基于机器学习的二氧化碳卫星反演方法

Abstract
Figure/Table
References
Related Citation (15)

Download: PDF (949 KB) HTML (1 KB)
Export: BibTeX | EndNote (RIS)

Abstract For the physical retrieval method of CO₂ from satellite observation in short-wave infrared band, the iterative inverse process is prone to non-convergence due to the nonlinear radiation transfer equation and the difficulty in accurately obtaining atmospheric state parameters in actual observation, which makes it impossible to obtain accurate CO₂ retrieval. Based on the advantages of machine learning model in data analysis and prediction learning, a new retrieval method using machine learning was proposed. The training data set was constructed including observed radiance, aerosol optical thickness and temperature. The feedforward neural network and quantitative conjugate gradient algorithm are used for training and learning to develop the new retrieval model. Then, the CO₂ concentration was retrieved using GOSAT observations. Validation of satellite retrieval is made by comparing with measurements from TCCON sites. Comparison shows that the correlation was better than 0.86, and the mean bias is less than 2.5x10^-6, which prove effectivity of the new retrieval method.

Key words： CO₂ satellite remote sensing machine learning neural network

Received: 29 March 2023

PACS:

X511

	Service

	E-mail this article
	Add to my bookshelf
	Add to citation manager
	E-mail Alert
	RSS
	Articles by authors
	MIAO Yun-fei
	ZOU Ming-min
	SHENG Shu-li
	ZHU Ke-wei
	DING Wen-qiao
	LIN Jun-jie
	QU Zheng
	LI Da-cheng

Cite this article:

MIAO Yun-fei,ZOU Ming-min,SHENG Shu-li等. CO₂ satellite inversion methocl based on machine learning[J]. CHINA ENVIRONMENTAL SCIENCECE, 2023, 43(S1): 20-27.

URL:

http://www.zghjkx.com.cn/EN/ OR http://www.zghjkx.com.cn/EN/Y2023/V43/IS1/20

[1] 沈永平,王国亚.IPCC第一工作组第五次评估报告对全球气候变化认知的最新科学要点[J].冰川冻土, 2013,35(5):1068-1076. Shen Yongping, Wang Guoya. Recent scientific highlights on global climate change understanding from the Fifth Assessment Report of IPCC Working Group I [J].Journal of Glaciology and Geocryology, 2013,35(5):1068-1076.
[2] 侯炜烨,金佳鑫,严涛,等.全球大气二氧化碳浓度2°x2.5°栅格模拟数据集(1992~2020) [J].全球变化数据学报(中英文), 2022,6(2): 191-199,363-371. Hou Weiye, Jin Jiaxin, Yan Tao, et al. Global atmospheric carbon dioxide concentration 2°x2.5° raster simulation dataset (1992-2020) [J].Journal of Global Change Data, 2002,6(2):191-199,363-371.
[3] Meng Q W, Qiu Y. Remote sensing analysis of multi-years spatial and temporal variation of CO₂ in China [J].Remote Sensing Technology and Application, 2016,31(2):203‒213.
[4] Menzel, W P, Schmit, T J, Zhang P, et al. Satellite-based atmospheric infrared sounder development and applications [J].Bulletin of the American Meteorological Society, 2018,99(3):583-603.
[5] Liu Y, Lv D R, Chen H B, et al., Advances in technologies and methods for satellite remote sensing of atmospheric CO₂ [J].Remote Sensing Technology and Application, 2011,26(2):247‒254.
[6] Thompson.O E, Hou Y-T, Dazlich D D. The ill-posed nature of the satellite temperature retrieval problem and the limits of retrievability [J].Journal of Atmospheric and Oceanic Technology, 1986,3:643-649.
[7] 雷连发,卢建平,朱磊,等.多通道地基微波辐射计大气遥感[J].遥感学报, 2014,18(1):180-191. Lei Lianfa, Lu Jianping, Zhu Lei, et al. Multi-channel ground-based microwave radiometer atmospheric remote sensing [J].Journal of remote sensing, 2014,17(1):180-191.
[8] 李云飞,李军,贺霖.2022.单样本对卷积神经网络遥感图像时空融合.遥感学报, 26(8):1614-1623. Li Y F, Li J, He L. Spatiotemporal fusion of remote sensing images with single sample pairs of convolutional neural networks. Journal of Remote Sensing, 2022,26(8):1614-1623.
[9] 彭豪杰,周杨,胡校飞,等.基于深度学习与随机森林的PM_2.5浓度预测模型.遥感学报, 2023,27(2):430-440. Peng H J, Zhou Y, Hu Z F, et al Prediction model of PM_2.5 concentration based on deep learning and random forest. Journal of Remote Sensing, 2023,27(2):430-440.
[10] Hamrani A, Akbarzadeh A, Madramootoo C A, et al. Machine learning for predicting greenhouse gas emissions from agricultural soils [J].Science of the Total Environment, 2020,741:140338.
[11] Zhang X, Yan F, Liu H, et al. Towards low carbon cities: A machine learning method for predicting urban blocks carbon emissions (UBCE) based on built environment factors (BEF) in Changxing City, China [J].Sustainable Cities and Society, 2021,69:102875.
[12] Zhu B, Ye S, Jiang M, et al. Achieving the carbon intensity target of China: A least squares support vector machine with mixture kernel function approach [J].Applied Energy, 2019,233:196-207.
[13] Wang H W, Li X B, Wang D, et al. Regional prediction of ground-level ozone using a hybrid sequence-to-sequence deep learning approach [J].Journal of Cleaner Production, 2020,253:119841.
[14] Shabani E, Hayati B, Pishbahar E, et al. A novel approach to predict CO₂ emission in the agriculture sector of Iran based on Inclusive Multiple Model [J].Journal of Cleaner Production, 2021,279:123708.
[15] Chen X F, de Leeuw G, Arola A, et al. Joint retrieval of the aerosol fine mode fraction and opti-cal depth using MODIS spectral reflectance over northern and eastern China: artificial neural network method. Remote Sensing of Environment, 2020,249:112006.
[16] Ranganathan A. The levenberg-marquardt algorithm [J].Tutoral on LM algorithm, 2004,11(1):101-110.
[17] Caballero J, Fernández M. Artificial neural networks from MATLAB® in medicinal chemistry. Bayesian-regularized genetic neural networks (BRGNN): Application to the prediction of the antagonistic activity against human platelet thrombin receptor (PAR-1) [J].Current topics in medicinal chemistry, 2008,8(18):1580-1605.
[18] Møller M F. A scaled conjugate gradient algorithm for fast supervised learning [J].Neural networks, 1993,6(4):525-533.
[19] Wennberg P O, Roehl C M, Wunch D, et al. 2014. TCCON data from Park Falls, Wisconsin, USA, Release GGG2014RO.TCCON data archive [Z].hosted by the Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee USA, http: //ldx.dloi.orgl10.14291/tccon.ggg2014.parkfallsO1.R0/1149161.
[20] Wennberg P O, Wunch D, Roehl C M, et al. 2014. TCCON data from Lamont, Oklahoma, USA, Release GGG2014RO [Z].TCCON data archive, hosted by the Carbon Dioxide Information Analysis Center, Oak Ridge National Laboratory, Oak Ridge, Tennessee, USA, http:/dx.doi.org/10.14291/tccon.ggg2014.lamont01.RO/1149159.
[21] Yoshida Y, Ota Y, Eguchi N, et al. Retrieval algorithm for CO₂ and CH₄ column abundances from short-wavelength spectral observation by the Greenhouse gases observing satellite [J].Atmos.Meas.Tech., 2010,3(6):4791-4833.
[22] C.W.O’Dell, B. Connor, H. Bosh, et al. The ACOS CO₂ retrieval algorithm-Part 1: Description and validation against synthetic observations [J].Atmos. Meas. Tech., 5,99-121,2012.
[23] 马良玉,左晓桐,王永军.神经网络在风烟系统含氧量软测量中的应用[J].自动化仪表, 2019,40(3):12-16,22. Ma Liangyu, Zuo Xiaotong, Wang Yongjun. Neural network in the specific application of oxygen content of soft measurement system [J].Automation instrument, 2019,40(3):12-16,22.