Rapid Adjustments of Cloud and Hydrological Cycle to Increasing CO2: a Review
- 1.1k Downloads
Rapid cloud response to instantaneous radiative perturbation in the troposphere due to change in CO2 concentration is called cloud adjustment. Cloud adjustment develops on a short timescale because it is separated from surface temperature-mediated changes in cloud. Adjustments in cloud and tropospheric properties including the hydrological cycle have attracted considerable attention because of their importance in the interpretation of mechanisms of climate change and the identification of sources of uncertainty in climate sensitivity. Modeling studies have clearly revealed that major aspects of the tropospheric adjustment including the warming and drying of the troposphere, associated reduction of low cloud and increasing shortwave cloud radiative forcing, downward shift of the low-cloud layer, and slowdown of the global hydrological cycle, are common among many climate model simulations. Combinations of model simulations with realistic and idealized aqua-planet settings have helped demonstrate the roles of land and robust aspects of the tropospheric adjustment.
KeywordsClimate sensitivity Effective radiative forcing Instantaneous radiative forcing Land–sea thermal contrast Low cloud Hydrological cycle
Equilibrium climate sensitivity (ECS), defined as global-mean surface air temperature (SAT) change due to a doubling of atmospheric CO2 concentration, is a metric commonly used to explore the Earth’s climate response to externally induced forcing . Despite the considerable efforts expended in quantifying and constraining ECS, its range of uncertainty has not been reduced effectively during these three decades [1, 2, 3]. The response of cloud properties to climate warming is one of the largest contributors to ECS uncertainty . Increasing CO2 results in a radiative perturbation that affects cloud properties directly [3, 4, 5, 6]. This direct cloud response to increasing levels of CO2 is independent of the global-mean SAT increase and it should be distinguished from SAT-mediated cloud changes. It has been derived from atmosphere–ocean coupled general circulation models (AOGCMs) that this cloud response, called “cloud adjustment,” generates radiative forcing of about 0.5 W m−2 per the doubling of CO2 concentration [7••]. This radiative effect is of great importance to the effective radiative forcing (ERF), determined by a radiative imbalance at the top of the atmosphere (TOA) without any changes in global-mean SAT under the conditions of doubled CO2 [1, 8], and it has a non-negligible contribution to the ECS [7••, 9•, 10•].
The rapid adjustments of cloud and tropospheric properties can be interpreted as part of the direct climate response to increasing CO2. Here, the “direct response” means the response to radiative forcing without any perturbations in sea surface temperature (SST) and sea ice. The roles of the direct climate response to radiative forcing for some aspects of climate change have been examined thoroughly (e.g., land–sea temperature contrast, atmospheric circulation [11, 12, 13, 14, 15, 16•, 17, 18, 19•, 20••, 21, 22, 23], and tropical cyclone activity [24, 25, 26]). Quantifying the radiative effect of the tropospheric cloud adjustment is essential for better understanding of the uncertainty in the ECS [7••, 27, 28, 29]. It has also been widely accepted that the role of the direct response to CO2 increase is substantial for the global hydrological cycle [19•, 20••, 21, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40], including evapotranspiration over land [5, 41, 42•].
The mechanism of tropospheric adjustment to increasing CO2 has been examined using two methodologies: the fixed-SST method (Hansen method) and the regression method (Gregory method). In the fixed-SST method, ERF can be evaluated by the difference between two equilibrium states simulated in atmospheric general circulation models (AGCMs) with fixed SST and sea ice, but with differing CO2 concentrations in the atmosphere [8, 43]. It should be noted that influences attributed to land surface warming in response to the CO2 increase are also included within this estimate (discussed below) . The timescales of the cloud adjustment to abrupt increases in CO2 are thought to be of the order of days to weeks [1, 41]. To examine the rapid adjustment processes in the cloud and hydrological cycle, a short ensemble simulation based on the fixed-SST method has been employed [19•, 20••]. In this method, short-term and large-member ensemble simulations were performed with the AGCM 1× and 4× the concentration of atmospheric CO2. The ensemble-means of the two model outputs are compared at a fixed lead-time (e.g., 1 day, 5 days, and so on) and their difference can be evaluated as the rapid adjustment. In the regression method, on the contrary, SST and sea ice are allowed to vary. In abrupt CO2 concentration increase experiments of AOGCM, the decrease in TOA net radiation is approximately linear with SAT increase. The regression line is extrapolated to estimate the ERF at zero global-mean SAT change (i.e., without feedbacks) [4, 5, 29, 45]. To improve the signal-to-noise ratio, a large ensemble (>10) of integrations with abrupt increase in CO2 is preferred to diagnose the ERF using the regression method [7••, 20••, 46]. Note that change in local SAT is not necessarily equal to zero even when the global-mean change in SAT is zero. The direct response to increasing CO2 estimated in the regression method is influenced by the local SAT change partly .
Improvements in the evaluation of the rapid adjustment have provided new insights into the adjustment processes and sources of uncertainty. In this paper, the progress made by studies on the rapid adjustment and the direct climate responses to increasing CO2 is reviewed. Note that carbon cycle feedback is not quantified in the ECS concept  and is not included in this paper. We focused mainly on the following: (1) the physical mechanism of adjustments in cloud and hydrological cycle, (2) timescales of the processes related to rapid adjustment, (3) role of land surface to the direct responses in the cloud and troposphere, (4) sources of uncertainty in the cloud adjustment, and (5) some applications of the concept of the direct climate response to externally induced radiative forcing. Here, the attention is on climate responses to spatially uniform CO2 concentration increase; the influences of horizontally inhomogeneous radiative forcing agents (e.g., aerosols) on clouds have been detailed in other papers [3, 6].
Physical Mechanisms of Adjustments in Cloud and Hydrological Cycle
Although some previous studies have estimated the longwave and shortwave components of the radiative effect of rapid cloud adjustment, the signs of the estimated forcings varied among the different studies and models (particularly in the shortwave component) [4, 29]. However, interpretations of the physical processes of the rapid cloud adjustment and the robustness of its radiative effect have improved considerably because of the increased signal-to-noise ratio associated with quantifying methodologies (see previous subsection).
Summary of ERF and cloud radiative forcing estimates due to doubling of CO2 concentration and their spreads among multi model ensembles
ERF (W m−2)
Net cloud forcing (W m−2)
Longwave cloud forcing (W m−2)
Shortwave cloud forcing (W m−2)
3.33 ± 0.47
−0.36 ± 0.47a
−0.85 ± 0.21a
0.49 ± 0.41a
Gregory and Webb (2008) 
3.45 ± 0.56
−0.39 ± 0.39a
−0.78 ± 0.30a
0.39 ± 0.60a
Andrews et al. (2012) 
3.51 ± 0.47
Andrews et al. (2012) 
3.77 ± 0.45
−0.15 ± 0.26a
−0.76 ± 0.11a
0.62 ± 0.24a
Kamae and Watanabe (2012) [9•]
3.75 ± 0.42
0.41 ± 0.28
−0.13 ± 0.24
0.54 ± 0.45
Vial et al. (2013) [10•]
0.53 ± 0.34
−0.03 ± 0.13
0.56 ± 0.29
Zelinka et al. (2013) [7••]
Timescales of Global and Regional Adjustments
One of the reasons for the spatial non-uniformity in the adjustment is change in large-scale atmospheric circulation [16•, 20••, 21]. In particular, gradual change in land–sea thermal contrast (Fig. 4a–c) [20••, 41] is one of the reasons for the differences between the spatial patterns of adjustments found on a daily timescale (Fig. 4e, h, k, and n) and the equilibrium response (Fig. 4f, i, l, and o). The role of the gradual response to CO2 forcing over land (Fig. 4a–c) on cloud change is discussed in the following subsection.
Role of Land Surface Response
In the fixed-SST method, the land surface warms in response to the increased CO2 concentration, resulting in a land–sea surface warming contrast (Fig. 4c) [4, 16•, 20••, 41]. It is not strictly accurate to include additional responses associated with land surface warming as part of the adjustment processes. Land surface temperature should be fixed in order to isolate the radiative perturbation due to increasing CO2 ; however, this method has not been applied to other model simulations because of the difficulty of its implementation . This land–sea thermal contrast can evoke anomalous atmospheric circulation (upward/downward motions over the land/ocean), cloud [16•, 20••, 62], and precipitation patterns . Increasing cloud amount via enhanced deep convection over the land [16•] results in a negative SWCRE response over the land (Fig. 4f) despite the positive global-mean response of SWCRE (Fig. 3b, Table 1) [20••]. However, tropospheric cloud adjustment can also be found in aqua-planet models [20••, 57, 63] (detailed below), suggesting that the effect of changing land–sea thermal contrast is not necessary for cloud adjustment.
ΔSWCRE (W m−2)
2.74 ± 0.18
2.28 ± 0.52
1.29 ± 0.21
2.74 ± 0.18
2.86 ± 0.51
2.04 ± 0.14
0.48 ± 0.51
−0.52 ± 0.41
ΔLWCRE (W m−2)
−1.95 ± 0.18
−2.06 ± 0.46
−1.67 ± 0.18
−1.95 ± 0.18
−2.46 ± 0.45
−2.18 ± 0.22
−0.77 ± 0.45
−0.43 ± 0.35
It should be noted that the land effect on the climate response is also associated with CO2 physiological processes [5, 41, 60, 64, 65, 66]. In response to increasing CO2 concentration in the atmosphere, CO2 used in photosynthesis gains easier entry into the leaf via the stomata, resulting in the closures of stomata and the reduction of evapotranspiration from land plants, particularly in the low-latitude regions. The reduction of evapotranspiration can be found just after the CO2 increase (Fig. 4m) [20••, 41, 42•]. This dampening of water vapor release from the land surface suppresses precipitation over tropical land areas (Fig. 4j). These CO2-induced physiological processes have been examined and compared with physical processes via sensitivity simulations with fixed and variable stomatal resistances [41, 42•, 49, 60]. The reduction of land surface evapotranspiration due to physiological processes results in (1) additional surface and tropospheric warming over the land and associated changes in atmospheric circulation and (2) reduction of low cloud over land and increasing positive SWCRE [41, 49]. The influence of CO2 physiological forcing is an important factor in the changes of the hydrological cycle and cloud amount particularly over tropical land areas.
The responses to increasing CO2 over the land detailed above (e.g., SAT; Fig. 4a, b) are also important for the transient climate changes. In response to the transient radiative forcing, the land surface warming simulated in AOGCMs is generally larger than the sea surface warming [67, 68] because of differences in thermal inertia, boundary-layer lapse rate, and surface energy budget between the land and ocean [67, 69, 70]. Here, the direct response to the CO2 forcing over the land (Fig. 4a, b) is essential for the land surface warming particularly over the Northern Hemisphere (NH) extratropics [22, 71, 72]. As a result, projected future changes in the land–sea thermal contrast and associated atmospheric circulation over the NH extratropics [73, 74, 75] are clearly affected by the direct land response to the CO2 forcing . In addition, the direct response to the CO2 forcing is also important for the past climate changes [14, 22, 72]. Contribution of the direct land response to anthropogenically induced external forcing (CO2, aerosols, and land-use change) on the recent increase in SAT and frequency of hot summers over the NH land areas was found to be robust [22, 72]. These findings [14, 22, 71, 72, 73, 74, 75] explained new aspects of the direct climate response to the external forcing including the rapid tropospheric adjustment.
Possible Factors for Intermodel Spread of Adjustment
Combinations of the fixed-SST method, regression method, and aqua-planet simulations (see previous subsection) have enabled the quantitative assessment of the cloud adjustment and the evaluation of possible factors related to intermodel spread. Assessments of the radiative effect of the cloud adjustment have revealed that the influence of intermodel spread of the cloud adjustment on the uncertainty in the ECS is smaller than that of SAT-mediated cloud change [9•, 10•, 29, 53]. In response to increasing levels of CO2, the rapid adjustments of lower tropospheric stability (LTS) and estimated inversion strength (EIS) were found to be generally positive , except for one model in which biases in radiative heating rate resulted in unusual adjustments of LTS and EIS . As LTS and EIS are measures that are highly correlated with the amount of stratus in the present climate, these results suggest that the vertical profile of instantaneous radiative forcing could be a factor in the uncertainty of the cloud adjustment.
The relationship between the intermodel spread of the cloud radiative component of the ERF and that of cloud feedback parameter has received considerable attention . The relationship between the ERF and total feedback parameter of the CMIP5 GCMs shows a significant anticorrelation . Namely, models with larger ERF tend to have larger negative feedback parameter and vice versa. Such a relationship has also been confirmed in physics parameter perturbation ensembles [29, 78]. The cancelation results in a reduction of the intermodel spread of the ECS from that suggested by the intermodel spreads of the feedback parameter or ERF alone. In addition, clear anticorrelation between the CRE components of the ERF and feedback can also be found in aqua-planet models . Further studies are needed to explore both the possible mechanisms for such anticorrelation and the reasons for the uncertainties in the cloud adjustment and resultant ECS.
Our understanding of the physical mechanisms of rapid cloud and climate adjustments has advanced considerably because of extensive research. One reason for the progress is the development of a systematic experimental framework (including fixed-SST and regression methods and aqua-planet configuration) [7••, 9•, 19•, 20••, 53, 57, 63] with which to examine the rapid cloud adjustment under the Cloud Feedback Model Intercomparison Project phase 2 (CFMIP2)/CMIP5 protocol . The short ensemble simulations [19•, 20••] also contribute to improve the process-based interpretation of the rapid adjustment. The derived results have revealed the robustness of the tropospheric cloud adjustment [7••, 9•, 57]. However, the remaining issues on the sources of uncertainty in the cloud adjustment (e.g., instantaneous radiative perturbation, PBL dehydration, and associated changes in cloud radiative properties) [7••, 9•, 29, 48, 63, 77] are still important for the assessment of uncertainty in the ECS [29, 63]. Unfortunately, it is not easy to assess reliability of the adjustment processes derived from the modeling studies by comparing with observational records. However, detailed comparisons with responses simulated in less parameterized models [16•, 51, 52] may contribute to assess confidence of the current understanding of the mechanism of the cloud adjustment.
In this review, processes associated with adjustment to solar irradiance [21, 36, 37, 42•] and spatially inhomogeneous forcing factors such as aerosols and ozone [36, 79, 80, 81, 82] have not been included. Both the direct responses and SAT-mediated changes to the external forcing agents (including CO2 and other forcing factors) are important for the changes in large-scale atmospheric circulations (e.g., tropical atmospheric circulation, mid-latitude westerlies, and stratospheric circulation) [12, 13, 14, 15, 16•, 17, 18, 19•, 20••, 21, 22, 23, 80] and hydrological cycle [36, 80, 81]. Comprehensive approaches for better understanding of the direct and SAT-mediated climate responses are important future works. Such framework may provide significant implications for broad scientific areas not limited to the uncertainties in the ECS and transient climate response  but detection and attribution, atmospheric chemistry, paleoclimate, and geoengineering.
This work was supported by the Program for Risk Information on Climate Change (SOUSEI program) and Grant-in-Aid 26281013 from the Ministry of Education, Culture, Sports, Science, and Technology (MEXT), Japan, and the Environment Research and Technology Development Fund (S-10) of the Ministry of the Environment of Japan.
Conflict of Interest
On behalf of all authors, the corresponding author states that there is no conflict of interest.
Papers of particular interest, published recently, have been highlighted as: • Of importance •• Of major importance
- 3.Boucher O, Randall D, Artaxo P, et al. Clouds and aerosols. In: Stocker TF, editor. Climate change 2013: the physical science basis. Contribution of working group I to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge: Cambridge University Press; 2014. p. 571–658.Google Scholar
- 7.••Zelinka MD, Klein SA, Taylor KE, Andrews T, Webb MJ, Gregory JM, et al. Contributions of different cloud types to feedbacks and rapid adjustments in CMIP5. J Clim. 2013;26:5007–27. Using a cloud radiative kernel, this study revealed roles of different cloud heights and cloud optical depths in the rapid adjustment.CrossRefGoogle Scholar
- 10.•Vial J, Dufresne J-L, Bony S. On the interpretation of inter-model spread in CMIP5 climate sensitivity estimates. Clim Dyn. 2013;41:3339–62. Using NCAR model’s radiative kernel, this study explored relative contributions of the adjustment on the temperature change and the ECS uncertainty.CrossRefGoogle Scholar
- 16.•Wyant MC, Bretherton CS, Blossey PN, Khairoutdinov M. Fast cloud adjustment to increasing CO2 in a superparameterized climate model. J Adv Model Earth Syst. 2012;4:M05001. doi:10.1029/2011MS000092 This study examined cloud adjustment to CO 2 increase using super parameterized climate model and explored the processes related to PBL shoaling.
- 19.•Bony S, Bellon G, Klocke D, Sherwood S, Fermepin S, Denvil S. Robust direct effect of carbon dioxide on tropical circulation and regional precipitation. Nat Geosci. 2013;6:447–451. doi:10.1038/ngeo1799 This study revealed robustness of direct responses to CO 2 increase in tropical atmospheric circulation and precipitation on different timescales.
- 20.••Kamae Y, Watanabe M. Tropospheric adjustment to increasing CO2: its timescale and the role of land–sea contrast. Clim Dyn. 2013;41:3007–24. Using short ensemble and idealized simulations, this study cleary explored the rapid processes of adjustment and role of the increasing land–sea thermal contrast.CrossRefGoogle Scholar
- 42.•Cao L, Bala G, Caldeira K. Climate response to changes in atmospheric carbon dioxide and solar irradiance on the time scale of days to weeks. Environ Res Lett. 2012;7, 034015. Using a large-member ensemble simulation conducted in an AOGCM, this study examined processes of the rapid adjustment in hydrological cycle.CrossRefGoogle Scholar
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.