Satellite-retrieved Aerosol Optical Depth Research Articles

Regional PM2.5 Estimation for Southern Ontario through Geographically Weighted Regression

In this study, a geographically weighted regression (GWR) approach was adopted to forecast regional concentration of particulate matter 2.5 (PM2.5) for the southern Ontario based on both in situ meteorological measurement and Satellite retrievals of aerosol optical depth (AOD). The correlation between monitored concentration of PM2.5 and Satellite-retrieved AOD would be quantified. The ground-level PM2.5 for South Ontario area was then predicted using GWR with AOD and meteorological variables considered as inputs. The results indicated that performance of GWR was slightly better than the ordinary least squares (OLS) model, indicating spatial variations between independent and dependent variables. Consequently, the GWR model can help us to predict the PM2.5 concentration in terms of time or region with satellite data, and also help improve satellite data inversion.

Assessment of Satellite AOD during the 2020 Wildfire Season in the Western U.S.

Satellite remote sensing of aerosol optical depth (AOD) is essential for detection, characterization, and forecasting of wildfire smoke. In this work, we evaluate the AOD (550 nm) retrievals during the extreme wildfire events over the western U.S. in September 2020. Three products are analyzed, including the Moderate-resolution Imaging Spectroradiometers (MODIS) Multi-Angle Implementation of Atmospheric Correction (MAIAC) product collections C6.0 and C6.1, and the NOAA-20 Visible Infrared Imaging Radiometer (VIIRS) AOD from the NOAA Enterprise Processing System (EPS) algorithm. Compared with the Aerosol Robotic Network (AERONET) data, all three products show strong linear correlations with MAIAC C6.1 and VIIRS presenting overall low bias (<0.06). The accuracy of MAIAC C6.1 is found to be substantially improved with respect to MAIAC C6.0 that drastically underestimated AOD over thick smoke, which validates the effectiveness of updates made in MAIAC C6.1 in terms of an improved representation of smoke aerosol optical properties. VIIRS AOD exhibits comparable uncertainty with MAIAC C6.1 with a slight tendency of increased positive bias over the AERONET AOD range of 0.5–3.0. Averaging coincident retrievals from MAIAC C6.1 and VIIRS provides a lower root mean square error and higher correlation than for the individual products, motivating the benefit of blending these datasets. MAIAC C6.1 and VIIRS are further compared to provide insights on their retrieval strategy. When gridded at 0.1° resolution, MAIAC C6.1 and VIIRS provide similar monthly AOD distribution patterns and the latter exhibits a slightly higher domain average. On daily scale, over thick plumes near fire sources, MAIAC C6.1 reports more valid retrievals where VIIRS tends to have retrievals designated as low or medium quality, which tends to be due to internal quality checks. Over transported smoke near scattered clouds, VIIRS provides better retrieval coverage than MAIAC C6.1 owing to its higher spatial resolution, pixel-level processing, and less strict cloud masking. These results can be used as a guide for applications of satellite AOD retrievals during wildfire events and provide insights on future improvement of retrieval algorithms under heavy smoke conditions.

Spatiotemporal high-resolution imputation modeling of aerosol optical depth for investigating its full-coverage variation in China from 2003 to 2020

Investigating spatiotemporal variations of atmospheric aerosols is important for climate change and environmental research. Although satellite aerosol optical depth (AOD) retrieved by the MAIAC (Multiangle Implementation of Atmospheric Correct) algorithm provides a unique opportunity to represent global aerosol loading with high spatiotemporal resolution, accurate assessment of long-term aerosol loading countrywide is still challenging due to its non-random missingness. This study aimed to develop an adaptive spatiotemporal high-resolution imputation modeling framework for AOD that incorporates random forest models and multisource data (the simulated AOD, meteorological, and surface condition data) to support full-coverage long- and short-term aerosol studies in China. Aided by the time-stratified approach, the imputation model was constructed for each day, and the MAIAC AOD was used as the target variable. The proposed approach could effectively capture the massive spatiotemporal variability in a large amount of data and deliver full-coverage AODs with high accuracies at a daily timescale (i.e., overall validation R2 against ground-level AOD measurements of 0.77). We then employed the proposed approach to impute the daily MAIAC retrieved AOD towards complete coverage for China for 2003–2020. Due to the complete coverage, the spatial pattern of monthly/seasonal/yearly mean AOD imputations has better representativeness than that of original MAIAC retrievals. Comparison analysis shows that the monthly/seasonal/yearly aerosol loading over most of China tends to be underestimated by temporal aggregates of original satellite-retrieved AODs. Such underestimation is particularly severe in summer and over the North China Plain (the amount of underestimation >0.2). Consequently, our full-coverage AOD imputations can advance scientific research and environmental management by supporting national and local complete pictures of both short-term episodes and long-term trends in atmospheric aerosols.

Spatiotemporal continuous estimates of daily 1 km PM2.5 from 2000 to present under the Tracking Air Pollution in China (TAP) framework

Abstract. High spatial resolution PM2.5 data covering a long time period are urgently needed to support population exposure assessment and refined air quality management. In this study, we provided complete-coverage PM2.5 predictions with a 1 km spatial resolution from 2000 to the present under the Tracking Air Pollution in China (TAP, http://tapdata.org.cn/, last access: 3 October 2022) framework. To support high spatial resolution modeling, we collected PM2.5 measurements from both national and local monitoring stations. To correctly reflect the temporal variations in land cover characteristics that affected the local variations in PM2.5, we constructed continuous annual geoinformation datasets, including the road maps and ensemble gridded population maps, in China from 2000 to 2021. We also examined various model structures and predictor combinations to balance the computational cost and model performance. The final model fused 10 km TAP PM2.5 predictions from our previous work, 1 km satellite aerosol optical depth retrievals, and land use parameters with a random forest model. Our annual model had an out-of-bag R2 ranging between 0.80 and 0.84, and our hindcast model had a by-year cross-validation R2 of 0.76. This open-access, 1 km resolution PM2.5 data product, with complete coverage, successfully revealed the local-scale spatial variations in PM2.5 and could benefit environmental studies and policymaking.

An integrated approach of Belief Rule Base and Convolutional Neural Network to monitor air quality in Shanghai

Accurate monitoring of air quality can reduce its adverse impact on earth. Ground-level sensors can provide fine particulate matter (PM2.5) concentrations and ground images. But, such sensors have limited spatial coverage and require deployment cost. PM2.5 can be estimated from satellite-retrieved Aerosol Optical Depth (AOD) too. However, AOD is subject to uncertainties associated with its retrieval algorithms and constrain the spatial resolution of estimated PM2.5. AOD is not retrievable under cloudy weather as well. In contrast, satellite images provide continuous spatial coverage with no separate deployment cost. Accuracy of monitoring from such satellite images is hindered due to uncertainties of sensor data of relevant enviromental parameters, such as, relative humidity, temperature, wind speed and wind direction. Belief Rule Based Expert System (BRBES) is an efficient algorithm to address these uncertainties. Convolutional Neural Network (CNN) is suitable for image analytics. Hence, we propose a novel model by integrating CNN with BRBES to monitor air quality from satellite images with improved accuracy. We customized CNN and optimized BRBES to increase monitoring accuracy further. An obscure image has been differentiated between polluted air and cloud in our model. Valid environmental data (temperature, wind speed and wind direction) have been adopted to further strengthen the monitoring performance of our proposed model. Three-year observation data (satellite images and environmental parameters) from 2014 to 2016 of Shanghai have been employed to analyze and design our proposed model. The results conclude that the accuracy of our model to monitor PM2.5 of Shanghai is higher than only CNN and other conventional Machine Learning methods. Real-time validation of our model on near real-time satellite images of April-2021 of Shanghai shows average difference between our calculated PM2.5 concentrations and the actual one within ± 5.51.

Joint features random forest (JFRF) model for mapping hourly surface PM2.5 over China

Ambient PM2.5 exerts strong regional pattern for its ability for long-range transport, implying that including the features of surrounding stations may improve the accuracy of machine-learning based model to estimate surface PM2.5 from the satellite-retrieved aerosol optical depth (AOD). However, most of current models either just use single point features, or simply average the observed surface PM2.5 from adjacent stations based on a fixed spatial proportional relationship. The question that how to properly take advantage of the features of surrounding stations for retrieving PM2.5 is still not well addressed. Here we propose an integrated algorithm called joint features random forest (JFRF) model which includes complex feature differences with surrounding stations and the observation of stations to learn the dynamic relations with the PM2.5 of target pixel, rather than the weighted average feature (WAF) only by surface PM2.5 as traditional models (with WAF) used. Results of cross validation suggest better performance of JFRF (R2 = 0.61–0.8; RMSE = 15.97–20.91 μg/m3) than single point feature model (ΔR2 = 0.09–0.3). JFRF also exhibits better performance than traditional models (with WAF) (ΔR2 = 0.05–0.11), particularly in regions with large AOD gradient (accounts for 33% of the total test set), which is of great significance for accurately representing the spatial heterogeneity of PM2.5 (e.g., pollution edging and hot spots areas). And the exclusion of AOD from the features significantly reduced the model performance (ΔR2 = −0.07 ∼ −0.1). Therefore, our study demonstrates the important of the feature differences of surrounding stations and satellite-retrieved AOD in representing the regional pattern of PM2.5 and further helping the machine-learning based model to improve the accuracy in estimating surface PM2.5.

Evaluation and comparison of MODIS and VIIRS aerosol optical depth (AOD) products over regions in the Eastern Mediterranean and the Black Sea

Aerosol loading has direct and indirect impacts on radiative forcing and climate; additionally, exposure to high aerosol concentrations has adverse effects on human health. Several inversion algorithms are used to retrieve aerosol properties, primarily aerosol optical depth (AOD). Satellite aerosol products are available from several moderate spatial resolution sensors and provide information on the spatial and temporal dynamics of aerosol concentration. This study provides an evaluation and inter-comparison of multiple aerosol products from Terra and Aqua Moderate Resolution and Imaging Spectroradiometer (MODIS) and Suomi National Polar-Orbiting Partnership (Suomi NPP) Visible-Infrared Imager Radiometer Suite (VIIRS) between 2014 and 2018 in the Eastern Mediterranean and the Black Sea. This evaluation focuses on four MODIS aerosol products including Dark Target (DT) AOD retrievals at 10-km and 3-km spatial resolutions, Deep Blue (DB) AOD retrievals at10-km spatial resolution, Multi-Angle Implementation of Atmospheric Correction (MAIAC) AOD retrievals at 1-km spatial resolution, and VIIRS DB AOD retrievals at 6-km spatial resolution. Satellite-based AOD retrievals are validated against in-situ AOD measurements from three AERONET ground stations located in urban/land, rural/coastal, and ocean surfaces.The results show that Pearson's correlation coefficient (r) between the satellite AOD retrievals and AERONET AOD measurements range between 0.45 and 0.93 whilst the Root-Mean-Squared Errors (RMSE) varies from 0.047 to 0.159 which highlights the reliability of satellite AOD retrievals' in the region. The best performing products over the urban/land surface are the MODIS MAIAC and VIIRS DB aerosol products with the RMSE of 0.048–0.061 and a higher percentage (81%–89%) of retrievals falling within expected error for land. Over the ocean, the MODIS DT aerosol products have the lowest errors (RMSE 0.047–0.061) and the highest percentage (70%–80%) of retrievals falling within the expected error. The performances of all algorithms over the rural/coastal region are poorer in terms of over- and under-estimations of AOD with a small percentage of retrievals (48%–87%) within expected error. Among the satellite aerosol products, the VIIRS aerosol product performs better over the coastal area in terms of high correlation coefficient (r = 0.88), low error (RMSE = 0.059), and the large percentage of retrievals within expected error (88%). The poor performance found over the more complex coastal regions suggests further improvements are needed in land-water-sediment masks, surface reflectance estimation, and aerosol model assumptions over mixed surface types.The aerosol products show the highest agreement (r ∼ 0.77–0.98) over the ocean surface whilst the coastal site aerosol products verified the greatest dissimilarity in aerosol loading variations (r ∼ 0.50–0.91). Lastly, the good correlation (r > 0.73) between the VIIRS and MODIS AOD products illustrates the potential of VIIRS to provide aerosol data continuity with MODIS.

Temporal and Spatial Autocorrelation as Determinants of Regional AOD-PM2.5 Model Performance in the Middle East

Exposure to fine particulate matter (PM2.5) air pollution has been shown in numerous studies to be associated with detrimental health effects. However, the ability to conduct epidemiological assessments can be limited due to challenges in generating reliable PM2.5 estimates, particularly in parts of the world such as the Middle East where measurements are scarce and extreme meteorological events such as sandstorms are frequent. In order to supplement exposure modeling efforts under such conditions, satellite-retrieved aerosol optical depth (AOD) has proven to be useful due to its global coverage. By using AODs from the Multiangle Implementation of Atmospheric Correction (MAIAC) of the MODerate Resolution Imaging Spectroradiometer (MODIS) and the Multiangle Imaging Spectroradiometer (MISR) combined with meteorological and assimilated aerosol information from the Modern-Era Retrospective analysis for Research and Applications, Version 2 (MERRA-2), we constructed machine learning models to predict PM2.5 in the area surrounding the Persian Gulf, including Kuwait, Bahrain, and the United Arab Emirates (U.A.E). Our models showed regional differences in predictive performance, with better results in the U.A.E. (median test R2 = 0.66) than Kuwait (median test R2 = 0.51). Variable importance also differed by region, where satellite-retrieved AOD variables were more important for predicting PM2.5 in Kuwait than in the U.A.E. Divergent trends in the temporal and spatial autocorrelations of PM2.5 and AOD in the two regions offered possible explanations for differences in predictive performance and variable importance. In a test of model transferability, we found that models trained in one region and applied to another did not predict PM2.5 well, even if the transferred model had better performance. Overall the results of our study suggest that models developed over large geographic areas could generate PM2.5 estimates with greater uncertainty than could be obtained by taking a regional modeling approach. Furthermore, development of methods to better incorporate spatial and temporal autocorrelations in machine learning models warrants further examination.

Estimating ground-level PM2.5 concentrations using two-stage model in Beijing-Tianjin-Hebei, China

In this study, a two-stage model (linear mixed-effect [LME] model + geographically weighted regression model) that could account for the spatiotemporal variability of the relationship of particulate matter less than 2.5 μm in diameter (PM2.5) and aerosol optical depth (AOD) was developed to estimate ground-level PM2.5 concentrations based on satellite-retrieved AOD products. The Beijing–Tianjin–Hebei (BTH) region was set as an example. In situ PM2.5 measurements from 80 monitoring stations and AODs from the Moderate Resolution Imaging Spectroradiometer in 2016 were used to evaluate and validate the model performance. The spatiotemporal coverage of AOD products was expanded, and the bandwidth parameters of the model were appropriately selected. Results reveal that the two-stage model effectively improves PM2.5 estimation precision. The overall coefficient of determination (R2) of the model after being cross-validated against ground observations is 0.890. The root mean squared prediction error, the mean absolute prediction error, and the relative prediction error are 19.752 μg/m3, 11.343 μg/m3, and 28.477%, respectively. The performance of the two-stage model is much better than that of the LME model. Our estimates indicate that the overall gridded annual mean PM2.5 estimation is 59.240 μg/m3 (range of 24.531–94.581 μg/m3). High values are mainly found in southern plains, while low values are observed in northwestern plateau mountain areas. The most heavily polluted season is winter, followed by fall and spring, whereas the least polluted season is summer. Therefore, the proposed two-stage model may serve as a reference for monitoring the BTH region with heavy PM2.5 pollution.

High resolution aerosol optical depth retrieval over urban areas from Landsat-8 OLI images

The satellite-retrieved aerosol optical depth (AOD) provides unique estimation of aerosol loading across a continuous space. However, current AOD products with a relatively coarse resolution (≥1 km) can hardly capture the details in urban areas with large spatial AOD gradients. To address this issue, here we developed a novel retrieval algorithm for retrieving AOD with extra fine spatial resolution (30 m) from Landsat-8 satellite OLI images. In the algorithm, the three land surface reflectance (LSR) estimation schemes and four aerosol types are adopted to improve the retrieval accuracy. The algorithm is applied on Beijing and Wuhan city in China during 2014–2019. Results suggest that the retrieved AOD with the new algorithm exhibits good agreement with the ground-based measured AOD (R2 = 0.920), and 81.63% of the AODs fall within the expected error line with a root-mean-square error of 0.112. Moreover, the high-resolution AOD products also successfully identified polluted sources and discrepancy of aerosol loading over different land cover types in two megacities in China, implying its great potential on detecting fine aerosol emission sources over the complex urban area. Such improvement of the new algorithm for retrieving 30 m AOD developed in this study demonstrates its substantial potentials in supporting further studies of air pollution management and human exposure at extra-fine spatial scale.

Variability, predictability, and uncertainty in global aerosols inferred from gap-filled satellite observations and an econometric modeling approach

Time series analyses and stochastic modeling assessments of aerosols are critical for climate change and human health studies. However, the precise characterization of the aerosol optical depth (AOD), its variability, trends, and predictability, and its associated uncertainty at the global scale is largely unexplored. In addition, gaps in satellite-retrieved AODs across space and time remain an obstacle to accurately revealing aerosol properties. This study uses MODIS Collection 6 AOD retrievals for a time series analysis and modeling of global aerosols from 2003 to 2015. Random forest (RF) regression is first applied to replace the missing data in the satellite AOD retrievals. AOD variations and trends are then investigated and future values are predicted using a Mann–Kendall analysis approach and an autoregressive integrated moving average (ARIMA) model, respectively. The results indicate that the developed RF model enhances the AOD data coverage significantly, with the root mean square error and mean absolute error statistical metrics well below 0.13 and 0.08, respectively. High AOD loadings are found over East, South, and Southwest Asia, West and Central Africa, and northern South America. Prominent discrepancies are shown in aerosol variations and trends likely because of dust emissions, biomass burning, fossil fuel combustion, and socioeconomic practices, which has significant implications for climate systems and mitigation policy-making. The ARIMA model delineates AOD features with clear annual and seasonal variations and with high accuracy over most regions. The performance of the model is jointly impacted by the data quality and data values. Overall, our study suggests the feasibility and applicability of the RF model in reconstructing area-scale satellite missing AOD retrievals, as well as the ability of the stochastic ARIMA model to accurately depict and forecast AOD profiles. The globally simulated and predicted aerosols will improve evaluations of aerosol effects for climate and epidemiological studies.

Satellite-Based Mapping of High-Resolution Ground-Level PM2.5 with VIIRS IP AOD in China through Spatially Neural Network Weighted Regression

Satellite-retrieved aerosol optical depth (AOD) data are extensively integrated with ground-level measurements to achieve spatially continuous fine particulate matters (PM2.5). Current satellite-based methods however face challenges in obtaining highly accurate and reasonable PM2.5 distributions due to the inability to handle both spatial non-stationarity and complex non-linearity in the PM2.5–AOD relationship. High-resolution (<1 km) PM2.5 products over the whole of China for fine exposure assessment and health research are also lacking. This study aimed to predict 750 m resolution ground-level PM2.5 in China with the high-resolution Visible Infrared Imaging Radiometer Suite (VIIRS) intermediate product (IP) AOD data using a newly developed geographically neural network weighted regression (GNNWR) model. The performance evaluations demonstrated that GNNWR achieved higher prediction accuracy than the widely used methods with cross-validation and predictive R2 of 0.86 and 0.85. Satellite-derived monthly 750 m resolution PM2.5 data in China were generated with robust prediction accuracy and almost complete coverage. The PM2.5 pollution was found to be greatly improved in 2018 in China with annual mean concentration of 31.07 ± 17.52 µg/m3. Nonetheless, fine-scale PM2.5 exposures at multiple administrative levels suggested that PM2.5 pollution in most urban areas needed further control, especially in southern Hebei Province. This work is the first to evaluate the potential of VIIRS IP AOD in modeling high-resolution PM2.5 over large-scale. The newly satellite-derived PM2.5 data with high spatial resolution and high prediction accuracy at the national scale are valuable to advance environmental and health researches in China.

Merging ground-based sunshine duration observations with satellite cloud and aerosol retrievals to produce high-resolution long-term surface solar radiation over China

Abstract. Although great progress has been made in estimating surface solar radiation (Rs) from meteorological observations, satellite retrieval, and reanalysis, getting best-estimated long-term variations in Rs are sorely needed for climate studies. It has been shown that Rs data derived from sunshine duration (SunDu) can provide reliable long-term variability, but such data are available at sparsely distributed weather stations. Here, we merge SunDu-derived Rs with satellite-derived cloud fraction and aerosol optical depth (AOD) to generate high-spatial-resolution (0.1∘) Rs over China from 2000 to 2017. The geographically weighted regression (GWR) and ordinary least-squares regression (OLS) merging methods are compared, and GWR is found to perform better. Based on the SunDu-derived Rs from 97 meteorological observation stations, which are co-located with those that direct Rs measurement sites, the GWR incorporated with satellite cloud fraction and AOD data produces monthly Rs with R2=0.97 and standard deviation =11.14 W m−2, while GWR driven by only cloud fraction produces similar results with R2=0.97 and standard deviation =11.41 W m−2. This similarity is because SunDu-derived Rs has included the impact of aerosols. This finding can help to build long-term Rs variations based on cloud data, such as Advanced Very High Resolution Radiometer (AVHRR) cloud retrievals, especially before 2000, when satellite AOD retrievals are not unavailable. The merged Rs product at a spatial resolution of 0.1∘ in this study can be downloaded at https://doi.org/10.1594/PANGAEA.921847 (Feng and Wang, 2020).

A model framework to reduce bias in ground-level PM2.5 concentrations inferred from satellite-retrieved AOD

We present a new method to infer ground-level fine particulate matter (PM2.5) from satellite remote sensing observations of aerosol optical depth (AOD). The conventional method generally uses a range of modelling approaches to determine PM2.5:AOD relationships that are subsequently used to infer ground-level PM2.5 concentrations from satellite-retrieved AOD. Here, we use a high-resolution atmospheric chemistry simulation to explore how changes in the vertical distribution of aerosol extinction coefficients affects the PM2.5:AOD relationship and how we can use that information to improve the robustness of inferred estimates of ground-level PM2.5 over eastern China. We define a metric, ΓPBLAOD, that describes the fraction of AOD that resides in the planetary boundary layer compared with the total columnar AOD. We determine physically-meaningful PM2.5:AOD relationships using data for which ΓPBLAOD≥50%, a criterion based on sensitivity analyses on data clusters that we identify using a hierarchical clustering method. We use statistical and machine learning methods to develop independent models that describe these PM2.5:AOD relationships, and use a Monte Carlo approach to quantify the improvement after our selection of more physically relevant data records. Benefiting from the improved representativeness of AOD for ground-level PM2.5, our method effectively reduces bias in inferred estimates of ground-level PM2.5 by 10–15% (9–12%) for space-borne sensors passing over in the morning (afternoon). It also captures more variations in ground-level PM2.5 by up to 8% (5%) for space-borne sensors passing over in the morning (afternoon), particularly over areas dominated by natural aerosols such as dust. Accordingly, our method improves the seasonal ground-level PM2.5 maps, e.g. the bias of the autumn (winter) mean of ground-level PM2.5 estimates over Qinghai and Gansu (Shaaxi, Shanxi, and Henan) provinces reduces from −8% to −5% (11%–6%).

Theoretical Uncertainty Analysis of Satellite Retrieved Aerosol Optical Depth Associated with Surface Albedo and Aerosol Optical Properties

Deriving aerosol optical depth (AOD) from space-borne observations is still challenging due to uncertainties associated with sensor calibration drift, cloud screening, aerosol type classification, and surface reflectance characterization. As an initial step to understanding the physical processes impacting these uncertainties in satellite AOD retrievals, this study outlines a theoretical approach to estimate biases in the satellite aerosol retrieval algorithm affected by surface albedo and prescribed aerosol optical properties using a simplified radiative transfer model with a traditional error propagation approach. We expand the critical surface reflectance concept to obtain the critical surface albedo (CSA), critical single scattering albedo (CSSA), and critical asymmetry parameter (CAP). The top-of-atmosphere (TOA) reflectance is not sensitive to significant variability in aerosol loading (AOD) at the critical value; thus, the AOD cannot be determined. Results show that 5% bias in surface albedo (A), single scattering albedo (SSA), or asymmetry parameter (g) lead to large retrieved AOD errors, especially high under conditions when A, SSA, or g are close to their critical values. The results can be useful for future research related to improvements of satellite aerosol retrieval algorithms and provide a preliminary framework to analytically quantify AOD uncertainties from satellite retrievals.

The comparison of AOD-based and non-AOD prediction models for daily PM2.5 estimation in Guangdong province, China with poor AOD coverage

The large amount of missing values has challenged the application of satellite-retrieved aerosol optical depth (AOD) in mapping surface PM2.5 concentrations. In this study, we developed a non-AOD random forest model to estimate daily concentrations of PM2.5 in Guangdong Province, China, where more than 80% of AOD data were missing. The predictive ability of the non-AOD model was compared with that of a AOD-based model. Daily ground-based measurements of PM2.5 were obtained from 148 ground-monitoring sites in Guangdong with a 60 km rectangle buffer from January 2016 to December 2018. Daily MODIS MAIAC AOD were downloaded from NASA at a resolution of approximately 1 km. Two random forest models were developed to predict ground-level PM2.5 with one included AOD as a predictor and the other one without AOD. The two random forest models were developed using the same dataset and their predictive abilities were compared. The results of 10-fold cross validation (CV) showed that the non-AOD and AOD-based random forest models yielded similar performance. The CV R2 (RMSE) for the non-AOD model in 2016–2018 were 0.82 (8.4 μg/m3), 0.81 (9.5 μg/m3) and 0.78 (9.4 μg/m3), and those for AOD-based model were 0.83 (8.2 μg/m3), 0.82 (9.2 μg/m3) and 0.80 (9.0 μg/m3), respectively. Higher consistency of estimated PM2.5 were observed in areas close to monitoring sites than those far away from sites, and in southern coastal than northern areas. As the non-AOD random forest model is not affected by AOD missingness, it can be used for epidemiological studies to estimate individual-level exposure to PM2.5 at a high resolution without spatial or temporal gaps.

PM2.5 Estimation and Spatial-Temporal Pattern Analysis Based on the Modified Support Vector Regression Model and the 1 km Resolution MAIAC AOD in Hubei, China

The satellite-retrieved Aerosol Optical Depth (AOD) is widely used to estimate the concentrations and analyze the spatiotemporal pattern of Particulate Matter that is less than or equal to 2.5 microns (PM2.5), also providing a way for the related research of air pollution. Many studies generated PM2.5 concentration networks with resolutions of 3 km or 10 km. However, the relatively coarse resolution of the satellite AOD products make it difficult to determine the fine-scale characteristics of PM2.5 distributions that are important for urban air quality analysis. In addition, the composition and chemical properties of PM2.5 are relatively complex and might be affected by many factors, such as meteorological and land cover type factors. In this paper, an AOD product with a 1 km spatial resolution derived from the Multi-Angle Implementation of Atmospheric Correction (MAIAC) algorithm, the PM2.5 measurements from ground sites and the meteorological data as the auxiliary variable, are integrated into the Modified Support Vector Regression (MSVR) model that proposed in this paper to estimate the PM2.5 concentrations and analyze the spatiotemporal pattern of PM2.5. Considering the relatively small dataset and the somewhat complex relationship between the variables, we propose a Modified Support Vector Regression (MSVR) model that based on SVR to fit and estimate the PM2.5 concentrations in Hubei province of China. In this paper, we obtained Cross Correlation Coefficient (R²) of 0.74 for the regression of independent and dependent variables, and the conventional SVR model obtained R² of 0.60 as comparison. We think our MSVR model obtained relatively good performance in spite of many complex factors that might impact the accuracy. We then utilized the optimal MSVR model to perform the PM2.5 estimating, analyze their spatiotemporal patterns, and try to explain the possible reasons for these patterns. The results showed that the PM2.5 estimations retrieved from 1 km MAIAC AOD could reflect more detailed spatial distribution characteristics of PM2.5 and have higher accuracy than that from 3 km MODIS AOD. Therefore, the proposed MSVR model can be a better method for PM2.5 estimating, especially when the dataset is relatively small.

Assessment of COVID-19 effects on satellite-observed aerosol loading over China with machine learning

Aerosols are a critical component of the climate system and a risk to human health. Here, the lockdown response to the coronavirus outbreak is used to analyse effects of dramatic reduction in anthropogenic aerosol sources on satellite-retrieved aerosol optical depth (AOD). A machine learning model is applied to estimate daily AOD during the initial lockdown in China in early 2020. The model uses information on aerosol climatology, geography and meteorological conditions, and explains 69% of the day-to-day AOD variability. A comparison of model-expected and observed AOD shows that no clear, systematic decrease in AOD is apparent during the lockdown in China. During March 2020, regional AOD is observed to be significantly lower than expected by the machine learning model in some coastal regions of the North China Plains and extending to the Korean peninsula. While this may possibly indicate a small lockdown effect on regional AOD, and potentially pointing trans-boundary effects of the lockdown measures, due to uncertainties associated with the method and the limited sample sizes, this AOD decrease cannot be unequivocally attributed to reduced anthropogenic emissions. Climatologically expected AOD is compared to a weather-adjusted expectation of AOD, indicating that meteorological influences have acted to significantly increase AOD during this time, in agreement with recent literature. The findings highlight the complexity of aerosol variability and the challenges of observation-based attribution of columnar aerosol changes.

Variable Impact of COVID-19 Lockdown on Air Quality across 91 Indian Cities

AbstractIndia responded to the severe acute respiratory syndrome (SARS) coronavirus disease 2019 (COVID-19) pandemic through a three-phase nationwide lockdown: 25 March–14 April, 15 April–3 May, and 4–17 May 2020. We utilized this unique opportunity to assess the impact of restrictions on the air quality of Indian cities. We conducted comprehensive statistical assessments for the air quality index (AQI) and criteria pollutant concentrations for 91 cities during the lockdown phases relative to the preceding seven days (prelockdown phase of 18–24 March 2020) and to corresponding values from the same days of the year in 2019. Both comparisons show statistically significant countrywide mean decrease in AQI (33%), PM2.5 (36%), PM10 (40%), NO2 (58%), O3 (5%), SO2 (25%), NH3 (28%), and CO (60%). These reductions represent a background or the lower bound of air quality burden of industrial and transportation sectors. The northern region was most impacted by the first two phases of the lockdown, whereas the southern region was most affected in the last phase. The northeastern region was least affected, followed by the eastern region, which also showed an increase in O3 during the lockdown. Analysis of satellite-retrieved aerosol optical depth (AOD) shows that effects of restrictions on particulate pollution were variable—locally confined in some areas or having a broader impact in other regions. Anomalous behavior over the eastern region suggests a differing role of regional societal response or meteorological conditions. The study results have policy implications because they provide the observational background values for the industrial and transportation sector’s contribution to urban pollution.

Estimating PM2.5 concentrations in Northeastern China with full spatiotemporal coverage, 2005–2016

Predicting long-term spatiotemporal characteristics of fine particulate matter (PM2.5) is important in China to understand historical levels of PM2.5, to support health effects research of both long-term and short-term exposures to PM2.5, and to evaluate the efficacy of air pollution control policies. Satellite-retrieved aerosol optical depth (AOD) provides a unique opportunity to characterize the long-term trends of ground-level PM2.5 at high spatial resolution. However, the missing rate of AOD in Northeastern China (NEC) is very high, especially in winter, and challenges the accuracy of long-term predictions of PM2.5 if left unresolved. Using random forest algorithms, this study developed a gap-filling approach combing satellite AOD, meteorological data, land use parameters, population and visibility in the NEC during 2005–2016. The model, including all predictors, combined with a model without AOD was able to fill the gap of PM2.5 predictions caused by missing AOD at 1-km resolution. The R2 (RMSE) of the full-coverage predictions was 0.81 (18.5 μg/m3) at the daily level. Gap-filled PM2.5 predictions on days with missing AOD reduced the relative prediction error from 28% to 2.5% in winter. The leave-one-year-out-cross-validation R2 (RMSE) of the full-coverage predictions was 0.65 (16.3 μg/m3) at the monthly level, indicating relatively high accuracy of predicted historical PM2.5 concentrations. Our results suggested that AOD helped increase the reliability of historical PM2.5 prediction when ground PM2.5 measurements were unavailable, even though predictions from the AOD model only accounted for approximate 37% of the whole dataset. Predicted PM2.5 level in NEC have increased since 2005, reached its peak during 2013–2015, then saw a major decline in 2016. Our high-resolution predictions also showed a south to north gradient and many pollution hot spots in the city clusters surrounding provincial capitals, as well as within large cities. Overall, by combining predictions from the AOD model with higher accuracy and predictions from the non-AOD model to achieve full coverage, our modeling approach could produce long-term, full-coverage historical PM2.5 levels in high-latitude areas in China, despite the widespread and persistent AOD missingness.