Artificial reefs (ARs) have been widely used to restore the seabed habitat and protect biodiversity. They can effectively increase the dissolved oxygen content in the bottom water layer by their disturbing effect of upwelling and downwelling. The bottom water is prone to hypoxia in summer due to the extreme weather of the global climate and excessive biomass in some marine ranching in northern China. Therefore, how to effectively use the upwelling effect of artificial reefs to alleviate this problem is a necessary subject of research. Generally, ARs are arranged by different intervals in a unit form on the seafloor, and the flow field effect is different from that of the individual reefs. However, few studies have been focused on the effect of layout mode on the flow field of a unit reef (UR). In this paper, we selected the interval between reefs (IR) and the angle of inflow (AI) as the influencing factors to study the flow field effect of UR. The upwelling and wake regions of 64 URs were presented by the efficiency and disturbance indices related to the flow characteristics and proposed an optimal layout mode having the best performance of the upwelling effect. The results showed that the interactions among the AI, the transverse, and longitudinal IRs were significant, and the AI has a significant influence on the flow field. These indices were effective and contribute to the layout optimization of UR. The AI close to 45° has a significant influence on the flow field effect of UR.

Concrete and wood are commonly used to manufacture artificial reefs (ARs) worldwide for marine resource enhancement and habitat restoration. Although microbial biofilms play an important role in marine ecosystems, the microbial communities that colonize concrete and wooden ARs and their temporal succession have rarely been studied. In this study, the temporal succession of the microbial communities on concrete and wooden AR blocks and the driving factors were investigated. The composition of the microbial communities underwent successive shifts over time: among the six dominant phyla, the relative abundances of Proteobacteria, Cyanobacteria and Gracilibacteria significantly decreased in wood, as did that of Cyanobacteria in concrete. Operational taxonomic units (OTU) richness and Shannon index were significantly higher in concrete than in wood. Non-metric multidimensional scaling ordination placed the microbial communities in two distinct clusters corresponding to the two substrate materials. The macrobenthic compositions of concrete and wood were broadly similar and shifted over time, especially in the first five weeks. The Shannon index of the microbial communities in concrete and wood increased significantly with the organism coverage. The results provide fundamental data on microbial community succession during the initial deployment of ARs and contribute to understanding the ecological effects of ARs.

The anaerobic membrane bioreactors (AnMBR) with ring membrane module were operated under mesophilic temperature (M-AnMBR) and atmospheric temperature (A-AnMBR). Compared to the M-AnMBR, the removal efficiency of the A-AnMBR was found to be lower and the faster membrane fouling occurred in the A-AnMBR under corresponding hydraulic retention time (HRT). The MiSeq high-throughput sequencing was applied to analyze the microbial community structure. The HRT change had different effects on the community richness and diversity of the cake and bulk sludge. The abundance of phylum Proteobacteria in the M-AnMBR was higher than that in the A-AnMBR, which should account for the higher removal of nutrients in the M-AnMBR. The faster membrane fouling would occur in the A-AnMBR due to the relatively high abundance of Bacteroidetes in the bulk sludge and cake sludge. Moreover, specific comparison down to the genus level showed that the dominant abundant bacterial genera were Candidate division OP8 norank and Anaerolineaceae uncultured in the cake sludge for M-AnMBR, and were VadinHA17 norank, WCHB1-69 norank, VadinBC27 wastewater-sludge group, and Synergistaceae uncultured in the cake sludge for A-AnMBR The different representative genera with the variation of the HRTs for the two bioreactors might indicate the different performance between the two AnMBRs.

In deep-learning-based speech enhancement (SE) systems, trained models are often used to handle unseen noise types and language environments in real-life scenarios. However, since production environments differ from training conditions, mismatch problems arise that may cause a serious decrease in the performance of an SE system. In this study, a domain adaptive method combining two adaptation strategies is proposed to improve the generalization of unlabeled noisy speech. In the proposed encoder-decoder-based SE framework, a domain discriminator and a domain confusion adaptation layer are introduced to conduct adversarial training. The model has two main innovations. First, the algorithm optimizes adversarial training by introducing a relativistic discriminator that relies on relative values by applying the difference, thus avoiding possible bias and better reflecting domain differences. Second, the multi-kernel maximum mean discrepancy (MK-MMD) between domains is taken as the regularization term of the domain adversarial loss, thereby further decreasing the edge distribution distance between domains. The proposed model improves the adaptability to unseen noises by encouraging the feature encoder to generate domain-invariant features. The model was evaluated using cross-noise and cross-language-and-noise experiments, and the results show that the proposed method provides considerable improvements over the baseline without an adaptation in the perceptual evaluation of speech quality (PESQ), the short time objective intelligibility (STOI) and the frequency-weighted signal-to-noise ratio (FWSNR).

A microalgae membrane bioreactor (MMBR) with internal circulating fluidized bed (ICFB) was constructed at room temperature to study the removal efficiency of marine aquaculture wastewater pollutants and continuously monitor the biomass of microalgae. Within 40 days of operation, the removal efficiency of NO₃⁻–N and NH₄⁺–N in the ICFB-MMBR reached 52% and 85%, respectively, and the removal amount of total nitrogen (TN) reached 16.2 mg/(L·d). In addition, the reactor demonstrated a strong phosphorus removal capacity. The removal efficiency of PO₄³⁻–P reached 80%. With the strengthening of internal circulation, the microalgae could be distributed evenly and enriched quickly. The maximum growth rate and biomass concentration reached 60 mg/(L·d) and 1.4 g/L, respectively. The harvesting of microalgae did not significantly affect the nitrogen and phosphorus removal efficiency of ICFB-MMBR. The membrane fouling of the reactor was investigated by monitoring transmembrane pressure difference (TMP). Overall, the membrane fouling cycle of ICFB-MMBR system was more than 40 days.