Forests are the most important carbon reservoirs on land, and forest carbon sinks can effectively reduce atmospheric CO${}_2$ concentrations and mitigate climate change. In recent years, various satellites have been launched that provide opportunities for identifying forest types with low cost and high time efficiency. Using multi-temporal remote sensing images and combining them with vegetation indices takes into account the vegetation growth pattern and substantially improves the identification accuracy, but it has high requirements for imaging, such as registration, multiple times, etc. Sometimes, it is difficult to satisfy, the plateau area is severely limited by the influence of clouds and rain, and Gaofen (GF) data require more control points for orthophoto correction. The study area was chosen to be Huize County, situated in Qujing City of Yunnan Province, China. The analysis was using the GF and Landsat images. According to deep learning and remote sensing image feature extraction methods, the semantic segmentation method of F-Pix2Pix was proposed, and the domain adaptation method according to transfer learning effectively solved the class imbalance in needleleaf/broadleaf forest identification. The results showed that (1) this method had the best performance and a higher accuracy than the existing products, 21.48% in non-forest/forest and 29.44% in needleleaf/broadleaf forest for MIoU improvement. (2) Applying transfer learning domain adaptation to semantic segmentation showed significant benefits, and this approach utilized satellite images of different resolutions to solve the class imbalance problem. (3) It can be used for long-term monitoring of multiple images and has strong generalization. The identification of needleleaf and broadleaf forests combined with the actual geographical characteristics of the forest provides a foundation for the accurate estimation of regional carbon sources/sinks.

The outbreak and prevalence of SARS-CoV-2 have severely affected social security. Physical isolation is an effective control that affects the short-term human-to-human transmission of the epidemic, although weather presents a long-term effect. Understanding the effect of weather on the outbreak allow it to be contained at the earliest possible. China is selected as the study area, and six weather factors that receive the most attention from January 20, 2020 to April 30, 2020 are selected to investigate the correlation between weather and SARS-CoV-2 to provide a theoretical basis for long-term epidemic prevention and control. The results show that (1) the average growth rate (GR) of SARS-CoV-2 in each province is logarithmically distributed with a mean value of 5.15%. The GR of the southeastern region is higher than that of the northwestern region, which is consistent with the Hu Line. (2) The specific humidity, 2-m temperature (T), ultraviolet (UV) radiation, and wind speed (WS) adversely affect the GR. By contrast, the total precipitation (TP) and surface pressure (SP) promote the GR. (3) For every 1 unit increase in UV radiation, the GR decreases by 0.30% in 11 days, and the UV radiation in China is higher than that worldwide (0.92% higher per day). Higher population aggregation and urbanization directly affect the epidemic, and weather is an indirect factor.

Monitoring CO₂ concentrations is believed to be an effective measure for assisting in the control of greenhouse gas emissions. Satellite measurements compensate for the sparse and uneven spatial distribution of ground observation stations, allowing for the collection of a wide range of CO₂ concentration data. However, satellite monitoring’s spatial coverage remains limited. This study fills the knowledge gaps of column-averaged dry-air mole fraction of CO₂ (XCO₂) products retrieved from the Greenhouse Gases Observing Satellite (GOSAT) and Orbiting Carbon Observatory Satellite (OCO-2) based on the normalized output of atmospheric chemical models, WRF-Chem, in Southern China during 2010–2019. Hefei (HF)/Total Carbon Column Observing Network (TCCON), Lulin (LLN)/World Data Centre for Greenhouse Gases (WDCGG) station observations were used to validate the results of void filling with an acceptable accuracy for spatiotemporal analysis (R = 0.96, R² = 0.92, RMSE = 2.44 ppm). Compared to the IDW (inverse distance weighting) and Kriging (ordinary Kriging) interpolation methods, this method has a higher validation accuracy. In addition, spatiotemporal distributions of CO₂, as well as the sensitivity of CO₂ concentration to the urban built-up areas and urban green space areas in China’s major southern cities during 2010–2019, are discussed. The approximate annual average concentrations have gradually increased from 388.56 to 414.72 ppm, with an annual growth rate of 6.73%, and the seasonal cycle presents a maximum in spring and a minimum in summer or autumn from 2010 to 2019. CO₂ concentrations have a strong positive correlation with the impervious area to city area ratio, while anomaly values of the impervious area to urban green area ratio occurred in individual cities. The experimental findings demonstrate the viability of the study hypothesis that combines remote sensing data with the WRF-Chem model to produce a local area dataset with high spatial resolution and an extracted urban unit from statistical data.

Natural imagery segmentation has been transferred to land cover classification in remote sensing imagery with excellent performance. However, two key issues have been overlooked in the transfer process: (1) some objects were easily overwhelmed by the complex backgrounds; (2) interclass information for indistinguishable classes was not fully utilized. The attention mechanism in the transformer is capable of modeling long-range dependencies on each sample for per-pixel context extraction. Notably, per-pixel context from the attention mechanism can aggregate category information. Therefore, we proposed a semantic segmentation method based on pixel representation augmentation. In our method, a simplified feature pyramid was designed to decode the hierarchical pixel features from the backbone, and then decode the category representations into learnable category object embedding queries by cross-attention in the transformer decoder. Finally, pixel representation is augmented by an additional cross-attention in the transformer encoder under the supervision of auxiliary segmentation heads. The results of extensive experiments on the aerial image dataset Potsdam and satellite image dataset Gaofen Image Dataset with 15 categories (GID-15) demonstrate that the cross-attention is effective, and our method achieved the mean intersection over union (mIoU) of 86.2% and 62.5% on the Potsdam test set and GID-15 validation set, respectively. Additionally, we achieved an inference speed of 76 frames per second (FPS) on the Potsdam test dataset, higher than all the state-of-the-art models we tested on the same device.

Urban-scale aerosol pollution monitoring is of critical importance for both climate regulation and public health. To overcome the limitations of conventional kilometer-scale satellite aerosol optical depth (AOD) products in resolving urban pollution heterogeneity, this study develops a physical-guided transfer deep neural network (PT-DNN) model based on high-resolution Landsat 8 data. The PT-DNN introduces a novel physics-guided training framework, in which radiative transfer simulations are integrated to physically constrain the AOD retrieval. Pre-training was conducted using multi-scenario radiative transfer simulations, with subsequent fine-tuning via ground-based AERONET measurements. The model architecture integrates convolutional neural network (CNN) with residual connection. Validation results over impervious surfaces indicate that the PT-DNN model outperforms conventional data-driven models, with the coefficient of determination (R²) increasing from 0.81 to 0.86 and root mean square error (RMSE) decreasing from 0.122 to 0.104. Moreover, the AOD distributions retrieved at a high spatial resolution of 30 m effectively reveal fine-scale pollution gradients within urban environments, especially in densely built-up and industrial areas.

Unlocking the secrets of habitable urban areas is crucial to improve the quality of life for urban dwellers. Accurate assessment of the ever-changing dynamics of a modern metropolis remains a challenging task. Previous studies have failed to reveal the dynamics of urban building spatial configuration at the micro-level. By analyzing high-resolution satellite imagery, this study has developed new direct and indirect metrics to better understand building density dynamics. We also applied these metrics to a study area located in Nanning City, China, revealing fascinating insights into the evolving spatial patterns of building density over the past 17 years. Our Q/R analysis uncovered areas with high habitability and suggested strategic improvements for sustainable building spatial configuration. This study is a valuable addition to the growing body of urban development research and provides scientific references for measurements of sustainable urban planning worldwide.