• Adler, R. F., and et al. , 2003: The Version-2 Global Precipitation Climatology Project (GPCP) Monthly Precipitation Analysis (1979–present). J. Hydrometeor., 4, 11471167, doi:10.1175/1525-7541(2003)004<1147:TVGPCP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Allen, M. R., , and W. J. Ingram, 2002: Constraints on future changes in climate and the hydrologic cycle. Nature, 419, 224232, doi:10.1038/nature01092.

    • Search Google Scholar
    • Export Citation
  • Anav, A., and et al. , 2013: Evaluating the land and ocean components of the global carbon cycle in the CMIP5 Earth system models. J. Climate, 26, 68016843, doi:10.1175/JCLI-D-12-00417.1.

    • Search Google Scholar
    • Export Citation
  • Arora, V. K., and et al. , 2011: Carbon emission limits required to satisfy future representative concentration pathways of greenhouse gases. Geophys. Res. Lett., 38, L05805, doi:10.1029/2010GL046270.

    • Search Google Scholar
    • Export Citation
  • Arora, V. K., and et al. , 2013: Carbon–concentration and carbon– climate feedbacks in CMIP5 Earth system models. J. Climate, 26, 52895314, doi:10.1175/JCLI-D-12-00494.1.

    • Search Google Scholar
    • Export Citation
  • Bacastow, R., 1976: Modulation of atmospheric carbon dioxide by the Southern Oscillation. Nature, 261, 116118, doi:10.1038/261116a0.

  • Behrenfeld, M. J., and et al. , 2001: Biospheric primary production during an ENSO transition. Science, 291, 25942597, doi:10.1126/science.1055071.

    • Search Google Scholar
    • Export Citation
  • Bellenger, H., , E. Guilyardi, , J. Leloup, , M. Lengaigne, , and J. Vialard, 2013: ENSO representation in climate models: From CMIP3 to CMIP5. Climate Dyn., 42, 19992018, doi:10.1007/s00382-013-1783-z.

    • Search Google Scholar
    • Export Citation
  • Booth, B. B. B., and et al. , 2012: High sensitivity of future global warming to land carbon cycle processes. Environ. Res. Lett., 7, 024002, doi:10.1088/1748-9326/7/2/024002.

    • Search Google Scholar
    • Export Citation
  • Bousquet, P., , P. Peylin, , P. Ciais, , C. Le Quere, , P. Friedlingstein, , and P. P. Tans, 2000: Regional changes in carbon dioxide fluxes of land and oceans since 1980. Science, 290, 13421346, doi:10.1126/science.290.5495.1342.

    • Search Google Scholar
    • Export Citation
  • Braswell, B. H., , D. S. Schimel, , E. Linder, , and B. Moore, 1997: The response of global terrestrial ecosystems to interannual temperature variability. Science, 278, 870872, doi:10.1126/science.278.5339.870.

    • Search Google Scholar
    • Export Citation
  • Brovkin, V., , T. Raddatz, , C. H. Reick, , M. Claussen, , and V. Gayler, 2009: Global biogeophysical interactions between forest and climate. Geophys. Res. Lett., 36, L07405, doi:10.1029/2009GL037543.

    • Search Google Scholar
    • Export Citation
  • Cao, M. K., , S. D. Prince, , B. Tao, , J. Small, , and K. R. Li, 2005: Regional pattern and interannual variations in global terrestrial carbon uptake in response to changes in climate and atmospheric CO2. Tellus, 57B, 210217, doi:10.1111/j.1600-0889.2005.00146.x.

    • Search Google Scholar
    • Export Citation
  • Chang, C.-P., , Z. Wang, , J. McBride, , and C.-H. Liu, 2005: Annual cycle of Southeast Asia—Maritime Continent rainfall and the asymmetric monsoon transition. J. Climate, 18, 287301, doi:10.1175/JCLI-3257.1.

    • Search Google Scholar
    • Export Citation
  • Clark, D. A., , S. C. Piper, , C. D. Keeling, , and D. B. Clark, 2003: Tropical rain forest tree growth and atmospheric carbon dynamics linked to interannual temperature variation during 1984–2000. Proc. Natl. Acad. Sci. USA, 100, 58525857, doi:10.1073/pnas.0935903100.

    • Search Google Scholar
    • Export Citation
  • Clark, D. A., , D. B. Clark, , and S. F. Oberbauer, 2013: Field-quantified responses of tropical rainforest aboveground productivity to increasing CO2 and climatic stress, 1997–2009. J. Geophys. Res. Biogeosci., 118, 783794, doi:10.1002/jgrg.20067.

    • Search Google Scholar
    • Export Citation
  • Collins, M., , R. E. Chandler, , P. M. Cox, , J. M. Huthnance, , J. Rougier, , and D. B. Stephenson, 2012: Quantifying future climate change. Nat. Climate Change, 2, 403409, doi:10.1038/nclimate1414.

    • Search Google Scholar
    • Export Citation
  • Cox, P. M., , D. Pearson, , B. B. Booth, , P. Friedlingstein, , C. Huntingford, , C. D. Jones, , and C. M. Luke, 2013: Sensitivity of tropical carbon to climate change constrained by carbon dioxide variability. Nature, 494, 341344, doi:10.1038/nature11882.

    • Search Google Scholar
    • Export Citation
  • Doughty, C. E., , and M. L. Goulden, 2008: Are tropical forests near a high temperature threshold? J. Geophys. Res. Biogeosci., 113, G00B07, doi:10.1029/2007JG000632.

    • Search Google Scholar
    • Export Citation
  • Dunne, J. P., and et al. , 2012: GFDL’s ESM2 global coupled climate–carbon Earth system models. Part I: Physical formulation and baseline simulation characteristics. J. Climate, 25, 66466665, doi:10.1175/JCLI-D-11-00560.1.

    • Search Google Scholar
    • Export Citation
  • Dunne, J. P., and et al. , 2013: GFDL’s ESM2 global coupled climate–carbon Earth system models. Part II: Carbon system formulation and baseline simulation characteristics. J. Climate, 26, 22472267, doi:10.1175/JCLI-D-12-00150.1.

    • Search Google Scholar
    • Export Citation
  • Fang, J. Y., , S. L. Piao, , Z. Y. Tang, , C. H. Peng, , and J. Wei, 2001: Interannual variability in net primary production and precipitation. Science, 293, 1723, doi:10.1126/science.293.5536.1723a.

    • Search Google Scholar
    • Export Citation
  • Feely, R. A., and et al. , 2002: Seasonal and interannual variability of CO2 in the equatorial Pacific. Deep-Sea Res. II, 49, 24432469, doi:10.1016/S0967-0645(02)00044-9.

    • Search Google Scholar
    • Export Citation
  • Friedlingstein, P., and et al. , 2006: Climate–carbon cycle feedback analysis: Results from the C4MIP model intercomparison. J. Climate, 19, 33373353, doi:10.1175/JCLI3800.1.

    • Search Google Scholar
    • Export Citation
  • Friedlingstein, P., , M. Meinshausen, , V. K. Arora, , C. D. Jones, , A. Anav, , S. K. Liddicoat, , and R. Knutti, 2014: Uncertainties in CMIP5 climate projections due to carbon cycle feedbacks. J. Climate, 27, 511526, doi:10.1175/JCLI-D-12-00579.1.

    • Search Google Scholar
    • Export Citation
  • Gu, G., , and R. F. Adler, 2011: Precipitation and temperature variations on the interannual time scale: Assessing the impact of ENSO and volcanic eruptions. J. Climate, 24, 22582270, doi:10.1175/2010JCLI3727.1.

    • Search Google Scholar
    • Export Citation
  • Guan, K. Y., , A. Wolf, , D. Medvigy, , K. K. Caylor, , M. Pan, , and E. F. Wood, 2013: Seasonal coupling of canopy structure and function in African tropical forests and its environmental controls. Ecosphere, 4, 35, doi:10.1890/ES12-00232.1.

    • Search Google Scholar
    • Export Citation
  • Guan, K. Y., and et al. , 2015: Photosynthetic seasonality of global tropical forests constrained by hydroclimate. Nat. Geosci., 8, 284289, doi:10.1038/ngeo2382.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., , H. Bellenger, , M. Collins, , S. Ferrett, , W. Cai, , and A. Wittenberg, 2012: A first look at ENSO in CMIP5. CLIVAR Exchanges, No. 17, International CLIVAR Project Office, Southampton, United Kingdom, 2932.

  • Gurney, K. R., , D. Baker, , P. Rayner, , and S. Denning, 2008: Interannual variations in continental-scale net carbon exchange and sensitivity to observing networks estimated from atmospheric CO2 inversions for the period 1980 to 2005. Global Biogeochem. Cycles, 22, GB3025, doi:10.1029/2007GB003082.

    • Search Google Scholar
    • Export Citation
  • Gurney, K. R., , K. Castillo, , B. Li, , and X. Zhang, 2012: A positive carbon feedback to ENSO and volcanic aerosols in the tropical terrestrial biosphere. Global Biogeochem. Cycles, 26, GB1029, doi:10.1029/2011GB004129.

    • Search Google Scholar
    • Export Citation
  • Halpert, M. S., , and C. F. Ropelewski, 1992: Surface temperature patterns associated with the Southern Oscillation. J. Climate, 5, 577593, doi:10.1175/1520-0442(1992)005<0577:STPAWT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hashimoto, H., and et al. , 2004: El Niño–Southern Oscillation-induced variability in terrestrial carbon cycling. J. Geophys. Res., 109, D23110, doi:10.1029/2004JD004959.

    • Search Google Scholar
    • Export Citation
  • Hoffman, F. M., and et al. , 2014: Causes and implications of persistent atmospheric carbon dioxide biases in Earth system models. J. Geophys. Res. Biogeosci., 119, 141162, doi:10.1002/2013JG002381.

    • Search Google Scholar
    • Export Citation
  • Huete, A. R., , N. Restrepo-Coupe, , P. Ratana, , K. Didan, , S. R. Saleska, , K. Ichii, , S. Panuthai, , and M. Gamo, 2008: Multiple site tower flux and remote sensing comparisons of tropical forest dynamics in monsoon Asia. Agric. For. Meteor., 148, 748760, doi:10.1016/j.agrformet.2008.01.012.

    • Search Google Scholar
    • Export Citation
  • Hurrell, J. W., and et al. , 2013: The Community Earth System Model: A framework for collaborative research. Bull. Amer. Meteor. Soc., 94, 13391360, doi:10.1175/BAMS-D-12-00121.1.

    • Search Google Scholar
    • Export Citation
  • Jones, C. D., , and P. M. Cox, 2005: On the significance of atmospheric CO2 growth rate anomalies in 2002–2003. Geophys. Res. Lett., 32, L14816, doi:10.1029/2005GL023027.

    • Search Google Scholar
    • Export Citation
  • Jones, C. D., , M. Collins, , P. M. Cox, , and S. A. Spall, 2001: The carbon cycle response to ENSO: A coupled climate–carbon cycle model study. J. Climate, 14, 41134129, doi:10.1175/1520-0442(2001)014<4113:TCCRTE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Jungclaus, J. H., and et al. , 2006: Ocean circulation and tropical variability in the coupled model ECHAM5/MPI-OM. J. Climate, 19, 39523972, doi:10.1175/JCLI3827.1.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and et al. , 1996: The NCEP/NCAR Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Keeling, C. D., , and R. Revelle, 1985: Effects of El Niño/Southern Oscillation on the atmospheric content of carbon dioxide. Meteoritics, 20, 437450.

    • Search Google Scholar
    • Export Citation
  • Keeling, C. D., , T. P. Whorf, , M. Wahlen, , and J. Vanderplicht, 1995: Interannual extremes in the rate of rise of atmospheric carbon dioxide since 1980. Nature, 375, 666670, doi:10.1038/375666a0.

    • Search Google Scholar
    • Export Citation
  • Keppel-Aleks, G., and et al. , 2014: Separating the influence of temperature, drought, and fire on interannual variability in atmospheric CO2. Global Biogeochem. Cycles, 28, 12951310, doi:10.1002/2014GB004890.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G. N., , and H. F. Diaz, 1989: Global climatic anomalies associated with extremes in the Southern Oscillation. J. Climate, 2, 10691090, doi:10.1175/1520-0442(1989)002<1069:GCAAWE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., , B. J. Soden, , and N. C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917932, doi:10.1175/1520-0442(1999)012<0917:RSSTVD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Krinner, G., and et al. , 2005: A dynamic global vegetation model for studies of the coupled atmosphere-biosphere system. Global Biogeochem. Cycles, 19, GB1015, doi:10.1029/2003GB002199.

    • Search Google Scholar
    • Export Citation
  • Kug, J.-S., , Y.-G. Ham, , J.-Y. Lee, , and F.-F. Jin, 2012: Improved simulation of two types of El Niño in CMIP5 models. Environ. Res. Lett., 7, 034002, doi:10.1088/1748-9326/7/3/034002.

    • Search Google Scholar
    • Export Citation
  • Lasslop, G., , K. Thonicke, , and S. Kloster, 2014: SPITFIRE within the MPI Earth system model: Model development and evaluation. J. Adv. Model. Earth Syst., 6, 740755, doi:10.1002/2013MS000284.

    • Search Google Scholar
    • Export Citation
  • Lee, K., , R. Wanninkhof, , T. Takahashi, , S. C. Doney, , and R. A. Feely, 1998: Low interannual variability in recent oceanic uptake of atmospheric carbon dioxide. Nature, 396, 155159, doi:10.1038/24139.

    • Search Google Scholar
    • Export Citation
  • Le Quéré, C., and et al. , 2003: Two decades of ocean CO2 sink and variability. Tellus, 55B, 649656, doi:10.1034/j.1600-0889.2003.00043.x.

    • Search Google Scholar
    • Export Citation
  • Le Quéré, C., and et al. , 2009: Trends in the sources and sinks of carbon dioxide. Nat. Geosci., 2, 831836, doi:10.1038/ngeo689.

    • Search Google Scholar
    • Export Citation
  • Le Quéré, C., and et al. , 2015: Global carbon budget 2014. Earth Syst. Sci. Data, 7, 4785, doi:10.5194/essd-7-47-2015.

  • Li, F., , S. Levis, , and D. S. Ward, 2013: Quantifying the role of fire in the Earth system – Part 1: Improved global fire modeling in the Community Earth System Model (CESM1). Biogeosciences, 10, 22932314, doi:10.5194/bg-10-2293-2013.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , and J. A. Taylor, 1994: On the temperature dependence of soil respiration. Funct. Ecol., 8, 315323, doi:10.2307/2389824.

  • Malhi, Y., , and J. Wright, 2004: Spatial patterns and recent trends in the climate of tropical rainforest regions. Philos. Trans. Roy. Soc., 359B, 311329, doi:10.1098/rstb.2003.1433.

    • Search Google Scholar
    • Export Citation
  • Nagai, S., , K. Ichii, , and H. Morimoto, 2007: Interannual variations in vegetation activities and climate variability caused by ENSO in tropical rainforests. Int. J. Remote Sens., 28, 12851297, doi:10.1080/01431160600904972.

    • Search Google Scholar
    • Export Citation
  • 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, 15601563, doi:10.1126/science.1082750.

    • Search Google Scholar
    • Export Citation
  • Obata, A., , and K. Shibata, 2012: Damage of land biosphere due to intense warming by 1000-fold rapid increase in atmospheric methane: Estimation with a climate-carbon cycle model. J. Climate, 25, 85248541, doi:10.1175/JCLI-D-11-00533.1.

    • Search Google Scholar
    • Export Citation
  • Peel, M. C., , B. L. Finlayson, , and T. A. McMahon, 2007: Updated world map of the Köppen–Geiger climate classification. Hydrol. Earth Syst. Sci., 11, 16331644, doi:10.5194/hess-11-1633-2007.

    • Search Google Scholar
    • Export Citation
  • Peylin, P., and et al. , 2005: Multiple constraints on regional CO2 flux variations over land and oceans. Global Biogeochem. Cycles, 19, GB1011, doi:10.1029/2003GB002214.

    • Search Google Scholar
    • Export Citation
  • Piao, S. L., , P. Ciais, , P. Friedlingstein, , N. de Noblet-Ducoudre, , P. Cadule, , N. Viovy, , and T. Wang, 2009: Spatiotemporal patterns of terrestrial carbon cycle during the 20th century. Global Biogeochem. Cycles, 23, GB4026, doi:10.1029/2008GB003339.

    • Search Google Scholar
    • Export Citation
  • Piao, S. L., and et al. , 2013: Evaluation of terrestrial carbon cycle models for their response to climate variability and to CO2 trends. Global Change Biol., 19, 21172132, doi:10.1111/gcb.12187.

    • Search Google Scholar
    • Export Citation
  • Potter, C., , S. Klooster, , M. Steinbach, , P. Tan, , V. Kumar, , S. Shekhar, , R. Nemani, , and R. Myneni, 2003: Global teleconnections of climate to terrestrial carbon flux. J. Geophys. Res., 108, 4556, doi:10.1029/2002JD002979.

    • Search Google Scholar
    • Export Citation
  • Poulter, B., and et al. , 2014: Contribution of semi-arid ecosystems to interannual variability of the global carbon cycle. Nature, 509, 600603, doi:10.1038/nature13376.

    • Search Google Scholar
    • Export Citation
  • Poulter, B., and et al. , 2015: Sensitivity of global terrestrial carbon cycle dynamics to variability in satellite-observed burned area. Global Biogeochem. Cycles, 29, 207222, doi:10.1002/2013GB004655.

    • Search Google Scholar
    • Export Citation
  • Prentice, I. C., , D. I. Kelley, , P. N. Foster, , P. Friedlingstein, , S. P. Harrison, , and P. J. Bartlein, 2011: Modeling fire and the terrestrial carbon balance. Global Biogeochem. Cycles, 25, GB3005, doi:10.1029/2010GB003906.

    • Search Google Scholar
    • Export Citation
  • Qian, H., , R. Joseph, , and N. Zeng, 2008: Response of the terrestrial carbon cycle to the El Niño–Southern Oscillation. Tellus, 60B, 537550, doi:10.1111/j.1600-0889.2008.00360.x.

    • Search Google Scholar
    • Export Citation
  • Raddatz, T. J., and et al. , 2007: Will the tropical land biosphere dominate the climate–carbon cycle feedback during the twenty-first century? Climate Dyn., 29, 565574, doi:10.1007/s00382-007-0247-8.

    • Search Google Scholar
    • Export Citation
  • Raupach, M. R., , J. G. Canadell, , and C. Le Quéré, 2008: Anthropogenic and biophysical contributions to increasing atmospheric CO2 growth rate and airborne fraction. Biogeosciences, 5, 16011613, doi:10.5194/bg-5-1601-2008.

    • Search Google Scholar
    • Export Citation
  • Rayner, P. J., , and R. M. Law, 1999: The interannual variability of the global carbon cycle. Tellus, 51B, 210212, doi:10.1034/j.1600-0889.1999.t01-1-00007.x.

    • Search Google Scholar
    • Export Citation
  • Rayner, P. J., , R. M. Law, , and R. Dargaville, 1999: The relationship between tropical CO2 fluxes and the El Niño–Southern Oscillation. Geophys. Res. Lett., 26, 493496, doi:10.1029/1999GL900008.

    • Search Google Scholar
    • Export Citation
  • Renwick, J. A., , and J. M. Wallace, 1996: Relationships between North Pacific wintertime blocking, El Niño, and the PNA pattern. Mon. Wea. Rev., 124, 20712076, doi:10.1175/1520-0493(1996)124<2071:RBNPWB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rödenbeck, C., , S. Houweling, , M. Gloor, , and M. Heimann, 2003: CO2 flux history 1982–2001 inferred from atmospheric data using a global inversion of atmospheric transport. Atmos. Chem. Phys., 3, 19191964, doi:10.5194/acp-3-1919-2003.

    • Search Google Scholar
    • Export Citation
  • Ropelewski, C. F., , and M. S. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., 115, 16061626, doi:10.1175/1520-0493(1987)115<1606:GARSPP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sarmiento, J. L., , M. Gloor, , N. Gruber, , C. Beaulieu, , A. R. Jacobson, , S. E. Mikaloff, , Fletcher, , S. Pacala, , and K. Rodgers, 2010: Trends and regional distributions of land and ocean carbon sinks. Biogeosciences, 7, 23512367, doi:10.5194/bg-7-2351-2010.

    • Search Google Scholar
    • Export Citation
  • Schwalm, C. R., , C. A. Williams, , K. Schaefer, , I. Baker, , G. J. Collatz, , and C. Rödenbeck, 2011: Does terrestrial drought explain global CO2 flux anomalies induced by El Niño? Biogeosciences, 8, 24932506, doi:10.5194/bg-8-2493-2011.

    • Search Google Scholar
    • Export Citation
  • Shao, P., , X. B. Zeng, , K. Sakaguchi, , R. K. Monson, , and X. D. Zeng, 2013: Terrestrial carbon cycle: Climate relations in eight CMIP5 Earth system models. J. Climate, 26, 87448764, doi:10.1175/JCLI-D-12-00831.1.

    • Search Google Scholar
    • Export Citation
  • Sitch, S., and et al. , 2008: Evaluation of the terrestrial carbon cycle, future plant geography, and climate-carbon cycle feedbacks using five dynamic global vegetation models (DGVMs). Global Change Biol., 14, 20152039, doi:10.1111/j.1365-2486.2008.01626.x.

    • Search Google Scholar
    • Export Citation
  • Smith, T. M., , and R. W. Reynolds, 2004: Improved extended reconstruction of SST (1854–1997). J. Climate, 17, 24662477, doi:10.1175/1520-0442(2004)017<2466:IEROS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., , R. J. Stouffer, , and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485498, doi:10.1175/BAMS-D-11-00094.1.

    • Search Google Scholar
    • Export Citation
  • Thoning, K. W., , P. P. Tans, , and W. D. Komhy, 1989: Atmospheric carbon dioxide at Mauna Loa Observatory: 2. Analysis of the NOAA GMCC data, 1974–1985. J. Geophys. Res., 94, 85498565, doi:10.1029/JD094iD06p08549.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1997: The definition of El Niño. Bull. Amer. Meteor. Soc., 78, 27712777, doi:10.1175/1520-0477(1997)078<2771:TDOENO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., , J. M. Caron, , D. P. Stepaniak, , and S. Worley, 2002: Evolution of El Niño–Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107, 4065, doi:10.1029/2000JD000298.

    • Search Google Scholar
    • Export Citation
  • Tsai, C., , S. K. Behera, , and T. Waseda, 2015: South Asia monsoon indices. Sci. Rep., 5, 8107, doi:10.1038/srep08107.

  • Van der Werf, G. R., , J. T. Randerson, , L. Giglio, , G. J. Collatz, , P. S. Kasibhatla, , and A. F. Arellano, 2006: Interannual variability in global biomass burning emissions from 1997 to 2004. Atmos. Chem. Phys., 6, 34233441, doi:10.5194/acp-6-3423-2006.

    • Search Google Scholar
    • Export Citation
  • Van der Werf, G. R., and et al. , 2010: Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires (1997–2009). Atmos. Chem. Phys., 10, 11 70711 735, doi:10.5194/acp-10-11707-2010.

    • Search Google Scholar
    • Export Citation
  • Wang, J., , P. M. Rich, , and K. P. Price, 2003: Temporal responses of NDVI to precipitation and temperature in the central Great Plains, USA. Int. J. Remote Sens., 24, 23452364, doi:10.1080/01431160210154812.

    • Search Google Scholar
    • Export Citation
  • Wang, J., , N. Zeng, , Y. Liu, , and Q. Bao, 2014: To what extent can interannual CO2 variability constrain carbon cycle sensitivity to climate change in CMIP5 Earth system models? Geophys. Res. Lett., 41, 35353544, doi:10.1002/2014GL060004.

    • Search Google Scholar
    • Export Citation
  • Wang, W. L., and et al. , 2013: Variations in atmospheric CO2 growth rates coupled with tropical temperature. Proc. Natl. Acad. Sci. USA, 110, 13 06113 066, doi:10.1073/pnas.1219683110.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., and et al. , 2010: Improved climate simulation by MIROC5: Mean states, variability, and climate sensitivity. J. Climate, 23, 63126335, doi:10.1175/2010JCLI3679.1.

    • Search Google Scholar
    • Export Citation
  • Watanabe, S., and et al. , 2011: MIROC-ESM 2010: Model description and basic results of CMIP5-20c3m experiments. Geosci. Model Dev., 4, 845872, doi:10.5194/gmd-4-845-2011.

    • Search Google Scholar
    • Export Citation
  • Weare, B. C., 2013: El Niño teleconnections in CMIP5 models. Climate Dyn., 41, 21652177, doi:10.1007/s00382-012-1537-3.

  • Wenzel, S., , P. M. Cox, , V. Eyring, , and P. Friedlingstein, 2014: Emergent constraints on climate-carbon cycle feedbacks in the CMIP5 Earth system models. J. Geophys. Res. Biogeosci., 119, 794807, doi:10.1002/2013JG002591.

    • Search Google Scholar
    • Export Citation
  • Wetzel, P., , A. Winguth, , and E. Maier-Reimer, 2005: Sea-to-air CO2 flux from 1948 to 2003: A model study. Global Biogeochem. Cycles, 19, GB2005, doi:10.1029/2004GB002339.

    • Search Google Scholar
    • Export Citation
  • Wittenberg, A. T., , A. Rosati, , N. C. Lau, , and J. J. Ploshay, 2006: GFDL’s CM2 global coupled climate models. Part III: Tropical Pacific climate and ENSO. J. Climate, 19, 698722, doi:10.1175/JCLI3631.1.

    • Search Google Scholar
    • Export Citation
  • Yeh, S. W., , Y. G. Ham, , and J. Y. Lee, 2012: Changes in the tropical Pacific SST trend from CMIP3 to CMIP5 and its implication of ENSO. J. Climate, 25, 77647771, doi:10.1175/JCLI-D-12-00304.1.

    • Search Google Scholar
    • Export Citation
  • Zeng, N., , A. Mariotti, , and P. Wetzel, 2005: Terrestrial mechanisms of interannual CO2 variability. Global Biogeochem. Cycles, 19, GB1016, doi:10.1029/2004GB002273.

    • Search Google Scholar
    • Export Citation
  • Zeng, N., and et al. , 2008: Dynamical prediction of terrestrial ecosystems and the global carbon cycle: A 25-year hindcast experiment. Global Biogeochem. Cycles, 22, GB4015, doi:10.1029/2008GB003183.

    • Search Google Scholar
    • Export Citation
  • Zhao, M., , and S. W. Running, 2010: Drought-induced reduction in global terrestrial net primary production from 2000 through 2009. Science, 329, 940943, doi:10.1126/science.1192666.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    (top) Global CO2 growth rate and (bottom) land–atmosphere CO2 flux anomalies from the ESMs (colors), MME (thick black), and observations (thick gray) during (a),(c) El Niño and (b),(d) La Niña years.

  • View in gallery

    (top)–(bottom) Carbon flux anomalies (PgC yr−1 °C−1) due to NPP, Rh, and fire from the ESMs (colors), MME (thick black), and observations (thick gray) during (a),(c),(e) El Niño and (b),(d),(f) La Niña events.

  • View in gallery

    MME composite maps of CO2 flux due to NPP anomalies during El Niño events: (a) JJA(0), (b) SON(0), (c) D(0)JF(1), (d) MAM(1), (e) JJA(1), and (f) SON(1). Boxes in (d) indicate regions in Table 3 [Amazonia (AMZ), Australia (AUS), equatorial Asia (EQA), South Asia (SOA), and Africa (AFR)].

  • View in gallery

    Composite maps of mean precipitation anomaly for July(0)–June(1) during (a),(b) El Niño and (c),(d) La Niña years from the (left) observations and (right) MME.

  • View in gallery

    As in Fig. 4, but for the surface temperature anomaly.

  • View in gallery

    Scatterplots of D(0)JF(1) Niño-3.4 index vs mean precipitation anomaly over tropical land (20°S–20°N) for July(0)–June(1): (a) observations, (b) CanESM2, (c) CESM1-BGC, (d) GFDL-ESM2M, (e) MIROC-ESM, (f) MPI-ESM-LR, and (g) MRI-ESM1. The correlation and regression coefficient are shown in the top right of each panel. Red (blue) dots show El Niño (La Niña) years, classified using Niño-3.4 anomaly magnitudes >1°C (signs positive and negative, respectively). Black dots indicate normal years.

  • View in gallery

    As in Fig. 6, but for mean surface temperature anomaly.

  • View in gallery

    Scatterplots of mean precipitation anomaly over the tropical land (20°S–20°N) for July(0)–June(1) vs NPP anomalies in the tropical regions (20°S–20°N): (a) CanESM2, (b) CESM1-BGC, (c) GFDL-ESM2M, (d) MIROC-ESM, (e) MPI-ESM-LR, (f) MPI-ESM-LR, and (g) MRI-ESN1. The correlation and regression coefficient are shown in the top right of each panel. Red (blue) dots show El Niño (La Niña) years, classified using Niño-3.4 anomaly magnitudes >1°C (signs positive and negative, respectively). Black dots indicate normal years.

  • View in gallery

    As in Fig. 8, but for surface temperature anomaly.

  • View in gallery

    Composites of seasonal NPP anomalies averaged over the globe and regions denoted in Fig. 3d from MME (bars) and each ESM (dots) during (a) El Niño and (b) La Niña events. Star-like symbols indicate the season containing maximum NPP anomalies in each region.

  • View in gallery

    Regression coefficients of NPP anomalies on precipitation (blue) and temperature (red) in South Asia. Error bars indicate 95% significance levels based on the Student’s t test.

  • View in gallery

    (a) Multiple regression coefficients of NPP anomalies on precipitation and temperature in Amazonia, Australia, South Asia, and Africa. (b) Regression coefficients of precipitation and temperature to the Niño-3.4 index in each region. Bars, colored dots, and circles denote MME, each model, and observations, respectively.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 571 571 16
PDF Downloads 355 355 12

Increased Atmospheric CO2 Growth Rate during El Niño Driven by Reduced Terrestrial Productivity in the CMIP5 ESMs

View More View Less
  • 1 School of Environmental Science and Engineering, Pohang University of Science and Technology (POSTECH), Pohang, South Korea
  • | 2 Pacific Northwest National Laboratory, Richland, Washington
  • | 3 Jet Propulsion Laboratory, California Institute of Technology, Pasadena, California
© Get Permissions
Full access

Abstract

Better understanding of factors that control the global carbon cycle could increase confidence in climate projections. Previous studies found good correlation between the growth rate of atmospheric CO2 concentration and El Niño–Southern Oscillation (ENSO). The growth rate of atmospheric CO2 increases during El Niño but decreases during La Niña. In this study, long-term simulations of the Earth system models (ESMs) in phase 5 of the Coupled Model Intercomparison Project archive were used to examine the interannual carbon flux variability associated with ENSO. The ESMs simulate the relationship reasonably well with a delay of several months between ENSO and the changes in atmospheric CO2. The increase in atmospheric CO2 associated with El Niño is mostly caused by decreasing net primary production (NPP) in the ESMs. It is suggested that NPP anomalies over South Asia are at their maxima during boreal spring; therefore, the increase in CO2 concentration lags 4–5 months behind the peak phase of El Niño. The decrease in NPP during El Niño may be caused by decreased precipitation and increased temperature over tropical regions. Furthermore, systematic errors may exist in the ESM-simulated temperature responses to ENSO phases over tropical land areas, and these errors may lead to an overestimation of ENSO-related NPP anomalies. In contrast, carbon fluxes from heterotrophic respiration and natural fires are likely underestimated in the ESMs compared with offline model results and observational estimates, respectively. These uncertainties should be considered in long-term projections that include climate–carbon feedbacks.

Current affiliation: School of Earth Sciences and Environmental Engineering, Gwangju Institute of Science and Technology, Gwangju, South Korea.

Current affiliation: School of Environmental Science and Engineering, South University of Science and Technology of China, Shenzhen, China.

Corresponding author address: Prof. Jong-Seong Kug, School of Environment Science and Engineering, Pohang University of Science and Technology, Pohang 790-784, South Korea. E-mail: jskug1@gmail.com

Abstract

Better understanding of factors that control the global carbon cycle could increase confidence in climate projections. Previous studies found good correlation between the growth rate of atmospheric CO2 concentration and El Niño–Southern Oscillation (ENSO). The growth rate of atmospheric CO2 increases during El Niño but decreases during La Niña. In this study, long-term simulations of the Earth system models (ESMs) in phase 5 of the Coupled Model Intercomparison Project archive were used to examine the interannual carbon flux variability associated with ENSO. The ESMs simulate the relationship reasonably well with a delay of several months between ENSO and the changes in atmospheric CO2. The increase in atmospheric CO2 associated with El Niño is mostly caused by decreasing net primary production (NPP) in the ESMs. It is suggested that NPP anomalies over South Asia are at their maxima during boreal spring; therefore, the increase in CO2 concentration lags 4–5 months behind the peak phase of El Niño. The decrease in NPP during El Niño may be caused by decreased precipitation and increased temperature over tropical regions. Furthermore, systematic errors may exist in the ESM-simulated temperature responses to ENSO phases over tropical land areas, and these errors may lead to an overestimation of ENSO-related NPP anomalies. In contrast, carbon fluxes from heterotrophic respiration and natural fires are likely underestimated in the ESMs compared with offline model results and observational estimates, respectively. These uncertainties should be considered in long-term projections that include climate–carbon feedbacks.

Current affiliation: School of Earth Sciences and Environmental Engineering, Gwangju Institute of Science and Technology, Gwangju, South Korea.

Current affiliation: School of Environmental Science and Engineering, South University of Science and Technology of China, Shenzhen, China.

Corresponding author address: Prof. Jong-Seong Kug, School of Environment Science and Engineering, Pohang University of Science and Technology, Pohang 790-784, South Korea. E-mail: jskug1@gmail.com

1. Introduction

Interannual variability of the global carbon cycle is closely related to El Niño–Southern Oscillation (ENSO), which is characterized by anomalous sea surface warming and cooling in the eastern and central Pacific (Bacastow 1976; Keeling and Revelle 1985; Braswell et al. 1997; Rayner and Law 1999). ENSO is a conspicuous component of ocean–atmosphere interactions in the equatorial climate system, and has an enormous influence on the ecosystem and global carbon cycle (Potter et al. 2003; Nagai et al. 2007; Schwalm et al. 2011).

Based on a recent analysis of the global carbon budget, Le Quéré et al. (2015) reported that the amounts of CO2 absorbed by the oceans (1.9 ± 0.5 PgC yr−1) and the land (2.1 ± 0.8 PgC yr−1) have been roughly equivalent over the past 55 yr. However, on the interannual time scale, an anomalous oceanic uptake of carbon of ~0.5 PgC yr−1 occurs during an El Niño event, because of the reduced upwelling of dissolved inorganic carbon-enriched water in the equatorial eastern Pacific (Feely et al. 2002; Le Quéré et al. 2003; Wetzel et al. 2005), but these changes in oceanic carbon flux cannot explain the overall interannual variability of the global atmospheric CO2 concentration (Lee et al. 1998; Le Quéré et al. 2003, 2009). Actually, without the linear trend due to anthropogenic effects, the standard deviations of ocean and land uptake for the most recent 55 yr are 0.17 and 0.96 PgC yr−1, respectively (Le Quéré et al. 2015). Comparison of these values indicates that the oceanic fluxes are less sensitive than land–atmosphere fluxes to ENSO. In contrast, inverse modeling studies have indicated that terrestrial carbon flux is dominant regarding the variability of atmospheric CO2 associated with ENSO (Bousquet et al. 2000; Rödenbeck et al. 2003). Gurney et al. (2008) used an inverse model to estimate that ~4.0 PgC yr−1 of CO2 was released by land processes during the 1997/98 El Niño but that the uptake by the oceans was much smaller than this.

Regional temperature, precipitation, and radiation anomalies are important parameters that affect terrestrial carbon pools and fluxes, and that provide the conditions under which land–atmosphere CO2 flux increases under El Niño conditions (Malhi and Wright 2004; Jones and Cox 2005). Three major terrestrial mechanisms that affect atmospheric CO2 concentration are associated with ENSO: net primary production (NPP), microbial soil and litter decomposition respiration, and fire variation (Zeng et al. 2005).

Plants consume CO2 and emit O2 during photosynthesis, but they also emit CO2 during autotrophic respiration Ra. Thus, the NPP amount is conditioned by the difference between gross primary production (GPP; i.e., total biomass produced by photosynthesis) and Ra. Because NPP is the result of photosynthesis that consumes CO2, the atmospheric CO2 concentration decreases as the NPP anomaly increases (Zeng et al. 2005, 2008; Qian et al. 2008). Conversely Rh, which reflects respiration by microbes in the soil and litter decomposition, consumes O2 and releases CO2, and it is strongly dependent on surface temperature (Lloyd and Taylor 1994). During El Niño years, the surface temperature in tropical regions increases widely over ocean and land surfaces, especially in Amazonia and northeast Australia (Halpert and Ropelewski 1992). Therefore, the interannual variability of Rh in the tropics is closely correlated to ENSO (Zeng et al. 2005).

Natural fires also emit CO2 from terrestrial ecosystems into the atmosphere. Severe drought that persists for several months causes increases in the frequencies and intensities of fires in tropical forests (Van der Werf et al. 2006). Wide regions of forest burn during periods of continued drought; therefore, emissions from fire contribute to the CO2 growth rate during El Niño years.

The dominant processes that control terrestrial CO2 fluxes associated with ENSO have not been clearly identified, and two different perspectives have been expressed. The first is that temperature variation, by its effects on soil respiration and photosynthesis, contributes to the terrestrial variability associated with ENSO (Piao et al. 2009; Wang et al. 2013). The other is that precipitation change has a critical influence on terrestrial CO2 fluxes because it affects photosynthesis (Keeling et al. 1995; Behrenfeld et al. 2001; Fang et al. 2001; Jones et al. 2001; Rödenbeck et al. 2003; Peylin et al. 2005; Zeng et al. 2005; Qian et al. 2008; Piao et al. 2009).

The temperature of the entire tropical atmosphere in the eastern Pacific warms during an El Niño event, with a time delay (Malhi and Wright 2004; Gu and Adler 2011); thus, the tropical land temperature tends to increase during El Niño events. Satellite data and short-term leaf-level results reveal that NPP decreases as temperature increases (Zhao and Running 2010; Clark et al. 2013). This relationship might indicate that the current climate state in the tropics had already exceeded the high-temperature optimum threshold, such that the additional warming reduced the carbon uptake in tropical rain forests (Doughty and Goulden 2008).

Precipitation decreases significantly over equatorial land during El Niño events but increases during La Niña years (Ropelewski and Halpert 1987). This countertrend correlates with the reduced NPP anomaly during El Niño events and, therefore, suggests that decreased precipitation might be responsible for it.

Furthermore, natural fires that occur frequently during El Niño phases can contribute to CO2 variability; for example, 2.8 PgC was emitted by natural fires during the 1997/98 El Niño, whereas only 2.1 PgC yr−1 was emitted during the non–El Niño years of 2002–07 (Van der Werf et al. 2010). Therefore, the occurrence of natural fires should be considered a major factor that affects the interannual variation of the growth rate of atmospheric CO2 (Keppel-Aleks et al. 2014).

General circulation models have improved to the point where ENSO and the related teleconnections can be simulated reliably (Jungclaus et al. 2006; Wittenberg et al. 2006; Watanabe et al. 2010; Guilyardi et al. 2012; Kug et al. 2012; Bellenger et al. 2013; Weare 2013). Therefore, they can be expected to provide plausible oceanic and atmospheric conditions for carbon-cycle modeling. Many experiments concerning the carbon-cycle variability associated with ENSO have been conducted using offline carbon models; however, no results of fully coupled carbon cycle models have been reported. A new generation of climate models [Earth system models (ESMs)], which are incorporated as part of phase 5 of the Coupled Model Intercomparison Project (CMIP5), include not only the full atmospheric and oceanic physics and dynamics, but also interactive biogeochemistry and the carbon cycle (Taylor et al. 2012; Anav et al. 2013; Shao et al. 2013).

Because ESMs consider the interactions between the physics and the carbon cycle, long-term ESM simulations provide a great opportunity both to test existing hypotheses regarding the sources of ENSO-related CO2 variability, and to examine how well these ESMs simulate the observed carbon cycle. Simulations of interactions between the atmosphere and the other climatic components in the ESMs have the advantage that the results can be used to investigate the effects of ENSO on CO2 variation, because various feedback processes are included.

Some recent studies have used ESMs to examine how atmosphere–land interaction affects climate–carbon feedback (Friedlingstein et al. 2006, 2014; Hoffman et al. 2014). Friedlingstein et al. (2014) revealed that the uncertainty in the projected atmospheric CO2 mixing ratio by the end of the twenty-first century is >350 ppmv for the representative concentration pathway 8.5 (RCP8.5) scenario. The magnitude of this value is largely attributable to the uncertainty in the contribution from land components (Arora et al. 2013). The sensitivity of land carbon to temperature changes might affect future projections considerably (Booth et al. 2012).

The observable features of the Earth system can be used to constrain the projections; this is the so-called emergent constraint. It identifies the relationship between observable contemporary variability and future sensitivity, such that the relationship may enable observations to constrain the estimate of the model’s sensitivity (Allen and Ingram 2002; Collins et al. 2012). For example, interannual CO2 growth rate has a strong linear relationship with tropical land carbon uptake in both the observations and the ESMs; therefore, a comparison between the observations and the ESM output could be used to constrain the ESM carbon-cycle projections. Simulation by the Coupled Climate–Carbon Cycle Model Intercomparison Project (C4MIP) revealed that the long-term sensitivity of tropical land carbon storage to climate warming had an emergent linear relationship with the short-term sensitivity of atmospheric CO2 to interannual temperature variability (Cox et al. 2013). The CMIP5 simulations suggest that a tight correlation between the long-term sensitivity and the short-term sensitivity enables the projections to be constrained by reference to observations (Wenzel et al. 2014). In contrast, Wang et al. (2014) found that changes to the carbon cycle in the terrestrial tropics were weakly linearly linked to the sensitivity of the historical atmospheric CO2 growth rate to tropical temperature variability. Therefore, they concluded that this emergent constraint is less effective than expected in reducing the models’ uncertainties. These contrasting results imply that current knowledge regarding both the long-term and short-term sensitivities of the carbon cycle and the processes in ESMs is immature.

Here, we examine the outcomes of ESM simulations to analyze how the interannual CO2 growth rate is associated with ENSO. In particular, we focused on terrestrial anomalies related to changes in temperature and precipitation with detailed analyses of regional contributions, which can be good metrics for the evaluation of emergent model fidelity to the interannual carbon cycle. Analyses of the ENSO-related carbon cycle as simulated by the ESMs provided an evaluation of the carbon-cycle simulation and suggested possible mechanisms by which ENSO modulates carbon-cycle variability.

2. Data and method

The CMIP5 includes various ESM simulations. The ESMs consider physical, chemical, biological, and anthropogenic processes that occur within the Earth system, including the interactions among the atmosphere, ocean, land surface, and sea ice. CMIP5 has been used for various experiments of long-term integrations to improve our understanding of climate change and variability. In the present study, the “esmControl” run of the ESM was used to examine the ENSO-related interannual CO2 variability without anthropogenic effects. This run is based on the preindustrial climate status and relevant greenhouse gas forcing, and estimates the unforced variability of the model and diagnoses climate drift in the unforced case (Taylor et al. 2012). Hence, anthropogenic greenhouse gas emission is kept at the preindustrial level, and total atmospheric CO2 is determined by the balance between carbon sources and sinks.

The simulations of the esmControl run were analyzed using six different ESMs [CanESM2, Community Earth System Model, version 1–Biogeochemistry (CESM1-BGC), GFDL-ESM2M, MIROC-ESM, MPI-ESM-LR, and MRI-ESM1; Table 1; expansions of acronyms are available online at http://www.ametsoc.org/PubsAcronymList)], all of which consider the global carbon cycle. To investigate the interannual variability of atmospheric CO2 associated with ENSO, El Niño and La Niña years were defined based on the Niño-3.4 (5°S–5°N, 170°–120°W) index of sea surface temperature (SST) in the ESMs and the observations. El Niño years were defined when the Niño-3.4 SST anomaly during December–February (DJF) was >1°C, and La Niña years were defined when it was <−1°C (Trenberth 1997). In this paper, year (0) refers to the developing year and year (1) refers to the decaying year of ENSO.

Table 1.

Description of the CMIP5 ESMs used in this study.

Table 1.

In this study, atmospheric CO2 (total mass and mole fraction), precipitation, land and sea surface temperatures, and four measures of carbon mass flux into the atmosphere (Table 2) were considered. The CMIP5 ESMs provide variables for carbon uptake, which are defined as carbon fluxes between the atmosphere and each domain (negative values mean flux into the atmosphere from the land). Because the magnitudes of the ENSO events differ among the models, anomalies of all variables are divided by the Niño-3.4 index of the composites to show normalized anomalies with respect to the Niño-3.4 index of magnitude 1°C; therefore, the units of carbon flux anomalies are PgC yr−1 °C−1. The multimodel ensemble (MME) was calculated by taking the mean of the composite results of the six ESMs.

Table 2.

Variables used in the CMIP5 ESMs.

Table 2.

The observational SST dataset used was the NOAA extended reconstruction of historical SST version 3 (Smith and Reynolds 2004) from the National Climatic Data Center. Monthly CO2 growth rate at Mauna Loa, Hawaii, was taken from NOAA’s Earth System Research Laboratory website (ftp://aftp.cmdl.noaa.gov/products/trends/co2/co2_mm_mlo.txt) (Thoning et al. 1989). Precipitation data were obtained from the Global Precipitation Climatology Project (Adler et al. 2003). Surface temperature anomalies are described using the National Centers for Environmental Prediction–National Center for Atmospheric Research reanalysis (Kalnay et al. 1996). Observational datasets for the 34 yr from 1979 to 2012 inclusive were used to evaluate the model results and for comparison purposes. To exclude the effects of anthropogenic emissions on atmospheric CO2 growth rate, linear trends in the magnitude of the CO2 growth rate over time were removed.

3. Carbon fluxes associated with ENSO

To examine how each model captures the relationship between ENSO and atmospheric CO2, the composite atmospheric CO2 growth rate (i.e., the time derivative of the monthly CO2 concentrations for El Niño and La Niña events) was calculated (Fig. 1). The CO2 growth rate was mostly positive during El Niño events. The MME showed that the maximum CO2 growth rate increase was ~1.5 PgC yr−1 °C−1, which is higher than the rate of ~1.22 PgC yr−1 °C−1 calculated from the observational data (Qian et al. 2008). Assuming that 56% of anthropogenic emissions remain airborne, the sensitivity of the annual CO2 growth rate to the annual mean Niño-3 index was 1.27 PgC yr−1 °C−1 during the interval 1967–81, which was free from volcanic influence (Jones et al. 2001). Compared with previous observational estimates (e.g., Hashimoto et al. 2004; Qian et al. 2008), the MME overestimates the CO2 growth rate associated with ENSO. However, the ESM simulations were performed under preindustrial conditions, so they did not include ongoing climate change signals. Therefore, caution must be applied when comparing the simulated relation to the present-day observational estimate.

Fig. 1.
Fig. 1.

(top) Global CO2 growth rate and (bottom) land–atmosphere CO2 flux anomalies from the ESMs (colors), MME (thick black), and observations (thick gray) during (a),(c) El Niño and (b),(d) La Niña years.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

It is found that the maximum CO2 growth rate appears during the boreal spring of an El Niño decay phase; that is, the pattern is delayed by several months from the El Niño peak, which occurs during boreal winter. This result is consistent with observations (Qian et al. 2008). During La Niña events (Fig. 1b), all models simulated both a negative CO2 growth rate and a delayed timing of the maximum CO2; these results are also consistent with observations (Rayner et al. 1999).

The seasonal evolution and magnitude of the composites of CO2 fluxes from the land to the atmosphere during El Niño (Fig. 1c) and La Niña (Fig. 1d) are quite similar to the global CO2 mass values, not only for the MME, but also for the individual models (e.g., Figs. 1a,b versus Figs. 1c,d). Oceanic fluxes vary little and do not clearly change with ENSO phases (not shown). This conclusion is consistent with the findings of previous observational studies (Sarmiento et al. 2010) and implies that carbon flux from the land is a major contributor to the interannual variability of the atmospheric CO2 growth rate associated with ENSO.

Although the MME captures the changes in CO2 growth rate associated with ENSO reasonably well, the magnitudes of the CO2 growth rates differ greatly among the models. For example, the maximum CO2 growth rate is about 3 times larger in GFDL-ESM2M than in CESM1-BGC. Furthermore, some models do not simulate the delayed maximum; for instance, the peaks in both CESM1-BGC and MPI-ESM-LR appear near the mature phase of El Niño. However, CanESM2, MIROC-ESM, and GFDL-ESM2M have magnitudes greater than does the MME and they simulate the timing of the maximum anomalies in April well; these models are the source of the simulated peak during boreal spring, because the MME results are calculated by simply averaging the outputs of all models.

These comparisons indicate that the current ESMs still have large uncertainties in simulating the magnitude and peak timing of the observed carbon cycle. Therefore, two problems must be addressed: 1) the processes that contribute to the relationship between atmospheric CO2 growth rate and ENSO in the observations and the ESMs should be identified and 2) the cause of the diversity among the model predictions of the relationship between ENSO and the carbon cycle should be determined.

ESMs simulate NPP, Rh, and CO2 emission due to fire separately; therefore, the contributions of these parameters to the model physics can be compared quantitatively. CanESM2 and MIROC-ESM do not include any parameterization for carbon fluxes associated with fire. To estimate those processes, which are important for ENSO-related carbon-cycle variability, composites of the carbon flux anomalies associated with NPP, Rh, and carbon emission due to fire are shown (Fig. 2).

Fig. 2.
Fig. 2.

(top)–(bottom) Carbon flux anomalies (PgC yr−1 °C−1) due to NPP, Rh, and fire from the ESMs (colors), MME (thick black), and observations (thick gray) during (a),(c),(e) El Niño and (b),(d),(f) La Niña events.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

The NPP anomalies explain most of the carbon flux anomalies associated with ENSO in the ESMs (Figs. 2a,b). The individual ESMs, as well as the MME, mostly exhibit similar magnitudes and patterns of evolution in NPP as compared to the CO2 growth rate (Figs. 1a,b). This similarity implies that the NPP anomalies contribute most of the variation in ENSO-related CO2 growth rate among the ESMs.

In contrast, the Rh anomaly contributes only ~12% of the variability in the CO2 growth rate in the MME (Figs. 2c,d). Notably, the Rh anomaly response lags ENSO, largely because of the lagged temperature response in the tropics to El Niño (Klein et al. 1999). Even though most ESMs (except MPI-ESM-LR) simulate the same sign of the Rh anomaly, the individual models exhibit a large ensemble spread from the MME. The MME Rh anomalies were −0.11 and 0.08 PgC yr−1 °C−1 during El Niño and La Niña years, respectively; these deviations are smaller than those observed in previous model studies (Jones et al. 2001; Hashimoto et al. 2004; Qian et al. 2008). The diversity of simulated Rh anomalies could be attributed to the lack of understanding and the physical representation of soil carbon dynamics in the ESMs.

Observational data suggest that forest fires might make a major contribution to the CO2 anomalies (Van der Werf et al. 2006). Global fire emissions averaged over 1997–2009 were 2.0 PgC yr−1 (range of 1.6–2.8 PgC yr−1) (Van der Werf et al. 2010); however, all the ESMs simulate very weak responses of emission due to fire compared with the observational estimates. Specifically, the ESMs do not simulate the increase in frequency and intensity of wildfires in equatorial Asia during El Niño events, although this increase is a major source of CO2 emissions during El Niño years (Van der Werf et al. 2010). The parameterization for wildfires should be improved in the next phase of CMIP (Keppel-Aleks et al. 2014).

4. Effects of ENSO-related precipitation and temperature on NPP anomalies

ENSO-related CO2 anomalies can be explained in large part by the NPP anomalies in the ESMs (Fig. 2). To examine how NPP variation is affected by ENSO, the spatial pattern of NPP anomalies associated with ENSO was examined (Fig. 3). The MME composite was applied after NPP anomaly data from each model were interpolated to a 1° × 1° grid (Fig. 3). During El Niño, strong NPP anomalies appear over most tropical land regions, and they are particularly distinctive over Amazonia, Australia, South Asia, and the Maritime Continent, where tropical rain forests exist. The rain forest makes a large contribution to the terrestrial carbon uptake (Clark et al. 2003). Positive anomalies indicate anomalous carbon uptake from the land to the atmosphere. Negative anomalies over extratropical regions, especially over North America, partly compensate for this tropical carbon release.

Fig. 3.
Fig. 3.

MME composite maps of CO2 flux due to NPP anomalies during El Niño events: (a) JJA(0), (b) SON(0), (c) D(0)JF(1), (d) MAM(1), (e) JJA(1), and (f) SON(1). Boxes in (d) indicate regions in Table 3 [Amazonia (AMZ), Australia (AUS), equatorial Asia (EQA), South Asia (SOA), and Africa (AFR)].

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

Although the tropics tend to release carbon all year round, they release it at an increasing rate as El Niño develops from June–August (JJA) to the following winter (DJF). In Amazonia, the NPP anomaly is greatest during the El Niño mature phase in D(0)JF(1). In contrast, the NPP anomalies are strongest during March–May (MAM) over South Asia, the Maritime Continent, and northern Australia. The lagged response of global mean carbon flux might be related to NPP anomalies in these regions.

To assess the efficacy of the ESMs in simulating the precipitation and temperature responses to ENSO, composites of the mean precipitation and mean surface temperature from July(0) to June(1) for both El Niño and La Niña years were constructed (Figs. 4 and 5). During El Niño years, precipitation increased over the central and eastern equatorial Pacific, and weak positive anomalies occurred over North America and high-latitude regions of South America, whereas dry conditions appeared in most land regions (e.g., Amazonia, Australia, South Asia, and the Maritime Continent). These patterns are approximately reversed during La Niña years.

Fig. 4.
Fig. 4.

Composite maps of mean precipitation anomaly for July(0)–June(1) during (a),(b) El Niño and (c),(d) La Niña years from the (left) observations and (right) MME.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

Fig. 5.
Fig. 5.

As in Fig. 4, but for the surface temperature anomaly.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

The major El Niño and La Niña precipitation patterns over the oceans and land are captured well by the CMIP5 models. The pattern correlation between the simulated and observed precipitation anomalies over the tropical domain (30°S–30°N) are 0.66 and 0.57 for El Niño and La Niña composites, respectively. In addition to the tropical precipitation, the ESMs tend to simulate well the anomalous precipitation patterns in the midlatitudes. The opposite carbon fluxes in North America to those in equatorial regions might be related to precipitation anomalies in North America. The Pacific–North American (PNA) teleconnection pattern, which is a typical change in the atmospheric circulation during ENSO phases, is related to precipitation anomalies (Renwick and Wallace 1996). This partial cancelation effect by the midlatitude anomalies for the ENSO-related carbon cycle has also been highlighted in a previous offline model study (Zeng et al. 2005), and its occurrence suggests that ESMs simulate both atmospheric midlatitude teleconnections to ENSO and relevant carbon flux anomalies.

Temperatures increase over Amazonia, eastern Australia, and South Asia during El Niño years, and decrease over those regions during La Niña years (Kiladis and Diaz 1989; Klein et al. 1999). The ESMs capture the sign of these land temperature responses well but tend to overestimate the magnitudes (Fig. 5); this bias suggests that the models overestimate the land surface response. The movement of the center of convection from the western to the central Pacific during El Niño years means that temperatures decrease in equatorial Asia during the development of an El Niño event (Halpert and Ropelewski 1992; Trenberth et al. 2002) but increase in equatorial Asia and Australia during the decay phase (Trenberth et al. 2002). Although these effects cancel each other in the observations when averaging the developing and decaying phases in El Niño events (Fig. 5a), temperature decreases simulated by the ESMs during the developing phase tend to be weaker than observed, especially in equatorial Asia and Australia. Continental-scale ENSO-related temperature anomalies in Amazonia and Africa also have greater amplitudes in the ESMs than in the observations.

To further examine whether the model simulates the precipitation and temperature anomalies over the land regions during ENSO events, the correlation between the D(0)JF(1) Niño-3.4 index and the mean precipitation anomalies over the tropical land between 20°S and 20°N for July(0)–June(1) was computed for the observations and the ESMs (Fig. 6). Each model shows a significant relationship between precipitation and the Niño-3.4 index. In particular, the correlations for CanESM2, CESM1-BGC, and GFDL-ESM2M are as high as for the observations. MIROC-ESM and MRI-ESM1 simulate relatively weak ENSO variability, which leads to a relatively weak relation between the land precipitation and the Niño-3.4 index.

Fig. 6.
Fig. 6.

Scatterplots of D(0)JF(1) Niño-3.4 index vs mean precipitation anomaly over tropical land (20°S–20°N) for July(0)–June(1): (a) observations, (b) CanESM2, (c) CESM1-BGC, (d) GFDL-ESM2M, (e) MIROC-ESM, (f) MPI-ESM-LR, and (g) MRI-ESM1. The correlation and regression coefficient are shown in the top right of each panel. Red (blue) dots show El Niño (La Niña) years, classified using Niño-3.4 anomaly magnitudes >1°C (signs positive and negative, respectively). Black dots indicate normal years.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

Clearly, precipitation and temperature contribute to the carbon flux anomalies in the limited observations, but quantitative estimation of these contributions is difficult. For the ESMs, the effects of precipitation and temperature on the carbon flux can be estimated to some extent, because various variables related to the carbon cycle are provided and relatively long-term outputs are available. Because the ESMs tend to simulate the observed relation between ENSO and atmospheric CO2 anomalies adequately, at least in terms of phase, the model outputs provide an appropriate opportunity to explore the role of precipitation anomalies on the carbon flux. First, the linear relationships based on the linear regressions of NPP anomalies with respect to the local precipitation and temperature anomalies were analyzed.

Precipitation and surface temperature anomalies over tropical land have strong linear relationships with the equatorial total NPP anomalies (Figs. 8 and 9). An increase in precipitation leads to an increase in photosynthesis, which can be linked directly to positive NPP anomalies (Nemani et al. 2003; Piao et al. 2009). Four ESMs have correlation coefficients >0.75 (Fig. 8), but CESM1-BGC (r = 0.57) and MRI-ESM1 (r = 0.54) have relatively low correlations, possibly as a result of the weak NPP variability. The change in NPP anomaly per precipitation anomaly is >9 PgC mm−1 for the MME, with 4.42 PgC mm−1 for CESM1-BGC and 5.73 PgC mm−1 for MRI-ESM1. Because land precipitation anomalies are controlled significantly by El Niño and La Niña events (Fig. 6), this linear regression suggests that a considerable proportion of NPP changes might be impacted by ENSO-related precipitation changes. However, NPP anomalies are also strongly correlated with temperature anomalies (Fig. 9); this relationship is consistent with previous findings that NPP decreases as temperature increases (Clark et al. 2013). All models except CESM1-BGC (r = −0.17) show strong negative relationships between temperature anomalies and NPP anomalies. The average change in NPP anomaly per temperature anomaly of the ESMs was −3.27 PgC K−1 yr−1, but individually, it was −0.61 PgC K−1 yr−1 for CESM1-BGC. Because ENSO-related temperature responses in the ESMs are more than twice as large as the observed responses (Fig. 7), this high sensitivity is linked directly to the overestimation of temperature-induced NPP anomalies. CESM1-BGC output is closest to the observations (Fig. 1), because it shows an exceptionally weak response of NPP to temperature change. Therefore, the ESMs’ overestimation of temperature responses to ENSO should be considered when the models are used to examine interannual variations in the carbon cycle. Both increased temperature and reduced precipitation during El Niño provide favorable conditions for decreased NPP. However, temperature and precipitation also have a strong linear relationship with each other; therefore, the simple linear regression (Figs. 8 and 9) cannot be used directly to quantify the contributions of the temperature and precipitation anomalies. The relative influences of the temperature and precipitation anomalies are discussed further in section 6.

Fig. 7.
Fig. 7.

As in Fig. 6, but for mean surface temperature anomaly.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

Fig. 8.
Fig. 8.

Scatterplots of mean precipitation anomaly over the tropical land (20°S–20°N) for July(0)–June(1) vs NPP anomalies in the tropical regions (20°S–20°N): (a) CanESM2, (b) CESM1-BGC, (c) GFDL-ESM2M, (d) MIROC-ESM, (e) MPI-ESM-LR, (f) MPI-ESM-LR, and (g) MRI-ESN1. The correlation and regression coefficient are shown in the top right of each panel. Red (blue) dots show El Niño (La Niña) years, classified using Niño-3.4 anomaly magnitudes >1°C (signs positive and negative, respectively). Black dots indicate normal years.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

Fig. 9.
Fig. 9.

As in Fig. 8, but for surface temperature anomaly.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

5. Regional contributions and lagged mechanism

The models suggest that CO2 uptake by vegetation decreases during El Niño because of the increased temperature and decreased precipitation. ENSO events tend to peak during winter, and maximum carbon flux occurs during the following spring (Figs. 1 and 2); thus, the change in atmospheric CO2 concentration lags the ENSO peak by 4–5 months. Other studies have noted the same lagged response (e.g., Jones and Cox 2005; Raupach et al. 2008; Wang et al. 2013). Zeng et al. (2005) suggested that the lag in the observed CO2 growth rate is due to a delayed response of soil moisture to precipitation. Because soil moisture rather than precipitation amount is associated more directly with the growth of vegetation, the effects of vegetation processes on the change in the CO2 growth rate can lag ENSO by about 6 months (Qian et al. 2008).

To explore further the delayed response of the CO2 flux, the seasonal evolutions of regional NPP anomalies associated with ENSO were investigated (Table 3; Fig. 3d). The contribution of NPP anomalies varies among regions (Figs. 10a,b). The NPP anomalies over Amazonia and Australia constitute a large proportion of the global NPP anomaly; that is, these regions are the primary contributors to these anomalies. The Amazonian contribution explains about 32% of the global NPP anomalies from September–November [SON(0)] to JJA(1), and the Australian contribution explains 37%, after averaging the contributions during El Niño and La Niña. This result is consistent with Poulter et al. (2014), who argued that precipitation anomalies in semiarid regions lead the recent interannual CO2 growth rate. Interestingly, despite the large contributions of Amazonia and Australia to the global NPP anomaly, they do not increase significantly from D(0)JF(1) to MAM(1), and even decrease in Amazonia, whereas the global NPP anomaly increases during this period. In fact, the NPP anomaly over South Asia is relatively small during boreal winter, but is at a maximum during the following boreal spring. The absolute differences in NPP anomalies over South Asia between D(0)JF(1) and MAM(1) were 0.24 and 0.14 PgC yr−1 °C−1 for El Niño and La Niña events, respectively. These differences are comparable with those of the global anomaly. These findings are also supported by the contrasting model result that South Asia leads the peak timing difference between the CO2 growth rate and ENSO (Gurney et al. 2012). In addition to the NPP anomaly in South Asia, the NPP anomaly over Africa contributes partly to the delayed response of the global NPP anomaly, although the intermodel diversity of the African NPP is relatively large (Guan et al. 2013).

Table 3.

Range information for each region in Figs. 1012.

Table 3.
Fig. 10.
Fig. 10.

Composites of seasonal NPP anomalies averaged over the globe and regions denoted in Fig. 3d from MME (bars) and each ESM (dots) during (a) El Niño and (b) La Niña events. Star-like symbols indicate the season containing maximum NPP anomalies in each region.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

Because the NPP is closely connected to ENSO-related precipitation and temperature, the seasonal evolution of the NPP anomaly might be explained by the seasonal evolution of the precipitation and temperature anomalies. NPP responses to precipitation and temperature variation show strong seasonality over South Asia (Fig. 11). The NPP change in response to given temperature and precipitation changes over South Asia is relatively weak during boreal winter, which is contemporaneous with the dry season associated with the monsoonal pattern. Even though South Asia is in the tropics, precipitation there has strong seasonality and the dry and wet seasons can be determined by changes in the wind direction from onshore to offshore (Chang et al. 2005). Local tower flux measurements in South Asia show abrupt increases in GPP in the drought-affected deciduous and dry evergreen forests at the time of the switch between the dry and rainy seasons (Huete et al. 2008). These observations suggest that the effects of precipitation and temperature on carbon uptake could operate more effectively during the growing season than during other seasons (Wang et al. 2003).

Fig. 11.
Fig. 11.

Regression coefficients of NPP anomalies on precipitation (blue) and temperature (red) in South Asia. Error bars indicate 95% significance levels based on the Student’s t test.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

The contribution of ENSO-related NPP change from equatorial Asia does not show significant seasonality from D(0)JF(1) to JJA(1), although it is distinctively larger during SON(0) than at other times (Figs. 10a,b). This result is consistent with a recent satellite-based study that showed the seasonality of vegetation activity depends on the precipitation regime (Guan et al. 2015). Because rainfall in equatorial Asia is sufficient for vegetation activity, even during El Niño years, the effect of precipitation anomalies related to El Niño is not great. Thus, the variability of NPP is more sensitive to precipitation anomalies in South Asia than in equatorial Asia (Fig. 10). However, across the Maritime Continent the seasonal cycle of monsoonal precipitation is the opposite of that South Asia (Chang et al. 2005). Therefore, to a considerable extent, NPP anomalies over equatorial Asia constitute global NPP anomalies only for SON(0), which is the driest season in this area.

In addition to the sensitivity of NPP to precipitation and temperature, ENSO-related anomalies also make considerable contributions to the NPP anomalies during MAM(1). Tsai et al. (2015) revealed the significant influence of ENSO on boreal spring precipitation anomalies over South Asia. Furthermore, many previous studies have shown that ENSO-related temperature responses over most tropical land lag a few months behind the ENSO event (e.g., Trenberth et al. 2002; Malhi and Wright 2004; Gu and Adler 2011). Hence, the seasonality of both NPP sensitivity and ENSO-related anomalies contributes to the amplification of NPP changes over South Asia that is associated with ENSO during boreal spring. In addition, excess precipitation during the previous season may also amplify NPP anomalies as a result of soil moisture memory (Zeng et al. 2005; Qian et al. 2008). However, lagged temperature responses also occur over Amazonia and Australia, which contribute to most of the ENSO-related CO2 growth rate, but they are not amplified during MAM(1) because of these out-of-phase relationships between the seasonality and the ENSO-related anomalies within these regions.

Previous studies have focused on CO2 flux distributions that correspond to latitude, based on zonal mean values in the tropics, midlatitudes, and high latitudes (e.g., Nemani et al. 2003; Hashimoto et al. 2004; Cao et al. 2005; Zeng et al. 2005; Qian et al. 2008). However, to explain why the CO2 growth rate lags behind the ENSO mature state, each tropical continent’s contribution should be examined. The results of the current study suggest that the seasonality of the NPP anomalies in South Asia is partly responsible for this delay; that is, the seasonality of the local terrestrial variability in the tropics has a critical influence on the response of atmospheric CO2 concentration to ENSO.

6. Contribution of precipitation and temperature to NPP anomalies

NPP variation associated with ENSO is closely related to ENSO-induced temperature and precipitation anomalies over tropical land areas (Figs. 8 and 9). However, the temperature and precipitation anomalies over the land area are highly correlated and their effects on the NPP anomalies cannot be easily distinguished from the simple linear regression. Therefore, a multiple regression approach was used in this study to investigate the individual contributions of the temperature and precipitation anomalies. The effects of these anomalies depend strongly on the regional climatological conditions and, therefore, multiple regressions were applied separately to each season and region. Using this approach, the NPP anomalies were separated into the contributions from the temperature and precipitation anomalies for the D(0)JF(1) and MAM(1) seasons, when ENSO-related NPP anomalies are large. In Amazonia and Australia, the multiple regression coefficients of the NPP anomalies related to temperature were strongly negative (Fig. 12a); that is, the NPP over these regions tends to decrease during positive temperature anomalies. Conversely, in these regions, the multiple regression coefficients for precipitation were positive, but with smaller amplitude than for temperature; that is, in the models, temperature effects are stronger than precipitation effects. In contrast, in South Asia, the signs of the multiple regression coefficients for the temperature and precipitation anomalies were the same as in Amazonia, but the precipitation had much larger amplitude than did temperature, particularly during MAM(1). This difference implies that in South Asia, precipitation anomalies affect NPP anomalies more than do temperature anomalies. These differing results suggest that the relative importance of the temperature and precipitation anomalies varies regionally.

Fig. 12.
Fig. 12.

(a) Multiple regression coefficients of NPP anomalies on precipitation and temperature in Amazonia, Australia, South Asia, and Africa. (b) Regression coefficients of precipitation and temperature to the Niño-3.4 index in each region. Bars, colored dots, and circles denote MME, each model, and observations, respectively.

Citation: Journal of Climate 29, 24; 10.1175/JCLI-D-14-00672.1

However, the multiple regression coefficients differed greatly among the models. In many cases, the spread was greater than the MME signal, and some models even produced opposite relationships to those of the MME. These differences emphasize that current land models have large uncertainties regarding NPP sensitivity to precipitation and temperature variations (Piao et al. 2013).

To consider the influence of ENSO on regional NPP anomalies, the regional temperature and precipitation anomalies related to ENSO forcing should be considered, in addition to the regional NPP sensitivities to the temperature and precipitation anomalies. The regression coefficients of the precipitation and temperature anomalies with respect to Niño-3.4 SST (Fig. 12b) differ among the models, but their spread is relatively small in comparison with the NPP sensitivities to temperature and precipitation (Fig. 12a). This difference suggests that the large spread in the ENSO-related NPP anomalies (Fig. 2) is largely due to the uncertainty in the NPP sensitivities to climatic factors. For example, GFDL-ESM2M has the highest NPP sensitivity to both temperature and precipitation among the ESMs used in this study; this model’s sensitivity is more than twice as large as the MME results (Fig. 12a). This model also displays the largest response both to NPP anomalies (Figs. 2a,b) associated with ENSO and to CO2 growth rate (Fig. 1). Thus, diverse NPP sensitivities to temperature and precipitation in the ESMs have a strong influence on the large spread of their ENSO-related CO2 fluxes.

Although the spread in both the temperature and the precipitation anomalies related to ENSO among the ESMs is small (Fig. 12b), the temperature responses to ENSO show some systematic biases. Most of the models overestimate the temperature responses in most tropical land regions. Specifically, most models simulate large-amplitude temperature anomalies in Australia, South Asia, and central Amazonia, whereas the observations show negligible anomalies. This overestimation of the temperature anomalies could lead to excessive estimates of NPP changes associated with ENSO. Irrespective of the large variations among ESMs in the sensitivity of the NPP estimate to temperature, the NPP variations are obviously overestimated because of the excessive temperature response to ENSO in the simulations. This overestimation might explain why the model predictions of CO2 growth rate associated with ENSO are greater than the observational estimates (Fig. 1).

The relative importance of precipitation and temperature on NPP varies among the models. An offline model with observed ENSO-related anomalies showed that precipitation affects NPP more than does temperature (Qian et al. 2008). However, in the ESM results, the temperature tends to affect NPP more than does precipitation. Unlike the ESMs, the offline model results do not overestimate the CO2 growth rate. Therefore, the difference in the relative importance of precipitation and temperature may occur because the ESMs overestimate ENSO-related temperature anomalies. Despite the high sensitivity of NPP to both precipitation and temperature in the current land models (Piao et al. 2013), successful simulation of the CO2 growth rate by the offline model implies that the systematic bias of ENSO-related temperature has a greater influence than precipitation on the overestimation of the CO2 growth rate.

7. Summary and discussion

In this study, the CMIP5 ESM control runs were analyzed to examine the interannual variability of the CO2 growth rate associated with ENSO. Similar to the results of previous studies based on observations (e.g., Hashimoto et al. 2004), the ESMs simulate a significant anomalous CO2 flux into the atmosphere from the land during ENSO events. Specifically, our results indicate that NPP anomalies could explain most of the carbon flux variation in the ESMs, and that anomalies in Rh and in the frequency and intensity of natural fires have relatively minor effects. Both ENSO-related precipitation and temperature have a negative regulatory effect on NPP variation within the tropics.

However, it is found that the ESM models overestimate temperature-induced NPP anomalies, because models overestimate the temperature anomalies associated with ENSO. Consequently, the models simulate a larger change in the CO2 growth rate than is observed. Results also demonstrate that the lagged response of the NPP anomalies is due mostly to an anomaly in uptake over South Asia, and that this anomaly may be linked to the coupling of NPP with seasonal cycles of vegetation growth (Krinner et al. 2005).

Although the ESM simulations indicate that NPP is a primary cause of the ENSO-related CO2 flux, the contribution of each physical process in nature remains debatable. Some previous studies have argued that temperature variability, which controls NPP and Rh anomalies, is the principal contributor to the global carbon cycle (e.g., Braswell et al. 1997; Wang et al. 2013). In particular, because the CO2 growth rate is more strongly correlated with land temperature than with precipitation, the CO2 growth rate may be coupled more strongly to temperature than to precipitation (Wang et al. 2013).

In this study, a multiple regression approach was used to quantify the relative contributions of the precipitation and temperature anomalies. The temperature-induced NPP anomalies were much larger than the precipitation-induced NPP anomalies in some regions, but only because the ESMs significantly overestimated ENSO-related temperature changes. Thus, we suggest that additional experiments with prescribed observational atmospheric data, as used in the offline Dynamic Global Vegetation Model project, would be helpful to quantify the sensitivity of ESMs’ carbon flux estimates to ENSO-related anomalies (Sitch et al. 2008; Piao et al. 2013).

Despite the overestimation of ENSO-related temperature anomalies in the ESMs, they simulate a weaker CO2 flux linked to Rh anomalies during El Niño events than has been found in previous offline model studies with observed ENSO-related temperature anomalies (e.g., Zeng et al. 2005; Qian et al. 2008). This change implies that the underestimation of the ENSO-related Rh anomalies is due to the sensitivity of Rh to a given temperature change in the parameterizations of the current ESMs, rather than as a result of local climate responses to ENSO forcing. The causes of this underestimation of the model parameterization should be further validated and investigated.

The ESMs tend to underestimate the effects of natural fires on the carbon cycle; two models do not even consider it. This underestimation contributes to the uncertainty in the simulated CO2 growth rate associated with ENSO. Although climatological biases and the effects of fire are acceptable for the ESM MME with respect to the observational results from the previous studies, the interannual variation of wildfires in the models is weak in equatorial Asia (Prentice et al. 2011; Li et al. 2013; Lasslop et al. 2014). Actually, the carbon emission by wildfires in equatorial Asia is 1.07 PgC yr−1 (i.e., 40% of the global fire emission estimate in 1997/98, which featured a significant El Niño event) (Van der Werf et al. 2010). However, the CMIP5 ESMs do not capture significant anomalies in this region during ENSO years. The CMIP5 simulations do not include the effects of biomass burning by anthropogenic activities and, therefore, may underestimate the total carbon emissions by fire. Furthermore, the fire parameterizations do not adequately simulate fire-driven deforestation in equatorial Asia (Prentice et al. 2011). Natural fire often occurs within this region, even though it is the wettest area in the world (Peel et al. 2007). The simulation of wildfires in equatorial Asia in the fire model could be quite uncertain because of the unrealistic parameters based on relative and specific humidity. Improving this component could improve the fidelity of ESM simulations of the global carbon cycle (Keppel-Aleks et al. 2014; Poulter et al. 2015).

Analyses presented here have identified several limitations in the current ESMs on an interannual time scale. By adopting the emergent constraint, such knowledge could be used to interpret and calibrate future climate projections by the ESMs. Cox et al. (2013) suggested using C4MIP to apply the emergent constraint approach to constrain the uncertainty of the tropical land climate projections that results from the sensitivity of observed carbon loss to tropical temperature variability. Furthermore, several studies have used the emergent constraint approach in CMIP5 ESMs when analyzing the sensitivity of terrestrial carbon uptake to tropical mean land temperature (Wang et al. 2014; Wenzel et al. 2014). However, here we showed that the NPP sensitivities to precipitation and temperature in each region and season are quite distinct on interannual time scales. This observation implies that the models could have regionally different systematic biases in the projected carbon flux changes and, therefore, that the emergent constraint approach could be applied based on regional relationships between the model and observations, rather than on the globally averaged relationship.

Furthermore, previous studies have focused only on the relationship between the carbon cycle and temperature in the emergent constraint on the carbon cycle (Cox et al. 2013; Friedlingstein et al. 2014). However, we suggested that precipitation anomalies are also important on interannual time scales. Because temperature anomalies have a strong linear relationship with precipitation on interannual time scales, the sensitivity of carbon loss to temperature might include the effects on variations in the precipitation-driven carbon flux. On long time scales, such as those involved in climatic change, the relationship between temperature and precipitation change could differ from the interannual relationship, and this difference could cause severe errors in the approach that uses emergent constraints. Therefore, the constraint should be chosen carefully where the precipitation anomaly, in addition to the temperature anomaly, has considerable influence on the NPP anomalies.

We also suggest that a variety of climate–carbon feedbacks occur, and that they may depend on tropical climate responses to ENSO-like future warming related to anthropogenic effects (Jones et al. 2001). If the interannual relationships between ENSO and tropical land temperature and between ENSO and precipitation could be applied to the long-term climatic response, the increased air temperature and reduced precipitation over tropical land by El Niño–like responses might exacerbate the reduction in terrestrial carbon uptake. Consequently, greenhouse warming could be accelerated further because of an increase in atmospheric CO2 concentration. In contrast, a La Niña–like response might decelerate future warming because of a reduction in atmospheric CO2 concentration (Yeh et al. 2012). However, Zeng et al. (2005) argued that the governing mechanisms and spatial patterns of long-term and interannual climatic change might differ. Therefore, further careful examination must be conducted to improve the understanding of long-term climate–carbon feedback loops.

Acknowledgments

This study was supported by the Korea Meteorological Administration Research and Development Program under Grant KMIPA 2015-2092 and the National Research Foundation (NRF-2014R1A2A2A01003827).

REFERENCES

  • Adler, R. F., and et al. , 2003: The Version-2 Global Precipitation Climatology Project (GPCP) Monthly Precipitation Analysis (1979–present). J. Hydrometeor., 4, 11471167, doi:10.1175/1525-7541(2003)004<1147:TVGPCP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Allen, M. R., , and W. J. Ingram, 2002: Constraints on future changes in climate and the hydrologic cycle. Nature, 419, 224232, doi:10.1038/nature01092.

    • Search Google Scholar
    • Export Citation
  • Anav, A., and et al. , 2013: Evaluating the land and ocean components of the global carbon cycle in the CMIP5 Earth system models. J. Climate, 26, 68016843, doi:10.1175/JCLI-D-12-00417.1.

    • Search Google Scholar
    • Export Citation
  • Arora, V. K., and et al. , 2011: Carbon emission limits required to satisfy future representative concentration pathways of greenhouse gases. Geophys. Res. Lett., 38, L05805, doi:10.1029/2010GL046270.

    • Search Google Scholar
    • Export Citation
  • Arora, V. K., and et al. , 2013: Carbon–concentration and carbon– climate feedbacks in CMIP5 Earth system models. J. Climate, 26, 52895314, doi:10.1175/JCLI-D-12-00494.1.

    • Search Google Scholar
    • Export Citation
  • Bacastow, R., 1976: Modulation of atmospheric carbon dioxide by the Southern Oscillation. Nature, 261, 116118, doi:10.1038/261116a0.

  • Behrenfeld, M. J., and et al. , 2001: Biospheric primary production during an ENSO transition. Science, 291, 25942597, doi:10.1126/science.1055071.

    • Search Google Scholar
    • Export Citation
  • Bellenger, H., , E. Guilyardi, , J. Leloup, , M. Lengaigne, , and J. Vialard, 2013: ENSO representation in climate models: From CMIP3 to CMIP5. Climate Dyn., 42, 19992018, doi:10.1007/s00382-013-1783-z.

    • Search Google Scholar
    • Export Citation
  • Booth, B. B. B., and et al. , 2012: High sensitivity of future global warming to land carbon cycle processes. Environ. Res. Lett., 7, 024002, doi:10.1088/1748-9326/7/2/024002.

    • Search Google Scholar
    • Export Citation
  • Bousquet, P., , P. Peylin, , P. Ciais, , C. Le Quere, , P. Friedlingstein, , and P. P. Tans, 2000: Regional changes in carbon dioxide fluxes of land and oceans since 1980. Science, 290, 13421346, doi:10.1126/science.290.5495.1342.

    • Search Google Scholar
    • Export Citation
  • Braswell, B. H., , D. S. Schimel, , E. Linder, , and B. Moore, 1997: The response of global terrestrial ecosystems to interannual temperature variability. Science, 278, 870872, doi:10.1126/science.278.5339.870.

    • Search Google Scholar
    • Export Citation
  • Brovkin, V., , T. Raddatz, , C. H. Reick, , M. Claussen, , and V. Gayler, 2009: Global biogeophysical interactions between forest and climate. Geophys. Res. Lett., 36, L07405, doi:10.1029/2009GL037543.

    • Search Google Scholar
    • Export Citation
  • Cao, M. K., , S. D. Prince, , B. Tao, , J. Small, , and K. R. Li, 2005: Regional pattern and interannual variations in global terrestrial carbon uptake in response to changes in climate and atmospheric CO2. Tellus, 57B, 210217, doi:10.1111/j.1600-0889.2005.00146.x.

    • Search Google Scholar
    • Export Citation
  • Chang, C.-P., , Z. Wang, , J. McBride, , and C.-H. Liu, 2005: Annual cycle of Southeast Asia—Maritime Continent rainfall and the asymmetric monsoon transition. J. Climate, 18, 287301, doi:10.1175/JCLI-3257.1.

    • Search Google Scholar
    • Export Citation
  • Clark, D. A., , S. C. Piper, , C. D. Keeling, , and D. B. Clark, 2003: Tropical rain forest tree growth and atmospheric carbon dynamics linked to interannual temperature variation during 1984–2000. Proc. Natl. Acad. Sci. USA, 100, 58525857, doi:10.1073/pnas.0935903100.

    • Search Google Scholar
    • Export Citation
  • Clark, D. A., , D. B. Clark, , and S. F. Oberbauer, 2013: Field-quantified responses of tropical rainforest aboveground productivity to increasing CO2 and climatic stress, 1997–2009. J. Geophys. Res. Biogeosci., 118, 783794, doi:10.1002/jgrg.20067.

    • Search Google Scholar
    • Export Citation
  • Collins, M., , R. E. Chandler, , P. M. Cox, , J. M. Huthnance, , J. Rougier, , and D. B. Stephenson, 2012: Quantifying future climate change. Nat. Climate Change, 2, 403409, doi:10.1038/nclimate1414.

    • Search Google Scholar
    • Export Citation
  • Cox, P. M., , D. Pearson, , B. B. Booth, , P. Friedlingstein, , C. Huntingford, , C. D. Jones, , and C. M. Luke, 2013: Sensitivity of tropical carbon to climate change constrained by carbon dioxide variability. Nature, 494, 341344, doi:10.1038/nature11882.

    • Search Google Scholar
    • Export Citation
  • Doughty, C. E., , and M. L. Goulden, 2008: Are tropical forests near a high temperature threshold? J. Geophys. Res. Biogeosci., 113, G00B07, doi:10.1029/2007JG000632.

    • Search Google Scholar
    • Export Citation
  • Dunne, J. P., and et al. , 2012: GFDL’s ESM2 global coupled climate–carbon Earth system models. Part I: Physical formulation and baseline simulation characteristics. J. Climate, 25, 66466665, doi:10.1175/JCLI-D-11-00560.1.

    • Search Google Scholar
    • Export Citation
  • Dunne, J. P., and et al. , 2013: GFDL’s ESM2 global coupled climate–carbon Earth system models. Part II: Carbon system formulation and baseline simulation characteristics. J. Climate, 26, 22472267, doi:10.1175/JCLI-D-12-00150.1.

    • Search Google Scholar
    • Export Citation
  • Fang, J. Y., , S. L. Piao, , Z. Y. Tang, , C. H. Peng, , and J. Wei, 2001: Interannual variability in net primary production and precipitation. Science, 293, 1723, doi:10.1126/science.293.5536.1723a.

    • Search Google Scholar
    • Export Citation
  • Feely, R. A., and et al. , 2002: Seasonal and interannual variability of CO2 in the equatorial Pacific. Deep-Sea Res. II, 49, 24432469, doi:10.1016/S0967-0645(02)00044-9.

    • Search Google Scholar
    • Export Citation
  • Friedlingstein, P., and et al. , 2006: Climate–carbon cycle feedback analysis: Results from the C4MIP model intercomparison. J. Climate, 19, 33373353, doi:10.1175/JCLI3800.1.

    • Search Google Scholar
    • Export Citation
  • Friedlingstein, P., , M. Meinshausen, , V. K. Arora, , C. D. Jones, , A. Anav, , S. K. Liddicoat, , and R. Knutti, 2014: Uncertainties in CMIP5 climate projections due to carbon cycle feedbacks. J. Climate, 27, 511526, doi:10.1175/JCLI-D-12-00579.1.

    • Search Google Scholar
    • Export Citation
  • Gu, G., , and R. F. Adler, 2011: Precipitation and temperature variations on the interannual time scale: Assessing the impact of ENSO and volcanic eruptions. J. Climate, 24, 22582270, doi:10.1175/2010JCLI3727.1.

    • Search Google Scholar
    • Export Citation
  • Guan, K. Y., , A. Wolf, , D. Medvigy, , K. K. Caylor, , M. Pan, , and E. F. Wood, 2013: Seasonal coupling of canopy structure and function in African tropical forests and its environmental controls. Ecosphere, 4, 35, doi:10.1890/ES12-00232.1.

    • Search Google Scholar
    • Export Citation
  • Guan, K. Y., and et al. , 2015: Photosynthetic seasonality of global tropical forests constrained by hydroclimate. Nat. Geosci., 8, 284289, doi:10.1038/ngeo2382.

    • Search Google Scholar
    • Export Citation
  • Guilyardi, E., , H. Bellenger, , M. Collins, , S. Ferrett, , W. Cai, , and A. Wittenberg, 2012: A first look at ENSO in CMIP5. CLIVAR Exchanges, No. 17, International CLIVAR Project Office, Southampton, United Kingdom, 2932.

  • Gurney, K. R., , D. Baker, , P. Rayner, , and S. Denning, 2008: Interannual variations in continental-scale net carbon exchange and sensitivity to observing networks estimated from atmospheric CO2 inversions for the period 1980 to 2005. Global Biogeochem. Cycles, 22, GB3025, doi:10.1029/2007GB003082.

    • Search Google Scholar
    • Export Citation
  • Gurney, K. R., , K. Castillo, , B. Li, , and X. Zhang, 2012: A positive carbon feedback to ENSO and volcanic aerosols in the tropical terrestrial biosphere. Global Biogeochem. Cycles, 26, GB1029, doi:10.1029/2011GB004129.

    • Search Google Scholar
    • Export Citation
  • Halpert, M. S., , and C. F. Ropelewski, 1992: Surface temperature patterns associated with the Southern Oscillation. J. Climate, 5, 577593, doi:10.1175/1520-0442(1992)005<0577:STPAWT>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Hashimoto, H., and et al. , 2004: El Niño–Southern Oscillation-induced variability in terrestrial carbon cycling. J. Geophys. Res., 109, D23110, doi:10.1029/2004JD004959.

    • Search Google Scholar
    • Export Citation
  • Hoffman, F. M., and et al. , 2014: Causes and implications of persistent atmospheric carbon dioxide biases in Earth system models. J. Geophys. Res. Biogeosci., 119, 141162, doi:10.1002/2013JG002381.

    • Search Google Scholar
    • Export Citation
  • Huete, A. R., , N. Restrepo-Coupe, , P. Ratana, , K. Didan, , S. R. Saleska, , K. Ichii, , S. Panuthai, , and M. Gamo, 2008: Multiple site tower flux and remote sensing comparisons of tropical forest dynamics in monsoon Asia. Agric. For. Meteor., 148, 748760, doi:10.1016/j.agrformet.2008.01.012.

    • Search Google Scholar
    • Export Citation
  • Hurrell, J. W., and et al. , 2013: The Community Earth System Model: A framework for collaborative research. Bull. Amer. Meteor. Soc., 94, 13391360, doi:10.1175/BAMS-D-12-00121.1.

    • Search Google Scholar
    • Export Citation
  • Jones, C. D., , and P. M. Cox, 2005: On the significance of atmospheric CO2 growth rate anomalies in 2002–2003. Geophys. Res. Lett., 32, L14816, doi:10.1029/2005GL023027.

    • Search Google Scholar
    • Export Citation
  • Jones, C. D., , M. Collins, , P. M. Cox, , and S. A. Spall, 2001: The carbon cycle response to ENSO: A coupled climate–carbon cycle model study. J. Climate, 14, 41134129, doi:10.1175/1520-0442(2001)014<4113:TCCRTE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Jungclaus, J. H., and et al. , 2006: Ocean circulation and tropical variability in the coupled model ECHAM5/MPI-OM. J. Climate, 19, 39523972, doi:10.1175/JCLI3827.1.

    • Search Google Scholar
    • Export Citation
  • Kalnay, E., and et al. , 1996: The NCEP/NCAR Reanalysis Project. Bull. Amer. Meteor. Soc., 77, 437471, doi:10.1175/1520-0477(1996)077<0437:TNYRP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Keeling, C. D., , and R. Revelle, 1985: Effects of El Niño/Southern Oscillation on the atmospheric content of carbon dioxide. Meteoritics, 20, 437450.

    • Search Google Scholar
    • Export Citation
  • Keeling, C. D., , T. P. Whorf, , M. Wahlen, , and J. Vanderplicht, 1995: Interannual extremes in the rate of rise of atmospheric carbon dioxide since 1980. Nature, 375, 666670, doi:10.1038/375666a0.

    • Search Google Scholar
    • Export Citation
  • Keppel-Aleks, G., and et al. , 2014: Separating the influence of temperature, drought, and fire on interannual variability in atmospheric CO2. Global Biogeochem. Cycles, 28, 12951310, doi:10.1002/2014GB004890.

    • Search Google Scholar
    • Export Citation
  • Kiladis, G. N., , and H. F. Diaz, 1989: Global climatic anomalies associated with extremes in the Southern Oscillation. J. Climate, 2, 10691090, doi:10.1175/1520-0442(1989)002<1069:GCAAWE>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., , B. J. Soden, , and N. C. Lau, 1999: Remote sea surface temperature variations during ENSO: Evidence for a tropical atmospheric bridge. J. Climate, 12, 917932, doi:10.1175/1520-0442(1999)012<0917:RSSTVD>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Krinner, G., and et al. , 2005: A dynamic global vegetation model for studies of the coupled atmosphere-biosphere system. Global Biogeochem. Cycles, 19, GB1015, doi:10.1029/2003GB002199.

    • Search Google Scholar
    • Export Citation
  • Kug, J.-S., , Y.-G. Ham, , J.-Y. Lee, , and F.-F. Jin, 2012: Improved simulation of two types of El Niño in CMIP5 models. Environ. Res. Lett., 7, 034002, doi:10.1088/1748-9326/7/3/034002.

    • Search Google Scholar
    • Export Citation
  • Lasslop, G., , K. Thonicke, , and S. Kloster, 2014: SPITFIRE within the MPI Earth system model: Model development and evaluation. J. Adv. Model. Earth Syst., 6, 740755, doi:10.1002/2013MS000284.

    • Search Google Scholar
    • Export Citation
  • Lee, K., , R. Wanninkhof, , T. Takahashi, , S. C. Doney, , and R. A. Feely, 1998: Low interannual variability in recent oceanic uptake of atmospheric carbon dioxide. Nature, 396, 155159, doi:10.1038/24139.

    • Search Google Scholar
    • Export Citation
  • Le Quéré, C., and et al. , 2003: Two decades of ocean CO2 sink and variability. Tellus, 55B, 649656, doi:10.1034/j.1600-0889.2003.00043.x.

    • Search Google Scholar
    • Export Citation
  • Le Quéré, C., and et al. , 2009: Trends in the sources and sinks of carbon dioxide. Nat. Geosci., 2, 831836, doi:10.1038/ngeo689.

    • Search Google Scholar
    • Export Citation
  • Le Quéré, C., and et al. , 2015: Global carbon budget 2014. Earth Syst. Sci. Data, 7, 4785, doi:10.5194/essd-7-47-2015.

  • Li, F., , S. Levis, , and D. S. Ward, 2013: Quantifying the role of fire in the Earth system – Part 1: Improved global fire modeling in the Community Earth System Model (CESM1). Biogeosciences, 10, 22932314, doi:10.5194/bg-10-2293-2013.

    • Search Google Scholar
    • Export Citation
  • Lloyd, J., , and J. A. Taylor, 1994: On the temperature dependence of soil respiration. Funct. Ecol., 8, 315323, doi:10.2307/2389824.

  • Malhi, Y., , and J. Wright, 2004: Spatial patterns and recent trends in the climate of tropical rainforest regions. Philos. Trans. Roy. Soc., 359B, 311329, doi:10.1098/rstb.2003.1433.

    • Search Google Scholar
    • Export Citation
  • Nagai, S., , K. Ichii, , and H. Morimoto, 2007: Interannual variations in vegetation activities and climate variability caused by ENSO in tropical rainforests. Int. J. Remote Sens., 28, 12851297, doi:10.1080/01431160600904972.

    • Search Google Scholar
    • Export Citation
  • 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, 15601563, doi:10.1126/science.1082750.

    • Search Google Scholar
    • Export Citation
  • Obata, A., , and K. Shibata, 2012: Damage of land biosphere due to intense warming by 1000-fold rapid increase in atmospheric methane: Estimation with a climate-carbon cycle model. J. Climate, 25, 85248541, doi:10.1175/JCLI-D-11-00533.1.

    • Search Google Scholar
    • Export Citation
  • Peel, M. C., , B. L. Finlayson, , and T. A. McMahon, 2007: Updated world map of the Köppen–Geiger climate classification. Hydrol. Earth Syst. Sci., 11, 16331644, doi:10.5194/hess-11-1633-2007.

    • Search Google Scholar
    • Export Citation
  • Peylin, P., and et al. , 2005: Multiple constraints on regional CO2 flux variations over land and oceans. Global Biogeochem. Cycles, 19, GB1011, doi:10.1029/2003GB002214.

    • Search Google Scholar
    • Export Citation
  • Piao, S. L., , P. Ciais, , P. Friedlingstein, , N. de Noblet-Ducoudre, , P. Cadule, , N. Viovy, , and T. Wang, 2009: Spatiotemporal patterns of terrestrial carbon cycle during the 20th century. Global Biogeochem. Cycles, 23, GB4026, doi:10.1029/2008GB003339.

    • Search Google Scholar
    • Export Citation
  • Piao, S. L., and et al. , 2013: Evaluation of terrestrial carbon cycle models for their response to climate variability and to CO2 trends. Global Change Biol., 19, 21172132, doi:10.1111/gcb.12187.

    • Search Google Scholar
    • Export Citation
  • Potter, C., , S. Klooster, , M. Steinbach, , P. Tan, , V. Kumar, , S. Shekhar, , R. Nemani, , and R. Myneni, 2003: Global teleconnections of climate to terrestrial carbon flux. J. Geophys. Res., 108, 4556, doi:10.1029/2002JD002979.

    • Search Google Scholar
    • Export Citation
  • Poulter, B., and et al. , 2014: Contribution of semi-arid ecosystems to interannual variability of the global carbon cycle. Nature, 509, 600603, doi:10.1038/nature13376.

    • Search Google Scholar
    • Export Citation
  • Poulter, B., and et al. , 2015: Sensitivity of global terrestrial carbon cycle dynamics to variability in satellite-observed burned area. Global Biogeochem. Cycles, 29, 207222, doi:10.1002/2013GB004655.

    • Search Google Scholar
    • Export Citation
  • Prentice, I. C., , D. I. Kelley, , P. N. Foster, , P. Friedlingstein, , S. P. Harrison, , and P. J. Bartlein, 2011: Modeling fire and the terrestrial carbon balance. Global Biogeochem. Cycles, 25, GB3005, doi:10.1029/2010GB003906.

    • Search Google Scholar
    • Export Citation
  • Qian, H., , R. Joseph, , and N. Zeng, 2008: Response of the terrestrial carbon cycle to the El Niño–Southern Oscillation. Tellus, 60B, 537550, doi:10.1111/j.1600-0889.2008.00360.x.

    • Search Google Scholar
    • Export Citation
  • Raddatz, T. J., and et al. , 2007: Will the tropical land biosphere dominate the climate–carbon cycle feedback during the twenty-first century? Climate Dyn., 29, 565574, doi:10.1007/s00382-007-0247-8.

    • Search Google Scholar
    • Export Citation
  • Raupach, M. R., , J. G. Canadell, , and C. Le Quéré, 2008: Anthropogenic and biophysical contributions to increasing atmospheric CO2 growth rate and airborne fraction. Biogeosciences, 5, 16011613, doi:10.5194/bg-5-1601-2008.

    • Search Google Scholar
    • Export Citation
  • Rayner, P. J., , and R. M. Law, 1999: The interannual variability of the global carbon cycle. Tellus, 51B, 210212, doi:10.1034/j.1600-0889.1999.t01-1-00007.x.

    • Search Google Scholar
    • Export Citation
  • Rayner, P. J., , R. M. Law, , and R. Dargaville, 1999: The relationship between tropical CO2 fluxes and the El Niño–Southern Oscillation. Geophys. Res. Lett., 26, 493496, doi:10.1029/1999GL900008.

    • Search Google Scholar
    • Export Citation
  • Renwick, J. A., , and J. M. Wallace, 1996: Relationships between North Pacific wintertime blocking, El Niño, and the PNA pattern. Mon. Wea. Rev., 124, 20712076, doi:10.1175/1520-0493(1996)124<2071:RBNPWB>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Rödenbeck, C., , S. Houweling, , M. Gloor, , and M. Heimann, 2003: CO2 flux history 1982–2001 inferred from atmospheric data using a global inversion of atmospheric transport. Atmos. Chem. Phys., 3, 19191964, doi:10.5194/acp-3-1919-2003.

    • Search Google Scholar
    • Export Citation
  • Ropelewski, C. F., , and M. S. Halpert, 1987: Global and regional scale precipitation patterns associated with the El Niño/Southern Oscillation. Mon. Wea. Rev., 115, 16061626, doi:10.1175/1520-0493(1987)115<1606:GARSPP>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Sarmiento, J. L., , M. Gloor, , N. Gruber, , C. Beaulieu, , A. R. Jacobson, , S. E. Mikaloff, , Fletcher, , S. Pacala, , and K. Rodgers, 2010: Trends and regional distributions of land and ocean carbon sinks. Biogeosciences, 7, 23512367, doi:10.5194/bg-7-2351-2010.

    • Search Google Scholar
    • Export Citation
  • Schwalm, C. R., , C. A. Williams, , K. Schaefer, , I. Baker, , G. J. Collatz, , and C. Rödenbeck, 2011: Does terrestrial drought explain global CO2 flux anomalies induced by El Niño? Biogeosciences, 8, 24932506, doi:10.5194/bg-8-2493-2011.

    • Search Google Scholar
    • Export Citation
  • Shao, P., , X. B. Zeng, , K. Sakaguchi, , R. K. Monson, , and X. D. Zeng, 2013: Terrestrial carbon cycle: Climate relations in eight CMIP5 Earth system models. J. Climate, 26, 87448764, doi:10.1175/JCLI-D-12-00831.1.

    • Search Google Scholar
    • Export Citation
  • Sitch, S., and et al. , 2008: Evaluation of the terrestrial carbon cycle, future plant geography, and climate-carbon cycle feedbacks using five dynamic global vegetation models (DGVMs). Global Change Biol., 14, 20152039, doi:10.1111/j.1365-2486.2008.01626.x.

    • Search Google Scholar
    • Export Citation
  • Smith, T. M., , and R. W. Reynolds, 2004: Improved extended reconstruction of SST (1854–1997). J. Climate, 17, 24662477, doi:10.1175/1520-0442(2004)017<2466:IEROS>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Taylor, K. E., , R. J. Stouffer, , and G. A. Meehl, 2012: An overview of CMIP5 and the experiment design. Bull. Amer. Meteor. Soc., 93, 485498, doi:10.1175/BAMS-D-11-00094.1.

    • Search Google Scholar
    • Export Citation
  • Thoning, K. W., , P. P. Tans, , and W. D. Komhy, 1989: Atmospheric carbon dioxide at Mauna Loa Observatory: 2. Analysis of the NOAA GMCC data, 1974–1985. J. Geophys. Res., 94, 85498565, doi:10.1029/JD094iD06p08549.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., 1997: The definition of El Niño. Bull. Amer. Meteor. Soc., 78, 27712777, doi:10.1175/1520-0477(1997)078<2771:TDOENO>2.0.CO;2.

    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., , J. M. Caron, , D. P. Stepaniak, , and S. Worley, 2002: Evolution of El Niño–Southern Oscillation and global atmospheric surface temperatures. J. Geophys. Res., 107, 4065, doi:10.1029/2000JD000298.

    • Search Google Scholar
    • Export Citation
  • Tsai, C., , S. K. Behera, , and T. Waseda, 2015: South Asia monsoon indices. Sci. Rep., 5, 8107, doi:10.1038/srep08107.

  • Van der Werf, G. R., , J. T. Randerson, , L. Giglio, , G. J. Collatz, , P. S. Kasibhatla, , and A. F. Arellano, 2006: Interannual variability in global biomass burning emissions from 1997 to 2004. Atmos. Chem. Phys., 6, 34233441, doi:10.5194/acp-6-3423-2006.

    • Search Google Scholar
    • Export Citation
  • Van der Werf, G. R., and et al. , 2010: Global fire emissions and the contribution of deforestation, savanna, forest, agricultural, and peat fires (1997–2009). Atmos. Chem. Phys., 10, 11 70711 735, doi:10.5194/acp-10-11707-2010.

    • Search Google Scholar
    • Export Citation
  • Wang, J., , P. M. Rich, , and K. P. Price, 2003: Temporal responses of NDVI to precipitation and temperature in the central Great Plains, USA. Int. J. Remote Sens., 24, 23452364, doi:10.1080/01431160210154812.

    • Search Google Scholar
    • Export Citation
  • Wang, J., , N. Zeng, , Y. Liu, , and Q. Bao, 2014: To what extent can interannual CO2 variability constrain carbon cycle sensitivity to climate change in CMIP5 Earth system models? Geophys. Res. Lett., 41, 35353544, doi:10.1002/2014GL060004.

    • Search Google Scholar
    • Export Citation
  • Wang, W. L., and et al. , 2013: Variations in atmospheric CO2 growth rates coupled with tropical temperature. Proc. Natl. Acad. Sci. USA, 110, 13 06113 066, doi:10.1073/pnas.1219683110.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., and et al. , 2010: Improved climate simulation by MIROC5: Mean states, variability, and climate sensitivity. J. Climate, 23, 63126335, doi:10.1175/2010JCLI3679.1.

    • Search Google Scholar
    • Export Citation
  • Watanabe, S., and et al. , 2011: MIROC-ESM 2010: Model description and basic results of CMIP5-20c3m experiments. Geosci. Model Dev., 4, 845872, doi:10.5194/gmd-4-845-2011.

    • Search Google Scholar
    • Export Citation
  • Weare, B. C., 2013: El Niño teleconnections in CMIP5 models. Climate Dyn., 41, 21652177, doi:10.1007/s00382-012-1537-3.

  • Wenzel, S., , P. M. Cox, , V. Eyring, , and P. Friedlingstein, 2014: Emergent constraints on climate-carbon cycle feedbacks in the CMIP5 Earth system models. J. Geophys. Res. Biogeosci., 119, 794807, doi:10.1002/2013JG002591.

    • Search Google Scholar
    • Export Citation
  • Wetzel, P., , A. Winguth, , and E. Maier-Reimer, 2005: Sea-to-air CO2 flux from 1948 to 2003: A model study. Global Biogeochem. Cycles, 19, GB2005, doi:10.1029/2004GB002339.

    • Search Google Scholar
    • Export Citation
  • Wittenberg, A. T., , A. Rosati, , N. C. Lau, , and J. J. Ploshay, 2006: GFDL’s CM2 global coupled climate models. Part III: Tropical Pacific climate and ENSO. J. Climate, 19, 698722, doi:10.1175/JCLI3631.1.

    • Search Google Scholar
    • Export Citation
  • Yeh, S. W., , Y. G. Ham, , and J. Y. Lee, 2012: Changes in the tropical Pacific SST trend from CMIP3 to CMIP5 and its implication of ENSO. J. Climate, 25, 77647771, doi:10.1175/JCLI-D-12-00304.1.

    • Search Google Scholar
    • Export Citation
  • Zeng, N., , A. Mariotti, , and P. Wetzel, 2005: Terrestrial mechanisms of interannual CO2 variability. Global Biogeochem. Cycles, 19, GB1016, doi:10.1029/2004GB002273.

    • Search Google Scholar
    • Export Citation
  • Zeng, N., and et al. , 2008: Dynamical prediction of terrestrial ecosystems and the global carbon cycle: A 25-year hindcast experiment. Global Biogeochem. Cycles, 22, GB4015, doi:10.1029/2008GB003183.

    • Search Google Scholar
    • Export Citation
  • Zhao, M., , and S. W. Running, 2010: Drought-induced reduction in global terrestrial net primary production from 2000 through 2009. Science, 329, 940943, doi:10.1126/science.1192666.

    • Search Google Scholar
    • Export Citation
Save