In an increasingly interconnected world, human–environment interactions involving flows of people, organisms, goods, information, and energy are expanding in magnitude and extent, often over long distances. As a universal paradigm for examining these interactions, the telecoupling framework (published in 2013) has been broadly implemented across the world by researchers from diverse disciplines. We conducted a systematic review of the first five years of telecoupling research to evaluate the state of telecoupling science and identify strengths, areas to be improved, and promising avenues for future study. We identified 89 studies using any derivation of the term telecoupling. These works emphasize trade flows, information transfer, and species dispersal at international, national, and regional scales involving one or a few countries, with China, Brazil, and the United States being the most frequently studied countries. Our review showed a rising trend in publications and citations on telecoupling, with 63% of identified telecoupling studies using the framework’s specific language (e.g., “flows”, “agents”). This result suggests that future telecoupling studies could apply the standardized telecoupling language and terminology to better coordinate, synthesize, and operationalize interdisciplinary research. Compelling topics for future research include operationalization of the telecoupling framework, commonalities among telecouplings, telecoupling mechanisms and causality, and telecoupled systems governance. Overall, the first five years of telecoupling research have improved our understanding of human–environment interactions, laying a promising foundation for future social–ecological research in a telecoupled world.

In recent decades, the wide use of deep learning-based methods has consistently improved the performance of remote sensing images and is widely used for hyperspectral change detection (HCD) tasks. However, most of the existing HCD method is based on the convolutional neural network (CNN), which shows limitations in long-range dependencies and also cannot mine sequence features well. The CD performance still has margins for improvement. In this study, inspired by the excellent performance of transformers in computer vision and which has shown a significant ability to model global dependencies to attenuate the loss of long-range information, we built a hybrid spatial-spectral convolutional vision transformer (SSViT) for HCD. Our proposed method combines the merits of CNN and transformer to fulfill effective and efficient HCD. This study focused on highly reliable pseudo-sample data generation by selection scenario. To generate a pseudo sample, we have used different methods: (1) we predict change and no-change areas by using Euclidean distance, (2) thresholding by Chan-Vese segmentation method for determining change and no-change pixels for intensity maps, (3) sorting of change and no-change pixels, and (4) selection of the minimum value of initial no-change pixels as pseudo change sample data, in addition to, choosing the maximum intensity value for change candidate pixels as change sample data. The highly reliable change pixels were selected, and then pseudo-training data was used to train the SSViT model. At last, the change map is generated by training the SSViT network based on pseudo-training data. The performance of the SSViT model is evaluated for real-world hyperspectral (HS) datasets with different change landcover types. Furthermore, a new series of HS images is introduced for CD purposes. The results of CD show that the HS images have a high potential for detecting subtle changes. The experimental results demonstrate that the proposed SSViT could outperform the advanced HCD methods.

Extensive, severe wildfires, and wildfire‐induced smoke occurred across the western and central United States since August 2020. Wildfires resulting in the loss of habitats and emission of particulate matter and volatile organic compounds pose serious threatens to wildlife and human populations, especially for avian species, the respiratory system of which are sensitive to air pollutions. At the same time, the extreme weather (e.g., snowstorms) in late summer may also impact bird migration by cutting off their food supply and promoting their migration before they were physiologically ready. In this study, we investigated the environmental drivers of massive bird die‐offs by combining socioecological earth observations data sets with citizen science observations. We employed the geographically weighted regression models to quantitatively evaluate the effects of different environmental and climatic drivers, including wildfire, air quality, extreme weather, drought, and land cover types, on the spatial pattern of migratory bird mortality across the western and central US during August‐September 2020. We found that these drivers affected the death of migratory birds in different ways, among which air quality and distance to wildfire were two major drivers. Additionally, there were more bird mortality events found in urban areas and close to wildfire in early August. However, fewer bird deaths were detected closer to wildfires in California in late August and September. Our findings highlight the important impact of extreme weather and natural disasters on bird biology, survival, and migration, which can provide significant insights into bird biodiversity, conservation, and ecosystem sustainability.

In this article, we propose a novel difference image (DI) creation method for unsupervised change detection in multi-temporal multi-spectral remote-sensing images based on deep learning theory. First, we apply deep belief network to learn local and high-level features from the local neighbour of a given pixel in an unsupervised manner. Second, a back propagation algorithm is improved to build a DI based on selected training samples, which can highlight the difference on changed regions and suppress the false changes on unchanged regions. Finally, we get the change trajectory map using simple clustering analysis. The proposed scheme is tested on three remote-sensing data sets. Qualitative and quantitative evaluations show its superior performance compared to the traditional pixel-level and texture-level-based approaches.