Suffering from speckle noise and complex scattering phenomena, classification results of SAR images are usually noisy and shattered, which makes them difficult to use in practical applications. Deep-learning-based semantic segmentation realizes segmentation and categorization at the same time, and thus can obtain smooth and fine-grained classification maps. However, this kind of methods require large data sets with pixel-wise categorical annotations, which are time consuming and tedious to retrieve. Compared with photographs and optical remote sensing images, manually annotating SAR data is even harder, which results in a delay of using relevant techniques in this field. In this letter, a new data set is proposed to support semantic segmentation for high-resolution PolSAR images. Limited by the aforementioned problems, the data set is only a small one with 50 image patches. Therefore, two transfer learning strategies are proposed, which adopt the fully convolutional network (FCN) and U-net architecture, respectively, and use distinct pretraining data sets to adapt to different situations. The experiments demonstrate the good performance of both methods and a promising applicability of using small training sets. Moreover, although trained with small patches, both networks can perfectly apply on large images. The new data set and methods are hopeful to support various PolSAR applications as baselines.

World Heritage sites in general are exposed to the impacts of natural hazards, which threaten their integrity and may compromise their value. Floods are a severe threat to the Angkor World Heritage site. Studies of regional floods and flood hazard zoning have played an increasingly important role in ensuring sustainability of the Angkor site. This study developed a flood hazard index (FHI) model based on a geographic information system (GIS) and used synthetic aperture radar (SAR) data to extract historical floods at Angkor from 2007 to 2013. Four indices (flood affected frequency, absolute elevation, elevation standard deviation, drainage density) were used to identify flood-prone areas. The Analytic Hierarchy Process (AHP) and the Delphi method were employed to determine the weight of each index. The weighted indices were then used to develop a distribution map of flood hazards at Angkor. The results show that 9 monuments are at risk by potential floods among the 52 components of the Angkor monuments. The high hazard and moderate-to-high hazard areas in the core zone are mainly located surrounding the West Baray but will not bring direct risk impact on the monuments located in the core archaeological zone. The moderate hazard areas are located on both sides of the Siem Reap and Roluos rivers and in the flooded area of the Tonle Sap Lake in the core archaeological zone. These areas cover 19.4 km², accounting for 9.13% of the total area of the core zone. This moderate hazard zone poses a greater flood threat to the core zone and must be given higher attention. The buffer zone is a small area with fewer sites. As such, flooding has a low impact on the buffer zone. The methods used in this study can be applied to flood hazard assessments of other heritage sites in Southeast Asia.

The Yellow River Delta (YRD) wetland is one of the largest and youngest wetland ecosystems in the world. It plays an important role in regulating climate and maintaining ecological balance in the region. This study analyzes the spatiotemporal changes in land use, wetland migration, and landscape pattern from 2013 to 2022 using Landsat-8 and Sentinel-1 data in YRD. Then wetland landscape changes and the impact of human activities are determined by analyzing correlation between landscape and socio-economic indicators including nighttime light centroid, total light intensity, cultivated land area and centroid, building area and centroid, economic and population. The results show that the total wetland area increased 1426 km² during this decade. However, the wetland landscape pattern tended to be fragmented from 2013 to 2022, with wetlands of different types interlacing and connectivity decreasing, and distribution becoming more concentrated. Different types of human activities had influences on different aspects of wetland landscape, with the expansion of cultivated land mainly compressing the core area of wetlands from the edge, the expansion of buildings mainly disrupting wetland connectivity, and socio-economic indicators such as total light intensity and the centroid mainly causing wetland fragmentation. The results show the changes of the YRD wetland and provide an explanation of how human activities effect the change of its landscape, which provides available data to achieve sustainable development goals 6.6 and may give an access to measure the change of wetland using human-activity data, which could help to adject behaviors to protect wetlands.

Biodiversity in China coexists alongside large ethnically diverse rural human populations within linked ‘biocultural’ social–ecological systems. Cumulative and changing local anthropogenic pressures linked to human population growth and economic development are driving unsustainable resource exploitation, habitat loss and extensive species declines. However, these pressures vary between species and systems, and efforts to conserve regional biodiversity may also have unintended negative impacts on local communities and their long‐term relationship with nature. Based on our extensive experience working across China's diverse but vulnerable biocultural landscapes, we showcase a series of case studies that illustrate the differing interactions between biodiversity and local community use of natural resources, and the differing consequences of conservation management on local livelihoods. We highlight that some social–ecological relationships in China can maintain and support biodiversity, and we advocate for management underpinned by interdisciplinary conservation research that engages local communities, associated with robust evidence‐based assessment and evaluation frameworks to ensure effective monitoring and optimization of impacts. We frame a series of steps required to understand and mitigate local pressures on threatened species in China and identify potential ‘win–win’ approaches for the regional maintenance of biodiversity, ecosystem services and sustainable livelihoods, with directions proposed for future research, engagement and management. Read the free Plain Language Summary for this article on the Journal blog.

Worldwide, there are only a handful of reintroduction programs for threatened salamander species, and very few have conducted postrelease studies to examine survival, habitat selection, and dispersal. Limitations in postrelease monitoring are primarily due to size constraints of amphibians and to dimensions of the radiotransmitters available for implantation. However, due to the large size of the critically endangered Chinese giant salamander (Andrias davidianus), these animals make optimal candidates for surgical implantation of radiotransmitters prior to reintroduction or translocation. The objective of this study was to develop an anesthetic protocol using tricane methanesulfonate (MS-222) and test a surgical procedure for coelomic implantation of radiotransmitters for this species. A total of 32 Chinese giant salamanders from two age groups (Group A: 4.7 yr old, n = 16; Group B: 2.7 yr old, n = 16) were implanted with 4-g radiotransmitters designed for underwater monitoring of fish. Group A w...