No abstract

No abstract

In recent years, the accelerated pace of urbanization has increased patch fragmentation, which has had a certain impact on the structure and ecological environment of forest–grass ecological networks, and certain protection measures have been taken in various regions. Therefore, studying the spatiotemporal changes and correlations of ecological service functions and forest–grass ecological networks can help to better grasp the changes in landscape ecological structure and function. This paper takes the Wuding River Basin as the research area and uses the windbreak and sand fixation service capacity index, soil conservation capacity, and net primary productivity (NPP) to evaluate the ecological service capacity of the research area from the three dimensions of windbreak and sand fixation, soil conservation, and carbon sequestration. The Regional Sustainability and Environment Index (RSEI) is used to extract ecological source areas, and GIS spatial analysis and the minimum cumulative resistance (MCR) model are used to extract potential ecological corridors. Referring to complex network theory, topology metrics such as degree distribution and clustering coefficient are calculated, and their correlation with ecological service capacity is explored. The results show that the overall ecological service capacity of sand fixation, soil fixation, and carbon sequestration in the research area in 2020 has increased compared to 2000, and the ecological flow at the northern and northwest boundaries of the river basin has been enhanced, but there are still shortcomings such as fragmented ecological nodes, a low degree of clustering, and poor connectivity. In terms of the correlation between topology indicators and ecological service functions, the windbreak and sand fixation service capacity index have the strongest correlation with clustering and the largest grasp, while the correlation between soil conservation capacity and eigencentrality is the strongest and has the largest grasp. The correlation between NPP and other indicators is not obvious, and its correlation with eccentricity and eigencentrality is relatively large.

Ultrasound has been used for antifouling on the surface of medical devices or food utensils, but it is rarely applied in marine anti-biofouling on underwater instruments. To understand whether ultrasonic antifouling is suitable for underwater optical windows, the effect of ultrasonic conditions including frequency, power and duration on the removal of microbiofouling on the surface of polymethyl methacrylate (PMMA), a type of common optical material, was investigated in this study by three-factor and three-level orthogonal experiments. Before and after the ultrasonic treatment, both surface morphology and fouling degree of PMMA samples immersed in Escherichia coli suspension and seawater were characterized and quantified using laser scanning microscope. The results showed that ultrasonic treatment can effectively remove microfouling from the PMMA surface under suitable conditions. Ultrasonic technology has a great potential for the control of microfouling on the marine optical instruments. When compared with power and duration, ultrasonic frequency has a more significant effect on antifouling efficacy of ultrasound. It is useful for PMMA samples exposed to seawater within 2 days to conduct an antifouling treatment under the condition of an ultrasonic frequency of 20 kHz, ultrasonic power of 40 W, and ultrasonic duration of 7 min.This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Graph neural networks (GNNs) have been widely applied to numerous fields. A recent work which combines layered structure and residual connection proposes an improved deep architecture to extend CAmouflage-REsistant GNN (CARE-GNN) to deep models named as Residual Layered CARE-GNN (RLC-GNN), which forms self-correcting and incremental learning mechanism, and achieves significant performance improvements on fraud detection task. However, we spot three issues of RLC-GNN, which are the usage of neighboring information reaching limitation, the training difficulty which is inherent problem to deep models and lack of comprehensive consideration about node features and external patterns. In this work, we propose three approaches to solve those three problems respectively. First, we suggest conducting similarity measure via cosine distance to take both local features and external patterns into consideration. Then, we combine the similarity measure module and the idea of adjacency-wise normalization with node-wise and batch-wise normalization and then propound partial neighborhood normalization methods to overcome the training difficulty while mitigating the impact of too much noise caused by high-density of graph. Finally, we put forward intermediate information supplement to solve the information limitation. Experiments are conducted on Yelp and Amazon datasets. And the results show that our proposed methods effectively solve the three problems. After applying the three methods, we achieve 4.81%, 6.62% and 6.81% improvements in the metrics of recall, AUC and Macro-F1 respectively on the Yelp dataset. And we obtain 1.65% and 0.29% improvements in recall and AUC respectively on the Amazon datasets.