Atmospheric Pollution Research最新文献

{"title":"Estimating and characterizing spatiotemporal distributions of elemental PM2.5 using an ensemble machine learning approach in Taiwan","authors":"Chun-Sheng Huang ,&nbsp;Kang Lo ,&nbsp;Yee-Lin Wu ,&nbsp;Fu-Cheng Wang ,&nbsp;Yi-Shiang Shiu ,&nbsp;Chu-Chih Chen ,&nbsp;Yuan-Chien Lin ,&nbsp;Cheng-Pin Kuo ,&nbsp;Ho-Tang Liao ,&nbsp;Tang-Huang Lin ,&nbsp;Chang-Fu Wu","doi":"10.1016/j.apr.2025.102463","DOIUrl":"10.1016/j.apr.2025.102463","url":null,"abstract":"<div><div>This paper presents an ensemble machine learning approach that combines Generalized Additive Model (GAM) with eXtreme Gradient Boosting (XGBoost) to estimate and characterize the spatiotemporal distributions of elemental PM_2.5 in Taiwan. Daily field measurements of 12 PM_2.5 elemental components were collected from 28 air quality monitoring stations between June 2021 and May 2022. Time-variant meteorological factors and time-invariant land-use patterns were incorporated as predictors. Results showed that the ensemble model effectively captured spatial variations in elemental PM_2.5 levels, as demonstrated by the identification of numerous time-invariant features using Shapley additive explanations analysis. A comparative analysis was conducted with a model using only XGBoost, which outperformed the ensemble model with higher cross-validated <em>R</em>^<em>2</em> and lower prediction errors. While the XGBoost-only model is recommended for exposure prediction, the ensemble model offers superior interpretability for investigating air pollution sources and aids in formulating air quality strategies from a spatial perspective.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102463"},"PeriodicalIF":3.9,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143437102","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reducing PM2.5 and O3 through optimizing urban ecological land form based on its size thresholds","authors":"Xin Chen,&nbsp;Fang Wei","doi":"10.1016/j.apr.2025.102466","DOIUrl":"10.1016/j.apr.2025.102466","url":null,"abstract":"<div><div>Optimizing the size and form of urban ecological land (UEL) is an effective approach to addressing PM_2.5-O₃ composite pollution in China. However, existing strategies are usually proposed based on the impact of one type of UEL on individual pollutants, while overlooking UEL forms’ different pollution reduction effects across its size intervals. This study identifies UELs (including forest, shrub, grassland, water, and wetland) of 1068 counties within the Yangtze River Economic Belt (YREB) and calculates their size and form metrics. Then the cross-sectional threshold regression model is used to analyze the threshold effect of UEL size on fitting models of pollutant concentrations. Finally, quadrant analysis is extended to categorize counties and provide differentiated planning strategies. The conclusions show: (1) UEL size presents a triple threshold effect on PM_2.5 concentrations at 4.302%, 8.055%, and 23.742%, and a single threshold effect on O₃ concentrations at 3.275%. Size and form metrics are not always significant across UEL size intervals. (2) Counties are categorized into 6 types based on their primary pollutants and UEL sizes, showing spatial clustering within each type. (3) With size increasing, dispersed and irregular UEL form helps more in reducing PM_2.5, while O₃ reduction prefers aggregated one, thereby the evolutionary UEL planning strategy is proposed.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102466"},"PeriodicalIF":3.9,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143430080","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Modelling lung deposition of fine particulate matter in males and females during urban cycle commuting","authors":"Gustavo Oneda ,&nbsp;Gabriel Moresco ,&nbsp;Danilo Fonseca Leonel ,&nbsp;Leonardo Hoinaski ,&nbsp;Joseph F. Welch ,&nbsp;Sarah Koch ,&nbsp;Ramon Cruz","doi":"10.1016/j.apr.2025.102467","DOIUrl":"10.1016/j.apr.2025.102467","url":null,"abstract":"<div><div>Exposure to fine particulate matter (PM_2.5) from urban areas may be modified by structural (e.g., airway anatomy) and functional (e.g., ventilatory pattern) sex-related physiological differences during exercise, resulting in greater PM_2.5 deposition in females versus males. Beyond the total PM_2.5 deposition, further insights concerning regional differences in PM_2.5 deposition are needed to understand females’ hyperresponsiveness to PM_2.5. Thus, a modelling-based analysis of structural and functional characteristics of PM_2.5 deposition in the human respiratory tract was conducted simulating an urban cycle commute of 30 min. Two scenarios were considered to estimate the PM_2.5 deposition: 1) greater minute ventilations in females versus males (p &lt; 0.001); and 2) minute ventilations matched between males and females (p = 0.710). We found that females experience 51.32% and 0.62% greater total PM_2.5 deposition for Scenarios 1 and 2, respectively (both p &lt; 0.001). Regardless of total minute ventilation, there was greater PM_2.5 deposition into the bronchiolar and alveolar region in females compared to males (p &lt; 0.001 for both). These data indicate a greater likelihood of bronchial hyperresponsiveness in females compared with males when exposed to PM_2.5 while cycle commuting in urban areas.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102467"},"PeriodicalIF":3.9,"publicationDate":"2025-02-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143419420","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Environmental impacts and costs of ozone formation in Bangkok Metropolitan Region","authors":"Soe Htet Aung ,&nbsp;Shabbir H. Gheewala ,&nbsp;Ekbordin Winijkul ,&nbsp;Sirima Panyametheekul ,&nbsp;Trakarn Prapaspongsa","doi":"10.1016/j.apr.2025.102450","DOIUrl":"10.1016/j.apr.2025.102450","url":null,"abstract":"<div><div>Ozone formation is an important environmental factor causing impacts on human health and ecosystem. Previous research relating to ozone formation often had limited scopes on direct emissions or focused on limited sectors of cities. This study aimed to quantify environmental impacts and costs due to ozone formation caused by energy generation, industry, agriculture, residential and commercial sectors, transport, fugitive gas emissions and waste treatment in the Bangkok Metropolitan Region (BMR) in 2022. The assessment applied spatially differentiated life cycle assessment framework, quantifying impacts on human health and ecosystem using local and global factors for on-site and supply chain emissions. The baseline situation in 2022 revealed that total emissions (on-site and supply chain) were 3.97E+05 tonnes of NO_x and 1.15E+05 tonnes of NMVOC. NO_x and the transport sector were the main stressor and hotspot causing impacts and costs of ozone formation in BMR. Total impact scores (on-site and supply chain) were 4.39E+02 disability-adjusted life year (human health impact) and 3.50E+01 species.year (ecosystem damage). The impacts were mainly contributed by on-site activities in BMR costing 6 billion Thai Baht. Scenarios were developed focusing primarily on the on-road transport since it was the hotspot causing health impacts and ecosystem damage. The scenarios indicated that upgrading fuel technology from diesel to compressed natural gas and modification of vehicles from diesel to electric were found to be very effective for overall reduction by more than 50% on average for health impacts and by more than 40% on average for ecosystem damage.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102450"},"PeriodicalIF":3.9,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143487697","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A significant reduction in the coal contribution to PM2.5 and exposed health risks due to the energy structure transition","authors":"Dongmei Hu ,&nbsp;Mingyang Yuan ,&nbsp;Yulong Yan ,&nbsp;Xiaolin Duan ,&nbsp;Yafei Guo ,&nbsp;Yueyuan Niu ,&nbsp;Wen Yan ,&nbsp;Lin Peng","doi":"10.1016/j.apr.2025.102457","DOIUrl":"10.1016/j.apr.2025.102457","url":null,"abstract":"<div><div>By collecting PM_2.5 samples containing heavy metals in a typical coal resource-based city, we analyzed the interannual variation of heavy metal concentrations over an extended time period (2018–2022). This analysis involved apportioning the sources of these heavy metals and evaluating the carcinogenic and non-carcinogenic health risks posed to different populations via the respiratory route. Results showed that Cd (83.99%), Zn (62.56%), Pb (56.97%), and As (2.60%) were associated with coal combustion, exhibiting decreasing trends. The maximal information coefficient (MIC) indicated that most of the elements with strong correlations were associated with coal combustion. Four sources, namely coal combustion, resuspended dust, traffic emission, and industry, were determined using positive matrix factorization. Cr posed the highest carcinogenic risk, particularly among adults. Coal consumption and its contribution showed significant reductions due to the energy structure transition of coal reduction. Notably, the top three metals in terms of carcinogenic risk were all associated with coal combustion. The carcinogenic risk associated with Cd and As from coal combustion was significantly lower in 2022 than in 2018.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102457"},"PeriodicalIF":3.9,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143453113","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Research on the impact of land use and meteorological factors on the spatial distribution characteristics of PM2.5 concentration","authors":"Xiaoxia Wang ,&nbsp;Hongtao Zhang ,&nbsp;Zhihai Fan ,&nbsp;Hong Ding","doi":"10.1016/j.apr.2025.102462","DOIUrl":"10.1016/j.apr.2025.102462","url":null,"abstract":"<div><div>Precisely capturing the spatial distribution characteristics of fine particulate matter (PM_2.5) is the key to air pollution prevention and control. Researches suggests that PM_2.5 concentrations is jointly affected by meteorological factors and urban land use. This study comprehensively considered these factors, with land cover, meteorology and road traffic as potential independent variables. The land use regression (LUR) model was used to identify the main influences on the spatial distribution of PM_2.5 concentration under different types of land use, and the importance of these factors was assessed using a random forest (RF) model. The research findings indicate that: (1) The fluctuations of PM_2.5 concentration in residential and industrial land use are relatively severe, ranging from 0 to 50 μg/m³. The changes in commercial and public service land use range from 0 to 40 μg/m³. And the in green space range from 15 to 33 μg/m³ (2) The distribution of PM_2.5 concentration in residential land is primarily influenced by precipitation and relative humidity, with relative importance of 36.3% and 23.9%, respectively. And that in industrial land use is primarily influenced by wind speed, with a relative importance of 30.6%. (3) The most influential determinant of spatial distribution of PM_2.5 concentration is meteorological factors, with relative importance exceeding 62%. The relative importance of road traffic ranges from 15.6% to 24.9%. And that of land cover factors ranges from 9.9% to 22.4%. This study analyzes the coupling connection between urban land use and spatial distribution of PM_2.5 concentration, and elaborates the specific influence of the former on the latter.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102462"},"PeriodicalIF":3.9,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143479229","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Household PM2.5 in a South African urban and rural setting: A comparative analysis using low-cost sensors","authors":"Matthew Benyon ,&nbsp;Ngwako Kwatala ,&nbsp;Tracey Laban ,&nbsp;Thandi Kapwata ,&nbsp;Chiara Batini ,&nbsp;Samuel Cai ,&nbsp;Lisa K. Micklesfield ,&nbsp;Rikesh Panchal ,&nbsp;Siyathemba Kunene ,&nbsp;Sizwe B. Zondo ,&nbsp;Brigitte Language ,&nbsp;Bianca Wernecke ,&nbsp;Scott Hazelhurst ,&nbsp;F. Xavier Gómez-Olivé ,&nbsp;Joshua Vande Hey ,&nbsp;Caradee Y. Wright","doi":"10.1016/j.apr.2025.102459","DOIUrl":"10.1016/j.apr.2025.102459","url":null,"abstract":"<div><div>Household air pollution (HAP) is responsible for millions of premature deaths each year<em>.</em> Exposure to household air pollutants as a risk factor for poor health has not been adequately quantified in many parts of the world, especially Sub-Saharan Africa. We aimed to assess HAP, specifically PM_2.5, and its associations with dwelling and household characteristics in urban (Soweto) and rural (Agincourt) settings in South Africa. We monitored indoor PM_2.5 concentrations in 40 unique households using low-cost sensors, across two study sites and seasons. Low-cost sensors were calibrated by collocation, and associations between dwelling and household characteristics with indoor PM_2.5 concentrations were assessed using a log-linear regression model. PM_2.5 concentrations were greater in urban households in the summer (50 μg/m³ (95% CI: 41–63) and in the winter (82 μg/m³ (95% CI: 62–109)) compared to rural households (summer: 19 μg/m³ (95%: CI 14–26) and winter: 48 μg/m³ (95% CI: 44–53)). The log-linear model (n = 39) explained 74% of the variance in leave-one-out cross validation. Significant associations with household PM_2.5 were observed with the following: the season, study setting, presence of tobacco smoking, presence of incense burning inside the dwelling, and the use of heating. This study found significant variations in HAP concentrations within and across the urban and rural communities, likely influenced by differences in ambient outdoor concentrations and individual behaviours such as incense burning. It is crucial to enhance community and policy maker awareness regarding the dangers of indoor smoking and the harmful effects of burning incense indoors.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102459"},"PeriodicalIF":3.9,"publicationDate":"2025-02-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143437105","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Unprecedented impacts of meteorological and photolysis rates on ozone pollution in a coastal megacity of northern China","authors":"Jianli Yang ,&nbsp;Chaolong Wang ,&nbsp;Yisheng Zhang ,&nbsp;Sufan Zhang ,&nbsp;Xing Peng ,&nbsp;Xiaofei Qin ,&nbsp;Jianhui Bai ,&nbsp;Lian Xue ,&nbsp;Guan Wang ,&nbsp;Shanshan Cui ,&nbsp;Wenxin Tao ,&nbsp;Jinhua Du ,&nbsp;Dasa Gu ,&nbsp;Xiaohan Su","doi":"10.1016/j.apr.2025.102461","DOIUrl":"10.1016/j.apr.2025.102461","url":null,"abstract":"<div><div>This study investigates the seasonal variations in O₃ levels in Qingdao, a typical coastal city, and quantifies the effects of key photolysis rate constants (<em>J</em>[O¹D] and <em>J</em>[NO₂]), meteorological parameters (RH, TEMP, and SF), and pollutants (ΔCO, PM_2.5, and NO₂) on O₃ levels across different seasons using machine learning. Additionally, the summer months, when photochemical reactions are most active, were analyzed in detail. The results indicate that the factors contributing to summer O₃ levels in order of importance, were RH, ΔCO, SF, PM_2.5, <em>J</em>[O¹D], NO₂, TEMP, WS, and <em>J</em>[NO₂]. RH was the most significant factor, with high humidity levels (&gt;75%) inhibiting O₃ formation. ΔCO, representing regional transport, was the second most influential, suggesting that direct O₃ transport and the delivery of high concentrations of precursors significantly promoted local O₃ production and accumulation. While <em>J</em>[O¹D] and <em>J</em>[NO₂] had different roles in O₃ promotion and depletion, <em>J</em>[O¹D] had a greater impact overall. The temperature in the range of 26 °C–32 °C inhibits O₃ production, When RH exceeded 90%, <em>J</em>[O¹D] accelerates while other photolysis rate constants decline, further suppressing the production of O₃. For comparison, multiple linear regression models were used to develop empirical equations for calculating hourly O₃ concentrations across the four seasons. The results showed that these factors explained 50%, 64%, 61%, and 63% of the O₃ sources in Qingdao for spring, summer, autumn, and winter, respectively. Sensitivity tests on factors influencing summer O₃ concentrations found that MLR could not quantify their contributions to O₃ levels.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102461"},"PeriodicalIF":3.9,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143419419","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Evaluating the spatiotemporal variations in atmospheric CO2 concentrations in China and identifying factors contributing to its increase","authors":"Weixin Zhu ,&nbsp;Hong Zhang ,&nbsp;Xiaoyu Zhang ,&nbsp;Haohao Guo ,&nbsp;Yong Liu","doi":"10.1016/j.apr.2025.102458","DOIUrl":"10.1016/j.apr.2025.102458","url":null,"abstract":"<div><div>Understanding the patterns and trends of atmospheric carbon dioxide (CO₂) is essential for comprehending the global carbon cycle and making accurate future climate predictions. CO₂ levels are influenced by complex and often interrelated factors, requiring innovative approaches that can tie place-specific factors with CO₂ concentrations. This study utilized the Orbiting Carbon Observatory-2 (OCO-2) data to explore the changes of CO₂ concentrations in China over the past decade. Additionally, climate parameters, vegetation cover, and anthropogenic activities were combined to explain temporal and spatial changes in CO₂ concentrations, using Geodetector and Multiscale Geographically Weighted Regression (MGWR) model. The results revealed a consistent increase (2.54 ppm/yr) and significant spatial agglomeration (High-High cluster in the east, Low-Low cluster in the west) of CO₂ concentrations in China. The spatial location (<em>q</em> = 0.68) emerged as the primary determinant of CO₂ levels, with population variable (<em>q</em> = 0.55) representing the secondary influencing factor. The interactions among natural elements and anthropogenic activities had substantially elevated CO₂ levels. Compared to the Geographically Weighted Regression (GWR), and Ordinary Least Squares (OLS) models, the MGWR model demonstrated superior capability in revealing the varying spatial scales of influence among different variables, making it more suitable for investigating the impacts of multiple factors on atmospheric CO₂ concentrations. The MGWR revealed significant variations in the optimal bandwidths among different explanatory variables, with temperature, precipitation, and LAI operating at much smaller scales. The findings are expected to provide valuable insights into regional processes influencing CO₂ concentrations and the development of targeted interventions.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102458"},"PeriodicalIF":3.9,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143474202","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"An integrated feature selection and machine learning framework for PM10 concentration prediction","authors":"Elham Kalantari ,&nbsp;Hamid Gholami ,&nbsp;Hossein Malakooti ,&nbsp;Dimitris G. Kaskaoutis ,&nbsp;Poorya Saneei","doi":"10.1016/j.apr.2025.102456","DOIUrl":"10.1016/j.apr.2025.102456","url":null,"abstract":"<div><div>The Sistan Basin, east Iran is a major dust source, presenting significant atmospheric, ecological, socio-economic, and health challenges. This study employed machine learning (ML) algorithms, including Random Forest (RF), K-Nearest Neighbor (KNN), Weighted K-Nearest Neighbor (WKNN), Support Vector Regression (SVR), and Least Absolute Shrinkage and Selection Operator (LASSO), to model and predict PM₁₀ concentrations in Zabol City (2013–2022), utilizing independent meteorological variables such as temperature, relative humidity, wind speed and direction. Feature selection methods — Filter (Information Gain, F-Test, Correlation Coefficient), Wrapper (Recursive Feature Elimination, Sequential Forward/Backward Selection), and Embedded (LASSO, Elastic Net, Ridge Regression, RF Importance) — were applied to identify significant predictors, with embedded methods providing the best balance of simplicity, accuracy, and cost-efficiency. Among the models, RF demonstrated the highest seasonal performance (R² = 0.75) during summer. RF's prediction R² values for PM₁₀ remained above 0.5 in all seasons, consistently outperformed the other models. The WKNN model performed reasonably well across all seasons, ranking second among the models, while the LASSO model demonstrated weaker performance. The SVR model showed satisfactory performance in specific seasons, such as summer and autumn. A common feature of all models was their better performance during summer. Importantly, the models relied solely on readily available meteorological data, enabling accurate predictions of PM₁₀ in this arid region of eastern Iran. The findings highlight the potential of ML techniques for developing air pollution prediction and warning systems, offering valuable support to policymakers in the design of effective pollution control strategies and safeguarding public health.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 5","pages":"Article 102456"},"PeriodicalIF":3.9,"publicationDate":"2025-02-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143479230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"环境科学与生态学","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}