Impact of different microenvironment on water quality during the litter decomposition
The variation trend of COD concentration was similar among the different treatment groups, Fig. 1-a showed the COD concentration at 35°C was higher than that at 25°C by comparing the A3 and B3 groups (p > 0.05). Initially, the COD in both temperature groups increased to 135.9 ± 0.00 mg/L and 135.53 ± 0.55 mg/L within the first 4 h due to the release of organic matter by the litter decomposition, and then gradually decreased to 13.17 ± 0.29 mg/L and 15.06 ± 0.12 mg/L with the microbial degradation. Higher temperature promotes the growth of aerobic cellulose-decomposing bacteria, thereby accelerating the rate of microbial decomposition of plant residues (Bornette et al. 2011). The concentration of NH3-N and TP at A3 increased to approximately 3.185 ± 0.01 mg/L and 1.053 ± 0.01 mg/L, respectively. There were no significant differences between A3 and B3 (p < 0.05). Similarly, TN in the overlying water at A3 reached the maximum value of 6.81 ± 0.01 mg/L at 4 h and then fluctuated downward. The unit cumulative release of COD was highly significant greater in group B3 than in group A3 during the experiment, with 4.55 ± 0.00 mg/g/d (p < 0.001) (Table 2). The inhibition of microbial activities under low temperature conditions resulted in a reduce rate of nitrogen cycling and transport, leading to a high nitrogen content. Apoplasts, which contain organic nitrogen, decompose at a relatively slow rate. Additionally, nitrogen-fixing bacteria exist on the surface of apoplasts and immobilizes nitrogen in the surrounding medium through nitrogen fixation (Hernes et al. 2001), leading to the accumulation of nitrogen residues. The processes of nitrogen release, ammonification, nitrification, denitrification and nitrogen fixation by nitrogen-fixing bacteria exist simultaneously in the water body, contributing to fluctuation and changes observed in TN.
The decomposition of lotus leaves was impeded by high N concentration. In condition with NH3-N concentration of 0.8 mg/L (B1) and 1.2 mg/L (B3), the water TN showed no significantly difference with the measured concentration at 4.31 ± 0.01 mg/L and 4.14 ± 0.01 mg/L at the 648 h, respectively (p < 0.05). Previous research has indicated that increased concentration of NH3-N can influence the community structure of filamentous bacteria, resulting in the decrease in the rate of plant decomposition (Cornut et al. 2012). Moreover, the nitrogen application has been found to inhibit release of nutrients from leave litters under various nitrogen application conditions (Zhang et al. 2020a). However, a study by Menendze on the decomposition of Ruppia cirrhosa Petagna (Grande) suggested that the nitrogen application actually enhanced litter decomposition, possibly due to the presence of different microbial species in the experiment (Menéndez et al. 2003). In this study, it was observed that the rate of litter decomposition was higher under low nitrogen concentrations, suggesting that the microorganisms in the system might be better adapted to survive under such conditions.
The addition of sediment treatments (A2) enhanced litter decomposition rate compared with A3 treatment. The concentrations of water TN, TP and COD with sediment addition increased to the maximum value of 7.79 ± 0.01 mg/L, 1.12 ± 0.00 mg/L and 145.73 ± 0.64 mg/L earlier compared to the A3 treatment. The sediment serves as a habitat for abundant microorganisms that contribute to the consumption of organic matter, thereby facilitating plant decomposition. And the unit cumulative release of nutrient was significantly greater in the group A2 than in the group A3(p < 0.05). As microorganisms decompose plant residues, nitrogen-fixing bacteria also thrive on the surface of the litter, giving rise to a variety of microbial symbiotic networks and lead to fluctuating changes in NH3-N levels. Furthermore, sediment demonstrate an adsorption effect on nitrogen and phosphorus, which leads to a decrease in TN and TP contents in water, as the adsorption effect surpasses the decomposition effect. The sediment contains enzymes such as β-glucosidase, cellulose hydrolase, and acetylglucosaminidase, which play a dominant role in litter decomposition by supplying sugars and nutrients for microorganisms, in turn, accelerates the rate of decomposition (Zhang et al. 2007). Moreover, sediment-associated functional microorganisms actively participated in initial microbial decomposition of insoluble components, facilitating the breakdown process and reducing its duration (Hui et al. 2019).
Table.2 The unit cumulative release of nutrients
mg/(g·d) | A2 | A3 | B1 | B3 |
COD | 168.42 ± 0.52a | 110.59 ± 0.06c | 147.44 ± 0.18b | 110.37 ± 0.03c |
NH3-N | 2.40 ± 0.01b | 1.35 ± 0.01c | 2.85 ± 0.03a | 1.14 ± 0.00d |
TN | 6.87 ± 0.01a | 3.68 ± 0.00d | 6.16 ± 0.01b | 4.55 ± 0.00c |
TP | 1.25 ± 0.016a | 0.72 ± 0.021d | 1.02 ± 0.039b | 0.84 ± 0.028c |
Changes in dissolved organic matter during decomposition of litter
The UV-Vis spectra of the dissolved organic matter at 72 h and 648 h were analyzed (Fig. 2), and the spectral curves consistently demonstrated a reduction in absorbance with increasing wavelength. The samples exhibited absorption peaks ranging from 220 to 300 nm, corresponding to the B absorption band caused by the π-π* jump of the benzene ring (Nishijima et al. 2004).This indicates the presence of aromatic ring structure, as well as the presence of lignin, phenols and other recalcitrant substances. The absorption peak at around 280 nm at 72 h was used to evaluated the extent of aromatic conformation (Chin et al. 1994), and the higher value correspond to a greater degree of aromatic conformation (Fan et al. 2006), which can further provide insight into the molecular weight of the organic matter. Figure 2 showed that the absorbance of B3 was higher than that of B1. It is worth noting that the magnitude of the absorbance is positively correlated with molecular weight of the organic matter (Wang et al. 2009). This suggested that organic matter exhibits a greater molecular weight under a concentration of 0.8 mg/L NH3-N, which in turn promotes litter decomposition.
SUVA254 provides an indication of the percentage of DOM that is in the form of highly photoreactive, relatively biologically stable aromatic carbon (Stubbins et al. 2008). The SUVA254 value indicates the extent of humification, with higher values suggesting a greater degree humification. The absorption peaks at 254 nm are associated with samples that contain compounds with unsaturated carbon-carbon bonds (Fig. 3-a), such as aromatic compounds (Schnitzler et al. 2007). This observation aligns with the explanation, which primarily attributes the presence of these compounds to the oxidation and fragmentation of lignin-derived compounds (Chefetz et al. 1998). Additionally, Boissier found a significant correlation between phenols, polyphenols, and the degree of organic matter decomposition(Boissier et al. 1993). The SUVA254 values of all experimental groups showed a slight increase during the decomposition process, suggesting an increase in unsaturated bonds within the organic matter, this also indicates the removal of non-aromatic compounds or the production of small amounts of aromatic compounds as the decomposition time advanced (Zhou et al. 2022b). At 648 h, the SUVA254 value was higher at 35 ℃ compared to 20 ℃, indicating a higher concentration of aromatic substances in the samples. The elevated temperature enhances microbial activity, leading to an accelerated decomposition rate of litter and a facilitated microbial-dominated humification process, ultimately increasing the degree of humification (Cotrufo et al. 2013; Zhou et al. 2022a). Throughout the experimental cycle, group B3 exhibited the maximum SUVA254 value of 1.231 at 648 h, suggesting the highest level of humification. A comparison of A250/365 (Peuravuori et al. 1997) revealed that the experimental group was less humified at 72 h, corresponding to the presence of humin in samples, which is known to be resistant to decomposition (Fig. 3-b). As the experiment progressed, the samples underwent a shift towards predominantly fulvic acid at 648 h, influenced by microbial activity (Minero et al., 2007). The values of A253/203 indicated a higher presence of unsubstituted aromatic ring structures in all treatment groups at 72 h (Fig. 3-c). In contrast, the substituents exhibited more carboxyl, carbonyl, hydroxyl, and esters at 648 h (Korshin et al. 1997), except in the experimental group with sediment, and the difference could be attributed to the influence of microorganisms.
Fourier infrared spectroscopy (FTIR) was employed to identify the functional groups and vibrational spectra of vibrational polar bonds (Dean et al. 2010), enabling the determination of the chemical components of the samples (Fig. 4). The curve trends in treatment groups were approximately similar, with variations primarily observed in transmittance magnitude. The presence of a broad absorption peak at 3750–3200 cm− 1 indicated a significant amount of O-H in the sample, originating from carbohydrates in the apoplast, including polysaccharides and proteins, thus indicating the presence of substances like cellulose (Dean et al. 2010; Olk et al. 2000). A minor absorption peak at 3000–2700 cm− 1 suggested the presence of an aliphatic alkane structure, further indicating that the sample contains sugars (glycosides) or vitamins (Filip et al. 1988; Her et al. 2008). The peak observed at 2500–2000 cm− 1 in the infrared spectrum is typically associated with C ≡ C stretching vibrations, indicating the presence of heptyne. Furthermore, the absorption peak at 1900–1630 cm− 1 confirms the presence of heptyne in the sample. However, variations in the intensity of the absorption peaks were observed, the absorption peaks at 1628–1411 cm− 1 indicated the presence of amides and possibly proteins, and most of the absorption peaks in B1 were higher compared to A3, indicating that rising temperature promoted litter decomposition and resulted in a higher concentration of organic matter. Moreover, there were fewer absorption peaks observed at 648 h compared to 72 h, which could be attributed to the secretion of extracellular enzymes by microorganisms during litter decomposition, and break down polymers such as cellulose and proteins into the soluble monomer materials. The narrower absorption peaks in group B1, as depicted in Fig. 3-a, suggested that the NH3-N concentration at 1.2 mg/L was unfavorable for the synthesis of amide or protein compared to B3. This finding was consistent with the research conducted by Brock et al on the Nymphaea alba L., which revealed that N supplementation promotes lignin content in white water lily residues, thus impeding the decay process (Brock et al. 1985). Based on the analysis of the graph comparing the A2 and A3 groups, it was evident that all three curves exhibit an absorption peak except for A3–72 h, which lacked an absorption peak at 1628–1411 cm− 1, could be assigned of the involvement of the substrate was advantageous for the synthesis of amide or protein compounds. Overall, the comparison result highlights the positive impact of substrate involvement on the production of amide or protein compounds, as well as the subsequent nutrient release and microbial activity enhancement in the water. The participation of microorganisms in plant decomposition further reinforces the significance of the substrate in ecological processes.
Fluorescence EEMS have proven critical to the description of DOM bioavailability, and therefore, its role in terrestrial and aquatic ecosystem nutrient cycling (Zhao et al. 2013). The decomposition of aquatic plant apoplast results in release of soluble organic matter into the overlying water (Abril et al. 2016), and three-dimensional fluorescence spectra are capable of representing organic matter such as humus and protein-like compounds (Cuss et al. 2015; Zhao et al. 2015). Figure 5 inllustrated that all the water samples contain tryptophan-like, aromatic proteins, tyrosine, humus, humic acid and fulvic acid substances. Humus-like substances are typically derived from decay and decomposition products of plants. Tryptophan production, apart from being influenced by microbial activity, mainly originates from polycyclic aromatic hydrocarbons or related substances and by-products, which affect the ratio of amine to protein-like substances. In the pre-decomposition stage, soluble components such as sugars, proteins, organic acids, phenols, etc., were rapidly released with high intensity fluorescence. Subsequently, the release rate of tyrosine and acid-like humic acids occurred at a slower pace, so that the concentration of dissolved substances decreased towards the end of the experiment, and the fluorescence peaks at 648 h exhibited a significant reduction compared with that at 72 h. The fluorescence intensity of the samples at 35°C (A groups) was greater than that at 20°C (B groups), this difference could be attributed to the fact that elevated temperatures promote the metabolic and growth rate of microorganisms, resulting in faster cycling of organic matter (Zhou et al. 2016). Moreover, elevated temperatures stimulate species interactions, leading to increased metabolism activity of individual microorganisms, higher abundance of degradation-dominant bacteria, and accelerated litter decomposition. Samples in B3 groups with NH3-N concentrations of 0.8 mg/L, as well as sediment involvement groups (A2), exhibited higher fluorescence intensity, contributing to the litter decomposition released greater amount of soluble organic matter at a faster rate.
The fluorescence index FI is commonly used to differentiate DOM originating from various sources (McKnight et al. 2001), the aquatic and microbial sources typically exhibit an FI value of 1.9, whereas terrestrial and soil sources tend to have an value of 1.3. As data presented in Table 3, the FI values of each treatment group surpassed 1.9, indicating that the DOM was a result of microbial action (Cui et al. 2022). The humification index (HIX) value at 72 h lower than that at 648 h and negatively correlated with TP but positively correlated with TN (p > 0.05) (Fig. 5-b), indicating an increase in complex molecules, such as aromatics, and suggested that humification intensified as the experiment progressed. The humification was higher in the groups exposed to 35 ℃, as indicated by the greater HIX value of B3, which indicated a promotion of litter decomposition at 35 ℃. On the other hand, the BIX representing the dissolved organic matter produced by microbial activities. The BIX values at 72 h for the A2 and B3 groups were smaller than that at 648 h and positively correlated with COD (p > 0.05), which was corresponding to a significant change in the source of organic matter, and primarily relied mainly on exogenous inputs released by litter decomposition in the early stage. At 648 h, the BIX value exceeded 1, which was predominantly related to the dissolved organic matter produced by bacteria (Huguet et al. 2008). Initially, microbial activity was limited due to leaching being the primary process, while towards the end of the experiment, microbial activity became more prominent. Dissolved organic matter showed strong primary characteristics, as indicated by the BIX value for A2-72 h ranging from 0.6 to 1, that likely originating from the sediment. Based on the significant increase of BIX value at 648 h, the presence of sediment facilitated the decomposition of apoptotic matter. According to previous research, which also demonstrated the presence of robust biogenic properties in wetland soils in the riparian zone of Chongming Island (Wang et al. 2015).
Table.3 Spectral parameters of dissolved organic matter under different conditions
Sample | FI | HIX | BIX |
A2-72 h | 2.41 | 1.67 | 0.66 |
A2-648 h | 2.64 | 3.30 | 2.82 |
A3-72 h | 2.59 | 1.58 | 0.55 |
A3-648 h | 2.53 | 1.251 | 0.50 |
B1-72 h | 2.87 | 1.28 | 0.48 |
B1-648 h | 2.35 | 1.88 | 0.55 |
B3-72 h | 2.83 | 2.51 | 0.33 |
B3-648 h | 2.63 | 2.58 | 1.20 |
Changes in bacterial community composition and diversity
The sequence detection data of all the samples in the 72 h and 648 h of the whole experimental cycle were statistically analyzed (Table S1), and the results after sequencing showed that the range of the number of sequences of all the valid sequenced genes was 39310 ~ 74772. Microbial abundance increased as decomposition time progressed, and the Chao and Shannon indexes ranged from 100.5 to 351.12 and from 2.31 to 4.00, respectively. The dilution curve of Shannon diversity index (Fig. S2) tended to flatten, and were always smooth state, representing that the change of the number of sequences can be enough to respond to the information of the diversity of bacteria within the sample. Chao index and Shannon of bacterial community diversity in the overlying water at 20℃ at 648 h (Table 4) were significantly lower than that at 35℃ (p < 0.05), which was consistent with previously study by Yuan that the increase in temperature favors microbial growth (Yuan et al. 2021a). The bacterial richness of the added sediment was higher, but there was no significant difference (p > 0.05) in Chao index with the unadded groups, whereas the Shannon's index was significant different (p < 0.05). In this regard, that the addition of sediment has been found to increase the microbial abundance and diversity in the water due to the presence of Chloroflexi, Proteobacteria, and Acidobacteria (Wang et al. 2018), which are released from the sediment. Increasing the concentration of NH3-N resulted in the Chao index, and a significant difference was observed between the two experimental groups (p < 0.05). In our study, higher NH3-N concentrations were found to reduce microbial abundance compared to lower NH3-N concentrations. While increased nitrogen availability provides nutrients to microorganisms, the pH reduction caused by added nitrogen has a detrimental effect on microorganisms (Sun et al. 2018).
As the decomposition time progressed, microbial abundance and diversity were higher at 648 h compared to 72 h in all experimental groups. The water quality indicators exhibited changes during the experiment, resulting in the suppression of certain microorganism activities and alterations in dominant microbial populations. In the later stages of litter decomposition, the insoluble material that remains undergoes humification in the water, leading to an increase trend in the Shannon index. Aneja investigated the variation of microbial communities during litter decomposition by using litterbags in streams, revealed that the litter quality played a crucial role in microbial succession (Aneja et al. 2006). During the process of decomposition, the levels of easily accessible carbon in the form of sugar and starch decrease, while the recalcitrant carbon such as lignin and phenol increase, with the change in the litter quality also result in the change of microorganism community composition (Xu et al. 2013).
Table.4 Microbial Alpha Diversity Index
Group | Sample | Chao | Shannon | Coverage |
A2 | A2-72 h | 208.15 | 3.01 | 99.88% |
A2-648 h | 351.12 | 3.29 | 99.93% |
A3 | A3-72 h | 135.88 | 2.43 | 99.93% |
A3-648 h | 246.71 | 3.14 | 99.94% |
B1 | B1-72 h | 128.00 | 2.31 | 99.95% |
B1-648 h | 314.24 | 4.00 | 99.96% |
B3 | B3-72 h | 100.50 | 2.84 | 99.95% |
B3-648 h | 237.60 | 3.30 | 99.96% |
Research has demonstrated that variations in environmental and nutritional conditions can disrupt the natural ecological balance of water, subsequently leading to alterations in the composition of bacterial communities (Ferreira et al. 2016). Specifically, the microbial community composition at the phylum level displayed structural variations across different decomposition periods and experimental groups (Fig. 6-a), however, Proteobacteria (44.0–96.8%) consistently emerged as the dominant phylum. The other dominant phyla were Bacteroidota (1.56–35.2%) and Firmicutes (0.24–41.83%). Significant quantities of nitrogen and phosphorus were released into the water during the initial stage of plant decomposition, primarily through the activity of Proteobacteria, and specifically, higher temperatures have been found to be associated with an increase in Proteobacteria abundance (Tao et al. 2019). Bacteroidota was the second most prevalent phylum, with the exception of the sediment supplemented groups with the more prevalent of Firmicutes in the initial stages. This was consistent with the previous study that have reported similar high abundant of Bacteroidota and Firmicutes during litter decomposition (Wang et al. 2020). It is noticeable that the abundance of Firmicutes can be markedly involved in the decomposition of various substances and elements, owing to their function of participate in the refractory substances decompose such as cellulose and lignin (Michaud et al. 2009). On the other hand, Bacteroidota is indeed associated with the transformation of organic compounds including DNA and proteins (Goffredi et al. 2010). In comparison with 20°C, warming significantly increased the abundance of Proteobacteria, decreased the number of Bacteroidota, and increased the bacterial species. This result ties well with previous studies wherein Wu, the contrasting temperatures would kill many of the original cold-adapted organisms, enabling the proliferation of organisms that are resistant and resilient to higher temperatures (Wu et al. 2015). Yuan found that warming significantly increases the complexity of microbial networks, including network size and number of key species, as compared to environmental controls (Yuan et al. 2021b). In the later stages of decomposition, there were no significant differences between the two experimental groups (A3-2 and B3-2) (p > 0.05). Moreover, γ-proteobacteria (20.74–86.42%) was found to be first dominant class (Fig. 6-b), which play a crucial role in metabolizing carbon sources and possess a denitrifying function (Purkhold et al. 2000). The clostridia (5.23%~30.30%) and Bacteroidia (3.50%~25.18%) were significantly abundant in A2 group due to the introduction of bacteria-containing sediment into the reaction vessel, which leading to a distinct microbial composition and micro-environment (Shen et al. 2016). This finding aligns with the research conducted by Wu, where an evaluation in the abundance of clostridia and Bacteroidia was observed following the addition of the substrate (Wu et al. 2017). The observed increase in α-proteobacteria and their functions, including organic matter consumption, NH3-N decomposition and nitrogen fixation, indicate their important role in maintaining ecosystem health and nutrient cycling (Purahong et al. 2016). Furthermore, the increase in α-proteobacteria abundance observed in response to higher temperature and reduced NH3-N concentration, along with the resulting acceleration of litter decomposition, highlights their important role in ecosystem nutrient dynamics.
According to the bacterial genus level data of litter decomposition, the relatively high abundance of Pseudomonas (0.34–43.75%), Curvibacter (1.95–57.87%) and Flavobacterium (0.04–28.74%) might be the result of the nitrogen-fixing abilities consistent with other studies (Zhu et al. 2022). There were 92 genera common to all experimental groups, indicating a highly similar microbial composition (Fig. 7). Among them, Flavobacterium showed the significantly higher abundance in A3-72 h (p < 0.05), which was classified among bacteriophages that actively contribute to the process of carbon cycling and reduce the COD content in water (Garcia-Pausas et al. 2011; Teeling et al. 2012). Additionally, there were substantial differences in bacterial community structures at the genus level among the different stages of decomposition. The abundance of Flavobacterium in anaphase of litter decomposition was lower compared to initial phase (p > 0.05), and the abundance of Pseudomonas and Curvibacter was higher in initial 72 h (p > 0.05) (Fig. S3). Nutrient release during litter decomposition drives the changes in bacterial communities, with certain microorganisms emerging as dominant species due to nutrient utilization. Pseudomonas, belongs to Proteobacteria, is a Dissimilatory phosphate-accumulating organism (DPAO), which has been involved in the P cycle and transformation, as well as the nitrification in the nitrogen (N) cycle. The genus-level bacterial diversity was higher in A3-2 (Fig. 6-b). Pseudomonas exhibited the highest abundance in the non-added sediment group and maintained a relatively high abundance throughout the early and late stages of litter decomposition (A3-1 and A3-2). Curvibacter exhibited the highest abundance in the high NH3-N group and played a role of denitrification, converting NH3-N to NH2-N, while utilizing NH3-N and organic matter as the nitrogen source. This finding aligns with previous studies that demonstrated an increase in N-related bacterial abundance due to N fertilization (Sun et al. 2018).
Identifying core microorganisms and interspecies relationships in the plant litter decomposition
To investigate the relationship between microbial communities, environmental variables with litter decomposition, bacterial community and environmental parameters were used for redundancy analysis (RDA and CCA). On phylum level, 60.64% of the variance was explained by the RDA1 axis and 14.00% by the RDA2 axis (Fig. 8-a), which collectively explained 74.64% of the variance difference. Among the factors, COD (p = 0.083) and TP (p = 0.117) exerted a stronger influence on the bacterial community compared to other variables. On genus level, 21.10% of the variance was explained by the CCA1 axis and 16.25% by the CCA2 axis (Fig. 8-b), which collectively explained 37.35% of the variance difference. Among the factors, COD (p = 0.048) and NH3-N (p = 0.049) exerted a stronger influence on the bacterial community compared to other variables. Previous studies have consistently identified COD as the primary determinant of microbial community changes (Fang et al. 2023). Meanwhile, TP has been recognized as another significant factor influencing the alterations in microbial communities (Zhang et al. 2012). Wang confirmed the significance of P as a key driver of biological metabolism in water (Wang et al. 2018). Despite the high nitrogen concentration, its impact on the microbial community was insignificant, possibly due to the microbial community's resistance to the effects of nutrient enrichment. In addition, N is a crucial element utilized by microorganisms. The TN content in the water showed higher variability during litter decomposition. It may be a result of increased microbial demand for N, which leads to the fixation of N and thus reduction of N release (Gessner 2010). Consequently, the decomposition of N in litter is primarily influenced by biological factors, nitrogen content fluctuates in correlation with the decomposition of organic matter.
The species and environmental factors (TP, TN, COD and NH3-N) with the top 20 total abundance at the phylum classification level were analyzed by Heatmap (Fig. 9). The relationship between environmental factors and microorganisms was different at different times. In the pre-experimental period, only COD and TP were associated with bacterial community composition, Proteobacteria showed a significant positive correlation with TP. Actinobacteria showed a significant negative correlation with COD. Firmicuyes showed a significant negative correlation with TP, but Bacteroidota did not show a significant correlation with the environmental factors (Fig. 9-a). The relationship between environmental factors and microorganisms was stronger at the end of the experiment than at the beginning of the experiment, and only NH3-N and TN were related to microbial community composition. Actinobacteria showed a significant positive correlation with TN. Protobacteria, Firmicuyes, and Bacteroidota did not show a significant correlation with environmental factors.
During the process of litter decomposition, microbial community dynamics interact with various environmental factors. The network analysis was employed to identify the core bacteria at the genus level associated with water quality during litter decomposition, where each node was represented as a genus. As shown in Fig. 10-b, the majority of nodes associated with TP at 72 h showed a negative correlation with large proportion of bacteria, with over 60% belonging to Firmicutes genus, but only 5 nodes displayed the correlation at 648 h (Fig. 10-c). At the initial stage, the microorganisms were influenced by the environment and the adapted microorganisms proliferated, decomposing insoluble substances and causing the release of large amounts of phosphorus from plant residues (Ozalp et al. 2007). This process inhibited the growth of certain genus, while promoting proliferation of other adapted genus. The number of nodes involved in COD conversion was not significantly different between 72 h and 648 h (p > 0.05). Initially, only a few nodes were involved in the conversion of NH3-N and TN, but most of the nodes exhibited N fixation after 648 h had, which was consistent with Purahong’s findings(Purahong et al. 2016). Flavobacterium enhances the nitrogen fixation by nitrogen-fixing bacteria by influencing the secretion flavonoids (Dardanelli et al. 2010). A large number of Proteobacteria nodes appeared in both phases, suggesting that it was mainly Proteobacteria that played a role throughout the whole experiment period (Zhang et al. 2020b). In addition, co-occurrence patterns revealed a significant increase in the complexity and connectivity of bacterial interactions in 648 h than during that in 72 h (p < 0.05). Previous study states that there are complementarity effects between biotic and abiotic, and that changes in the abiotic environment interact with community composition and the loss diversity, ultimately influencing the efficiency of microbial carbon utilization efficiency within the community (Domeignoz-Horta et al. 2020).
A plethora of microorganisms participate in the decomposition process of aquatic plant litter, each performing distinct functions. Species with high abundance were identified as functional microorganisms decomposing aquatic plants through the utilization of LEfSe multilevel species difference discriminant analysis. It is evident that both the predominant species in 72 h (Acidovorax, Pantoea, Pelomonas, Herbaspirillum) and the dominant species in 648 h (Rhizobiales, Azospirillum, Pseudacidovorax, Xanthobacter, Phenylobacterium) belong to Proteobacteria, which is consistent with the results of co-occurrence network analysis (Fig. 10-b; Fig. 10-c). There were subtle variations in microbial abundance trends among different experimental groups (Fig. 10-a), probably attributed to differences in the rate of litter decomposition and nutrient content within the water. Most of the microorganisms observed in this experiment were nitrogen-fixing bacteria, suggesting that functional microbial taxa have a direct or transient influence on nitrogen cycling processes. Previous study has also highlighted that the abundance of nitrogen-fixing genes changed with vegetation type (de Sosa et al. 2018).