive gradients in thin films (DGT) technique, in conjunction with 16S rRNA gene sequencing, was applied to explain the interaction between the nutrient salts and bacterial community structure at nine representative sites. The objectives of the present study were (1) to quantify vertical nutrient salts distribution and diffusion fluxes in river deltas, (2) to investigate the bacterial community structure in the overlying water, and (3) to evaluate the effects of seasons and human factors on nutrient salts and the bacterial community structure.

2. Material and Method

2.1. Study Area and Sample Collection

The Pearl River is located in the southern part of China and has three major tributaries, namely, the Xijiang River, the Beijiang River, and the Dongjiang River, along with various rivers of the Pearl River Delta. The total basin area is 453,690 km², the total catchment area is 467,348 km², with an annual runoff of approximately 282.6 × 10⁹ m³, and the annual rainfall level is 1500–2000 mm, indicating that the PRD is an important water resource for the high-quality development of the PRD urban agglomeration [9]. The PRD, which contains eight major estuaries and 324 networked river channels, is characterized by intense economic activities and a dense human population. In 2022, the resident population was about 78.294 million, accounting for 61.9% of the population of Guangdong Province, and the gross domestic product (GDP) contributed 81.1% of Guangdong’s GDP [20]. Human activities such as sand mining in estuaries, channel dredging, and reclamation by building embankments have an important influence on water and sediment fluxes, the shoreline, and the topography. The river systems in the Pearl River Basin merge together, flowing westwards to the South China Sea through the eight major gates of Humen, Jiaomen, Hong Qimen, Hengmen, Mo Daomen, Ji Timen, Hu Tiaomen, and Ya’men.

The sediments of most rivers have been dredged, so it is very difficult to collect undisturbed sediments sampling sites. Therefore, 14 monitoring sites (sample collection locations) in the PRD were sampled in winter (December 2021) and summer (August 2022). Sampling sites S1, S2, S12, S13, and S14 were located in an upstream area and sites S2, S4, S5, S6, S7, S8, S9, S10, and S11 were located in a downstream area; salinity was >5‰ for the high-salinity regions and salinity was <5‰ for the freshwater regions. The upstream area was far from residential areas and pollution sources, while the downstream area has experienced a high intensity of anthropogenic pollution. In addition, the data for the river flow velocity and water quality monitoring in the past were used as a background reference. At each sampling point, 2 L of water samples and 2 kg of sediment samples were collected and immediately transported to the laboratory, where surface water from nine national sites was filtered through a 0.2-μm membrane for bacterial community structure analysis, and the rest was stored at ~4 °C until testing.

2.2. Determination of Physico-Chemical Parameters

The dissolved oxygen (DO, mg‧L⁻¹), water temperature (T, °C), and electrical conductivity (EC, us/cm) of the water body were measured in situ by a YSI ProDSS Multiparameter Digital Water Quality Meter. A SonTek M9/S5 intelligent multi-frequency Doppler current profiler was employed to measure the river flow velocity and water depth. Furthermore, samples of the overlying water and sediment were collected for the analysis of nitrate (NO₃⁻-N, ultraviolet spectrophotometry), chemical oxygen demand (COD, dichromate method), sulfate (SO₄²⁻, atomic absorption spectrometry), ammonia nitrogen (NH₄⁺-N, Nessler’s reagent spectrophotometry), 5-day biochemical oxygen demand (BOD₅, dilution and seeding method), total phosphorus (TP, molybdenum blue spectrophotometry), and chlorophyll a (Chla, spectrophotometry).

2.3. DGT Analysis

The double-sided DGT technique (AMP-TH and ZrO-CA) was used to measure NH₄⁺, NO₃⁻, PO₄³⁻, Fe²⁺, Mn²⁺, and SO₄²⁻ at the SWI. This technique requires placing the DGT in a sediment culture column (a culture column that mimics a sediment–water interface chamber) and exposing it to the overlying water at a height of about 2 cm for 24 h, during which time the temperature of the overlying water is measured periodically. After the removal of the DGT, the pellet is rinsed with distilled water, the SWI is measured and labeled, and the sample is placed in a humid sealed bag and stored in a cool, dark place. This method is mainly based on Fick’s first law of diffusion, simulating the diffusion and migration of target ions in the natural environment, and in this way, the accumulation of target ions per unit time and the diffusion flux of the SWI could be calculated.

The DGT-S fixed film was scanned with a flatbed scanner, and the ImageJ software (Version 1.53) was used to convert the scanned image to the corresponding grayscale. Using the established calibration curve, the accumulated DGT-S was calculated according to Equation (1) [21]:

$$y=-171e^{-x/7.23}+220$$

(1)

where x is the accumulated amount of ions within the retention time (μg/cm²), and y is the grayscale.

The accumulated amounts (M) of DGT-P, DGT-Fe, DGT-Mn, DGT-NH₄⁺, and DGT-NO₃⁻ were calculated using Equation (2), and the final effective DGT concentration (cDGT) was determined using Equation (3), while the P and N diffusion fluxes at the SWI can be estimated using Equation (4), according to Fick’s first law [22]:

$$M=c_{e}V/f$$

(2)

$$c_{DGT}=M\Delta g/DAT$$

(3)

$$F=F_w+F_s=-D_w\left({\displaystyle \frac{\partial C_w}{\partial x_w}}\right)+-\phi D_s\left({\displaystyle \frac{\partial C_s}{\partial x_s}}\right)$$

(4)

where M is the accumulated amount of target ions on the immobilized membrane (μg), c_e is the concentration of target ions in the extract (μg‧L⁻¹), V is the volume of the extract (mL), f is the extraction rate, c_DGT is the effective concentration of the target ions during the retention time (mg‧L⁻¹), $\Delta$g is the thickness of the diffusion layer (cm), D is the diffusion rate of the target ion to be measured in the diffusion layer (cm²‧s⁻¹), A is the effective area of the fixed membrane (cm²), T is the placement time (s), F_w is the diffusion flux of the target ions from the sediment to the overlying water, F_s is the diffusion flux of the target ions from the overlying water to the sediment (μg‧(cm²‧d)⁻¹), φ is the porosity of the sediment, (∂C_w)/(∂x_w) and (∂C_S)/(∂x_S) are the concentration gradients at the SWI, and c is the diffusion coefficient in the sediment, calculated by φ (when φ < 0.7, D_s = φD₀; when φ ≥ 0.7, D_s = φ²D₀, where D₀ is the ideal solution diffusion coefficient). At 25 °C, the ideal diffusion coefficients of PO₄³⁻, NH₄⁺, NO₃⁻, SO₄²⁻, and Fe²⁺ are 6.12 × 10⁻⁶, 19.8 × 10⁻⁶, 19.0 × 10⁻⁶, 10.7 × 10⁻⁶, and 7.19 × 10⁻⁶ cm²‧s⁻¹, respectively.

2.4. Bacterial Community Analysis

Microbial DNA was extracted from 54 surface water samples (2 seasons × 9 sites × 3 replicates) using the E.Z.N.A.^® Soil DNA Kit (Omega Bio-tek, Norcross, GA, USA) according to the kit instructions. The quality and concentration of the extracted DNA were inspected by agarose gel electrophoresis and a Nanodrop One spectrophotometer (Thermo Fisher Scientific, Waltham, MA USA). To characterize the bacterial community structure and composition of the water, the V3–V4 variable region of the 16S rRNA gene was amplified using universal primers 338F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806R (5′-ACTCCTACGGGAGGCAGCA-3′). Subsequently, paired-end sequencing (2 × 250 bp) was conducted on the Illumina MiSeq platform (Shanghai BIOZERON Co., Ltd., Shanghai, China) to obtain the sequencing data. Paired original sequences were joined after filtering adaptor sequences to generate clean sequences. Finally, Quantitative Insights Into Microbial Ecology 2 (QIIME2) were used for quality control and bioinformatic analysis. The high-quality sequences were clustered into operational taxonomic units (OTUs) with a 97% similarity threshold, and the taxonomic information for each sequence was annotated based on the Silva 16S rRNA database (V138). The alpha diversity was calculated by QIIME2 to estimate the species richness and diversity of the bacterial communities, including Chao1 and Shannon indices. A Mantel test was performed to explore the correlation between the environment and the bacterial communities (Vegan package). Functional Annotation of Prokaryotic Taxa (FAPROTAX) was used to predict the ecological functions of the bacterial communities.

3. Results and Discussion

3.1. Spatiotemporal Changes in Water Quality

Spatiotemporal variations in seven water quality physicochemical factors were analyzed (Figure 2). On the time scale, there were significant differences in DO, NH₄⁺, TP, TN, NO₃⁻, and TOC, except for BOD₅; the average water temperature in summer was 30.21 °C compared to 18.87 °C in winter; the average DO concentration in summer was 5.16 mg‧L⁻¹, lower than that in winter (8.01 mg‧L⁻¹); and the average concentrations of DO, BOD₅, NH₄⁺, TN, NO₃⁻, and TOC were higher in winter than in summer. The water quality was more favorable in summer than in winter, suggesting that the increased amount of rainfall during summer has a stronger impact on the dilution of pollutants in rivers than the non-point source pollution [23].

Spatially, heavy pollution by domestic and industrial sewage accumulates from the upstream to the downstream of the river, and the DO, TP, and TOC average concentrations showed a decreasing trend, while a gradually increasing trend was observed in BOD₅, NH₄⁺, TN, NO₃⁻, and TOC, translating into heavy environmental pressure on the PRD estuary region [4]. Most high-salinity areas were located in the middle and downstream areas, and the average concentrations of BOD₅, NH₄⁺, TN, NO₃⁻, and TOC in the high-salinity regions were higher than those in the freshwater regions. Among the 14 sampling points, the water quality of the upstream sites S2, S13, S12, and S14 was better than that of the downstream sites. The PRD is an important hub for economic development and human activities in China; due to population growth and economic development in the PRD region, algal bloom, caused by high nutrient salt levels, occurs frequently, and anoxic water areas are formed [24]. Sites S4, S5, S6, and S8 were typical anoxic sites, where the TN, NO₃⁻, and SO₄²⁻ concentrations reached their peak levels. Overall, in the study area, the difference between the physicochemical parameters of the water in summer and winter was greater than the difference between the different sampling points.

3.2. Spatiotemporal Changes in Sediment Composition

The surface sediment was composed of sand, silt, and clay particles, with significant seasonal variations (Figure 3). The sediment was dominated by sand particles at the time of both sampling events, and the average sand content in summer was 83.5%, which is higher than the 67.8% found in winter, which may be related to the increased surface runoff and the coarsening of the sediment caused by the rainfall during the summer. Previous studies have shown that the P adsorption capacity of sediments is closely related to OM: when the rainfall increases in summer, OM is carried into the sediment, and the TP concentration in sediment also increases; the average TP and OM concentrations in the sediment in summer (899.714 mg‧kg⁻¹ and 267 35.714 mg‧kg⁻¹) are approximately twice those found in winter [25]. The content of silty sand and clay particles in winter was higher compared to that found in summer, which is more conducive to the adsorption capacity of sediment, which can effectively adsorb organic carbon and nitrogen from the water [26,27]. That may be the reason why the TN in winter (789.929 mg‧kg⁻¹) was higher than in summer (481.12 mg‧kg⁻¹).

The sediment contained more silt and clay particles in winter, especially at points S7, S8, S9, and S10, as they were located in the estuary of the river, which made them more susceptible to channel narrowing, concentration of tidal forces, and stronger erosion by water flow [28]. It is now generally accepted that nutrient salts, heavy metals, and organic pollutants in overlying water can accumulate in sediment through adsorption, flocculation, and sedimentation, making sediment an important reservoir for many pollutants [29,30]. Since the structure and diversity of bacterial communities in estuarine sediments are closely related to contaminant concentrations, bacterial communities become an useful indicator of sediment health [31]. In our study, the bacterial community structure in the water was combined with previous reports on surface sediments to obtain an accurate representation of the transformations of nutrient salts in an estuarine area.

3.3. Vertical Distribution of DGT-P, Fe, S, and N

The high-resolution vertical distribution of DGT-P, Fe, S, and N at the SWI is illustrated in Figure 4, and the concentrations of PO₄³⁻, Fe²⁺, and SO₄²⁺ in the water and sediment showed a gradual increase with increasing depth. The exchange in phosphorus between the sediment and water involves the precipitation and dissolution of phosphates, the subsidence and resuspension of inorganic phosphorus particles, and the adsorption and desorption of dissolved organic phosphorus; these processes occur simultaneously [32]. Fe is a redox-sensitive element, and in an anoxia-driven reducing environment, Fe (III) is transformed into Fe (II), causing the adsorbed P to be released into the water; this is why Fe is regarded as an important variable in controlling P release [33]. In addition, the sulfide-mediated chemical iron reduction process is a major mechanism for the internal P release rate [34]. Therefore, there are strong correlations between the internal P, Fe, and S distribution [10].

The wind and waves, ship navigation, and industrial activities, along with rainfall, may have an impact on the nutrient salt concentration and distribution in the overlying water and sediment, especially in river deltas [35]. The tidal velocity of the fluctuation in summer (0.384 m⁻¹, 0.554 ms⁻¹) was significantly higher than that in winter (0.315 m⁻¹, 0.294 ms⁻¹), which may be relatively unstable due to the change in flow velocity, resulting in the concentrations of P, Fe, and S in the pore water in summer (0.208, 3.387 and 0.108 mg L⁻¹) being slightly lower than those in winter (0.189, 3.021 and 0.059 mg L⁻¹). At the same time, the higher the average temperature in summer, the lower the average DO concentration (30.21 °C, 5.87 mg L⁻¹) in winter (19.1 °C, 8.13 mg L⁻¹), which may be because the anaerobic environment is conducive to the release of Fe-P from the sediment.

Nitrogen transformation mainly occurs through nitrification and denitrification; nitrification means the oxidation of ammonia to nitrate via nitrite, linking biological nitrogen fixation with anaerobic denitrification [7]. The NH₄⁺-N concentration was low in the overlying water, but it gradually increased with depth from the SWI and exhibited a certain degree of fluctuation, while the NO₃⁻-N concentration gradually decreased with depth and stabilized at 2 cm below the SWI, which attributed to the decrease in DO concentrations and the proneness to anaerobic ammonium oxidation under anoxic conditions, with NO₃⁻-N being converted into NH³⁺-N [36]. Because the average chlorophyll a concentration in summer (7.67 mg‧L⁻¹) was considerably higher than that in winter (1.61 mg‧L⁻¹), the algae consumed more NH₄⁺-N in summer, the average NH₄⁺-N concentration in pore water in winter (0.372 mg‧L⁻¹) was higher than that in summer (0.226 mg‧L⁻¹), and the average NO₃⁻-N concentration in summer (0.686 mg‧L⁻¹) was higher than that in winter (0.223 mg‧L⁻¹). This process consumed a large amount of oxygen in the water and sediments, leading to the formation of low-oxygen or even hypoxic areas in the water bodies in summer.

3.4. Diffusion Fluxes of P, Fe, N, and S at the SWI

The pollutants in surface sediments are highly reactive and can be released across the SWI into the overlying water by gradient diffusion, resuspension, and biological disturbance [37]. To understand the pollutants’ budget balance and material cycling, the nutrient salt fluxes at the SWI were explored. The diffusion fluxes of P, Fe, and N at the SWI are shown in Figure 5. A positive flux indicates that nutrient salts are released from the sediment into the overlying water, whereas a negative flux indicates that nutrient salts are released from the overlying water into the sediment. In this study, PO₄³⁻, Fe²⁺, and NH₄⁺-N were released from the sediment into the overlying water, whereas NO₃⁻-N was released from the overlying water into the sediment, which confirmed that the sediment tended to be a reservoir of PO₄³⁻, Fe²⁺, and NH₄⁺-N [38]. The release intensity of PO₄³⁻, Fe²⁺, and NO₃⁻-N was much higher in summer than in winter, indicating the existence of an oxidation layer at the SWI in winter that prevented PO₄³⁻, Fe²⁺, and NO₃⁻-N from being released into the overlying water. However, during the anoxic periods in summer, the oxidation layer was thinning, allowing the release of PO₄³⁻, Fe²⁺, and NO₃⁻-N from the sediment [39]. The PO₄³⁻ and Fe²⁺ diffusion fluxes at S4, S5, and S6 were significantly higher than those at the other sites, which might be related to large amounts of industrial pollution emissions and the resuspension of partial sediments near estuaries.

3.5. Spatiotemporal Variations in Diversity of Microbial Communities in Water

Seasonality is a common feature in ecosystems, and seasonal changes cause corresponding changes in environmental factors, ultimately leading to variations in the bacterial community structure [40]. In this study, 16S rRNA gene sequencing was used to analyze the diversity of the bacterial communities in the overlying water at the nine representative points in summer and winter. Each sample was measured three times and purified by agarose gels. Clustering and comparison revealed that the average sequence length was 401–500 bp, and the sequences were divided by 97% sequence similarity; 24,739 OTUs were obtained. In total, 62 phyla were identified in the overlying water, and the number and composition of the bacterial communities varied considerably among the samples.

At the phylum level (Figure 6a), Proteobacteria were the dominant group, accounting for 40% of the total abundance, followed by Actinobacteriota, Cyanobacteria, Bacteroidota, and Chloroflexi, accounting for 29%, 18%, 3%, and 2%, respectively. The abundance of both Firmicutes and Verrucomicrobiota was about 1%, similar to previously obtained results in the Pearl River Basin [41]. The average abundance of the Proteobacteria in winter was 50% higher than that in summer (29%), and the abundances of Actinobacteriota and Cyanobacteria were 25% and 12% in winter compared to 33% and 23% in summer. At the genus level (Figure 6b), the group with the highest abundance was the CL500-29 marine group, accounting for 11.8%, followed by Psychrobacter, Chloroplast_norank, Cyanobium P CC-6307, and PeM15_norank, accounting for 10.88%, 10.65%, 6.47%, and 5.31%, respectively, and the abundances of CL500-29_marine_grou and hgcI_clade were 18.9% and 5.7% in summer, compared to 4.7% and 2.4% in winter. Previous studies have shown that Cyanobacteria, as primary producers, are an important component of the plankton and major contributors to the carbon, nitrogen, and phosphorus cycles [42]. Furthermore, the genera CL500-29_marine_grou and hgcI_clade genus are dominant genera in eutrophic waters and play an important role in the prokaryotic plankton community [43]. The PRD has a long-term low-oxygen environment and high-temperature hypoxia in summer, leading to a high abundance of Cyanobacteria, CL500-29_marine_grou, and hgcI_clade. From these results, it is clear that prolonged periods of high temperatures and anoxia in the water can also lead to cyanobacterial outbreaks even at low nutrient salt concentrations, while the increase in the amount of blue–green algae exacerbates the anoxia in the water and reduces biodiversity [44,45]. Our results demonstrated that it is necessary to strengthen the monitoring of the Pearl River Basin to prevent the abnormal proliferation of blue algae [46].

The Chao1 index and Shannon index values showed significant temporal differences in both indices (p < 0.001, p = 0.002), with higher values in summer compared to winter (Figure 6c,d), indicating that the richness and diversity of the bacterial communities in summer were higher. There was no significant difference in the spatial scale of the watershed; the abundance of microorganisms in the upstream sites was slightly higher compared to that in the downstream sites, which may be due to the accumulation of nutrient salts downstream areas with severe environmental pollution, which leads to a decrease in the abundance and diversity of microorganisms that are intolerant of pollutants [41]. It is now generally accepted that high salinity and ion levels in seawater can lead to a decrease in microbial diversity; therefore, freshwater shows a slightly higher bacterial abundance than saltwater [47].

3.6. Microbial Participation in P, Fe, N, and S Coupling Processes

Microorganisms drive energy flow and material circulation, making microbial ecology particularly important. Previous studies indicated that temperature, salinity, and nutrient salt concentration are factors affecting the microbial community structure in the Pearl River Basin [48]. In our study, a random forest experiment, principal coordinates analysis, and the Mantel test were used to determine the relationship between the environmental parameters and the relative abundances (Figure 7). The random forest experiment results show that TOC was the key factor affecting bacterial abundance, followed by DO and NO₃⁻. The result of the Mantel test showed that TOC, T, NO₃⁻, TP, and TN (p < 0.001) were important environmental factors in summer, while TOC, T, and NO₃⁻ (p < 0.001) were important environmental factors in winter. This suggested that TOC may be the main factor influencing the microbial community structure of the PRD and that it interacts with other factors to affect the composition and structure of the microbial community [8].

Low-valent sulfur can be oxidized by SOB under aerobic, anaerobic, or light conditions, eventually generating SO₄²⁻. The DO in the water was sufficient in winter, and higher levels of DO increase the activity of SOB. In this study, SOB belongs to the Mycobacterium phylum and played a critical role in the oxidation of sulfides (Figure 8a); the average abundance of Mycobacterium in summer was 1.257%, lower than that in winter (3.186%), resulting in a higher average sulfate concentration in winter (408.793 mg‧L⁻¹) than in summer (30.229 mg‧L⁻¹).

Denitrifying fungi, nitrobacteria, anaerobic ammonium oxidation bacteria, oxygen-consuming denitrifying bacteria, and anaerobic nitrifiers are the main participants in nitrogen recycling. Although anaerobic surface sediments are more suitable for denitrifying bacteria, our results demonstrated that the average abundance of denitrifying bacteria in summer was 0.318%, lower than the 0.513% found in winter, whereas the average abundance of nitrifying bacteria in summer was 0.327%, slightly higher than the 0.278% found in winter. The DO concentration was high in winter, and the abundant organic matter could provide sufficient substrate for denitrifying the bacteria. Nitrifying bacteria showed higher abundances in summer, when the water is hypoxic as well as less polluted; this may be because temperature is a major factor affecting the growth and reproduction of nitrifying bacteria [49].

PAOs consume oxygen to absorb excess phosphorus and release phosphorus under anaerobic conditions, making them important organisms in the biological phosphorus removal processes [50]. Through metabolic processes, these bacteria convert soluble phosphorus into insoluble phosphate compounds such as organic phosphates and Ca-P. The abundance of PAOs is related to temperature and DO. Overall, the average abundance of PAOs was 0.680% in summer, slightly higher than the 0.584% found in winter. Common genera involved in enhanced biological phosphorus removal include Acinetobacter and Acidovorax. In the present study, the average abundance of Acinetobacter was 0.356% in summer, slightly higher than the 0.325% found in winter, whereas the average abundance of Acidovorax was 0.3%, higher than the 0.109% seen in winter, possibly due to the release of organic phosphates from algae blooms during the summer, resulting in a higher average TP concentration than in winter. In addition, the abundant rainfall in summer leads to leaching and, consequently, to the loss of phosphorus from agricultural fields, providing a rich substrate for the growth and reproduction of PAOs [51,52].

Based on FAPROTAX, the bacterial community functions were annotated (Figure 8b). The dominant functions were N cycles, including nitrite_respiration and nitrate_reduction. Nitrate respiration is a process of denitrification and dissimilatory reduction of nitrate to ammonium [53]. The abundance of nitrite respiration bacteria in winter was significantly higher than in summer; this is important to correctly interpret the stronger denitrification in water and lower NO₃⁻ concentration. Combined with the DGT results, it was considered that the microbial activity could lead to more potential P being preserved in the form of Fe/Ca/Al-P, along with oxidation of sulfate under oxidizing conditions (in winter). Subsequently, the oxidizing conditions weakened (in summer), leading to the strong competitiveness of PAOs, with potential P hydrolyzed in the form of active P, as well as high concentrations and fluxes of P, Fe, and S in the pore water; this analysis found evidence of the risk of P and Fe being released into the water from the sediment.

4. Conclusions

The findings of this study highlight the impact of human activities and seasonal changes on the spatiotemporal distribution of nutrient salts and the bacterial community in the PRD. The PRD network region showed strong seasonal characteristics, with higher water quality in summer than in winter and lower nutrient salts in the upstream than in the downstream. The sediment composition was also significantly influenced by surface runoff and tidal scour, with higher sand and silt concentrations in summer and more clay and silty particles in winter. The sediment was identified as a source of PO₄³⁻, Fe²⁺, and NH₄⁺-N. Due to the impacts of sand excavation, shipping, and enterprise–industrial emissions, the average concentrations of P and Fe and their diffusion fluxes were higher in more densely populated regions. Microorganisms are the main drivers of the P, Fe, S, and N cycles. During summer, the abundant non-point source pollution from rainfall input provides a rich substrate for the bacteria in water, leading to a strong competitiveness of the PAOs and nitrifying bacteria. Meanwhile, a high temperature was favorable for the exchange of elements at the SWI, with a greater release of P, Fe, and N. During the low temperatures and high DO and nutrient salts seen in winter, the SOB and denitrifying bacteria were active, which correctly indicates the high concentration of SO₄²⁻ and NH4⁺-N in the water. Furthermore, TOC and DO were the most important factors affecting the bacterial community abundance.

This study provides a theoretical foundation for understanding the distribution of nutrient salts in the PRD and the microbe-mediated geochemical reactions. Due to the impact of land-sourced inputs, tides, and other factors on the microbial communities, further studies are needed, ideally using high-throughput sequencing methods, quantifying the impacts of the natural environment on water quality, and establishing a pollution management system for the Pearl River.