Advanced Search
Article Contents

Ocean–atmosphere Teleconnections Play a Key Role in the Interannual Variability of Seasonal Gross Primary Production in China


doi: 10.1007/s00376-021-1226-4

  • Since the 1950s, the terrestrial carbon uptake has been characterized by interannual variations, which are mainly determined by interannual variations in gross primary production (GPP). Using an ensemble of seven-member TRENDY (Trends in Net Land–Atmosphere Carbon Exchanges) simulations during 1951–2010, the relationships of the interannual variability of seasonal GPP in China with the sea surface temperature (SST) and atmospheric circulations were investigated. The GPP signals that mostly relate to the climate forcing in terms of Residual Principal Component analysis (hereafter, R-PC) were identified by separating out the significant impact from the linear trend and the GPP memory. Results showed that the seasonal GPP over China associated with the first R-PC1 (the second R-PC2) during spring to autumn show a monopole (dipole or tripole) spatial structure, with a clear seasonal evolution for their maximum centers from springtime to summertime. The dominant two GPP R-PC are significantly related to Sea Surface Temperature (SST) variability in the eastern tropical Pacific Ocean and the North Pacific Ocean during spring to autumn, implying influences from the El Niño–Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO). The identified SST and circulation factors explain 13%, 23% and 19% of the total variance for seasonal GPP in spring, summer and autumn, respectively. A clearer understanding of the relationships of China’s GPP with ocean–atmosphere teleconnections over the Pacific and Atlantic Ocean should provide scientific support for achieving carbon neutrality targets.
    摘要: 自1950年来,在初级生产力(GPP)年际变率影响下,陆地碳汇呈现显著的年际变化。在区域尺度上,这种年际变率的主要来源——特别是海气遥相关过程,仍不甚清楚。基于TRENDY陆-气间碳交换趋势比较计划的7个模式,时间跨度为1951至2010年,本文研究了中国各季GPP的年际变化与海表温度和大气环流之间的关系。通过分离受长期趋势和GPP记忆影响的部分,本文首先将春夏秋各个季节的GPP残差部分识别为更受气候强迫影响的信号。结果表明,春季至秋季的第一(第二)GPP残差模态呈现单极型(偶极型或三极型)的空间分布,中心位置随季节推移而变化。春夏秋各季的前两个GPP残差模态与赤道东太平洋海温和北太平洋海温显著相关,表明了厄尔尼诺-南方涛动和太平洋年代际涛动对中国GPP的重要影响。此外,大气环流与GPP的关系特征显示,北极涛动、西太平洋涛动、贝加尔湖阻塞以及西太平洋副热带高压分别对春夏秋三季、春季、夏季和秋季的中国大陆GPP有着重要影响。以上识别的海温及环流因子可分别解释春季13%,夏季23%和秋季19%的GPP季节平均的总方差。特别指出,以上因子在春夏秋季,主要与中国中部、西南部、东北部和南部的GPP变化有关,而这些区域主导了中国大陆GPP的年际变化。对中国GPP及其在太平洋和大西洋上海气遥相关背景的深入理解,将为有效实现中国碳循环的估算和预估,并为碳中和目标的实现提供科学依据。
  • 加载中
  • Figure 1.  (a) Temporal variations of China’s GPP for all the four seasons (black) and the total of the three seasons of spring-summer-autumn (red) (units: KgC m−2); and spatial distributions of the (b–d) total variance [units: (KgC m−2 month−1)2] of the seasonal mean GPP over China (e–g) residual variance [units: (KgC m−2 month−1)2] of the seasonal mean GPP from the trend and GPP memory, and (h–j) ratio of the residual variance to the total variance of GPP over China (units: %) for spring (left), summer (middle) and autumn (right), respectively.

    Figure 2.  Spatial distributions of the two dominant EOF modes of the total seasonal mean GPP field (with explained variance in brackets) for (a, g) spring, (c, i) summer, and (e, k) autumn, respectively; and temporal variations of the total PC time series (T-PC; black) and the residual PC time series from the trend and GPP memory (R-PC; red) for (b, h) spring, (d, j) summer, and (f, l) autumn, respectively. Noted the R-PC2 coincide with the T-PC2 in spring and autumn

    Figure 3.  Correlation maps of contemporary GPP (left-hand column), SST (middle column) and 500-hPa geopotential height (right-hand column) associated with the (a–c) spring R-PC1, (d–f) summer R-PC1, (g–i) autumn R-PC1, (j–l) spring R-PC2, (m–o) summer R-PC2 and (p–r) autumn R-PC2, respectively. The shaded areas in the correlation maps are significant at the 95% confidence level, using the Student’s t-test.

    Figure 4.  Correlation maps of contemporary soil moisture (left-hand column), precipitation (middle column) and temperature (right-hand column) associated with the (a–c) spring R-PC1, (d–f) summer R-PC1, (g– i) autumn R-PC1, (j– l) spring R-PC2, (m–o) summer R-PC2 and (p–r) autumn R-PC2, respectively. The shaded areas in the correlation maps are significant at the 95% confidence level, using the Student’s t-test.

    Figure 5.  Spatial distributions of the fraction of variance of GPP explained by the SST and circulation factors, for (a) spring, (b) summer and (c) autumn, respectively.

    Table 1.  Seven trendy models used in this study.

    Model nameAbbreviationSpatial resolution (lat × lon)Land surface modelFull nitrogen cycleFire simulationHarve fluxSource
    Community LandCLM4C0.5° × 0.5°YesNoYesNoOleson et al. (2010)
    Community LandCLM4CN0.5° × 0.5°YesYesYesNoOleson et al. (2010)
    Lund–Potsdam–JenaLPJ0.5° × 0.5°NoNoYesYesSitch et al. (2003)
    LPJ-GUESSLPJ-GUESS0.5° × 0.5°NoNoYesNoSmith et al. (2001)
    Sheffield-DGVMSDGVM3.75° × 2.5°NoNoYesNoWoodward et al. (1995)
    TRIFFIDTRI3.75° × 2.5°YesNoNoNoHughes et al. (2006)
    CABLECABLE0.5° × 0.5°NoYesNoNoWang et al. (2011)
    DownLoad: CSV

    Table 2.  Total variability of the mean seasonal variations in GPP pattern [leftmost column; (KgC m−2 month−1)2], the residual variability from the trend and the GPP memory of the seasonal mean GPP pattern [second column; (KgC m−2 month−1)2], and the percentage of the residual to total variability (third column; %).

    SeasonT- variabilityR- variabilityR/T variability
    spring0.0610.03964
    summer0.1950.11760
    autumn0.1200.07865
    DownLoad: CSV

    Table 3.  Correlation coefficients between the GPP R-PC and the contemporary (third column) and lead-time (fifth column) climate indices. The years used in the climate indices are denoted by (1) for the preceding year.

    GPP R-PCClimate indicesCorrelation (contemporary)Climate indicesCorrelation (lead-lag)
    spring R-PC1spring AO0.45***Feb AO0.43***
    spring PDO−0.36**spring(1) PDO−0.33**
    spring R-PC2spring Niño-3.40.28*Feb Niño-3.40.32*
    summer R-PC1summer AO0.26*May AO0.30*
    summer PDO−0.34**spring PDO−0.31*
    summer R-PC2summer Niño-3.40.28*D(1)JF Niño-3.4−0.34**
    autumn R-PC1autumn AO0.25**Aug AO0.25*
    autumn PDO−0.28*summer PDO−0.28*
    autumn R-PC2autumn Niño-3.40.47***summer Niño-3.40.49***
    *0.05, **0.01, *** 0.001
    DownLoad: CSV
  • Ahlström, A., and Coauthors, 2015: The dominant role of semi-arid ecosystems in the trend and variability of the land CO2 sink. Science, 348, 895−899, https://doi.org/10.1126/science.aaa1668.
    Barman, R., A. K. Jain, and M. L. Liang, 2014: Climate-driven uncertainties in modeling terrestrial gross primary production: A site level to global-scale analysis. Global Change Biology, 20, 1394−1411, https://doi.org/10.1111/gcb.12474.
    Beer, C., and Coauthors, 2010: Terrestrial gross carbon dioxide uptake: Global distribution and Covariation with climate. Science, 329, 834−838, https://doi.org/10.1126/science.1184984.
    Betts, R. A., C. A. Burton, R. A. Feely, M. Collins, C. D. Jones, and A. J. Wiltshire, 2021: ENSO and the carbon cycle. El Niño Southern Oscillation in a Changing Climate, M. J. McPhaden et al., Eds., John Wiley & Sons, Inc.
    Chen, X. J., X. G. Mo, S. Hu, and S. X. Liu, 2017: Contributions of climate change and human activities to ET and GPP trends over North China Plain from 2000 to 2014. Journal of Geographical Sciences, 27, 661−680,
    Cho, M.-H., G.-H. Lim, and H.-J. Song, 2014: The effect of the wintertime arctic oscillation on springtime vegetation over the northern high latitude region. Asia-Pacific Journal of Atmospheric Sciences, 50, 567−573, https://doi.org/10.1007/s13143-014-0046-1.
    Dannenberg, M. P., E. K. Wise, M. Janko, T. Hwang, and W. K. Smith, 2018: Atmospheric teleconnection influence on North American land surface phenology. Environmental Research Letters, 13, 034029, https://doi.org/10.1088/1748-9326/aaa85a.
    Forkel, M., N. Carvalhais, C. Rödenbeck, R. Keeling, M. Heimann, K. Thonicke, S. Zaehle, and M. Reichstein, 2016: Enhanced seasonal CO2 exchange caused by amplified plant productivity in northern ecosystems. Science, 351, 696−699, https://doi.org/10.1126/science.aac4971.
    Frederiksen, C. S., and X. Zheng, 2004: Variability of seasonal-mean fields arising from intraseasonal variability. Part 2, application to nh winter circulations. Climate Dyn., 23, 193−206,
    Friedlingstein, P., and Coauthors, 2020: Global carbon budget 2020. Earth System Science Data, 12, 3269−3340,
    Gong, D.-Y., J. Yang, S.-J. Kim, Y. Q. Gao, D. Guo, T. J. Zhou, and M. Hu, 2011: Spring Arctic Oscillation-East Asian summer monsoon connection through circulation changes over the western North Pacific. Climate Dyn., 37, 2199−2216, https://doi.org/10.1007/s00382-011-1041-1.
    Harris, I., T. J. Osborn, P. Jones, and D. Lister, 2020: Version 4 of the CRU TS monthly high-resolution gridded multivariate climate dataset. Scientific Data, 7, 109, https://doi.org/10.6084/m9.figshare.11980500.
    Houghton, R. A., 2000: Interannual variability in the global carbon cycle. J. Geophys. Res., 105, 20121−20130, https://doi.org/10.1029/2000JD900041.
    Hughes, J. K., P. J. Valdes, and R. Betts, 2006: Dynamics of a global-scale vegetation model. Ecological Modelling, 198, 452−462, https://doi.org/10.1016/j.ecolmodel.2006.05.020.
    Kalnay, E., and Coauthors, 1996: The NCEP/NCAR 40-year reanalysis project. Bull. Amer. Meteor. Soc., 77, 437−472, https://doi.org/10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.
    Kim, J.-S., J.-S. Kug, and S.-J. Jeong, 2017: Author Correction: Intensification of terrestrial carbon cycle related to El Niño–Southern Oscillation under greenhouse warming. Nature Communications, 8, 207, https://doi.org/10.1038/s41467-017-02461-9.
    Le Quéré, C., and Coauthors, 2018: Global carbon budget 2017. Earth System Science Data, 10, 405−448,
    Li, Y. Y., and Coauthors, 2021: Response of growing season gross primary production to El Niño in different phases of the pacific decadal oscillation over Eastern China based on Bayesian model averaging. Adv. Atmos. Sci., 38, 1580−1595, https://doi.org/10.1007/s00376-021-0265-1.
    Lorenz, E. N., 1956: Empirical orthogonal functions and statistical weather prediction. Statistical Forecast Project Report 1, 49 pp.
    Ma, J., X. M. Xiao, R. H. Miao, Y. Li, B. Q. Chen, Y. Zhang, and B. Zhao, 2019: Trends and controls of terrestrial gross primary productivity of China during 2000−2016. Environmental Research Letters, 14, 084032,
    Muller, W. A., C. Frankignoul, and N. Chouaib, 2008: Observed decadal tropical Pacific-North Atlantic teleconnections. Geophys. Res. Lett., 35, L24810, https://doi.org/10.1029/2008GL035901.
    Nemani, R. R., C. D. Keeling, H. Hashimoto, W. M. Jolly, S. C. Piper, C. J. Tucker, R. B. Myneni, and S. W. Running, 2003: Climate-driven increases in global terrestrial net primary production from 1982 to 1999. Science, 300, 1560−1563,
    Nian, D., N. M. Yuan, K. R. Ying, G. Liu, Z. T. Fu, Y. J. Qi, and C. L. F. Franzke, 2020: Identifying the sources of seasonal predictability based on climate memory analysis and variance decomposition. Climate Dyn., 55, 3239−3252, https://doi.org/10.1007/s00382-020-05444-7.
    Oleson, K. W., and Coauthors, 2010: Technical Description of version 4.0 of the Community Land Model (CLM) (No. NCAR/TN-478+STR). University Corporation for Atmospheric Research,
    Peng, J., L. Dan, and M. Huang, 2014: Sensitivity of global and regional terrestrial carbon storage to the direct CO2 effect and climate change based on the CMIP5 model intercomparison. PLoS One, 9, e95282, https://doi.org/10.1371/journal.pone.0095282.
    Peng, J., and L. Dan, 2015: Impacts of CO2 concentration and climate change on the terrestrial carbon flux using six global climate-carbon coupled models. Ecological Modelling, 304, 69−83, https://doi.org/10.1016/j.ecolmodel.2015.02.016.
    Peng, J., Y.-P. Wang, B. Z. Houlton, L. Dan, B. Pak, and X. B. Tang, 2020: Global carbon sequestration is highly sensitive to model-based formulations of nitrogen fixation. Global Biogeochemical Cycles, 34, e2019GB006296, https://doi.org/10.1029/2019GB006296.
    Peng, J., L. Dan, K. R. Ying, S. Yang, X. B. Tang, and F. Q. Yang, 2021: China's interannual variability of net primary production is dominated by the central china region. J. Geophys. Res., 126, e2020JD033362, https://doi.org/10.1029/2020JD033362.
    Piao, S. L., J. Y. Fang, P. Ciais, P. Peylin, Y. Huang, S. Sitch, and T. Wang, 2009a: The carbon balance of terrestrial ecosystems in China. Nature, 458, 1009−1013, https://doi.org/10.1038/nature07944.
    Piao, S. L., P. Ciais, P. Friedlingstein, N. de Noblet-Ducoudré, P. Cadule, N. Viovy, and T. Wang, 2009b: Spatiotemporal patterns of terrestrial carbon cycle during the 20th century. Global Biogeochemical Cycles, 23, GB4026, https://doi.org/10.1029/2008GB003339.
    Piao, S. L., and Coauthors, 2013: Evaluation of terrestrial carbon cycle models for their response to climate variability and to CO2 trends. Global Change Biology, 19, 2117−2132, https://doi.org/10.1111/gcb.12187.
    Piao, S. L., and Coauthors, 2020: Interannual variation of terrestrial carbon cycle: Issues and perspectives. Global Change Biology, 26, 300−318, https://doi.org/10.1111/gcb.14884.
    Rayner, N. A., D. E. Parker, E. B. Horton, C. K. Folland, L. V. Alexander, D. P. Rowell, E. C. Kent, and A. Kaplan, 2003: Global analyses of sea surface temperature, sea ice, and night marine air temperature since the late nineteenth century. J. Geophys. Res., 108, 4407, https://doi.org/10.1029/2002JD002670.
    Reimer, J. J., R. Vargas, D. Rivas, G. Gaxiola-Castro, J. M. Hernandez-Ayon, and R. Lara-Lara, 2015: Sea surface temperature influence on terrestrial gross primary production along the southern California current. PLoS One, 10, e0125177, https://doi.org/10.1371/journal.pone.0125177.
    Richardson, A. D., D. Y. Hollinger, J. D. Aber, S. V. Ollinger, and B. H. Braswell, 2007: Environmental variation is directly responsible for short- but not long-term variation in forest-atmosphere carbon exchange. Global Change Biology, 13, 788−803, https://doi.org/10.1111/j.1365-2486.2007.01330.x.
    Schaefer, K., A. S. Denning, and O. Leonard, 2005: The winter Arctic Oscillation, the timing of spring, and carbon fluxes in the Northern Hemisphere. Global Biogeochemical Cycles, 19, GB3017, https://doi.org/10.1029/2004GB002336.
    Shen, B. Z., Z. D. Lin, R. Y. Lu, and Y. Lian, 2011: Circulation anomalies associated with interannual variation of early- and late-summer precipitation in Northeast China. Science China Earth Sciences, 54, 1095−1104, https://doi.org/10.1007/s11430-011-4173-6.
    Sitch, S., and Coauthors, 2003: Evaluation of ecosystem dynamics, plant geography and terrestrial carbon cycling in the LPJ dynamic global vegetation model. Global Change Biology, 9, 161−185, https://doi.org/10.1046/j.1365-2486.2003.00569.x.
    Smith, B., I. C. Prentice, and M. T. Sykes, 2001: Representation of vegetation dynamics in the modelling of terrestrial ecosystems: comparing two contrasting approaches within European climate space. Global Ecology and Biogeography, 10, 621−637.
    Tharammal, T., G. Bala, N. Devaraju, and R. Nemani, 2019: A review of the major drivers of the terrestrial carbon uptake: Model-based assessments, consensus, and uncertainties. Environmental Research Letters, 14, 093005, https://doi.org/10.1088/1748-9326/ab3012.
    Wang, B., J. Yang, T. J. Zhou, and B. Wang, 2008: Interdecadal changes in the major modes of Asian-Australian Monsoon variability: Strengthening relationship with ENSO since the late 1970s. J. Climate, 21, 1771−1789, https://doi.org/10.1175/2007JCLI1981.1.
    Wang, B., J.-Y. Lee, and B. Q. Xiang, 2015: Asian summer monsoon rainfall predictability: A predictable mode analysis. Climate Dyn., 44, 61−74, https://doi.org/10.1007/s00382-014-2218-1.
    Wang, Y. P., and Coauthors, 2011: Diagnosing errors in a land surface model (CABLE) in the time and frequency domains. Journal of Geophysical Research: Biogeosciences, 116, G01034.
    Wieder, W. R., C. C. Cleveland, W. K. Smith, and K. Todd-Brown, 2015: Future productivity and carbon storage limited by terrestrial nutrient availability. Nature Geoscience, 8, 441−444, https://doi.org/10.1038/ngeo2413.
    Wilks, D. S., 1995: Statistical Methods in the Atmospheric Sciences. Academic Press, 467 pp.
    Woodward, F. I., T. M. Smith, and W. R. Emanuel, 1995: A global land primary productivity and phytogeography model. Global Biogeochemical Cycles, 9, 471−490, https://doi.org/10.1029/95GB02432.
    Wu, R. G., and B. Wang, 2002: A contrast of the East Asian summer Monsoon-ENSO relationship between 1962-77 and 1978-93. J. Climate, 15, 3266−3279, https://doi.org/10.1175/1520-0442(2002)015<3266:ACOTEA>2.0.CO;2.
    Wu, Z. W., B. Wang, J. P. Li, and F.-F. Jin, 2009: An empirical seasonal prediction model of the east Asian summer monsoon using ENSO and NAO. J. Geophys. Res., 114, D18120, https://doi.org/10.1029/2009JD011733.
    Yang, Q., Z. G. Ma, X. G. Fan, Z.-L. Yang, Z. F. Xu, and P. L. Wu, 2017: Decadal modulation of precipitation patterns over eastern China by sea surface temperature anomalies. J. Climate, 30, 7017−7033, https://doi.org/10.1175/JCLI-D-16-0793.1.
    Yao, Y. T., and Coauthors, 2018: Spatiotemporal pattern of gross primary productivity and its covariation with climate in China over the last thirty years. Global Change Biology, 24, 184−196, https://doi.org/10.1111/gcb.13830.
    Ying, K. R., T. B. Zhao, X.-W. Quan, X. G. Zheng, and C. S. Frederiksen, 2015: Interannual variability of autumn to spring seasonal precipitation in eastern China. Climate Dyn., 45, 253−271, https://doi.org/10.1007/s00382-014-2411-2.
    Ying, K. R., T. B. Zhao, X. G. Zheng, X.-W. Quan, C. S. Frederiksen, and M. X. Li, 2016: Predictable signals in seasonal mean soil moisture simulated with observation-based atmospheric forcing over China. Climate Dyn., 47, 2373−2395, https://doi.org/10.1007/s00382-015-2969-3.
    Ying, K. R., X. G. Zheng, T. B. Zhao, C. S. Frederiksen, and X.-W. Quan, 2017: Identifying the predictable and unpredictable patterns of spring-to-autumn precipitation over eastern China. Climate Dyn., 48, 3183−3206, https://doi.org/10.1007/s00382-016-3258-5.
    Ying, K. R., C. S. Frederiksen, T. B. Zhao, X. G. Zheng, Z. Xiong, X. Yi, and C. X. Li, 2018: Predictable and unpredictable modes of seasonal mean precipitation over Northeast China. Climate Dyn., 50, 3081−3095, https://doi.org/10.1007/s00382-017-3795-6.
    Zhang, X. Z., P. J. Rayner, Y.-P. Wang, J. D. Silver, X. J. Lu, B. Pak, and X. G. Zheng, 2016: Linear and nonlinear effects of dominant drivers on the trends in global and regional land carbon uptake: 1959 to 2013. Geophys. Res. Lett., 43, 1607−1614, https://doi.org/10.1002/2015GL067162.
    Zhang, L., and Coauthors, 2019: Interannual variability of terrestrial net ecosystem productivity over China: Regional contributions and climate attribution. Environmental Research Letters, 14, 014003, https://doi.org/10.1088/1748-9326/aaec95.
    Zhang, A. Z., and G. S. Jia, 2020: ENSO-driven reverse coupling in interannual variability of pantropical water availability and global atmospheric CO2 growth rate. Environmental Research Letters, 15, 034006, https://doi.org/10.1088/1748-9326/ab66cc.
    Zheng, X. G., and R. E. Basher, 1999: Structural time series models and trend detection in global and regional temperature series. J. Climate, 12, 2347−2358, https://doi.org/10.1175/1520-0442(1999)012<2347:STSMAT>2.0.CO;2.
    Zheng, X., D. M. Straus, and C. S. Frederiksen, 2008: Variance decomposition approach to the prediction of the seasonal mean circulation: Comparison with dynamical ensemble prediction using NCEP’s CFS. Quart. J. Roy. Meteor. Soc., 134, 1997−2009, https://doi.org/10.1002/qj.330.
    Zhu, Z. C., S. L. Piao, Y. Y. Xu, A. Bastos, P. Ciais, and S. S. Peng, 2017: The effects of teleconnections on carbon fluxes of global terrestrial ecosystems. Geophys. Res. Lett., 44, 3209−3218, https://doi.org/10.1002/2016GL071743.
  • [1] Xiaofei WU, Jiangyu MAO, 2019: Decadal Changes in Interannual Dependence of the Bay of Bengal Summer Monsoon Onset on ENSO Modulated by the Pacific Decadal Oscillation, ADVANCES IN ATMOSPHERIC SCIENCES, 36, 1404-1416.  doi: 10.1007/s00376-019-9043-8
    [2] Xinyi XING, Xianghui FANG, Da PANG, Chaopeng JI, 2024: Seasonal Variation of the Sea Surface Temperature Growth Rate of ENSO, ADVANCES IN ATMOSPHERIC SCIENCES, 41, 465-477.  doi: 10.1007/s00376-023-3005-x
    [3] CHEN Shangfeng, CHEN Wen, WEI Ke, 2013: Recent Trends in Winter Temperature Extremes in Eastern China and their Relationship with the Arctic Oscillation and ENSO, ADVANCES IN ATMOSPHERIC SCIENCES, 30, 1712-1724.  doi: 10.1007/s00376-013-2296-8
    [4] WU Zhiwei, LI Jianping, 2008: Prediction of the Asian-Australian Monsoon Interannual Variations with the Grid-Point Atmospheric Model of IAP LASG (GAMIL), ADVANCES IN ATMOSPHERIC SCIENCES, 25, 387-394.  doi: 0.1007/s00376-008-0387-8
    [5] CHEN Wei, LU Riyu, 2014: The Interannual Variation in Monthly Temperature over Northeast China during Summer, ADVANCES IN ATMOSPHERIC SCIENCES, 31, 515-524.  doi: 10.1007/s00376-013-3102-3
    [6] SU Qin, LU Riyu, LI Chaofan, 2014: Large-scale Circulation Anomalies Associated with Interannual Variation in Monthly Rainfall over South China from May to August, ADVANCES IN ATMOSPHERIC SCIENCES, 31, 273-282.  doi: 10.1007/s00376-013-3051-x
    [7] Minmin WU, Rong-Hua ZHANG, Junya HU, Hai ZHI, 2024: Synergistic Interdecadal Evolution of Precipitation over Eastern China and the Pacific Decadal Oscillation during 1951–2015, ADVANCES IN ATMOSPHERIC SCIENCES, 41, 53-72.  doi: 10.1007/s00376-023-3011-z
    [8] Xiaoxuan ZHAO, Riyu LU, 2020: Vertical Structure of Interannual Variability in Cross-Equatorial Flows over the Maritime Continent and Indian Ocean in Boreal Summer, ADVANCES IN ATMOSPHERIC SCIENCES, 37, 173-186.  doi: 10.1007/s00376-019-9103-0
    [9] ZHENG Fei, ZHANG Rong-Hua, ZHU Jiang, , 2014: Effects of Interannual Salinity Variability on the Barrier Layer in the Western-Central Equatorial Pacific: A Diagnostic Analysis from Argo, ADVANCES IN ATMOSPHERIC SCIENCES, 31, 532-542.  doi: 10.1007/s00376-013-3061-8
    [10] Kaiming HU, Yingxue LIU, Gang HUANG, Zhuoqi HE, Shang-Min LONG, 2020: Contributions to the Interannual Summer Rainfall Variability in the Mountainous Area of Central China and Their Decadal Changes, ADVANCES IN ATMOSPHERIC SCIENCES, 37, 259-268.  doi: 10.1007/s00376-019-9099-5
    [11] Philip E. BETT, Adam A. SCAIFE, Chaofan LI, Chris HEWITT, Nicola GOLDING, Peiqun ZHANG, Nick DUNSTONE, Doug M. SMITH, Hazel E. THORNTON, Riyu LU, Hong-Li REN, 2018: Seasonal Forecasts of the Summer 2016 Yangtze River Basin Rainfall, ADVANCES IN ATMOSPHERIC SCIENCES, 35, 918-926.  doi: 10.1007/s00376-018-7210-y
    [12] Ning JIANG, Congwen ZHU, 2021: Seasonal Forecast of South China Sea Summer Monsoon Onset Disturbed by Cold Tongue La Niña in the Past Decade, ADVANCES IN ATMOSPHERIC SCIENCES, 38, 147-155.  doi: 10.1007/s00376-020-0090-y
    [13] PENG Jingbei, CHEN Lieting, ZHANG Qingyun, 2014: The Relationship between the El Nio/La Nia Cycle and the Transition Chains of Four Atmospheric Oscillations. Part I: The Four Oscillations, ADVANCES IN ATMOSPHERIC SCIENCES, 31, 468-479.  doi: 10.1007/s00376-013-2275-0
    [14] WANG Zhiren, WU Dexing, CHEN Xue'en, QIAO Ran, 2013: ENSO Indices and Analyses, ADVANCES IN ATMOSPHERIC SCIENCES, 30, 1491-1506.  doi: 10.1007/s00376-012-2238-x
    [15] MENG Xiangfeng, WU Dexing, LIN Xiaopei, LAN Jian, 2006: A Further Investigation of the Decadal Variation of ENSO Characteristics with Instability Analysis, ADVANCES IN ATMOSPHERIC SCIENCES, 23, 156-164.  doi: 10.1007/s00376-006-0016-3
    [16] LI Gang*, LI Chongyin, TAN Yanke, and BAI Tao, 2014: The Interdecadal Changes of South Pacific Sea Surface Temperature in the Mid-1990s and Their Connections with ENSO, ADVANCES IN ATMOSPHERIC SCIENCES, 31, 66-84.  doi: 10.1007/s00376-013-2280-3
    [17] Yuanhai FU, Zhongda LIN, Tao WANG, 2021: Simulated Relationship between Wintertime ENSO and East Asian Summer Rainfall: From CMIP3 to CMIP6, ADVANCES IN ATMOSPHERIC SCIENCES, 38, 221-236.  doi: 10.1007/s00376-020-0147-y
    [18] Yadi LI, Xichen LI, Juan FENG, Yi ZHOU, Wenzhu WANG, Yurong HOU, 2024: Uncertainties of ENSO-related Regional Hadley Circulation Anomalies within Eight Reanalysis Datasets, ADVANCES IN ATMOSPHERIC SCIENCES, 41, 115-140.  doi: 10.1007/s00376-023-3047-0
    [19] CHEN Wen, WEI Ke, 2009: Interannual Variability of the Winter Stratospheric Polar Vortex in the Northern Hemisphere and their Relations to QBO and ENSO, ADVANCES IN ATMOSPHERIC SCIENCES, 26, 855-863.  doi: 10.1007/s00376-009-8168-6
    [20] JIANG Yuxin, TAN Benkui, 2015: Two Modes and Their Seasonal and Interannual Variation of the Baroclinic Waves/Storm Tracks over the Wintertime North Pacific, ADVANCES IN ATMOSPHERIC SCIENCES, 32, 1244-1254.  doi: 10.1007/s00376-015-4251-3
  • ESM-20210226-1.pdf

Get Citation+

Export:  

Share Article

Manuscript History

Manuscript received: 11 June 2021
Manuscript revised: 29 November 2021
Manuscript accepted: 06 December 2021
通讯作者: 陈斌, bchen63@163.com
  • 1. 

    沈阳化工大学材料科学与工程学院 沈阳 110142

  1. 本站搜索
  2. 百度学术搜索
  3. 万方数据库搜索
  4. CNKI搜索

Ocean–atmosphere Teleconnections Play a Key Role in the Interannual Variability of Seasonal Gross Primary Production in China

    Corresponding author: Jing PENG, pengjing@tea.ac.cn
  • Key Laboratory of Regional Climate-Environment for Temperate East Asia, Institute of Atmospheric Physics, Chinese Academy of Sciences, Beijing 100029, China

Abstract: Since the 1950s, the terrestrial carbon uptake has been characterized by interannual variations, which are mainly determined by interannual variations in gross primary production (GPP). Using an ensemble of seven-member TRENDY (Trends in Net Land–Atmosphere Carbon Exchanges) simulations during 1951–2010, the relationships of the interannual variability of seasonal GPP in China with the sea surface temperature (SST) and atmospheric circulations were investigated. The GPP signals that mostly relate to the climate forcing in terms of Residual Principal Component analysis (hereafter, R-PC) were identified by separating out the significant impact from the linear trend and the GPP memory. Results showed that the seasonal GPP over China associated with the first R-PC1 (the second R-PC2) during spring to autumn show a monopole (dipole or tripole) spatial structure, with a clear seasonal evolution for their maximum centers from springtime to summertime. The dominant two GPP R-PC are significantly related to Sea Surface Temperature (SST) variability in the eastern tropical Pacific Ocean and the North Pacific Ocean during spring to autumn, implying influences from the El Niño–Southern Oscillation (ENSO) and the Pacific Decadal Oscillation (PDO). The identified SST and circulation factors explain 13%, 23% and 19% of the total variance for seasonal GPP in spring, summer and autumn, respectively. A clearer understanding of the relationships of China’s GPP with ocean–atmosphere teleconnections over the Pacific and Atlantic Ocean should provide scientific support for achieving carbon neutrality targets.

摘要: 自1950年来,在初级生产力(GPP)年际变率影响下,陆地碳汇呈现显著的年际变化。在区域尺度上,这种年际变率的主要来源——特别是海气遥相关过程,仍不甚清楚。基于TRENDY陆-气间碳交换趋势比较计划的7个模式,时间跨度为1951至2010年,本文研究了中国各季GPP的年际变化与海表温度和大气环流之间的关系。通过分离受长期趋势和GPP记忆影响的部分,本文首先将春夏秋各个季节的GPP残差部分识别为更受气候强迫影响的信号。结果表明,春季至秋季的第一(第二)GPP残差模态呈现单极型(偶极型或三极型)的空间分布,中心位置随季节推移而变化。春夏秋各季的前两个GPP残差模态与赤道东太平洋海温和北太平洋海温显著相关,表明了厄尔尼诺-南方涛动和太平洋年代际涛动对中国GPP的重要影响。此外,大气环流与GPP的关系特征显示,北极涛动、西太平洋涛动、贝加尔湖阻塞以及西太平洋副热带高压分别对春夏秋三季、春季、夏季和秋季的中国大陆GPP有着重要影响。以上识别的海温及环流因子可分别解释春季13%,夏季23%和秋季19%的GPP季节平均的总方差。特别指出,以上因子在春夏秋季,主要与中国中部、西南部、东北部和南部的GPP变化有关,而这些区域主导了中国大陆GPP的年际变化。对中国GPP及其在太平洋和大西洋上海气遥相关背景的深入理解,将为有效实现中国碳循环的估算和预估,并为碳中和目标的实现提供科学依据。

    • Over the past six decades, China’s terrestrial ecosystem has played an important role in the global carbon sink (Piao et al., 2009a; Peng et al., 2014; Le Quéré et al., 2018; Friedlingstein et al., 2020), with considerable interannual variance in response to climate changes (Zhang et al., 2019; Peng et al., 2021). Gross primary production (GPP) is one of the largest carbon fluxes and is strongly associated with the annual variance of the terrestrial carbon sink (Houghton, 2000; Piao et al., 2020). Previous studies have shown that ocean–atmosphere teleconnections exert important influences on the interannual variabilities and predictabilities of various climate variables, such as precipitation and temperature, on global and regional scales (e.g., Ying et al., 2015, 2017; Nian et al., 2020). However, in China, as one of the regions with the fastest increase in carbon emissions, such an impact of these teleconnections on the interannual variances of GPP has not been well clarified. Furthermore, knowledge gaps in climate–carbon linkages may further contribute to conflicting model results concerning the future terrestrial carbon uptake in earth system models (Tharammal et al., 2019). Therefore, developing a better understanding of the relationships between the interannual variation of GPP in China and ocean–atmosphere teleconnections will not only offer the opportunity to better understand climate–carbon cycle relationships, but also increase the confidence in future projections of the carbon cycle under different carbon emission scenarios.

      As indicated by various climatological studies, external forcing and low-frequency (interannual to supra-annual) internal atmospheric dynamics are influential climatic factors governing the interannual variations of seasonal mean anomalies in China’s precipitation, temperature and soil moisture (e.g., Wang et al., 2008, 2015; Wu et al., 2009; Gong et al., 2011; Ying et al., 2016, 2017; Nian et al., 2020), including, for instance, global sea surface temperature (SST) anomalies and large-scale atmospheric oscillations. Because of the close relationships between GPP and local hydrothermal conditions of precipitation, temperature and soil moisture, as revealed by many previous studies (Nemani et al., 2003; Richardson et al., 2007; Beer et al., 2010; Barman et al., 2014; Zhang et al., 2019; Peng et al., 2020), the interannual variability of China’s GPP should also arise from these distant ocean–atmosphere teleconnection processes. It has long been noted that the well-known climatic modes, such as El Niño–Southern Oscillation (ENSO) and Arctic Oscillation (AO), can influence the global terrestrial carbon cycle (Reimer et al., 2015; Dannenberg et al., 2018). For instance, hot and dry climate conditions in El Niño years are the primary reasons for a lower carbon sink and even a carbon source at regional scales, particularly in the tropics (Nemani et al., 2003; Ahlström et al., 2015; Zhang and Jia, 2020). In addition, Schaefer et al. (2005) and Cho et al. (2014) found a close relationship among the winter AO and the vegetation activity and carbon fluxes in the following spring over the Northern Hemisphere. However, at regional scales, due to the different intensities of the ocean–land–atmosphere coupling, the response of the regional carbon flux to natural cyclic climatic phenomena varies substantially (Dannenberg et al., 2018; Betts et al., 2021). At present, in China, how and to what extent the interannual variability of GPP responds to these impacts from sea temperature and large-scale atmospheric circulation is not yet clear. Therefore, it is important to better understand the relationship between the interannual changes in China’s GPP and ocean–atmosphere teleconnection. This is a key process towards elucidating the carbon sequestration mechanisms of terrestrial ecosystems on regional scales, especially over China (Zhu et al., 2017).

      Considering the sources of interannual variability of China’s GPP, the significant trend in China’s GPP, owing to the effects of human activities and climate change, has been extensively examined (Piao et al., 2009b; Peng and Dan, 2015; Forkel et al., 2016; Chen et al., 2017; Yao et al., 2018; Ma et al., 2019; Piao et al., 2020). This long-term trend provides an important source of interannual variability for GPP in China. Furthermore, in the same way that the oceans can store heat and soil moisture can store water, vegetation is a carbon pool, with associated “memory” features (inertia of a climate variable that persists from the past conditions). This is reflected by the close relationship between the current and prior GPP anomalies. As the GPP can “remember” the anomalous status long after those anomalies are “forgotten” by the atmosphere, it possesses larger potential predictabilities [Fig. S1 in the Electronic Supplementary Materials, (ESM)] than atmospheric variables such as precipitation (Ying et al., 2017) and temperature (Nian et al., 2020). Thus, the memory of GPP also provides a source of its annual variance. However, this study is mostly interested in the linkage of China’s GPP with ocean–atmosphere teleconnections. Consequently, we separated out the significant effects from the linear trend and GPP memory as the first step.

      The objective of the present study is to better understand the sources of the interannual variability of spring-to-autumn GPP in China that derive from ocean–atmosphere teleconnections. To achieve this, using monthly GPP data obtained from a seven-member TRENDY (Trends in Net Land–Atmosphere Carbon Exchanges) simulation over China, we focus on the dominant GPP signals that are mostly related to the SST and circulations, by separating out the significant effects from the linear trend and GPP memory. This study aims to explore (1) the main ocean–atmosphere teleconnections related to the interannual variability of the seasonal mean GPP in the Chinese mainland from spring to autumn; and (2) to what extent the year-to-year fluctuations of China’s GPP are dominated by the identified key ocean–atmosphere teleconnection factors. The data and methods employed in the study are described in section 2. Analyses of seasonal contribution characteristics of China’s GPP interannual variabilities are reported in section 3.1. The GPP interannual variability arising from the trend, GPP memory and climate forcing are presented in section 3.2 and the main sources of the interannual variability from ocean–atmosphere teleconnections are discussed in section 3.3. The fractions of variance explained by the key ocean–atmosphere teleconnection factors are evaluated in section 3.4. The conclusions and some further discussion are presented in section 4.

    2.   Data and methods
    • To further assess the responses of China’s GPP to ocean–atmosphere teleconnections, TRENDY models that participated in the Global Carbon Project (Le Quéré et al., 2018) were prepared to simulate the monthly GPP from 1951 to 2010. The seven models included in this study were CABLE, CLM4C, CLM4CN, LPJ, LPJ_GUESS, SDGVM and TRI (Table 1). The TRENDY project includes datasets of offline experiments driven by constant or varying inputs (i.e., climate variables, atmospheric CO2, and land use/cover forcing). Its purpose was to distinguish between the effects of CO2, climate, and land use. We mainly evaluated the results of the S2 experiment, which only considers varying climate variables and CO2, without varying land use. Further details about the S2 TRENDY simulation protocol can be found in Piao et al. (2013). Importantly, models from TRENDY are broadly considered to be the world’s most advanced terrestrial ecosystem models. Through comparison with satellite remote sensing data, previous studies have demonstrated that carbon fluxes from TRENDY are simulated well in China (Zhang et al., 2016; Peng et al., 2021).

      Model nameAbbreviationSpatial resolution (lat × lon)Land surface modelFull nitrogen cycleFire simulationHarve fluxSource
      Community LandCLM4C0.5° × 0.5°YesNoYesNoOleson et al. (2010)
      Community LandCLM4CN0.5° × 0.5°YesYesYesNoOleson et al. (2010)
      Lund–Potsdam–JenaLPJ0.5° × 0.5°NoNoYesYesSitch et al. (2003)
      LPJ-GUESSLPJ-GUESS0.5° × 0.5°NoNoYesNoSmith et al. (2001)
      Sheffield-DGVMSDGVM3.75° × 2.5°NoNoYesNoWoodward et al. (1995)
      TRIFFIDTRI3.75° × 2.5°YesNoNoNoHughes et al. (2006)
      CABLECABLE0.5° × 0.5°NoYesNoNoWang et al. (2011)

      Table 1.  Seven trendy models used in this study.

      In order to better understand the air–ocean teleconnections associated with GPP in China, we used monthly mean datasets of (1) 500-hPa geopotential height from the National Centers for Environmental Prediction–National Center for Atmospheric Research Reanalysis 1 project (Kalnay et al., 1996), with a spatial grid resolution of 2.5°; (2) SST (on a 1° × 1° grid) from the UK Met Office Hadley Centre Sea Ice and Sea Surface Temperature dataset, version 1 (Rayner et al., 2003); (3) precipitation and temperature (on a 0.5° × 0.5° grid) from the Climate Research Unit time series datasets, version 4.03 (Harris et al., 2020); (4) soil moisture (on a 1° × 1° grid) from TRENDY simulations in China; and (5) climate indices from the National Oceanic and Atmospheric Administration’s Climate Prediction Center (available at https://psl.noaa.gov/data/climateindices/list/).

    • Firstly, the climatological mean was removed from the monthly data. Then, by separating out the significant influences from the prolonged trend and GPP memory, the following procedure was followed to derive the GPP signals that were mostly associated with the climate forcing component:

      (1) An empirical orthogonal function (EOF) analysis (Lorenz, 1956) was applied to the sample covariance matrix of the seasonal GPP field ${x_{y,o}}$ in year y (y = 1, 2, ..., Y; Y is the total number of years) and season o (a three-month season) to derive the EOF modes and their associated principal component (PC) time series ${t_y}$.

      (2) For each PC time series ${t_y}$, the dependence on a linear trend and a red noise process was modelled as

      Here, $\lambda $ is the linear trend (consecutive years from 1 to Y) within total Y years, $\mu $ is an intercept/mean term, and ${\varepsilon _y}$ is an autoregressive process of order 1 (AR1), i.e.,

      where $\alpha $ is the yearly autocorrelation coefficient and ${\eta _y}$ is the white noise [following Eq. (1) of Zheng and Basher, 1999].

      In this case, by removing those effects from the trend $\lambda $ and GPP memory component $\alpha {\varepsilon _{y - 1}}$ that were statistically significant (the Autoregressive Integrated Moving Average Model in R-code was used for calculations), the climate information was included in the residual component of the PC time series ${\eta _y}$. For convenience, the total and residual component of the PC time series are denoted by T-PC (${t_y}$) and R-PC (${\eta _y}$), respectively.

      (3) With R-PC ${\eta _y}$ derived, the main SST and circulation factors related to the GPP in China could then be identified by calculating the correlation coefficients of the SST or circulations associated with the GPP R-PC.

      Following Wilks (1995), the fractional variance was applied to evaluate the relative importance of the identified SST or circulation factors associated with the dominant GPP R-PC. Let ${{\boldsymbol{e}}_j}$ be the jth (j = 1,…., J, where J is the total number of EOF) EOF, and ${{\boldsymbol{\hat p}}_{j,y}}$ be the linear regression for the jth R-PC time series in year y based on its associated key ocean–atmosphere teleconnection factors. Then, the fractional variance of the seasonal mean anomalies explained by the key SST or circulation factors is

      where FV is the fractional variance of the seasonal mean anomalies explained by the key SST or circulation factors, ${{\boldsymbol{r}}_y}$ is the yth-year seasonal GPP, ${\boldsymbol{\bar r}} = \left[ {\bar r\left( {\text{1}} \right),...,\bar r\left( {{N}} \right)} \right]$ is the climatology of ${{\boldsymbol{r}}_y}$, and $ \left\| \,\right\|$ is the Euclidean distance operator; for example,

      Here, n (= 1,…., N) denotes a specific location.

    3.   Results
    • The interannual variation of the seasonal mean GPP exhibits considerable differences among seasons. In general, the interannual variability of China’s GPP from spring to autumn (March to November) – covering the whole growing season – dominates the variations for all seasons (Fig. 1a), accounting for 91% of total interannual variations. Particularly, the largest interannual variations are observed in June–July–August (summer) over most of the Chinese mainland, and the second largest are in September–October–November (autumn), followed by March–April–May (spring) (Table 2; Figs. 1bd). A similar east–west gradient in terms of spatial distribution is apparent in all these three seasons, with larger values mainly located over eastern China, where the influence of the East Asian monsoon is greatest (Figs. 1bd). With the development of the East Asian monsoon, the maximum centers show a clear seasonal migration from spring to autumn, with the largest values over central China and southwestern China in spring [larger than 0.002 (KgC m−2 month−1)2], central China and northeastern China in summer [larger than 0.004 (KgC m−2 month−1)2], and southeastern China in autumn [larger than 0.004 (KgC m−2 month−1)2]. Also, there are remarkable differences in the mean state and principal modes of SST and atmospheric circulations during different seasons (Frederiksen and Zheng, 2004; Zheng et al., 2008). Therefore, it is important to investigate the relationships of the interannual variability of the GPP in China associated with the SST and atmospheric circulation patterns in separate seasons. The discussions below will focus on the seasons from spring to autumn, which comprise the growing seasons and have the largest interannual variabilities.

      Figure 1.  (a) Temporal variations of China’s GPP for all the four seasons (black) and the total of the three seasons of spring-summer-autumn (red) (units: KgC m−2); and spatial distributions of the (b–d) total variance [units: (KgC m−2 month−1)2] of the seasonal mean GPP over China (e–g) residual variance [units: (KgC m−2 month−1)2] of the seasonal mean GPP from the trend and GPP memory, and (h–j) ratio of the residual variance to the total variance of GPP over China (units: %) for spring (left), summer (middle) and autumn (right), respectively.

      SeasonT- variabilityR- variabilityR/T variability
      spring0.0610.03964
      summer0.1950.11760
      autumn0.1200.07865

      Table 2.  Total variability of the mean seasonal variations in GPP pattern [leftmost column; (KgC m−2 month−1)2], the residual variability from the trend and the GPP memory of the seasonal mean GPP pattern [second column; (KgC m−2 month−1)2], and the percentage of the residual to total variability (third column; %).

    • First, the T-PC and R-PC of China’s GPP are estimated using Eq. (1). As shown in Fig. 2, apparent upward trends, which are statistically significant at the 95% confidence level based on the Student’s t-test, can be seen for the GPP T-PC1 during spring to autumn (black lines in Figs. 2b, d and f). Meanwhile, there is a statistically significant (at the 95% confidence level) memory signal in the GPP T-PC2 of summer (black line in Fig. 2j). Thus, to focus on the GPP signals that are mostly related to the climate forcing component, the R-PC needed to be extracted by removing the significant effects from the trend and GPP memory (red lines in right-hand panel of Fig. 2; only the two most dominant PC with the largest explained variances are displayed).

      Figure 2.  Spatial distributions of the two dominant EOF modes of the total seasonal mean GPP field (with explained variance in brackets) for (a, g) spring, (c, i) summer, and (e, k) autumn, respectively; and temporal variations of the total PC time series (T-PC; black) and the residual PC time series from the trend and GPP memory (R-PC; red) for (b, h) spring, (d, j) summer, and (f, l) autumn, respectively. Noted the R-PC2 coincide with the T-PC2 in spring and autumn

      To explore the relative importance of the trend, memory and climate forcing on the interannual variability of seasonal GPP in China, the total variance and the residual variance from the linear trend and the GPP memory of the GPP were then obtained. Overall, the average residual-to-total percentage variance is high for the GPP in China, with values of 64%, 60% and 65% from spring to autumn, respectively (Table 2). This indicates that there is still a large amount of the GPP interannual variance that cannot have originated from the prolonged trend and the GPP memory; that is, the GPP interannual variance may be largely affected by the lagged and simultaneous conditions of the climate forcing component.

      Compared to the spatial distributions of the total variance (Figs. 1bd), the residual variance from the trend and the GPP memory (Figs. 1eg) show similar distribution features, with the largest values over central China and southwestern China in spring, central China and northeastern China in summer, and southeastern China in autumn. In particular, the maximum centers of the residual-to-total ratio (Figs. 1hj), which contain over 50% of the percentage, include the above particular areas that possess the largest interannual variations (Figs. 1bg). Thus, it was necessary to identify the key ocean–atmosphere teleconnections associated with the seasonal GPP over China to further understand the main sources of the interannual variability of China’s GPP.

    • The focus in this section is the GPP signals that are mostly related to the climate forcing factors (R-PC), as determined by separating out the influences from the long-term trends and the GPP memory [Eq. (1)]. The spatial characteristics of the seasonal GPP in China associated with the dominant R-PC from spring to autumn are discussed. The main ocean–atmosphere teleconnections associated with the seasonal GPP in China from the SST and circulations were identified by calculating the correlation between the GPP R-PC and the seasonal mean SST or circulation anomalies. To strengthen the physical explanation of the GPP signals, we also examined the conditions of soil moisture, precipitation and temperature associated with the GPP R-PC (GPP interannual variations are strongly associated with those of the atmosphere and land).

    • The seasonal GPP field associated with the R-PC1 for spring, summer and autumn are displayed in Figs. 3a, d and g, respectively. In the phase shown here, positive loadings correspond to larger-than-normal GPP conditions in China for all three seasons. Meanwhile, the anomalous maximum centers demonstrate a clear seasonal migration. In particular, during spring, the amplitude center is situated over southern China and northeastern China, with the largest values along the Yangtze River (Fig. 3a). During boreal summer (summer), accompanied by the development of the East Asian summer monsoon and the movement of the monsoon rainfall belt, the local maximum is mainly located in the regions of the Yangtze–Huaihe River Valley and northeastern China (Fig. 3d). During autumn, as the East Asian monsoon retreats, the maximum GPP center is situated in southeastern China (Fig. 3g). In addition, similar maximum centers can be seen for the spatial patterns of the GPP associated with the R-PC1 (Figs. 3a, d and g), as well as those with the corresponding T-PC1 (Figs. 2a, c and e), while the latter has a hybrid structure associated with the trend and climate forcing. This implies that the climate forcing component may dominate the GPP’s interannual variabilities in the above areas.

      Figure 3.  Correlation maps of contemporary GPP (left-hand column), SST (middle column) and 500-hPa geopotential height (right-hand column) associated with the (a–c) spring R-PC1, (d–f) summer R-PC1, (g–i) autumn R-PC1, (j–l) spring R-PC2, (m–o) summer R-PC2 and (p–r) autumn R-PC2, respectively. The shaded areas in the correlation maps are significant at the 95% confidence level, using the Student’s t-test.

      The one-point correlation maps of the simultaneous SSTs associated with the GPP R-PC1 during spring–autumn display similar spatial characteristics, as shown in Figs. 3b, e, and h. The most remarkable feature of the SST correlation is a horseshoe-shaped PDO-like pattern [PDO: Pacific Decadal Oscillation; see, for example, Fig. 3c of Ying et al. (2018)] over the North Pacific Ocean, and a sandwich-like tripole structure — with significant positive loadings in latitudes between 20°N and 40°N and negative anomalies in its north and south — over the North Atlantic Ocean. The simultaneous 500-hPa height correlation maps associated with the leading modes of spring to autumn (Figs. 3c, f and i) all show an AO-like zonal structure over the mid-to-high latitudes of the Northern Hemisphere; this is characterized by opposite anomalies over Greenland and the North Atlantic Ocean/northern Europe [see, for example, S-REOF1 in Fig. 4 of Frederiksen and Zheng (2004)], and is closely linked to the North Atlantic tripolar SSTs. These suggest that the AO and PDO may be important factors influencing China’s GPP from spring to autumn. Consistently, the temporal correlation coefficients of the AO/PDO indices associated with the GPP R-PC1 in spring, summer and autumn are 0.45/−0.36, 0.26/−0.34 and 0.25/−0.28 when calculated simultaneously, and 0.43/−0.33, 0.30/−0.31 and 0.25/−0.28 when calculated at the lead time (Table 3), respectively, and are statistically significant at the 95% confidence level. Furthermore, the relationship between China’s GPP and the combined effects of the PDO and AO was further examined using correlation maps of the GPP in China associated with the simultaneous and lagged PDO and AO indices in Table 3 during the seasons from spring to autumn (figures not shown), and the results were found to be consistent.

      Figure 4.  Correlation maps of contemporary soil moisture (left-hand column), precipitation (middle column) and temperature (right-hand column) associated with the (a–c) spring R-PC1, (d–f) summer R-PC1, (g– i) autumn R-PC1, (j– l) spring R-PC2, (m–o) summer R-PC2 and (p–r) autumn R-PC2, respectively. The shaded areas in the correlation maps are significant at the 95% confidence level, using the Student’s t-test.

      GPP R-PCClimate indicesCorrelation (contemporary)Climate indicesCorrelation (lead-lag)
      spring R-PC1spring AO0.45***Feb AO0.43***
      spring PDO−0.36**spring(1) PDO−0.33**
      spring R-PC2spring Niño-3.40.28*Feb Niño-3.40.32*
      summer R-PC1summer AO0.26*May AO0.30*
      summer PDO−0.34**spring PDO−0.31*
      summer R-PC2summer Niño-3.40.28*D(1)JF Niño-3.4−0.34**
      autumn R-PC1autumn AO0.25**Aug AO0.25*
      autumn PDO−0.28*summer PDO−0.28*
      autumn R-PC2autumn Niño-3.40.47***summer Niño-3.40.49***
      *0.05, **0.01, *** 0.001

      Table 3.  Correlation coefficients between the GPP R-PC and the contemporary (third column) and lead-time (fifth column) climate indices. The years used in the climate indices are denoted by (1) for the preceding year.

      Figures 4ai show the correlation maps of simultaneous soil moisture, precipitation and temperature associated with the GPP R-PC1 during spring to autumn. Anomalously wetter conditions of precipitation and soil moisture observed over southern China in spring (Figs. 4a and b), the Yangtze–Huaihe River Valley and northeastern China in summer (Figs. 4d and e), and southeastern China in autumn (Figs. 4g and h) are responsible for the positive anomaly of GPP in those areas (Figs. 3a, d and g). Locally, the significantly wetter conditions are a response to an anomalous anticyclone centered over the northwestern Pacific around Japan (Figs. 3c, f and i), which brings moist air from the northwestern Pacific to eastern China. This anomalous high is consistent with the significant positive SST anomalies situated over the northwestern Pacific (Figs. 3b, e and h) – one of the major features of negative PDO phases (Muller et al., 2008), and known to be closely associated with the AO variability (Gong et al., 2011). Meanwhile, significantly higher-than-normal temperature anomalies are observed in northeastern China during spring (Fig. 4c). As the temperature has a generally positive feedback to GPP in northeastern China during spring (Peng et al., 2021), the GPP there is enhanced (Fig. 3a). The temperature associated with the leading GPP mode of summer shows cooler-than-normal conditions in central-eastern China (Fig. 4f), which increases the GPP there (Fig. 3d) via its water deficit effects (Kim et al., 2017; Li et al., 2021; Peng et al., 2021). The above results are generally consistent with many previous studies (e.g., Gong et al., 2011; Yang et al., 2017; Ying et al., 2017, 2018) in which relationships between both the PDO and AO with the East Asian monsoon and climate in China have been found during different seasons.

    • The GPP fields associated with the R-PC2 during spring, summer and autumn are displayed in Figs. 3j, m and p, respectively. Associated with the spring R-PC2, the GPP fields have a tripolar spatial structure, with negative anomalies in southeastern and northeastern China and positive anomalies between the lower reaches of the Yangtze and Yellow rivers (Fig. 3j) in the phase shown here. Corresponding to the summer R-PC2, meanwhile, there is a dipole pattern, with higher-than-normal conditions in southern China and lower-than-normal conditions in northeastern China (Fig. 3m). For the autumn R-PC2’s related GPP, there are negative loadings in central-eastern China and positive loadings in southeastern China (Fig. 3p). Again, there are large similarities between the spatial distributions of the GPP associated with the R-PC2 (Figs. 3j, m and p) and those with the T-PC2 (Figs. 2g, i and k), suggesting dominant effects from the climate forcing component in these regions.

      The most notable features that appear in the simultaneous SST correlation maps associated with the R-PC2 of GPP during spring to autumn (Figs. 3k, n and q) are the significant positive values over the eastern tropical Pacific Ocean, indicating that ENSO is the possible source of the interannual variability for China’s GPP during these seasons. The results are consistent with the correlation maps of China’s GPP associated with the Niño3.4 index (figures not shown), which also indicate a close linkage between ENSO and China’s GPP from spring to autumn. Previous studies have revealed that ENSO is one of the major factors affecting the interannual variations of China’s soil moisture (Ying et al., 2016) and temperature (Nian et al., 2020), as well as the East Asian summer monsoon and monsoonal rainfall (Wu et al., 2009; Wang et al., 2015; Ying et al., 2016, 2017), via its regulation of air–sea interactions over the Pacific–Eurasia region. However, there are much stronger ENSO–GPP correlations for autumn (autumn; Fig. 3q) than in the warm seasons (spring and summer; Figs. 3k and n). In particular, the temporal correlations of the Niño3.4 index associated with the second R-PC in spring, summer and autumn are 0.28, 0.28 and 0.47 when calculated simultaneously, and 0.32, −0.34 and 0.49 when calculated at the lead time (statistically significant at the 95% confidence level), respectively. Consistent with this, a previous study by Ying et al. (2017) reported a similar seasonality in the lead–lag and simultaneous relationship between ENSO and precipitation in eastern China from spring to autumn, based on precipitation observations, in which stronger ENSO–precipitation correlations were seen in autumn [Fig. 8b of Ying et al. (2017)] than that in spring and summer [Fig. 3j of Ying et al. (2017)]. Other studies have indicated that this seasonality in the behavior of ENSO–precipitation correlations might be due to the ENSO signal in autumn persisting throughout the entire period, whereas the signal in spring and summer has an apparent interdecadal change around the late 1970s (Wu and Wang, 2002; Wang et al., 2008; Ying et al., 2015). Explanations for the seasonality in the GPP–ENSO relationship is still an interesting and open question that needs further examination.

      When examining the atmospheric circulation anomalies, the simultaneous 500-hPa geopotential height associated with the spring R-PC2 (Fig. 3l) displays a distinct meridional dipole in the western Pacific, resembling the western Pacific Oscillation (WPO) pattern of Frederiksen and Zheng (2004; their S-REOF3 in Fig. 4; pattern correlation with canonical WPO for spring is 0.66), which is closely related to ENSO. Corresponding to this anomalous circulation in spring, the soil moisture and the precipitation show negative anomalies in southern China and positive anomalies in the area between the Yangtze and Yellow rivers (Figs. 4j and k); plus, the temperature displays warmer-than-normal conditions in southern China and cooler conditions in central eastern China (Fig. 4l), which is responsible for the anomaly centers of the GPP EOF2 in these regions (Fig. 3j). Based on in-situ observations, Ying et al. (2017; their Fig. 7e) concluded that El Niño and positive WPO are significantly related to drier (wetter) rainfall patterns in the south (north) over eastern China in spring, which is consistent with our results above.

      For the summer GPP R-PC2, the most noticeable feature in the associated 500-hPa height field is an anomalous positive center located around Lake Baikal (Fig. 3o). Previous studies have found that this dual blocking high condition over Lake Baikal is favorable for suppressed rainfall in northeastern China (Shen et al., 2011; Ying et al., 2018), which is consistent with our results. In particular, an increase (decrease) in soil moisture and rainfall (Figs. 4m and n), and decrease (increase) in temperature (Fig. 4o) over southern China (northeastern China), facilitate this particular GPP spatial pattern (Fig. 3m).

      Associated with the GPP R-PC2 of autumn, the simultaneous 500-hPa height shows significant positive anomalies over the northwestern Pacific Ocean (Fig. 3r), which is the region most influenced by the western Pacific subtropical high (WPSH), suggesting the WPSH is an important factor affecting the GPP anomalies over China during autumn. The anomalously drier conditions in central-eastern China (Figs. 4p and q), as well as the anomalously wetter (Figs. 4p and q) and cooler conditions (Fig. 4r) in southern China, correspond to the anomalous GPP dipole structure as shown in Fig. 3p. The results agree with Ying et al. (2017), in which ENSO and the WPSH were identified as the most important factors affecting the interannual variability of autumn seasonal mean rainfall in eastern China, and an ENSO-related precipitation pattern in autumn similar to our GPP EOF2 was observed.

    • Based on the above results, we calculated the fraction of variance of seasonal GPP explained by the key SST and circulation factors, using Eq. (3), in order to further evaluate the relative contributions of the identified ocean–atmosphere teleconnections of the interannual variability of China’s GPP. Here, the key SST and circulation factors are represented by a projection of seasonal mean SST and height fields on their corresponding correlation maps with the GPP R-PC. In general, the key SST and circulation teleconnection factors can produce a large amount of GPP interannual variance, with an average percentage explained variance of 13%, 23% and 19% over China during spring, summer and autumn, respectively. The spatial distributions of the fraction of variance explained by the key SST and circulation factors (Fig. 5) display maximum centers (more than 30%) over central-eastern China and southwestern China in spring (Fig. 5a), central-eastern China and northeastern China in summer (Fig. 5b), and southeastern China in autumn (Fig. 5c). Meanwhile, these regions make substantial absolute contributions to the carbon cycle for China, as they have the largest interannual variabilities (Figs. 1bd) and climatological means (figure not shown) in GPP across the country as a whole. Thus, aside from the prolonged trend and the GPP memory, the SST and atmospheric teleconnections are also crucial to the interannual variance of China’s GPP.

      Figure 5.  Spatial distributions of the fraction of variance of GPP explained by the SST and circulation factors, for (a) spring, (b) summer and (c) autumn, respectively.

    4.   Summary and discussion
    • A seven-member TRENDY simulation was used to estimate the GPP field over China for the period 1951–2010. By separating out the significant effects from the linear trend and GPP memory, the dominant GPP signals (R-PC) that are most related to climate forcing factors were derived, for the seasons of spring to autumn. The key sources of interannual variability from the SST and atmospheric circulations for the seasonal GPP in China were then identified. In this paper, we have gained a more quantitative understanding of the sources of interannual variability of China’s GPP from the ocean–atmosphere teleconnections. The main results can be summarized as follows:

      (1) The seasonal GPP over China associated with the R-PC1 (R-PC2) during spring to autumn show a monopole (dipole or tripolar) spatial structure, with a clear seasonal evolution for their maximum centers from spring to autumn.

      (2) The PDO and AO are closely linked with the GPP R-PC1 from spring to autumn. This combined impact on GPP in China is accompanied by a significant anomalous anticyclone or cyclone centered around Japan. ENSO is possibly the source of the interannual variability of the GPP R-PC2 of spring to autumn. In response, a WPO-like circulation, Lake Baikal blocking, and an anomalous WPSH are closely related to the GPP R-PC2 of spring, summer and autumn, respectively. Further analysis indicated that these remote relationships between GPP and both SST and large-scale circulation are regulated by the local hydrothermal conditions of rainfall, temperature, soil moisture, and so on.

      (3) The fractions of variance of seasonal GPP explained by the key SST and circulation factors are large, with an average of 13%, 23% and 19% over China during spring, summer and autumn, respectively. The spatial distributions of the fractions of variance explained by the key SST and circulation factors show maximum centers (larger than 30%) over central-eastern and southwestern China in spring, central-eastern and northeastern China in summer, and southern China in autumn. Meanwhile, these are the key regions that dominate the interannual GPP variability for the country as a whole.

      Our findings above are based on analyses of multi-model ensemble datasets from TRENDY. A critical question is whether the GPP–teleconnection relationships exist in each individual model. To address this question, we performed the same analysis for the seven individual models as those that were done for the multi-model ensemble from TRENDY. Based on the seven individual model outputs, the spatial structures of the contemporary GPP correlations associated with the two dominant R-PC from spring to autumn (Figs. S2 and S3 in the ESM) are quite similar to that for the multi-model ensemble from TRENDY (left column of Fig. 3). In addition, when examining the correlation coefficients between the GPP R-PC based on the seven individual model outputs and the simultaneous (or lagged) climate indices, the significant GPP–teleconnection relationships observed in the multi-model ensemble from TRENDY could also be seen for the seven individual models (Tables S1 and S2 in the ESM) in most cases. Although as expected, the GPP–teleconnection correlation values for the multi-model ensemble from TRENDY are in-between those for the individual models, this nevertheless indicates that the GPP–teleconnection relationships in TRENDY are generally robust for certain individual models.

      Our work suggests that the AO, PDO, ENSO, WPO, Lake Baikal blocking and WPSH are worthy of attention in terms of the interannual variability of the seasonal mean GPP in the Chinese mainland during spring–summer–autumn, spring–summer–autumn, spring–summer–autumn, spring, summer and autumn, respectively. Also, we suggest that there are considerable seasonal differences in GPP–teleconnection relationships from spring to autumn, corresponding to local hydrothermal conditions. These findings help improve understanding of the interannual variability of China’s GPP. Also, identifying the SST and large-scale circulation factors are key for improving the seasonal forecasting of GPP, as these slowly varying external forcing factors and internal dynamics could have persistent influences on China’s climate (see Table 3 and Text S1 in the ESM). Thus, it provides an excellent way to estimate and project changes in the carbon cycle across China and would also provide scientific support for achieving carbon neutrality targets.

      It is important to acknowledge that our finding of an effect of external forcing and internal dynamics on GPP has been reached without consideration of land-use change. If land-use change and field irrigation are included, it is possible that the effects of ocean–atmosphere teleconnections may prove to be smaller than found in our study. Nonetheless, our results highlight the effect of ocean–atmosphere teleconnections on national-scale GPP in natural ecosystems, leading to a recommendation for a more substantial focus on understanding this process in the biosphere. In particular, it should be noted that five out seven models did not include nitrogen (N) cycle in this study. In addition, the TRENDY S2 simulation in this study did not include the phosphorus (P) cycle. Thus, our simulation did not consider the effects of N and P limitations on GPP. Ignoring these limitations may have resulted in an overestimation of GPP (Peng et al., 2020; Wieder et al., 2015). Therefore, in future work, to further verify the present reported results, we plan to carry out a similar analysis of the simulations but with due consideration paid to P limitation.

      Acknowledgements. This work was supported by National Natural Science Foundation of China (Grant No. 42141017), National Basic Research Program of China (Grant No. 2020YFA0608904) and the National Natural Science Foundation of China (Grant Nos. 41975112, 42175142, 42175013, and 41630532).

      Electronic supplementary material: Supplementary material is available in the online version of this article at https://doi.org/10.1007/s00376-021-1226-4.

Reference

Catalog

    /

    DownLoad:  Full-Size Img  PowerPoint
    Return
    Return