Sustainable aquaculture has gained much attention in recent years. It is still lacking in the ecological dynamics of the microbial communities in the pond-rice field recirculating systems. The integrated systems can combine aquaculture with the constructed wetlands in the nutrient cycling and water quality. However, only a little is known about the succession of the microbial communities, their interaction patterns, and the process shaping their assembly over time and space. In this study, a systematic investigation was carried out to explore the spatiotemporal variation in the planktonic bacterial communities in such systems. An attempt was also made on their responses to environmental changes and the relative contributions of the deterministic and stochastic factors in the community assembly. A representative rice–fish co-culture system was selected as the experimental model. The recirculating water flow was divided into three functional zones: the inlet section (receiving nutrient-rich pond effluent), the field section (a rice paddy wetland functioning as a biofilter), and the outlet section (a discharge area with reduced nutrient concentrations). Two fixed sampling sites were established in each zone. The water samples were collected monthly in the peak period of the rice cultivation and fish farming from April to August 2022. Temporal dynamics were captured after sampling. High-throughput sequencing targeting the 16S rRNA gene was employed to identify the bacterial taxa. The community composition, richness, and evenness were assessed after comprehensive bioinformatic and statistical analysis. Furthermore, the alpha and beta diversity indices, principal coordinate analysis (PCoA) with Bray–Curtis dissimilarity, and canonical correspondence analysis (CCA) were utilized to explore the relationships between bacterial communities and environmental variables. In addition, the bacterial co-occurrence networks were constructed to examine the microbial interaction patterns. Ecological assembly was then inferred using both neutral community models and niche-based frameworks. The relative influence of the stochastic and deterministic factors was evaluated to shape the bacterial community structure. The results indicated that the rice field and constructed wetland played a crucial role in the nutrient removal and water quality. The average removal rates of the total phosphorus (TP), total nitrogen (TN), ammonia nitrogen (NH₃-N), nitrate nitrogen (NO₃⁻-N), and nitrite nitrogen (NO₂⁻-N) were 65.5%, 47.3%, 51.2%, 56.4%, and 53.3%, respectively, indicating the strong purification of the system. Temporally, there was a significantly seasonal variation in the composition of the bacterial community (P<0.05), particularly in response to the environmental gradients, such as the temperature and nutrient concentrations. Spatially, the overall structure of the bacterial community remained relatively consistent across the three sections; However, the bacterial taxa shared the localized changes in the relative abundance and community evenness. Proteobacteria, Actinobacteria, and Bacteroidetes were the dominant phyla in most samples. Co-occurrence network analysis revealed that the field section also exhibited the most complex network of the bacterial interaction, indicating the high modularity, centrality, and clustering coefficients. There was the strong environmental filtering and pronounced ecological niche differentiation in the rice wetland. The environmental variables, NH₃-N, NO₃⁻-N, and NO₂⁻-N were measured to identify them as the key driving factors on the bacterial community shifts. Moreover, the ecological modeling showed a gradual increase in the contribution of the stochastic factors along the direction of water flow. There was a shift from the predominantly deterministic assembly near the inlet to the more random community assembly downstream. The environmental filtering intensity was reduced to increase the microbial dispersal or ecological drift. In conclusion, this finding can provide a comprehensive overview of the spatiotemporal patterns and ecological assembly of the bacterial communities in an integrated rice-fish farming system. The theoretical insights can be offered to understand the influence of the environmental factors on the bacterial community structure in the recirculating agricultural wetlands. As such, the microbial community can also be optimized to support the ecological design and management of the integrated rice-fish farming systems.

The soil and water assessment tool (SWAT) has been widely used in distributed-parameter hydrological models in order to simulate agricultural non-point source pollution (NPS) and mitigation strategies. Although the SWAT has presented significant effectiveness in the humid and semi-humid areas, great difficulties still remain in the dry regions. Particularly, the essential parameters are often required to calibrate under water-limited settings. Some challenges are also attributed to the model's reduced depiction of the hydrological response, including the soil moisture dynamics and surface runoff production. In this study, a hybrid framework was established to integrate the "physics-constrained + data-driven" approaches. The SWAT model was also combined with a deep learning architecture, namely a convolutional neural network paired with a long short-term memory network (CNN-LSTM). An enhanced particle swarm optimization technique (IPSO) was included in this linked framework to optimize three elements: the CNN-LSTM network architecture, its hyperparameters, as well as the daily fusion weight that was employed to linearly amalgamate SWAT and CNN-LSTM outputs. Furthermore, the IPSO assessed the error between the fused prediction series and observed pollutant concentrations at each iteration using adaptive inertia weighting and stochastic perturbations. Thereafter, the particle placements were dynamically adjusted in the parameter space. The SWAT and CNN-LSTM models were then used to rotate daily between major (master) and secondary (auxiliary) roles. The error correction markedly diminished the bias in real time. The individual performance of the SWAT and CNN-LSTM was initially evaluated to compare the combined SWAT-IPSO-CNN-LSTM system on daily timescales. Subsequently, the efficacy of IPSO after hyperparameter optimization was assessed against nine metaheuristic algorithms. The Ningxia section of the Yellow River basin was selected as the research region. A multi-scale investigation was implemented on the watershed nitrogen and phosphorus dynamics. Total nitrogen (TN) and total phosphorus (TP) loads were simulated, and then the enhancement of the coupled model was simulated. The results show that better performance was achieved over the individual SWAT configuration. In TN simulations, the coefficient of determination (R²), and the Nash–Sutcliffe efficiency (NSE) rose by 14.1% and 14.5%, respectively, while the percent bias (PBIAS) decreased by 38.6%; In TP simulations, the R² and NSE were enhanced by 10.7% and 12.0%, respectively, while there was a 65.3% reduction in the PBIAS. Moreover, the centered root-mean-square error (CRMSE) decreased from 0.233 (TN) and 0.192 (TP) in the uncoupled SWAT simulations to 0.117 and 0.127, respectively, in the linked framework, indicating a significant reduction in the prediction error. A spatiotemporal study of the coupled model demonstrated that the peak TN (57% of annual maximum) and TP (68%) loads on the wet season were dominated by the synergistic impacts of the intense precipitation and fertilizer application. The NPS levels were relatively slightly reduced in the dry season, due mainly to the compensatory winter irrigation. At the sub-watershed level, the southern areas exhibited the pollution peaks that linked to the rainfall-induced runoff, whereas the northern irrigation districts shared a bimodal pollution pattern mostly driven by intensive agriculture practices. Specifically, the irrigation district of the Yellow River diversion accounted for 31%-37% of the TN and TP loads among river segments; The largest pollutant export per unit area was observed in the Hongliugou and Kushuihe segments, due to the concentrated agropastoral activities. The coupled SWAT-IPSO-CNN-LSTM technique can be expected to efficiently mitigate the calibration uncertainty in a dry watershed. Its adaptive error-correction can accurately and rapidly forecast the daily nutrient pollution. Thus, this hybrid framework can enhance the technological assistance for the water environment in drought-prone areas.