The multiangle polarimetric (MAP) measurement provides abundant information about aerosol microphysical properties, but its physical retrieval methods of aerosols usually rely on time-consuming optimal iterative calculations. This study introduces a robust and efficient MAP aerosol retrieval over eastern China based on a data-driven deep learning (DL) method. By directly training the function relationship between Polarization and Directionality of the Earth’s Reflectances (POLDER) measurements and matched aerosol products in typical Aerosol Robotic Network (AERONET) sites with the deep belief network (DBN) methods, aerosol optical depth (AOD), fine mode AOD (FAOD), coarse mode AOD (CAOD), and single scattering albedo (SSA) can be retrieved reliably. Ground validation shows very high accuracy for POLDER-3 DBN AOD (${R} = 0.917$) and FAOD (${R} = 0.942$) compared with AERONET results. Despite a decrease in retrieval accuracy, DBN CAOD and spectral SSA exhibit very consistent variations with ground inversions. In particular, POLDER-3 DBN retrievals over eastern China perform better than generalized retrieval of aerosol and surface properties (GRASP) products with optimized method. Our results demonstrate that DBN can well model the complex functional relationships between MAP measurements and aerosol optical/microphysical parameters. With the striking advantage in computational efficiency and modeling ability, the DL methods, such as DBN, have an enormous potential in operational aerosol retrieval of the emerging MAP satellite instruments.

Diffuse solar radiation (DSR) is essential for carbon absorption in ecosystems and clean energy. Due to the scarcity of DSR observation stations, obtaining spatially continuous and high-accuracy all-sky DSR is a significant challenge. To achieve high-accuracy DSR estimation with limited observational data, this study developed a physics-guided deep learning (DL) algorithm. The algorithm effectively combines the advantages of the radiative transfer model (RTM) and DL and utilizes Himawari-8 top-of-atmosphere (TOA) reflectance and angular data as inputs to estimate DSR. Independent Baseline Surface Radiation Network (BSRN) and Wuhan University station observation validation results show that the algorithm has a high and robust performance in estimating instantaneous (hourly and daily) DSR, with a Pearson correlation coefficient (R) of 0.88 (0.91 and 0.91), a root-mean-square error (RMSE) of 61.84 (50.66 and 17.2) W/m2, and a mean bias error (MBE) of 0.16 (0.5 and −4.43) W/m2. In addition, compared to five existing DSR products (JiEA, CHSSDR, Deep Space Climate Observatory (DSCOVR)/Earth Polychromatic Imaging Camera (EPIC), ERA5, and CERES-SYN1deg), the algorithm shows the highest consistency (hourly$R=0.84$and daily$R=0.86$) and the smallest biases (hourly MBE =9.22 W/m2 and daily MBE =4.9 W/m2) at China Meteorological Administration (CMA) stations. Furthermore, comparisons with Himawari-8’s cloud cover product and related DSR products confirm the spatial rationality and continuity of the estimated DSR by this algorithm. This study demonstrates the advantages of the physics-guided neural network (PGNN) over traditional DL in enhancing the accuracy and transferability of DSR estimation, highlighting its potential for application in DSR retrievals from other similar satellites.

Direct aerosol radiative effect (DARE) is crucial for atmospheric radiation balance and climate. DARE is usually calculated using instantaneous aerosol optical depth (AOD) from satellite overpasses instead of daily mean AOD. Most satellites pass over twice a day at fixed local times, and these high-frequency data may not be an effective substitute for daily averages, thus affecting calculations of the radiative effects of aerosols. This study undertakes a comprehensive global-scale exploration into the disparities between utilizing satellite AOD products as substitutes for time-averaged values in radiative effect simulations. DARE calculated by instantaneous AOD from satellites were compared with the daily DARE from AERONET. The DARE calculated using MISR AOD showed the lowest RMSE of 10.3719 Wm-2. MODIS (Terra/Aqua) performed well in representing daily means, with Terra (RMSE = 11.1417 Wm-2) and Aqua (RMSE = 10.9722 Wm-2) showing similar results. In contrast, using VIIRS instantaneous AOD resulted in a higher RMSE of 14.3992 Wm-2. Since aerosols vary greatly during the day, this paper evaluates the representativeness error at different moments of time by comparing the DARE calculated using the daily mean AOD and instantaneous AOD from AERONET. The results show that the DARE from 8:00 a.m. to 2:00 p.m. has the smallest deviation (less than 0.6 Wm-2), which provides better temporal representativeness for the DARE calculation. The differences in the effect of daily changes in seasonal AOD on DARE calculations were small for all seasons from 10:00 a.m. to 13:00 p.m., and more pronounced in the fall.

The Multi-angle Imaging SpectroRadiometer (MISR) measurement with a large range of scattering angles provides valuable information about aerosol microphysical properties. The current MISR algorithm utilizes pre-defined aerosol mixtures in lookup tables (LUT) to infer aerosol types and microphysical parameters, which performs well globally but remains subject to considerable uncertainties in regional scales. To make efficient use of MISR measurement, we developed a physics-informed Deep Learning (PDL) method to retrieve aerosol optical/microphysical parameters over land in eastern China. By combining the physical constraint of radiative transfer simulation and modeling ability of DL methods, each aerosol parameter can be modeled with the whole used MISR measurements separately with high computational efficiency. PDL Aerosol Optical Depth (AOD) and fine AOD(FAOD) have high correlation coefficients (R>0.95) with Aerosol Robotic Network (AERONET) observations, with 89% and 81% values falling into expected error (EE) envelope of ± (0.05+20%AODAERONET) respectively. Despite only a slightly higher accuracy than recent MISR Version 23 products, PDL retrievals have solved the underestimation problem of AOD and FAOD at moderate-high values (>0.4). Besides better constraint of abnormal values in coarse AOD(CAOD), PDL algorithm significantly improves retrieval accuracy of MISR Single Scattering Albedo (SSA). With reliable and robust performance, PDL algorithm provides a flexible and efficient aerosol retrieval framework for emerging multi-angle polarimetric measurements.

Cloud mask (CM) data are indispensable for estimating downward surface shortwave radiation (DSSR). Most DSSR products are generated using CM data derived from passive satellite observations through the threshold methods. Some uncertainty exists in these CM data, yet the impact of CM quality on the DSSR estimates has received limited attention. To address this gap, this study proposed a method to quantify and mitigate errors in DSSR estimates resulting from CM data error, using the Himawari-8 (H8) products as a case study. First, machine learning (ML) models were constructed for CM, DSSR, and needed atmospheric parameters for DSSR estimation. Then, high-reliability CM estimates were utilized to update the H8 CM. The missing atmospheric parameters resulting from the CM updates were filled by the constructed models. Subsequently, DSSR data were estimated based on the updated CM. Results show that the updated CM effectively corrects misclassifications in the H8 CM, and the differences are more than 600 Wm−2 between DSSR estimates and H8 DSSR for some pixels. Cloud-aerosol Lidar and infrared pathfinder satellite observations (CALIPSO) CM and in situ DSSR were used as truth references for validation. The improved accuracy of the updated CM compared to H8 CM is mainly observed for snow/ice, with clear-sky and cloudy-sky hit rates (HRs) increasing by 0.1 and 0.3, respectively. Besides, when the H8 CM is consistent with the updated CM, the estimated DSSR exhibits a slightly lower root mean square error (RMSE) compared to the H8 DSSR, with a difference of no more than 3 Wm−2. However, in cases where the two CM data are inconsistent, the reduction in RMSE for the estimated DSSR compared to the H8 DSSR is more significant, exceeding 9 Wm−2.

Satellite-based aerosol optical property retrieval over land, especially size-related parameters, is challenging. This study proposed a novel two-stage machine learning (ML) algorithm for retrieving aerosol optical depth (AOD), Ångström exponent (AE), fine mode fraction (FMF), and fine mode AOD (FAOD)) over land using MODIS observed reflectance. The new ML algorithm consists of three steps: (1) first, all samples extracted from AERONET measurements were used to train the ML model, (2) then, to reduce the extreme estimation bias of the model, divided low-value and high-value samples were used to train low-value and high-value ML models, respectively, and (3) finally, the three ML models were integrated into the final retrieval based on the weight interpolation. Independent site network validation results show that the new ML algorithm has a Pearson correlation coefficient (R) of 0.894 (0.638, 0.661, 0.865) and root mean square error (RMSE) of 0.146 (0.258, 0.245, 0.153) for the AOD (AE, FMF, FAOD) retrieval, which significantly outperforms the validation metrics of MODIS operational products, with AOD (AE, FMF, FAOD) RMSE of 0.130-0.156 (0.536-0.569, 0.313, 0.191). The inter-comparison of aerosol products shows that the spatial patterns of AOD, AE, FMF, and FAOD of the new ML algorithm are in good agreement with those of the MODIS and POLDER products. These results illustrate that the new ML algorithm has good performance and transferability and indicate the ability of ML methods to be applied to multispectral instruments (such as MODIS) to retrieve multiple aerosol properties.

The multi-angle polarimetric (MAP) instruments have been a focus of recent satellite missions dedicated to enhanced detection of global aerosol microphysical properties. Considering that satellite observations can hardly infer all the unknowns of atmosphere and surface, it’s crucial to know how many and which aerosol parameters can be accurately retrieved from these different MAP measurements as well as their uncertainties. In this study, we present a comprehensive insight into the information content of POLDER-3 and 3MI observations for aerosol retrievals and estimate posterior errors of corresponding parameters based on Bayesian theory. The total degree of freedom for signal (DFS) of aerosol retrievals is around 6-8 from POLDER-3, and is raised by ~1.8-3.5 with 3MI. The retrieval accuracy of volume concentration and effective radius are high (<4%) in the fine-dominant case for both POLDER-3 and 3MI, but get much lower (~8% and ~15%) in coarse-dominant conditions. Furthermore, the advanced 3MI measurements can upgrade the retrieval uncertainties of POLDER-3 by ~50%. Though additional shortwave infrared bands of 3MI provide more information regarding coarse particles, the influence of aerosols on surface BRDF leads to a decrease of the total DFS. With a prior assumption that variations of refractive index depending on wavelength, satellite retrieval accuracy of the real (<0.03) and imaginary part (<0.003) reaches close levels with that of ground-based Sun photometers. Our results can provide a fundamental reference for MAP satellite retrieval of aerosol microphysical properties.

Mainstream satellite cloud masking algorithms are prone to mis-masking in haze-polluted areas, which may cause errors in aerosol radiative effect calculations and attribution of surface solar radiance changes; thereby, distinguishing between clouds and haze is critical to obtaining accurate land and atmospheric data products. Existing cloud and haze mask algorithms based on the threshold method may require us to spend a lot of manpower to perform multiple threshold tests; in addition, the obtained thresholds are only applicable to particular sensors, which limits the generality of the threshold-based cloud and haze mask algorithms. In this study, a new cloud and haze mask algorithm based on a combination of radiative transfer simulations and machine learning text simulation-based cloud and haze masking (SCHM) is proposed and applied to MODIS images. When we simulated the apparent reflectance of the first seven visible and text near-infrared channels of MODIS, the CALIOP and AERONET data verification results showed that the SCHM algorithm achieved 85.16% and 90.08% hit rates for cloud and haze recognition, respectively. When we added three thermal infrared channels (20, 31, and 35 bands) for simulation, the cloud and haze hit rates were improved to approximately 85.72% and 90.62%, respectively. This indicates that the SCHM algorithm can improve the accuracy of detection results by improving the radiative transfer simulation parameters. Compared with existing threshold-based methods, the SCHM algorithm has the advantages of simple logic, convenient modification, and flexible configuration.

Accurate estimates of downward surface shortwave radiation (DSSR) are critical for hydrological, biogeochemical, and ecological studies and remote sensing-based estimation of DSSR is an important way to derive DSSR at different spatio-temporal ranges. However, current estimation algorithms usually somewhat rely on atmospheric parameters or in-situ measurements, further blocking the application of these methods. Inspired by the emerging DSSR reanalysis data from the model simulation, this study proposed an integrated method by initializing the estimation model with ERA5 reanalysis data and further refining the estimation through iterative training. The random forest regression method was applied in the estimation model to build the connection between DSSR with the MODIS top-of-atmosphere reflectance, cloud flag, geometry information, elevation, latitude, and coefficient of Sun-Earth distance as input features. To separately consider the impact from cloud cover, the estimation model was established for clear-sky and cloudy-sky conditions, respectively. The proposed method was applied to estimate instantaneous DSSR of MODIS daytime overpasses in the Southwest part of China in 2020. Comparison between the estimates of the initialized model and the finalized model shows that the iterative process improves the DSSR estimates on both spatial distribution and accuracy. Validated by the measurements from nine sites in the study area, the DSSR estimates of the finalized model show a 0.02 higher correlation coefficient (CC) and 7.35 W m-2lower root mean squared error (RMSE) than that of the initialized model. To better evaluate the performance of the proposed method, three popular DSSR products including ERA5, MCD18A1, and Himawari-8 were introduced to make an inter-comparison with the estimation of this study. The validation results showed that the all-sky DSSR estimated in this study had the best accuracy, with a CC of 0.90, a mean bias error of 37.80 W m-2, a RMSE of 125.30 W m-2, and a relative root mean squared error of 42.73%. Obvious improvements can be observed under cloudy-sky and clear-sky conditions, respectively. Because of the simplicity and reliable performance of the proposed method, it shows good potential for DSSR estimation.

The Advanced Geostationary Radiation Imager (AGRI) onboard the Fengyun 4A (FY-4A) satellite has high spatiotemporal resolution and provides useful spectral information that can be used to monitor aerosols and air pollution. The objective of this study is to propose the Land General Aerosol (LaGA) algorithm for retrieving aerosol information using AGRI data in the Asia region. First, the sensitivity analysis indicated that the AGRI blue band is more suitable for aerosol retrieval, and its red band is sensitive under high aerosol loading. Then, a real-time surface reflectance (SR) database was established using the atmosphere-corrected technique based on the background AOD library and regional aerosol model parameters. By comparing the AGRI observed reflectance with that calculated using a lookup table, the AGRI aerosol optical depth (AOD) with a 1-h resolution was obtained. The validation results indicated that the AGRI AOD, both at all moments (data volume: 12,102) and the daily mean (data volume: 1,766), exhibit a good agreement with AERONET AOD (R > 0.830). Its performance was comparable to that of the MOdIs dark target (DT) AOD (expected error (EE), ± (0.05 + 20%τAERONET): AGRI = 0.673 vs. DT = 0.666) and Himawari-8 (H8) AOD (EE: AGRI = 0.698 vs. H8 = 0.658). The pixel-by-pixel comparison demonstrated that the R between the AGRI and MODIS AODs was >0.6, and the mean bias between them was within ±0.05 in most of the study area. These results suggest the robustness of the proposed algorithm, and it has great potential for application in the follow-up Fengyun 4 series satellites.

High temporal resolution aerosol optical depth (AOD) observations derived from new-generation geostationary (GEO) satellite possess unique advantages in analyzing aerosol fast variation processes and thereby providing more accurate assessments on their climate effects and health risks. Unfortunately, the expected advantages and values are dramatically limited by relatively large proportion of data missing in the GEO AOD products due to cloud obscuration and intrinsic retrieval algorithm. Although several data recovery algorithms have been proposed in recent years to improve the spatial coverage for GEO AOD products, yet most of them aims at filling up the data blanks rather than reconstructing the temporally continuous variation of aerosol. Accordingly, in this study, a novel framework of nested spatiotemporal fusion blending GEO with sun-synchronous orbit (SSO) satellite observations based on Bayesian maximum entropy (BME) theorem is developed for GEO Advanced Himawari-8 Imager (AHI) AOD recovery with the sufficient excavation of complementary information from GEO and SSO satellite observations, where the minute-stage and hour-stage BME fusion are jointly employed to reconcile temporal inconsistency and data discrepancies between GEO and SSO observations. The results demonstrate that the AOD spatial coverage is dramatically increased by 240.9% (from 20.5% to 70%) with ensured accuracy after Nested-BME fusion. Additionally, two case analyses, during the development and dispersion processes of haze respectively, both demonstrate that the proposed Nested-BME fusion framework could reconstruct the reliable aerosol diurnal variation trends on the basis of recovering missing data for Himawari-8 AHI AOD datasets, while the AHI official level-2 and level-3 AOD products fail to capture these key trends. Furthermore, the developed Nested-BME AOD fusion framework is also applicable for other geostationary satellites over other regions, which could substantially enhance the availability and value of high temporal resolution AOD products for better scientific applications.

High-frequency aerosol observation from new-generation geostationary meteorological satellite is capable to capture and monitor the spatiotemporal dynamic variation of aerosols, which is of vital significance to environmental research and climate studies. Due to the diversity and complexity of land cover, it is a challenge to retrieve aerosol properties with high accuracy over land especially over heterogeneous land surfaces. In this study, a Geometry-Discrete Minimum Reflectance Aerosol Retrieval Algorithm (GeoMRA) has been proposed to retrieve 10-min high temporal resolution aerosol optical depth (AOD) datasets for geostationary Himawari-8 AHI sensor, aiming at providing universal bidirectional reflectance distribution function (BRDF) descriptions for different land surfaces with different heterogeneous extent. The AOD retrievals from GeoMRA demonstrate good consistency against the ground-based AERONET measurements in the East Asia from 2015 to 2020, with a correlation coefficient (R) of 0.883 and approximately 65.6% of matchups falling within the expected error envelope of ±(0.05 + 15%). Intercomparison between the GeoMRA retrieved AOD and other operational AOD products shows that the GeoMRA AOD retrievals, which generally possess similar spatial distribution and accuracy as MODIS AOD products, have better performances than the Japan Aerospace Exploration Agency (JAXA) AOD products by providing more accurate AOD retrievals with higher spatial coverage. Moreover, the AOD bias analyses further demonstrate the robustness of GeoMRA algorithm, and an extreme haze event shows that the continuous GeoMRA AOD images illustrate smoother temporal variations than JAXA AOD products, demonstrating its efficacy and reliability in capturing the process of haze transport and monitoring the continuous spatiotemporal variation of aerosol. The above results suggest the considerable accuracy of GeoMRA algorithm for scientific application requirement, and demonstrate the robustness of proposed BRDF scheme in describing heterogeneous surfaces with diverse reflectance distribution.

Satellite-derived land surface temperatures (LSTs) are a critical parameter in various fields. Unfortunately, there are numerous gaps in LST products due to cloud contamination and orbital gaps. In previous studies, various gapfilling methods have been developed. However, most of those methods use only spatiotemporal information to fill gaps. In this study, a gapfilling method called the enhanced hybrid (EH) method that integrates spatiotemporal information and information from other similar LST products was proposed. The accuracy of the EH method was compared with the accuracies of three other gapfilling methods that only use spatiotemporal information: Remotely Sensed DAily land Surface Temperature reconstruction (RSDAST), interpolation of the mean anomalies (IMAs), and Gapfill. It was found that the correlations between the four LST products were strong, indicating that using information from other products may improve the accuracy of gapfilling. On average, the mean absolute errors (MAEs) of the data filled using the EH method were 23.7%–52.7% lower than those of RSDAST, 35.4%–38.7% lower than those of IMA, and 38.5%–46.9% lower than those of the Gapfill method. The usage of information from other similar LST products was the main reason for the high accuracy observed for the EH method. In addition, the LST images filled using the RSDAST and IMA methods had some outliers, while there were fewer obvious outliers in the LST images filled with the EH method. It was concluded that the EH method is a robust gapfilling method with a high accuracy.

This paper proposes a new method for automatic co-registration of digital elevation models (DEMs) based on centroids of subwatersheds. Subwatersheds are stable physical features, making their centroids more reliable and accurate as control points (CPs) than the other features. In the present method, subwatersheds are delineated from DEMs using hydrological analysis procedures. Modified invariant moments are employed to measure the similarity of subwatersheds for determining the correspondences between the reference and input. Centroids of matched subwatersheds are then derived as CP candidates, where the root-mean-square error is applied to eliminate mismatches using a global consistency check method. The established CP pairs are used to estimate parameters of a 3-D conformal transformation model, which is employed to rectify the input DEM. The accuracy of CP detection was assessed using two DEM subsets of different terrains. The results of six tests showed that the maximum shift and rotation errors were about 2 m (1/45 pixel) and 0.006°. The developed approach was used to co-register the Advanced Spaceborne Thermal Emission and Reflection Radiometer Global DEM to the Shuttle Radar Topography Mission DEM at three locations. The results revealed that both the mean absolute error and the standard deviation of the elevation differences were reduced for all the tests after co-registration, showing the good performance of the proposed method. Comparisons have also been made against previous works, which suggested that our results were consistent with the previous studies.

In this article, we introduce the genetic algorithm into the inversion of aerosol size distribution. We are often faced with limited or insufficient observations in remote sensing and the observations are contaminated. The particle spectrum extinction equation is an ill-posed integral equation of the extinction inversion method[1]. To overcome the ill-posed nature, we use a double logarithmic normal distribution function to express the aerosol size distribution. To obtain the optimal solution, we introduce the genetic algorithm to gain the minimum sum of squared errors. Our method can improve accuracy and reduce the computational difficulty. The assumption of parameters in the bimodal distribution function is important to the inversion results. The aerosol size distribution obtained from the GRIMM 180 PM monitor and the TSI Scanning mobility particle sizers is compared with that computed via the method proposed by Dubovik and King(2000)[2]. Obvious difference has been discovered between aerosol size distribution on the ground and in the total atmospheric column. As a result, it is necessary to develop multi-wavelength and multi-function lidar to get observe the three-dimensional distribution characteristics of aerosol.

This study investigates the multifractal scaling behavior of global land surface air temperature. Multifractal detrended fluctuation analysis is employed to examine the long-range correlation and multiscaling behavior of the CRUTEM4 near-surface air temperature anomaly over land from 1901 to 2013. The scaling exponents of the temperature series averaged over the global earth, Northern Hemisphere, and Southern Hemisphere are 0.85, 0.80, and 0.94, respectively. The results suggest that the global temperature anomaly is long-range correlated or long-term persistent and that the temperature of the Northern Hemisphere is higher than that of the Southern Hemisphere. Compared with the inner continental land, the island and coastal areas show stronger long-range correlation or long-term persistence. The dependence of h(q) on q indicates that the global land near-surface temperature is multifractal. In addition, the degree of multifractality is heterogeneously distributed in global lands.