A general regression neural network approach to reconstruct global 1°×1° resolution sea surface <i>p</i>CO<sub>2</sub>

Zhong Guorong; Li Xuegang; Qu Baoxiao; Wang Yanjun; Yuan Huamao; Song Jinming

doi:10.3969/j.issn.0253-4193.2020.10.007

Volume 42 Issue 10

Nov. 2020

Turn off MathJax

Article Contents

Article Navigation > Haiyang Xuebao > 2020 > 42(10): 70-79

Zhong Guorong，Li Xuegang，Qu Baoxiao, et al. A general regression neural network approach to reconstruct global 1°×1° resolution sea surface pCO2[J]. Haiyang Xuebao，2020, 42(10)：70–79 doi: 10.3969/j.issn.0253-4193.2020.10.007

Citation:

Zhong Guorong，Li Xuegang，Qu Baoxiao, et al. A general regression neural network approach to reconstruct global 1°×1° resolution sea surface p CO₂[J]. Haiyang Xuebao，2020, 42(10)：70–79 doi: 10.3969/j.issn.0253-4193.2020.10.007

Citation:

PDF( 33365 KB)

A general regression neural network approach to reconstruct global 1°×1° resolution sea surface pCO₂

doi: 10.3969/j.issn.0253-4193.2020.10.007

Zhong Guorong^{1, 2, 3, 4
,},
Li Xuegang^{1, 2, 3, 4
,
,},
Qu Baoxiao^{1, 3, 4},
Wang Yanjun⁴,
Yuan Huamao^{1, 2, 3, 4},
Song Jinming^{1, 2, 3}

1.
Key Laboratory of Marine Ecology and Environmental Sciences, Institute of Oceanology, Chinese Academy of Sciences, Qingdao 266071, China
2.
University of Chinese Academy of Sciences, Beijing 100049, China
3.
Marine Ecology and Environmental Science Laboratory, Pilot National Laboratory for Marine Science and Technology (Qingdao), Qingdao 266237, China
4.
Center for Ocean Mega-Science, Chinese Academy of Sciences, Qingdao 266071, China

Received Date: 2019-12-29
Rev Recd Date: 2020-03-23

Available Online: 2020-11-13

Publish Date: 2020-10-25

Abstract

Abstract

Sea surface partial pressure of carbon dioxide (pCO₂) is a crucial parameter for estimating ocean carbon source and sink term, but its sparse and uneven in situ measurements in space and time lead to large uncertainty in the estimate of sea-air CO₂ flux and characteristics of ocean carbon source and sink. To eliminate this uncertainty, a general regression neural network approach using the Surface Ocean CO₂ Atlas (SOCAT) dataset, based on the non-liner regression of pCO₂ and longitude, latitude, time, temperature, salinity and concentration of chlorophyll, was successfully used in the reconstruction of global 1°×1° resolution monthly sea surface pCO₂ from 1998 to 2018, with a root mean square error (RMSE) of 16.93 μatm and a mean relative error (MRE) of 2.97%, lower than existing feed-forward neural network (FFNN), self-organizing neural network (SOM) and machine learning approaches. The global distribution of pCO₂ obtained by this approach agrees well with existing researches.
- general regression neural network,
- sea surface pCO₂,
- global ocean grid data

FullText(HTML)

References(24)

References

[1]	曲宝晓, 宋金明, 袁华茂, 等. 东海海−气界面二氧化碳通量的季节变化与控制因素研究进展[J]. 地球科学进展, 2013, 28(7): 783−793. doi: 10.11867/j.issn.1001-8166.2013.07.0783 Qu Baoxiao, Song Jinming, Yuan Huamao, et al. Advances of seasonal variations and controlling factors of the air-sea CO₂ flux in the East China Sea[J]. Advances in Earth Science, 2013, 28(7): 783−793. doi: 10.11867/j.issn.1001-8166.2013.07.0783
[2]	Takahashi T, Sutherland S C, Feely R A, et al. Decadal change of the surface water pCO₂ in the North Pacific: a synthesis of 35 years of observations[J]. Journal of Geophysical Research: Oceans, 2006, 111(C7): C07S05.
[3]	Song Jinming. Biogeochemical Processes of Biogenic Elements in China Marginal Seas[M]. Berlin, Heidelberg: Springer Science & Business Media, 2010: 140−144.
[4]	宋金明, 李学刚, 袁华茂, 等. 渤黄东海生源要素的生物地球化学[M]. 北京: 科学出版社, 2019: 45-47. Song Jinming, Li Xuegang, Yuan Huamao, et al. Biogeochemistry of Biogenic Elements in the Bohai Sea, Yellow Sea and East China Sea[M]. Beijing: Science Press, 2019: 45−47.
[5]	Takahashi T, Sutherland S C, Wanninkhof R, et al. Climatological mean and decadal change in surface ocean pCO₂, and net sea-air CO₂ flux over the global oceans[J]. Deep-Sea Research Part II: Topical Studies in Oceanography, 2009, 56(8/10): 554−577.
[6]	Takamura T R, Inoue H Y, Midorikawa T, et al. Seasonal and inter-annual variations in $ {{p{\rm{CO}}}}_2^{{\rm{sea}}} $ and air-sea CO₂ fluxes in mid-latitudes of the western and eastern North Pacific during 1999−2006: recent results utilizing voluntary observation ships[J]. Journal of the Meteorological Society of Japan, 2010, 88(6): 883−898. doi: 10.2151/jmsj.2010-602
[7]	Sarma V V S S, Saino T, Sasaoka K, et al. Basin-scale pCO₂ distribution using satellite sea surface temperature, Chl a, and climatological salinity in the North Pacific in spring and summer[J]. Global Biogeochemical Cycles, 2006, 20(3): GB3005.
[8]	Lefévre N, Watson A J, Watson A R. A comparison of multiple regression and neural network techniques for mapping in situ pCO₂ data[J]. Tellus B: Chemical and Physical Meteorology, 2005, 57(5): 375−384. doi: 10.1111/j.1600-0889.2005.00164.x
[9]	Zeng J Y, Nojiri Y, Nakaoka S I, et al. Surface ocean CO₂ in 1990−2011 modelled using a feed-forward neural network[J]. Geoscience Data Journal, 2015, 2(1): 47−51. doi: 10.1002/gdj3.26
[10]	Zeng J, Nojiri Y, Landschützer P, et al. A global surface ocean fCO₂ climatology based on a feed-forward neural network[J]. Journal of Atmospheric and Oceanic Technology, 2014, 31(8): 1838−1849. doi: 10.1175/JTECH-D-13-00137.1
[11]	Telszewski M, Chazottes A, Schuster U, et al. Estimating the monthly pCO₂ distribution in the North Atlantic using a self-organizing neural network[J]. Biogeosciences, 2009, 6(8): 1405−1421. doi: 10.5194/bg-6-1405-2009
[12]	Nakaoka S, Telszewski M, Nojiri Y, et al. Estimating temporal and spatial variation of ocean surface pCO₂ in the North Pacific using a self-organizing map neural network technique[J]. Biogeosciences, 2013, 10(9): 6093−6106. doi: 10.5194/bg-10-6093-2013
[13]	Denvil-Sommer A, Gehlen M, Vrac M, et al. LSCE-FFNN-v1: a two-step neural network model for the reconstruction of surface ocean pCO₂ over the global ocean[J]. Geoscientific Model Development, 2019, 12(5): 2091−2105. doi: 10.5194/gmd-12-2091-2019
[14]	Körtzinger A. Determination of carbon dioxide partial pressure (p(CO₂))[M]//Grasshoff K, Kremling K, Ehrhardt M. Methods of Seawater Analysis. 3rd ed. New York: Wiley, 1999: 149−158.
[15]	Specht D F. A general regression neural network[J]. IEEE Transactions on Neural Networks, 1991, 2(6): 568−576. doi: 10.1109/72.97934
[16]	陈明. MATLAB神经网络原理与实例精解[M]. 北京: 清华大学出版社, 2013: 208−237. Chen Ming. MATLAB Neural Network Principle and Example Fine Solution[M]. Beijing: Tsinghua University Press, 2013: 208−237.
[17]	Chen Shuangliang, Hu Chuanmin, Barnes B B, et al. A machine learning approach to estimate surface ocean pCO₂ from satellite measurements[J]. Remote Sensing of Environment, 2019, 228: 203−226. doi: 10.1016/j.rse.2019.04.019
[18]	Friedrich T, Oschlies A. Neural network-based estimates of North Atlantic surface pCO₂ from satellite data: a methodological study[J]. Journal of Geophysical Research: Oceans, 2009, 114(C3): C03020.
[19]	Hales B, Strutton P G, Saraceno M, et al. Satellite-based prediction of pCO₂ in coastal waters of the eastern North Pacific[J]. Progress in Oceanography, 2012, 103: 1−15. doi: 10.1016/j.pocean.2012.03.001
[20]	Landschützer P, Gruber N, Bakker D C E, et al. A neural network-based estimate of the seasonal to inter-annual variability of the Atlantic Ocean carbon sink[J]. Biogeosciences, 2013, 10(11): 7793−7815. doi: 10.5194/bg-10-7793-2013
[21]	Laruelle G G, Landschützer P, Gruber N, et al. Global high-resolution monthly pCO₂ climatology for the coastal ocean derived from neural network interpolation[J]. Biogeosciences, 2017, 14(19): 4545−4561. doi: 10.5194/bg-14-4545-2017
[22]	Zeng Jiye, Matsunaga T, Saigusa N, et al. Technical note: Evaluation of three machine learning models for surface ocean CO₂ mapping[J]. Ocean Science, 2017, 13(2): 303−313. doi: 10.5194/os-13-303-2017
[23]	Landschützer P, Gruber N, Bakker D C E. Decadal variations and trends of the global ocean carbon sink[J]. Global Biogeochemical Cycles, 2016, 30(10): 1396−1417. doi: 10.1002/2015GB005359
[24]	Takahashi T, Sutherland S C, Chipman D W, et al. Climatological distributions of pH, pCO₂, total CO₂, alkalinity, and CaCO₃ saturation in the global surface ocean, and temporal changes at selected locations[J]. Marine Chemistry, 2014, 164: 95−125. doi: 10.1016/j.marchem.2014.06.004