• Anderegg, W. R. L., J. M. Kane, and L. D. L. Anderegg, 2013: Consequences of widespread tree mortality triggered by drought and temperature stress. Nat. Climate Change, 3, 3036, doi:10.1038/nclimate1635.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Andrews, T., M. Doutriaux-Boucher, O. Boucher, and P. M. Forster, 2011: A regional and global analysis of carbon dioxide physiological forcing and its impact on climate. Climate Dyn., 36, 783792, doi:10.1007/s00382-010-0742-1.

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

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bathiany, S., M. Claussen, and V. Brovkin, 2014: CO2-induced Sahel greening in three CMIP5 Earth system models. J. Climate, 27, 71637184, doi:10.1175/JCLI-D-13-00528.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, R. A., P. M. Cox, S. E. Lee, and F. I. Woodward, 1997: Contrasting physiological and structural vegetation feedbacks in climate change simulations. Nature, 387, 796799, doi:10.1038/42924.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, R. A., P. M. Cox, M. Collins, P. P. Harris, C. Huntingford, and C. D. Jones, 2004: The role of ecosystem–atmosphere interactions in simulated Amazonian precipitation decrease and forest dieback under global climate warming. Theor. Appl. Climatol., 78, 157175, doi:10.1007/s00704-004-0050-y.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, R. A., and Coauthors, 2007: Projected increase in continental runoff due to plant responses to increasing carbon dioxide. Nature, 448, 10371041, doi:10.1038/nature06045.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bitz, C. M., and Coauthors, 2012: Climate sensitivity of the Community Climate System Model, version 4. J. Climate, 25, 30533070, doi:10.1175/JCLI-D-11-00290.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Boucher, O., A. Jones, and R. A. Betts, 2009: Climate response to the physiological impact of carbon dioxide on plants in the Met Office Unified Model HadCM3. Climate Dyn., 32, 237249, doi:10.1007/s00382-008-0459-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cao, L., G. Bala, K. Caldeira, R. Nemani, and G. Ban-Weiss, 2010: Importance of carbon dioxide physiological forcing to future climate change. Proc. Natl. Acad. Sci. USA, 107, 95139518, 10.1073/pnas.0913000107.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chadwick, R., I. Boutle, and G. Martin, 2013: Spatial patterns of precipitation change in CMIP5: Why the rich do not get richer in the tropics. J. Climate, 26, 38033822, doi:10.1175/JCLI-D-12-00543.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chadwick, R., H. Douville, and C. B. Skinner, 2017: Timeslice experiments for understanding regional climate projections: Applications to the tropical hydrological cycle and European winter circulation. Climate Dyn., doi:10.1007/s00382-016-3488-6, in press.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cowling, S. A., and C. B. Field, 2003: Environmental control of leaf area production: Implications for vegetation and land-surface modeling. Global Biogeochemical Cycles, 17, 1007, doi:10.1029/2002GB001915.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cox, P. M., and Coauthors, 2008: Increasing risk of Amazonian drought due to decreasing aerosol pollution. Nature, 453, 212215, doi:10.1038/nature06960.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Douville, H., S. Planton, J.-F. Royer, D. B. Stephenson, S. Tyteca, L. Kergoat, S. Lafont, and R. A. Betts, 2000: Importance of vegetation feedbacks in doubled-CO2 climate experiments. J. Geophys. Res., 105, 14 84114 861, doi:10.1029/1999JD901086.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eltahir, E. A. B., 1998: A soil moisture–rainfall feedback mechanism: 1. Theory and observations. Water Resour. Res., 34, 765776, doi:10.1029/97WR03499.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Field, C. B., R. B. Jackson, and H. A. Mooney, 1995: Stomatal responses to increased CO2: Implications from the plant to the global scale. Plant Cell Environ., 18, 12141225, doi:10.1111/j.1365-3040.1995.tb00630.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Field, C. B., and Coauthors, 2012: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. Cambridge University Press, 582 pp.

  • Frank, D., and Coauthors, 2015: Effects of climate extremes on the terrestrial carbon cycle: Concepts, processes and potential future impacts. Global Change Biol., 21, 28612880, doi:10.1111/gcb.12916.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fu, R., and W. Li, 2004: The influence of the land surface on the transition from dry to wet season in Amazonia. Theor. Appl. Climatol., 78, 97110, doi:10.1007/s00704-004-0046-7.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gedney, N., P. M. Cox, R. A. Betts, O. Boucher, C. Huntingford, and P. A. Stott, 2006: Detection of a direct carbon dioxide effect in continental river runoff records. Nature, 439, 835838, doi:10.1038/nature04504.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gent, P. R., and Coauthors, 2011: The Community Climate System Model version 4. J. Climate, 24, 49734991, doi:10.1175/2011JCLI4083.1.

  • Good, S. P., D. Noone, and G. Bowen, 2015: Hydrologic connectivity constrains partitioning of global terrestrial water fluxes. Science, 349, 175177, doi:10.1126/science.aaa5931.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gu, H., Z. Zong, and K. C. Hung, 2004: A modified superconvergent patch recovery method and its application to large deformation problems. Finite Elem. Anal. Des., 40, 665687, doi:10.1016/S0168-874X(03)00109-4.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Held, I. M., and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 56865699, doi:10.1175/JCLI3990.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hirabayashi, Y., and Coauthors, 2013: Global flood risk under climate change. Nat. Climate Change, 3, 816821, doi:10.1038/nclimate1911.

  • Huffman, G. J., R. Adler, M. Morrissey, D. Bolvin, S. Curtis, R. Joyce, B. McGavock, and J. Susskind, 2001: Global precipitation at one-degree daily resolution from multisatellite observations. J. Hydrometeor., 2, 3650, doi:10.1175/1525-7541(2001)002<0036:GPAODD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hunke, E. C., and W. H. Lipscomb, 2008: CICE: The Los Alamos sea ice model: Documentation and software user’s manual. Tech. Rep. LA-CC-06-012, v.4.0, 72 pp.

  • Jones, C. D., and Coauthors, 2011: The HadGEM2-ES implementation of CMIP5 centennial simulations. Geosci. Model Dev., 4, 543570, doi:10.5194/gmd-4-543-2011.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kala, J., M. G. De Kauwe, A. J. Pitman, B. E. Medlyn, Y. P. Wang, R. Lorenz, and S. E. Perkins-Kirkpatrick, 2016: Impact of the representation of stomatal conductance on model projections of heatwave intensity. Sci. Rep., 6, 23418, doi:10.1038/srep23418.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kent, C., R. Chadwick, and D. P. Rowell, 2015: Understanding uncertainties in future projections of seasonal tropical precipitation. J. Climate, 28, 43904413, doi:10.1175/JCLI-D-14-00613.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kergoat, L., S. Lafont, H. Douville, B. Berthelot, G. Dedieu, S. Planton, and J.-F. Royer, 2002: Impact of doubled CO2 on global-scale leaf area index and evapotranspiration: Conflicting stomatal conductance and LAI responses. J. Geophys. Res., 107, 4808, doi:10.1029/2001JD001245.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kloster, S., F. Dentener, J. Feichter, F. Raes, U. Lohmann, E. Roeckner, and I. Fischer-Bruns, 2010: A GCM study of future climate response to aerosol pollution reductions. Climate Dyn., 34, 11771194, doi:10.1007/s00382-009-0573-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lammertsma, E. I., H. J. de Boer, S. C. Dekker, D. L. Dilcher, A. F. Lotter, and F. Wagner-Cremer, 2011: Global CO2 rise leads to reduced maximum stomatal conductance in Florida vegetation. Proc. Natl. Acad. Sci. USA, 108, 40354040, doi:10.1073/pnas.1100371108.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lau, W. K.-M., H.-T. Wu, and K.-M. Kim, 2013: A canonical response of precipitation characteristics to global warming from CMIP5 models. Geophys. Res. Lett., 40, 31633169, doi:10.1002/grl.50420.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., P. E. Thornton, K. W. Oleson, and G. B. Bonan, 2007: The partitioning of evapotranspiration into transpiration, soil evaporation, and canopy evaporation in a GCM: Impacts on land–atmosphere interaction. J. Hydrometeor., 8, 862880, doi:10.1175/JHM596.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, J.-E., and Coauthors, 2012: Reduction of tropical land region precipitation variability via transpiration. Geophys. Res. Lett., 39, L19704, doi:10.1029/2012GL053417.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Levis, S., J. A. Foley, and D. Pollard, 2000: Large-scale vegetation feedbacks on a doubled CO2 climate. J. Climate, 13, 13131325, doi:10.1175/1520-0442(2000)013<1313:LSVFOA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, J.-L., and Coauthors, 2006: Tropical intraseasonal variability in 14 IPCC AR4 climate models. Part I: Convective signals. J. Climate, 19, 26652690, doi:10.1175/JCLI3735.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, Y.-S., and Coauthors, 2015: Optimal stomatal behaviour around the world. Nat. Climate Change, 5, 459464, doi:10.1038/nclimate2550.

  • Lindsay, K., and Coauthors, 2014: Preindustrial-control and twentieth-century carbon cycle experiments with the Earth System Model CESM1(BGC). J. Climate, 27, 89819005, doi:10.1175/JCLI-D-12-00565.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lloyd, J., and G. D. Farquhar, 2008: Effects of rising temperatures and [CO2] on the physiology of tropical forest trees. Philos. Trans. Roy. Soc. London, 363B, 18111817, doi:10.1098/rstb.2007.0032.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahowald, N., F. Lo, Y. Zheng, L. Harrison, C. Funk, D. Lombardozzi, and C. Goodale, 2016: Projections of leaf area index in Earth system models. Earth Syst. Dynam., 7, 211229, doi:10.5194/esd-7-211-2016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Malhi, Y., J. T. Roberts, R. A. Betts, T. J. Killeen, W. Li, and C. A. Nobre, 2008: Climate change, deforestation, and the fate of the Amazon. Science, 319, 169172, doi:10.1126/science.1146961.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Milly, P. C. D., and K. A. Dunne, 2016: Potential evapotranspiration and continental drying. Nat. Climate Change, 6, 946949, doi:10.1038/nclimate3046.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Neale, R. B., J. H. Richter, A. J. Conley, S. Park, P. H. Lauritzen, A. Gettleman, and D. L. Williamson, 2010: Description of the NCAR Community Atmosphere Model (CAM5.0). NCAR Tech. Rep. NCAR/TN-486+STR, 274 pp.

  • Notaro, M., S. Vavrus, and Z. Liu, 2007: Global vegetation and climate change due to future increases in CO2 as projected by a fully coupled model with dynamic vegetation. J. Climate, 20, 7090, doi:10.1175/JCLI3989.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • O’Gorman, P. A., and T. Schneider, 2009: The physical basis for increases in precipitation extremes in simulations of 21st-century climate change. Proc. Natl. Acad. Sci. USA, 106, 14 77314 777, doi:10.1073/pnas.0907610106.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Oleson, K. W., and Coauthors, 2010: Technical description of version 4.0 of the Community Land Model (CLM). NCAR Tech. Note NCAR/TN-478+STR, 257 pp.

  • Pearson, R. G., S. J. Phillips, M. M. Loranty, P. S. A. Beck, T. Damoulas, S. J. Knight, and S. J. Goetz, 2013: Shifts in Arctic vegetation and associated feedbacks under climate change. Nat. Climate Change, 3, 673677, doi:10.1038/nclimate1858.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peng, J., L. Dan, and W. Dong, 2014: Are there interactive effects of physiological and radiative forcing produced by increased CO2 concentration on changes of land hydrological cycle? Global Planet. Change, 112, 6478, doi:10.1016/j.gloplacha.2013.11.007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pielke, R. A., 2001: Influence of the spatial distribution of vegetation and soils on the prediction of cumulus convective rainfall. Rev. Geophys., 39, 151177, doi:10.1029/1999RG000072.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pitman, A. J., and M. Zhao, 2000: The relative impact of observed change in land cover and carbon dioxide as simulated by a climate model. Geophys. Res. Lett., 27, 12671270, doi:10.1029/1999GL011029.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pu, B., and R. E. Dickinson, 2014: Hydrological changes in the climate system from leaf responses to increasing CO2. Climate Dyn., 42, 19051923, doi:10.1007/s00382-013-1781-1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schär, C., D. Lüthi, U. Beyerle, and E. Heise, 1999: The soil–precipitation feedback: A process study with a regional climate model. J. Climate, 12, 722741, doi:10.1175/1520-0442(1999)012<0722:TSPFAP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., D. Lüthi, M. Litschi, and C. Schär, 2006: Land–atmosphere coupling and climate change in Europe. Nature, 443, 205209, doi:10.1038/nature05095.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Skinner, C. B., M. Ashfaq, and N. S. Diffenbaugh, 2012: Influence of twenty-first-century atmospheric and sea surface temperature forcing on West African climate. J. Climate, 25, 527542, doi:10.1175/2011JCLI4183.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sun, Y., S. Solomon, A. Dai, and R. W. Portmann, 2007: How often will it rain? J. Climate, 20, 48014818, doi:10.1175/JCLI4263.1.

  • Swann, A. L. S., F. M. Hoffman, C. D. Koven, and J. T. Randerson, 2016: Plant responses to increasing CO2 reduce estimates of climate impacts on drought severity. Proc. Natl. Acad. Sci. USA, 113, 10 01910 024, doi:10.1073/pnas.1604581113.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Taylor, C. M., R. A. de Jeu, F. Guichard, P. P. Harris, and W. A. Dorigo, 2012: Afternoon rain more likely over drier soils. Nature, 489, 423426, doi:10.1038/nature11377.

    • Crossref
    • 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.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., A. Dai, R. M. Rasmussen, and D. B. Parsons, 2003: The changing character of precipitation. Bull. Amer. Meteor. Soc., 84, 12051217, doi:10.1175/BAMS-84-9-1205.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van der Ent, R. J., L. Wang-Erlandsson, P. W. Keys, and H. H. G. Savenije, 2014: Contrasting roles of interception and transpiration in the hydrological cycle—Part 2: Moisture recycling. Earth Syst. Dynam., 5, 471489, doi:10.5194/esd-5-471-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van Vuuren, D. P., and Coauthors, 2011: The representative concentration pathways: An overview. Climatic Change, 109, 531, doi:10.1007/s10584-011-0148-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, S., Y. Yang, A. P. Trishchenko, A. G. Barr, T. A. Black, and H. McCaughey, 2009: Modeling the response of canopy stomatal conductance to humidity. J. Hydrometeor., 10, 521532, doi:10.1175/2008JHM1050.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhu, Z., and Coauthors, 2016: Greening of the Earth and its drivers. Nat. Climate Change, 6, 791795, doi:10.1038/nclimate3004.

  • View in gallery

    Zonal mean of annual (a) evapotranspiration and (b) evapotranspiration change from CCSM4 and 15 CMIP5 models. Thick black line in (a) denotes the historical period CCSM4 simulation HistRad_HistStomata_HistLAI. Thick black line in (b) denotes the change (CO2_Total) between CCSM4 simulations FutRad_FutStomata_FutLAI and HistRad_HistStomata_HistLAI. Blue lines in (a) denote the historical period average (1980–2005) from the CMIP5 models. Blue lines in (b) denote the change between the future RCP8.5 (2080–2100) and historical simulations from the CMIP5 models. Only land grid points are used in the zonal mean calculations. All models are regridded to a common 1° × 1° grid. CMIP5 models are listed in Table S1.

  • View in gallery

    Change in annual mean evapotranspiration from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. Only statistically significant differences are shown.

  • View in gallery

    Percent contribution to annual mean evapotranspiration change from (a) CO2_Rad, (b) CO2_Stomata, (c) CO2_LAI, and (d) the nonlinear interaction term (see section 2).

  • View in gallery

    Change in annual mean (left) transpiration and (right) evaporation from (a),(b) CO2_Total, (c),(d) CO2_Rad, (e),(f) CO2_Stomata, and (g),(h) CO2_LAI. Only statistically significant differences are shown. Evaporation over the ocean is masked out.

  • View in gallery

    Annual mean precipitation from (a) GPCP (1997–2014) and (b) the 30-yr historical period CCSM4 simulation HistRad_HistStomata_HistLAI. Zonal mean of annual (c) precipitation and (d) precipitation change from CCSM4 and 15 CMIP5 models. Thick black line in (c) denotes the historical period CCSM4 simulation HistRad_HistStomata_HistLAI and in (d) denotes the change (CO2_Total) between CCSM4 simulation FutRad_FutStomata_FutLAI and simulation HistRad_HistStomata_HistLAI. Blue lines in (c) denote the historical period average (1980–2005) from the CMIP5 models and in (d) denote the change between the future RCP8.5 (2080–2100) and historical simulations from the CMIP5 models. Only land grid points are used in zonal mean calculations for (c) and (d). All models are regrid to a common 1° × 1° grid. CMIP5 models are listed in Table S1.

  • View in gallery

    Change in annual mean precipitation from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. Only statistically significant differences are shown.

  • View in gallery

    Percent contribution to annual mean precipitation change from (a) CO2_Rad, (b) CO2_Stomata, (c) CO2_LAI and (d) the nonlinear interaction term (see section 2).

  • View in gallery

    Change in seasonal mean precipitation from CO2_Stomata: (a) December–February, (b) March–May, (c) June–August, and (d) September–November.

  • View in gallery

    Change in annual number of days without precipitation from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. A day is considered to have no precipitation when the accumulated daily precipitation is less than 0.1 mm. Only statistically significant differences are shown.

  • View in gallery

    Change in the annual number of days with heavy precipitation from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. Heavy rainfall days are those that exceed the 95th percentile event from the historical control simulation (year 2000 conditions). Only statistically significant differences are shown.

  • View in gallery

    Change in the (a) annual number of days with light precipitation and (b) annual mean low-cloud fraction from CO2_Stomata. A day is considered to have light precipitation when the accumulated daily precipitation is greater than 0.1 mm and less than 2 mm. Only statistically significant differences are shown in (a). Brown shading indicates fewer days with light precipitation in (a).

  • View in gallery

    Change in the annual number of (a),(b) dry days and (c),(d) days with heavy precipitation from CO2_Stomata simulations with (left) HadGEM2-A and (right) CESM1-BGC_esmFixClim1 (see section 2). A day is considered to have no precipitation when the accumulated daily precipitation is less than 0.1 mm. Heavy rainfall days are those that exceed the 95th percentile event from the historical period (first 30 yr). Only statistically significant differences are shown.

  • View in gallery

    Change in annual mean (a) surface latent heat flux, (b) surface temperature, (c) vertically integrated atmospheric moisture and moisture flux, and (d) CAPE from CO2_Stomata.

  • View in gallery

    Box-and-whisker plots represent the number of dry days (<0.1 mm) for each month. Lines represent the average daily precipitation. Red (blue) coloring indicates data from a simulation with (without) future changes in stomatal conductance from CO2. Box-and-whisker plots show min, max, and interquartile range, and consist of 30 yr of data. Average daily precipitation is averaged over 30 years. Area weighted averages from (a) eastern Amazon (10°S–0°, 65°–50°W), (b) northern Congo (0°–10°N, 15°–30°E), (c) Europe (40°–55°N, 0°–15°E), and (d) midlatitude South America (30°–40°S, 65°–55°W).

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 156 156 31
PDF Downloads 134 134 38

The Role of Plant CO2 Physiological Forcing in Shaping Future Daily-Scale Precipitation

View More View Less
  • 1 Department of Earth and Environmental Sciences, University of Michigan, Ann Arbor, Michigan
  • 2 Met Office Hadley Centre, Exeter, United Kingdom
  • 3 Department of Earth System Science, and Woods Institute for the Environment, Stanford University, Stanford, California
  • 4 Department of Earth and Environmental Sciences, University of Michigan, Ann Arbor, Michigan, and Department of Geology and Geophysics, University of Utah, Salt Lake City, Utah
© Get Permissions
Full access

Abstract

Continued anthropogenic CO2 emissions are expected to drive widespread changes in precipitation characteristics. Nonetheless, projections of precipitation change vary considerably at the regional scale between climate models. Here, it is shown that the response of plant physiology to elevated CO2, or CO2 physiological forcing drives widespread hydrologic changes distinct from those associated with CO2 radiative forcing and has a role in shaping regional-scale differences in projected daily-scale precipitation changes. In a suite of simulations with the Community Climate System Model, version 4 (CCSM4), reduced stomatal conductance from projected physiological forcing drives large decreases in transpiration and changes the distribution of daily-scale precipitation within and adjacent to regions of dense vegetation and climatologically high transpiration. When atmospheric conditions are marginally favorable for precipitation, reduced transpiration dries the boundary layer and increases the likelihood of dry day occurrence. In CCSM4, the annual number of dry days increases by upward of 15 days yr−1 over tropical land and the continental midlatitudes. Decreases in transpiration from physiological forcing also increase the number of heavy precipitation events by up to 8 days yr−1 in many tropical forest regions. Despite reductions in the land surface contribution to atmospheric moisture, diminished surface latent heat fluxes warm the forest boundary layer and increase moisture convergence from nearby oceans, enhancing instability. The results suggest that consideration of the radiative impacts of CO2 alone cannot account for projected regional-scale differences in daily precipitation changes, and that CO2 physiological forcing may contribute to differences in projected precipitation characteristics among climate models.

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author e-mail: Christopher B. Skinner, chrisbs@umich.edu

Supplemental information related to this paper is available at the Journals Online website: http://dx.doi.org/10.1175/JCLI-D-16-0603.s1.

Abstract

Continued anthropogenic CO2 emissions are expected to drive widespread changes in precipitation characteristics. Nonetheless, projections of precipitation change vary considerably at the regional scale between climate models. Here, it is shown that the response of plant physiology to elevated CO2, or CO2 physiological forcing drives widespread hydrologic changes distinct from those associated with CO2 radiative forcing and has a role in shaping regional-scale differences in projected daily-scale precipitation changes. In a suite of simulations with the Community Climate System Model, version 4 (CCSM4), reduced stomatal conductance from projected physiological forcing drives large decreases in transpiration and changes the distribution of daily-scale precipitation within and adjacent to regions of dense vegetation and climatologically high transpiration. When atmospheric conditions are marginally favorable for precipitation, reduced transpiration dries the boundary layer and increases the likelihood of dry day occurrence. In CCSM4, the annual number of dry days increases by upward of 15 days yr−1 over tropical land and the continental midlatitudes. Decreases in transpiration from physiological forcing also increase the number of heavy precipitation events by up to 8 days yr−1 in many tropical forest regions. Despite reductions in the land surface contribution to atmospheric moisture, diminished surface latent heat fluxes warm the forest boundary layer and increase moisture convergence from nearby oceans, enhancing instability. The results suggest that consideration of the radiative impacts of CO2 alone cannot account for projected regional-scale differences in daily precipitation changes, and that CO2 physiological forcing may contribute to differences in projected precipitation characteristics among climate models.

© 2017 American Meteorological Society. For information regarding reuse of this content and general copyright information, consult the AMS Copyright Policy (www.ametsoc.org/PUBSReuseLicenses).

Corresponding author e-mail: Christopher B. Skinner, chrisbs@umich.edu

Supplemental information related to this paper is available at the Journals Online website: http://dx.doi.org/10.1175/JCLI-D-16-0603.s1.

1. Introduction

Terrestrial evapotranspiration (ET) accounts for roughly 35% of the moisture that falls as precipitation on land (van der Ent et al. 2014). The contribution of regional evapotranspiration to total regional rainfall, the precipitation recycling ratio, varies in space and time and is generally highest where surface and root zone soil moisture is abundant and when advective moisture fluxes are relatively small (Trenberth et al. 2003). Regions of high precipitation recycling over land rely on terrestrial moisture sources for precipitation and are particularly sensitive to any forcing that impacts local or upwind land ET rates, such as changes in land cover or greenhouse gas concentrations (Pielke 2001; Malhi et al. 2008; Seneviratne et al. 2006). Further, variations in the partitioning of total ET between transpiration and evaporation can impact surface energy flux variability and drive changes in the intensity and frequency of individual rainfall events, particularly in regions of high ET and precipitation recycling (Lawrence et al. 2007; Lee et al. 2012). Estimates of future changes in ET and its partitioning into evaporation and transpiration are essential for projecting rainfall changes and the associated impacts on humans and ecosystems worldwide.

Enhanced radiative forcing due to increased carbon dioxide (CO2) concentrations, among other greenhouse gases, is expected to drive widespread changes in ET (evaporation over ocean, and evaporation plus transpiration over land) (Milly and Dunne 2016; Swann et al. 2016). Plant physiological responses to increased CO2 also promote changes in ET unrelated to the radiative effects of CO2. In response to high ambient CO2 concentrations, many plants decrease stomatal conductance and transpiration in order to minimize water loss, increasing their plant water-use efficiency (the ratio of carbon assimilation rate to transpiration rate) (Field et al. 1995; Lammertsma et al. 2011). These decreases in transpiration affect the amount and variability of soil moisture and ET, both of which have been shown to influence the frequency, intensity, and duration of precipitation events (Eltahir 1998; Lawrence et al. 2007; C. Taylor et al. 2012; Lee et al. 2012). Given that transpiration accounts for roughly 64% of total terrestrial ET (Good et al. 2015), understanding the complex and often competing responses of vegetation–water interactions to CO2 variations is critical for future hydrologic assessment.

Climate model experiments designed to isolate the impact of projected CO2 physiological forcing (stomata changes) on future climate, with vegetation structure and vegetation location fixed, exhibit widespread reductions in transpiration in response to a doubling of atmospheric CO2 (Betts et al. 1997; Douville et al. 2000; Cao et al. 2010; Andrews et al. 2011; Pu and Dickinson 2014). Reduced transpiration rates are highest in the midlatitudes during the summer season, and in tropical forest regions (Douville et al. 2000; Pu and Dickinson 2014). Reductions in transpiration impact local precipitation directly by diminishing a source of moisture to the atmosphere and indirectly by altering regional-scale circulation (Betts et al. 2004, 2007; Boucher et al. 2009). Climate models project the greatest changes in precipitation (positive and negative) from future CO2 physiological forcing in densely vegetated areas of the tropics (Betts et al. 1997; Swann et al. 2016).

Climate model sensitivity studies suggest that reduced transpiration leads to greater precipitation variability in regions of high transpiration and high precipitation recycling (Lawrence et al. 2007; Lee et al. 2012). Lee et al. (2012) set the transpiration flux to zero within modern-day climate simulations with the Community Climate System Model, version 3 (CCSM3), and found that tropical forests experience a higher frequency of both precipitation-free days and days with extremely heavy precipitation (>97th percentile). Similarly, Lawrence et al. (2007) modified vegetation and soil hydrology parameterizations in a present-day climate simulation with the Community Land Model, version 3 (CLM3), and found that local increases in transpiration at the expense of canopy evaporation shift the distribution of rainfall toward more frequent, lighter events. The results of Lawrence et al. (2007) and Lee et al. (2012) suggest that reductions in transpiration from CO2 physiological forcing may contribute to increased precipitation variability in regions of high climatological transpiration and in regions that rely heavily on transpiration for rainfall. However, the prescribed transpiration changes in these studies are highly idealized, and the response of precipitation to projected, physically based variations in transpiration from CO2 physiological forcing is unclear.

While elevated CO2 is expected to reduce stomatal conductance, projected CO2 fertilization and the associated changes in vegetation density and location further complicate the response of climate to CO2 physiological forcing. In environments where CO2 supply limits gross primary production, increases in CO2 stimulate enhanced biomass production and greater leaf area index (LAI) (Cowling and Field 2003; Lloyd and Farquhar 2008; Zhu et al. 2016). Within climate model simulations of a high CO2 climate, this CO2 fertilization effect can lead to greater canopy-scale transpiration and partially or completely negate the diminished transpiration from reduced stomatal conductance, impacting regional-scale precipitation change (Betts et al. 1997; Levis et al. 2000; Kergoat et al. 2002).

Changes in precipitation characteristics have the potential to cause widespread impacts on natural and human systems (Field et al. 2012; Anderegg et al. 2013; Hirabayashi et al. 2013; Frank et al. 2015). Efforts to anticipate and prepare for these impacts are made easier when projections of precipitation change and the underlying processes that drive these changes are consistent among climate model simulations. At regional scales, there remains considerable uncertainty in projections of future precipitation characteristics and the processes that shape them (O’Gorman and Schneider 2009). To date, the potential impact of CO2 physiological forcing and CO2 fertilization on future changes in precipitation characteristics remains largely unexplored. Here, we investigate the link between future CO2–vegetation interactions and the hydrologic cycle with the goal of providing insights into the mechanisms that shape projected regional-scale changes in precipitation characteristics on daily time scales. We utilize a suite of sensitivity experiments with the Community Climate System Model, version 4 (CCSM4), to investigate the individual contributions of CO2 radiative forcing, CO2 physiological forcing, and CO2 fertilization on future hydrologic changes. Our results suggest that consideration of CO2 physiological forcing is critical for understanding regional differences in the projected response of precipitation characteristics to elevated CO2 concentrations.

2. Data and methods

To analyze the contributions of CO2 radiative forcing, CO2 physiological forcing, and CO2 fertilization to projected future hydrologic changes, we conduct a set of five 60-yr climate model simulations with the CCSM4 (Gent et al. 2011). Within CCSM4, the Community Atmosphere Model, version 4 (CAM4; Neale et al. 2010), and the Community Land Model, version 4 (CLM4; Oleson et al. 2010), are run on a 0.9° × 1.25° finite-volume grid. Sea ice is modeled with the fully active Community Ice Code, version 4 (CICE4; Hunke and Lipscomb 2008). The ocean is represented with a mixed layer, slab ocean model run at a nominal resolution of 1° (Bitz et al. 2012). The atmosphere has 26 terrain-following vertical levels, and the soil column is divided into 15 vertical layers. All simulations are initialized from the same point in a preindustrial CCSM4 run (denoted as b40.1850.track1.1deg.006) and therefore have identical initial conditions (land surface, atmosphere, ocean, and sea ice). We analyze the final 30 years of each 60-yr simulation. Our experimental design is similar to that of Pu and Dickinson (2014), although here our goal is to better understand changes in daily precipitation characteristics from future CO2–vegetation interactions.

We vary CO2 concentrations in the land and atmosphere models separately to identify the relative contributions of CO2 radiative versus physiological forcing (Table 1). Changes in atmospheric CO2 used in the atmosphere model capture the radiative effect, and are denoted as “Rad” in the simulation name. Changes in atmospheric CO2 used in the land model capture the plant physiological effect, and are denoted as “Stomata” in the simulation name. The atmospheric CO2 concentration in the atmosphere and land models is either prescribed at 367 ppm (historical, ~year 2000), denoted as “Hist” in the simulation name, or 900 ppm [future, ~year 2095 in the representative concentration pathway 8.5 (RCP8.5) scenario], denoted as “Fut” in the simulation name (van Vuuren et al. 2011). We also vary vegetation area between historical or future predicted values to estimate climate change due solely to changes in CO2 fertilization. Climatological (monthly varying) LAI and stem area index (SAI) values in each grid box and plant functional type (PFT) are prescribed from an end of the twentieth-century time period [mean of years 1996–2005 from phase 5 of the Coupled Model Intercomparison Project (CMIP5) CCSM4 “historical” scenario] denoted as “HistLAI,” or end of twenty-first-century time period (mean of years 2091–2100 from the CMIP5 CCSM4 future RCP8.5 scenario) denoted as “FutLAI.” Changes in LAI or SAI between the Hist and Fut simulations arise from the direct effects of CO2 fertilization on vegetation structure, and from the indirect effects of elevated CO2 on vegetation structure through changes in climate (temperature, moisture, etc.) (Fig. S1 in the supplemental material). The changes in LAI and SAI do not include changes in vegetation location; the type and amount of PFTs in each grid cell are the same in the Hist and Fut simulations.

Table 1.

Simulation names and climate forcings.

Table 1.

We isolate the influences of CO2 changes on hydroclimate due to radiative forcing, stomatal conductance changes (physiological forcing), and LAI changes (CO2 fertilization) by comparing simulations where one of these factors are changed to an historical control simulation (Table 2). By subtracting each experimental simulation from the future simulation (FutRad_FutStomata_FutLAI), we are able to assess the impact of both combined and individual forcings within the context of future end of twenty-first century climate. To assess the full impact of projected future changes in CO2 (CO2_Total), including the combined influences of radiative forcing, stomatal conductance, and LAI changes, we subtract the historical control simulation HistRad_HistStomata_HistLAI from the future simulation FutRad_FutStomata_FutLAI. To isolate the impact of future CO2 radiative forcing (CO2_Rad), we subtract HistRad_FutStomata_FutLAI from simulation FutRad_FutStomata_FutLAI. To isolate the impact of future CO2 physiological forcing (CO2_Stomata), we subtract simulation FutRad_HistStomata_FutLAI from simulation FutRad_FutStomata_FutLAI. Last, to isolate the impact of future CO2 fertilization (CO2_LAI), we subtract FutRad_FutStomata_HistLAI from simulation FutRad_FutStomata_FutLAI.

Table 2.

Experiment names and climate forcings.

Table 2.

CLM4 predicts leaf stomatal conductance separately for C3 and C4 plants using the Ball–Berry conductance model (Oleson et al. 2010). The model relates stomatal conductance to the net photosynthesis rate, atmospheric relative humidity, and the CO2 concentration at the leaf surface. Soil water indirectly impacts stomatal conductance via photosynthesis. Specifically, soil moisture stress (low soil moisture) reduces photosynthesis, which then limits stomatal conductance (Oleson et al. 2010). Increases in atmospheric CO2 promote decreases in stomatal conductance, assuming other environmental variables do not change. All other greenhouse gases and aerosol concentrations in the simulations are prescribed and fixed at year 2000 values for all simulations (CH4 = 1753 ppb and N2O = 319 ppb). Therefore, radiative forcing in our experiments refers only to the projected energy perturbation from changes in CO2. Consequently, our experiments underestimate the full radiative forcing expected from future greenhouse gas changes during the twenty-first century.

We follow the methodology of Pitman and Zhao (2000) and Skinner et al. (2012) to estimate the relative contribution of CO2 radiative forcing, CO2 physiological forcing, and CO2 fertilization to the total simulated climate response to elevated CO2. The calculation for the percent contribution of physiological forcing is as follows:
eq1
where
eq2
Likewise, we replace |CO2_Stomata| in the numerator with |CO2_Rad| to calculate the percent contribution of radiative forcing to the total climate response, etc. The final term in the denominator represents the nonlinear interaction between the CO2 radiative forcing, CO2 physiological forcing, and CO2 fertilization. The use of absolute values in the calculation allows both negative and positive changes to contribute to the total change.

Additionally, to test the robustness of our results to the selection of climate model and to the role of ocean–atmosphere coupling, we make use of two similar climate model experiments with the Hadley Centre Global Environment Model, version 2, for atmosphere and land (HadGEM2-A; Jones et al. 2011) and the Community Earth System Model, version 1, with biogeochemistry [CESM1(BGC); Lindsay et al. 2014]. First, we compare two 30-yr simulations from HadGEM2-A to estimate the combined impact of projected changes in CO2 fertilization and CO2 physiological forcing on future hydrologic change (Chadwick et al. 2017). The only difference between the two HadGEM2-A simulations is the concentration of atmospheric CO2 seen by the land surface model. In the first simulation, the land model sees prescribed, time-varying CO2 concentrations from the historical period 1979–2008 (335–385 ppm). In the second simulation, the time-varying CO2 concentrations in the land model are quadrupled (1340–1540 ppm). Both simulations utilize the same prescribed historical period sea surface temperatures, and quadrupled time-varying atmospheric CO2 concentrations (1340–1540 ppm). Unlike the CCSM4 simulations, the effects of changing LAI and stomatal conductance are included simultaneously in the HadGEM2-A simulation. Further details of the HadGEM2-A simulations are described in Chadwick et al. (2017).

We also compare our CCSM4 results to those from the coupled atmosphere–ocean CESM1(BGC) simulation, which is archived as part of the carbon–climate feedback experiment within CMIP5 (Arora et al. 2013; Swann et al. 2016). CESM1(BGC) shares many of the same components as CCSM4, including the atmosphere (CAM4) and land (CLM4) models. The primary difference between the CCSM4 simulations and the CESM1(BGC) simulation is the use of a dynamic ocean model in CESM1(BGC). By comparing our CCSM4 simulation with CESM1(BGC) we are able to assess the role of deep ocean circulation in shaping the climate response to CO2 physiological forcing. We analyze a single, transient (140 yr) simulation, denoted CESM1-BGC_esmFixClim1 in the CMIP5 archive, where the CESM1(BGC) land and ocean models see an increase in atmospheric CO2 concentration from 284 to 1140 ppm (increases at a rate of 1% yr−1), but atmospheric CO2 is fixed at 284 ppm in the atmosphere model. Stomatal conductance and LAI/SAI (as well as ocean biogeochemistry) are free to respond to the CO2 increase. Similar to the HadGEM2-A analysis, the experimental design of the CESM1(BGC) simulation does not allow us to quantify the individual contributions of CO2 fertilization and CO2 physiological forcing to the total CO2–vegetation forcing. Instead, we quantify the combined influence of CO2 fertilization and CO2 physiological forcing on future climate. To estimate the role of CO2–vegetation interactions on projected hydroclimate changes in CESM1(BGC), we compare data from the first and final 30 years of the transient simulation.

Finally, we evaluate the simulation of precipitation in CCSM4 by comparing our simulations to the daily resolution Global Precipitation Climatology Project (GPCP), version 1.2 dataset (Huffman et al. 2001), and to 15 coupled atmosphere–ocean climate models from the CMIP5 historical and RCP8.5 experiments (K. Taylor et al. 2012) (see Table S1 in the supplemental material). The daily GPCP observation dataset covers the time period from 1997 through 2014 at a global 1° resolution. We interpolate all CMIP5 model data to a common 1° × 1° grid using a patch recovery method (Gu et al. 2004). All models from CMIP5 include the effects of future radiative forcing from CO2, aerosols, and other greenhouse gases. The majority of CMIP5 models also include the effects of CO2 physiological forcing (Table S1). We assess the statistical significance of changes (at the 95% confidence level) with the use of a permutation test. The permutation test is a nonparametric method of significance testing that makes no assumptions about the underlying distribution of the data.

3. Results

a. Evapotranspiration changes

The latitudinal distribution of historical period terrestrial ET (HistRad_HistStomata_HistLAI) and of the projected terrestrial ET change in CO2_Total is consistent with historical period ET (1980–2005) and projected ET change [(2080–2100) − (1980–2005)] in 15 coupled atmosphere–ocean climate models from CMIP5 (Figs. 1a,b; see also Table S1). Although CCSM4 exhibits relatively high terrestrial ET change in the Southern Hemisphere and relatively low ET change in the Northern Hemisphere, it is important to note that unlike the RCP8.5 simulations, our CCSM4 simulations do not include projected changes in aerosol and non-CO2 greenhouse gas concentrations, which may account for some of the disparity (Fig. 1b) (see section 2). For example, prescribed reductions in aerosol concentrations in the RCP8.5 simulations enhance Northern Hemisphere warming and may lead to greater ET change compared with our fixed aerosol CCSM4 simulations (Cox et al. 2008; Kloster et al. 2010). Overall, the simulated change in terrestrial evapotranspiration in CO2_Total closely matches that from the fully coupled (dynamic ocean) CCSM4 experiment from CMIP5 (2080–2100) − (1980–2005) (Gent et al. 2011), which suggests our idealized slab ocean simulations serve as a good approximation for the fully coupled dynamic ocean CCSM4 model (Fig. S2 in the supplemental material).

Fig. 1.
Fig. 1.

Zonal mean of annual (a) evapotranspiration and (b) evapotranspiration change from CCSM4 and 15 CMIP5 models. Thick black line in (a) denotes the historical period CCSM4 simulation HistRad_HistStomata_HistLAI. Thick black line in (b) denotes the change (CO2_Total) between CCSM4 simulations FutRad_FutStomata_FutLAI and HistRad_HistStomata_HistLAI. Blue lines in (a) denote the historical period average (1980–2005) from the CMIP5 models. Blue lines in (b) denote the change between the future RCP8.5 (2080–2100) and historical simulations from the CMIP5 models. Only land grid points are used in the zonal mean calculations. All models are regridded to a common 1° × 1° grid. CMIP5 models are listed in Table S1.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Over land, projected future CO2 radiative and CO2 physiological forcing drive large but opposing changes in annual mean ET (Fig. 2). Surface warming from CO2 radiative forcing (CO2_Rad) enhances ET across most continental regions (Fig. 2b). In contrast, changes in terrestrial ET from CO2 physiological forcing (CO2_Stomata) are largely negative (Figs. 2c). Within CCSM4, the response of ET to future CO2–vegetation interactions (CO2 physiological forcing + CO2 fertilization) is dominated by projected changes in stomatal conductance (as opposed to changes in LAI) (Figs. 2c,d). Changes in stomatal conductance reduce ET by −0.024 and −0.108 mm day−1 averaged across the globe and land, respectively. In contrast, changes in LAI enhance global and land ET by 0.008 and 0.026 mm day−1, respectively. The overall response of ET to future elevated CO2 (CO2_Total) resembles a balance between the individual responses to radiative forcing (CO2_Rad) and stomatal conductance (CO2_Stomata) (Fig. 2).

Fig. 2.
Fig. 2.

Change in annual mean evapotranspiration from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. Only statistically significant differences are shown.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

The relative contributions of radiative forcing and physiological forcing to the total CO2-driven response of annual mean ET vary considerably at the regional scale (Figs. 3a,b). Throughout much of the densely vegetated tropics (10°S–10°N), CO2-driven changes in stomatal conductance are as important as or more important than CO2 radiative forcing in driving projected future ET change (Figs. 2 and 3). Over central and eastern North America, western and central Europe, and portions of Southeast Asia and the Congo River basin, stomata-driven reductions in ET largely counteract radiative-driven increases in ET, yielding little to no annual mean ET response to future elevated CO2. Over the oceans and sparsely vegetated land, future changes in ET are almost entirely a response to CO2 radiative forcing. To more fully understand the ET responses, we next assess the individual contributions of transpiration and evaporation to ET change.

Fig. 3.
Fig. 3.

Percent contribution to annual mean evapotranspiration change from (a) CO2_Rad, (b) CO2_Stomata, (c) CO2_LAI, and (d) the nonlinear interaction term (see section 2).

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Differences between ET changes from future CO2 radiative and physiological forcing are largely driven by changes in transpiration (Fig. 4). Future radiative forcing and reduced stomatal conductance yield transpiration changes of opposite sign in similar locations across the globe (Figs. 4c,e). In CCSM4, reduced stomatal conductance from elevated CO2 diminishes transpiration in all vegetated areas, particularly in the tropics, Europe, southeastern Asia, the La Plata basin of Argentina, and the eastern United States (Fig. 4e). Warming from CO2 radiative forcing increases the vapor pressure deficit and enhances transpiration in these same regions (Fig. 4c). In contrast, both CO2 radiative forcing and reductions in stomatal conductance increase surface evaporation over most continental areas (with the prominent exception of radiative-driven evaporation changes in Central America and northern South America) (Figs. 4d,f). Evaporation increases with reduced stomatal conductance in part because diminished transpiration rates leave larger quantities of moisture available for evaporation from the soil (Figs. S4a–d in the supplemental material) (Swann et al. 2016).

Fig. 4.
Fig. 4.

Change in annual mean (left) transpiration and (right) evaporation from (a),(b) CO2_Total, (c),(d) CO2_Rad, (e),(f) CO2_Stomata, and (g),(h) CO2_LAI. Only statistically significant differences are shown. Evaporation over the ocean is masked out.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Although stomatal conductance-driven changes in transpiration and evaporation are often found in the same locations (Figs. 4e,f), the overall response of ET to projected changes in stomatal conductance is dominated by the larger magnitude reductions in transpiration. In contrast, LAI-driven changes in evaporation and transpiration counteract one another, minimizing the impact of changes in CO2 fertilization on future ET in our simulations (Figs. 2d, 3e, and 4g,h). Specifically, increases in LAI enhance transpiration at the expense of soil moisture, limiting evaporation. Additionally, greater vegetation coverage can limit evaporation by shading the soil from sunlight. Overall, our results indicate that future reductions in stomatal conductance contribute substantially to projected regional-scale changes in ET and that the combined influences of radiative forcing and physiological forcing dominate the future ET response to CO2.

b. Precipitation changes

The CCSM4 historical control simulation (HistRad_HistStomata_HistLAI) captures the observed global spatial pattern of annual mean precipitation well (Figs. 5a,b). Additionally, the latitudinal distribution of terrestrial precipitation resembles that of the 15 CMIP5 models (Fig. 5c). Precipitation biases in the CCSM4 historical control simulation are consistent with those found in the twentieth-century climate simulation from the fully coupled dynamic ocean CCSM4 model (Gent et al. 2011). Compared with GPCP data, CCSM4 overestimates precipitation along the intertropical convergence zones (ITCZs) of the Indian and Pacific Oceans and in topographically complex regions including the Himalayas, the Andes, and Rwenzori Mountains of eastern equatorial Africa (Figs. 5a,b). Gent et al. (2011) note that CCSM4 exhibits stronger than observed amplitude in El Niño–Southern Oscillation (ENSO) variability, and weaker than observed Madden–Julian oscillation (MJO) variability, both of which impact precipitation, particularly in the tropics. A detailed evaluation of twentieth-century climate in CCSM4 can be found in Gent et al. (2011). Terrestrial precipitation change in CO2_Total is similar, although slightly greater in magnitude than that in the fully coupled (dynamic ocean) CCSM4 experiment from CMIP5 (Fig. S3 in the supplemental material). Similar to ET change, the simulated precipitation change in the Southern Hemisphere in CO2_Total is relatively high compared with the CMIP5 climate change simulations (RCP8.5 − historical) (Fig. 5d).

Fig. 5.
Fig. 5.

Annual mean precipitation from (a) GPCP (1997–2014) and (b) the 30-yr historical period CCSM4 simulation HistRad_HistStomata_HistLAI. Zonal mean of annual (c) precipitation and (d) precipitation change from CCSM4 and 15 CMIP5 models. Thick black line in (c) denotes the historical period CCSM4 simulation HistRad_HistStomata_HistLAI and in (d) denotes the change (CO2_Total) between CCSM4 simulation FutRad_FutStomata_FutLAI and simulation HistRad_HistStomata_HistLAI. Blue lines in (c) denote the historical period average (1980–2005) from the CMIP5 models and in (d) denote the change between the future RCP8.5 (2080–2100) and historical simulations from the CMIP5 models. Only land grid points are used in zonal mean calculations for (c) and (d). All models are regrid to a common 1° × 1° grid. CMIP5 models are listed in Table S1.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

CO2 radiative and CO2 physiological forcing drive distinct patterns in regional precipitation change (Figs. 6b,c). Projected changes in CO2 radiative forcing impact precipitation across most of the globe (Figs. 6b and 7a). Enhanced radiative forcing brings increased precipitation to the high latitudes and large portions of the midlatitudes, including much of Asia, Australia, North America, and South America. Rainfall decreases over portions of the subtropics, as well as northern South America. Future CO2 radiative forcing drives particularly large precipitation increases in the tropics, including over the Maritime Continent, eastern equatorial Africa, and the Indian and Pacific Oceans. These tropical precipitation increases are in part driven by thermodynamic increases in atmospheric moisture convergence (the rich-get-richer hypothesis) (Held and Soden 2006) and in part by shifts in convergence zones due to sea surface temperature (SST) pattern changes and changes in regional-scale atmospheric dynamics (e.g., changes in regional Walker circulations) (Chadwick et al. 2013; Kent et al. 2015).

Fig. 6.
Fig. 6.

Change in annual mean precipitation from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. Only statistically significant differences are shown.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Fig. 7.
Fig. 7.

Percent contribution to annual mean precipitation change from (a) CO2_Rad, (b) CO2_Stomata, (c) CO2_LAI and (d) the nonlinear interaction term (see section 2).

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Similar to changes in ET, the response of precipitation to CO2–vegetation interactions is dominated by changes in stomatal conductance (Figs. 6c,d and 7b,c). The largest changes in precipitation from stomatal conductance are located in the tropics and in portions of the midlatitude Northern Hemisphere, where ET changes are of the greatest magnitude (Figs. 2c, 6c, and 7b). With the exception of the eastern Amazon and far northern South America, changes in stomatal conductance enhance mean annual rainfall over tropical land. Precipitation increases throughout central Africa, the Maritime Continent, southeastern Asia, the equatorial Andes, and the western Amazon. In these areas, enhanced precipitation combined with greater soil moisture from reduced transpiration leads to an increase in surface runoff (Fig. S4) (Gedney et al. 2006; Cao et al. 2010). Over the tropical oceans, annual mean rainfall decreases, particularly in the equatorial Atlantic and eastern Pacific. Drying in the Northern Hemisphere midlatitudes is primarily concentrated in the boreal summer season when vegetation LAI, and hence transpiration, is at its climatological peak (Fig. 8). Projected changes in LAI have little impact on future annual mean precipitation within CCSM4 (Figs. 6d and 7c). Overall, although CO2 radiative forcing is the primary driver of future precipitation change across much of the globe, CO2 physiological forcing is critical in shaping the processes that control future precipitation in regions of high ET such as the tropics and the summer season midlatitudes.

Fig. 8.
Fig. 8.

Change in seasonal mean precipitation from CO2_Stomata: (a) December–February, (b) March–May, (c) June–August, and (d) September–November.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Over most terrestrial regions, the individual precipitation responses to CO2 radiative forcing, CO2 physiological forcing, and CO2 fertilization add nearly linearly to give the total precipitation response to future CO2 (Fig. 7d and Fig. S5 in the supplemental material). This is consistent with the results of Peng et al. (2014), who found minimal interactive effects between CO2 radiative and physiological forcing within a subset of CMIP5 models. However, there are some regions where interactions between the different forcings in the CO2_Total simulation result in nonlinearities. In CCSM4, the largest magnitude deviations on land occur over the highlands of eastern Africa and in portions of southeastern Asia (Fig. S5). Because these regions are topographically complex and located within the humid tropics, even small changes in the location and magnitude of moisture convergence can drive large changes in precipitation, making the linear decomposition approximation less appropriate in these areas.

In addition to changes in annual mean precipitation, elevated CO2 intensifies the hydrologic cycle, increasing the frequency of dry days and heavy precipitation days in many regions (Sun et al. 2007). Across most of the globe, CO2 radiative forcing is the primary driver of projected changes in daily precipitation characteristics (Figs. 9 and 10). Future radiative forcing increases annual dry day occurrence (<0.1 mm) over northern South America, portions of the terrestrial Northern Hemisphere subtropics (~20°–40°N), and along the northern margin of the SH extratropical storm track (~50°S) (Fig. 9b). Dry days become less common in the Sahel, the subtropical Southern Hemisphere, and the high latitudes. Radiative forcing increases the frequency of heavy rainfall days (>95th percentile event from the historical control simulation using only days with greater than 0.1 mm of precipitation) across broad stretches of the globe, including the mid- and high latitudes, tropical Pacific, tropical Africa, and Maritime Continent, and decreases heavy rainfall days over Central America and the tropical and subtropical Atlantic (Fig. 10b).

Fig. 9.
Fig. 9.

Change in annual number of days without precipitation from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. A day is considered to have no precipitation when the accumulated daily precipitation is less than 0.1 mm. Only statistically significant differences are shown.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Fig. 10.
Fig. 10.

Change in the annual number of days with heavy precipitation from (a) CO2_Total, (b) CO2_Rad, (c) CO2_Stomata, and (d) CO2_LAI. Heavy rainfall days are those that exceed the 95th percentile event from the historical control simulation (year 2000 conditions). Only statistically significant differences are shown.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Despite its dominant influence on global precipitation change, radiative forcing alone does not fully account for the projected regional-scale changes in future precipitation characteristics in CO2_Total (Figs. 9 and 10). Within CCSM4, CO2 physiological forcing plays an important role in shaping the geographic distribution of projected daily-scale precipitation changes (Figs. 9 and 10). Future increases in the annual number of dry days in the Northern Hemisphere midlatitudes, including Europe and North America, as well as in the Amazon, Andes, and central Africa are largely a response to changes in stomatal conductance (Fig. 9c). These regions experience upward of 15 more dry days per year when future CO2-driven changes in stomatal conductance are included in CCSM4. Increases in the annual number of dry days are similar, although slightly more widespread and robust, when using a dry day threshold of 0.01 mm (as opposed to 0.1 mm), a value closer to the dry day threshold of 0.024 mm used in Lau et al. (2013) (Fig. S6 in the supplemental material). Interestingly, in many regions, the increase in dry day occurrence comes at the expense of very light rainfall or drizzle days (0.1–2 mm) (Fig. 11a), suggesting that decreases in stomatal conductance reduce the likelihood of rainfall on days when atmospheric conditions are only marginally favorable for precipitation. The reductions in light rainfall days are consistent with decreases in low cloud fraction (Fig. 11b). In the midlatitudes, the increase in dry days exceeds the reduction in light precipitation days, suggesting that some dry day increases also come at the expense of heavier rainfall events. Although limited in geographic extent, changes in LAI also impact precipitation characteristics in CCSM4, reducing the number of dry days in parts of East Africa, the Sahel, northern Europe, and Southeast Asia (Fig. 9d). However, the geographic pattern of projected LAI change (Fig. S1) does not closely resemble the pattern of LAI-driven dry day change across the globe (Fig. 9d), highlighting the limited direct contribution of LAI changes in driving future dry day frequency in CCSM4.

Fig. 11.
Fig. 11.

Change in the (a) annual number of days with light precipitation and (b) annual mean low-cloud fraction from CO2_Stomata. A day is considered to have light precipitation when the accumulated daily precipitation is greater than 0.1 mm and less than 2 mm. Only statistically significant differences are shown in (a). Brown shading indicates fewer days with light precipitation in (a).

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

In the tropics, physiological forcing also contributes to future changes in the frequency of heavy precipitation events (Fig. 10c). The CO2-driven reductions in stomatal conductance drive mean annual increases in excess of eight heavy rainfall days per year in large parts of the Andes, western Amazon, central Africa, and the Maritime Continent. In the eastern Amazon and far northeastern South America, projected changes in stomatal conductance are instrumental in reducing the number of future heavy rainfall days.

In a similar set of climate model experiments with HadGEM2-ES (see section 2), the projected combined impact of future CO2 physiological forcing and CO2 fertilization yields similar changes in daily precipitation characteristics to our CO2_Stomta experiment (Figs. 12a,c). Specifically, the number of dry days increases over tropical land and the continental Northern Hemisphere midlatitudes (although dry day changes in North America are located north and west of those in CCSM4) (Fig. 12a). Additionally, as in CCSM4, the annual frequency of heavy rainfall events decreases over the eastern Amazon, and increases over the rest of the world’s tropical forests (Fig. 12c). Similarly, in a coupled atmosphere–ocean simulation with CESM1(BGC), changes in daily precipitation characteristics from combined CO2 physiological forcing and CO2 fertilization (see section 2) resemble those of CCSM4, suggesting that the use of a slab ocean model in our CCSM4 simulations does not significantly impact the results (Figs. 12b,d).

Fig. 12.
Fig. 12.

Change in the annual number of (a),(b) dry days and (c),(d) days with heavy precipitation from CO2_Stomata simulations with (left) HadGEM2-A and (right) CESM1-BGC_esmFixClim1 (see section 2). A day is considered to have no precipitation when the accumulated daily precipitation is less than 0.1 mm. Heavy rainfall days are those that exceed the 95th percentile event from the historical period (first 30 yr). Only statistically significant differences are shown.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

c. Temperature, moisture, instability, and circulation changes

Previous theoretical and climate modeling work has generated considerable insights into the mechanisms that shape the response of precipitation characteristics to CO2 radiative forcing (Trenberth et al. 2003; O’Gorman and Schneider 2009). Here, we seek to provide a physical explanation for changes in precipitation characteristics associated with future CO2–vegetation interactions. We focus primarily on the climate system response to future CO2 physiological forcing, as future CO2 fertilization does not significantly contribute to future hydrologic changes in our CCSM4 experiments (Figs. 3 and 7).

Changes in annual mean precipitation and daily precipitation characteristics from future physiological forcing are located in or adjacent to regions of projected large ET reduction (cf. Figs. 2, 6, 9 and 10). Diminished ET arising from reduced stomatal conductance decreases the latent heat flux from the surface to the atmosphere warming the global and terrestrial surface by 0.26° and 0.56°C, respectively (Figs. 13a,b). Warming on land occurs everywhere except in desert regions. Reductions in ET (Fig. 2) and surface latent heat flux (Fig. 13a) drive the greatest warming over tropical forests, the La Plata basin, southeastern Asia, and the continental Northern Hemisphere midlatitudes (Fig. 13b). Reductions in low cloud fraction (Fig. 11b), which are driven by increased atmospheric temperatures and reduced boundary layer relative humidity (Fig. S7 in the supplemental material), likely amplify surface warming in these regions. Warming also peaks in the high latitudes east of Greenland, far from any terrestrial ET and surface latent heat flux change, and is driven in part by increased low cloud cover during the winter season (Fig. 11b) (Pu and Dickinson 2014).

Fig. 13.
Fig. 13.

Change in annual mean (a) surface latent heat flux, (b) surface temperature, (c) vertically integrated atmospheric moisture and moisture flux, and (d) CAPE from CO2_Stomata.

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

Despite reductions in surface latent heat flux, and hence a source of moisture to the atmosphere, total vertically integrated precipitable water increases considerably over tropical land (excluding eastern South America and the western Sahel), and along eastern Asia (Fig. 13c). As noted by Pu and Dickinson (2014), changes in stomatal conductance drive a negative surface moisture feedback, whereby reduced ET warms the land surface and draws in moisture from nearby oceans. This reallocation of moisture is most prominent over western South America, central Africa, the Maritime Continent, and eastern Asia (Fig. 13c). In these regions, warmer temperatures and enhanced moisture convergence lead to increased convective available potential energy (CAPE) and a greater likelihood of heavy precipitation events (Figs. 10c and 13d). The large reduction in precipitation and heavy precipitation days over the eastern Amazon is unique within tropical forest regions. Anomalous westerly moisture flux directed toward the Amazon is partially blocked by the Andes, leading to substantial increases in annual precipitation and heavy precipitation events in the present day semiarid areas of western Ecuador and Peru, at the expense of the Amazon (Figs. 6c, 10c, and 13c). Further, greater Northern Hemisphere warming likely draws the ITCZ and associated moisture northward and away from the Amazon (although the northward ITCZ shift in the Atlantic is weaker than that in the Pacific) (Figs. 8 and 13c), and moisture export from the eastern Amazon toward the Andes and La Plata basin is enhanced (Fig. 13c). This reduction in eastern Amazon precipitation is a consistent response to CO2 physiological forcing across climate models (Betts et al. 2004; Swann et al. 2016). In the mid- and high latitudes, moisture convergence does not balance the reduced source of ET to the atmosphere, and mean precipitation, as well as the number of heavy rainfall events, remains constant or decreases slightly (Figs. 6c, 10c, and 13c).

Future physiological forcing increases the frequency of dry days in months when the local or upwind land surface acts as an important source of moisture for rainfall. In the tropics, dry days increase most when the ITCZ and associated moisture convergence are located outside the region of interest. For example, in the northern Congo and eastern Amazon, increases in dry days are concentrated during the dry season, and in particular during the transition from the dry season to the wet season (Figs. 14a,b). As noted by Lee et al. (2012), this is a time when atmospheric conditions are marginal for the development of rainfall and regional terrestrial ET may have an important role in supplying moisture to initiate the rainy season (Fu and Li 2004). Regions of the tropics that exhibit little to no increase in dry days, for instance the Maritime Continent and parts of Southeast Asia, rely less on terrestrial ET for precipitation (van der Ent et al. 2014). Instead, nearby oceanic regions supply nearly all of the moisture that falls as precipitation in these areas.

Fig. 14.
Fig. 14.

Box-and-whisker plots represent the number of dry days (<0.1 mm) for each month. Lines represent the average daily precipitation. Red (blue) coloring indicates data from a simulation with (without) future changes in stomatal conductance from CO2. Box-and-whisker plots show min, max, and interquartile range, and consist of 30 yr of data. Average daily precipitation is averaged over 30 years. Area weighted averages from (a) eastern Amazon (10°S–0°, 65°–50°W), (b) northern Congo (0°–10°N, 15°–30°E), (c) Europe (40°–55°N, 0°–15°E), and (d) midlatitude South America (30°–40°S, 65°–55°W).

Citation: Journal of Climate 30, 7; 10.1175/JCLI-D-16-0603.1

In the midlatitudes, increases in dry day frequency from physiological forcing are greatest during times of the year when large-scale circulation is less conducive for rainfall and the frequency of dry days is already high. For example, in Europe and the South American midlatitudes, dry days increase most during the summer months, when large-scale moisture advection is relatively low, and transpiration and precipitation recycling ratios are relatively high (Schär et al. 1999; Trenberth et al. 2003) (Figs. 14c,d). Overall, the diminished moisture source (Fig. 2c), as well as warmer temperatures (Fig. 13b) and associated increases in saturation vapor pressure, increase the likelihood of rain-free days when atmospheric conditions are only marginally favorable for precipitation. These results are consistent with a reduction in light precipitation days in many regions of diminished ET (Fig. 11a).

4. Discussion

a. Summary and implications

Previous idealized climate model sensitivity studies have shown that changes in transpiration flux can impact the frequency and intensity of rainfall in vegetated regions (Lawrence et al. 2007; Lee et al. 2012). The imposed transpiration changes in these experiments were large [in the case of Lee et al. (2012), transpiration flux was completely removed] and not intended to reflect realistic changes in response to past or future climate forcings. Here, we demonstrate that transpiration changes from projected CO2 physiological forcing are of sufficient magnitude to influence daily and annual precipitation characteristics and drive large regional-scale impacts in climate model simulations of the twenty-first century. Although warming from radiative forcing is the primary driver of future hydrologic changes at the global scale, regionally, physiological forcing can dominate the hydrologic response to changes in CO2.

The identification of CO2 physiological forcing as an important driver of daily-scale hydrologic change helps to more fully understand geographic and seasonal patterns of future projected changes in precipitation characteristics. Hydrologic changes from physiological forcing are primarily constrained to regions and times of the year with large transpiration fluxes, such as the tropics and the summer season Northern Hemisphere midlatitudes (Figs. 2, 4, 6, and 810). In many densely vegetated tropical regions, specifically those that exhibit high transpiration but rely on oceanic sources of moisture for a large proportion of rainfall, future CO2 physiological forcing enhances the likelihood of heavy rainfall events through a negative surface moisture feedback, whereby diminished transpiration from reduced stomatal conductance warms the surface and enhances moisture convergence from surrounding regions (Figs. 10 and 13). In these areas, annual ET decreases while annual precipitation increases. By definition, the local precipitation recycling ratio goes down, and these regions become more reliant on external sources of moisture for rainfall.

Further, in the tropics, reduced stomatal conductance and lower transpiration increases the likelihood of dry days during the transition between the dry and wet seasons, when moisture fluxes from the land surface are particularly important for initiating rainfall (Figs. 9 and 14) (Fu and Li 2004). Likewise, in vegetated areas of the midlatitudes, reduced stomatal conductance increases the frequency of dry days, particularly during the summer season when transpiration and precipitation recycling are relatively high (Figs. 9 and 14). Interestingly, the number of summer season dry days increases in midlatitude Europe, South America, and portions of North America despite increases in summer soil moisture (Fig. S4). The increase in soil moisture reduces soil moisture stress on vegetation, promotes stomatal conductance, and serves as a negative feedback on the original reduction in transpiration from physiological forcing. However, the combined influences of enhanced stomatal conductance from low soil moisture stress and the moderate increase in soil evaporation from excess soil moisture (Fig. 4f) do not balance the much larger reduction in transpiration from the CO2-induced limitation of stomatal conductance, leading to low relative humidity and an increase in dry days (Fig. 9; see also Fig. S7). In some midlatitude regions, including the central plains of the United States and southeastern Canada, drier soils from future physiological forcing (from temperature, cloud, and rainfall feedbacks) amplify the reduction in stomatal conductance and reduce soil evaporation, leading to more frequent dry day occurrence. Importantly, the changes in precipitation characteristics from future CO2–vegetation interactions identified in this study are robust across several different climate model configurations (e.g., simulations with slab and dynamic ocean models, simulations where CO2 physiological forcing and CO2 fertilization are simulated separately and in tandem, and simulations with different CO2 concentrations) (Figs. 9, 10, and 12).

Our results highlight the influential role of transpiration in shaping daily hydrologic characteristics, particularly in the tropics. Lee et al. (2012) propose that transpiration serves to reduce the variability of daily precipitation throughout tropical forests. In contrast, our findings suggest that the role of transpiration is not the same across all tropical forest regions. For example, in the eastern Amazon, reduced transpiration increases the annual number of dry days, but it also decreases the annual number of heavy rainfall days (Figs. 9 and 10). We find that the role of transpiration in shaping daily precipitation depends on a number of factors, including regional geography and time of year. In addition to its impact on precipitation, transpiration also serves as an important control on surface runoff. Specifically, reduced transpiration from future CO2 physiological forcing increases surface runoff, particularly in the tropics (Fig. S4) (Betts et al. 2007; Gedney et al. 2006; Cao et al. 2010). Our results suggest that enhanced runoff from CO2 physiological forcing may also be a consequence of more frequent heavy rainfall events.

b. Uncertainties, caveats, and future work

Although our results point to the important role of CO2 physiological forcing in shaping projected future precipitation changes in climate model simulations, our findings should be interpreted cautiously. Many uncertainties regarding the interactions among vegetation, carbon dioxide, and hydrology remain (Wang et al. 2009). Most climate models, including those employed in this study, assume that plant water use strategies depend only on the plant’s photosynthetic pathway (classified as C3 or C4) (Kala et al. 2016). However, observations indicate that plants within C3 and C4 groupings exhibit large species-specific differences in stomatal behavior (Lin et al. 2015). Kala et al. (2016) find substantial differences in future climate projections between climate model simulations that use PFT-level water use strategy parameterizations and simulations that use traditional photosynthetic pathway-level parameterizations. Specifically, they find that the implementation of PFT-level water use strategies leads to a greater reduction in ET over Europe in response to future elevated CO2, suggesting that our simulations may underestimate the regional hydrological impact of future CO2 physiological forcing there (Kala et al. 2016). Further, our model experiments do not include the impacts of enhanced CO2 concentrations on down-regulation of photosynthesis, which Pu and Dickinson (2014) have shown can substantially reduce stomatal conductance and transpiration within CCSM4.

Earth system models project widespread increases in LAI under elevated greenhouse gas concentrations (Mahowald et al. 2016). However, by the end of the twenty-first century, global mean LAI changes vary by a factor of about 14 across models under the RCP8.5 scenario (Mahowald et al. 2016). Relationships between projected changes in LAI and changes in vegetation carbon or climate variables are not straightforward and the impact of projected LAI differences on the simulated spread in twenty-first-century precipitation projections across models is unclear. Although the hydroclimate impacts from projected LAI in our CCSM4 simulation are minor, future work should assess the contribution of LAI to hydrologic change within the full suite of CMIP climate models. Additionally, our simulations do not include dynamic vegetation models, and therefore do not capture the potential impacts of future changes in vegetation type or location from climate change or land use change on ET and precipitation, which may be large (Notaro et al. 2007; Pearson et al. 2013; Bathiany et al. 2014).

Last, it is important to note that climate models struggle to properly simulate the full distribution of observed precipitation rates (Lin et al. 2006; O’Gorman and Schneider 2009). In particular, models tend to overestimate the annual number of days with very light precipitation (Lin et al. 2006). A proportion of the dry day (with less than 0.1 mm of precipitation) increases found in our simulation come at the expense of days with very light rainfall (days with 0.1–2 mm of precipitation). Although we provide a mechanistic explanation for the changes in rainfall distribution in CCSM4 [and find consistent results in HadGEM2-A and CESM1(BGC)], the robustness of our results should be tested with other models, particularly those with higher-resolution, resolved convection, and different land surface schemes.

5. Conclusions

Theoretical and climate modeling studies project a robust change in global precipitation characteristics in response to future increases in CO2 (Trenberth et al. 2003; O’Gorman and Schneider 2009; Lau et al. 2013). Given the potential for acute impacts on human and natural systems, understanding the physical mechanisms that drive these hydrologic changes is critical. To date, most studies have focused on the role of radiative forcing in driving these hydrologic changes. Our modeling results suggest that consideration of the radiative impacts of CO2 alone cannot account for projected regional-scale differences in daily precipitation changes. Within our simulations with CCSM4, we find that projected CO2 physiological forcing substantially reduces future transpiration in vegetated areas such as the midlatitudes during the summer season and the tropics. The reductions in transpiration are of sufficient magnitude to drive changes in the annual frequency of dry days and heavy rainfall days in these regions. The physiologically driven changes have the potential to reinforce or greatly diminish the influence of radiative forcing on future precipitation characteristics. These results suggest that the representation of plant water use and its relationship to atmospheric CO2 within climate models likely contributes to the persistent spread in future regional-scale precipitation projections. Given the potential for future CO2 physiological forcing to shape projected changes in precipitation characteristics, future work should aim to improve the representation of the complex interactions among CO2, vegetation, and hydrology within climate models in order to refine estimates of hydrology and daily dry and wet extremes.

Acknowledgments

This work was supported by the Turner Postdoctoral Fellowship awarded to Christopher Skinner, and by National Science Foundation Award 1602956. Robin Chadwick was supported by the Newton Fund through the Met Office Climate Science for Service Partnership Brazil (CSSP Brazil). We wish to thank Hervé Douville, Justin Mankin, Bing Pu, Deepti Singh, and Abigail Swann for helpful discussions and three anonymous reviewers whose feedback greatly improved the manuscript. The CESM project is supported by the National Science Foundation and the Office of Science (BER) of the U.S. Department of Energy. We acknowledge the World Climate Research Programme and the climate modeling groups for making available their model output, and the U.S. Department of Energy’s Program for Climate Model Diagnosis and Intercomparison for coordinating and supporting database development. We thank Stanford University and the Stanford Research Computing Center for providing computational resources and support that have contributed to these research results. Model data are available upon request.

REFERENCES

  • Anderegg, W. R. L., J. M. Kane, and L. D. L. Anderegg, 2013: Consequences of widespread tree mortality triggered by drought and temperature stress. Nat. Climate Change, 3, 3036, doi:10.1038/nclimate1635.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Andrews, T., M. Doutriaux-Boucher, O. Boucher, and P. M. Forster, 2011: A regional and global analysis of carbon dioxide physiological forcing and its impact on climate. Climate Dyn., 36, 783792, doi:10.1007/s00382-010-0742-1.

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

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bathiany, S., M. Claussen, and V. Brovkin, 2014: CO2-induced Sahel greening in three CMIP5 Earth system models. J. Climate, 27, 71637184, doi:10.1175/JCLI-D-13-00528.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, R. A., P. M. Cox, S. E. Lee, and F. I. Woodward, 1997: Contrasting physiological and structural vegetation feedbacks in climate change simulations. Nature, 387, 796799, doi:10.1038/42924.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, R. A., P. M. Cox, M. Collins, P. P. Harris, C. Huntingford, and C. D. Jones, 2004: The role of ecosystem–atmosphere interactions in simulated Amazonian precipitation decrease and forest dieback under global climate warming. Theor. Appl. Climatol., 78, 157175, doi:10.1007/s00704-004-0050-y.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Betts, R. A., and Coauthors, 2007: Projected increase in continental runoff due to plant responses to increasing carbon dioxide. Nature, 448, 10371041, doi:10.1038/nature06045.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Bitz, C. M., and Coauthors, 2012: Climate sensitivity of the Community Climate System Model, version 4. J. Climate, 25, 30533070, doi:10.1175/JCLI-D-11-00290.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Boucher, O., A. Jones, and R. A. Betts, 2009: Climate response to the physiological impact of carbon dioxide on plants in the Met Office Unified Model HadCM3. Climate Dyn., 32, 237249, doi:10.1007/s00382-008-0459-6.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cao, L., G. Bala, K. Caldeira, R. Nemani, and G. Ban-Weiss, 2010: Importance of carbon dioxide physiological forcing to future climate change. Proc. Natl. Acad. Sci. USA, 107, 95139518, 10.1073/pnas.0913000107.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chadwick, R., I. Boutle, and G. Martin, 2013: Spatial patterns of precipitation change in CMIP5: Why the rich do not get richer in the tropics. J. Climate, 26, 38033822, doi:10.1175/JCLI-D-12-00543.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Chadwick, R., H. Douville, and C. B. Skinner, 2017: Timeslice experiments for understanding regional climate projections: Applications to the tropical hydrological cycle and European winter circulation. Climate Dyn., doi:10.1007/s00382-016-3488-6, in press.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cowling, S. A., and C. B. Field, 2003: Environmental control of leaf area production: Implications for vegetation and land-surface modeling. Global Biogeochemical Cycles, 17, 1007, doi:10.1029/2002GB001915.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Cox, P. M., and Coauthors, 2008: Increasing risk of Amazonian drought due to decreasing aerosol pollution. Nature, 453, 212215, doi:10.1038/nature06960.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Douville, H., S. Planton, J.-F. Royer, D. B. Stephenson, S. Tyteca, L. Kergoat, S. Lafont, and R. A. Betts, 2000: Importance of vegetation feedbacks in doubled-CO2 climate experiments. J. Geophys. Res., 105, 14 84114 861, doi:10.1029/1999JD901086.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Eltahir, E. A. B., 1998: A soil moisture–rainfall feedback mechanism: 1. Theory and observations. Water Resour. Res., 34, 765776, doi:10.1029/97WR03499.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Field, C. B., R. B. Jackson, and H. A. Mooney, 1995: Stomatal responses to increased CO2: Implications from the plant to the global scale. Plant Cell Environ., 18, 12141225, doi:10.1111/j.1365-3040.1995.tb00630.x.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Field, C. B., and Coauthors, 2012: Managing the Risks of Extreme Events and Disasters to Advance Climate Change Adaptation. Cambridge University Press, 582 pp.

  • Frank, D., and Coauthors, 2015: Effects of climate extremes on the terrestrial carbon cycle: Concepts, processes and potential future impacts. Global Change Biol., 21, 28612880, doi:10.1111/gcb.12916.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Fu, R., and W. Li, 2004: The influence of the land surface on the transition from dry to wet season in Amazonia. Theor. Appl. Climatol., 78, 97110, doi:10.1007/s00704-004-0046-7.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gedney, N., P. M. Cox, R. A. Betts, O. Boucher, C. Huntingford, and P. A. Stott, 2006: Detection of a direct carbon dioxide effect in continental river runoff records. Nature, 439, 835838, doi:10.1038/nature04504.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gent, P. R., and Coauthors, 2011: The Community Climate System Model version 4. J. Climate, 24, 49734991, doi:10.1175/2011JCLI4083.1.

  • Good, S. P., D. Noone, and G. Bowen, 2015: Hydrologic connectivity constrains partitioning of global terrestrial water fluxes. Science, 349, 175177, doi:10.1126/science.aaa5931.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gu, H., Z. Zong, and K. C. Hung, 2004: A modified superconvergent patch recovery method and its application to large deformation problems. Finite Elem. Anal. Des., 40, 665687, doi:10.1016/S0168-874X(03)00109-4.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Held, I. M., and B. J. Soden, 2006: Robust responses of the hydrological cycle to global warming. J. Climate, 19, 56865699, doi:10.1175/JCLI3990.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hirabayashi, Y., and Coauthors, 2013: Global flood risk under climate change. Nat. Climate Change, 3, 816821, doi:10.1038/nclimate1911.

  • Huffman, G. J., R. Adler, M. Morrissey, D. Bolvin, S. Curtis, R. Joyce, B. McGavock, and J. Susskind, 2001: Global precipitation at one-degree daily resolution from multisatellite observations. J. Hydrometeor., 2, 3650, doi:10.1175/1525-7541(2001)002<0036:GPAODD>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Hunke, E. C., and W. H. Lipscomb, 2008: CICE: The Los Alamos sea ice model: Documentation and software user’s manual. Tech. Rep. LA-CC-06-012, v.4.0, 72 pp.

  • Jones, C. D., and Coauthors, 2011: The HadGEM2-ES implementation of CMIP5 centennial simulations. Geosci. Model Dev., 4, 543570, doi:10.5194/gmd-4-543-2011.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kala, J., M. G. De Kauwe, A. J. Pitman, B. E. Medlyn, Y. P. Wang, R. Lorenz, and S. E. Perkins-Kirkpatrick, 2016: Impact of the representation of stomatal conductance on model projections of heatwave intensity. Sci. Rep., 6, 23418, doi:10.1038/srep23418.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kent, C., R. Chadwick, and D. P. Rowell, 2015: Understanding uncertainties in future projections of seasonal tropical precipitation. J. Climate, 28, 43904413, doi:10.1175/JCLI-D-14-00613.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kergoat, L., S. Lafont, H. Douville, B. Berthelot, G. Dedieu, S. Planton, and J.-F. Royer, 2002: Impact of doubled CO2 on global-scale leaf area index and evapotranspiration: Conflicting stomatal conductance and LAI responses. J. Geophys. Res., 107, 4808, doi:10.1029/2001JD001245.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Kloster, S., F. Dentener, J. Feichter, F. Raes, U. Lohmann, E. Roeckner, and I. Fischer-Bruns, 2010: A GCM study of future climate response to aerosol pollution reductions. Climate Dyn., 34, 11771194, doi:10.1007/s00382-009-0573-0.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lammertsma, E. I., H. J. de Boer, S. C. Dekker, D. L. Dilcher, A. F. Lotter, and F. Wagner-Cremer, 2011: Global CO2 rise leads to reduced maximum stomatal conductance in Florida vegetation. Proc. Natl. Acad. Sci. USA, 108, 40354040, doi:10.1073/pnas.1100371108.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lau, W. K.-M., H.-T. Wu, and K.-M. Kim, 2013: A canonical response of precipitation characteristics to global warming from CMIP5 models. Geophys. Res. Lett., 40, 31633169, doi:10.1002/grl.50420.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lawrence, D. M., P. E. Thornton, K. W. Oleson, and G. B. Bonan, 2007: The partitioning of evapotranspiration into transpiration, soil evaporation, and canopy evaporation in a GCM: Impacts on land–atmosphere interaction. J. Hydrometeor., 8, 862880, doi:10.1175/JHM596.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lee, J.-E., and Coauthors, 2012: Reduction of tropical land region precipitation variability via transpiration. Geophys. Res. Lett., 39, L19704, doi:10.1029/2012GL053417.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Levis, S., J. A. Foley, and D. Pollard, 2000: Large-scale vegetation feedbacks on a doubled CO2 climate. J. Climate, 13, 13131325, doi:10.1175/1520-0442(2000)013<1313:LSVFOA>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, J.-L., and Coauthors, 2006: Tropical intraseasonal variability in 14 IPCC AR4 climate models. Part I: Convective signals. J. Climate, 19, 26652690, doi:10.1175/JCLI3735.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lin, Y.-S., and Coauthors, 2015: Optimal stomatal behaviour around the world. Nat. Climate Change, 5, 459464, doi:10.1038/nclimate2550.

  • Lindsay, K., and Coauthors, 2014: Preindustrial-control and twentieth-century carbon cycle experiments with the Earth System Model CESM1(BGC). J. Climate, 27, 89819005, doi:10.1175/JCLI-D-12-00565.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Lloyd, J., and G. D. Farquhar, 2008: Effects of rising temperatures and [CO2] on the physiology of tropical forest trees. Philos. Trans. Roy. Soc. London, 363B, 18111817, doi:10.1098/rstb.2007.0032.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Mahowald, N., F. Lo, Y. Zheng, L. Harrison, C. Funk, D. Lombardozzi, and C. Goodale, 2016: Projections of leaf area index in Earth system models. Earth Syst. Dynam., 7, 211229, doi:10.5194/esd-7-211-2016.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Malhi, Y., J. T. Roberts, R. A. Betts, T. J. Killeen, W. Li, and C. A. Nobre, 2008: Climate change, deforestation, and the fate of the Amazon. Science, 319, 169172, doi:10.1126/science.1146961.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Milly, P. C. D., and K. A. Dunne, 2016: Potential evapotranspiration and continental drying. Nat. Climate Change, 6, 946949, doi:10.1038/nclimate3046.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Neale, R. B., J. H. Richter, A. J. Conley, S. Park, P. H. Lauritzen, A. Gettleman, and D. L. Williamson, 2010: Description of the NCAR Community Atmosphere Model (CAM5.0). NCAR Tech. Rep. NCAR/TN-486+STR, 274 pp.

  • Notaro, M., S. Vavrus, and Z. Liu, 2007: Global vegetation and climate change due to future increases in CO2 as projected by a fully coupled model with dynamic vegetation. J. Climate, 20, 7090, doi:10.1175/JCLI3989.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • O’Gorman, P. A., and T. Schneider, 2009: The physical basis for increases in precipitation extremes in simulations of 21st-century climate change. Proc. Natl. Acad. Sci. USA, 106, 14 77314 777, doi:10.1073/pnas.0907610106.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Oleson, K. W., and Coauthors, 2010: Technical description of version 4.0 of the Community Land Model (CLM). NCAR Tech. Note NCAR/TN-478+STR, 257 pp.

  • Pearson, R. G., S. J. Phillips, M. M. Loranty, P. S. A. Beck, T. Damoulas, S. J. Knight, and S. J. Goetz, 2013: Shifts in Arctic vegetation and associated feedbacks under climate change. Nat. Climate Change, 3, 673677, doi:10.1038/nclimate1858.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Peng, J., L. Dan, and W. Dong, 2014: Are there interactive effects of physiological and radiative forcing produced by increased CO2 concentration on changes of land hydrological cycle? Global Planet. Change, 112, 6478, doi:10.1016/j.gloplacha.2013.11.007.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pielke, R. A., 2001: Influence of the spatial distribution of vegetation and soils on the prediction of cumulus convective rainfall. Rev. Geophys., 39, 151177, doi:10.1029/1999RG000072.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pitman, A. J., and M. Zhao, 2000: The relative impact of observed change in land cover and carbon dioxide as simulated by a climate model. Geophys. Res. Lett., 27, 12671270, doi:10.1029/1999GL011029.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Pu, B., and R. E. Dickinson, 2014: Hydrological changes in the climate system from leaf responses to increasing CO2. Climate Dyn., 42, 19051923, doi:10.1007/s00382-013-1781-1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Schär, C., D. Lüthi, U. Beyerle, and E. Heise, 1999: The soil–precipitation feedback: A process study with a regional climate model. J. Climate, 12, 722741, doi:10.1175/1520-0442(1999)012<0722:TSPFAP>2.0.CO;2.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Seneviratne, S. I., D. Lüthi, M. Litschi, and C. Schär, 2006: Land–atmosphere coupling and climate change in Europe. Nature, 443, 205209, doi:10.1038/nature05095.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Skinner, C. B., M. Ashfaq, and N. S. Diffenbaugh, 2012: Influence of twenty-first-century atmospheric and sea surface temperature forcing on West African climate. J. Climate, 25, 527542, doi:10.1175/2011JCLI4183.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Sun, Y., S. Solomon, A. Dai, and R. W. Portmann, 2007: How often will it rain? J. Climate, 20, 48014818, doi:10.1175/JCLI4263.1.

  • Swann, A. L. S., F. M. Hoffman, C. D. Koven, and J. T. Randerson, 2016: Plant responses to increasing CO2 reduce estimates of climate impacts on drought severity. Proc. Natl. Acad. Sci. USA, 113, 10 01910 024, doi:10.1073/pnas.1604581113.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Taylor, C. M., R. A. de Jeu, F. Guichard, P. P. Harris, and W. A. Dorigo, 2012: Afternoon rain more likely over drier soils. Nature, 489, 423426, doi:10.1038/nature11377.

    • Crossref
    • 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.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Trenberth, K. E., A. Dai, R. M. Rasmussen, and D. B. Parsons, 2003: The changing character of precipitation. Bull. Amer. Meteor. Soc., 84, 12051217, doi:10.1175/BAMS-84-9-1205.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van der Ent, R. J., L. Wang-Erlandsson, P. W. Keys, and H. H. G. Savenije, 2014: Contrasting roles of interception and transpiration in the hydrological cycle—Part 2: Moisture recycling. Earth Syst. Dynam., 5, 471489, doi:10.5194/esd-5-471-2014.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • van Vuuren, D. P., and Coauthors, 2011: The representative concentration pathways: An overview. Climatic Change, 109, 531, doi:10.1007/s10584-011-0148-z.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Wang, S., Y. Yang, A. P. Trishchenko, A. G. Barr, T. A. Black, and H. McCaughey, 2009: Modeling the response of canopy stomatal conductance to humidity. J. Hydrometeor., 10, 521532, doi:10.1175/2008JHM1050.1.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Zhu, Z., and Coauthors, 2016: Greening of the Earth and its drivers. Nat. Climate Change, 6, 791795, doi:10.1038/nclimate3004.

Supplementary Materials

Save