A machine learning model for predicting PM2.5 and nitrate concentrations based on long-term water-soluble inorganic salts datasets at a road site station

Mapping nationwide concentrations of sulfate and nitrate in ambient PM2.5 in South Korea using machine learning with ground observation data

Particulate matter (PM) is a major air pollutant in Northeast Asia, with frequent high PM episodes. To investigate the nationwide spatial distribution maps of PM2.5 and secondary inorganic aerosols in South Korea, prediction models for mapping SO42- and NO3- concentrations in PM2.5 were developed using machine learning with ground-based observation data. Specifically, the random forest algorithm was used in this study to predict the SO42- and NO3- concentrations at 548 air quality monitoring stations located within the representative radii of eight intensive air quality monitoring stations. The average concentrations of PM2.5, SO42-, and NO3- across the entire nation were 17.2 ± 2.8, 3.0 ± 0.6, and 3.4 ± 1.2 μg/m3, respectively. The spatial distributions of SO42- and NO3- concentrations in 2021 revealed elevated concentrations in both the western and central regions of South Korea. This result suggests that SO42- concentrations were primarily influenced by industrial activities rather than vehicle emissions, whereas NO3- concentrations were more associated with vehicle emissions. During a high PM2.5 event (November 19-21, 2021), the concentration of SO42- was primarily influenced by SOX emissions from China, while the concentration of NO3- was affected by NOX emissions from both China and Korea. The methodology developed in this study can be used to explore the chemical characteristics of PM2.5 with high spatiotemporal resolution. It can also provide valuable insights for the nationwide mitigation of secondary PM2.5 pollution.

Impact assessment of spatial-temporal distribution of riverine dust on air quality using remote sensing data and numerical modeling

Soil erosion is a severe problem in Taiwan due to the steep terrain, fragile geology, and extreme climatic events resulting from global warming. Due to the rapidly changing hydrological conditions affecting the locations and the amount of transported sand and fine particles, timely impact evaluation and riverine dust control are difficult, particularly when resources are limited. To comprehend the impact of desertification in estuarine areas on the variation of air pollutant concentrations, this study utilized remote sensing technology coupled with an air pollutant dispersion model to determine the unit contribution of potential pollution sources and quantify the effect of riverine dust on air quality. The images of the downstream area of the Beinan River basin captured by Formosat-2 in May 2006 were used to analyze land use and land cover (LULC) composition. Subsequently, the diffusion model ISCST-3 based on Gaussian distribution was utilized to simulate the transport of PM across the study area. Finally, a mixed-integer programming model was developed to optimize resource allocation for dust control. Results reveal that sand deposition in specific river sections significantly influences regional air quality, owing to the unique local topography and wind field conditions. The present optimal plan model for regional air quality control further showed that after implementing engineering measures including water cover, revegetation, armouring cover, and revegetation, total PM concentrations would be reduced by 51%. The contribution equivalent calculation, using the air pollution diffusion model, was effectively integrated into the optimization model to formulate a plan for reducing riverine dust with limited resources based on air quality requirements.

Prediction of atmospheric PM2.5 level by machine learning techniques in Isfahan, Iran

With increasing levels of air pollution, air quality prediction has attracted more attention. Mathematical models are being developed by researchers to achieve precise predictions. Monitoring and prediction of atmospheric PM2.5 levels, as a predominant pollutant, is essential in emission mitigation programs. In this study, meteorological datasets from 9 years in Isfahan city, a large metropolis of Iran, were applied to predict the PM2.5 levels, using four machine learning algorithms including Artificial Neural |Networks (ANNs), K-Nearest-Neighbors (KNN), Support Vector |Machines (SVMs) and ensembles of classification trees Random Forest (RF). The data from 7 air quality monitoring stations located in Isfahan City were taken into consideration. The Confusion Matrix and Cross-Entropy Loss were used to analyze the performance of classification models. Several parameters, including sensitivity, specificity, accuracy, F1 score, precision, and the area under the curve (AUC), are computed to assess model performance. Finally, by introducing the predicted data for 2020 into ArcGIS software and using the IDW (Inverse Distance Weighting) method, interpolation was conducted for the area of Isfahan city and the pollution map was illustrated for each month of the year. The results showed that, based on the accuracy percentage, the ANN model has a better performance (90.1%) in predicting PM2.5 grades compared to the other models for the applied meteorological dataset, followed by RF (86.1%), SVM (84.6%) and KNN (82.2%) models, respectively. Therefore, ANN modelling provides a feasible procedure for the managerial planning of air pollution control.

Enhanced nitrate contribution to light extinction during haze pollution in Chengdu: Insights based on an improved multiple linear regression model

In recent years, the annual mean concentration of PM_2.5 has decreased in Chengdu, China; however, atmospheric visibility has not improved accordingly. Low-visibility events occurred even when the PM_2.5 mass concentrations were below the national ambient air quality secondary standard (daily mean concentration, 75 μg/m³). In this study, the non-linear relationship between PM_2.5 and visibility was analyzed under different NO₃^- mass fractions in PM_2.5 based on 2-year field observation data. The results indicated that NO₃^- formation contributed to particulate pollution events and reduced atmospheric visibility. Multiple linear regression was used to propose a localized reconstruction equation for the light-scattering coefficient. According to the maximum likelihood estimation method and log-transformed residuals, the mass scattering coefficients (MSEs) of organic matter (OM), NH₄NO₃, and (NH₄)₂SO₄ in Chengdu were 7.42, 3.83, and 3.80, respectively. OM and NH₄NO₃ contributed to more than 50% of the light-extinction coefficient (b_ext). NH₄NO₃ was the main pollutant causing the substantial increase in b_ext. Chengdu has a high relative humidity (annual mean 70%), and under such conditions, the contribution of NH₄NO₃ to b_ext was considerably enhanced through hygroscopic growth and heterogeneous reactions. This study estimated the localized MSEs of OM, NH₄NO₃, and (NH₄)₂SO₄ in Chengdu and emphasized that effective control measures to reduce nitrate and its precursors could simultaneously ameliorate air quality and visibility in humid regions with poor atmospheric visibility.

Chang impact analysis of level 3 COVID-19 alert on air pollution indicators using artificial neural network

In this study, mean monthly and diurnal variations in fine particulate matters (PM_2.5), nitrate, sulfate, and gaseous precursors were investigated during the Level 3 COVID-19 alert from May 19 to July 27 in 2021. For comparison, the historical data during the identical period in 2019 and 2020 were also provided to determine the effect of the Level 3 COVID-19 alert on aerosols and gaseous pollutants concentrations in Taichung City. A machine learning model using the artificial neural network technique coupled with a kinetic model was applied to predict NO_x, O₃, nitrate (NO₃ ^-), and sulfate (SO₄ ^2-) to investigate potential emission sources and chemical reaction mechanism. D during the Level 3 COVID-19 alert, a decrease in NO_x concentration due to a decrease in traffic flow under the NO_x-saturated regime was observed to enhance the secondary NO₃ ^- and O₃ formation. The present models were shown to predict 80.1, 77.0, 72.6, and 67.2% concentrations of NO_x, O₃, NO₃ ^-, and SO₄ ^2-, respectively, which could help decision-makers for pollutant emissions reduction policies development and air pollution control strategies. It is recommended that more long-term datasets, including water soluble inorganic salts (WIS), precursors including OH radicals, NH₃, HNO₃, and H₂SO₄, be provided by regulatory air quality monitoring stations to further improve the prediction model accuracy.

Historically understanding the spatial distributions of particle surface area concentrations over China estimated using a non-parametric machine learning method

A non-parametric ensemble model was proposed to estimate the long-term (2015-2019) particle surface area concentrations (S_A) over China for the first time on basis of a vilification dataset of measured particle number size distribution. This ensemble model showed excellent cross-validation R² value (CV R² = 0.83) as well as a relatively low root-mean-square error (RMSE = 195.0 μm²/cm³). No matter in which year, considerable spatial heterogeneity of S_A was found over China with higher S_A in Beijing-Tianjin-Hebei (BTH), Yangtze River Delta (YRD), and Middle Lower Reaches of Yangtze River (MLYR). From 2015 to 2019, S_A significantly decreased in representative city clusters. The reduction rates were 140.1 μm²·cm^-3·a^-1 in BTH, 110.7 μm²·cm^-3·a^-1 in Pearl River Delta (PRD), 105.2 μm²·cm^-3·a^-1 in YRD, and 92.4 μm²·cm^-3·a^-1 in Sichuan Basin (SCB), respectively. Even though such quick reduction, high S_A (ranged from ~800 μm²/cm³ to ~1750 μm²/cm³) during the heavy pollution period (PM_2.5 > 75 μg/m³) still existed in the above-mentioned city clusters and may provide rich reaction vessels for multiphase chemistry. A dichotomy of enhanced annual 4th maximum daily 8-h average O₃ concentrations (4MDA8 O₃) and decreased S_A during summertime was found in Shanghai, a representative city of YRD. In Chengdu (SCB), increased 4MDA8 O₃ concentration was associated with a synchronous increase of S_A from 2017 to 2019. Differently, 4MDA8 O₃ concentrations enhanced in Beijing (BTH) and Guangzhou (PRD), while not significant for S_A before 2018. This work will greatly deepen our understanding of the historical variation and spatial distributions of S_A over China.