• Andrews, T., , and P. M. Forster, 2008: CO2 forcing induces semi-direct effects with consequences for climate feedback interpretations. Geophys. Res. Lett., 35, L04802, doi:10.1029/2007GL032273.

    • 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.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., , J. M. Gregory, , P. M. Forster, , and M. J. Webb, 2012a: Cloud adjustment and its role in CO2 radiative forcing and climate sensitivity: A review. Surv. Geophys., 33, 619635, doi:10.1007/s10712-011-9152-0.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., , J. M. Gregory, , M. J. Webb, , and K. E. Taylor, 2012b: Forcing, feedbacks and climate sensitivity in CMIP5 coupled atmosphere–ocean climate models. Geophys. Res. Lett., 39, L09712, doi:10.1029/2012GL051607.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., , M. A. Ringer, , M. Doutriaux-Boucher, , M. J. Webb, , and W. J. Collins, 2012c: Sensitivity of an Earth system climate model to idealized radiative forcing. Geophys. Res. Lett., 39, L10702, doi:10.1029/2012GL051942.

    • Search Google Scholar
    • Export Citation
  • Armour, K. C., , I. Eisenman, , E. Blanchard-Wrigglesworth, , K. E. McCusker, , and C. M. Bitz, 2011: The reversibility of sea ice loss in a state-of-the-art climate model. Geophys. Res. Lett., 38, L16705, doi:10.1029/2011GL048739.

    • Search Google Scholar
    • Export Citation
  • Armour, K. C., , C. M. Bitz, , and G. H. Roe, 2013: Time-varying climate sensitivity from regional feedbacks. J. Climate, 26, 4518–4534.

  • Boer, G. J., , and B. Yu, 2003: Climate sensitivity and response. Climate Dyn., 20, 415429.

  • 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.

    • Search Google Scholar
    • Export Citation
  • Boucher, O., and Coauthors, 2012: Reversibility in an Earth system model in response to CO2 concentration changes. Environ. Res. Lett., 7, 024013, doi:10.1088/1748-9326/7/2/024013.

    • Search Google Scholar
    • Export Citation
  • Bouttes, N., , J. M. Gregory, , and J. A. Lowe, 2013: The reversibility of sea level rise. J. Climate, 26, 25022513.

  • 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.

    • Search Google Scholar
    • Export Citation
  • Cao, L., , G. Bala, , and K. Caldeira, 2011: Why is there a short-term increase in global precipitation in response to diminished CO2 forcing? Geophys. Res. Lett., 38, L06703, doi:10.1029/2011GL046713.

    • Search Google Scholar
    • Export Citation
  • Chadwick, R., , P. Wu, , P. Good, , and T. Andrews, 2013: Asymmetries in tropical rainfall and circulation patterns in idealised CO2 removal experiments. Climate Dyn., 40, 295316, doi:10.1007/s00382-012-1287-2.

    • Search Google Scholar
    • Export Citation
  • Collins, W. J., and Coauthors, 2011: Development and evaluation of an Earth-System model—HadGEM2. Geosci. Model Dev., 4, 10511075, doi:10.5194/gmd-4-1051-2011.

    • Search Google Scholar
    • Export Citation
  • Colman, R., , and B. McAvaney, 2011: On tropospheric adjustment to forcing and climate feedbacks. Climate Dyn., 36, 16491658, doi:10.1007/s00382-011-1067-4.

    • Search Google Scholar
    • Export Citation
  • Danabasoglu, G., , and P. R. Gent, 2009: Equilibrium climate sensitivity: Is it accurate to use a slab ocean model? J. Climate, 22, 24942499.

    • Search Google Scholar
    • Export Citation
  • Dong, B., , J. M. Gregory, , and R. T. Sutton, 2009: Understanding land–sea warming contrast in response to increasing greenhouse gases. Part I: Transient adjustment. J. Climate, 22, 30793097.

    • Search Google Scholar
    • Export Citation
  • Doutriaux-Boucher, M., , M. J. Webb, , J. M. Gregory, , and O. Boucher, 2009: Carbon dioxide induced stomatal closure increases radiative forcing via a rapid reduction in low cloud. Geophys. Res. Lett., 36, L02703, doi:10.1029/2008GL036273.

    • Search Google Scholar
    • Export Citation
  • Forster, P. M., , and K. Taylor, 2006: Climate forcings and climate sensitivities diagnosed from coupled climate model integrations. J. Climate, 19, 61816194.

    • Search Google Scholar
    • Export Citation
  • Forster, P. M., , T. Andrews, , P. Good, , J. M. Gregory, , L. S. Jackson, , and M. Zelinka, 2013: Evaluating adjusted forcing and model spread for historical and future scenarios in the CMIP5 generation of climate models. J. Geophys. Res., 118, 1139–1150, doi:10.1002/jgrd.50174.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., , K. H. Cook, , and M. E. Schlesinger, 1981: Preliminary analysis of experiments on the climatic effects of increased CO2 with an atmospheric general circulation model and a climatological ocean. J. Geophys. Res., 86 (C7), 63856393.

    • Search Google Scholar
    • Export Citation
  • Good, P., , J. M. Gregory, , and J. A. Lowe, 2011: A step-response simple climate model to reconstruct and interpret AOGCM projections. Geophys. Res. Lett., 38, L01703, doi:10.1029/2010GL045208.

    • Search Google Scholar
    • Export Citation
  • Good, P., , J. M. Gregory, , J. A. Lowe, , and T. Andrews, 2013: Abrupt CO2 experiments as tools for predicting and understanding CMIP5 representative concentration pathway projections. Climate Dyn., 40, 10411053, doi:10.1007/s00382-012-1410-4.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., 2000: Vertical heat transports in the ocean and their effect on time-dependent climate change. Climate Dyn., 16, 501515.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., , and J. F. B. Mitchell, 1997: The climate response to CO2 of the Hadley Centre coupled AOGCM with and without flux adjustment. Geophys. Res. Lett., 24, 19431946.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., , and P. Forster, 2008: Transient climate response estimated from radiative forcing and observed temperature change. J. Geophys. Res., 113, D23105, doi:10.1029/2008JD010405.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., , and M. J. Webb, 2008: Tropospheric adjustment induces a cloud component in CO2 forcing. J. Climate, 21, 5871.

  • Gregory, J. M., and Coauthors, 2004: A new method for diagnosing radiative forcing and climate sensitivity. Geophys. Res. Lett., 31, L03205, doi:10.1029/2003GL018747.

    • Search Google Scholar
    • Export Citation
  • Hansen, J., and Coauthors, 2005: Efficacy of climate forcings. J. Geophys. Res., 110, D18104, doi:10.1029/2005JD005776.

  • Hasselmann, K., , R. Sausen, , E. Maier-Reimer, , and R. Voss, 1993: On the cold start problem in transient simulations with coupled atmosphere–ocean models. Climate Dyn., 9, 5361.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., and Coauthors, 2010: Probing the fast and slow components of global warming by returning abruptly to preindustrial forcing. J. Climate, 23, 24182427.

    • Search Google Scholar
    • Export Citation
  • Jiang, J. H., and Coauthors, 2012: Evaluation of cloud and water vapor simulations in CMIP5 climate models using NASA “A-Train” satellite observations. J. Geophys. Res., 117, D14105, doi:10.1029/2011JD017237.

    • Search Google Scholar
    • Export Citation
  • Kamae, Y., , and M. Watanabe, 2012: On the robustness of tropospheric adjustment in CMIP5 models. Geophys. Res. Lett., 39, L23808, doi:10.1029/2012GL054275.

    • Search Google Scholar
    • Export Citation
  • Kamae, Y., , and M. Watanabe, 2013: Tropospheric adjustment to increasing CO2: Its timescale and the role of land–sea contrast. Climate Dyn., 41, 3007–3024, doi:10.1007/s00382-012-1555-1.

    • Search Google Scholar
    • Export Citation
  • Keen, A. B., , and J. M. Murphy, 1997: Influence of natural variability and the cold start problem on the simulated transient response to increasing CO2. Climate Dyn., 13, 847864, doi:10.1007/s003820050201.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., , and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6, 15871606.

  • Klein, S. A., , and C. Jakob, 1999: Validation and sensitivities of frontal clouds simulated by the ECMWF model. Mon. Wea. Rev., 127, 25142531.

    • Search Google Scholar
    • Export Citation
  • Kuhlbrodt, T., , and J. M. Gregory, 2012: Ocean heat uptake and its consequences for the magnitude of sea level rise and climate change. Geophys. Res. Lett., 39, L18608, doi:10.1029/2012GL052952.

    • Search Google Scholar
    • Export Citation
  • Lambert, F. H., , M. J. Webb, , and M. M. Joshi, 2011: The relationship between land–ocean surface temperature contrast and radiative forcing. J. Climate, 24, 32393256.

    • Search Google Scholar
    • Export Citation
  • Li, C., , J.-S. V. Storch, , and J. Marotzke, 2013: Deep-ocean heat uptake and equilibrium climate response. Climate Dyn., 40, 1071–1086, doi:10.1007/s00382-012-1350-z.

    • Search Google Scholar
    • Export Citation
  • Lock, A. P., and Coauthors, 2000: A new boundary layer mixing scheme. Part I: Scheme description and single-column model tests. Mon. Wea. Rev., 128, 31873199.

    • Search Google Scholar
    • Export Citation
  • Long, D. J., , and M. Collins, 2013: Quantifying global climate feedbacks, responses and forcing under abrupt and gradual CO2 forcing. Climate Dyn., 41, 2471–2479, doi:10.1007/s00382-013-1677-0.

    • Search Google Scholar
    • Export Citation
  • Martin, G. M., and Coauthors, 2000: A new boundary layer mixing scheme. Part II: Tests in climate and mesoscale models. Mon. Wea. Rev., 128, 32003217.

    • Search Google Scholar
    • Export Citation
  • Martin, G. M., and Coauthors, 2011: The HadGEM2 family of Met Office Unified Model Climate configurations. Geosci. Model Dev., 4, 723757, doi:10.5194/gmd-4-723-2011.

    • Search Google Scholar
    • Export Citation
  • Mitchell, J. F. B., 1983: The seasonal response of a general circulation model to changes in CO2 and sea temperatures. Quart. J. Roy. Meteor. Soc., 109, 113152.

    • Search Google Scholar
    • Export Citation
  • Mitchell, J. F. B., , and W. J. Ingram, 1992: Carbon dioxide and climate: Mechanisms of changes in cloud. J. Climate, 5, 521.

  • Myhre, G., , E. J. Highwood, , K. P. Shine, , and F. Stordal, 1998: New estimates of radiative forcing due to well mixed greenhouse gases. Geophys. Res. Lett., 25, 27152718.

    • Search Google Scholar
    • Export Citation
  • Raper, S. C. B., , J. M. Gregory, , and R. J. Stouffer, 2002: The role of climate sensitivity and ocean heat uptake on AOGCM transient temperature response. J. Climate, 15, 124130.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and Coauthors, 1996: Comparison of radiative and physiological effects of doubled atmospheric CO2 on climate. Science, 271, 14021406.

    • Search Google Scholar
    • Export Citation
  • Senior, C. A., , and J. F. B. Mitchell, 2000: The time-dependence of climate sensitivity. Geophys. Res. Lett., 21, 26852688, doi:10.1029/2000GL011373.

    • Search Google Scholar
    • Export Citation
  • Shine, K. P., , J. Cook, , E. J. Highwood, , and M. J. Joshi, 2003: An alternative to radiative forcing for estimating the relative importance of climate change mechanisms. Geophys. Res. Lett., 30, 2047, doi:10.1029/2003GL018141.

    • Search Google Scholar
    • Export Citation
  • Tomassini, L., and Coauthors, 2013: The respective roles of surface temperature driven feedbacks and tropospheric adjustment to CO2 in CMIP5 transient climate simulations. Climate Dyn., 41, 3103–3126, doi:10.1007/s00382-013-1682-3.

    • Search Google Scholar
    • Export Citation
  • Vial, J., , J.-L. Dufresne, , and S. Bony, 2013: On the interpretation of inter-model spread in CMIP5 climate sensitivity estimates. Climate Dyn., 41, 3339–3362, doi:10.1007/s00382-013-1725-9.

    • Search Google Scholar
    • Export Citation
  • Vilà-Guerau de Arellano, J., , C. C. van Heerwaarden, , and J. Lelieveld, 2012: Modelled suppression of boundary-layer clouds by plants in a CO2-rich atmosphere. Nat. Geosci., 5, 701704, doi:10.1038/ngeo1554.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , H. Shiogama, , M. Yoshimori, , T. Ogura, , T. Yokohata, , H. Okamoto, , S. Emori, , and M. Kimoto, 2012: Fast and slow timescales in the tropical low-cloud response to increasing CO2 in two climate models. Climate Dyn., 39, 16271641, doi:10.1007/s00382-011-1178-y.

    • Search Google Scholar
    • Export Citation
  • Webb, M. J., , and A. P. Lock, 2013: Coupling between subtropical cloud feedback and the local hydrological cycle in a climate model. Climate Dyn., 41, 1923–1939, doi:10.1007/s00382-012-1608-5.

    • Search Google Scholar
    • Export Citation
  • Webb, M. J., , C. A. Senior, , S. Bony, , and J.-J. Morcrette, 2001: Combining ERBE and ISCCP data to assess cloud in the Hadley Centre, ECMWF and LMD atmospheric climate models. Climate Dyn., 17, 905922.

    • Search Google Scholar
    • Export Citation
  • Webb, M. J., , F. H. Lambert, , and J. M. Gregory, 2013: Origins of differences in climate sensitivity, forcing and feedback in climate models. Climate Dyn., 40, 677–707, doi:10.1007/s00382-012-1336-x.

    • Search Google Scholar
    • Export Citation
  • Williams, K. D., , W. J. Ingram, , and J. M. Gregory, 2008: Time variation of effective climate sensitivity in GCMs. J. Climate, 21, 50765090.

    • Search Google Scholar
    • Export Citation
  • Winton, M., , K. Takahashi, , and I. M. Held, 2010: Importance of ocean heat uptake efficacy to transient climate change. J. Climate, 23, 23332344.

    • Search Google Scholar
    • Export Citation
  • Wu, P., , R. Wood, , J. Ridley, , and J. Lowe, 2010: Temporary acceleration of the hydrological cycle in response to a CO2 rampdown. Geophys. Res. Lett., 37, L12705, doi:10.1029/2010GL043730.

    • Search Google Scholar
    • Export Citation
  • Wyant, M. C., , C. S. Bretherton, , P. N. Blossey, , and M. Khairoutdinov, 2012: Fast cloud adjustment to increasing CO2 in a superparameterized climate model. J. Adv. Model. Earth Syst., 4, M05001, doi:10.1029/2011MS000092.

    • Search Google Scholar
    • Export Citation
  • Zelinka, M. D., , S. A. Klein, , K. E. Taylor, , T. Andrews, , M. J. Webb, , J. M. Gregory, , and P. M. Forster, 2013: Contributions of different cloud types to feedbacks and rapid adjustments in CMIP5. J. Climate, 26, 5007–5027.

    • Search Google Scholar
    • Export Citation
  • View in gallery

    Time series of (a) CO2 concentration and effective radiative forcing (ERF) and (b) global-mean surface air-temperature change (ΔT). (c) The relationship between forcing and ΔT for the period of increasing (black) and decreasing (red) forcing, the slope giving the climate resistance (ρ). (d) Change in TOA radiative fluxes. Note that all data are globally and annually averaged, and radiative fluxes are defined as positive downward.

  • View in gallery

    Comparison of the global annual-mean time series of (a) effective radiative forcing (ERF) and its (b) clear-sky and (c) cloud radiative effect (CRE) components. Each term has been diagnosed via two methods using fixed-SST experiments: (i) from the change in TOA flux in a 1% yr−1 compound CO2 increase/decrease experiment with fixed climatological 1860 SSTs (shown as the thin lines exhibiting interannual variability) and (ii) as a time slice step 4×CO2 at the point of CO2 quadrupling (year 140), linearly scaled with time (thick lines with no interannual variability).

  • View in gallery

    Change in global annual-mean TOA radiative fluxes, ΔNi (from the fully coupled AOGCM simulation), minus the associated component of the effective radiative forcing, Fi (from the fixed-SST experiment), as a function of global annual-mean surface air temperature change, ΔT. Subscript i denotes the radiation component: net, LW clear-sky, SW clear-sky, and net (LW + SW) CRE. Points are global annual means for the periods of increasing (black) and decreasing (red) forcing. Radiative fluxes are defined as positive downward. The slope of the lines measures the climate feedback parameter.

  • View in gallery

    Change in global annual-mean TOA LW clear-sky radiation from the fully coupled AOGCM as a function of global-mean surface air temperature change, ΔT (i.e., as in Fig. 3, but without removing the associated component of the effective radiative forcing from the fixed-SST experiment). Points are global annual means for the periods of increasing (black) and decreasing (red) forcing. Radiative fluxes are defined as positive downward.

  • View in gallery

    CO2-driven cloud adjustments in the nine basic ISCCP cloud categories at the end (years 125–140) of the 1% ramp-up simulation for (a)–(c) high-, (d)–(f) middle-, and (g)–(i) low-level clouds. The adjustments are estimated from the fixed-SST experiment. Units are fractional coverage measured in percent.

  • View in gallery

    Total (a)–(c) high- and (d)–(f) low-level cloud adjustments at the end (years 125–140) of the 1% ramp-up simulation and their separation into radiative (RAD) and plant physiological effects (PHYS). The adjustments are estimated from the 4×CO2, 4×CO2RAD, and 4×CO2PHYS fixed-SST experiments described in the text. FULL indicates both radiative and physiological effects are present. Units are fractional coverage measured in percent.

  • View in gallery

    CO2-driven adjustments in various parameters related to the cloud changes at the end (years 125–140) of the 1% ramp-up simulation. The adjustments are estimated from the fixed-SST experiment. (bottom) The physiological component of the surface latent and sensible heating to adjustments from the 4×CO2PHYS fixed-SST experiment (negative indicates a reduction in flux from the surface to the atmosphere for the latent and sensible heat terms).

  • View in gallery

    The CO2-driven adjustment of the boundary layer types defined by the boundary layer mixing scheme at the end (years 125–140) of the 1% ramp-up simulation. For clarity, the seven types used by the scheme have been reduced to five by combining the two stable types (1 and 2) and the two decoupled types (4 and 5). Units are fractional coverage measured in percent.

  • View in gallery

    Cloud feedback contribution to the ISCCP-defined (a)–(c) high- and (d)–(f) low-level cloud types at the end (years 125–140) of the 1% ramp-up simulation. These are estimated as a residual between the fully coupled AOGCM cloud changes and the adjustments estimated from the fixed-SST experiment. Units are fractional coverage measured in percent.

  • View in gallery

    The boundary-layer-type feedback contributions at the end (years 125–140) of the 1% ramp-up simulation. These are estimated as a residual between the fully coupled AOGCM cloud changes and the adjustments estimated from the fixed-SST experiment. Units are fractional coverage measured in percent.

  • View in gallery

    The zonal-mean cloud amount adjustment and feedback changes over (a),(d) ocean and (b),(e) land with respect to the control at the end (years 125–140) of the 1% ramp-up experiment. (c),(f) Also shown are the adjustments over land in the RAD and PHYS experiments. The cloud amount changes are shown on the model’s native vertical grid from the surface up to a height of 18 km. Units are fractional coverage measured in percent. The vertical axis is height in meters.

  • View in gallery

    CO2-driven adjustments in relative humidity (RH) over ocean and land and its breakdown into components due to changes in specific humidity (q) and temperature (T) as described in the text. The bottom row shows the plant physiological impact on the RH adjustment and its components over land. Changes are estimated from the relevant fixed-SST experiment at the end (years 125–140) of the 1% ramp-up experiment. The vertical axis is height in meters.

  • View in gallery

    (a),(b) High- and (c),(d) low-level cloud amount feedbacks at the point of CO2 doubling (year 70 and 210, relative to the control) during the ramp-up and ramp-down phases of the simulation. These are estimated from the cloud amount changes in the fully coupled AOGCM minus the corresponding adjustment from the fixed-SST experiment, then normalized by the global-mean temperature change.

  • View in gallery

    The variation of local (a)–(c) surface temperature change (K K−1) and (d)–(f) clear-sky absorbed SW radiation (W m−2 K−1) over the Southern Hemisphere high latitudes, with global-mean surface temperature change, for three consecutive periods of equal decline in forcing during the ramp-down phase. The plots are calculated from the slope of the regression of local annual-mean ΔT, or clear-sky absorbed SW radiation, against global annual-mean ΔT during the relevant period. The sign has been reversed, so that in (a)–(c) negative values indicate that local surface temperature is declining, in line with the global mean. Negative values in (d)–(f) indicate that absorbed SW radiation is declining, in line with the global-mean cooling.

All Time Past Year Past 30 Days
Abstract Views 0 0 0
Full Text Views 61 61 7
PDF Downloads 49 49 10

Cloud Feedbacks, Rapid Adjustments, and the Forcing–Response Relationship in a Transient CO2 Reversibility Scenario

View More View Less
  • 1 Met Office Hadley Centre, Exeter, United Kingdom
© Get Permissions
Full access

Abstract

The Hadley Centre Global Environment Model, version 2–Earth System (HadGEM2-ES) climate model is forced by a 1% yr−1 compound increase in atmospheric CO2 for 140 years, followed by a 1% yr−1 CO2 decrease back to the starting level. Analogous atmosphere-only simulations are performed to diagnose the component of change associated with the effective radiative forcing and rapid adjustments. The residual change is associated with radiative feedbacks that are shown to be linearly related to changes in global-mean surface air temperature and are found to be reversible under this experimental design, even for regional cloud feedback changes. The cloud adjustment is related to changes in cloud amount, with little indication of any large-scale changes in cloud optical depth. Plant physiological forcing plays a significant role in determining the cloud adjustment in this model and is the dominant contribution to the low-level cloud changes over land. Low-level cloud adjustments are associated with changes in surface turbulent fluxes and lower tropospheric stability, with significant adjustments in boundary layer cloud types and in the depth of the boundary layer itself. The linearity of simple forcing–response frameworks are examined and found to be generally applicable. Small regional departures from linearity occur during the early part of the ramp-down phase, where the Southern Ocean and eastern tropical Pacific continue to warm for a few decades, despite the reversal in radiative forcing and global temperatures. The importance of considering time-varying patterns of warming and regional phenomena when diagnosing and understanding feedbacks in a coupled atmosphere–ocean framework is highlighted.

Corresponding author address: Timothy Andrews, Met Office Hadley Centre, FitzRoy Road, Exeter EX1 3PB, United Kingdom. E-mail: timothy.andrews@metoffice.gov.uk

Abstract

The Hadley Centre Global Environment Model, version 2–Earth System (HadGEM2-ES) climate model is forced by a 1% yr−1 compound increase in atmospheric CO2 for 140 years, followed by a 1% yr−1 CO2 decrease back to the starting level. Analogous atmosphere-only simulations are performed to diagnose the component of change associated with the effective radiative forcing and rapid adjustments. The residual change is associated with radiative feedbacks that are shown to be linearly related to changes in global-mean surface air temperature and are found to be reversible under this experimental design, even for regional cloud feedback changes. The cloud adjustment is related to changes in cloud amount, with little indication of any large-scale changes in cloud optical depth. Plant physiological forcing plays a significant role in determining the cloud adjustment in this model and is the dominant contribution to the low-level cloud changes over land. Low-level cloud adjustments are associated with changes in surface turbulent fluxes and lower tropospheric stability, with significant adjustments in boundary layer cloud types and in the depth of the boundary layer itself. The linearity of simple forcing–response frameworks are examined and found to be generally applicable. Small regional departures from linearity occur during the early part of the ramp-down phase, where the Southern Ocean and eastern tropical Pacific continue to warm for a few decades, despite the reversal in radiative forcing and global temperatures. The importance of considering time-varying patterns of warming and regional phenomena when diagnosing and understanding feedbacks in a coupled atmosphere–ocean framework is highlighted.

Corresponding author address: Timothy Andrews, Met Office Hadley Centre, FitzRoy Road, Exeter EX1 3PB, United Kingdom. E-mail: timothy.andrews@metoffice.gov.uk

1. Introduction

a. Conceptual framework

The earth’s energy balance provides a convenient framework for understanding the global response of the climate system to natural and anthropogenic forcings. During transient climate change, the global surface air temperature change ΔT (K) is determined by forcing, feedback, and heat uptake processes. These quantities can be related through a simple linear relationship whereby the net heat flux into the climate system N (W m−2)—which is overwhelmingly into the ocean—is partitioned between an external radiative perturbation F (W m−2) and a heat flux lost to space that is determined by climate feedback processes [−αΔT (W m−2)],
e1
where −α (W m−2 K−1) is the climate feedback parameter (e.g., Gregory et al. 2004).
Under scenarios of steadily increasing forcing, such as a 1% compound increase in CO2 per year, models suggest a further simplification so that F is proportional to ΔT,
e2
where ρ (W m−2 K−1) is the “climate resistance” (Gregory and Forster 2008). This views the deep ocean as a heat sink, which takes heat away from the mixed layer proportionally to ΔT, so that N = κΔT, where κ (W m−2 K−1) is the “ocean heat uptake efficiency” (Gregory and Mitchell 1997; Raper et al. 2002). Equations (1) and (2) show that climate resistance depends on both the ocean heat uptake efficiency and climate feedbacks, ρ = κ + α.

Linearity of these relationships is clearly only an approximation. Their applicability and limitations have been explored in fully coupled atmosphere–ocean general circulation models (AOGCMs) (e.g., Gregory et al. 2004; Forster and Taylor 2006; Williams et al. 2008; Danabasoglu and Gent 2009; Winton et al. 2010; Held et al. 2010; Li et al. 2013; Andrews et al. 2012b). Yet most have focused on the global mean, and little has been done on the regional processes. In addition, it is unclear how well these concepts apply to mitigation scenarios that aim to avoid, or even reverse, potentially dangerous climate change (Bouttes et al. 2013). For example, are feedback processes that influence the global radiation balance reversible?

b. Effective radiative forcing and adjustments

A recent development has been the realization that some aspects of cloud changes, when driven by increased CO2, are better thought of as a rapid atmospheric adjustment to be included in the external perturbation F, rather than a surface temperature mediated feedback [see Andrews et al. (2012a) for a review of these developments]. The easiest way to demonstrate these atmospheric adjustments is with an atmospheric general circulation model (i.e., with fixed SSTs). In such experiments, the atmosphere and land surface are free to respond to the perturbation, but the large-scale surface temperature mediated feedbacks are inhibited. Such experiments have long been used to investigate the atmospheric response to forcing (e.g., Gates et al. 1981; Mitchell, 1983).

Fixed-SST experiments provide a convenient way of diagnosing F from the change in top-of-atmosphere radiation balance [ΔT ~ 0 implies N ~ F from Eq. (1)] (Shine et al. 2003; Hansen et al. 2005). We refer to this as the effective radiative forcing (ERF). There is significant intermodel spread in ERF, seen in both transient scenarios (Forster and Taylor 2006; Forster et al. 2013) and idealized CO2 experiments, whereby cloud adjustments have been identified as a significant contributing uncertainty to the ERF and climate sensitivity (Gregory and Webb 2008; Andrews and Forster 2008; Andrews et al. 2012b; Webb et al. 2013; Kamae and Watanabe 2012; Vial et al. 2013). It is therefore important to investigate cloud adjustments and associated processes in models in more detail.

c. Aims of this study

In this paper we aim to explore these concepts further. We will build on previous studies by examining the simple forcing–response relationships [Eqs. (1) and (2)] at both a global and regional scale in a transient reversibility scenario. We will additionally focus on the role of transient cloud changes, and the processes involved in their adjustments and feedbacks, for a number of reasons.

  1. Clouds continue to be a major source of intermodel spread in climate feedbacks and climate sensitivity in models (Andrews et al. 2012b; Vial et al. 2013).
  2. Little has been done on the physical processes behind cloud adjustments to CO2 and how they are manifested regionally and in a transient scenario.
  3. Clouds are believed to be a source of nonlinearity in the forcing–feedback relationship (Senior and Mitchell 2000; Williams et al. 2008; Andrews et al. 2012b; Zelinka et al. 2013).
  4. Boucher et al. (2012) identified potential hysteresis behavior in global low-level cloud amount with reversibility experiments using the same model used in this study [the Hadley Centre Global Environment Model, version 2–Earth System (HadGEM2-ES); see section 2].
  5. There is a sound basis for examining cloud changes in HadGEM2-ES, as its present-day simulation of clouds has been shown to be one of the best—compared to a range of observational metrics—of the models participating in phase 5 of the Coupled Model Intercomparison Project (CMIP5) (Jiang et al. 2012).
We will use the Met Office Hadley Centre climate model HadGEM2-ES forced by a 1% yr−1 compound increase in CO2, followed by a 1% yr−1 decrease (see section 2). While clearly not a realistic scenario of CO2 removal, the design provides a framework for understanding processes relevant to a hypothetical mitigation scenario. Boucher et al. (2012) used the same experiment to provide an overview of the reversibility and “hysteresis type” behavior of various processes, focusing on “Earth system” components. Chadwick et al. (2013) described asymmetries in the spatial patterns of precipitation changes over the tropical oceans.

Section 2 describes the model and experiments. Section 3 details the global radiation balance and the applicability of the linear forcing–feedback relationships. Section 4 focuses on the role of clouds, describing the transient cloud changes in terms of adjustments and feedbacks and associated processes. Section 5 shows where some of the above relationships break down and how this relates to changing patterns of warming. Section 6 presents a summary and discussion.

2. Experiments and method

a. Climate model description

We use the earth system (ES) configuration of the Met Office Hadley Centre Global Environmental Model version 2 (Martin et al. 2011; Collins et al. 2011). HadGEM2-ES is a state-of-the-art coupled earth system model that includes atmospheric, land surface, hydrology, aerosol, ocean, and sea ice processes. The earth system component includes terrestrial and oceanic ecosystems, as well as interactive tropospheric chemistry and the various couplings between them. The atmospheric component has 38 levels in the vertical and a horizontal resolution of 1.25° latitude × 1.875° longitude. A detailed description of the model can be found in Collins et al. (2011) and Martin et al. (2011).

In response to CO2 forcing, HadGEM2-ES has as a transient climate response (TCR), defined as ΔT averaged over years 61–80 in a 1% yr−1 compound increase in CO2 scenario, of 2.5 K (Andrews et al. 2012c) and an effective climate sensitivity of ~4.6 K (Andrews et al. 2012b), which is close to the upper bound of CMIP5 models. It has a relatively large cloud feedback, with large positive contributions coming from the low cloud regions off the west coasts of the tropical continents (Webb and Lock 2013). Its 4×CO2 ERF (7.0 W m−2) is close to the mean of CMIP5 models (Andrews et al. 2012b; Kamae and Watanabe 2012).

b. Climate model experiments

1) Fully coupled simulations

From a stable preindustrial control (CO2 concentration 286 ppmv), the CO2 concentration is increased by 1% yr−1 for 140 years, reaching 4 times its preindustrial level (1144 ppmv). CO2 levels are then reduced by 1% yr−1 back to preindustrial levels (see Fig. 1a, blue line). The global annual mean response ΔT is shown in Fig. 1b.

Fig. 1.
Fig. 1.

Time series of (a) CO2 concentration and effective radiative forcing (ERF) and (b) global-mean surface air-temperature change (ΔT). (c) The relationship between forcing and ΔT for the period of increasing (black) and decreasing (red) forcing, the slope giving the climate resistance (ρ). (d) Change in TOA radiative fluxes. Note that all data are globally and annually averaged, and radiative fluxes are defined as positive downward.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

2) Atmosphere-only simulations

To diagnose the ERF (section 3a) and related adjustments we repeat the control and CO2 ramp-up/ down experiments with just the atmospheric component of the model (i.e., we replace the ocean and sea ice models with climatological SST and sea ice distributions representative of the fully coupled preindustrial control simulation).

We perform three additional atmosphere-only simulations. In all cases the CO2 levels are instantaneously increased to 4 times their preindustrial levels (equivalent to a time slice at year 140 in the transient experiments) and run for 30 years. The three experiments are coupled with different aspects of the climate system:

  1. 4×CO2: a step 4 × CO2 increase, equivalent to a time slice at year 140 in the transient CO2 ramp-up/down fixed-SST experiment;
  2. 4×CO2RAD: as in 4×CO2 but only the radiation scheme sees the 4 × CO2; plants continue to see 1 × CO2;
  3. 4×CO2PHYS: as in 4×CO2 but only the plant scheme sees the 4 × CO2; the radiation scheme continues to see 1 × CO2.
These three experiments have two purposes: (i) different couplings with plants/radiation allow us to diagnose the component of ERF and adjustments that are radiatively driven, versus those which come about owing to CO2 plant physiological effects [similar experimental designs have been used previously by Doutriaux-Boucher et al. (2009), Boucher et al. (2009), Andrews et al. (2011), and others; see also section 4], and (ii) 30-yr experiments provide a clean signal of the response, which is needed for regional analysis where variability can be large.

3. Global climate change

a. Effective radiative forcing

The effective radiative forcing, calculated as the change in net top-of-atmosphere (TOA) radiative flux in the fixed-SST simulations, increases/decreases linearly with time (Fig. 1a). This is expected for the instantaneous radiative forcing, due to its logarithmic dependence on the change in CO2 (e.g., Myhre et al. 1998), but here we show that it also holds for the ERF. The 4×CO2 ERF calculations can then be used to calibrate the logarithmic formula to determine the ERF for any CO2 level. The CO2 forcing is given by F = β ln(C/C0), where C (C0) are the perturbed (control) CO2 levels and β is a constant to be determined (Myhre et al. 1998). Using the ERF determined from the 4×CO2 fixed-SST step experiment (F = 7.0 W m−2) and C/C0 = 4 gives β = 5.05 for this model.

We compare the transient ERF with F—determined from the abrupt 4×CO2 time slice—linearly scaled with time [e.g., F(t) = Ft/140 for the ramp up] in Fig. 2a. The agreement is excellent, even for the longwave and shortwave clear-sky and cloud radiative effect (CRE), defined as the difference between all-sky and clear-sky, contributions (Figs. 2b,c). In the rest of the analysis we use the linearly scaled 4×CO2 fixed-SST experiments to represent the ERF and adjustments as they have a better signal-to-noise ratio, especially at the regional scale.

Fig. 2.
Fig. 2.

Comparison of the global annual-mean time series of (a) effective radiative forcing (ERF) and its (b) clear-sky and (c) cloud radiative effect (CRE) components. Each term has been diagnosed via two methods using fixed-SST experiments: (i) from the change in TOA flux in a 1% yr−1 compound CO2 increase/decrease experiment with fixed climatological 1860 SSTs (shown as the thin lines exhibiting interannual variability) and (ii) as a time slice step 4×CO2 at the point of CO2 quadrupling (year 140), linearly scaled with time (thick lines with no interannual variability).

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

b. Climate resistance

For both ramp up and ramp down a constant ρ is a good approximation (Fig. 1c), although the relationship between F and ΔT is different between the ramp-up (ρ = 1.22 W m−2 K−1) and ramp-down (ρ = 1.55 W m−2 K−1) periods. As F is symmetrical (Fig. 1a), this must be because ΔT decreases less in the ramp-down period than it increased in the ramp up (Fig. 1b). This could arise from either more stabilizing feedbacks (i.e., α is larger) in the ramp down and/or because κ is larger. In section 3c we will show that α is consistent between ramp up and ramp down; hence differences in ρ must arise from differences in κ.

The importance of the heat flux into the deep ocean can be seen by considering the evolution of ΔT (Fig. 1b) in relation to the TOA radiative fluxes (Fig. 1d). Throughout most of the simulation the system is taking up heat, that is, N > 0 (Fig. 1d, black line); yet (counterintuitively), ΔT is declining soon after CO2 begins to be removed. Hence the surface is cooling despite a positive energy imbalance at the TOA. It must be the case that the surface and ocean mixed layer is losing more heat to the deep ocean than it is gaining from the atmosphere. We do not pursue this further, but it could be explored with two-layer models that represent fluxes between the mixed layer and deep ocean (e.g., Gregory 2000; Held et al. 2010; Bouttes et al. 2013) or step-response models (Good et al. 2011; Bouttes et al. 2013).

As described by Gregory and Forster (2008), the small but nonzero F intercept (Fig. 1c) is likely due to the “cold start” effect (Hasselmann et al. 1993; Keen and Murphy 1997), which gives rise to a nonzero intercept of N against ΔT, meaning a constant κ is not exact. We see that an analogous effect occurs during the first few years of the ramp-down period (Fig. 1c), where the points deviate from the straight line. It will be shown (section 5) that during this period most of globe cools immediately after the CO2 forcing is reversed, but the Southern Ocean continues to warm for a number of years.

c. Climate feedback and “hysteresis”

The transient changes in TOA radiative fluxes (Fig. 1d) are the result of forcing and feedback processes. The fixed-SST experiments allow us to diagnose the component of these changes that are associated with forcing and adjustment processes. The feedback component can then be determined by resolving Eq. (1) into its radiative component terms (e.g., Gregory and Webb 2008): Ni = Fi − αiΔT, where subscript i denotes the longwave (LW) clear-sky, shortwave (SW) clear-sky, or net (LW + SW) CRE. Figure 3 shows the relationship between Ni − Fi and ΔT, the slope giving the feedback parameter −αi. The ramp down (red points) largely returns along the same path as the ramp up in all components; that is, the feedback processes that determine the global radiation balance are largely reversible under this scenario.

Fig. 3.
Fig. 3.

Change in global annual-mean TOA radiative fluxes, ΔNi (from the fully coupled AOGCM simulation), minus the associated component of the effective radiative forcing, Fi (from the fixed-SST experiment), as a function of global annual-mean surface air temperature change, ΔT. Subscript i denotes the radiation component: net, LW clear-sky, SW clear-sky, and net (LW + SW) CRE. Points are global annual means for the periods of increasing (black) and decreasing (red) forcing. Radiative fluxes are defined as positive downward. The slope of the lines measures the climate feedback parameter.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

Andrews et al. (2012b) determined F and the −αi components for various CMIP5 models forced by an instantaneous 4 × CO2. Their F (determined from a linear regression method) is slightly different from that determined through fixed-SST experiments (see Andrews et al. 2012b). If we use their F, we then determine feedback components on both the ramp up and down almost identical to those determined from the abrupt 4×CO2 simulation for HadGEM2-ES. Kuhlbrodt and Gregory (2012) similarly showed that this is applicable widely across CMIP5 models, although Long and Collins (2013) point to small nonlinearities. Consistent forcing and feedback behavior of the climate system under large step and realistic transient forcing experiments underpins the simple model of Good et al. (2011, 2013), who showed that transient 1% CO2 experiments can be derived from a linear convolution of scaled down 4×CO2 results.

Recent studies (e.g., Wu et al. 2010; Boucher et al. 2012) have used reversibility scenarios to identify potential hysteresis behavior. Such an example is demonstrated in Fig. 4 for the LW clear-sky radiative flux, which does not return down the same path as it went up (i.e., there appear to be two possible radiative states for the same global-mean temperature state). However, such simple analysis can be misleading because it ignores the role of forcing and adjustments (Boucher et al. 2012). When the forcing is correctly considered (Fig. 3), the hysteresis-type behavior is removed. An analogous effect has been described in the context of global-mean precipitation (Cao et al. 2011).

Fig. 4.
Fig. 4.

Change in global annual-mean TOA LW clear-sky radiation from the fully coupled AOGCM as a function of global-mean surface air temperature change, ΔT (i.e., as in Fig. 3, but without removing the associated component of the effective radiative forcing from the fixed-SST experiment). Points are global annual means for the periods of increasing (black) and decreasing (red) forcing. Radiative fluxes are defined as positive downward.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

4. Cloud adjustment and feedback

In this section we consider these issues in relation to the cloud changes in HadGEM2-ES. We will describe and examine in detail the adjustments and feedbacks and the dominant physical processes in operation. We also use clouds as a test for potentially important questions regarding the regional nature of these relationships. For example, how are adjustments characterized in terms of specific physical processes? Is the division between the adjustments and feedbacks valid at regional scales? Does the consistency between the global mean feedbacks in the ramp-up and ramp-down scenarios also apply to the feedbacks at the regional scale?

Note that the changes in global-mean cloud radiative effect are small compared to clear-sky fluxes (Figs. 13). However, the global results mask large regional cloud changes that compensate in the global mean (see below).

a. Global, land, and ocean cloud adjustments and feedbacks

We first consider the global-mean cloud changes defined by the International Satellite Cloud Climatology Project (ISCCP) simulator cloud type diagnostics (Klein and Jakob 1999; Webb et al. 2001). The ISCCP cloud types allow the split between changes in cloud amount and cloud optical depth to be examined in a straightforward manner and, because they are radiatively defined, the cloud type can be related to changes in top-of-atmosphere fluxes if desired (e.g., Zelinka et al. 2013).

To illustrate the relative contribution of cloud adjustments to cloud feedbacks we consider the changes over the last 15 years of the 1% ramp-up experiment. The total change in any quantity is considered to be the sum of the adjustment and the feedback; hence, we define the feedback term as the total change relative to the control less the (appropriately scaled) adjustment derived from the 4×CO2 fixed-SST experiment. The remaining feedback term is therefore more physically related to large-scale changes in surface temperature than simply considering the total cloud change because it is not contaminated by adjustments.

Table 1 shows the global-mean total, adjustment, and feedback terms for the ISCCP-defined high-, middle-, and low-level cloud amounts, together with the split between ocean and land grid points. The adjustment is negligible at the global-mean scale for both high- and low-level cloud but it contributes roughly one-third of the total change for middle-level cloud (Table 1). Separating into land and ocean shows that there is a significant cancellation between increases and decreases in different regions for high- and low-level cloud adjustments. For oceanic high-level cloud the magnitude of the adjustment is approximately 25% that of the feedback but the effects are of opposite sign. Over land, adjustment and feedback contribute almost equally to the total increase in high cloud. For low-level cloud there is a high degree of cancellation between adjustment (positive) and feedback (negative) over ocean, whereas over land adjustment dominates and the feedback is close to zero.

Table 1.

Global-mean cloud amount changes with respect to the control simulation at the end (years 125–140) of the 1% ramp-up simulation. The total changes (TOT) are broken down into the contributions from the adjustment (ADJ) and the feedback (F/B). Adjustments are determined from the fixed-SST experiment. The feedback contribution is estimated as the residual between the total change in the fully coupled simulation and the adjustment. Units are fractional coverage measured in percent.

Table 1.

The ISCCP cloud diagnostics allow us to divide the cloud changes into seven optical depth bins for each of the defined cloud-top pressure levels. For high, middle, and low cloud, over both land and ocean, the sign of adjustment is the same for almost all optical depth bins. This indicates that the adjustment is primarily associated with increases or decreases of cloud amount, rather than large-scale shifts of the optical depth distribution itself.

The largest adjustments occur for low-level cloud, and the detailed changes in the optical depth bins are shown in Table 2. Over ocean the adjustments are dominated by increases in medium thickness (stratocumulus) cloud whereas over land there are reductions in all optical depth bins. The low cloud feedback indicates a shift to optically thicker cloud over both land and ocean. Note that the high cloud feedback (not shown) also shows an increase in optical depth, but only for the optically thinnest (cirrus) cloud.

Table 2.

Breakdown of the adjustment and feedback contributions at the end (years 125–140) of the 1% ramp-up simulation to the total changes in global-mean low-level cloud as a function of the seven ISCCP visible optical depth (τ) categories. Results are shown separately for land and ocean. Units are fractional coverage measured in percent.

Table 2.

A feature of previous Hadley Centre climate models was the relatively large contribution of plant physiological forcing to the fast cloud adjustments (e.g., Doutriaux-Boucher et al. 2009; Dong et al. 2009). This results from the closing of plant stomata at higher CO2 concentrations and a consequent reduction in evapotranspiration—that is, a reduction in the flux of moisture from land to the atmosphere (Sellers et al. 1996). Here we use the additional fixed-SST experiments (4×CO2RAD and 4×CO2PHYS, described in section 2) to isolate this effect in HadGEM2-ES. Table 3 shows the breakdown of the global-mean cloud adjustments over land between the radiative and plant physiological effects. The CO2 radiative effect dominates the middle- and high-level cloud adjustments, but the plant physiological effect makes a nonnegligible contribution. For low-level cloud the physiological effect is by far the dominant effect, and this is discussed in more detail below.

Table 3.

Cloud adjustments averaged over land for high-, middle-, and low-level cloud at the end (years 125–140) of the 1% ramp-up simulation and their separation into radiative (RAD) and plant physiological effects (PHYS). The adjustments are estimated from the 4×CO2, 4×CO2RAD, and 4×CO2PHYS fixed-SST experiments. FULL indicates both radiative and physiological effects are present. The final column indicates that these two effects add approximately linearly at the global land scale (i.e., FULL-RAD ≈ PHYS). Units are fractional coverage measured in percent.

Table 3.

b. Closer examination of the cloud adjustments

1) Regional cloud adjustments

Figure 5 shows the geographical distributions of the ISCCP low-, middle-, and high-level cloud adjustments in the three main optical depth categories (thin, medium thickness, and thick cloud). High-level cloud adjustments are largest in the tropics and are generally positive over land and negative over ocean. The high cloud changes are of roughly comparable magnitude in all three optical depth categories, but there is some evidence for a thickening of high cloud over the warm pool and a thinning over Northern Hemisphere midlatitude land areas.

Fig. 5.
Fig. 5.

CO2-driven cloud adjustments in the nine basic ISCCP cloud categories at the end (years 125–140) of the 1% ramp-up simulation for (a)–(c) high-, (d)–(f) middle-, and (g)–(i) low-level clouds. The adjustments are estimated from the fixed-SST experiment. Units are fractional coverage measured in percent.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

The middle-level cloud adjustment is negative almost everywhere for all three optical depth categories. This explains the dominance of the middle-level cloud adjustment in the global means (Table 1). Low-level cloud adjustments are dominated by medium thickness cloud (referred to as “stratocumulus” by ISCCP), with large reductions over land and large increases over the oceans, the latter being particularly apparent in the well-known areas of marine stratocumulus cloud (e.g., Klein and Hartmann 1993). Similar results are also seen in the CMIP5 multimodel mean analysis of Zelinka et al. (2013).

Figure 6 shows the total (i.e., the sum over all optical depth bins) low- and high-level cloud adjustments, together with the contributions from the radiative and plant physiological effects. The high cloud changes are largest over the tropics and subtropics and are dominated by the radiative effect. The physiological effect is important over certain land areas (e.g., South America).

Fig. 6.
Fig. 6.

Total (a)–(c) high- and (d)–(f) low-level cloud adjustments at the end (years 125–140) of the 1% ramp-up simulation and their separation into radiative (RAD) and plant physiological effects (PHYS). The adjustments are estimated from the 4×CO2, 4×CO2RAD, and 4×CO2PHYS fixed-SST experiments described in the text. FULL indicates both radiative and physiological effects are present. Units are fractional coverage measured in percent.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

The low-level cloud adjustment shows a clear distinction between the dominance of the radiative effect (increased cloud) over the ocean and the physiological effect (reduced cloud) over land. There is some evidence for the “spreading” of the physiological effect over ocean areas, for example, off the west coast of South America, which results from the lower-tropospheric warming over land being advected over adjacent oceanic areas (Mitchell 1983; Dong et al. 2009; Kamae and Watanabe 2013). This is predominantly a radiative effect: the global-mean land surface temperature increase is 1.30 K whereas it is 0.87 and 0.36 K in the RAD and PHYS experiments, respectively. However, in this model the land warming owing to the physiological effect alone is large enough for the impact over ocean to be apparent even in the climatological annual mean.

In many regions there is an anticorrelation between the changes in these ISCCP satellite-equivalent high-level and low-level cloud amounts. We have examined the model cloud amounts (both large-scale and convective) and have confirmed that this anticorrelation is not an artifact of any covering or uncovering effect, for example, less high cloud revealing more low cloud or vice versa, but is indeed representative of the model’s true behavior.

2) Associated processes

Figure 7 shows the CO2 adjustments from the fixed-SST experiment of a selection of parameters related to the cloud changes. The land surface warming (Fig. 7a) is largest at middle latitudes and over South America (where there is a strong local influence of the plant physiological effect). Comparison of the net CRE adjustment (Fig. 7b) with Fig. 6 shows the dominant impact of low-level cloud changes on the net top-of-atmosphere radiation balance (the SW effect is much larger than the LW effect) and the strong top-of-atmosphere cancellation between the SW and LW impacts of high-level cloud. Although the plant physiological effect on the ERF is relatively small (Andrews et al. 2012c)—it is 13%, 1 W m−2 out of 7.5 W m−2, over land globally in our model—it is the dominant contribution to the CRE adjustment over land, particularly over the Northern Hemisphere midlatitudes (Table 4).

Fig. 7.
Fig. 7.

CO2-driven adjustments in various parameters related to the cloud changes at the end (years 125–140) of the 1% ramp-up simulation. The adjustments are estimated from the fixed-SST experiment. (bottom) The physiological component of the surface latent and sensible heating to adjustments from the 4×CO2PHYS fixed-SST experiment (negative indicates a reduction in flux from the surface to the atmosphere for the latent and sensible heat terms).

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

Table 4.

Change in net TOA radiative fluxes (W m−2) and CRE (W m−2) associated with CO2 effective radiative forcing (FULL, RAD, and PHYS) and associated adjustments at the end (years 125–140) of the 1% ramp-up simulation over global, tropical (30°N–30°S) and Northern Hemisphere midlatitude (30°–60°N) land areas.

Table 4.

The adjustment in lower tropospheric stability (LTS)—defined as the difference between the potential temperature at 700 hPa and the surface—(Fig. 7c) confirms a strong association between changes in low-level cloud and LTS over both land and ocean. Note that Wyant et al. (2012) and others (e.g., Watanabe et al. 2012) have used LTS as a compositing variable for examining low cloud changes in these types of experiment, but generally only over the tropical oceans. Similarly, Webb et al. (2013) found a strong correlation between changes in net CRE cloud adjustments and changes in lower tropospheric stability across the generation of models from phase 3 of the Coupled Model Intercomparison Project (CMIP3). LTS is related to the strength of the temperature inversion capping the boundary layer: increased LTS corresponds to a stronger inversion and increased low-level cloud (Klein and Hartmann 1993). Over the ocean the SSTs are fixed so that the increased warming aloft resulting from the increased CO2 leads to a more stable boundary layer and increased cloud (see also Wyant et al. 2012; Watanabe et al. 2012). In contrast, the land surface temperature is able to warm, the LTS decreases, and there is a reduction in low-level cloud amount.

Both surface latent heat (LH) and sensible heat (SH) fluxes (Figs. 7d,e) are reduced over the oceans, consistent with much earlier (e.g., Mitchell 1983) and more recent (Wyant et al. 2012) studies. Over land we see the importance of the plant physiological effect on both LH and SH in this model (Figs. 7f,g): stomatal closure leads to decreased evapotranspiration and consequent reductions (increases) in latent (sensible) heat fluxes (e.g., Boucher et al. 2009). The similarity between Figs. 7d,e and 7f,g shows that this is the dominant influence over land; only certain parts of Africa see a substantially larger response in LH and SH fluxes due to the CO2 radiative effect.

3) Boundary layer changes

The changes in low-level cloud and associated energy fluxes point to rapid responses to the CO2 increase in the atmospheric boundary layer. This can be examined further by looking at the changes in specific boundary layer “types” defined by the boundary layer mixing scheme (Lock et al. 2000). There are seven boundary layer types, with the classification at any grid point and at any time step depending on diagnoses made by the mixing scheme. The diagnosed type is then set to unity and the other six to zero: the fractional occurrence of each of the different types is then calculated by averaging over the relevant time interval (daily, monthly, seasonal mean, etc.). Following Martin et al. (2000) we here combine types 1 and 2 (stable), and types 4 and 5 (decoupled), with no significant loss of information in this context.

The 4×CO2 adjustments to these boundary layer types are shown in Fig. 8. Over the ocean the increase in tropical and subtropical stratocumulus is related to an increase in decoupled and, closer to the coasts, well-mixed boundary layers at the expense of the cumulus type. At higher latitudes in both hemispheres, but most prominently over the Southern Ocean, the increase in low-level cloud is associated with an increase in stable boundary layers, with the occurrence of all other types reducing. Changes in shear-driven boundary layers are predominantly confined to the Arctic and closely match the increase in low-level cloud seen in the RAD experiment. Over land the dominant feature is the increase in the well-mixed boundary layer type, corresponding to the decrease in both low-level cloud and latent heat flux, and the increase in sensible heat flux. All other boundary layer types tend to decrease over land. The increase in well-mixed boundary layers over land results mostly from the plant physiological effect and occurs at the expense of the cumulus type (not shown).

Fig. 8.
Fig. 8.

The CO2-driven adjustment of the boundary layer types defined by the boundary layer mixing scheme at the end (years 125–140) of the 1% ramp-up simulation. For clarity, the seven types used by the scheme have been reduced to five by combining the two stable types (1 and 2) and the two decoupled types (4 and 5). Units are fractional coverage measured in percent.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

Accompanying the changes to the low-level cloud and boundary layer mixing regimes are changes in the depth of the planetary boundary layer (PBL) itself (Table 5). Our results agree with both Wyant et al. (2012) and Watanabe et al. (2012) in suggesting a shallower PBL over the tropical oceans as part of the CO2 adjustment process. Using the model’s own boundary layer height diagnostic gives a reduction of ~30 m over the tropical ocean. If we use the same method as Wyant et al. (2012)—who estimate the boundary layer height from a relative humidity threshold—we obtain a reduction of 62 m, in closer agreement with their estimate of 80 m over the most stable regions. Clearly there is some sensitivity to the method used in determining the boundary layer height and changes to it, and the models themselves may simulate different magnitudes of this change.

Table 5.

Change in mean boundary layer height (m) from the 4×CO2, 4×CO2RAD, and 4×CO2PHYS fixed-SST experiments (i.e., as adjustments to 4 × CO2) over tropical (30°N–30°S) ocean/land and Northern Hemisphere midlatitude (30°–60°N) land areas.

Table 5.

Consistent with the changes to the boundary layer mixing regimes and surface heat fluxes, there is an increase in PBL depth over tropical land areas that is dominated by the physiological effect, whereas at midlatitudes the radiative and physiological effects are of comparable magnitude (Table 5). The geographical distributions of the PBL depth changes (not shown) closely match those in low-level cloud, boundary layer types, and surface heat fluxes.

It is important to interpret these annual mean changes carefully, particularly the increases and decreases between the different boundary layer types, owing to both seasonal and diurnal variations. Land areas during summer, for example, tend to be mainly characterized by a stable boundary layer at night and a well-mixed regime during the day.

Finally, it is also important to keep in mind the large-scale dynamical context and the changes in circulation patterns associated with the CO2 increase (Mitchell 1983). The responses of the moisture and surface fluxes as well as the temperature increase over land relative to fixed SSTs leads to reductions in surface pressure over almost all land areas in the annual mean (not shown). Closer examination suggests that these reductions and the corresponding increases in surface pressure over midlatitude oceans are very similar to those described in Mitchell (1983). This suggests that the large-scale dynamical response is very robust—at least in successive generations of Hadley Centre models—and is in agreement with other modeling studies (e.g., Lambert et al. 2011).

c. Regional pattern of cloud feedbacks

A similar approach can be used to diagnose and examine the contribution from cloud feedbacks. Here we define the feedback contribution as the difference between the total change and the adjustment near the end of the 1% ramp-up simulation, that is, as above for the global means but for each grid box. The high- and low-level cloud feedback contributions are shown in Fig. 9.

Fig. 9.
Fig. 9.

Cloud feedback contribution to the ISCCP-defined (a)–(c) high- and (d)–(f) low-level cloud types at the end (years 125–140) of the 1% ramp-up simulation. These are estimated as a residual between the fully coupled AOGCM cloud changes and the adjustments estimated from the fixed-SST experiment. Units are fractional coverage measured in percent.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

For high cloud the changes in the three basic visible optical depth categories are generally of the same sign (increased cloud), suggesting little evidence for wide-scale high-cloud optical depth feedbacks in this model, consistent with what was seen in the global mean. Reductions in high cloud are restricted to certain tropical areas. There is no dominant pattern of association between the high cloud feedback and adjustments.

The low-level cloud changes indicate a large optical depth feedback (cloud becoming thicker) over the Southern Ocean and the Arctic. There are reductions in low cloud over the main marine stratocumulus regions, including the eastern equatorial Pacific, which are greatest (in both magnitude and geographical extent) over the northeastern Pacific. Over land the low cloud feedback is generally quite small, so the total cloud response is dominated by the adjustment. These patterns are also evident in the CMIP5 multimodel mean analysis of Zelinka et al. (2013), who estimated cloud feedback changes from regressing the local cloud anomalies against global-mean ΔT in abrupt 4×CO2 experiments. The consistency of the results suggests our method of estimating patterns of cloud feedback as a residual is in agreement with other methods.

The feedback response of the boundary layer mixing regimes (Fig. 10) indicates that the reductions in low-level cloud are generally linked to a transition from decoupled to cumulus regimes, most prominently over the tropical and subtropical Pacific and Atlantic Oceans. The responses off the west coast of South America are likely influenced by local minima in the SST warming that occurs in this region (see section 5 below). Further details of subtropical low cloud feedback mechanisms using the atmospheric component of HadGEM2 are given in Webb and Lock (2013).

Fig. 10.
Fig. 10.

The boundary-layer-type feedback contributions at the end (years 125–140) of the 1% ramp-up simulation. These are estimated as a residual between the fully coupled AOGCM cloud changes and the adjustments estimated from the fixed-SST experiment. Units are fractional coverage measured in percent.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

d. Vertical structure of the cloud feedbacks and adjustments

Here we briefly consider the vertical distribution of the cloud responses by looking at zonal mean–height cross sections of the adjustments and feedbacks over land and ocean on the model’s full vertical resolution (38 levels). This provides both greater detail and a check on the analysis based on the ISCCP cloud diagnostics.

The cloud feedback (Fig. 11) is very similar over land and ocean (cf. Zelinka et al. 2013). This structure corresponds to the “classical” response of cloud to surface warming described in much earlier studies (e.g., Mitchell and Ingram 1992): at tropical and middle latitudes cloud decreases in the lower and middle troposphere and increases aloft, with the division between the decreases and increases matching the zonal variation of the tropopause. This is similarly seen in other CMIP5 models (e.g., Zelinka et al. 2013; Tomassini et al. 2013). At higher latitudes, poleward of ~60° in both hemispheres, cloud increases throughout the depth of the troposphere. Unlike Mitchell and Ingram (1992), there is an increase in cloud near the surface over the tropical oceans, which is almost certainly linked to the far greater resolution of the PBL in this model. This is opposite to the behavior reported by Colman and McAvaney (2011) and is not apparent in the ensemble mean changes shown by Zelinka et al. (2013).

Fig. 11.
Fig. 11.

The zonal-mean cloud amount adjustment and feedback changes over (a),(d) ocean and (b),(e) land with respect to the control at the end (years 125–140) of the 1% ramp-up experiment. (c),(f) Also shown are the adjustments over land in the RAD and PHYS experiments. The cloud amount changes are shown on the model’s native vertical grid from the surface up to a height of 18 km. Units are fractional coverage measured in percent. The vertical axis is height in meters.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

In common with many GCMs, cloud fraction in HadGEM2-ES is diagnosed from relative humidity, so changes in cloud and RH should generally be quite strongly correlated (Mitchell and Ingram 1992; Colman and McAvaney 2011). The zonal mean–height cross sections of the adjustments to RH over land and ocean (Fig. 12) are closely related to the cloud fraction changes in Fig. 11 as expected, the lack of a perfect match-up being indicative of the nonlinearity in the specified RH–cloud relationship.

Fig. 12.
Fig. 12.

CO2-driven adjustments in relative humidity (RH) over ocean and land and its breakdown into components due to changes in specific humidity (q) and temperature (T) as described in the text. The bottom row shows the plant physiological impact on the RH adjustment and its components over land. Changes are estimated from the relevant fixed-SST experiment at the end (years 125–140) of the 1% ramp-up experiment. The vertical axis is height in meters.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

We decompose the RH adjustments ΔRH into components due to the changes in temperature ΔT and specific humidity Δq following the method of Kamae and Watanabe (2012). This allows us to write
eq1
There is a high degree of compensation between the drying due to increasing temperature and the moistening due to increasing specific humidity (Fig. 12). Over both the ocean and land we can identify areas where either ΔRH(ΔT) or ΔRH(Δq) dominates, or indeed where they reinforce (e.g., the midtroposphere in the tropics). A feature over land is the importance of the plant physiological effect at low and midlatitudes in the boundary layer (Figs. 12g–i). Here the strong reinforcement between the warming effect and the reduced moisture leads to the largest reductions in RH seen anywhere: this is strongly correlated with the reduced boundary layer cloud fraction and associated changes (described above) in these areas.

5. Time variation of feedbacks and the pattern of warming

In section 3 we showed how the conceptual framework for estimating global mean climate feedbacks was valid in both ramp-up and ramp-down phases of our experiments. We found that by accounting for the forcing/adjustments we could explain the apparent asymmetric (or hysteresis-type) behavior of some parameters between the two phases. Here we consider if this also applies at the regional scale.

Figure 13 shows the geographical distributions of high-level and low-level cloud feedbacks when the CO2 concentration is double that of the control in both ramp-up and ramp-down phases of the experiment (i.e., 20-yr averages centered on years 70 and 210 relative to the start). The cloud feedbacks are calculated as before by removing the appropriately scaled adjustments from the fixed-SST experiment and then normalizing by the global mean surface temperature change relative to the control (2.5 K in the ramp up and 3.8 K in the ramp down).

Fig. 13.
Fig. 13.

(a),(b) High- and (c),(d) low-level cloud amount feedbacks at the point of CO2 doubling (year 70 and 210, relative to the control) during the ramp-up and ramp-down phases of the simulation. These are estimated from the cloud amount changes in the fully coupled AOGCM minus the corresponding adjustment from the fixed-SST experiment, then normalized by the global-mean temperature change.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

Although not identical, the patterns are remarkably similar for both high- and low-level clouds. This suggests that the global-mean behavior seen in section 3—that is, the reversibility of radiative feedbacks—also applies to the geographical patterns of the feedbacks. This is even the case for cloud feedbacks, which are generally thought to be one of the most complex and uncertain processes operating in climate models.

Note that the feedbacks calculated in this way are “the local contribution” to the global-mean feedbacks; that is, they indicate the local cloud response with respect to the global-mean surface temperature change rather than the local surface temperature change (Boer and Yu 2003). Armour et al. (2013) have recently suggested that the global-mean feedbacks should instead be estimated from a fixed geographical distribution of the local feedbacks (i.e., the local response to the local surface temperature change) convolved with the time-varying pattern of surface temperature change.

We investigate how the pattern of surface warming evolves in our experimental design during the ramp down in Figs. 14a–c, which shows how local surface temperature change varies with global-mean ΔT (relative to the control) for the ramp-down phase of the experiment. This is shown for three consecutive periods (covering equal declines in forcing) after the CO2 has started to decrease. The plots are calculated from the regression slope of annual-mean local ΔT against annual global-mean ΔT during each period.

Fig. 14.
Fig. 14.

The variation of local (a)–(c) surface temperature change (K K−1) and (d)–(f) clear-sky absorbed SW radiation (W m−2 K−1) over the Southern Hemisphere high latitudes, with global-mean surface temperature change, for three consecutive periods of equal decline in forcing during the ramp-down phase. The plots are calculated from the slope of the regression of local annual-mean ΔT, or clear-sky absorbed SW radiation, against global annual-mean ΔT during the relevant period. The sign has been reversed, so that in (a)–(c) negative values indicate that local surface temperature is declining, in line with the global mean. Negative values in (d)–(f) indicate that absorbed SW radiation is declining, in line with the global-mean cooling.

Citation: Journal of Climate 27, 4; 10.1175/JCLI-D-13-00421.1

The local temperature change follows the direction of the global mean and declines in the early part of the ramp-down phase over the majority of the globe (Fig. 14a). The exceptions are the southeastern tropical Pacific and the Southern Ocean, which continue to warm for a few decades after the CO2 reversal. After a few decades these regions then begin to cool and local temperature change scales with the global mean almost everywhere (Figs. 14b,c).

If local warming does not scale with global-mean ΔT, then this may have implications for estimating feedbacks and their time variation (Armour et al. 2013). We illustrate this using the relationship between global and local surface temperature change, sea ice, and SW radiation in the Southern Ocean (see also Armour et al. 2011). Figures 14d–f show how the corresponding change in absorbed clear-sky SW radiation during the same periods of the ramp down (calculated as for surface temperature) varies with global-mean ΔT. As the globe is cooling, we expect to see increased sea ice and less SW absorption owing to the associated increase in surface albedo. This is indeed what happens during most of the ramp down (Fig. 14e,f) except during the early period where the opposite occurs (Fig. 14d). In this period the globe is cooling but the sea ice is still decreasing (and consequently absorbed SW radiation is increasing). This counterintuitive result is reconciled when one considers that it is local temperature change in this region that is “unusual”; that is, the Southern Ocean is still warming despite global cooling (Fig. 14a). This highlights the dangers of expressing feedbacks using global-mean ΔT without also considering the local physical processes involved and how they vary in time.

6. Summary and discussion

We have used a suite of fully coupled and atmosphere-only (fixed SST) experiments with HadGEM2-ES to diagnose the effective radiative forcing, rapid adjustments, feedbacks, and heat uptake in a 1% yr−1 CO2 increase scenario, followed by a 1% yr−1 CO2 decrease. We have focused on the applicability of simple conceptual models under a transient reversibility scenario and the role of cloud adjustments and associated processes.

We have shown that an abrupt 4×CO2 fixed-SST experiment—linearly scaled in time—can be used to represent the CO2 effective radiative forcing and adjustments in the transient CO2 experiment. Once these forcing and adjustments have been removed from the fully coupled AOGCM output, it is found that the global radiation balance and its components—and by implication the radiative feedbacks—scale linearly with global-mean surface temperature change, ΔT. We have provided a general method for separating forcing and feedback in a transient scenario where forcing and feedback evolve together. Tomassini et al. (2013) looked at radiative feedbacks in the CMIP5 1% CO2 fully coupled ramp-up simulations but could not clearly separate adjustment and feedback. We found that the feedback parameters are reversible under this idealized mitigation scenario. This even applied to the regional distribution of cloud feedback, which is regarded as one of the least well understood feedback processes.

The total cloud response to CO2 in the transient simulation is found to depend on both the adjustments and the feedbacks. At the global scale the cloud adjustment is related to changes in cloud amount, with no indication of any large-scale change in cloud optical depth, in agreement with Colman and McAvaney (2011) who used a different model, although in contrast to Zelinka et al. (2013) who found small optical depth changes associated with adjustments in their multimodel-mean cloud changes. The cloud feedback indicates changes in both cloud amount and optical depth, with the latter—a thickening—particularly apparent for low-level cloud. In HadGEM2-ES the plant physiological forcing plays a significant role in determining the cloud adjustment and is the dominant contribution to the low-level cloud changes over land. A reduction in boundary layer clouds in response to elevated CO2 levels has recently been highlighted in the process modeling study of Vilà-Guerau del Arellano et al. (2012). Here we have presented a more complete account of the cloud response to plant physiological forcing than previous GCM studies (e.g., Boucher et al. 2009; Cao et al. 2010).

Regionally, cloud feedbacks can compensate or combine with the adjustments to either dampen or enhance the total cloud changes. Low-level cloud adjustments are found to be associated with changes in surface turbulent fluxes and lower tropospheric stability. Over land, the physiological forcing reduces the latent heat flux to the atmosphere and increases the sensible heat flux. This drying and warming of the boundary layer is associated with changes in the frequency of occurrence of different boundary layer mixing regimes and an increase in the depth of the boundary layer itself. Over the ocean the CO2 adjustment leads to increases in tropical and subtropical stratocumulus cloud and a shallowing of the boundary layer.

Note that correctly separating cloud adjustments and feedback—or forcing and feedback more generally—is important for correctly predicting transient climate change (Gregory and Webb 2008). However, in situations where F = ρΔT is appropriate (as we have shown in these experiments) the separation of forcing and feedback has less practical value, as F and ΔT covary. It becomes more important when the scenario changes, or to understand the underlying physical processes driving climate change.

We have shown that a constant ratio between radiative forcing and global-mean surface temperature change (the climate resistance ρ) is a good approximation during both the ramp-up or ramp-down phases of the experiment. The exceptions are in the early periods of both phases, when the forcing is changing but global-mean temperature takes a few decades to respond. We found that ρ is larger during the ramp-down period than the ramp up due to differences in ocean heat uptake efficiency: surface temperatures do not decline as quickly as they increase.

Linearity of the simple forcing–response relationships cannot be expected to hold in all circumstances. For example, some feedback processes may have their own time scales not closely tied with ΔT, such as changes in vegetation, ice sheets, and soil moisture. In addition, changing patterns of surface warming could either amplify or dampen local feedback effects and give rise to time variation in climate sensitivity parameters (Senior and Mitchell 2000; Armour et al. 2013). The eastern tropical Pacific and the Southern Ocean—regions important for cloud feedback and the global energy balance—have been identified as regions where local warming does not scale well with global ΔT (Senior and Mitchell 2000; Williams et al. 2008; Held et al. 2010; Li et al. 2013).

We examined the transient evolution of the pattern of surface warming during the ramp-down phase of the experiment and found some regional anomalies. During the first few decades of the ramp down most of the globe cools immediately once the CO2 increase reverses. The exceptions are the Southern Ocean and eastern tropical Pacific, which continue to warm for a few decades before cooling in line with global temperatures. We have demonstrated how feedback analysis techniques in these regions using global-mean ΔT could give rise to time-varying sea ice and SW radiative feedback parameters.

The above results suggest that, while it might be necessary to consider the possible time variation of feedbacks when normalized by global-mean ΔT in these idealized experiments, such behavior may well be restricted to relatively short time periods, for example, following a reversal in forcing, and confined to certain geographical locations. This implies that our relatively simple conceptual framework for analyzing the global-mean relationships between forcing and feedbacks is generally applicable, at least to first order, even at the regional scale. Furthermore, “unusual” or “anomalous” behavior—in either time or space—might usefully be considered as deviations from this simple framework rather than necessitating the design of a more complex analysis method. On the other hand, we have highlighted the dangers and limitations of interpreting simple conceptual forcing–response frameworks applied to transient fully coupled scenarios, especially in the presence of a nonconstant pattern of warming.

Acknowledgments

This work was supported by the Joint DECC/Defra Met Office Hadley Centre Climate Programme (GA01101). We thank two reviewers for their positive and constructive reviews.

REFERENCES

  • Andrews, T., , and P. M. Forster, 2008: CO2 forcing induces semi-direct effects with consequences for climate feedback interpretations. Geophys. Res. Lett., 35, L04802, doi:10.1029/2007GL032273.

    • 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.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., , J. M. Gregory, , P. M. Forster, , and M. J. Webb, 2012a: Cloud adjustment and its role in CO2 radiative forcing and climate sensitivity: A review. Surv. Geophys., 33, 619635, doi:10.1007/s10712-011-9152-0.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., , J. M. Gregory, , M. J. Webb, , and K. E. Taylor, 2012b: Forcing, feedbacks and climate sensitivity in CMIP5 coupled atmosphere–ocean climate models. Geophys. Res. Lett., 39, L09712, doi:10.1029/2012GL051607.

    • Search Google Scholar
    • Export Citation
  • Andrews, T., , M. A. Ringer, , M. Doutriaux-Boucher, , M. J. Webb, , and W. J. Collins, 2012c: Sensitivity of an Earth system climate model to idealized radiative forcing. Geophys. Res. Lett., 39, L10702, doi:10.1029/2012GL051942.

    • Search Google Scholar
    • Export Citation
  • Armour, K. C., , I. Eisenman, , E. Blanchard-Wrigglesworth, , K. E. McCusker, , and C. M. Bitz, 2011: The reversibility of sea ice loss in a state-of-the-art climate model. Geophys. Res. Lett., 38, L16705, doi:10.1029/2011GL048739.

    • Search Google Scholar
    • Export Citation
  • Armour, K. C., , C. M. Bitz, , and G. H. Roe, 2013: Time-varying climate sensitivity from regional feedbacks. J. Climate, 26, 4518–4534.

  • Boer, G. J., , and B. Yu, 2003: Climate sensitivity and response. Climate Dyn., 20, 415429.

  • 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.

    • Search Google Scholar
    • Export Citation
  • Boucher, O., and Coauthors, 2012: Reversibility in an Earth system model in response to CO2 concentration changes. Environ. Res. Lett., 7, 024013, doi:10.1088/1748-9326/7/2/024013.

    • Search Google Scholar
    • Export Citation
  • Bouttes, N., , J. M. Gregory, , and J. A. Lowe, 2013: The reversibility of sea level rise. J. Climate, 26, 25022513.

  • 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.

    • Search Google Scholar
    • Export Citation
  • Cao, L., , G. Bala, , and K. Caldeira, 2011: Why is there a short-term increase in global precipitation in response to diminished CO2 forcing? Geophys. Res. Lett., 38, L06703, doi:10.1029/2011GL046713.

    • Search Google Scholar
    • Export Citation
  • Chadwick, R., , P. Wu, , P. Good, , and T. Andrews, 2013: Asymmetries in tropical rainfall and circulation patterns in idealised CO2 removal experiments. Climate Dyn., 40, 295316, doi:10.1007/s00382-012-1287-2.

    • Search Google Scholar
    • Export Citation
  • Collins, W. J., and Coauthors, 2011: Development and evaluation of an Earth-System model—HadGEM2. Geosci. Model Dev., 4, 10511075, doi:10.5194/gmd-4-1051-2011.

    • Search Google Scholar
    • Export Citation
  • Colman, R., , and B. McAvaney, 2011: On tropospheric adjustment to forcing and climate feedbacks. Climate Dyn., 36, 16491658, doi:10.1007/s00382-011-1067-4.

    • Search Google Scholar
    • Export Citation
  • Danabasoglu, G., , and P. R. Gent, 2009: Equilibrium climate sensitivity: Is it accurate to use a slab ocean model? J. Climate, 22, 24942499.

    • Search Google Scholar
    • Export Citation
  • Dong, B., , J. M. Gregory, , and R. T. Sutton, 2009: Understanding land–sea warming contrast in response to increasing greenhouse gases. Part I: Transient adjustment. J. Climate, 22, 30793097.

    • Search Google Scholar
    • Export Citation
  • Doutriaux-Boucher, M., , M. J. Webb, , J. M. Gregory, , and O. Boucher, 2009: Carbon dioxide induced stomatal closure increases radiative forcing via a rapid reduction in low cloud. Geophys. Res. Lett., 36, L02703, doi:10.1029/2008GL036273.

    • Search Google Scholar
    • Export Citation
  • Forster, P. M., , and K. Taylor, 2006: Climate forcings and climate sensitivities diagnosed from coupled climate model integrations. J. Climate, 19, 61816194.

    • Search Google Scholar
    • Export Citation
  • Forster, P. M., , T. Andrews, , P. Good, , J. M. Gregory, , L. S. Jackson, , and M. Zelinka, 2013: Evaluating adjusted forcing and model spread for historical and future scenarios in the CMIP5 generation of climate models. J. Geophys. Res., 118, 1139–1150, doi:10.1002/jgrd.50174.

    • Search Google Scholar
    • Export Citation
  • Gates, W. L., , K. H. Cook, , and M. E. Schlesinger, 1981: Preliminary analysis of experiments on the climatic effects of increased CO2 with an atmospheric general circulation model and a climatological ocean. J. Geophys. Res., 86 (C7), 63856393.

    • Search Google Scholar
    • Export Citation
  • Good, P., , J. M. Gregory, , and J. A. Lowe, 2011: A step-response simple climate model to reconstruct and interpret AOGCM projections. Geophys. Res. Lett., 38, L01703, doi:10.1029/2010GL045208.

    • Search Google Scholar
    • Export Citation
  • Good, P., , J. M. Gregory, , J. A. Lowe, , and T. Andrews, 2013: Abrupt CO2 experiments as tools for predicting and understanding CMIP5 representative concentration pathway projections. Climate Dyn., 40, 10411053, doi:10.1007/s00382-012-1410-4.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., 2000: Vertical heat transports in the ocean and their effect on time-dependent climate change. Climate Dyn., 16, 501515.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., , and J. F. B. Mitchell, 1997: The climate response to CO2 of the Hadley Centre coupled AOGCM with and without flux adjustment. Geophys. Res. Lett., 24, 19431946.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., , and P. Forster, 2008: Transient climate response estimated from radiative forcing and observed temperature change. J. Geophys. Res., 113, D23105, doi:10.1029/2008JD010405.

    • Search Google Scholar
    • Export Citation
  • Gregory, J. M., , and M. J. Webb, 2008: Tropospheric adjustment induces a cloud component in CO2 forcing. J. Climate, 21, 5871.

  • Gregory, J. M., and Coauthors, 2004: A new method for diagnosing radiative forcing and climate sensitivity. Geophys. Res. Lett., 31, L03205, doi:10.1029/2003GL018747.

    • Search Google Scholar
    • Export Citation
  • Hansen, J., and Coauthors, 2005: Efficacy of climate forcings. J. Geophys. Res., 110, D18104, doi:10.1029/2005JD005776.

  • Hasselmann, K., , R. Sausen, , E. Maier-Reimer, , and R. Voss, 1993: On the cold start problem in transient simulations with coupled atmosphere–ocean models. Climate Dyn., 9, 5361.

    • Search Google Scholar
    • Export Citation
  • Held, I. M., and Coauthors, 2010: Probing the fast and slow components of global warming by returning abruptly to preindustrial forcing. J. Climate, 23, 24182427.

    • Search Google Scholar
    • Export Citation
  • Jiang, J. H., and Coauthors, 2012: Evaluation of cloud and water vapor simulations in CMIP5 climate models using NASA “A-Train” satellite observations. J. Geophys. Res., 117, D14105, doi:10.1029/2011JD017237.

    • Search Google Scholar
    • Export Citation
  • Kamae, Y., , and M. Watanabe, 2012: On the robustness of tropospheric adjustment in CMIP5 models. Geophys. Res. Lett., 39, L23808, doi:10.1029/2012GL054275.

    • Search Google Scholar
    • Export Citation
  • Kamae, Y., , and M. Watanabe, 2013: Tropospheric adjustment to increasing CO2: Its timescale and the role of land–sea contrast. Climate Dyn., 41, 3007–3024, doi:10.1007/s00382-012-1555-1.

    • Search Google Scholar
    • Export Citation
  • Keen, A. B., , and J. M. Murphy, 1997: Influence of natural variability and the cold start problem on the simulated transient response to increasing CO2. Climate Dyn., 13, 847864, doi:10.1007/s003820050201.

    • Search Google Scholar
    • Export Citation
  • Klein, S. A., , and D. L. Hartmann, 1993: The seasonal cycle of low stratiform clouds. J. Climate, 6, 15871606.

  • Klein, S. A., , and C. Jakob, 1999: Validation and sensitivities of frontal clouds simulated by the ECMWF model. Mon. Wea. Rev., 127, 25142531.

    • Search Google Scholar
    • Export Citation
  • Kuhlbrodt, T., , and J. M. Gregory, 2012: Ocean heat uptake and its consequences for the magnitude of sea level rise and climate change. Geophys. Res. Lett., 39, L18608, doi:10.1029/2012GL052952.

    • Search Google Scholar
    • Export Citation
  • Lambert, F. H., , M. J. Webb, , and M. M. Joshi, 2011: The relationship between land–ocean surface temperature contrast and radiative forcing. J. Climate, 24, 32393256.

    • Search Google Scholar
    • Export Citation
  • Li, C., , J.-S. V. Storch, , and J. Marotzke, 2013: Deep-ocean heat uptake and equilibrium climate response. Climate Dyn., 40, 1071–1086, doi:10.1007/s00382-012-1350-z.

    • Search Google Scholar
    • Export Citation
  • Lock, A. P., and Coauthors, 2000: A new boundary layer mixing scheme. Part I: Scheme description and single-column model tests. Mon. Wea. Rev., 128, 31873199.

    • Search Google Scholar
    • Export Citation
  • Long, D. J., , and M. Collins, 2013: Quantifying global climate feedbacks, responses and forcing under abrupt and gradual CO2 forcing. Climate Dyn., 41, 2471–2479, doi:10.1007/s00382-013-1677-0.

    • Search Google Scholar
    • Export Citation
  • Martin, G. M., and Coauthors, 2000: A new boundary layer mixing scheme. Part II: Tests in climate and mesoscale models. Mon. Wea. Rev., 128, 32003217.

    • Search Google Scholar
    • Export Citation
  • Martin, G. M., and Coauthors, 2011: The HadGEM2 family of Met Office Unified Model Climate configurations. Geosci. Model Dev., 4, 723757, doi:10.5194/gmd-4-723-2011.

    • Search Google Scholar
    • Export Citation
  • Mitchell, J. F. B., 1983: The seasonal response of a general circulation model to changes in CO2 and sea temperatures. Quart. J. Roy. Meteor. Soc., 109, 113152.

    • Search Google Scholar
    • Export Citation
  • Mitchell, J. F. B., , and W. J. Ingram, 1992: Carbon dioxide and climate: Mechanisms of changes in cloud. J. Climate, 5, 521.

  • Myhre, G., , E. J. Highwood, , K. P. Shine, , and F. Stordal, 1998: New estimates of radiative forcing due to well mixed greenhouse gases. Geophys. Res. Lett., 25, 27152718.

    • Search Google Scholar
    • Export Citation
  • Raper, S. C. B., , J. M. Gregory, , and R. J. Stouffer, 2002: The role of climate sensitivity and ocean heat uptake on AOGCM transient temperature response. J. Climate, 15, 124130.

    • Search Google Scholar
    • Export Citation
  • Sellers, P. J., and Coauthors, 1996: Comparison of radiative and physiological effects of doubled atmospheric CO2 on climate. Science, 271, 14021406.

    • Search Google Scholar
    • Export Citation
  • Senior, C. A., , and J. F. B. Mitchell, 2000: The time-dependence of climate sensitivity. Geophys. Res. Lett., 21, 26852688, doi:10.1029/2000GL011373.

    • Search Google Scholar
    • Export Citation
  • Shine, K. P., , J. Cook, , E. J. Highwood, , and M. J. Joshi, 2003: An alternative to radiative forcing for estimating the relative importance of climate change mechanisms. Geophys. Res. Lett., 30, 2047, doi:10.1029/2003GL018141.

    • Search Google Scholar
    • Export Citation
  • Tomassini, L., and Coauthors, 2013: The respective roles of surface temperature driven feedbacks and tropospheric adjustment to CO2 in CMIP5 transient climate simulations. Climate Dyn., 41, 3103–3126, doi:10.1007/s00382-013-1682-3.

    • Search Google Scholar
    • Export Citation
  • Vial, J., , J.-L. Dufresne, , and S. Bony, 2013: On the interpretation of inter-model spread in CMIP5 climate sensitivity estimates. Climate Dyn., 41, 3339–3362, doi:10.1007/s00382-013-1725-9.

    • Search Google Scholar
    • Export Citation
  • Vilà-Guerau de Arellano, J., , C. C. van Heerwaarden, , and J. Lelieveld, 2012: Modelled suppression of boundary-layer clouds by plants in a CO2-rich atmosphere. Nat. Geosci., 5, 701704, doi:10.1038/ngeo1554.

    • Search Google Scholar
    • Export Citation
  • Watanabe, M., , H. Shiogama, , M. Yoshimori, , T. Ogura, , T. Yokohata, , H. Okamoto, , S. Emori, , and M. Kimoto, 2012: Fast and slow timescales in the tropical low-cloud response to increasing CO2 in two climate models. Climate Dyn., 39, 16271641, doi:10.1007/s00382-011-1178-y.

    • Search Google Scholar
    • Export Citation
  • Webb, M. J., , and A. P. Lock, 2013: Coupling between subtropical cloud feedback and the local hydrological cycle in a climate model. Climate Dyn., 41, 1923–1939, doi:10.1007/s00382-012-1608-5.

    • Search Google Scholar
    • Export Citation
  • Webb, M. J., , C. A. Senior, , S. Bony, , and J.-J. Morcrette, 2001: Combining ERBE and ISCCP data to assess cloud in the Hadley Centre, ECMWF and LMD atmospheric climate models. Climate Dyn., 17, 905922.

    • Search Google Scholar
    • Export Citation
  • Webb, M. J., , F. H. Lambert, , and J. M. Gregory, 2013: Origins of differences in climate sensitivity, forcing and feedback in climate models. Climate Dyn., 40, 677–707, doi:10.1007/s00382-012-1336-x.

    • Search Google Scholar
    • Export Citation
  • Williams, K. D., , W. J. Ingram, , and J. M. Gregory, 2008: Time variation of effective climate sensitivity in GCMs. J. Climate, 21, 50765090.

    • Search Google Scholar
    • Export Citation
  • Winton, M., , K. Takahashi, , and I. M. Held, 2010: Importance of ocean heat uptake efficacy to transient climate change. J. Climate, 23, 23332344.

    • Search Google Scholar
    • Export Citation
  • Wu, P., , R. Wood, , J. Ridley, , and J. Lowe, 2010: Temporary acceleration of the hydrological cycle in response to a CO2 rampdown. Geophys. Res. Lett., 37, L12705, doi:10.1029/2010GL043730.

    • Search Google Scholar
    • Export Citation
  • Wyant, M. C., , C. S. Bretherton, , P. N. Blossey, , and M. Khairoutdinov, 2012: Fast cloud adjustment to increasing CO2 in a superparameterized climate model. J. Adv. Model. Earth Syst., 4, M05001, doi:10.1029/2011MS000092.

    • Search Google Scholar
    • Export Citation
  • Zelinka, M. D., , S. A. Klein, , K. E. Taylor, , T. Andrews, , M. J. Webb, , J. M. Gregory, , and P. M. Forster, 2013: Contributions of different cloud types to feedbacks and rapid adjustments in CMIP5. J. Climate, 26, 5007–5027.

    • Search Google Scholar
    • Export Citation
Save