Modeling of PM2.5 in the Yangtze River delta based on geographically weighted random forest
CHEN Yi-min1,2, SU Zhang-wen1, CHEN Yi-ping1, LIN Zi-peng1
1. School of Petrochemical Engineering, Zhangzhou Institute of Technology, Zhangzhou 363000, China; 2. Collaborative Innovation Center of Fine Chemicals in Fujian Province, Zhangzhou 363000, China
Abstract：This study used Random Forest (RF) and Geographically Weighted Random Forest (GWRF) to train, verify, and test the PM2.5 concentration and data of its driving factors in the Yangtze River delta region from 2003 to 2019 and explore their relationships. The results show that: (1) Compared with the RF model, the GWRF model performs better in training and predicting PM2.5 concentration. The evaluation indicators of the GWRF model are better, and the residual spatial autocorrelation is lower than the RF model. (2) The GWRF model predicts a better PM2.5 concentration distribution in 2019 than the RF model, which is basically consistent with the actual observed concentration distribution. However, overestimation in the north, underestimation in the south, and the area of overestimation are more significant than the underestimation for the two models. (3) The RF model is global in studying the most significant driving factors of PM2.5concentration distribution. In contrast, the GWRF model's interpretation of PM2.5 influence by drought, temperature, temperature difference, wind, and human interference shows a localized effect. On a large scale, this localized effect has practical guidance for preventing and controlling PM2.5. In addition, in the context of global climate warming and spatial heterogeneity of regional climate, integrating drought into PM2.5 prediction and establishing models with localized effects can help environmental regulatory agencies and decision-makers formulate prevention and control measures.
陈艺敏, 苏漳文, 陈移萍, 林子彭. 基于地理加权随机森林的长三角PM2.5建模[J]. 中国环境科学, 2024, 44(8): 4240-4248.
CHEN Yi-min, SU Zhang-wen, CHEN Yi-ping, LIN Zi-peng. Modeling of PM2.5 in the Yangtze River delta based on geographically weighted random forest. CHINA ENVIRONMENTAL SCIENCECE, 2024, 44(8): 4240-4248.
[1] Yang D Y, Ye C, Wang, X M, et al. Global distribution and evolvement of urbanization and PM2.5 [J]. Atmospheric Environment, 2018,182: 171-178. [2] He W H, Meng H, Han J, et al. Spatiotemporal PM2.5 estimations in China from 2015 to 2020 using an improved gradient boosting decision tree [J]. Chemosphere, 2022,296:134003. [3] Liu G X, Dong X C, Kong Z Y, et al. Does national air quality monitoring reduce local air pollution? The case of PM2.5 for China [J]. Journal of Environmental Management, 2021,296:113232. [4] 许珊,邹滨,胡晨霞.面向场景的城市PM2.5浓度空间分布精细模拟[J]. 中国环境科学, 2019,39(11):4570-4579. Xu S, Zou B, Hu C X. Urban scene-oriented simulation of the spatial distribution of PM2.5 concentration in an intra-urban area at fine scale [J]. China Environmental Science, 2023,43(12):6246-6260. [5] Yan X, Zang Z, Jiang Y Z, et al. A spatial-temporal interpretable deep learning model for improving interpretability and predictive accuracy of satellite-based PM2.5 [J]. Environmental Pollution, 2021,273:116459. [6] Liu J J, Weng F Z, Li Z Q. Satellite-based PM2.5 estimation directly from reflectance at the top of the atmosphere using a machine learning algorithm [J]. Atmospheric Environment, 2019,208:113-122. [7] Kerckhoffs J, Hoek G, Portengen L, et al. Performance of prediction algorithms for modeling outdoor air pollution spatial surfaces [J]. Environmental Science & Technology, 2019,53:1413-1421. [8] 廖彤,林玉君,季文浩,等.珠江三角洲光化学活跃期甲醛及其前体物的源解析研究[J]. 中国环境科学, 2022,42(12):5533-5541. Liao T, Lin Y J, Ji W H, et al. Research on the sources of formaldehyde and its precursors during the active period of atmospheric photochemistry in the Pearl River Delta. China Environmental Science, 2022,42(12):5533-5541. [9] Wang X P, Sun W B. Meteorological parameters and gaseous pollutant concentrations as predictors of daily continuous PM2.5 concentrations using deep neural network in Beijing-Tianjin-Hebei, China [J]. Atmospheric Environment, 2019,211:128-137. [10] Su Z W, Xu Z H, Lin L, et al. Exploration of the Contribution of Fire Carbon Emissions to PM2.5 and Their Influencing Factors in Laotian Tropical Rainforests [J]. Remote Sensing, 2022,14:4052. [11] Maiti A, Zhang Q, Sannigrahi S, et al. Exploring spatiotemporal effects of the driving factors on COVID-19incidences in the contiguous United States [J]. Sustainable Cities and Society, 2021,68:102784. [12] Danesh Yazd, M, Kuang Z, Dimakopoulou K, et al. Predicting fine particulate matter (PM2.5) in the greater london area: an ensemble approach using machine learning methods [J]. Remote Sensing, 2020,12:914. [13] Zhang T N, He W H, Zheng H, et al. Satellite-based ground PM2.5 estimation using a gradient boosting decision tree [J]. Chemosphere, 2021,268:128801. [14] Luo Z Y, Huang F F, Liu H. PM2.5 concentration estimation using convolutional neural network and gradient boosting machine [J]. Journal of Environmental Sciences, 2020,98:85-93. [15] Hu X F, Belle J H, Meng X, et al. Estimating PM2.5 concentrations in the conterminous United States using the random forest approach [J]. Environmental Science & Technology, 2017,51(12):6936. [16] Su Z W, Lin L, Chen Y M, et al. Understanding the distribution and drivers of PM2.5 concentrations in the Yangtze River Delta from 2015 to 2020 using Random Forest Regression [J]. Environmental Monitoring and Assessment, 2022,94(4):284.1-284.17. [17] He Q Q, Huang B. Satellite-based mapping of daily high-resolution ground PM2.5 in China via space-time regression modeling [J]. Remote Sensing of Environment, 2018,206:72-83. [18] Georganos S, Brousse O, Dujardin S, et al. Modelling and mapping the intra-urban spatial distribution of Plasmodium falciparum parasite rate using very-high-resolution satellite derived indicators [J]. International Journal of Health Geographics, 2020,19:1-18. [19] Georganos S, Grippa T, Gadiaga A N, et al. Geographical random forests: a spatial extension of the random forest algorithm to address spatial heterogeneity in remote sensing and population modeling [J]. Geocarto International, 2019,36(2):121-136. [20] Quiñones S, Goyal A, Ahmed Z U. Geographically weighted machine learning model for untangling spatial heterogeneity of type 2diabetes mellitus (T2D) prevalence in the USA [J]. Scientific Reports, 2021,11: 6955. [21] Donkelaar V, Martin A, Randall V,et al. Regional estimates of chemical composition of fine particulate matter using a combined geosciencestatistical method with information from satellites, models, and monitors [J]. Environmental Science & Technology, 2019,53:2843. [22] 吴舒祺,么嘉棋,杨冉,等.长三角城市群PM2.5时空变化和影响因素分析[J]. 环境科学, 2023,44(10):5325-5334. Wu S Q, Yao J Q, Yang R, et al. Spatio-temporal variations in PM2.5 and its influencing factors in the Yangtze River delta urban agglomeration [J]. Environmental Science, 2023,44(10):5325-5334. [23] 张瑞旭,芮守娟,王伟军,等.气象条件对西安市夏季和冬季近地面大气环境污染特征影响[J]. 生态环境学报, 2020,29(1):165-174. Zhang R X, Rui S J, Wamg W J, et al. Effects of meteorological conditions on the characteristics of near-surface atmospheric environmental pollution in summer and winter in Xi'an, China [J]. Ecology and Environmental Sciences, 2020,29(1):165-174. [24] 高婵娟,赵啟超,丁若男,等.2018年吉林市大气污染物浓度变化及其与气象因素的相关性分析[J]. 环境工程, 2021,39(5):71-79. Gao C J, Zhao Q C, Ding R N, et al. Variations of atmospheric pollutants concent rations and their correlation with meteorological factor in Jilin city in 2018[J]. Environmental Engineering, 2021,39(5):71-79. [25] Demetillo M A G, Anderson J F, Geddes J A, et al. Observing severe drought influences on ozone air pollution in California [J]. Environmental Science & Technology, 2019,53:4695-4706. [26] Berg A, McColl K A. No projected global drylands expansion under greenhouse warming [J]. Nature Climate Change, 2021,11:331-337. [27] 李梦真,张廷斌,易桂花,等.四川盆地PM2.5浓度时空变化特征遥感监测与影响因子分析[J]. 环境科学, 2021,42(7):3136-3146. Li M Z, Zhang T B, Yi G H, et al. Spatio-temporal variation characteristics monitored by remotely sensed technique of PM2.5 concentration and its influencing factor analysis in Sichuan basin, China [J]. Environmental Science, 2021,42(7):3136-3146. [28] 刘世梁,刘芦萌,武雪,等.区域生态效应研究中人类活动强度定量化评价[J]. 生态学报, 2018,38(19):6797-6809. Liu S L, Liu L M, Wu X, et al. Quantitative evaluation of human activity intensity on the regional ecological impact studies. Acta Ecologica Sinica, 2018,38(19):6797-6809. [29] Martinez A I, Labib S M. Demystifying normalized difference vegetation index (NDVI) for greenness exposure assessments and policy interventions in urban greening [J]. Environmental Research, 2023,220:115155. [30] Su Z W, Zheng L J, Luo S S, et al. Modeling wildfire drivers in Chinese tropical forest ecosystems using global logistic regression and geographically weighted logistic regression [J]. Natural Hazards, 2021, 108:1317-1345. [31] Ma X Q,Yin Z C,Cao B F, et al. Meteorological influences on co-occurrence of O3 and PM2.5 pollution and implication for emission reductions in Beijing-Tianjin-Hebei [J]. Science China (Earth Sciences), 2023,66(6):1258-1267. [32] Prosperi P, Bloise M, Tubiello F N, et al. New estimates of greenhouse gas emissions from biomass burning and peat fires using MODIS Collection 6burned areas [J]. Climatic Change, 2020,161:415-432. [33] Schag G M, Stow D A, Riggan P J, et al. Spatial-statistical analysis of landscape-level wildfire rate of spread [J]. Remote Sensing, 2022, 14(16):3980. [34] Edwards J, Santosmedellin C, Liechty Z, et al. Compositional shifts in root-associated bacterial and archaeal microbiota track the plant life cycle in field-grown rice [J]. PLOS Biology, 2018,16(2):2003862. [35] Rodrigues M, de la Riva J, Fotheringham S. Modeling the spatial variation of the explanatory factors of human-caused wildfires in Spain using geographically weighted logistic regression [J]. Applied Geography, 2014,48:52-63. [36] Chen J M, Li C L, Ristovski Z, et al. A review of biomass burning: emissions and impacts on air quality, health and climate in China [J]. Science of the Total Environment, 2017,579:1000-1034. [37] Feng H H, Zou B, Tang Y M. Scale- and regiondependence in landscape-PM2.5 correlation: Implications for urban planning [J]. Remote Sensing, 2017,9(9):918. [38] She Q N, Peng X, Xu Q, et al. Air quality and its response to satellite-derived urban form in the Yangtze River Delta, China [J]. Ecological Indicators, 2017,75:297-306. [39] Yan J W, Tao F, Zhang S Q, et al. Spatiotemporal distribution characteristics and driving forces of PM2.5 in three urban agglomerations of the Yangtze River Economic Belt [J]. International Journal of Environmental Research and Public Health, 2021,18(5):2222. [40] Wei J, Li Z Q, Pinker R T, et al. Himawari-8-derived diurnal variations in ground-level PM2.5 pollution across China using the fast space-time Light Gradient Boosting Machine (LightGBM) [J]. Atmospheric Chemistry and Physics, 2021,21:7863-7880. [41] Wang Y X, Xie Y Y, Dong W H, et al. Adverse effects of increasing drought on air quality via natural processes [J]. Atmospheric Chemistry and Physics, 2017,17:12827-12843. [42] Feng X Y, Wei S M, Wang S G. Temperature inversions in the atmospheric boundary layer and lower troposphere over the Sichuan Basin, China: climatology and impacts on air pollution [J]. Science of the Total Environment, 2020,726:138579. [43] 孙璐,田文寿,张文煜,等.上海周边地区海陆风特征及其对PM2.5和O3浓度的影响研究[J]. 环境科学学报, 2022,42(9):339-350. Sun L, Tian W S, Zhang W Y, et al. Characteristics of sea-land breezes and their impact on PM2.5 and O3 in Shanghai and adjacent areas [J]. Acta Scientiae Circumstantiae, 2022,42(9):339-350. [44] 杨显玉,吕雅琼,邵平,等.四川盆地大气氨与氮氧化物排放对细颗粒物污染的影响及减排潜力[J]. 中国环境科学, 2022,42(8): 3502-3511. Yang X Y, Tian W S, Zhang W L, et al. Impacts of ammonia and nitrogen oxides emissions on fine particle matter pollution and mitigation potential over the Sichuan Basin [J]. China Environmental Science, 2022,42(8):3502-3511. [45] Geng G N, Zhang Q, Martin RV, et al. Estimating long-term PM2.5 concentrations in China using satellite-based aerosol optical depth and a chemical transport model [J]. Remote Sensing of Environment, 2015,166:262-270.
