Traditional villages face many difficulties in the era of globalization, especially in light of fast industrialization and urbanization. The breakdown of settlement patterns and the erosion of local characteristics and cultural identities pose critical issues for the sustainable development of these communities. While research on traditional villages and dwellings in core cultural areas is relatively advanced, there remains a significant gap in studies focusing on traditional villages and dwellings in multicultural intermingling regions. By clarifying the characteristics of traditional villages and the cultural landscapes of dwellings under the influence of multiple cultures, as well as their differentiation and underlying mechanisms, this research aims to provide theoretical support for the protective planning of world cultural heritage, which is increasingly characterized by clustering and regionalization. Taking the traditional villages and dwellings in the Jiangxi and Anhui junction area as a case study, we developed a cultural landscape factor system for traditional villages and dwellings across four dimensions: natural environment, spatial configuration, dwelling typology, and historical and cultural context. Using geographic information systems (GIS) zoning methods and statistical spatial analysis, we divided the area into three distinct cultural landscape zones. The findings indicate that the cultural landscapes within each zone exhibit unique regional characteristics at both the village and dwelling levels, particularly in site selection, settlement patterns, and architectural aesthetics. Differentiation across zones is shaped by natural factors, such as topography and water systems, as well as by regional culture, historical migration, the chronological sequence of regional development, commerce and trade growth, and the evolution of administrative systems, alongside broader cultural, economic, and social factors, showing consistent patterns. This study demonstrates that utilizing a scientific and objective zoning approach to accurately identify the cultural landscape characteristics and differentiation patterns across various cultural zones, while clarifying the historical evolution of villages and the transformation of dwelling forms, provides practical insights for cultural landscape zoning in other multicultural regions. Furthermore, it provides scientific guidance to advance China’s rural revitalization strategy and supports the regional protection and sustainable development of world cultural heritage.

Hyperspectral images (HSI) contain abundant spectral information. Efficient extraction and utilization of this information for image classification remain prominent research topics. Previously, hyperspectral classification techniques primarily relied on statistical attributes and mathematical models of spectral data. Deep learning classification techniques have recently been extensively utilized for hyperspectral data classification, yielding promising outcomes. This study proposes a deep learning approach that uses polarization feature maps for classification. Initially, the polar co-ordinate transformation method was employed to convert the spectral information of all pixels in the image into spectral feature maps. Subsequently, the proposed Deep Context Feature Fusion Network (DCFF-NET) was utilized to classify these feature maps. The model was validated using three open-source hyperspectral datasets: Indian Pines, Pavia University, and Salinas. The experimental results indicated that DCFF-NET achieved excellent classification performance. Experimental results on three public HSI datasets demonstrated that the proposed method accurately recognized different objects with an overall accuracy (OA) of 86.68%, 94.73%, and 95.14% based on the pixel method, and 98.15%, 99.86%, and 99.98% based on the pixel-patch method.

The spatiotemporal patterns and shifts of net ecosystem productivity (NEP) play a pivotal role in ecological conservation and addressing climate change. For example, by quantifying the NEP information within ecosystems, we can achieve the protection and restoration of natural ecological balance. Monitoring the changes in NEP enables a more profound understanding and prediction of ecosystem alterations caused by global warming, thereby providing a scientific basis for formulating policies aimed at mitigating and adapting to climate change. The accurate prediction of NEP sheds light on the ecosystem’s response to climatic variations and aids in formulating targeted carbon sequestration policies. While traditional ecological process models provide a comprehensive approach to predicting NEP, they often require extensive experimental and empirical data, increasing research costs. In contrast, machine-learning models offer a cost-effective alternative for NEP prediction; however, the delicate balance in algorithm selection and hyperparameter tuning is frequently overlooked. In our quest for the optimal prediction model, we examined a combination of four mainstream machine-learning algorithms with four hyperparameter-optimization techniques. Our analysis identified that the backpropagation neural network combined with Bayesian optimization yielded the best performance, with an R² of 0.68 and an MSE of 1.43. Additionally, deep-learning models showcased promising potential in NEP prediction. Selecting appropriate algorithms and executing precise hyperparameter-optimization strategies are crucial for enhancing the accuracy of NEP predictions. This approach not only improves model performance but also provides us with new tools for a deeper understanding of and response to ecosystem changes induced by climate change.

Under the framework of the United Nations Sustainable Development Goals (SDGs), mitigating global PM_2.5 exposure inequality has become a worldwide task. However, the relationship between PM_2.5 exposure inequality and the SDGs remains poorly understood. In this study, we evaluated the spatial distribution of global PM_2.5 exposure inequality from 2000 to 2019 on a global scale and investigated the interactions between PM_2.5 exposure inequality and the SDGs. The results indicate that global PM_2.5 exposure inequality is increasingly worsening. Compared with high-income and low-income regions, middle- and lower-income regions have higher levels of PM_2.5 exposure inequality and face greater risks of PM_2.5 exposure and associated health impacts. Our analysis of the relationship between PM_2.5 exposure inequality and the SDGs reveals that the positive impact of air-SDGs related to air pollution and health, such as SDG3, SDG5, SDG8, and SDG13, on global PM_2.5 exposure inequality is becoming increasingly significant. In contrast, the negative impact of SDG2, SDG7, and SDG10 on global PM_2.5 exposure inequality has been growing annually. These findings provide valuable insights into the complex interactions between PM_2.5 exposure inequality and the SDGs and offer a reference for future efforts to build healthier and more equitable urban air environments.

Hyperspectral images (HSI) contain abundant spectral information. Efficient extraction and utilization of this information for image classification remain prominent research topics. Previously, hyperspectral classification techniques primarily relied on statistical attributes and mathematical models of spectral data. Deep learning classification techniques have recently been extensively utilized for hyperspectral data classification, yielding promising outcomes. This study proposes a deep learning approach that uses polarization feature maps for classification. Initially, the polar coordinate transformation method is employed to convert the spectral information of all pixels in the image into spectral feature maps. Subsequently, the proposed Deep Context Feature Fusion Network (DCFF-NET) is utilized to classify these feature maps. The model is validated using three open-source hyperspectral datasets: Indian Pines, Pavia University, and Salinas. The experimental results indicated that DCFF-NET achieves excellent classification performance. Experimental results on three public HSI datasets demonstrate that the proposed method accurately recognizes different objects with an overall accuracy (OA) of 86.68%, 94.73%, and 95.14% based on the pixel method, and 98.15%, 99.86%, and 99.98% based on the pixel-patch method.