PM2.5 remote sensing estimation based on spatiotemporal factor optimization model
ZHANG Na1, CHEN Wen-qian1, BAI Xue-song1, CAO Xiao-yi2
1. School of Information and Control Engineering, Qingdao University of Technology, Qingdao, 266520; 2. Key Laboratory for Semi-Arid Climate Change, Ministry of Education, College of Atmospheric Sciences, Lanzhou University, Lanzhou 730000, China
Abstract：In order to obtain the continuous spatiotemporal distribution of PM2.5 concentration and improve the estimation accuracy, this paper proposes a new PM2.5 estimation model (SFRF) based on the optimization of spatiotemporal factors. The SFRF model integrates spatiotemporal factors into a random forest (RF) algorithm by integrating high-resolution (1km) satellite-retrieved aerosol optical depth (AOD) products, as well as meteorological data, nighttime light data, and vegetation. Using these data to build an SFRF model to accurately predict the PM2.5 concentration in Shandong Province in 2019 and generate high spatial resolution (1km) PM2.5 concentration in Shandong Province. The performance of the SFRF model was evaluated using the ten-fold cross-validation method and compared with the BPNN, SVM, XGBoost, RF and PCA-RF models. The results showed that the coefficient of determination and root mean square error (RMSE) values of the SFRF model verification are 0.85 and 8.10μg/m3, respectively, which are better than other models. The SFRF model can estimate PM2.5 concentration in Shandong Province with high spatial resolution on daily, seasonal, and annual scales.
[1] Xing Y F, Xu Y H, Shi M H, et al. The impact of PM2.5 on the human respiratory system [J]. Journal of Thoracic Disease, 2016,8(1):E69. [2] Lin Y, Zou J, Yang W, et al. A review of recent advances in research on PM2.5 in China [J]. International Journal of Environmental Research and Public Health, 2018,15(3):438. [3] Yin H, Pizzol M, Xu L. External costs of PM2.5 pollution in Beijing, China: Uncertainty analysis of multiple health impacts and costs [J]. Environmental Pollution, 2017,226:356-369. [4] Lee H J, Coull B A, Bell M L, et al. Use of satellite~based aerosol optical depth and spatial clustering to predict ambient PM2.5 concentrations [J]. Environmental Research, 2012,118:8-15. [5] Chang H H, Hu X, Liu Y. Calibrating MODIS aerosol optical depth for predicting daily PM2.5 concentrations via statistical downscaling [J]. Journal of Exposure Science & Environmental Epidemiology, 2014, 24(4):398-404. [6] You W, Zang Z, Pan X, et al. Estimating PM2.5 in Xi'an, China using aerosol optical depth: A comparison between the MODIS and MISR retrieval models [J]. Science of the Total Environment, 2015,505:1156-1165. [7] Gao M, Han Z, Tao Z, et al. Air quality and climate change, Topic 3 of the Model Inter~Comparison Study for Asia Phase III (MICS~Asia III)–Part 2: Aerosol radiative effects and aerosol feedbacks [J]. Atmospheric Chemistry and Physics, 2020,20(2):1147-1161. [8] Gupta P, Khan M N, da Silva A, et al. MODIS aerosol optical depth observations over urban areas in Pakistan: Quantity and quality of the data for air quality monitoring [J]. Atmospheric Pollution Research, 2013,4(1):43-52. [9] Zhao R, Gu X, Xue B, et al. Short period PM2.5 prediction based on multivariate linear regression model [J]. PloS One, 2018,13(7):e0201011. [10] Chelani A B. Estimating PM2.5 concentration from satellite derived aerosol optical depth and meteorological variables using a combination model [J]. Atmospheric Pollution Research, 2019,10(3):847-857. [11] Ma Z, Liu Y, Zhao Q, et al. Satellite-derived high resolution PM2.5 concentrations in Yangtze River Delta Region of China using improved linear mixed effects model [J]. Atmospheric Environment, 2016,133:156-164. [12] Hu X, Waller L A, Al-Hamdan M Z, et al. Estimating ground-level PM2.5 concentrations in the southeastern US using geographically weighted regression [J]. Environmental Research, 2013,121:1-10. [13] Guo B, Wang X, Pei L, et al. Identifying the spatiotemporal dynamic of PM2.5 concentrations at multiple scales using geographically and temporally weighted regression model across China during 2015~2018[J]. Science of The Total Environment, 2021,751:141765. [14] 康新礼,张文豪,刘原萍,等.基于随机森林的京津冀地区PM2.5遥感反演及变化分析[J]. 遥感技术与应用, 2022,37(2):424-435. Kang X L, Zhang W H, Liu Y P, et al. Remote sensing inversion and change analysis of PM2.5 in the Beijing-Tianjin-Hebei region based on random forest [J]. Remote Sensing Technology and Applications, 2022,37(2):424-435. [15] 郭骐嘉,姚宜斌,周永江.融合GNSS气象参数的PM2.5随机森林预测模型[J]. 测绘科学, 2021,46(4):37-42,56. Guo Q J, Yao Y B, Zhou Y J. PM2.5 random forest prediction model integrating GNSS meteorological parameters [J]. Surveying and Mapping Science, 2021,46(4):37-42,56. [16] 卫星君,赵晓萌,王琦,等.基于特征指标的气象因子对PM2.5浓度的影响分析[J]. 中国环境监测, 2022,38(6):90-100. Wei X J, Zhao X M, Wang Q, et al. Analysis of the impact of meteorological factors on PM2.5 concentration based on characteristic indicators [J]. China Environmental Monitoring, 2022,38(6):90-100. [17] 任才溶,谢刚.基于随机森林和气象参数的PM2.5浓度等级预测[J]. 计算机工程与应用, 2019,55(2):213-220. Ren C R, Xie G. PM2.5 concentration level prediction based on random forest and meteorological parameters [J]. Computer Engineering and Applications, 2019,55(2):213-220. [18] 周琪,于洋,刘苗苗,等.基于机器学习和非参数估计的PM2.5风险评估[J]. 中国环境科学, 2022,42(8):3554-3560. Zhou Q, Yu Y, Liu M M, et al. PM2.5 risk assessment based on machine learning and non-parametric estimation [J]. Chinese Environmental Science, 2022,42(8):3554-3560. [19] 李卓建,赵尚民,郭鹏程.太原市城区PM2.5浓度时空分布特征研究[J]. 环境污染与防治, 2021,43(3):353-358. Li Z J, Zhao S M, Guo P C. Research on the spatial and temporal distribution characteristics of PM2.5 concentration in urban areas of Taiyuan City [J]. Environmental Pollution and Prevention, 2021,43(3):353-358. [20] 范丽行,杨晓辉,宋春杰,等.基于时空混合效应模型的京津冀PM2.5浓度变化模拟[J]. 环境科学, 2022,43(5):2262-2273. Fan L X, Yang X H, Song C J, et al. Simulation of PM2.5 concentration changes in Beijing-Tianjin-Hebei based on spatiotemporal mixed effect model [J]. Environmental Science, 2022,43(5):2262-2273. [21] Zhou Y, Chang F J, Chang L C, et al. Multi~output support vector machine for regional multi~step~ahead PM2.5 forecasting [J]. Science of the Total Environment, 2019,651:230-240. [22] 李建新,刘小生,刘静,等.基于MRMR~HK~SVM模型的PM2.5浓度预测[J]. 中国环境科学, 2019,39(6):2304-2310. Li J X, Liu X S, Liu J, et al. PM2.5 concentration prediction based on MRMR~HK~SVM model [J]. Chinese Environmental Science, 2019,39(6):2304-2310. [23] 宋国君,国潇丹,杨啸,等.沈阳市PM2.5浓度ARIMA~SVM组合预测研究[J]. 中国环境科学, 2018,38(11):4031-4039. Song G J, Guo X D, Yang X, et al. Research on ARIMA~SVM combination prediction of PM2.5 concentration in Shenyang [J]. Chinese Environmental Science, 2018,38(11):4031-4039. [24] Gao Y, ang Z, Li C, et al. Assessing neighborhood variations in ozone and PM2.5 concentrations using decision tree method [J]. Building and Environment, 2021,188:107479. [25] Huang K, Xiao Q, Meng X, et al. Predicting monthly high~resolution PM2.5 concentrations with random forest model in the North China Plain [J]. Environmental Pollution, 2018,242:675-683. [26] Chen W, Ran H, Cao X, et al. Estimating PM2.5 with high-resolution 1-km AOD data and an improved machine learning model over Shenzhen, China [J]. Science of the Total Environment, 2020,746:141093. [27] Jing Y, Pan L, Sun Y. Estimating PM2.5 concentrations in a central region of China using a three-stage model [J]. International Journal of Digital Earth, 2023,16(1):578-592. [28] Yang Y, Christakos G. Spatiotemporal characterization of ambient PM2.5 concentrations in Shandong province (China) [J]. Environmental Science & Technology, 2015,49(22):13431-13438. [29] Eck T F, Holben B N, Reid J S, et al. Observations of the interaction and transport of fine mode aerosols with cloud and/or fog in Northeast Asia from Aerosol Robotic Network and satellite remote sensing [J]. Journal of Geophysical Research: Atmospheres, 2018,123(10):5560-5587. [30] Oliver M A, Webster R. Kriging: A method of interpolation for geographical information systems [J]. International Journal of Geographical Information System, 1990,4(3):313-332. [31] Yang J, Hu M. Filling the missing data gaps of daily MODIS AOD using spatiotemporal interpolation [J]. Science of the Total Environment, 2018,633:677-683. [32] 王敏,邹滨,郭宇,等.基于BP人工神经网络的城市PM2.5浓度空间预测[J]. 环境污染与防治, 2013,35(9):63-66,70. Wang M, Zou B, Guo Y, et al. Spatial prediction of urban PM2.5 concentration based on BP artificial neural network [J]. Environmental Pollution and Prevention, 2013,35(9):63-66,70. [33] Chen Y. Prediction algorithm of PM2.5 mass concentration based on adaptive BP neural network [J]. Computing, 2018,100(8):825-838. [34] Weizhen H, Zhengqiang L, Yuhuan Z, et al. Using support vector regression to predict PM10 and PM2.5 [C]//IOP conference series: Earth and Environmental Science, IOP Publishing, 2014,17(1):012268. [35] Ma J, Yu Z, Qu Y, et al. Application of the XGBoost machine learning method in PM2.5 prediction: A case study of Shanghai [J]. Aerosol and Air Quality Research, 2020,20(1):128-138. [36] Gupta I, Sharma V, Kaur S, et al. PCA-RF: an efficient Parkinson's disease prediction model based on random forest classification [J]. arXiv Preprint arXiv: 2203,11287,2022. [37] Chen W, Ran H, Cao X, et al. Estimating PM2.5 with high~resolution 1~km AOD data and an improved machine learning model over Shenzhen, China [J]. Science of the Total Environment, 2020,746: 141093. [38] Wei J, Huang W, Li Z, et al. Estimating 1~km~resolution PM2.5 concentrations across China using the space~time random forest approach [J]. Remote Sensing of Environment, 2019,231:111221. [39] Jiang G, Wang W. Error estimation based on variance analysis of k~fold cross~validation [J]. Pattern Recognition, 2017,69:94-106. [40] 卢鋆镆,曾穗平,曾坚,等.基于随机森林的高分辨率PM2.5浓度时空变化模拟——以中原城市群核心区为例[J]. 中国环境科学, 2023,43(7):32993311. Lu J R, Zeng S P, Zeng J, et al. High-resolution PM2.5 concentration spatiotemporal change simulation based on random forest - taking the core area of the Central Plains urban agglomeration as an example [J]. Chinese Environmental Science, 2023,43(7):32993311. [41] Guo B, Zhang D, Pei L, et al. Estimating PM2.5 concentrations via random forest method using satellite, auxiliary, and ground~level station dataset at multiple temporal scales across China in 2017[J]. Science of The Total Environment, 2021,778:146288. [42] 贾松林,苏林,陶金花,等.卫星遥感监测近地表细颗粒物多元回归方法研究[J]. 中国环境科学, 2014,34(3):565-573. Jia S L, Su L, Tao J H, et al. Research on multiple regression method for near-surface fine particulate matter monitoring by satellite remote sensing [J]. Chinese Environmental Science, 2014,34(3):565-573. [43] Qin K, Zou J, Guo J, et al. Estimating PM1 concentrations from MODIS over Yangtze River Delta of China during 2014~2017[J]. Atmospheric Environment, 2018,195:149-158. [44] Ma Z, Hu X, Huang L, et al. Estimating ground~level PM2.5 in China using satellite remote sensing [J]. Environmental Science & Technology, 2014,48(13):7436-7444. [45] 徐发昭,李净,褚馨德,等.基于MODIS数据与多机器学习法的日PM2.5模拟研究[J]. 中国环境科学, 2022,42(6):2523-2529. Xu F Z, Li J, Chu X D, et al. Research on daily PM2.5 simulation based on MODIS data and multi-machine learning method [J]. Chinese Environmental Science, 2022,42(6):2523-2529. [46] Yu M Y, Xu Y, Li J Q, et al. Geographic Detector-Based Spatiotemporal Variation and Influence Factors Analysis of PM2.5 in Shandong, China [J]. Polish Journal of Environmental Studies, 2021,30(1). [47] 刘新玲.2000~2005年山东省大气污染变化特征分析[D]. 济南:山东大学, 2008. Liu X L. Analysis of changing characteristics of air pollution in Shandong Province from 2000 to 2005[D]. Jinan: Shandong University, 2008. [48] Zhang H, Jiang Q, Wang J, et al. Analysis on the impact of two winter precipitation episodes on PM2.5 in Beijing [J]. Environmental Science and Ecotechnology, 2021,5:100080.
