Atmospheric Pollution Research最新文献_第8页

Using a chemical transport model and satellite measurements to assess ship-induced NO2 concentrations in the Barents and Kara Sea region 利用化学输送模型和卫星测量评估巴伦支海和喀拉海地区船舶引起的NO2浓度

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-12 DOI: 10.1016/j.apr.2025.102745

Nikolai Figenschau, Jinmei Lu

In the past decade, ship traffic and the associated atmospheric emissions of nitrogen oxides (NO_x) have significantly increased in the Barents and Kara Sea region. Despite the rapid growth in ship traffic and NO_x emissions, research on the impact of these developments on regional air quality in the data-sparse Arctic remains limited. This study assesses the impact of recent changes in ship traffic on regional atmospheric nitrogen dioxide (NO₂) concentrations for the years 2015, 2019, and 2021, utilizing the European Monitoring and Evaluation Programme's Meteorological Synthesizing Center – West (EMEP) chemical transport model (CTM). The feasibility of using Copernicus Sentinel-5P Tropospheric Monitoring Instrument (TROPOMI) NO₂ tropospheric vertical column densities (VCDs), adjusted with EMEP CTM vertical profiles as an independent reference for model evaluation in 2019 and 2021. The study reveals a continuous increase in ship-induced NO₂ surface concentrations from 2015 to 2021. The contribution of the shipping sector is substantial (40–90 %), particularly during summer and in areas with dense shipping activities, such as the Gulf of Ob, where shipping has shifted from having a seasonal to a year-round impact. The comparison between modelled NO₂ concentrations and TROPOMI NO₂ VCDs showed weak to moderate positive correlations, with stronger relationships observed in specific subareas with high ship traffic, like the Gulf of Ob. This demonstrates the potential of combining satellite data with CTMs to assess the impact of ship traffic on regional air quality in high latitudes. Ultimately, this study highlights the significant impact the shipping sector already has on regional air quality, underscoring the need for continued monitoring and assessment of ongoing human activities and their future impact on the Arctic environment.

在过去十年中，巴伦支和喀拉海地区的船舶交通和相关的大气氮氧化物（NOx）排放量显著增加。尽管船舶交通和氮氧化物排放迅速增长，但在数据稀少的北极地区，关于这些发展对区域空气质量影响的研究仍然有限。本研究利用欧洲监测和评估计划的气象综合中心-西部（EMEP）化学输送模型（CTM），评估了2015年、2019年和2021年船舶交通的近期变化对区域大气二氧化氮（NO2）浓度的影响。2019年和2021年利用哥白尼哨兵- 5p对流层监测仪器（TROPOMI）对流层NO2垂直柱密度（vcd）作为独立参考进行模式评估的可行性，并与EMEP CTM垂直剖面进行调整。该研究显示，从2015年到2021年，船舶引起的二氧化氮表面浓度持续增加。航运业的贡献是巨大的（40 - 90%），特别是在夏季和航运活动密集的地区，如鄂毕湾，在那里航运已经从季节性影响转变为全年影响。模拟的二氧化氮浓度与TROPOMI二氧化氮vcd之间的比较显示出弱至中度的正相关，在船舶交通量大的特定子区域（如Ob湾）观察到更强的关系。这表明将卫星数据与CTMs结合起来评估船舶交通量对高纬度地区区域空气质量的影响的潜力。最后，本研究强调了航运业已经对区域空气质量产生的重大影响，强调需要继续监测和评估正在进行的人类活动及其未来对北极环境的影响。

{"title":"Using a chemical transport model and satellite measurements to assess ship-induced NO2 concentrations in the Barents and Kara Sea region","authors":"Nikolai Figenschau, Jinmei Lu","doi":"10.1016/j.apr.2025.102745","DOIUrl":"10.1016/j.apr.2025.102745","url":null,"abstract":"<div><div>In the past decade, ship traffic and the associated atmospheric emissions of nitrogen oxides (NO<sub>x</sub>) have significantly increased in the Barents and Kara Sea region. Despite the rapid growth in ship traffic and NO<sub>x</sub> emissions, research on the impact of these developments on regional air quality in the data-sparse Arctic remains limited. This study assesses the impact of recent changes in ship traffic on regional atmospheric nitrogen dioxide (NO<sub>2</sub>) concentrations for the years 2015, 2019, and 2021, utilizing the European Monitoring and Evaluation Programme's Meteorological Synthesizing Center – West (EMEP) chemical transport model (CTM). The feasibility of using Copernicus Sentinel-5P Tropospheric Monitoring Instrument (TROPOMI) NO<sub>2</sub> tropospheric vertical column densities (VCDs), adjusted with EMEP CTM vertical profiles as an independent reference for model evaluation in 2019 and 2021. The study reveals a continuous increase in ship-induced NO<sub>2</sub> surface concentrations from 2015 to 2021. The contribution of the shipping sector is substantial (40–90 %), particularly during summer and in areas with dense shipping activities, such as the Gulf of Ob, where shipping has shifted from having a seasonal to a year-round impact. The comparison between modelled NO<sub>2</sub> concentrations and TROPOMI NO<sub>2</sub> VCDs showed weak to moderate positive correlations, with stronger relationships observed in specific subareas with high ship traffic, like the Gulf of Ob. This demonstrates the potential of combining satellite data with CTMs to assess the impact of ship traffic on regional air quality in high latitudes. Ultimately, this study highlights the significant impact the shipping sector already has on regional air quality, underscoring the need for continued monitoring and assessment of ongoing human activities and their future impact on the Arctic environment.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102745"},"PeriodicalIF":3.5,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145963308","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}

引用次数: 0

First long-term air quality assessment in Luanda, Angola: Performance evaluation of a low-cost monitoring station against reference equipment 安哥拉罗安达首次长期空气质量评估：对照参考设备对一个低成本监测站的性能进行评估

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-12 DOI: 10.1016/j.apr.2025.102746

Alan Victor da Silva , Leonardo Furst , Yago A. Cipoli , Marlene J.S. Soares , Anabela G.A. Leitão , Manuel Feliciano , Célia A. Alves

Low-cost air quality monitoring stations (LCMS), which integrate sensors for gases and particulate matter (PM), offer an economical solution for expanding monitoring networks. However, their reliability requires validation, particularly in rapidly urbanising regions with limited infrastructure. This study presents the first long-term, continuous, multi-pollutant air quality assessment in Luanda, Angola - where no monitoring stations currently exist - by evaluating the performance of an LCMS against reference-grade equipment. Daily averages, correlation metrics (R², RMSE), and a hybrid Bland-Altman/regression analyses were used to evaluate the agreement. Results indicated strong correlation for CO (R² = 0.96; RMSE = 0.24 ppm) and good for NO₂ (R² = 0.81; RMSE = 6.35 ppb), although limitations near detection limits were noted. Significant challenges were identified in O₃ measurements (R² = 0.77, RMSE = 7.13 ppb), primarily due to strong cross-sensitivity to high ambient NO₂ levels and potential sensor ageing. For PM₁₀ and PM_2.5, although good linear correlations (R² ∼ 0.82) were observed with reference methods, the LCMS exhibited considerable systematic bias (RMSE over 46 μg/m⁻³) and consistently underestimate concentrations. The study also registered frequent and severe exceedances of WHO AQG and EU standards for PM₁₀, PM_2.5, and NO₂, underscoring significant public health risks. Despite limitations, particularly for O₃ measurements and biases in PM data, the LCMS demonstrates potential as a cost-effective tool to complement reference networks, enhance spatial monitoring coverage, identify pollution hotspots, and support air quality management in resource-constrained settings, since continuous calibration and validation procedures are implemented to mitigate measurement uncertainties.

低成本的空气质量监测站（LCMS）集成了气体和颗粒物（PM）传感器，为扩大监测网络提供了一种经济的解决方案。然而，它们的可靠性需要验证，特别是在基础设施有限的快速城市化地区。本研究通过评估LCMS与参考级设备的性能，首次对安哥拉罗安达（目前没有监测站）进行了长期、连续、多污染物的空气质量评估。使用日平均值、相关指标（R2、RMSE）和混合Bland-Altman/回归分析来评估一致性。结果表明，CO （R2 = 0.96, RMSE = 0.24 ppm）和NO2 （R2 = 0.81, RMSE = 6.35 ppb）具有很强的相关性，尽管注意到接近检出限的局限性。在O3测量中发现了重大挑战（R2 = 0.77, RMSE = 7.13 ppb），主要是由于对高环境NO2水平和潜在传感器老化的强交叉敏感性。对于PM10和PM2.5，尽管用参考方法观察到良好的线性相关性（R2 ~ 0.82），但LCMS表现出相当大的系统偏差（RMSE超过46 μg/m−3），并且一直低估浓度。该研究还记录了世卫组织空气质量标准和欧盟PM10、PM2.5和二氧化氮标准的频繁和严重超标，强调了重大的公共卫生风险。尽管存在局限性，特别是对于O3测量和PM数据的偏差，但由于实施了连续校准和验证程序以减轻测量不确定性，LCMS显示出作为一种具有成本效益的工具的潜力，可以补充参考网络，扩大空间监测覆盖范围，识别污染热点，并支持资源受限环境下的空气质量管理。

{"title":"First long-term air quality assessment in Luanda, Angola: Performance evaluation of a low-cost monitoring station against reference equipment","authors":"Alan Victor da Silva , Leonardo Furst , Yago A. Cipoli , Marlene J.S. Soares , Anabela G.A. Leitão , Manuel Feliciano , Célia A. Alves","doi":"10.1016/j.apr.2025.102746","DOIUrl":"10.1016/j.apr.2025.102746","url":null,"abstract":"<div><div>Low-cost air quality monitoring stations (LCMS), which integrate sensors for gases and particulate matter (PM), offer an economical solution for expanding monitoring networks. However, their reliability requires validation, particularly in rapidly urbanising regions with limited infrastructure. This study presents the first long-term, continuous, multi-pollutant air quality assessment in Luanda, Angola - where no monitoring stations currently exist - by evaluating the performance of an LCMS against reference-grade equipment. Daily averages, correlation metrics (R<sup>2</sup>, RMSE), and a hybrid Bland-Altman/regression analyses were used to evaluate the agreement. Results indicated strong correlation for CO (R<sup>2</sup> = 0.96; RMSE = 0.24 ppm) and good for NO<sub>2</sub> (R<sup>2</sup> = 0.81; RMSE = 6.35 ppb), although limitations near detection limits were noted. Significant challenges were identified in O<sub>3</sub> measurements (R<sup>2</sup> = 0.77, RMSE = 7.13 ppb), primarily due to strong cross-sensitivity to high ambient NO<sub>2</sub> levels and potential sensor ageing. For PM<sub>10</sub> and PM<sub>2.5</sub>, although good linear correlations (R<sup>2</sup> ∼ 0.82) were observed with reference methods, the LCMS exhibited considerable systematic bias (RMSE over 46 μg/m<sup>−3</sup>) and consistently underestimate concentrations. The study also registered frequent and severe exceedances of WHO AQG and EU standards for PM<sub>10</sub>, PM<sub>2.5</sub>, and NO<sub>2</sub>, underscoring significant public health risks. Despite limitations, particularly for O<sub>3</sub> measurements and biases in PM data, the LCMS demonstrates potential as a cost-effective tool to complement reference networks, enhance spatial monitoring coverage, identify pollution hotspots, and support air quality management in resource-constrained settings, since continuous calibration and validation procedures are implemented to mitigate measurement uncertainties.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102746"},"PeriodicalIF":3.5,"publicationDate":"2025-09-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145963400","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}

引用次数: 0

A proposal of indoor air pollutant limit values for Turkish schools based on a literature review of emission sources, concentrations, health effects, and limits/guidelines 根据对排放源、浓度、健康影响和限值/准则的文献审查，提出土耳其学校室内空气污染物极限值的建议

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-11 DOI: 10.1016/j.apr.2025.102743

Sait Cemil Sofuoglu , Akif Ari , Mihriban Civan , Yetkin Dumanoglu , Gulen Gullu , Sibel Mentese , Burcu Onat , Ülkü Alver Şahin , Macit Toksoy

Limit Values Working Group (LVWG) was established under Indoor Air Quality Committee of Turkish Climatization Assembly of the Union of Chambers and Commodity Exchanges of Türkiye. LVWG was tasked with reviewing the pertinent literature on indoor emission sources, concentrations in schools, health effects, and existing limit and guideline values to identify the indoor air pollutants that need to be addressed and to be recommended a limit value for Turkish schools. LVWG members took responsibilities based on their individual expertise. The recommendations were concluded in consensus decision-making after in-group discussions. A total of 19 pollutants/pollutant groups (carbon dioxide, carbon monoxide, nitrogen dioxide, ozone, radon, volatile organic compounds, formaldehyde, trihalomethanes, polycyclic aromatic hydrocarbons, polychlorinated biphenyls, brominated flame retardants, organophosphate esters, phthalate esters, particulate matter, bioaerosols (bacteria, fungi, viruses), microbial pollutants and allergens) were reviewed. Limit values were recommended for 11 pollutants/groups based on the current knowledge, i.e. pollutant health effects and indoor air concentrations taking into account the exposure duration, the prevalence of existing limit/guideline values and the health effects on which they are based.

限值工作组（LVWG）是在土耳其商会和商品交易所联盟土耳其气候大会室内空气质量委员会下成立的。工作小组的任务是审查有关室内排放源、学校浓度、健康影响以及现有限值和指导值的相关文献，以确定需要解决的室内空气污染物，并为土耳其学校建议一个限值。LVWG成员根据各自的专业知识承担责任。这些建议是在小组讨论后以协商一致的决策方式得出的。综述了19种污染物/污染物组（二氧化碳、一氧化碳、二氧化氮、臭氧、氡、挥发性有机物、甲醛、三卤甲烷、多环芳烃、多氯联苯、溴化阻燃剂、有机磷酸酯、邻苯二甲酸酯、颗粒物、生物气溶胶（细菌、真菌、病毒）、微生物污染物和过敏原）。根据目前的知识，即污染物对健康的影响和室内空气浓度，建议了11种污染物/组的极限值，同时考虑到接触时间、现有极限值/指标值的普遍程度以及它们所依据的健康影响。

{"title":"A proposal of indoor air pollutant limit values for Turkish schools based on a literature review of emission sources, concentrations, health effects, and limits/guidelines","authors":"Sait Cemil Sofuoglu , Akif Ari , Mihriban Civan , Yetkin Dumanoglu , Gulen Gullu , Sibel Mentese , Burcu Onat , Ülkü Alver Şahin , Macit Toksoy","doi":"10.1016/j.apr.2025.102743","DOIUrl":"10.1016/j.apr.2025.102743","url":null,"abstract":"<div><div>Limit Values Working Group (LVWG) was established under Indoor Air Quality Committee of Turkish Climatization Assembly of the Union of Chambers and Commodity Exchanges of Türkiye. LVWG was tasked with reviewing the pertinent literature on indoor emission sources, concentrations in schools, health effects, and existing limit and guideline values to identify the indoor air pollutants that need to be addressed and to be recommended a limit value for Turkish schools. LVWG members took responsibilities based on their individual expertise. The recommendations were concluded in consensus decision-making after in-group discussions. A total of 19 pollutants/pollutant groups (carbon dioxide, carbon monoxide, nitrogen dioxide, ozone, radon, volatile organic compounds, formaldehyde, trihalomethanes, polycyclic aromatic hydrocarbons, polychlorinated biphenyls, brominated flame retardants, organophosphate esters, phthalate esters, particulate matter, bioaerosols (bacteria, fungi, viruses), microbial pollutants and allergens) were reviewed. Limit values were recommended for 11 pollutants/groups based on the current knowledge, i.e. pollutant health effects and indoor air concentrations taking into account the exposure duration, the prevalence of existing limit/guideline values and the health effects on which they are based.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102743"},"PeriodicalIF":3.5,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962826","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}

引用次数: 0

Understanding the evolution of municipal solid waste incineration emissions in China: Patterns, determinants, and control measures 了解中国城市生活垃圾焚烧排放的演变：模式、决定因素和控制措施

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-11 DOI: 10.1016/j.apr.2025.102741

Abdoulaye Boré , Jicui Cui , Guanyi Chen , Nickolas J. Themelis , Wenchao Ma

Municipal solid waste (MSW) incineration has become central to China's waste management strategy, driven by rapid urbanization, environmental constraints, and energy recovery goals. This study evaluates the evolution of MSW incineration emissions between 2005 and 2020 across 510 waste-to-energy (WTE) facilities nationwide, using a multi-pollutant framework involving nine major emissions. These pollutants include NO_x, SO₂, PM, HCl, CO, dioxins, Hg, Cd + Tl, and Sb + As + Pb + Cr + Co + Cu + Mn + Ni. The Logarithmic Mean Divisia Index (LMDI) was used to explore the impact of socioeconomic factors, treatment structures, and emission intensities. The results reveal a multi-phase, policy-driven emission trajectory shaped by technological transformation, regulatory milestones, and regional inequalities. Three policy phases were identified, corresponding to national Five-Year Plans (FYPs). During the 13^th FYP (2016–2020), emission intensity (EI) reductions accounted for over 80 % of net SO₂ and PM emission declines, and more than 60 % for dioxins and HCl, driven by advanced flue gas treatment and moving grate (MG) incinerator adoption. Despite national improvements, regional disparities persist. Larger incinerators (≥1000 t/d) showed 40–60 % lower emissions than smaller ones (≤300 t/d), while NO_x concentrations remained high across all plant sizes. This research highlights the critical role of policy and technology in driving emission reductions, while also exposing governance and infrastructure gaps that undermine progress.

在快速城市化、环境约束和能源回收目标的推动下，城市固体废物（MSW）焚烧已成为中国废物管理战略的核心。本研究采用涉及九种主要排放的多污染物框架，评估了2005年至2020年间全国510个垃圾焚烧能源（WTE）设施的城市生活垃圾焚烧排放演变。这些污染物包括NOx、SO2、PM、HCl、CO、二恶英、Hg、Cd + Tl和Sb + As + Pb + Cr + CO + Cu + Mn + Ni。采用对数平均分裂指数（LMDI）探讨社会经济因素、治理结构和排放强度的影响。研究结果揭示了一个由技术转型、监管里程碑和区域不平等形成的多阶段、政策驱动的排放轨迹。根据国家五年计划确定了三个政策阶段。在“十三五”规划（2016-2020年）期间，由于采用先进的烟气处理和移动炉排（MG）焚烧炉，排放强度（EI）的减少占二氧化硫和PM净排放量减少的80%以上，二恶英和HCl的减少超过60%。尽管各国有所改善，但地区差异依然存在。大型焚烧炉（≥1000t /d）的排放量比小型焚烧炉（≤300t /d）低40 - 60%，而氮氧化物浓度在所有工厂规模中都保持较高水平。这项研究强调了政策和技术在推动减排方面的关键作用，同时也暴露了阻碍进展的治理和基础设施差距。

{"title":"Understanding the evolution of municipal solid waste incineration emissions in China: Patterns, determinants, and control measures","authors":"Abdoulaye Boré , Jicui Cui , Guanyi Chen , Nickolas J. Themelis , Wenchao Ma","doi":"10.1016/j.apr.2025.102741","DOIUrl":"10.1016/j.apr.2025.102741","url":null,"abstract":"<div><div>Municipal solid waste (MSW) incineration has become central to China's waste management strategy, driven by rapid urbanization, environmental constraints, and energy recovery goals. This study evaluates the evolution of MSW incineration emissions between 2005 and 2020 across 510 waste-to-energy (WTE) facilities nationwide, using a multi-pollutant framework involving nine major emissions. These pollutants include NO<sub>x</sub>, SO<sub>2</sub>, PM, HCl, CO, dioxins, Hg, Cd + Tl, and Sb + As + Pb + Cr + Co + Cu + Mn + Ni. The Logarithmic Mean Divisia Index (LMDI) was used to explore the impact of socioeconomic factors, treatment structures, and emission intensities. The results reveal a multi-phase, policy-driven emission trajectory shaped by technological transformation, regulatory milestones, and regional inequalities. Three policy phases were identified, corresponding to national Five-Year Plans (FYPs). During the 13<sup>th</sup> FYP (2016–2020), emission intensity (EI) reductions accounted for over 80 % of net SO<sub>2</sub> and PM emission declines, and more than 60 % for dioxins and HCl, driven by advanced flue gas treatment and moving grate (MG) incinerator adoption. Despite national improvements, regional disparities persist. Larger incinerators (≥1000 t/d) showed 40–60 % lower emissions than smaller ones (≤300 t/d), while NO<sub>x</sub> concentrations remained high across all plant sizes. This research highlights the critical role of policy and technology in driving emission reductions, while also exposing governance and infrastructure gaps that undermine progress.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102741"},"PeriodicalIF":3.5,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145963296","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}

引用次数: 0

Spatiotemporal heterogeneity assessment of provincial carbon emissions in China based on a dynamically weighted ensemble model 基于动态加权集合模型的中国省际碳排放时空异质性评价

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-11 DOI: 10.1016/j.apr.2025.102742

Wenjing Zhu , Quanfeng Wang , Bin Liu , Xiaoyu Zhang , Qinxiang Wang , Yue Li

Amid intensifying climate change, monitoring carbon emissions is vital for regional sustainable development. To address pronounced spatiotemporal heterogeneity in China’s provincial emissions (2000–2022), we propose an ensemble framework integrating Geographically and Temporally Weighted Regression (GTWR), Geographical Optimal Similarity (GOS), and Random Forest (RF). Beyond geographic proximity, we build a multi-dimensional spatiotemporal weight tensor from geographic distance and annually varying economic–energy–social Manhattan distances. This tensor feeds a Spatiotemporal Weighted Ensemble Neural Network (STWENN), which, via backpropagation, adaptively allocates base-learner weights across space and time. Predictions are fused by combining these adaptive weights with annually estimated global weights from Ordinary Least Squares (OLS). Results show: (1) higher predictive precision (test R² = 0.8217, where R² represents the goodness-of-fit of the model) than spatially homogeneous benchmarks and accurate characterization of high-emission areas; (2) spatial linkages are shifting from geographic to functional associations driven by supply chains and energy flows; the learned weight dynamics align with this transition, supporting interpretability; (3) Bootstrap-based uncertainty analysis indicates robust trend prediction with sensitivity to external shocks. Overall, the study offers a modeling pathway that balances accuracy and interpretability, supplies evidence on heterogeneous regional drivers and spillovers, and informs regionally differentiated low-carbon and collaborative governance policies.

在气候变化加剧的背景下，碳排放监测对区域可持续发展至关重要。为了解决中国各省排放（2000-2022）明显的时空异质性，我们提出了一个整合地理和时间加权回归（GTWR）、地理最优相似性（GOS）和随机森林（RF）的集成框架。除了地理邻近，我们从地理距离和每年变化的经济-能源-社会曼哈顿距离建立了多维时空权重张量。该张量为时空加权集成神经网络（STWENN）提供信息，该网络通过反向传播自适应地在空间和时间上分配基础学习器权重。通过将这些自适应权重与普通最小二乘（OLS）每年估计的全球权重相结合，将预测融合在一起。结果表明：(1)预测精度（检验R2 = 0.8217， R2代表模型拟合优度）高于空间均匀基准和高排放区准确表征；(2)空间联系由地理联系向供应链和能量流驱动的功能联系转变；习得的权重动态与这种转变保持一致，支持可解释性；(3)基于bootstrap的不确定性分析具有对外部冲击敏感的鲁棒性趋势预测。总体而言，本研究提供了一种平衡准确性和可解释性的建模途径，提供了异质性区域驱动因素和溢出效应的证据，并为区域差异化的低碳和协作治理政策提供了信息。

{"title":"Spatiotemporal heterogeneity assessment of provincial carbon emissions in China based on a dynamically weighted ensemble model","authors":"Wenjing Zhu , Quanfeng Wang , Bin Liu , Xiaoyu Zhang , Qinxiang Wang , Yue Li","doi":"10.1016/j.apr.2025.102742","DOIUrl":"10.1016/j.apr.2025.102742","url":null,"abstract":"<div><div>Amid intensifying climate change, monitoring carbon emissions is vital for regional sustainable development. To address pronounced spatiotemporal heterogeneity in China’s provincial emissions (2000–2022), we propose an ensemble framework integrating Geographically and Temporally Weighted Regression (GTWR), Geographical Optimal Similarity (GOS), and Random Forest (RF). Beyond geographic proximity, we build a multi-dimensional spatiotemporal weight tensor from geographic distance and annually varying economic–energy–social Manhattan distances. This tensor feeds a Spatiotemporal Weighted Ensemble Neural Network (STWENN), which, via backpropagation, adaptively allocates base-learner weights across space and time. Predictions are fused by combining these adaptive weights with annually estimated global weights from Ordinary Least Squares (OLS). Results show: (1) higher predictive precision (test R<sup>2</sup> = 0.8217, where R<sup>2</sup> represents the goodness-of-fit of the model) than spatially homogeneous benchmarks and accurate characterization of high-emission areas; (2) spatial linkages are shifting from geographic to functional associations driven by supply chains and energy flows; the learned weight dynamics align with this transition, supporting interpretability; (3) Bootstrap-based uncertainty analysis indicates robust trend prediction with sensitivity to external shocks. Overall, the study offers a modeling pathway that balances accuracy and interpretability, supplies evidence on heterogeneous regional drivers and spillovers, and informs regionally differentiated low-carbon and collaborative governance policies.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102742"},"PeriodicalIF":3.5,"publicationDate":"2025-09-11","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145963303","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}

引用次数: 0

Quantitative analysis to define baseline criteria for introducing reduced-emission firecrackers 定量分析以厘定引入减放鞭炮的基准准则

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-10 DOI: 10.1016/j.apr.2025.102740

Shilpa Kumari , Rahul Wadichar , Payal Mane , Sadhana Rayalu , Penumaka Nagababu

This research paper emphasizes the crucial role of statistical methods in validating the proposed methodology for the state-of-the-art emission testing facility. This facility is specifically designed for monitoring emissions from developed reduced-emission firecrackers and commercial crackers. Establishing baseline values, derived through statistical analysis of data collected by the emission testing facility, is pivotal in ensuring the production of less polluting firecrackers by the fireworks industry. This, in turn, supports sustainable festival celebrations and events in the future. Statistical techniques such as frequency distribution, regression equations, and the Spearman correlation coefficient were employed to understand the significance of the data and its distribution for calculating standard error and deviation. The baseline values, identified through this statistical analysis, serve as crucial parameters in the evaluation of emission levels. According to the study's findings, the P-value indicates a significant result at P < 0.05. Furthermore, the correlation coefficient between PM₁₀ and PM_2.5 is reported to have an R² value of 0.99, highlighting a strong correlation. This robust statistical foundation strengthens the credibility of the proposed methodology and underscores its importance in advancing the development and monitoring of environmentally sustainable firecrackers.

这篇研究论文强调了统计方法在验证最先进的排放测试设施所提出的方法中的关键作用。该设施是专门为监测开发的减排鞭炮和商用鞭炮的排放而设计的。通过对排放测试设施收集的数据进行统计分析得出的基准值，对于确保烟花工业生产污染较低的鞭炮至关重要。这反过来又支持了未来可持续的节日庆祝活动。使用频率分布、回归方程和Spearman相关系数等统计技术来了解数据的重要性及其分布，以计算标准误差和偏差。通过这一统计分析确定的基线值是评价排放水平的关键参数。根据研究结果，P值表示P <； 0.05的显著结果。此外，PM10与PM2.5的相关系数R2值为0.99，显示出很强的相关性。这一坚实的统计基础加强了所提议方法的可信度，并强调了其在促进开发和监测环境可持续鞭炮方面的重要性。

{"title":"Quantitative analysis to define baseline criteria for introducing reduced-emission firecrackers","authors":"Shilpa Kumari , Rahul Wadichar , Payal Mane , Sadhana Rayalu , Penumaka Nagababu","doi":"10.1016/j.apr.2025.102740","DOIUrl":"10.1016/j.apr.2025.102740","url":null,"abstract":"<div><div>This research paper emphasizes the crucial role of statistical methods in validating the proposed methodology for the state-of-the-art emission testing facility. This facility is specifically designed for monitoring emissions from developed reduced-emission firecrackers and commercial crackers. Establishing baseline values, derived through statistical analysis of data collected by the emission testing facility, is pivotal in ensuring the production of less polluting firecrackers by the fireworks industry. This, in turn, supports sustainable festival celebrations and events in the future. Statistical techniques such as frequency distribution, regression equations, and the Spearman correlation coefficient were employed to understand the significance of the data and its distribution for calculating standard error and deviation. The baseline values, identified through this statistical analysis, serve as crucial parameters in the evaluation of emission levels. According to the study's findings, the P-value indicates a significant result at P < 0.05. Furthermore, the correlation coefficient between PM<sub>10</sub> and PM<sub>2.5</sub> is reported to have an R<sup>2</sup> value of 0.99, highlighting a strong correlation. This robust statistical foundation strengthens the credibility of the proposed methodology and underscores its importance in advancing the development and monitoring of environmentally sustainable firecrackers.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 12","pages":"Article 102740"},"PeriodicalIF":3.5,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145044271","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}

引用次数: 0

Comparison of PM2.5 trends and source factors in urban and rural locations in Bangladesh 孟加拉国城市和农村地区PM2.5趋势和来源因素的比较

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-10 DOI: 10.1016/j.apr.2025.102744

Shahid Uz Zaman , Shatabdi Roy , Benjamin de Foy , Rakibul Omar , Md Abdullah Al-Amin , Prakash Bhave , Michael Howard Bergin , James Jay Schauer , Abdus Salam

Air pollution remains a critical environmental and public health concern, particularly in developing countries like Bangladesh. Although urban air pollution has received significant attention, rural areas also experience high PM_2.5 concentrations. In this study, a low-cost sensor (LCS) network was deployed across five locations: Dhaka, Rajshahi, Panchagarh, Netrokona, and Bhola from April 2022 to September 2023 to assess local and regional sources of PM_2.5. A Generalized Additive Model (GAM) was applied to analyze the influence of meteorology and source contributions on observed PM_2.5 concentrations. The highest PM_2.5 levels were recorded in Netrokona (212.81 ± 64.5 μgm⁻³), followed by Panchagarh (128.6 ± 71.7 μgm⁻³), Rajshahi (110.4 ± 48.5 μgm⁻³), Dhaka (105.1 ± 55.7 μgm⁻³) and Bhola (82.2 ± 36.2 μgm⁻³). A consistent diurnal pattern was observed across all sites, characterized by two peaks in the morning and evening. GAM analysis revealed that the boundary layer height had the lowest influence in Bhola and Panchagarh, while Dhaka exhibited the highest contribution. The contribution of long-range transport was found uniform at all the sites. The Trajectory Cluster Concentration Impact (TCCI) showed that the Indo-Gangetic Plain (IGP) is responsible for the enhancement of 50 μgm⁻³ at all the sites. However, wind transported from the Bay of Bengal associates PM_2.5 reduction of 20–40 μgm⁻³. Impacts of local winds on the PM_2.5 concentrations in the GAM simulations suggested that winds from the northwest are associated with higher PM_2.5. These findings emphasize the need for comprehensive air quality management strategies that extend beyond major urban centers to include rural and semi-urban areas.

空气污染仍然是一个严重的环境和公共卫生问题，特别是在孟加拉国等发展中国家。尽管城市空气污染受到了广泛关注，但农村地区的PM2.5浓度也很高。在这项研究中，从2022年4月至2023年9月，在达卡、拉杰沙希、潘查加尔、奈特罗科纳和博拉五个地点部署了一个低成本传感器（LCS）网络，以评估当地和区域的PM2.5来源。采用广义加性模型（GAM）分析了气象和源对PM2.5浓度的影响。PM2.5浓度最高的城市为Netrokona(212.81±64.5 μgm−3)，其次为Panchagarh（128.6±71.7 μgm−3）、Rajshahi（110.4±48.5 μgm−3）、Dhaka（105.1±55.7 μgm−3）和Bhola（82.2±36.2 μgm−3）。在所有地点观察到一致的日模式，其特征是早上和晚上两个高峰。GAM分析显示，边界层高度对Bhola和Panchagarh的影响最小，而达卡的贡献最大。远距离输运的贡献在所有地点都是一致的。轨迹聚类浓度影响（TCCI）表明，印度-恒河平原（IGP）对所有位点的50 μgm−3的增强负责。而来自孟加拉湾的风导致PM2.5减少20 ~ 40 μgm−3。局地风对GAM模拟中PM2.5浓度的影响表明，西北风与较高的PM2.5有关。这些研究结果强调需要制定全面的空气质量管理战略，将范围扩大到主要城市中心以外，包括农村和半城市地区。

{"title":"Comparison of PM2.5 trends and source factors in urban and rural locations in Bangladesh","authors":"Shahid Uz Zaman , Shatabdi Roy , Benjamin de Foy , Rakibul Omar , Md Abdullah Al-Amin , Prakash Bhave , Michael Howard Bergin , James Jay Schauer , Abdus Salam","doi":"10.1016/j.apr.2025.102744","DOIUrl":"10.1016/j.apr.2025.102744","url":null,"abstract":"<div><div>Air pollution remains a critical environmental and public health concern, particularly in developing countries like Bangladesh. Although urban air pollution has received significant attention, rural areas also experience high PM<sub>2.5</sub> concentrations. In this study, a low-cost sensor (LCS) network was deployed across five locations: Dhaka, Rajshahi, Panchagarh, Netrokona, and Bhola from April 2022 to September 2023 to assess local and regional sources of PM<sub>2.5</sub>. A Generalized Additive Model (GAM) was applied to analyze the influence of meteorology and source contributions on observed PM<sub>2.5</sub> concentrations. The highest PM<sub>2.5</sub> levels were recorded in Netrokona (212.81 ± 64.5 μgm<sup>−3</sup>), followed by Panchagarh (128.6 ± 71.7 μgm<sup>−3</sup>), Rajshahi (110.4 ± 48.5 μgm<sup>−3</sup>), Dhaka (105.1 ± 55.7 μgm<sup>−3</sup>) and Bhola (82.2 ± 36.2 μgm<sup>−3</sup>). A consistent diurnal pattern was observed across all sites, characterized by two peaks in the morning and evening. GAM analysis revealed that the boundary layer height had the lowest influence in Bhola and Panchagarh, while Dhaka exhibited the highest contribution. The contribution of long-range transport was found uniform at all the sites. The Trajectory Cluster Concentration Impact (TCCI) showed that the Indo-Gangetic Plain (IGP) is responsible for the enhancement of 50 μgm<sup>−3</sup> at all the sites. However, wind transported from the Bay of Bengal associates PM<sub>2.5</sub> reduction of 20–40 μgm<sup>−3</sup>. Impacts of local winds on the PM<sub>2.5</sub> concentrations in the GAM simulations suggested that winds from the northwest are associated with higher PM<sub>2.5</sub>. These findings emphasize the need for comprehensive air quality management strategies that extend beyond major urban centers to include rural and semi-urban areas.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102744"},"PeriodicalIF":3.5,"publicationDate":"2025-09-10","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145963304","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}

引用次数: 0

Evaluating magnetic properties of atmospheric particulate matter (PM10) at Brasília bus station, central Brazil 巴西中部Brasília公交车站大气颗粒物（PM10）磁性评价

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-09 DOI: 10.1016/j.apr.2025.102739

Poliana Dutra Maia , Elder Yokoyama , Luiza de Souza Romano , Guilherme Gomide , Jerome Depeyrot , Sebastião William da Silva , Myller de Sousa Tonhá

This study analyzes the Saturation Magnetization (M_S) in PM₁₀ particles collected at a central bus station in Brasília from 2018 to 2019, exploring its relationship with chemical elements (total metals - Fe, V, Cu, Mn, Sr, Cd - and ions - Ca²⁺, Mg²⁺, Na⁺, K⁺, Cl⁻, NO₃⁻, SO₄²⁻) and meteorological variables (relative humidity and global radiation). M_S was measured using a S700X SQUID magnetometer and modeled with the Langevin equation, yielding values between 0.114 and 1.407 Am²/kg (n = 13), with a standard deviation of 0.01–0.1. Statistical analysis revealed similar median values between M_S and the studied variables, indicating shared sources of PM₁₀. The higher total concentrations of metals (Fe > Sr > Mn) and ions (Ca²⁺ = 30 %, Na⁺ = 23 % of total ions) suggest a natural origin of PM₁₀, likely from soil minerals. In contrast, metals (Cu, V, Cd, Ni, Co and Zn) and ions (SO₄²⁻, NO₃⁻) of anthropogenic origin were associated with vehicle emissions and road dust. FTIR spectroscopy identified magnetite as the dominant magnetic carrier in PM₁₀, with absorption peaks at 570 and 390 cm⁻¹, and particle sizes ranging from 8.5 to 11 nm, although a broader size range (8–33 nm) was also observed. This supports the formation of heterogeneous particles through mineral-metal or carbonaceous adsorption, as confirmed by morphological data. The temporal distribution of M_S in PM₁₀ samples showed peak values during the rainy season, despite no direct correlation with the studied variables, suggesting that iron solubility in minerals may influence the saturation magnetization.

本研究分析了2018 - 2019年Brasília某公交车站PM10颗粒的饱和磁化强度（MS），探讨了其与化学元素（总金属Fe、V、Cu、Mn、Sr、Cd）和离子（Ca2+、Mg2+、Na+、K+、Cl−、NO3−、SO42−）以及气象变量（相对湿度和全球辐射）的关系。用S700X SQUID磁强计测量质谱，用Langevin方程建模，所得值为0.114 ~ 1.407 Am2/kg (n = 13)，标准差为0.01 ~ 0.1。统计分析显示MS和研究变量之间的中位数相似，表明PM10的来源相同。较高的金属（Fe > Sr > Mn）和离子（Ca2+ = 30%, Na+ = 23%的总离子）的总浓度表明PM10的自然来源，可能来自土壤矿物质。相反，人为来源的金属（Cu、V、Cd、Ni、Co和Zn）和离子（SO42−、NO3−）与车辆排放和道路粉尘有关。FTIR光谱分析发现，PM10中的主要磁性载体为磁铁矿，吸收峰位于570和390 cm−1，粒径范围为8.5 ~ 11 nm，但也观察到更宽的粒径范围（8 ~ 33 nm）。这支持通过矿物-金属或碳质吸附形成的非均质颗粒，正如形态学数据所证实的那样。PM10样品中MS的时间分布在雨季达到峰值，尽管与所研究的变量没有直接相关，这表明铁在矿物中的溶解度可能影响饱和磁化强度。

{"title":"Evaluating magnetic properties of atmospheric particulate matter (PM10) at Brasília bus station, central Brazil","authors":"Poliana Dutra Maia , Elder Yokoyama , Luiza de Souza Romano , Guilherme Gomide , Jerome Depeyrot , Sebastião William da Silva , Myller de Sousa Tonhá","doi":"10.1016/j.apr.2025.102739","DOIUrl":"10.1016/j.apr.2025.102739","url":null,"abstract":"<div><div>This study analyzes the Saturation Magnetization (M<sub>S</sub>) in PM<sub>10</sub> particles collected at a central bus station in Brasília from 2018 to 2019, exploring its relationship with chemical elements (total metals - Fe, V, Cu, Mn, Sr, Cd - and ions - Ca<sup>2+</sup>, Mg<sup>2+</sup>, Na<sup>+</sup>, K<sup>+</sup>, Cl<sup>−</sup>, NO<sub>3</sub><sup>−</sup>, SO<sub>4</sub><sup>2−</sup>) and meteorological variables (relative humidity and global radiation). M<sub>S</sub> was measured using a S700X SQUID magnetometer and modeled with the Langevin equation, yielding values between 0.114 and 1.407 Am<sup>2</sup>/kg (n = 13), with a standard deviation of 0.01–0.1. Statistical analysis revealed similar median values between M<sub>S</sub> and the studied variables, indicating shared sources of PM<sub>10</sub>. The higher total concentrations of metals (Fe > Sr > Mn) and ions (Ca<sup>2+</sup> = 30 %, Na<sup>+</sup> = 23 % of total ions) suggest a natural origin of PM<sub>10</sub>, likely from soil minerals. In contrast, metals (Cu, V, Cd, Ni, Co and Zn) and ions (SO<sub>4</sub><sup>2−</sup>, NO<sub>3</sub><sup>−</sup>) of anthropogenic origin were associated with vehicle emissions and road dust. FTIR spectroscopy identified magnetite as the dominant magnetic carrier in PM<sub>10</sub>, with absorption peaks at 570 and 390 cm<sup>−1</sup>, and particle sizes ranging from 8.5 to 11 nm, although a broader size range (8–33 nm) was also observed. This supports the formation of heterogeneous particles through mineral-metal or carbonaceous adsorption, as confirmed by morphological data. The temporal distribution of M<sub>S</sub> in PM<sub>10</sub> samples showed peak values during the rainy season, despite no direct correlation with the studied variables, suggesting that iron solubility in minerals may influence the saturation magnetization.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102739"},"PeriodicalIF":3.5,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145962824","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}

引用次数: 0

Enhancing Air Quality forecasting with functional neural networks: A case study of PM2.5 in Seoul 用功能神经网络增强空气质量预报：以首尔市PM2.5为例

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-09 DOI: 10.1016/j.apr.2025.102732

Yaeji Lim , Yeonjoo Park

Reliable prediction of PM_2.5 levels is essential due to their substantial impacts on public health, the environment, and society. This is especially critical in regions like South Korea, where air quality is often compromised by elevated PM_2.5 concentrations resulting from domestic emissions and transboundary pollution. This study considers a Functional Neural Network (FNN) model that combines Functional Data Analysis (FDA) with deep learning techniques to predict PM_2.5 levels. The FNN model is applied to data from 13 monitoring stations in Seoul and compared with traditional multivariate-based neural networks (NN) and functional regression (FM) models. The enhanced predictive accuracy was observed from the FNN model by integrating dynamic temporal patterns in pollutant and meteorological trajectories as functional inputs. Additionally, this study proposes a model selection procedure within the FNN framework to identify a subset of functional inputs that significantly enhances prediction performance. Comprehensive comparison studies confirm that the proposed FNN, combined with the input selection procedure, offers a reliable tool for PM_2.5 prediction. This functional approach holds potential for supporting air quality management and protecting public health.

由于PM2.5水平对公共健康、环境和社会的重大影响，可靠的预测至关重要。这在韩国等地区尤其重要，这些地区的空气质量经常受到国内排放和跨境污染造成的PM2.5浓度升高的影响。这项研究考虑了一个功能神经网络（FNN）模型，该模型结合了功能数据分析（FDA）和深度学习技术来预测PM2.5水平。将该模型应用于首尔13个监测站的数据，并与传统的基于多变量的神经网络（NN）和函数回归（FM）模型进行了比较。通过将污染物和气象轨迹的动态时间模式作为功能输入，FNN模型的预测精度得到了提高。此外，本研究提出了FNN框架内的模型选择程序，以识别显著提高预测性能的功能输入子集。综合比较研究证实，所提出的FNN与输入选择程序相结合，为PM2.5预测提供了可靠的工具。这种功能性方法具有支持空气质量管理和保护公众健康的潜力。

{"title":"Enhancing Air Quality forecasting with functional neural networks: A case study of PM2.5 in Seoul","authors":"Yaeji Lim , Yeonjoo Park","doi":"10.1016/j.apr.2025.102732","DOIUrl":"10.1016/j.apr.2025.102732","url":null,"abstract":"<div><div>Reliable prediction of PM<sub>2.5</sub> levels is essential due to their substantial impacts on public health, the environment, and society. This is especially critical in regions like South Korea, where air quality is often compromised by elevated PM<sub>2.5</sub> concentrations resulting from domestic emissions and transboundary pollution. This study considers a Functional Neural Network (FNN) model that combines Functional Data Analysis (FDA) with deep learning techniques to predict PM<sub>2.5</sub> levels. The FNN model is applied to data from 13 monitoring stations in Seoul and compared with traditional multivariate-based neural networks (NN) and functional regression (FM) models. The enhanced predictive accuracy was observed from the FNN model by integrating dynamic temporal patterns in pollutant and meteorological trajectories as functional inputs. Additionally, this study proposes a model selection procedure within the FNN framework to identify a subset of functional inputs that significantly enhances prediction performance. Comprehensive comparison studies confirm that the proposed FNN, combined with the input selection procedure, offers a reliable tool for PM<sub>2.5</sub> prediction. This functional approach holds potential for supporting air quality management and protecting public health.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"16 12","pages":"Article 102732"},"PeriodicalIF":3.5,"publicationDate":"2025-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145044272","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}

引用次数: 0

Exploring homogeneous learnware market for PM2.5 prediction: A case study in the Yangtze River Delta region of China 同质学习软件市场在PM2.5预测中的应用——以中国长三角地区为例

IF 3.5 3区环境科学与生态学 Q2 ENVIRONMENTAL SCIENCES

Atmospheric Pollution Research

Pub Date : 2025-09-08 DOI: 10.1016/j.apr.2025.102738

Yuchen Zhang , Yanchuan Shao , Riyang Liu , Jun Bi , Zongwei Ma

In recent years, many machine learning techniques have been applied to the spatio-temporal predictions of PM_2.5 and achieved remarkable results. However, machine learning techniques come with several limitations in spatio-temporal extrapolation tasks, such as high modeling costs and unstable performance. The recently introduced learnware paradigm involves establishing a market of pre-trained models (learnwares) that can be applied in prediction tasks to provide optimal solutions. This approach is capable of overcoming the limitations in PM_2.5 modeling by offering tailored recommendations and operational frameworks for better predictions. In this study, a homogeneous learnware market is first developed in the Yangtze River Delta region, which contains several learnwares trained on PM_2.5 prediction datasets from 2015 to 2020. Various experiments were designed to explore the benefits and drawbacks of different deployment strategies offered by the learnware market. The results show that the learnware market can typically recommend the most suitable learnware or the combination of learnwares for a user’s task. The applications of multiple learnwares with appropriate strategies yield better outcomes than directly reusing a single learnware. In addition, adopting labeled data to reuse learnwares through the learnware market can get better performance than self-trained models developed by users. We demonstrate that the learnware market holds significant potential for accelerating the transfer of machine learning models to experts or nonexperts in air pollution exposure studies.

近年来，许多机器学习技术被应用于PM2.5的时空预测，并取得了显著的效果。然而，机器学习技术在时空外推任务中存在一些局限性，例如高建模成本和不稳定的性能。最近引入的learnware范式涉及建立一个预训练模型（learnware）市场，可以应用于预测任务以提供最佳解决方案。这种方法能够克服PM2.5建模的局限性，为更好的预测提供量身定制的建议和操作框架。本研究首先在长三角地区开发了一个同构的学习软件市场，该市场包含多个基于2015 - 2020年PM2.5预测数据集训练的学习软件。设计了各种实验来探索学习软件市场提供的不同部署策略的优点和缺点。结果表明，学习软件市场通常可以为用户的任务推荐最合适的学习软件或学习软件的组合。采用适当策略的多个学习器的应用比直接重用单个学习器产生更好的结果。此外，通过learnware market采用标记数据复用learnware可以获得比用户自行开发的训练模型更好的性能。我们证明了学习软件市场在加速机器学习模型向空气污染暴露研究中的专家或非专家的转移方面具有巨大的潜力。

{"title":"Exploring homogeneous learnware market for PM2.5 prediction: A case study in the Yangtze River Delta region of China","authors":"Yuchen Zhang , Yanchuan Shao , Riyang Liu , Jun Bi , Zongwei Ma","doi":"10.1016/j.apr.2025.102738","DOIUrl":"10.1016/j.apr.2025.102738","url":null,"abstract":"<div><div>In recent years, many machine learning techniques have been applied to the spatio-temporal predictions of PM<sub>2.5</sub> and achieved remarkable results. However, machine learning techniques come with several limitations in spatio-temporal extrapolation tasks, such as high modeling costs and unstable performance. The recently introduced learnware paradigm involves establishing a market of pre-trained models (learnwares) that can be applied in prediction tasks to provide optimal solutions. This approach is capable of overcoming the limitations in PM<sub>2.5</sub> modeling by offering tailored recommendations and operational frameworks for better predictions. In this study, a homogeneous learnware market is first developed in the Yangtze River Delta region, which contains several learnwares trained on PM<sub>2.5</sub> prediction datasets from 2015 to 2020. Various experiments were designed to explore the benefits and drawbacks of different deployment strategies offered by the learnware market. The results show that the learnware market can typically recommend the most suitable learnware or the combination of learnwares for a user’s task. The applications of multiple learnwares with appropriate strategies yield better outcomes than directly reusing a single learnware. In addition, adopting labeled data to reuse learnwares through the learnware market can get better performance than self-trained models developed by users. We demonstrate that the learnware market holds significant potential for accelerating the transfer of machine learning models to experts or nonexperts in air pollution exposure studies.</div></div>","PeriodicalId":8604,"journal":{"name":"Atmospheric Pollution Research","volume":"17 2","pages":"Article 102738"},"PeriodicalIF":3.5,"publicationDate":"2025-09-08","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145963302","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}

引用次数: 0