The coastal zone presents complex hydrodynamic interactions among inland groundwater, reservoir water, and intruding seawater, with important implications for ecosystem functioning and water quality. However, the relative roles of hydraulic connectivity and seawater-driven salinity gradients in shaping microbial communities at the aquifer–reservoir interface remain unclear. Here, we integrated hydrochemical analyses with high-throughput 16S rRNA gene sequencing to investigate bacterial community composition, assembly processes, and co-occurrence network patterns across groundwater_in (entering the reservoir), groundwater_out (exiting the reservoir), and reservoir water in a coastal system. Our findings reveal that seawater intrusion exerts a stronger influence on groundwater_out, leading to distinct chemical profiles and salinity-driven environmental filtering, whereas hydraulic connectivity promotes greater microbial similarity between groundwater_in and reservoir water. Groundwater samples exhibited higher alpha and beta diversity compared to the reservoir, with dominant taxa such as Comamonadaceae, Flavobacteriaceae, and Rhodobacteraceae serving as indicators of seawater intrusion. Community assembly analyses showed that homogeneous selection predominated, especially under strong salinity gradients, while dispersal limitation and spatial distance also contributed in areas of reduced connectivity. Key chemical factors, including TDS, Na⁺, Cl⁻, Mg²⁺, and K⁺, strongly shaped groundwater communities. Additionally, groundwater bacterial networks were more complex and robust than those in reservoir water, suggesting enhanced resilience to salinity stress. Collectively, this study demonstrates that salinity gradients can override the effects of hydraulic connectivity in structuring bacterial communities and their networks at coastal interfaces. Our findings provide novel microbial insights relevant for understanding biogeochemical processes and support the use of microbial indicators for more sensitive monitoring and management of coastal groundwater resources.

Reservoir water and the adjacent soil are ecologically interconnected yet distinct microhabitats in saline coastal wetland ecosystems, but direct comparisons of their bacterial community composition and assembly remain limited. Here, we integrated high-throughput 16S rRNA gene sequencing with statistical, null model, and network analyses to compare diversity patterns, assembly mechanisms, and interactions of abundant and rare bacterial taxa in both habitats. Soil communities exhibited greater taxonomic diversity but a lower overall abundance, while reservoir communities displayed a pronounced vertical stratification, in contrast to the more spatially uniform soil communities at the sampled scale. Key environmental drivers differed: salinity (reflecting the harsh saline context) and nutrient levels structured reservoir communities, whereas the nutrient availability and cation exchange capacity predominated in soils. Stochastic processes mainly governed the assembly of abundant taxa in both habitats, whereas deterministic selection more strongly structured rare taxa, especially in soils subject to harsh saline conditions. The co-occurrence network analysis revealed higher connectivity and modularity in soils, with moderate taxa acting as critical connectors between modules. In contrast, rare taxa played a pivotal role in sustaining network stability in the reservoir. Together, these findings demonstrate distinct, habitat-dependent assembly mechanisms and ecological roles of abundant and rare bacterial taxa in saline coastal wetland microhabitats, providing insights that can inform wetland conservation and ecosystem management.

Reservoir and adjacent soil are ecologically interconnected yet distinct microhabitats in wetland ecosystems, but direct comparisons of their bacterial community composition and assembly remain limited. In this study, we integrated high-throughput 16S rRNA gene sequencing with statistical, null model, and network analyses to compare the diversity patterns, assembly mechanisms, and interactions of abundant and rare bacterial taxa in both habitats. Soil communities exhibited greater taxonomic diversity but lower overall abundance, while reservoir communities displayed pronounced vertical stratification and soil communities were more spatially homogeneous. Key environmental drivers differed: salinity and nutrient levels structured reservoir communities, whereas nutrient availability and cation exchange capacity dominated in soil. Stochastic processes mainly governed the assembly of abundant taxa in both environments, whereas deterministic selection was more prominent in structuring rare taxa, especially in soil. Co-occurrence network analysis revealed higher connectivity and modularity in soils, with moderate taxa acting as critical connectors between modules, while rare taxa played a pivotal role in sustaining the network stability in reservoir. These findings underscore that distinct assembly mechanisms and ecological roles of abundant and rare bacterial taxa in different wetland microhabitats, providing novel insights to inform effective strategies for wetland conservation and scientific ecosystem management.

Dissolved organic nitrogen (DON) is an important component of the aquatic nitrogen pool, which is composed of countless molecular individuals with different ecological functions. Yet, its sources and molecular composition remain poorly understood in reservoir ecosystems. In this study, water samples were collected from different regions of the Beidagang Reservoir. Hydrochemical measurements and Fourier transform ion cyclotron resonance mass spectrometry (FT-ICR MS) were used to characterize the molecular composition, spatial distribution, and potential sources of DON. The results demonstrated that DON constituted a substantial fraction of total dissolved nitrogen and exhibited pronounced spatial variability in both concentration and molecular composition. CHO and CHON formulas were abundant across the reservoir, with carboxylic-rich alicyclic molecules (CRAM-like) representing the dominant molecular class. A total of 16,338 CHON formulas were identified, of which only 407 were shared by all samples, indicating pronounced spatial variation in DON composition. Paired mass distance (PMD) analysis further showed that approximately 60% of the DON formulas could be classified as potential precursors or products. Amine reactions accounted for approximately 40% of all putative transformation pairs, suggesting that transformations involving nitrogen functional groups may represent an important pathway during DON transport. Upstream DON was likely influenced by the combined effects of surface runoff, atmospheric deposition, and internal reservoir processes, whereas downstream DON appeared to be more strongly affected by anthropogenic activities and seawater-related sources. These findings revealed the molecular diversity and potential transformation pathways of DON in the Beidagang Reservoir and provided molecular-level evidence for understanding nitrogen cycling in reservoirs exposed to multiple environmental inputs, which helped predict the response of nitrogen cycling to global climate change.