Global temperatures are continuing to rise as atmospheric carbon dioxide (CO₂) concentrations increase, and climate warming has become a major challenge to global sustainable development. The Cross-Track Infrared Sounder (CrIS) instrument is a Fourier transform spectrometer with 0.625 cm⁻¹ spectral resolution covering a 15 μm CO₂-absorbing band, providing a way of monitoring CO₂ with on a large scale twice a day. This paper proposes a method to predict the concentration of column-averaged CO₂ (XCO₂) from thermal infrared satellite data using ensemble learning to avoid the iterative computations of radiative transfer models, which are necessary for optimization estimation (OE). The training data set is constructed with CrIS satellite data, European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis v5 (ERA5) meteorological parameters, and ground-based observations. The training set was processed using two methods: correlation significance analysis (abbreviated as CSA) and principal component analysis (PCA). Extreme Gradient Boosters (XGBoost), Extreme Random Trees (ERT), and Gradient Boost Regression Tree (GBRT) are used for training and learning to develop the new retrieval model. The results showed that the R² of XCO₂ prediction built from the PCA dataset was bigger than that from the CSA dataset. These three learning models were verified by validation sets, and the ERT model showed the best agreement between model predictions and the truth (R² = 0.9006, RMSE = 0.7994 ppmv, MAE = 0.5804 ppmv). The ERT model was finally selected to estimate the concentrations of XCO₂. The deviation of XCO₂ predictions of 12 TCCON sites in 2019 was within ±1 ppm. The monthly averages of XCO₂ concentrations in close agreement with TCCON ground observations were grouped into four regions: Asia (R² = 0.9671, RMSE = 0.7072 ppmv), Europe (R² = 0.9703, RMSE = 0.8733 ppmv), North America (R² = 0.9800, RMSE = 0.6187 ppmv), and Oceania (R² = 0.9558, RMSE = 0.4614 ppmv).

Precipitable water vapor (PWV) is a critical component of the Earth’s atmosphere, playing a pivotal role in weather systems, climate dynamics, and hydrological cycles. Accurate estimation of PWV is essential for numerical weather prediction, climate modeling, and atmospheric correction in remote sensing. Ground-based observation stations can only provide PWV measurements at discrete points, whereas spaceborne infrared remote sensing enables spatially continuous coverage, but its retrieval algorithm is restricted to clear-sky conditions. This study proposes an innovative approach that uses ensemble learning models to integrate infrared and microwave satellite data and other geographic features to achieve all-weather PWV retrieval. The proposed product shows strong consistency with IGRA radiosonde data, with correlation coefficients (R) of 0.96 for the ascending orbit and 0.95 for the descending orbit, and corresponding RMSE values of 5.65 and 5.68, respectively. Spatiotemporal analysis revealed that the retrieved PWV product exhibits a clear latitudinal gradient and seasonal variability, consistent with physical expectations. Unlike MODIS PWV products, which suffer from cloud-induced data gaps, the proposed method provides seamless spatial coverage, particularly in regions with frequent cloud cover, such as southern China. Temporal consistency was further validated across four east Asian climate zones, with correlation coefficients exceeding 0.88 and low error metrics. This algorithm establishes a novel all-weather approach for atmospheric water vapor retrieval that does not rely on ground-based PWV measurements for model training, thereby offering a new solution for estimating water vapor in regions lacking ground observation stations.

Nitrogen dioxide (NO2) is a critical air pollutant with significant health and environmental implications, particularly in urban areas where high levels of emissions are prevalent. Accurate nowcasting of ground-level NO2 concentrations is essential for effective air quality management and timely public health interventions. Traditional methods often struggle with balancing the spatial accuracy of ensemble learning models and the temporal forecasting strengths of time-series models like long short-term memory (LSTM) networks. In this study, we propose an improved hybrid framework, GC-LSTM, to nowcast regional ground-level NO2 concentrations on an hourly scale based on satellite-derived NO2 vertical column densities (VCDs), meteorological data, and on-site observations. GC-LSTM integrates the spatial learning capabilities of grained cascade forest (gcForest) with the temporal prediction strengths of LSTM networks, leveraging the strengths of both spatial inference and time-series prediction. This study focuses on the Beijing-Tianjin–Hebei (BTH) region, one of China’s most polluted areas, as a case study. Our results indicate that the GC-LSTM framework performs a strong correlation between predicted and observed ground-level NO2 concentrations, with an $R^{2}$ of 0.746 and a mean absolute percentage error (MAPE) of 18.4% at a 1-h prediction interval. Even as the prediction intervals extended to 2 and 3 h, the GC-LSTM consistently outperforms the gcForest model across all evaluated metrics, with $R^{2}$ values higher by 0.097 and 0.117, and root mean square error (RMSE) values lower by 0.666 and $1.76~\mu \text {g/m}^{3}$ than those nowcasted by using the standalone gcForest model, respectively, highlighting its robustness and adaptability. Furthermore, the capacity of the GC-LSTM framework for continual learning and adaptation ensures its effectiveness in dynamic environments, making it a valuable tool for real-time air quality forecasting and environmental management.