Browse > Article
http://dx.doi.org/10.7780/kjrs.2020.36.5.3.5

Monitoring Ground-level SO2 Concentrations Based on a Stacking Ensemble Approach Using Satellite Data and Numerical Models  

Choi, Hyunyoung (Department of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology)
Kang, Yoojin (Department of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology)
Im, Jungho (Environmental Satellite Center, Climate and Air Quality Research Department, National Institute of Environmental Research)
Shin, Minso (Department of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology)
Park, Seohui (Department of Urban and Environmental Engineering, Ulsan National Institute of Science and Technology)
Kim, Sang-Min (Environmental Satellite Center, Climate and Air Quality Research Department, National Institute of Environmental Research)
Publication Information
Korean Journal of Remote Sensing / v.36, no.5_3, 2020 , pp. 1053-1066 More about this Journal
Abstract
Sulfur dioxide (SO2) is primarily released through industrial, residential, and transportation activities, and creates secondary air pollutants through chemical reactions in the atmosphere. Long-term exposure to SO2 can result in a negative effect on the human body causing respiratory or cardiovascular disease, which makes the effective and continuous monitoring of SO2 crucial. In South Korea, SO2 monitoring at ground stations has been performed, but this does not provide spatially continuous information of SO2 concentrations. Thus, this research estimated spatially continuous ground-level SO2 concentrations at 1 km resolution over South Korea through the synergistic use of satellite data and numerical models. A stacking ensemble approach, fusing multiple machine learning algorithms at two levels (i.e., base and meta), was adopted for ground-level SO2 estimation using data from January 2015 to April 2019. Random forest and extreme gradient boosting were used as based models and multiple linear regression was adopted for the meta-model. The cross-validation results showed that the meta-model produced the improved performance by 25% compared to the base models, resulting in the correlation coefficient of 0.48 and root-mean-square-error of 0.0032 ppm. In addition, the temporal transferability of the approach was evaluated for one-year data which were not used in the model development. The spatial distribution of ground-level SO2 concentrations based on the proposed model agreed with the general seasonality of SO2 and the temporal patterns of emission sources.
Keywords
ground-level $SO_2$ concentrations; OMI; machine learning; stacking ensemble;
Citations & Related Records
Times Cited By KSCI : 2  (Citation Analysis)
연도 인용수 순위
1 Ren, Y., L. Zhang, and P. N. Suganthan, 2016. Ensemble classification and regression-recent developments, applications and future directions, IEEE Computational Intelligence Magazine, 11(1): 41-53.   DOI
2 Saini, R., and S. Ghosh, 2017. Ensemble classifiers in remote sensing, Proc. of A review, paper presented at 2017 International Conference on Computing, Communication and Automation (ICCCA), IEEE, Greater Noida, India, May. 5-6, pp. 1148-1152.
3 Scherbart, A., and T. W. Nattkemper, 2008. The diversity of regression ensembles combining bagging and random subspace method, paper presented at International Conference on Neural Information Processing, Springer, Berlin, Heidelberg, Germany, pp. 911-918.
4 Seo, J., J. Yoon, G.-H. Choo, D.-r. Kim, and D.-W. Lee, 2020. Long-term Trend Analysis of NOx and SOx over in East Asia Using OMI Satellite Data and National Emission Inventories (2005-2015), Korean Journal of Remote Sensing, 36(2-1): 121-137 (in Korean with abstract).   DOI
5 Seung-Woo, L., C.-H. Lee, D.-H. Ji, and H.-J. Youn, 2010. Temporal and spatial distribution of ambient sulfur dioxide concentration in forest areas, Korea, Korean Journal of Soil Science and Fertilizer, 43(6): 1035-1039 (in Korean with abstract).
6 Shang, F., D. Chen, X. Guo, J. Lang, Y. Zhou, Y. Li, and X. Fu, 2019. Impact of Sea Breeze Circulation on the Transport of Ship Emissions in Tangshan Port, China, Atmosphere, 10(11): 723.   DOI
7 Shin, M., Y. Kang, S. Park, J. Im, C. Yoo, and L. J. Quackenbush, 2020. Estimating ground-level particulate matter concentrations using satellitebased data: a review, GIScience & Remote Sensing, 57(2): 174-189.   DOI
8 Wang, X., and W. Sun, 2019. Meteorological parameters and gaseous pollutant concentrations as predictors of daily continuous PM2.5 concentrations using deep neural network in Beijing-Tianjin-Hebei, China, Atmospheric Environment, 211: 128-137.   DOI
9 Sulla-Menashe, D., and M. A. Friedl, 2018. User guide to collection 6 MODIS land cover (MCD12Q1 and MCD12C1) product, USGS: Reston, VA, USA, 1-18.
10 Sun, W., and Z. Li, 2020. Hourly PM2.5 concentration forecasting based on feature extraction and stacking-driven ensemble model for the winter of the Beijing-Tianjin-Hebei area, Atmospheric Pollution Research, Elsevier, Amsterdam, NL, Jun. 6, vol. 11, pp. 110-121.
11 Westervelt, D., A. Conley, A. Fiore, J. F. Lamarque, D. Shindell, M. Previdi, G. Faluvegi, G. Correa, and L. Horowitz, 2017. Multimodel precipitation responses to removal of US sulfur dioxide emissions, Journal of Geophysical Research: Atmospheres, 122(9): 5024-5038.   DOI
12 Wu, Y., R. Li, L. Cui, Y. Meng, H. Cheng, and H. Fu, 2020. The high-resolution estimation of sulfur dioxide ($SO_2$) concentration, health effect and monetary costs in Beijing, Chemosphere, 241: 125031.   DOI
13 Xiao, Y., J. Wu, Z. Lin, and X. Zhao, 2018. A deep learning-based multi-model ensemble method for cancer prediction, Computer methods and programs in biomedicine, 153: 1-9.   DOI
14 Yun, S.-Y., 2014. Seasonal washout effect of precipitation on major air pollutants ($O_3$, CO, $NO_2$, $SO_2$, PM10) during summer and winter, Unpublished master's thesis, Ewha Womans University, Seoul, Korea.
15 Zhai, B., and J. Chen, 2018. Development of a stacked ensemble model for forecasting and analyzing daily average PM2.5 concentrations in Beijing, China, Science of The Total Environment, 635: 644-658.   DOI
16 Chen, J., J. Yin, L. Zang, T. Zhang, and M. Zhao, 2019. Stacking machine learning model for estimating hourly PM2. 5 in China based on Himawari 8 aerosol optical depth data, Science of The Total Environment, 697: 134021.   DOI
17 Zhang, D., Y. Zhou, L. Zheng, R. Shi, and M. Chen, 2018. The spatial distribution characteristics and ground-level estimation of $NO_2$ and $SO_2$ over Huaihe River Basin and Shanghai based on satellite observations, Proc. of SPIE 10767, Remote Sensing and Modeling of Ecosystems for ustainability 54, 107670L SPIE optical Engineening, California, USA, Sep. 18, p. 10767.
18 Zhang, H., B. Di, D. Liu, J. Li, and Y. Zhan, 2019. Spatiotemporal distributions of ambient $SO_2$ across China based on satellite retrievals and ground observations: Substantial decrease in human exposure during 2013-2016, Environmental Research, 179: 108795.   DOI
19 Adhikari, R, 2015. A neural network based linear ensemble framework for time series forecasting, Neurocomputing, 157: 231-242.   DOI
20 Bauduin, S., L. Clarisse, J. Hadji-Lazaro, N. Theys, C. Clerbaux, and P.-F. Coheur, 2016. Retrieval of near-surface sulfur dioxide ($SO_2$) concentrations at a global scale using IASI satellite observations.
21 Divina, F., A. Gilson, F. Goméz-Vela, M. García Torres, and J. F. Torres, 2018. Stacking ensemble learning for short-term electricity consumption forecasting, Energies, 11(4): 949.   DOI
22 Kang, H., J. Park, J. Yang, W. Choi, D. Kim, and H. Lee, 2019. Uncertainties of $SO_2$ Vertical Column Density Retrieval from Ground-based Hyperspectral UV Sensor Based on Direct Sun Measurement Geometry, Korean Journal of Remote Sensing, 35(2): 289-298.   DOI
23 Feng, L., Y. Li, Y. Wang, and Q. Du, 2020. Estimating hourly and continuous ground-level PM2.5 concentrations using an ensemble learning algorithm: The ST-stacking model, Atmospheric Environment, 223: 117242.   DOI
24 Fernandes, A., M. Riffler, J. Ferreira, S. Wunderle, C. Borrego, and O. Tchepel, 2019. Spatial analysis of aerosol optical depth obtained by air quality modelling and SEVIRI satellite observations over Portugal, Atmospheric Pollution Research, 10(1): 234-243.   DOI
25 Fioletov, V., C. McLinden, N. Krotkov, M. Moran, and K. Yang, 2011. Estimation of $SO_2$ emissions using OMI retrievals, Geophysical Research Letters, 38(21).
26 Herrera, F., F. Charte, A. J. Rivera, and M. J. Del Jesus, 2016. Multilabel classification, in Multilabel Classification, edited, pp. 17-31, Springer.
27 Huang, K., Q. Xiao, X. Meng, G. Geng, Y. Wang, A. Lyapustin, D. Gu, and Y. Liu, 2018. Predicting monthly high-resolution PM2. 5 concentrations with random forest model in the North China Plain, Environmental Pollution, 242: 675-683.   DOI
28 Khan, A., K.-H. Kim, J. E. Szulejko, R. J. Brown, E.-C. Jeon, J.-M. Oh, Y. S. Shin, and A. A. Adelodun, 2017. Long-term trends in airborne $SO_2$ in an air quality monitoring station in Seoul, Korea, from 1987 to 2013, Journal of the Air & Waste Management Association, 67(8): 923-932.   DOI
29 Kim, B.-W., and K.-H. Kim, 2014. The Analysis of Time Series of $SO_2$ Concentration and the Control Factor in An Urban Area of Yongsan-gu, Seoul, Journal of the Korean Earth Science Society, 35(7): 543-553.   DOI
30 Kharol, S. K., C. A. McLinden, C. E. Sioris, M. W. Shephard, V. Fioletov, A. van Donkelaar, P. Sajeev, and R. V. Martin, 2017. OMI satellite observations of decadal changes in ground-level sulfur dioxide over North America, Atmospheric Chemistry and Physics, 17(9): 5921.   DOI
31 Li, H., J. Wang, R. Li, and H. Lu, 2019. Novel analysisforecast system based on multi-objective optimization for air quality index, Journal of Cleaner Production, 208: 1365-1383.   DOI
32 Lee, H.-D., G.-H. Lee, I.-D. Kim, J.-S. Kang, and K.-J. Oh, 2013. The influences of concentration distribution and movement of air pollutants by sea breeze and mist around Onsan industrial complex, Clean Technology, 19(2): 95-104.   DOI
33 Lee, H. D., J. W. Yoo, M. K. Kang, J. S. Kang, J. H. Jung, and K. J. Oh, 2014. Evaluation of concentrations and source contribution of PM10 and $SO_2$ emitted from industrial complexes in Ulsan, Korea: Interfacing of the WRF-CALPUFF modeling tools, Atmospheric Pollution Research, 5(4): 664-676.   DOI
34 Li, H., F. Faruque, W. Williams, M. Al-Hamdan, J. Luvall, W. Crosson, D. Rickman, and A. Limaye, 2009. Optimal temporal scale for the correlation of AOD and ground measurements of PM2.5 in a realtime air quality estimation system, Atmospheric Environment, 43(28): 4303-4310.   DOI
35 Oh, I., Y. Kim, and M. Hwang, 2004. Effects of late sea-breeze on ozone distributions in the coastal urban area, J. Kor. Soc. Atmos. Environ, 20: 345-360.
36 Li, R., L. Cui, J. Liang, Y. Zhao, Z. Zhang, and H. Fu, 2020. Estimating historical $SO_2$ level across the whole China during 1973-2014 using random forest model, Chemosphere, 247: 125839.   DOI
37 Liu, Q., S. Wang, W. Zhang, J. Li, and G. Dong, 2019. The effect of natural and anthropogenic factors on PM2.5: Empirical evidence from Chinese cities with different income levels, Science of the Total Environment, 653: 157-167.   DOI
38 Oh, I., J.-H. Bang, S. Kim, E. Kim, M.-K. Hwang, and Y. Kim, 2016. Spatial distribution of air pollution in the Ulsan metropolitan region, Journal of Korean Society for Atmospheric Environment, 32(4): 394-407.   DOI
39 Park, J., J. Ryu, D. Kim, J. Yeo, and H. Lee, 2016. Long-range transport of $SO_2$ from continental Asia to northeast Asia and the northwest Pacific ocean: flow rate estimation using OMI data, surface in situ data, and the HYSPLIT model, Atmosphere, 7(4): 53.   DOI
40 Pandey, A. K., R. P. Kumar, and K. Kumar, 2015. Satellite and ground-based seasonal variability of $NO_2$ and $SO_2$ over New Delhi, India, paper presented at Remote Sensing of Clouds and the Atmosphere XX, International Society for Optics and Photonics.
41 Park, S., M. Shin, J. Im, C.-K. Song, M. Choi, J. Kim, S. Lee, R. Park, J. Kim, and D.-W. Lee, 2019. Estimation of ground-level particulate matter concentrations through the synergistic use of satellite observations and process-based models over South Korea.