Electron microscopy (EM) image denoising is critical for visualization and subsequent analysis. Despite the remarkable achievements of deep learning-based non-blind denoising methods, their performance drops significantly when domain shifts exist between the training and testing data. To address this issue, unpaired blind denoising methods have been proposed. However, these methods heavily rely on image-to-image translation and neglect the inherent characteristics of EM images, limiting their overall denoising performance. In this paper, we propose the first unsupervised domain adaptive EM image denoising method, which is grounded in the observation that EM images from similar samples share common content characteristics. Specifically, we first disentangle the content representations and the noise components from noisy images and establish a shared domain-agnostic content space via domain alignment to bridge the synthetic images (source domain) and the real images (target domain). To ensure precise domain alignment, we further incorporate domain regularization by enforcing that: the pseudo-noisy images, reconstructed using both content representations and noise components, accurately capture the characteristics of the noisy images from which the noise components originate, all while maintaining semantic consistency with the noisy images from which the content representations originate. To guarantee lossless representation decomposition and image reconstruction, we introduce disentanglement-reconstruction invertible networks. Finally, the reconstructed pseudo-noisy images, paired with their corresponding clean counterparts, serve as valuable training data for the denoising network. Extensive experiments on synthetic and real EM datasets demonstrate the superiority of our method in terms of image restoration quality and downstream neuron segmentation accuracy. Our code is publicly available at https://github.com/sydeng99/DADn.

As the mainstream technique for capturing images of biological specimens at nanometer resolution, electron microscopy (EM) is extremely time-consuming for scanning wide field-of-view (FOV) specimens. In this paper, we investigate a challenging task of large-factor EM image super-resolution (EMSR), which holds great promise for reducing scanning time, relaxing acquisition conditions, and expanding imaging FOV. By exploiting the repetitive structures and volumetric coherence of EM images, we propose the first generative learning-based framework for large-factor EMSR. Specifically, motivated by the predictability ofrepetitive structures and textures in EM images, we first learn a discrete codebook in the latent space to represent highresolution (HR) cell-specific priors and a latent vector indexer to map low-resolution (LR) EM images to their corresponding latent vectors in a generative manner. By incorporating the generative cell-specific priors from HR EM images through a multi-scale prior fusion module, we then deploy multi-image feature alignment and fusion to further exploit the inter-section coherence in the volumetric EM data. Extensive experiments demonstrate that our proposed framework outperforms advanced single-image and video super-resolution methods for 8× and 16× EMSR (i.e., with 64 times and 256 times less data acquired, respectively), achieving superior visual reconstruction quality and down-stream segmentation accuracy on benchmark EM datasets. Code is available at https://github.com/jtshou/GPEMSR.

Electron microscopy (EM) images are notoriously challenging to segment due to their complex structures and lack of effective annotations. Fortunately, large-scale self-supervised pretraining offers a promising solution by allowing us to acquire prior knowledge of cell and subcellular tissue structures, which can significantly improve EM instance segmentation results. However, most existing pretraining methods fail to capture the crucial local information that is essential for EM images, instead focusing only on high-level semantic information. In this paper, we propose a novel pretraining framework that leverages multiscale visual representations to adapt to the complex structures of EM images. Our framework achieves instance-level alignment by maximizing the consistency between strongly and weakly augmented images, while also incorporating a cross-attention mechanism to match multiscale features and encode more low-level information into high-level semantics. Most importantly, our approach employs multi-task optimization on the feature pyramid, enabling multiscale pixel restoration and feature comparison. We extensively pretrain our method on four large-scale EM datasets and demonstrate significant gains on neuron and mitochondria segmentation tasks. Code is available at https://github.com/ydchen0806/MS-Con-EM-Seg.

The spatial variation of forest aboveground carbon sequestration (ACS) is crucial for assessing the carbon storage and time of carbon-emission peaking. Unfortunately, a systematic assessment of accurate ACS and its change has not yet been developed, especially at a large scale. Here, we evaluate the spatiotemporal ACS trend based on LANDIS PRO and PnET-II models from 2000 to 2020 in Western Sichuan, China, coupled with multi-source remote sensing datasets (e.g., MODIS and Lansat) and in situ forest inventory to obtain the landscape information. The Mann-Kendall trend test is used to examine the ACS trend. The total ACS increased from 344.62 to 466.99 Tg due to large-scale reforestation programs from 2000 to 2020. The evergreen coniferous forest contributed the most ACS (85.28%) in Western Sichuan compared with other species. The developed methodology can be applied to analyzing the ACS at the watershed and regional levels promptly and helping decision-makers and managers develop effective forest management measures.

Recently studies usually used remote sensing products and reanalysis datasets to obtain the precipitation. Yet, due to sparse ground observations and complex topography, the accuracy of precipitation products in mountainous regions is one of the great concerns in hydrology. Thus, assessing the accuracy and availability of precipitation products is vital to hydrological research in mountainous regions. In this study, we evaluated the precision of precipitation products including IMERG, ERA5-Land, and AERA5-Asia at multiple temporal and spatial scales using ground observations of meteorological stations from 2001 to 2015 in Southwest China, a typical mountainous region. The key findings are: (1) The IMERG datasets had the best performance on a monthly scale. (2) The IMERG and AERA5-Asia data products had higher precision at seasonal and annual scales. (3) The AERA5-Asia achieved better performance in the high- altitude areas compared to others. In addition, the ERA5- Land has the lowest accuracy. This study can provide important guidelines for the selection of precipitation products for both scientific and management purposes in mountainous areas, especially in Southwest China.

The Zagunao, as a typical alpine forested watershed in the upper Yangtze River Basin, China, plays an important role in providing valuable forest ecosystem services such as water purification, water supply, climate regulation, and carbon sequestration for downstream cities in the Chengdu plain. However, alpine forests in the Zagunao watershed are confronted with growing disturbances including climate changes and anthropogenic disturbances. These disturbances can impact ecological patterns and processes, resulting in degraded forest ecosystem functions and services. Thus, detecting degraded forests and associated driving factors in the Zagunao watershed is urgently needed. In this study, we combined the meteorological data and multiple remote sensing datasets to investigate the spatial distribution of forest degradation and its driving factors in the Zagunao watershed. The key findings are: (1) from 2001 to 2017, degraded forests accounted for 9.71% of the forested area in the Zagunao watershed, and the natural coniferous broadleaved mixed forest was the main degraded vegetation type; (2) 95% of the degradation was driven by anthropogenic activities while only 5% of it was caused by climate drivers. This study is beneficial for forestry sectors to effectively detect forest degradation and provides scientific guidance for future forest restorations and management in alpine forested watersheds.

Climate variability and vegetation disturbance as critical driving factors significantly affect the regional hydrology in forested watersheds. Yet, due to landscape heterogeneities such as topography, soil characteristics, climate conditions, and vegetation types, hydrological responses to climate and forest changes and associated mechanisms have not been fully understood. The forested Stellako watershed, a typical forest-disturbed area in Canada, has been studied to assess the annual streamflow affected by climate variability and forest disturbance. The study period was from 1951 to 2018. The methods include the modified double mass curve (MDMC), Autoregressive Integrated Moving Average (ARIMA) intervention, and multivariate ARIMA. In the Stellako watershed, 1980 was identified as a breakpoint in the yearly time series between 1951 and 2018. The 1951-1979 period was considered the reference. From 1980 to 2018, the streamflow decreased by 14.69 and 26.18 mm due to climate and forest changes, respectively. The annual runoff variation was mainly attributed to forest disturbances contributing 60.78% of the total variations. Thus, a timely assessment of the runoff variety at a watershed scale has been obtained. The developed methodology can be applied to quantify the disturbance effects at a watershed or regional scale and develop watershed and forest management strategies under future climate and forest changes.

Homogeneous instance segmentation aims to identify each instance in an image where all interested instances belong to the same category, such as plant leaves and microscopic cells. Recently, proposal-free methods, which straightforwardly generate instance-aware information to group pixels into different instances, have received increasing attention due to their efficient pipeline. However, they often fail to distinguish adjacent instances due to similar appearances, dense distribution and ambiguous boundaries of instances in homogeneous images. In this paper, we propose a pixel-embedded affinity modeling method for homogeneous instance segmentation, which is able to preserve the semantic information of instances and improve the distinguishability of adjacent instances. Instead of predicting affinity directly, we propose a self-correlation module to explicitly model the pairwise relationships between pixels, by estimating the similarity between embeddings generated from the input image through CNNs. Based on the self-correlation module, we further design a cross-correlation module to maintain the semantic consistency between instances. Specifically, we map the transformed input images with different views and appearances into the same embedding space, and then mutually estimate the pairwise relationships of embeddings generated from the original input and its transformed variants. In addition, to integrate the global instance information, we introduce an embedding pyramid module to model affinity on different scales. Extensive experiments demonstrate the versatile and superior performance of our method on three representative datasets. Code and models are available at https://github.com/weih527/Pixel-Embedded-Affinity.

Understanding the temporal variations of terrestrial water storage (TWS) in large river basins is crucial for water resource management and ecosystem protection. Yet, the variations of TWS in large river basins are often been assessed in term of temporal trends with limited studies on the stationarity of TWS. In this study, we investigated the temporal trends and stability of TWS in then large river basins in China from 2004 to 2014 using the corrected Gravity Recovery and Climate Experiment (GRACE) gravity satellite data. Key findings are: (1) the average TWS in China showed a significant downward trend during the study period; and (2) the TWS in ten large river basins was non-stationary across China. Water surpluses were observed in the Northeast and the Southeast, while water deficits were found in the Northwest and the Southwest. This study provides policy-makers with critical information to design adaptive strategies and plans for basin-scale water resources management.

Southwest China is an ecologically fragile area in China. Understanding temporal and spatial variations of vegetation degradation can help to formulate measures of ecosystem protection and ecological function restoration. In this study, we evaluated vegetation dynamics and identified vegetation degradation and its spatial patterns in Southwest China, based on land cover, DEM, and GLASS LAI datasets from 2001 to 2017. The key results are: (1) Though the average LAI in Southwest China showed an insignificant trend during the study period, based on grid-scale analysis, significant declines in LAI indicating vegetation degradation were identified in some areas such as western Sichuan, western and central Yunnan, and western Tibet; (2) about 10.75% of the vegetation experienced significant degradation during the study period in Southwest China; (3) degraded vegetation was mostly distributed in high elevation areas, and about 43% degraded vegetation was located in areas with the elevation between 3500m and 5000m; (4) the dominant degraded vegetation types included grassland, alpine vegetation, shrubland, and coniferous forest. Our findings can provide valuable management implications for vegetation restoration and ecological protection in Southwest China.

Serial section transmission electron micro-scopy (ssTEM) reveals biological information at a scale of nanometer and plays an important role in the ultrastructural analysis. However, due to the imperfect preparation of biological samples, ssTEM images are usually degraded with various artifacts that greatly challenge the subsequent analysis and visualization. In this paper, we introduce a unified deep learning framework for ssTEM image restoration which addresses three main types of artifacts, i.e., Support Film Folds (SFF), Staining Precipitates (SP), and Missing Sections (MS). To achieve this goal, we first model the appearance of SFF and SP artifacts by conducting comprehensive analyses on the statistics of real degraded images, relying on which we can then simulate a large number of paired images (degraded/artifacts-free) for training a deep restoration network. Then, we design a coarse-to-fine restoration network consisting of three modules, i.e., interpolation, correction, and fusion. The interpolation module exploits the adjacent artifacts-free images for an initial restoration, while the correction module resorts to the degraded image itself to rectify the artifacts. Finally, the fusion module jointly utilizes the above two results to further improve the restoration fidelity. Experimental results on both synthetic and real test data validate the significantly improved performance of our proposed framework over existing solutions, in terms of both image restoration fidelity and neuron segmentation accuracy. To the best of our knowledge, this is the first unified deep learning framework for ssTEM image restoration from different types of artifacts. Code is available at https://github.com/sydeng99/ssTEM-restoration.

The upper Minjiang River watershed provides critical ecological services to sustain downstream populated cities in the Chengdu plain of China. Understanding its environmental quality dynamics and associated response to vegetation change can help with vegetation restoration program designs for better protection of watershed ecosystems. In this study, five environmental quality indices were first calculated to evaluate environmental quality dynamics by use of MODIS MCD12Q1 and SPOT NDVI products. Then, responses of environment quality to vegetation change from 2001 to 2017 were assessed. Key findings were: (1) there was a slight decline in the integrated ecological index (EI), indicating a relatively stable environment quality in the study watershed; (2) EI had significant positive correlations with forest land or grassland area, while its relationship with the sparse woodland area was significantly negative; and (3) environmental quality changes were mainly driven by soil erosion, earthquake, landslides, vegetation change, and human activities. This study has important implications for the protection of environmental quality by vegetation restoration.