Plant Growth Modications Due to Coastal Embankments Affect Soil Bacterial and Archaeal Communities Over the Growing Season in Salt Marshes of Eastern China

Aims: Although the inuences of coastal embankments on physicochemical soil properties and carbon (C) and nitrogen (N) cycling have been widely studied, the mechanisms of their effects on soil microbial ecologies remain poorly understood. Thus, the aim of this study was to investigate variations in the diversity and composition of soil bacterial and archaeal communities between natural and embanked saltmarshes, as well as the determinants that drive these variations. Methods: 16S rRNA gene sequence analysis was performed to assess the impacts of embankments on the bacterial and archaeal communities of native Suaeda salsa, Phragmites australis, and invasive Spartina alterniora saltmarshes on the east coast of China. Results: Embankments were found to signicantly decrease the microbial diversity of the S. alterniora salt marsh, while they increased the OTU richness of the P. australis salt marsh. Embankments modied the compositions of soil bacterial and archaeal communities in both the S. alterniora and P. australis salt marshes. However, variations in the microbial diversity, richness, and community compositions between the native and embanked S. salsa salt marshes were insignicant. Conclusions: These results were possibly because the embankment signicantly altered soil nutrient substrate levels (e.g., soil organic C and N) by variations in plant residues and physiochemical soil properties in S. alterniora and P. australis saltmarshes, whereas the embankment had no observable changes in the soil nutrient substrate and the plant residue in S. salsa saltmarsh. This study also elucidated the effects of coastal embankments on biogeochemical cycles, and highlighted their potential hazards to ecosystems. of and in the S. and P. salt in and archaeal diversity, richness, and community between and salt results the S. and salt noted salt This study provides further insights toward a better understanding of the variations and driving patterns of soil following the establishment of embankments, which the effects of coastal embankments on and highlighted their potential hazards


Introduction
The reclamation coastal land through the establishment of embankments has caused considerable ecological problems in natural coastal ecosystems. For instance, they have greatly threatened coastal habitats and biodiversity (particularly soil microbial communities) and altered ecosystem processes and functions (particularly C and N cycling) ( Moreover, bacteria and archaea are sensitive soil elements that can rapidly detect and react to various conditions and changes in the soil environment (Gryta and Frac 2020). Although the impacts of coastal embankments on aboveground ecosystems have been widely studied, their potential advantages and risks to belowground ecosystems, particularly soil bacterial and archaeal communities, have not been fully elucidated.
Soil nutrient substrates are one of the primary driving factors for bacterial and archaeal diversity and community compositions (Peralta et al. 2013; Kuang et al. 2021; Sun et al. 2021). In particular, the decomposition of soil organic C and N (SOC and SON, respectively) provide energy and nutrition for soil microorganisms; thus, they are overarching driving factors for the presence and support of soil bacteria and archaea (Zechmeister-Boltenstern et al. 2015; Yang et al. 2020). However, various bacterial and archaeal taxa respond differently to the availability of nutrients. For instance, copiotrophic bacteria (e.g., phylum Bacteroidetes) prefer nutrient-rich environments, as labile organic C and N (LOC and LON, respectively) inputs can signi cantly promote their growth (Fierer et al. 2007a; Newton and McMahon 2011;Philippot et al. 2013).
In contrast, because LOC and LON inputs have negative effects on oligotrophic bacteria (e.g., Acidobacterium and some species of Actinobacteria) and can limit their growth, these organisms prefer nutrient-poor environments (Podosokorskaya et  Consequently, the responses of soil bacterial and archaeal communities to coastal embankments may differ contingent on the shifts of nutrient substrates. Soil microbial communities are considerably in uenced by biotic (e.g., the quantity and quality of plant materials) (Angel et al. 2010; Cotrufo et al. 2019) and abiotic (e.g., climate, soil type, and soil physicochemical properties) factors (Bainard et al. 2016;Nguyen et al. 2018). Plant properties, including residues (e.g., litter and roots) and aboveground (e.g., leaves, stems and litter) biomass, strongly affect soil bacteria and archaea (Brotosudarmo et al. 2015;Li et al. 2014a; Urbanová et al. 2015). The former dictates the availability of soil C and N, in addition to providing unique environments for soil bacteria and archaea (Kuske et al. 2002; Urbanová et al. 2015), whereas the latter in uences the presence of photosynthetic bacteria (PSB) by affecting the extent of light that reaches the soil surface ( Notably, the establishment of coastal embankments considerably shifts the net primary production and physicochemical properties of coastal wetland soils (Yang et al. 2016;Yang et al. 2019). Thus, the identi cation of biotic and abiotic factors that affect soil bacterial and archaeal communities in environments with coastal embankments may improve our understanding of how the coastal embankments impact bacterial and archaeal communities in soils.
China has been increasing the intensive construction of coastal embankments, with approximately 60% of the total length of its mainland coastline being enclosed by thousands of kilometers of these levees (Ma et al. 2014;Sun et al. 2015). Jiangsu province has the highest abundance of coastal wetlands and embankments in Eastern China (Chung et al. 2004). Currently, a large proportion of the natural coastal wetlands in this region have been reclaimed by embankments (i.e., construction of dikes, seawalls, and barriers along the coastline), shponds, farmlands, and unban lands (Liu 2018).
The results of previous studies with a focus on the impacts of the conversion of coastal wetlands to other land-use types on soil microbial communities have been inconsistent (Xu et  Additionally, it is di cult to distinguish between the effects of coastal embankments and subsequent land-use conversions (e.g., shponds, agricultural, and urban land) on soil microbial communities (Bai et al. 2013;Bu et al. 2015). Therefore, the impacts of coastal embankments without associated land-use changes on the diversity and compositions of soil microbial communities remain unknown. Moreover, our previous studies revealed that there were diverse responses to coastal embankments by soil C and N pools for distinct vegetation types, as they have variable effects on residue biomasses of speci c plants (Feng et al. 2022).
Thus, we hypothesized that there would be different responses from microbial communities for diverse vegetation types (e.g., microbial community diversity and compositions) to coastal embankments, which was due to variations in the quantity and quality of soil nutrients. Further, we hypothesized that variations in soil physicochemical properties following the establishment of embankments would affect the diversity and compositions of microbial communities. To verify our hypothesis, we compared the diversity and compositions of soil bacterial and archaeal communities by extracting and sequencing their 16S RNA genes in embanked and adjacent unembanked Spartina alterni ora Loisel., Suaeda salsa (L.) Pall., and Phragmites australis (Cav.) Trin. ex Steud. salt marshes along the Jiangsu coastline.
In addition, the plant biomass and soil properties (i.e., soil moisture, salinity, pH, SOC, LOC, water-soluble organic C (WSOC), SON, LON, and water-soluble organic N (WSON)) were measured. The aim of this study was to: (1) understand whether the establishment of coastal embankments without land-use changes in uenced the diversity and compositions of bacterial and archaeal communities for various vegetation types, and if so, (2) identify the factors that drive these variations.
The mean annual temperature and precipitation of this area are 13.7-14.6°C and 980-1,070 mm, respectively (Wang and Liu 2005), and the salinity of the seawater is ~ 30.9% (Yang et al. 2020). S. alterni ora, S. salsa, and P. australis salt marshes comprise the main vegetation types in coastal Jiangsu. S. alterni ora is an invasive perennial grass that was introduced to China from North America in 1979, while S. salsa and P. australis are native plants ; Qin and Li 2012).
The coastal wetlands of Yancheng are one of the important stops on the migration route of shorebirds between East Asia and Australia. This is also the largest wintering area of endangered Red-crowned Cranes (Grus japonensis) (Liu et al. 2010; Wang et al. 2019b). However, since its introduction, S. alterni ora has been outcompeting native plants for resources such as space; therefore, several embankments have been constructed to control its expansion (

Field sampling
In September 2016, four parallel transects (40 × 40 m) were established in embanked S. alterni ora (ESA), S. salsa (ESS), and P. australis (EPA) salt marshes and unembanked S. alterni ora (USA), S. salsa (USS), and P. australis (UPA) salt marshes (Fig. 1). We randomly selected three plots (2 × 2 m) in each transect, and soil samples were randomly collected from three points (Ø 5 cm × 30 cm deep) in each plot. All the soil samples from the same transects of the same salt marsh were homogeneously mixed. Finally, 24 soil samples (four replicates × six treatments) were obtained. All visible plant litter, stones, and roots were removed to form the nal soil samples, which were divided into four subsamples after thorough mixing.
The rst soil subsample was placed in an aluminum box to determine soil moisture. The second was air-dried, passed through a 1 mm sieve, and then used for the measurement of soil pH, salinity, SOC, and SON. The third was passed through a 2 mm sieve, preserved at 4°C, and then used for the determination of WSOC and WSON concentrations. The fourth soil subsample was passed through a 2 mm sieve and stored at -80°C for the analyses of the diversity and compositions of soil bacterial and archaeal communities.
Three quadrats (50 × 50 cm) were established at each transect to collect the plant samples, including leaves, stems, litter, and roots. Roots were collected in each transect from three blocks of soil (10 cm length × 10 cm width × 30 cm depth), which were sifted through a 15 mm sieve. The aboveground biomass was calculated as the sum of the leaves, stems, and litter biomass, whereas the belowground biomass was represented by the root biomass.

Soil and plant properties
The soil pH was determined in a 1:2.5 soil:water suspension using a pH meter, whereas the soil salinity was measured in a 1:5 soil:water suspension with a conductivity meter (Yang et al. 2016). The total inorganic C and N were obtained from 10 g of air- The WSOC and WSON were determined following a previously established method (Cabrera and Beare 1993;Yu et al. 1994;Yang et al. 2016). In addition, the leaf, stem, litter, and root biomass were measured after oven-drying at 65°C.

DNA extraction and polymerase chain reactions (PCR)
The total genomic DNA of the 50 mg soil samples was extracted using the Omega Bio-Tek E.Z.N.A. Soil DNA Extraction Kit (Omega Bio-Tek, Atlanta, USA) following the manufacturer's protocol. The DNA concentration and purity were monitored using 1% agarose gels. According to the concentration, the DNA was diluted to 1 ng/µL using sterile water.

Quanti cation and sequencing of PCR products
Equal volumes of 1X loading buffer (containing SYBR TM Green) were mixed with the PCR products and electrophoresed on a 2% agarose gel for detection. The PCR products were mixed at equidensity ratios. Subsequently, the PCR products were puri ed using the Qiagen Gel Extraction Kit, according to the manufacturer's instructions (Qiagen, Germany).
Sequencing libraries were generated using the TruSeq DNA PCR-Free Sample Preparation Kit (Illumina, San Diego, CA, USA) following the manufacturer's recommendations, and index codes were added. The library quality was assessed on a Qubit 2.0 Fluorometer (Thermo Scienti c, CN) and Agilent Bioanalyzer 2100 system (Agilent, Santa Clara, CA, USA). Finally, the library was sequenced using an Illumina NovaSeq platform, and 250 bp paired-end reads were generated.

Data analysis
Paired-end reads were assigned to the soil subsamples based on their unique barcodes and truncated by cutting off the barcode and primer sequences. Paired-end reads were merged using FLASh v.1.2.7 (Center for Computational Biology, Baltimore, MD, USA), when at least some of the reads overlapped the read generated from the opposite end of the same DNA fragment; the splicing sequences were designated as raw tags (Magoč and Salzberg 2011). The quality ltering of raw tags was performed under speci c ltering conditions to obtain high-quality clean tags according to the QIIME v. Sequence analysis was performed using Uparse software v.7.0.1001 (Edgar 2013). Sequences with ≥ 97% similarity were assigned to the same optical transform unit (OTU). Representative sequences of each OTU were screened for further annotation. For each representative sequence, the Silva Database was used based on the Mothur algorithm to annotate the taxonomic data (Quast et al. 2013). To study the phylogenetic relationships of different OTUs and the variations in the dominant species of different samples, multiple sequence alignment was conducted using MUSCLE software v.3.8.31 (Edgar 2004). The OTU abundance data was normalized using a standard with sequence numbers that corresponded to the sample with the least number of sequences. The ensuing analyses of alpha and beta diversity were performed based on the normalized data output.
Indices were calculated with QIIME v.1.9.1 and displayed via R software v.2.15.3 (Caporaso et al. 2010). The Simpson index was employed to assess community diversity, and principal coordinate analysis (PCoA) and analysis of similarities (ANOSIM) were performed to detect differences in community structures (beta diversity). Signi cant differences in the soil bacterial and archaeal communities between USA and ESA, USS and ESS, or UPA and EPA were evaluated by linear discriminant analysis (LDA) and effect size (LEfSe) (Segata et al. 2011). Cladograms were created showing the differences in the microbial lineages between USA and ESA, ESS and ESS, or UPA and EPA with LDA values of 4 or higher.

Statistical analysis
The in uences of environmental factors (e.g., soil moisture, salinity, pH, LOC, ROC, WSOC, LON, RON, WSON, and C/N) on the bacterial and archaeal community structures at the phylum, class, order, family, genus, and species levels were evaluated using redundancy analyses (RDA) via CANOCO v.5.0 (Wageningen University & Research, Wageningen, Netherlands). The Monte Carlo permutation test (499 permutations) was performed, and the statistical signi cance was set at 0.05. Statistical software SPSS 22.0 (IBM Corp, Armonk, New York, USA) was used to analyze the data. The impacts of coastal embankments on the plant biomass, physicochemical soil properties, Simpson index, OTU richness, and relative abundance of dominant bacteria and archaea in USA, ESA, USS, ESS, UPS, and EPS were evaluated using one-way ANOVA (analysis of variance).
The relationships between the diversity and abundance of soil bacteria and archaea and the plant biomass and soil properties were evaluated using Pearson's correlation analysis. Furthermore, structural equation modeling (SEM) was performed to determine the direct and indirect effects of coastal embankments on parameters related to soil bacterial and archaeal abundance following the expectations of the a priori model (Fig. S1). The normality of all endogenous variables and the overall model was tested and veri ed.
The normal-distribution-based maximum likelihood method was used for parameter estimation (Boldea and Magnus 2009). The best-tting model was selected by the sequential removal of non-signi cant paths (p < 0.05). The overall goodness-of-t for the model was tested using the chi-squared test (χ 2 ) and the root-mean-square error of approximation (RMSEA; 0 ≤ RMSEA

Plant and soil properties
The litter, root, aboveground, residue, and total biomass of ESA were 70.642%, 57.901%, 57.773%, 61.001%, and 57.822% respectively, which were signi cantly lower than those of USA (Tables 1 and S1). However, the root, residue, and total biomass of EPA were 175.490%, 143.828%, and 104.229%, respectively, which was higher than those of UPA (Table 1). Of note is that there was a negligible difference between the residue and total biomass of the unembanked and embanked S. salsa salt marshes (Table 1). Table 1 Leaf, stem, litter, root, aboveground, and total biomass (mean ± SE, n = 4) in the unembanked and embanked Spartina alterni ora, Suaeda salsa, and Phragmites australis salt marshes. Different superscripted lower-case letters indicate statistical signi cance at p < 0.05 between the unembanked and embanked ranges in the same plant salt marshes. Following the establishment of the embankments, the soil moisture decreased signi cantly in the S. alterni ora and S. salsa salt marshes, whereas it increased dramatically in the P. australis salt marsh. Moreover, the embankments signi cantly increased the soil pH, and reduced soil salinity in all vegetation salt marshes ( was lower than those of USA (Table 2). Table 2 Soil physicochemical properties (a) and concentrations of soil total, labile, water-soluble, and recalcitrant organic C (b), and N (c) (mean ± SE, n = 4) in the unembanked and embanked S. alterni ora, S. salsa, and P. australis salt marshes. Different superscripted lower-case letters indicate statistical signi cance at p < 0.05 between the unembanked and embanked ranges in the same plant salt marsh. The WSOC, ROC, and RON were 34.375%, 78.808%, and 54.867%, respectively, which was higher in EPA than in UPA (Table 2).
Further, there were no signi cant differences in the concentrations of SOC, LOC, WSOC, ROC, SON, LON, WSON, and RON between the ESS and USS (Table 2). However, it is worthy of note that the WSOC concentrations in both ESS and USS were signi cantly higher than those in UPA and EPA, while the WSON concentrations in both the ESS and USS were markedly higher than those in USA and ESA (Table S1). The top two concentrations of ROC and RON were observed in USA and EPA (Table S1).

Alpha diversity of bacterial and archaeal communities
The OTU richness of EPA was signi cantly higher than that of UPA, and the Simpson diversity index of ESA was signi cantly lower than that of USA (p < 0.05; Fig. 2). However, both the OTU richness and the Simpson diversity index between USS and ESS were not signi cant (p ≥ 0.05; Fig. 2). Further, the Pearson's correlation analysis of all salt marshes revealed that the variations in Simpson diversity indices were signi cantly positively correlated with soil moisture, aboveground and root biomass, SOC, ROC, SON, LON, and RON concentrations, and OTU richness (Table S2).

Taxonomic composition of soil bacterial and archaeal communities
USA possessed an abundance of Deltaproteobacteria (from class to order, within the phylum Proteobacteria) and Epsilonproteobacteria (from class to order, within the phylum Proteobacteria), while the ESA had an abundance of Actinobacteria (from phylum to order), Betaproteobacteria (within the phylum Proteobacteria), and the order Xanthomonadales (from order to family, within the class Gammaproteobacteria, phylum Proteobacteria) (Fig. 3). Moreover, the abundances of the phylum Chlorobi, and classes Ignavibacteria, Chlorobia (within the phylum Chlorobi), and Chloro exi (within the phylum Chloro exi) were signi cantly higher in ESA than those in USA (p < 0.05), whereas the abundances of the phylum Fusobacteria and the class Rhodothermia (within the phylum Bacteroidetes) were signi cantly lower in ESA than in USA (Fig. 4).
ESS had a high occurrence of Gammaproteobacteria (from class to family, within the phylum Proteobacteria) and the order Burkholderiales (within the class Betaproteobacteria) (Fig. 3). The family Shewanellaceae (from family to genus, within the order Alteromonadales, class Gammaproteobacteria) was abundant in UPA, while the Betaproteobacteria (from class to order), the order Pseudomonadales (within the class Gammaproteobacteria), and the family Chromatiaceae were abundant in EPA (Fig. 3). Additionally, the abundance of class 4C0d-2 (within the phylum Cyanobacteria) was signi cantly lower in EPA than in UPA (p < 0.05; Fig. 4). Moreover, the highest abundance of Gammaproteobacteria was observed in ESS in contrast to USA, ESA, USS, UPA, and EPA (Table S3). The top two abundances of Deltaproteobacteria and Epsilonproteobacteria were observed in USA and EPA (Table S3).

Beta diversity of soil bacterial and archaeal communities
Bray-Curtis dissimilarity indices indicated variations in the bacterial and archaeal communities between USA and ESA and between UPA and EPA were 65.620% (p = 0.0230) and 45.830% (p = 0.0460), respectively (Fig. 5). However, variations in the bacterial and archaeal communities between USS and ESS were not signi cant (p ≥ 0.05; Fig. 5).

Controls on the soil bacterial communities
Ten environmental variables (i.e., soil moisture, pH, salinity, ROC, LOC, WSOC, RON, LON, WSON, and C/N) explained 23.500% and 28.000% of the total changes in the compositions of soil bacterial and archaeal communities at the phylum and class levels, respectively (Fig. 6). The results of Monte Carlo permutation tests revealed that the variations at the phylum level were closely related to C/N (F = 7.00, p = 0.002) and WSON (F = 3.10, p = 0.036) (Fig. 6a), while those at the class level were closely associated with WSOC (F = 7.10, p = 0.002), C/N (F = 3.70, p = 0.010), and RON (F = 2.60, p = 0.042) (Fig. 6b).
The SEM and Pearson's analysis revealed that the abundance of Gammaprotebacteria was highly correlated with the concentrations of RON, LOC, LON, and WSOC (Fig. 7). The abundances of Deltaproteobacteria, Betaproteobacteria, and Epsilonproteobacteria were signi cantly correlated with the concentrations of ROC and RON (Fig. 7). Moreover, soil salinity had a signi cant negative correlation with the abundance of Betaproteobacteria (Fig. 7). The abundance of Actinobacteria was signi cantly negatively correlated with the LON, WSON, and RON concentrations (Fig. 7). The abundance of Rhodothermi was signi cantly correlated with the ROC concentration (Fig. 7). Moreover, a signi cant negative correlation was found between the abundances of Chlorobi, Ignavibacteria, Chloro exi, and Chromatiaceae and aboveground biomass (Table S4).

Discussion
The establishment of coastal embankments signi cantly decreased the Simpson diversity index in the S. alterni ora salt marsh, while they increased the OTU richness in the P. australis salt marsh (Fig. 2). Interestingly, coastal embankments altered the microbial diversity and richness only slightly in the 0-30 cm soil layer of the S. salsa salt marsh (p ≥ 0.05; Fig. 2 Conversely, the lower soil salinity (e.g., 1.737‰) in the embanked salt marsh was more suitable for the growth of P. australis than the high soil salinity (e.g., 4.164‰) of the unembanked salt marsh ( Table 1). The increased residue biomass of P. australis caused by the low soil salinity led to an increase in the SOC and SON of the embanked P. australis salt marsh, which presumptively increased the OTU richness (Tables 1 and 2; Yang et al. 2020). However, coastal embankments had insigni cant impacts on plant residues, as well as soil SOC, LOC, WSOC, SON, LON, and WSON in the S. salsa salt marsh, which explained the negligible impact of these structures on bacterial and archaeal diversity and richness in the S. salsa salt marsh ( Fig. 2; Table 2).
Soil moisture is a vital driver of soil bacterial and archaeal diversity (Guo and Zhou 2020; Yang et al. 2020; Sun et al. 2021).
Previous studies reported that the higher availability of soil water in coastal zones was advantageous for the diversity of soil bacteria and archaea in these areas, which was consistent with the results of this study (Table S2; Gao et al. 2020;Novoa et al. 2020). The embankments prevented seawater from reaching the coastal wetlands, which decreased the soil moisture of the S. alterni ora salt marsh, thereby contributing to lower diversity (Table 2) In this study, the soil moisture was increased in the embanked P. australis salt marsh, which might have explained its higher OTU richness ( Fig. 2; Table 2). Soil pH shows an excellent correlation with bacterial and archaeal diversity; however, in our study, an insigni cant correlation was observed between the soil pH and bacterial and archaeal diversity ( Table 2). In summary, our results indicated that the establishment of coastal embankments in uenced the bacterial and archaeal diversity and richness in S. alterni ora and P. australis communities by shifting soil nutrient substrates, which was primarily determined by plant residues, and changes in soil moisture (Tables 1, 2 and S2).
The establishment of coastal embankments drastically modi ed the composition of soil bacterial and archaeal communities in S. alterni ora and P. australis salt marshes (Fig. 5). PCoA and Bray-Curtis dissimilarity indices revealed that the soil bacterial and archaeal community compositions in the embanked S. alterni ora and P. australis salt marshes were clustered and distinct from those in the unembanked S. alterni ora and P. australis salt marshes, respectively (Fig. 5). However, the variation in the compositions of soil bacterial and archaeal populations between the unembanked and embanked S. salsa communities was insigni cant (Fig. 5).
The WSOC, which is the rst to be rapidly utilized by bacteria and archaea, was the most direct driver of soil bacterial and Thus, in this study, the decrease of WSOC and RON in the S. alterni ora salt marsh following the establishment of coastal embankments caused a differentiation in the bacterial and archaeal communities between the unembanked and embanked S. alterni ora salt marshes (Figs. 5, 6 and S2; Table 2). Additionally, the higher concentrations of WSOC and RON in the embanked P. australis salt marsh in uenced the differentiation of the bacterial and archaeal communities between the unembanked and embanked P. australis salt marshes (Figs. 5, 6 and S2; Table 2).
Furthermore, the insigni cant difference in soil nutrient substrates explained the inconspicuous differentiation of bacterial and archaeal communities between the unembanked and embanked S. salsa salt marshes (Figs. 5, 6 and S2; Table 2). Besides, the microbial community composition was also primarily in uenced by changes in the soil C/N ratio (Fig. 6). An increased C/N ratio typically suggests that soils are degraded due to the loss of C, which limits microbial growth (Pereira et al. 2021). However, in this study, embankments did not change the soil C/N ratio for all vegetation types; thus, they did not modify the microbial community composition by altering the soil C/N ratio (Table 2).
Consequently, our results con rmed that the establishment of coastal embankments in uenced the soil bacterial and archaeal community composition primarily by altering the concentrations of nutrient substrates, which was determined signi cantly by plant residues (Fig. 7; Table 2; Liao et al. 2007;Yang et al. 2016). In this study, the establishment of coastal embankments had a negligible effect on the relative archaeal abundance at the phylum, class, order, family, genus, and species levels ( Table S5).
As archaea are extremophilic microorganisms, few environmental factors can have dramatic impacts on their growth, which might explain these results (Bates et al. 2011;Ventosa and Haba 2011).
Several chemoorganotrophic bacteria were signi cantly in uenced following the establishment of embankments. The phylum proteobacteria has been described in marine and terrestrial ecosystems as an important component of biogeochemical cycles, and was the most abundant phylum in the sediment of coastal Yancheng ( Fig. S3; Lee et al. 2020;Yang et al. 2020). In this study, the class Gammaproteobacteria was the most abundant, and its relative abundance was in uenced markedly following the establishment of embankments (Figs. 3 and S3).
The coastal embankment dramatically simulated the presence of Gammaprotebacteria in S. salsa saltmarshes (Fig. 3). Moreover, the abundance of Gammaproteobacteria in the embanked S. salsa saltmarsh was much higher than those in the other saltmarshes ( Fig. 3; Table S3). Gammaproteobacteria is classi ed as a copiotrophic bacterial taxon, which has the capacity to grow quickly by utilizing labile compounds, particularly dissolved organic matter ( In this study, we also observed that S. salsa saltmarshes possessed higher WSON concentrations than S. alterni ora saltmarshes, and higher WSOC concentrations than P. australis saltmarshes (Table S1) In this study, SEM analysis also veri ed that the abundance of Deltaproteobacteria was highly correlated with ROC and RON concentrations in the soil (Fig. 7). Since S. alterni ora and P. australis residues are rich in lignin and lignocellulosic materials, the coastal embankments altered the inputs of recalcitrant materials from plant residues in the S. alterni ora and P. australis Therefore, the coastal embankments decreased the quantities of plant residues, and soil ROC and RON, which resulted in the reduced abundance of Deltaproteobacteria in the S. alterni ora saltmarsh. The coastal embankments increased the amount of plant residues, as well as soil ROC and RON, which led to an increase in the abundance of Deltaproteobacteria in the P. australis saltmarsh (Tables 2 and S3). Similar to Deltaproteobacteria, we found that the abundances of Epsilonproteobacteria and Rhodothermi were obviously correlated with the ROC and RON concentrations, which supported the notion that Thus, the coastal embankments had similar in uences on Epsilonproteobacteria and Rhodothermi as on Deltaproteobacteria (Fig. 7). Moreover, some Actinobacteria are oligotrophic bacteria whose growth is restricted in nutrient-rich soils (Pascault et al. 2013;Trivedi et al. 2013;Verzeaux et al. 2016;Yang et al. 2020). Decreased S. alterni ora biomass in the soil of embanked salt marshes severely reduced the availability of nutrients (e.g., SOC and WSON), which was bene cial for the presence of Actinobacteria, and reduced its abundance in the embanked S. alterni ora sediment ( Fig. 3; Tables 1 and 2). Thus, our results suggested that the establishment of coastal embankments in uenced the distribution of chemoorganotrophic bacteria, as they modi ed the soil nutrient substrates (e.g., SOC, ROC, RON, and WSON) by appreciably altering the organic material inputs from plants.
Coastal embankments also affected the presence of photosynthetic bacteria by in uencing plant growth. Photosynthetic bacteria are widely distributed in coastal sediments (Okubo et al. 2006;Idi et al. 2014). In this study, the quantity of obligately phototrophic bacteria from the phylum Chlorobi and class Chloro exi (Garrity et al. 2005;Hanada 2014) was signi cantly higher in the embanked S. alterni ora salt marsh (Fig. 4). The light-limiting effect of the dense vegetation of S. alterni ora directly restricted the presence of PSB (He et al. 2012;Li et al. 2014b;Brotosudarmo et al. 2015).
Indeed, the abundances of the phylum Chlorobi and class Chloro exi were signi cantly negatively correlated with the aboveground biomass of plants (Table S4). Thus, the higher occurrence of the phylum Chlorobi and class Chloro exi might have been induced by the increased sunlight due to the lower aboveground biomass of plants in the embanked S. alterni ora salt marsh ( Fig. 4; Tables 1 and S4; Wang et al. 2012). In contrast, the shade provided by the increased aboveground biomass of plants might explain the lower abundance of 4C0d-2, which are also obligately phototrophic bacteria in the embanked P. australis salt marsh ( Fig. 4; Table 1).
Chromatiaceae, which is known as phototrophic purple sulfur bacteria, can grow via photolithoautotrophic metabolism (Imhoff 2005). In this study, the presence of Chromatiaceae was negatively correlated with the aboveground biomass of plants (Table   S4). Thus, the decreased aboveground biomass of S. salsa in the embanked salt marsh explained the greater occurrence of Chromatiaceae in this environment (Tables 1 and S4) Tables 1 and 2). Additionally, in embanked regions, the abundance of Betaproteobacteria were more pronounced near to shore, rather than at more seaward locations (Table S3). Betaproteobacteria has been suggested to originate from freshwater or terrestrial ecosystems and is rarely found in seawater Indeed, in this study, the abundance of Betaproteobacteria was negatively correlated with soil salinity and positively related with the soil pH (Fig. 7). Therefore, we speculated that coastal embankments signi cantly reduced soil salinity and increased the pH by blocking seawater, which ultimately promoted the growth of Betaproteobacteria. However, Betaproteobacteria have been reported to be highly active in the decomposition of organic matter and can utilize fresh and labile substrates (Dai et al. 2021). Thus, more abundant Betaproteobacteria in the embanked regions might have given rise to losses in labile nutrients (Fig. 7). In conclusion, the establishment of coastal embankments strongly in uenced the growth of some bacterial strains by stopping seawater from reaching the plant communities, and consequently changing the physicochemical soil properties, and varying bacterial populations may cause changes in the composition of soil nutrients.

Conclusion
This study endeavored to investigate alterations in soil bacteria and archaea communities to infer the deterministic processes that drove these variations following the establishment of embankments in the saltmarshes of coastal China. Embankments signi cantly decreased soil microbial diversity in the S. alterni ora salt marsh, while increasing their OTU richness in the P. australis salt marsh. Embankments signi cantly modi ed the compositions of soil bacterial and archaeal communities in the S. alterni ora and P. australis salt marshes. However, variations in soil bacterial and archaeal diversity, richness, and community compositions between native and embanked S. salsa salt marshes were insigni cant. These results were likely due to the drastic changes in the concentrations of soil nutrient substrates that were caused by the variations in plant residues and physiochemical soil properties in the embanked S. alterni ora and P. australis salt marshes, which were not noted in the S. salsa salt marshes. This study provides further insights toward a better understanding of the variations and driving patterns of soil microbial communities following the establishment of embankments, which elucidated the effects of coastal embankments on biogeochemical cycles, and highlighted their potential hazards to ecosystems.    Absolute abundance of dominant bacteria in the unembanked and embanked Spartina alterni ora, Suaeda salsa, and Phragmites australis salt marshes. Different superscripted lower-case letters indicate p < 0.05 between the unembanked and embanked ranges in the same plant community.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. 6PLSOAppendix.docx