[논문]부스팅 기반 기계학습기법을 이용한 지상 미세먼지 농도 산출

박서희; 김미애; 임정호

doi:10.7780/kjrs.2021.37.2.11

부스팅 기반 기계학습기법을 이용한 지상 미세먼지 농도 산출
Estimation of Ground-level PM10 and PM2.5 Concentrations Using Boosting-based Machine Learning from Satellite and Numerical Weather Prediction Data 원문보기

대한원격탐사학회지 = Korean journal of remote sensing, v.37 no.2, 2021년, pp.321 - 335

박서희 (울산과학기술원 도시환경공학부) , 김미애 (울산과학기술원 도시환경공학부) , 임정호 (울산과학기술원 도시환경공학부)

초록
AI-Helper

미세먼지 (PM10) 및 초미세먼지 (PM2.5)는 인체에 흡수 가능하여 호흡기 질환 및 심장 질환과 같이 인체건강에 악영향을 미치며, 심각할 경우 조기 사망에 영향을 줄 수 있다. 전 세계적으로 현장관측기반의 모니터링을 수행하고 있지만 미 관측지역에 대한 대기질 분포의 공간적인 한계점이 존재하여 보다 광범위한 지역에 대한 지속적이고 정확한 모니터링이 필요한 상황이다. 위성기반 에어로졸 정보를 사용함으로써 이러한 현장 관측자료의 한계점을 극복할 수 있다. 따라서 본 연구에서는 다양한 위성 및 모델자료를 활용하여 2019년도에 대해 한 시간 단위의 지상 PM10 및 PM2.5 농도를 추정하였다. GOCI 위성의 관측영역을 포함하는 동아시아 지역에 대해 트리 기반 앙상블 방법을 사용하는 Boosting 기법인 GBRTs (Gradient Boosted Regression Trees)와 LightGBM (Light Gradient Boosting Machine)을 활용하여 모델을 구축하였다. 또한, 기상변수 및 토지피복변수의 사용유무에 따른 모델의 성능을 비교하기 위해 두 가지 festure set으로 나누어 테스트하였다. 두 기법 모두 주요 변수인 AOD (Aerosol Optical Depth), SSA (Single Scattering Albedo), DEM (Digital Eelevation Model), DOY (Day of Year), HOD (Hour of Day)와 기상변수 및 토지피복변수를 함께 사용한 Feature set 1을 사용하였을 때 높은 정확도를 보였다. Feature set 1에 대해 GBRT 모델이 LightGBM에 비해서약 10%의 정확도 향상을 보였다. 가장 정확도가 높았던 기상 및 지표면 변수를 포함한 Feature set1을 사용한 GBRT기반 모델을 최종모델로 선정하였으며 (PM10: R2 = 0.82 nRMSE = 34.9%, PM2.5: R2 = 0.75 nRMSE = 35.6%), 계절별 및 연평균 PM10 및 PM2.5 농도에 대한 공간적인 분포를 확인해본 결과, 현장관측자료와 비슷한 공간 분포를 보였으며, 국가별 농도 분포와 계절에 따른 시계열 농도 패턴을 잘 모의하였다.

Abstract ▼ AI-Helper

Particulate matter (PM10 and PM2.5 with a diameter less than 10 and 2.5 ㎛, respectively) can be absorbed by the human body and adversely affect human health. Although most of the PM monitoring are based on ground-based observations, they are limited to point-based measurement sites, which leads to uncertainty in PM estimation for regions without observation sites. It is possible to overcome their spatial limitation by using satellite data. In this study, we developed machine learning-based retrieval algorithm for ground-level PM10 and PM2.5 concentrations using aerosol parameters from Geostationary Ocean Color Imager (GOCI) satellite and various meteorological parameters from a numerical weather prediction model during January to December of 2019. Gradient Boosted Regression Trees (GBRT) and Light Gradient Boosting Machine (LightGBM) were used to estimate PM concentrations. The model performances were examined for two types of feature sets-all input parameters (Feature set 1) and a subset of input parameters without meteorological and land-cover parameters (Feature set 2). Both models showed higher accuracy (about 10 % higher in R2) by using the Feature set 1 than the Feature set 2. The GBRT model using Feature set 1 was chosen as the final model for further analysis(PM10: R2 = 0.82, nRMSE = 34.9 %, PM2.5: R2 = 0.75, nRMSE = 35.6 %). The spatial distribution of the seasonal and annual-averaged PM concentrations was similar with in-situ observations, except for the northeastern part of China with bright surface reflectance. Their spatial distribution and seasonal changes were well matched with in-situ measurements.

주제어

표/그림 (12)

그림 Fig. 1. Study area with ground particulate matter (PM) observation sites (blue dots). The background image is elevation for study area.
표 Table 1. Two input feature sets used in the machine learning models in this study
표 Table 2. Hyperparameters of the GBRT model tested for model optimization in this study. Each element in square bracket for each hyperparameter was examined, and the final hyperparameters are shown by superscript letters
표 Table 3. Hyperparameters of the LightGBM model tested for model optimization in this study. Each element in square bracket for each hyperparameter was examined, and the final hyperparameters are shown by superscript letters
그림 Fig. 2. Scatter plots using the GBRT model results for independent test data with feature set 1 (All variables) for PM₁₀ (left) and PM_2.5 (right). Points are represented as the number of observations estimated by a Gaussian kernel density estimation. The brighter the point, the higher density of the samples.
그림 Fig. 3. Scatter plots using the GBRT model results for independent test data with feature set 2 (5 selected variables) for PM₁₀ (left) and PM_2.5 (right). Points are represented as the number of observations estimated by a Gaussian kernel density estimation. The brighter the point, the higher density of the samples.
그림 Fig. 4. Scatter plots using the LightGBM model for independent test data with feature set 1 (All variables) for PM₁₀ (left) and PM_2.5 (right). Points are represented as the number of observations estimated by a Gaussian kernel density estimation. The brighter the point, the higher density of the samples.
그림 Fig. 5. Scatter plots using the LightGBM model for independent test data with feature set 2 (5 selected variables) for PM₁₀ (left) and PM_2.5 (right). Points are represented as the number of observations estimated by a Gaussian kernel density estimation. The brighter the point, the higher density of the samples.
그림 Fig. 6. Relative variable importance of the GBRT models for PM₁₀ (left) and PM_2.5 (right).
그림 Fig. 7. Seasonal distribution of PM₁₀ concentrations estimated by the GBRT model.
그림 Fig. 8. Seasonal distribution of PM_2.5 concentrations estimated by the GBRT model.
그림 Fig. 9. Annual distribution of PM₁₀ (upper) and PM_2.5 (bottom) estimation with observed PM₁₀ and PM_2.5 concentrations, respectively.

AI 본문요약
AI-Helper

이론/모형

LightGBM 모델은 Python 3 sikit-learn 환경에서 수행되었고 제공되는 패키지인 lightgbm의 LGBMRegressor 를 이용하여 수행하였다. Table 3는 본 연구에서 테스트 된 LightGBM의 하이퍼파라미터 조합을 보여준다.
지상 PM 농도는 기상 등의 다양한 외부 요인에 의해서 영향을 받는다. 따라서 본 연구에서는 우리나라에 특화된 수치 모델인 RDAPS기반 기온(Air Temperature, Temp), 강수량(Precipitation, Precip), 상대습도(Relative Humidity, RH), 경계층고도(Planetary Boundary Layer Height, PBLH), 최대 풍속(Maximum wind speed, MaxWS) 을 입력 변수로 사용하였다. 시간에 따른 에어로졸 이류의 영향을 확인하기 위해서 1, 3, 5, 7일 동안 최대 풍속을 누적하여 만든 stack1_maxWS, stack3_maxWS, stack5_maxWS, stack7_maxWS을 추가적으로 고려해주었다.

참고문헌 (35)

Chen, Z.-Y., T.-H. Zhang, R. Zhang, Z.-M. Zhu, J. Yang, P.-Y. Chen, C.-Q. Ou, and Y. Guo, 2019. Extreme gradient boosting model to estimate PM 2.5 concentrations with missing-filled satellite data in China, Atmospheric Environment, 202: 180-189.

상세보기
Du, C., S. Liu, X. Yu, X. Li, C. Chen, Y. Peng, Y. Dong, Z. Dong, and F. Wang, 2013. Urban boundary layer height characteristics and relationship with particulate matter mass concentrations in Xi'an, central China, Aerosol and Air Quality Research, 13(5): 1598-1607.

상세보기
Friedman, J.H., 2002. Stochastic gradient boosting, Computational Statistics & Data Analysis, 38(4): 367-378.

상세보기
Geng, G., Q. Zhang, R.V. Martin, A. van Donkelaar, H. Huo, H. Che, J. Lin, and K. He, 2015. Estimating long-term PM 2.5 concentrations in China using satellite-based aerosol optical depth and a chemical transport model, Remote sensing of Environment, 166: 262-270.

상세보기
Ghotbi, S., S. Sotoudeheian, and M. Arhami, 2016. Estimating urban ground-level PM 10 using MODIS 3 km AOD product and meteorological parameters from WRF model, Atmospheric Environment, 141: 333-346.

상세보기
Gui, K., H. Che, Z. Zeng, Y. Wang, S. Zhai, Z. Wang, M. Luo, L. Zhang, T. Liao, and H. Zhao, 2020. Construction of a virtual PM 2.5 observation network in China based on high-density surface meteorological observations using the Extreme Gradient Boosting model, Environment International, 141: 105801.

상세보기
Guo, B., D. Zhang, L. Pei, Y. Su, X. Wang, Y. Bian, D. Zhang, W. Yao, Z. Zhou, and L. Guo, 2021. Estimating PM 2.5 concentrations via random forest method using satellite, auxiliary, and ground-level station dataset at multiple temporal scales across China in 2017, Science of The Total Environment, 778: 146288.

상세보기
Gupta, P. and S.A. Christopher, 2009a. Particulate matter air quality assessment using integrated surface, satellite, and meteorological products: Multiple regression approach, Journal of Geophysical Research: Atmospheres, 114(D14205): 1-13.
Gupta, P. and S.A. Christopher, 2009b. Particulate matter air quality assessment using integrated surface, satellite, and meteorological products: 2. A neural network approach, Journal of Geophysical Research: Atmospheres, 114(D20205): 1-14.
Jerrett, M., M.C. Turner, B.S. Beckerman, C.A. Pope III, A. Van Donkelaar, R.V. Martin, M. Serre, D. Crouse, S.M. Gapstur, and D. Krewski, 2017. Comparing the health effects of ambient particulate matter estimated using ground-based versus remote sensing exposure estimates, Environmental Health Perspectives, 125(4): 552-559.

상세보기
Jethva, H., O. Torres, and C. Ahn, 201., Global assessment of OMI aerosol single-scattering albedo using ground-based AERONET inversion, Journal of Geophysical Research: Atmospheres, 119(14): 9020-9040.

상세보기
Jin, K., 2018. LEO and GEO satellite programs for space-borne measurement of aerosol, Current Industrial and Technological Trends in Aerospace, 16(1): 53-62.
Ke, G., Q. Meng, T. Finley, T. Wang, W. Chen, W. Ma, Q. Ye, and T.-Y. Liu, 2017, Lightgbm: A highly efficient gradient boosting decision tree, Advances in Neural Information Processing Systems, 30: 3146-3154.
Kloog, I., M. Sorek-Hamer, A. Lyapustin, B. Coull, Y. Wang, A.C. Just, J. Schwartz, and D.M. Broday, 2015. Estimating daily PM 2.5 and PM 10 across the complex geo-climate region of Israel using MAIAC satellite-based AOD data, Atmospheric environment, 122: 409-416.

상세보기
Krasnov, H., I. Katra, and M. Friger, 2016. Increase in dust storm related PM 10 concentrations: A time series analysis of 2001-2015, Environmental Pollution, 213: 36-42.

상세보기
Kukkonen, J., M. Pohjola, R.S. Sokhi, L. Luhana, N. Kitwiroon, L. Fragkou, M. Rantamaki, E. Berge, V. Odegaard, and L.H. Slordal, 2005. Analysis and evaluation of selected local-scale PM 10 air pollution episodes in four European cities: Helsinki, London, Milan and Oslo, Atmospheric Environment, 39(15): 2759-2773.

상세보기
Lee, H.J., R.B. Chatfield, and A.W. Strawa, 2016. Enhancing the applicability of satellite remote sensing for PM 2.5 estimation using MODIS deep blue AOD and land use regression in California, United States, Environmental Science & Technology, 50(12): 6546-6555.

상세보기
Lee, P.S.-H., J.C. Park, and J.-Y. Seo, 2020. Estimation of ambient PM 10 and PM 2.5 concentrations in Seoul, South Korea, using empirical models based on MODIS and Landsat 8 OLI imagery, Korean Journal of Agricultural Science, 47(1): 59-66 (in Korean with English abstract).

원문보기 상세보기
Lim, H., M. Choi, M. Kim, J. Kim, and P. Chan, 2016. Retrieval and validation of aerosol optical properties using Japanese next generation meteorological satellite, Himawari-8, Korean Journal of Remote Sensing, 32(6): 681-691 (in Korean with English abstract).

원문보기 상세보기
Liu, Y., J.A. Sarnat, V. Kilaru, D.J. Jacob, and P. Koutrakis, 2005. Estimating ground-level PM 2.5 in the eastern United States using satellite remote sensing, Environmental Science & Technology, 39(9): 3269-3278.

상세보기
Lv, Z., W. Wei, S. Cheng, X. Han, and X. Wang, 2020, Meteorological characteristics within boundary layer and its influence on PM 2.5 pollution in six cities of North China based on WRF-Chem, Atmospheric Environment, 228: 117417.

상세보기
Ma, Z., X. Hu, L. Huang, J. Bi, and Y. Liu, 2014. Estimating ground-level PM 2.5 in China using satellite remote sensing, Environmental Science & Technology, 48(13): 7436-7444.

상세보기
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, Atmospheric Chemistry and Physics, 19(2): 1097-1113.

상세보기
Pedregosa, F., G. Varoquaux, A. Gramfort, V. Michel, B. Thirion, O. Grisel, M. Blondel, P. Prettenhofer, R. Weiss, and V. Dubourg, 2011. Scikit-learn: Machine learning in Python, The Journal of Machine Learning Research, 12: 2825-2830.

상세보기
Pope III, C.A., M. Ezzati, and D.W. Dockery, 2009, Fine-particulate air pollution and life expectancy in the United States, New England Journal of Medicine, 360(4): 376-386.

상세보기
Qadeer, K. and M. Jeon, 2019. Prediction of PM 10 Concentration in South Korea Using Gradient Tree Boosting Models, Proc. of 2019 the 3rd International Conference on Vision, Image and Signal Processing, Vancouver, CA, Aug. 26-28, pp. 1-6.
Rodriguez, S., X. Querol, A. Alastuey, G. Kallos, and O. Kakaliagou, 2001. Saharan dust contributions to PM 10 and TSP levels in Southern and Eastern Spain, Atmospheric Environment, 35(14): 2433-2447.

상세보기
She, Q., M. Choi, J.H. Belle, Q. Xiao, J. Bi, K. Huang, X. Meng, G. Geng, J. Kim, and K. He, 2020. Satellite-based estimation of hourly PM 2.5 levels during heavy winter pollution episodes in the Yangtze River Delta, China, Chemosphere, 239: 124678.

상세보기
Soni, M., S. Payra, and S. Verma, 2018. Particulate matter estimation over a semi arid region Jaipur, India using satellite AOD and meteorological parameters, Atmospheric Pollution Research, 9(5): 949-958.

상세보기
Sorek-Hamer, M., I. Kloog, P. Koutrakis, A.W. Strawa, R. Chatfield, A. Cohen, W.L. Ridgway, and D.M. Broday, 2015. Assessment of PM 2.5 concentrations over bright surfaces using MODIS satellite observations, Remote Sensing of Environment, 163: 180-185.

상세보기
Stirnberg, R., J. Cermak, S. Kotthaus, M. Haeffelin, H. Andersen, J. Fuchs, M. Kim, J.-E. Petit, and O. Favez, 2021. Meteorology-driven variability of air pollution (PM 1 ) revealed with explainable machine learning, Atmospheric Chemistry and Physics, 21(5): 3919-3948.

상세보기
Wang, S., W. Yuan, and K. Shang, 2006. The impacts of different kinds of dust events on PM 10 pollution in northern China, Atmospheric Environment, 40(40): 7975-7982.

상세보기
Xu, Y., H.C. Ho, M.S. Wong, C. Deng, Y. Shi, T.-C. Chan, and A. Knudby, 2018. Evaluation of machine learning techniques with multiple remote sensing datasets in estimating monthly concentrations of ground-level PM 2.5 , Environmental Pollution, 242: 1417-1426.

상세보기
Zhan, Y., Y. Luo, X. Deng, H. Chen, M.L. Grieneisen, X. Shen, L. Zhu, and M. Zhang, 2017. Spatiotemporal prediction of continuous daily PM 2.5 concentrations across China using a spatially explicit machine learning algorithm, Atmospheric Environment, 155: 129-139.

상세보기
Zhang, T., W. He, H. Zheng, Y. Cui, H. Song, and S. Fu, 2021. Satellite-based ground PM 2.5 estimation using a gradient boosting decision tree, Chemosphere, 268: 128801.

상세보기

표제어: PCR

동의어: Packet Collision Rate

용어 설명 출처 목록 (6)

용어 설명: PCR은 세균 특이성이 있는 primer를 이용하여 적은 수의 세균이 있을지라도 쉽게 검출할 수 있는 유용한 방법이며, 이를 이용하여 구강 내 치면세균막이나 타액에서 직접 세균을 검출할 수 있게 되었다[8].

내보내기 구분	파일저장 인쇄 메일전송
구성항목	기본정보 상세정보 관리번호, 논문명, 저널/프로시딩명, 저자 , 발행년, 권, 호, 시작페이지, 끝페이지, 발행기관 관리번호, 논문명, 대등논문명, 저자 , 저널/프로시딩명, 발행기관, 발행년, 발행언어, 권, 호, 시작페이지, 끝페이지, ISBN, ISSN, 주제분야, 키워드, 초록(한글), 초록(영문), 저자(소속기관)
저장형식	Text(ASCII format) Excel format RefWorks Direct Export RIS format (for Reference Manager, ProCite, EndNote), Scholar's Aids, Mendeley
메일정보	받는사람 (필수) @ 보내는사람 (선택) @ 제목 내용 KISTI 검색결과 이메일 서비스
안내	총 건의 자료가 검색되었습니다. 다운받으실 자료의 인덱스를 입력하세요. (1-10,000) 검색결과의 순서대로 최대 10,000건 까지 다운로드가 가능합니다. 데이타가 많을 경우 속도가 느려질 수 있습니다.(최대 2~3분 소요) 다운로드 파일은 UTF-8 형태로 저장됩니다. 파일의 내용이 제대로 보이지 않을실 때는 웹브라우저 상단의 보기 -> 인코딩 -> 자동선택 여부를 확인하십시오. ~ Text(ASCII format) Excel format

연합인증