Fill the Gaps of Eddy Covariance Fluxes Using Machine Learning Algorithms

  • Shaoying WANG ,
  • Yu ZHANG ,
  • Xianhong MENG ,
  • Minhong SONG ,
  • Lunyu SHANG ,
  • Youqi SU ,
  • Zhaoguo LI
Expand
  • <sup>1.</sup>Northwest Institute of Eco-Environment and Resources,CAS/Key Laboratory of Land Surface Process and Climate Change in Cold and Arid Regions,CAS,Lanzhou 730000,Gansu,China;<sup>2.</sup>Chengdu University of Information Technology/School of Atmospheric Sciences,Chengdu 610225,Sichuan,China;<sup>3.</sup>University of Chinese Academy of Sciences,Beijing 100049,China

Received date: 2019-05-18

  Online published: 2020-12-28

Abstract

The eddy covariance long-term measurements commonly include data gaps due to system failures, quality control and quality assurance.In this study, the marginal distribution sampling (MDS) algorithm and three machine learning algorithms (random forest RF, support vector machine SVM and artificial neural networks ANN) were applied to fill the gaps of sensible heat flux (H), latent heat flux (LE) and net ecosystem exchange in 2016 over an alpine ecosystem.Results indicate that the performance of RF is better than SVM and ANN.During the nighttime, the periods of sunrise and sunset, and in the winter and spring, the performance of three machine learning algorithms is relatively weak, compared to other periods or seasons.On the monthly and annual scales, the filled NEE budget is significantly influenced by the choice of gap-filling method, compared to H and LE.

Cite this article

Shaoying WANG , Yu ZHANG , Xianhong MENG , Minhong SONG , Lunyu SHANG , Youqi SU , Zhaoguo LI . Fill the Gaps of Eddy Covariance Fluxes Using Machine Learning Algorithms[J]. Plateau Meteorology, 2020 , 39(6) : 1348 -1360 . DOI: 10.7522/j.issn.1000-0534.2019.00142

References

[1]Agarwal D, Pastorello G, Poindexter C, et al, 2014.The data post-processing pipeline for AmeriFlux data products[C].AGU Fall Meeting Abstracts.
[2]Baldocchi D, 2014.Measuring fluxes of trace gases and energy between ecosystems and the atmosphere-the state and future of the eddy covariance method[J].Global Change Biology, 20(12): 3600-3609.
[3]Beringer J, McHugh I, Hutley L B, et al, 2017.Technical note: Dynamic integrated gap-filling and partitioning for OzFlux (DINGO)[J].Biogeosciences, 14(6): 1457-1460.
[4]Breiman L, 2001.Random forests[J].Machine Learning, 45(1): 5-32.
[5]Cortes C, Vapnik V, 1995.Support-vector networks[J].Machine Learning, 20(3): 273-297.
[6]Dengel S, Zona D, Sachs T, et al, 2013.Testing the applicability of neural networks as a gap-filling method using CH4 flux data from high latitude wetlands[J].Biogeosciences, 10(12): 8185-8200.
[7]Falge E, Baldocchi D, Olson R, et al, 2001.Gap filling strategies for defensible annual sums of net ecosystem exchange[J].Agricultural and Forest Meteorology, 107(1): 43-69.
[8]Foltynová L, Fischer M, McGloin R P, 2020.Recommendations for gap-filling eddy covariance latent heat flux measurements using marginal distribution sampling[J].Theoretical and Applied Climatology, 139(1): 677-688.
[9]J?nsson P, Eklundh L, 2004.TIMESAT—a program for analyzing time-series of satellite sensor data[J].Computers & Geosciences, 30(8): 833-845.
[10]Hui D, Wan S, Su B, et al, 2004.Gap-filling missing data in eddy covariance measurements using multiple imputation (MI) for annual estimations[J].Agricultural and Forest Meteorology, 121(1/2): 93-111.
[11]Kim Y, Johnson M S, Knox S H, et al, 2020.Gap-filling approaches for eddy covariance methane fluxes: A comparison of three machine learning algorithms and a traditional method with principal component analysis[J].Global Change Biology, 26(3): 1499-1518.
[12]Knox S H, Jackson R B, Poulter B, et al, 2019.FLUXNET-CH4 synthesis activity: objectives, observations, and future directions[J].Bulletin of the American Meteorological Society, 100(12): 2607-2632.
[13]Knox S H, Sturtevant C, Matthes J H, et al, 2015.Agricultural peatland restoration: Effects of land-use change on greenhouse gas (CO2 and CH4) fluxes in the Sacramento-San Joaquin Delta[J].Global Change Biology, 21(2): 750-765.
[14]Lee X, Massman W, Law B, 2004.Handbook of micrometeorology: A guide for surface flux measurement and analysis[M].Boston: Springer Science & Business Media.
[15]Mauder M, Foken T, Clement R, et al, 2008.Quality control of CarboEurope flux data-Part 2: Inter-comparison of eddy-covariance software[J].Biogeosciences, 5: 451-462.
[16]Moffat A M, Papale D, Reichstein M, et al, 2007.Comprehensive comparison of gap-filling techniques for eddy covariance net carbon fluxes[J].Agricultural and Forest Meteorology, 147(3-4): 209-232.
[17]Ooba M, Hirano T, Mogami J, et al, 2006.Comparisons of gap-filling methods for carbon flux dataset: A combination of a genetic algorithm and an artificial neural network[J].Ecological Modelling, 198(3/4): 473-486.
[18]Papale D, Reichstein M, Aubinet M, et al, 2006.Towards a standardized processing of Net Ecosystem Exchange measured with eddy covariance technique: algorithms and uncertainty estimation[J].Biogeosciences, 3: 571-583.
[19]Pastorello G, Trotta C, Canfora E, et al, 2020.The FLUXNET2015 dataset and the ONEFlux processing pipeline for eddy covariance data[J].Scientific Data, 7: 225.
[20]Reichstein M, Falge E, Baldocchi D, et al, 2005.On the separation of net ecosystem exchange into assimilation and ecosystem respiration: review and improved algorithm[J].Global Change Biology, 11(9): 1424-1439.
[21]Soloway A D, Amiro B D, Dunn A L, et al, 2017.Carbon neutral or a sink?Uncertainty caused by gap-filling long-term flux measurements for an old-growth boreal black spruce forest[J].Agricultural and Forest Meteorology, 233: 110-121.
[22]Wang S Y, Zhang Y, Lü S H, et al, 2016.Biophysical regulation of carbon fluxes over an alpine meadow ecosystem in the eastern Tibetan Plateau[J].International Journal of Biometeorology, 60(6): 801-812.
[23]Wutzler T, Lucas-Moffat A, Migliavacca M, et al, 2018.Basic and extensible post-processing of eddy covariance flux data with REddyProc[J].Biogeosciences, 15(16): 5015-5030.
[24]Xu T, Guo Z, Liu S, et al, 2018.Evaluating Different Machine Learning Methods for Upscaling Evapotranspiration from Flux Towers to the Regional Scale[J].Journal of Geophysical Research: Atmospheres, 123(16): 8674-8690.
[25]李春, 何洪林, 刘敏, 等, 2008.ChinaFLUX CO2 通量数据处理系统与应用[J].地球信息科学, 10(5): 557-565.
[26]刘立新, 董云社, 齐玉春, 2004.草地生态系统土壤呼吸研究进展[J].地理科学进展, 23(7): 35-42.
[27]刘绍辉, 方精云, 1997.土壤呼吸的影响因素及全球尺度下温度的影响[J].生态学报, 17(5): 469-476.
[28]吕少宁, 文军, 张宇, 等, 2012.不同平均时间对 LOPEX10 资料涡动相关湍流通量计算结果影响的探讨[J].高原气象, 31(6): 1530-1538.
[29]王少影, 张宇, 吕世华, 等, 2009.金塔绿洲湍流资料的质量控制研究[J].高原气象, 28(6): 1260-1273.
[30]王少影, 张宇, 吕世华, 等, 2012.玛曲高寒草甸地表辐射与能量收支的季节变化[J].高原气象, 31(3): 605-614.
[31]徐自为, 刘绍民, 宫丽娟, 等, 2008.涡动相关仪观测数据的处理与质量评价研究[J].地球科学进展, 23(4): 357-370.
[32]于贵瑞, 陈智, 张雷明, 等, 2017.通量观测事业的科学使命再认知——为解决全球可持续发展的生态学问题奠定坚实的数据基础[J].资源与生态学报, 8(2): 115-120.
[33]庄金鑫, 王维真, 王介民, 2013.涡动相关通量计算及三种主要软件的比较分析[J].高原气象, 32(1): 78-87.DOI: 10.7522/j.issn.1000-0534.2012.00009.
Outlines

/