Solar radiation is the Earth’s primary source of energy and has an important role in the surface radiation balance, hydrological cycles, vegetation photosynthesis, and weather and climate extremes. The accurate prediction of solar radiation is therefore very important in both the solar industry and climate research. We constructed 12 machine learning models to predict and compare daily and monthly values of solar radiation and a stacking model using the best of these algorithms were developed to predict solar radiation. The results show that meteorological factors (such as sunshine duration, land surface temperature, and visibility) are crucial in the machine learning models. Trend analysis between extreme land surface temperatures and the amount of solar radiation showed the importance of solar radiation in compound extreme climate events. The gradient boosting regression tree (GBRT), extreme gradient lifting (XGBoost), Gaussian process regression (GPR), and random forest models performed better (poor) prediction capabilities of daily and monthly solar radiation. The stacking model, which included the GBRT, XGBoost, GPR, and random forest models, performed better than the single models in the prediction of daily solar radiation but showed no advantage over the XGBoost model in the prediction of the monthly solar radiation. We conclude that the stacking model and the XGBoost model are the best models to predict solar radiation.

Precipitable water vapor (PWV) plays a vital role in climate research, especially for Antarctica in which meteorological observations are insufficient due to the adverse climate and topography therein. Reanalysis data sets provide a great opportunity for Antarctic water vapor research. This study investigates the climatological PWV means, variability and trends over Antarctica from four reanalyses, including the fifth generation of European Centre for Medium-Range Weather Forecasts (ECMWF) Reanalysis (ERA5), the Second Modern-Era Retrospective analysis for Research and Applications (MERRA-2), Japanese 55-year Reanalysis (JRA-55) and National Centers for Environmental Prediction/Department of Energy (NCEP/DOE), in the period of 2001–2018 based on radiosonde and GNSS observations. PWV data from the ERA5, MERRA-2, JRA-55 and NCEP/DOE have been evaluated by radiosonde and GNSS observations, showing that ERA5 and MERRA-2 perform better than JRA-55 and NCEP/DOE with mean root mean square (RMS) errors below 1.2 mm. The climatological PWV mean distribution over Antarctica roughly shows a decreasing trend from west to east, with the highest content in summer and the lowest content in winter. The PWV variability is generally small over Antarctica, showing a seasonal dependence that is larger in the cold season and smaller in the warm season. PWV trends for all reanalyses at most Antarctic regions are insignificant and most reanalyses present overall drying trends from 2001 to 2018, except for ERA5 exhibiting a moistening trend. PWV trends also show seasonal and regional dependence. All reanalyses are generally consistent with radiosonde and GNSS observations in reproducing the PWV means (mean differences within 1.1 mm), variability (mean differences within 3%) and trends (mean differences within 6.4% decade⁻¹) over Antarctica, except for NCEP/DOE showing spurious variability and trends in East Antarctica. Results can help us further understand these four reanalysis PWV products and promote climate research in Antarctica.

To assess the health risk of PM_2.5, it is necessary to accurately estimate the actual exposure level of the population to PM_2.5. However, the spatial distribution of PM_2.5 may be inconsistent with that of the population, making it necessary for a high-spatial-resolution and refined assessment of the population exposure to air pollution. This study takes the Yangtze River Delta (YRD) Region as an example since it has a high-density population and a high pollution level. The brightness reflectance of night-time light, and MODIS-based (Moderate Resolution Imaging Spectroradiometer) vegetation index, elevation, and slope information are used as independent variables to construct a random-forest (RF) model for the estimation of the population spatial distribution, before any combination with the PM_2.5 data retrieved from MODIS. This enables assessment of the population exposure to PM_2.5 (i.e., intensity of population exposure to PM_2.5 and population-weighted PM_2.5 concentration) at a 3-km resolution, using the year 2013 as an example. Results show that the variance explained for the RF-model-estimated population density reaches over 80%, while the estimated errors in half of counties are < 20%, indicating the high accuracy of the estimated population. The spatial distribution of population exposure to PM_2.5 exhibits an obvious urban–suburban–rural difference consistent with the population distribution but inconsistent with the PM_2.5 concentration. High and low PM_2.5 concentrations are mainly distributed in the northern and southern YRD Region, respectively, with the mean proportions of the population exposed to PM_2.5 concentrations > 35μg/m³ close to 100% in all four seasons. A high-level population exposure to PM_2.5 is mainly found in Shanghai, most of the Jiangsu Province, the central Anhui Province, and some coastal cities of the Zhejiang Province. The highest risk of population exposure to PM_2.5 occurs in winter, followed by spring and autumn, and the lowest in summer, consistent with the PM_2.5 seasonal variation_. Seasonal-averaged population-weighted PM_2.5 concentrations are different from PM_2.5 concentrations in the region, which are closely related to the urban-exposed population density and pollution levels. This work provides a novel assessment of the proposed population-density exposure to PM_2.5 by using multi-satellite retrievals to determine the high-spatial-resolution risk of air pollution and detailed regional differences in the population exposure to PM_2.5.

Accurate forecasting of wind speed is essential for daily life and social production. While numerical weather prediction products are widely used, they rely on global data and mathematical models to solve atmospheric dynamics' equations, often failing to capture localized micrometeorological phenomena accurately. Factors such as surface conditions, land‐sea differences, and topography, particularly in coastal areas, further impact the accuracy of wind speed forecasts. This study presents a new method to enhance short‐term wind speed forecasting along China's coast by incorporating local and neighborhood spatiotemporal information. The approach integrates meteorological data from adjacent grid points as new inputs in the LightGBM, CatBoost, and XGBoost algorithms. Stacking ensemble technique is then employed to effectively combine with the aforementioned foundational models. Two sets of experiments are conducted: Experiments 1 exclude surrounding information, while Experiments 2 include it. Each set consists of five experiment groups: annual, spring, summer, autumn, and winter. Within each group, four models are tested: XGBoost, LightGBM, CatBoost, and stacking. Results show that incorporating surrounding site information improves forecast accuracy. In all five groups with added surrounding site information, the stacking model performs best. Compared to ECMWF forecast data, the stacking model improves wind speed forecast accuracy from 53.3%, 50.9%, 55.2%, 53.0%, and 54.0% to 77.2%, 73.1%, 76.7%, 78.2%, and 77.1%, respectively. These findings demonstrate the potential effectiveness of the proposed method for improving short‐term wind speed forecasts in China's coastal areas.

Surface visibility (SV), a key indicator of atmospheric transparency, is used widely in the fields of environmental monitoring, transportation, and aviation. However, the sparse distribution and limited number of SV monitoring sites make it difficult to fulfill the urgent need for spatiotemporally seamless fine-scale monitoring. Here, we developed the operational real-time SV retrieval (RT-SVR) framework for China that incorporates information from multiple data sources, including Chinese Land Data Assimilation System meteorological data, in situ observations, and other ancillary data. Seamless hourly SV data with 6.25-km spatial resolution are available in real time via the operational RT-SVR model, which was built using a two-layer stacked ensemble approach that combines multiple machine learning algorithms and a deep learning module. Sample-based cross-validation of the RT-SVR model on approximately 41.3 million data pairs revealed strong robustness and high accuracy, with a Pearson correlation coefficient (R) value of 0.95 and a root mean square error (RMSE) of 3.17 km. An additional hindcast-validation experiment, performed with continuous observations obtained over one year (approximately 20.8 million data pairs), demonstrated the powerful generalization capabilities of the RT-SVR model, albeit with slight degradation in performance (R = 0.85, RMSE = 5.28 km). The seamless hourly SV data with real-time update capability enable tracking of the generation, development, and dissipation of various low-SV events (e.g., fog, haze, and dust storms) in China. The developed framework might also prove useful for quantitative retrieval of aerosol-related parameters (e.g., PM_2.5, PM₁₀, and aerosol optical depth).

The verification and correction of CALIPSO aerosol products is key to understanding the atmospheric environment and climate change. However, CALIPSO often cannot detect the full profile of aerosol for the low instrument sensitivity near the surface. Thus, a correction scheme for the aerosol extinction coefficient (AECs) in the planetary boundary layer (PBL) is proposed to improve the quality of the CALIPSO-based aerosol optical depth (AOD) at 532 nm. This scheme assumed that the aerosol is vertically and uniformly distributed below the PBL, and that the AECs in the whole PBL are equal to those at the top of the PBL; then, the CALIPSO AOD was obtained by vertically integrating AECs throughout the whole atmosphere. Additionally, the CALIPSO AOD and corrected CALIPSO AOD were validated against seven ground-based sites across eastern China during 2007–2015. Our results show that the initial CALIPSO AOD obtained by cloud filtering was generally lower than that of the ground-based observations. After accounting for the AECs in the PBL, the adjustment method tended to improve the CALIPSO AOD data quality. The average R (slope) value from all sites was improved by 7% (46%). Further, the relative distance between the ground track of CALIPSO and the ground station exhibited an influence on the validation result of CALIPSO AOD. The retrieval precision of CALIPSO AOD worsened with the increase in water vapor in the atmosphere. Our findings indicate that our scheme significantly improves the accuracy of CALIPSO AOD, which will help to provide alternative AOD products in the presence of severe atmospheric pollution.