Understanding global changes in fine-mode aerosols during 2008-2017 using statistical methods and deep learning approach

Estimation of pan-European, daily total, fine-mode and coarse-mode Aerosol Optical Depth at 0.1° resolution to facilitate air quality assessments

Aerosol Optical Depth (AOD) data derived from satellites is crucial for estimating spatially-resolved PM concentrations, but existing AOD data over land remain affected by several limitations (e.g., data gaps, coarser resolution, higher uncertainty or lack of size fraction data), which weakens the AOD-PM relationship. We developed a 0.1° resolution daily AOD data set over Europe over the period 2003-2020, based on two-stage Quantile Machine Learning (QML) frameworks. Our approach first fills gaps in satellite AOD data and then constructs three components' models to obtain reliable full-coverage AOD along with Fine-mode AOD (fAOD) and Coarse-mode AOD (cAOD). These models are based on AERONET (AErosol RObotic NETwork) observations, Gap-filled satellite AOD, climate and atmospheric composition reanalyses. Our QML AOD products exhibit better quality with an out-of-sample R2 equal to 0.68 for AOD, 0.66 for fAOD and 0.65 for cAOD, which is 23-92 %, 11-13 % and 115-132 % higher than the corresponding satellite or reanalysis products, respectively. Over 91.6 %, 81.6 %, and 88.9 % of QML AOD, fAOD and cAOD predictions fall within ±20 % Expected Error (EE) envelopes, respectively. Previous studies reported that a weak satellite AOD-PM correlation across Europe (Pearson correlation coefficient (PCC) around 0.1). Our QML products exhibit higher correlations with ground-level PMs, particularly when broadly matched by size: AOD with PM10, fAOD with PM2.5, cAOD with PM coarse (R = 0.41, 0.45 and 0.26, respectively). Different AOD fractions more effectively distinct PM size fractions, than total AOD. Our QML aerosol dataset and models pioneer full-coverage, daily high-resolution monitoring of fine-mode and coarse-mode aerosols, effectively addressing existing AOD challenges for further PMs exposures' estimations. This dataset opens avenues for more in-depth exploration of the impacts of aerosols on human health, climate, visibility, and biogeochemical processes, offering valuable insights for air quality management and environmental health risk assessment.

Columnar optical-radiative properties and components of aerosols in the Arctic summer from long-term AERONET measurements

Aerosols as an external factor have an important role in the amplification of Arctic warming, yet the geography of this harsh region has led to a paucity of observations, which has limited our understanding of the Arctic climate. We synthesized the latest decade (2010-2021) of data on the microphysical-optical-radiative properties of aerosols and their multi-component evolution during the Arctic summer, taking into consideration the important role of wildfire burning. Our results are based on continuous observations from eight AERONET sites across the Arctic region, together with a meteorological reanalysis dataset and satellite observations of fires, and utilize a back-trajectory model to track the source of the aerosols. The summer climatological characteristics within the Arctic Circle showed that the aerosols are mainly fine-mode aerosols (fraction >0.95) with a radius of 0.15-0.20 μm, a slight extinction ability (aerosol optical depth ∼ 0.11) with strong scattering (single scattering albedo ∼0.95) and dominant forward scattering (asymmetry factor ∼ 0.68). These optical properties result in significant cooling at the Earth's surface (∼-13 W m-2) and a weak cooling effect at the top of the atmosphere (∼-5 W m-2). Further, we found that Arctic region is severely impacted by wildfire burning events in July and August, which primarily occur in central and eastern Siberia and followed in subpolar North America. The plumes from wildfire transport aerosols to the Arctic atmosphere with the westerly circulation, leading to an increase in fine-mode aerosols containing large amounts of organic carbon, with fraction as high as 97-98 %. Absorptive carbonaceous aerosols also increase synergistically, which could convert the instantaneous direct aerosol radiative effect into a heating effect on the Earth-atmosphere system. This study provides insights into the complex sources of aerosol loading in the Arctic atmosphere in summer and emphasizes the important impacts of the increasingly frequent occurrence of wildfire burning events in recent years.

Deep neural networks for spatiotemporal PM2.5 forecasts based on atmospheric chemical transport model output and monitoring data

Reliable long-horizon PM_2.5 forecasts are crucial and beneficial for health protection through early warning against air pollution. However, the dynamic nature of air quality makes PM_2.5 forecasts at long horizons very challenging. This study proposed a novel machine learning-based model (MCNN-BP) that fused multiple convolutional neural networks (MCNN) with a back-propagation neural network (BPNN) for making spatiotemporal PM_2.5 forecasts for the next 72 h at 74 stations covering the whole Taiwan simultaneously. Model configuration involved an ensemble of massive hourly air quality and meteorological monitoring datasets and the existing publicly-available PM_2.5 simulated (forecasted) datasets from an atmospheric chemical transport (ACT) model. The proposed methodology collaboratively constructed two CNNs to mine the observed data (the past) and the forecasted data from ACT (the future) separately. The results showed that the MCNN-BP model could significantly improve the accuracy of spatiotemporal PM_2.5 forecasts and substantially reduce the forecast biases of the ACT model. We demonstrated that the proposed MCNN-BP model with effective feature extraction and good denoising ability could overcome the curse of dimensionality and offer satisfactory regional long-horizon PM_2.5 forecasts. Moreover, the MCNN-BP model has considerably shorter computational time (5 min) and lower computational load than the compute-intensive ACT model. The proposed approach hits a milestone in multi-site and multi-horizon forecasting, which significantly contributes to early warning against regional air pollution.

Determination of the human impact on the drop in NO2 air pollution due to total COVID-19 lockdown using Human-Influenced Air Pollution Decrease Index (HIAPDI)

This study investigates the relationship between territorial human influence and decreases in NO2 air pollution during a total COVID-19 lockdown in Metropolitan France. NO2 data from the confinement period and the Human Influence Index (HII) were implemented to address the problem. The relative change in tropospheric NO2 was calculated using Sentinel-5P (TROPOMI) satellite data. Hotspot-Coldspot analysis was performed to examine the change in NO2. Moreover, the novel Human-Influenced Air Pollution Decrease Index (HIAPDI) was developed. Weather bias was investigated by implementing homogeneity analysis with χ2 test. The correlations between variables were tested with the statistical T-test. Likewise, remote observations were validated with data from in-situ monitoring stations. The study showed a strong correlation between the NO2 decrease during April 2020 under confinement measures and HII. The greater the anthropogenic influence, the greater the reduction of NO2 in the regions (R2 = 0.62). The new HIAPDI evidenced the degree of anthropogenic impact on NO2 change. HIAPDI was found to be a reliable measure to determine the correlation between human influence and change in air pollution (R2 = 0.93). It is concluded that the anthropogenic influence is a determining factor in the phenomenon of near-surface NO2 reduction. The implementation of HIAPDI is recommended in the analysis of other polluting gases.

Health risk and disease burden attributable to long-term global fine-mode particles

Particulate matter 2.5 (PM_2.5) pollution has long been a global environmental problem and still poses a great threat to public health. This study investigates global spatiotemporal variations in PM_2.5 using the newly developed satellite-derived PM_2.5 dataset from 1998 to 2018. An integrated exposure-response (IER) model was employed to examine the characteristics of PM_2.5-related deaths caused by chronic obstructive pulmonary disease (COPD), ischemic heart disease (IHD), lung cancer (LC), and stroke in adults (age≥25), as well as lower respiratory infection (LRI) in children (age≤5). The results showed that high annual PM_2.5 concentrations were observed mainly in East Asia and South Asia. Over the 19-year period, PM_2.5 concentrations constantly decreased in developed regions, but increased in most developing regions. Approximately 84% of the population lived in regions where PM_2.5 concentrations exceeded 10 μg/m³. Meanwhile, the vast majority of the population (>60%) in East and South Asia was consistently exposed to PM_2.5 levels above 35 μg/m³. PM_2.5 exposure was linked to 3.38 (95% UI: 3.05-3.70) million premature deaths globally in 2000, a number that increased to 4.11 (95% UI: 3.55-4.69) million in 2018. Premature deaths related to PM_2.5 accounted for 6.54%-7.79% of the total cause of deaths worldwide, with a peak in 2011. Furthermore, developing regions contributed to the majority (85.95%-95.06%) of PM_2.5-related deaths worldwide, and the three highest-ranking regions were East Asia, South Asia, and Southeast Asia. Globally, IHD and stroke were the two main contributors to total PM_2.5-related deaths, followed by COPD, LC, and LRI.