On the interpretation of inter-model spread in CMIP5 climate sensitivity estimates

This study diagnoses the climate sensitivity, radiative forcing and climate feedback estimates from eleven general circulation models participating in the Fifth Phase of the Coupled Model Intercomparison Project (CMIP5), and analyzes inter-model differences. This is done by taking into account the fact that the climate response to increased carbon dioxide (CO₂) is not necessarily only mediated by surface temperature changes, but can also result from fast land warming and tropospheric adjustments to the CO₂ radiative forcing. By considering tropospheric adjustments to CO₂ as part of the forcing rather than as feedbacks, and by using the radiative kernels approach, we decompose climate sensitivity estimates in terms of feedbacks and adjustments associated with water vapor, temperature lapse rate, surface albedo and clouds. Cloud adjustment to CO₂ is, with one exception, generally positive, and is associated with a reduced strength of the cloud feedback; the multi-model mean cloud feedback is about 33 % weaker. Non-cloud adjustments associated with temperature, water vapor and albedo seem, however, to be better understood as responses to land surface warming. Separating out the tropospheric adjustments does not significantly affect the spread in climate sensitivity estimates, which primarily results from differing climate feedbacks. About 70 % of the spread stems from the cloud feedback, which remains the major source of inter-model spread in climate sensitivity, with a large contribution from the tropics. Differences in tropical cloud feedbacks between low-sensitivity and high-sensitivity models occur over a large range of dynamical regimes, but primarily arise from the regimes associated with a predominance of shallow cumulus and stratocumulus clouds. The combined water vapor plus lapse rate feedback also contributes to the spread of climate sensitivity estimates, with inter-model differences arising primarily from the relative humidity responses throughout the troposphere. Finally, this study points to a substantial role of nonlinearities in the calculation of adjustments and feedbacks for the interpretation of inter-model spread in climate sensitivity estimates. We show that in climate model simulations with large forcing (e.g., 4 × CO₂), nonlinearities cannot be assumed minor nor neglected. Having said that, most results presented here are consistent with a number of previous feedback studies, despite the very different nature of the methodologies and all the uncertainties associated with them.     

Climate feedbacks determined using radiative kernels in a multi-thousand member ensemble of AOGCMs

The use of radiative kernels to diagnose climate feedbacks is a recent development that may be applied to existing climate change simulations. We apply the radiative kernel technique to transient simulations from a multi-thousand member perturbed physics ensemble of coupled atmosphere-ocean general circulation models, comparing distributions of model feedbacks with those taken from the CMIP-3 multi GCM ensemble. Although the range of clear sky longwave feedbacks in the perturbed physics ensemble is similar to that seen in the multi-GCM ensemble, the kernel technique underestimates the net clear-sky feedbacks (or the radiative forcing) in some perturbed models with significantly altered humidity distributions. In addition, the compensating relationship between global mean atmospheric lapse rate feedback and water vapor feedback is found to hold in the perturbed physics ensemble, but large differences in relative humidity distributions in the ensemble prevent the compensation from holding at a regional scale. Both ensembles show a similar range of response of global mean net cloud feedback, but the mean of the perturbed physics ensemble is shifted towards more positive values such that none of the perturbed models exhibit a net negative cloud feedback. The perturbed physics ensemble contains fewer models with strong negative shortwave cloud feedbacks and has stronger compensating positive longwave feedbacks. A principal component analysis used to identify dominant modes of feedback variation reveals that the perturbed physics ensemble produces very different modes of climate response to the multi-model ensemble, suggesting that one may not be used as an analog for the other in estimates of uncertainty in future response. Whereas in the multi-model ensemble, the first order variation in cloud feedbacks shows compensation between longwave and shortwave components, in the perturbed physics ensemble the shortwave feedbacks are uncompensated, possibly explaining the larger range of climate sensitivities observed in the perturbed simulations. Regression analysis suggests that the parameters governing cloud formation, convection strength and ice fall speed are the most significant in altering climate feedbacks. Perturbations of oceanic and sulfur cycle parameters have relatively little effect on the atmospheric feedbacks diagnosed by the kernel technique.     

The respective roles of surface temperature driven feedbacks and tropospheric adjustment to CO2 in CMIP5 transient climate simulations

An overview of radiative climate feedbacks and ocean heat uptake efficiency diagnosed from idealized transient climate change experiments of 14 CMIP5 models is presented. Feedbacks explain about two times more variance in transient climate response across the models than ocean heat uptake efficiency. Cloud feedbacks can clearly be identified as the main source of inter-model spread. Models with strong longwave feedbacks in the tropics feature substantial increases in cloud ice around the tropopause suggestive of changes in cloud-top heights. The lifting of the tropical tropopause goes together with a general weakening of the tropical circulation. Distinctive inter-model differences in cloud shortwave feedbacks occur in the subtropics including the equatorward flanks of the storm-tracks. Related cloud fraction changes are not confined to low clouds but comprise middle level clouds as well. A reduction in relative humidity through the lower and mid troposphere can be identified as being the main associated large-scale feature. Experiments with prescribed sea surface temperatures are analyzed in order to investigate whether the diagnosed feedbacks from the transient climate simulations contain a tropospheric adjustment component that is not conveyed through the surface temperature response. The strengths of the climate feedbacks computed from atmosphere-only experiments with prescribed increases in sea surface temperatures, but fixed CO₂ concentrations, are close to the ones derived from the transient experiment. Only the cloud shortwave feedback exhibits discernible differences which, however, can not unequivocally be attributed to tropospheric adjustment to CO₂. Although for some models a tropospheric adjustment component is present in the global mean shortwave cloud feedback, an analysis of spatial patterns does not lend support to the view that cloud feedbacks are dominated by their tropospheric adjustment part. Nevertheless, there is positive correlation between the strength of tropospheric adjustment processes and cloud feedbacks across different climate models.     

Origins of differences in climate sensitivity, forcing and feedback in climate models

We diagnose climate feedback parameters and CO₂ forcing including rapid adjustment in twelve atmosphere/mixed-layer-ocean (“slab”) climate models from the CMIP3/CFMIP-1 project (the AR4 ensemble) and fifteen parameter-perturbed versions of the HadSM3 slab model (the PPE). In both ensembles, differences in climate feedbacks can account for approximately twice as much of the range in climate sensitivity as differences in CO₂ forcing. In the AR4 ensemble, cloud effects can explain the full range of climate sensitivities, and cloud feedback components contribute four times as much as cloud components of CO₂ forcing to the range. Non-cloud feedbacks are required to fully account for the high sensitivities of some models however. The largest contribution to the high sensitivity of HadGEM1 is from a high latitude clear-sky shortwave feedback, and clear-sky longwave feedbacks contribute substantially to the highest sensitivity members of the PPE. Differences in low latitude ocean regions (30°N/S) contribute more to the range than those in mid-latitude oceans (30–55°N/S), low/mid latitude land (55°N/S) or high latitude ocean/land (55–90°N/S), but contributions from these other regions are required to account fully for the higher model sensitivities, for example from land areas in IPSL CM4. Net cloud feedback components over the low latitude oceans sorted into percentile ranges of lower tropospheric stability (LTS) show largest differences among models in stable regions, mainly due to their shortwave components, most of which are positive in spite of increasing LTS. Differences in the mid-stability range are smaller, but cover a larger area, contributing a comparable amount to the range in climate sensitivity. These are strongly anti-correlated with changes in subsidence. Cloud components of CO₂ forcing also show the largest differences in stable regions, and are strongly anticorrelated with changes in estimated inversion strength (EIS). This is qualitatively consistent with what would be expected from observed relationships between EIS and low-level cloud fraction. We identify a number of cases where individual models show unusually strong forcings and feedbacks compared to other members of the ensemble. We encourage modelling groups to investigate unusual model behaviours further with sensitivity experiments. Most of the models fail to correctly reproduce the observed relationships between stability and cloud radiative effect in the subtropics, indicating that there remains considerable room for model improvements in the future.     

Atmospheric radiative feedbacks associated with transient climate change and climate variability

This study examines in detail the ‘atmospheric’ radiative feedbacks operating in a coupled General Circulation Model (GCM). These feedbacks (defined as the change in top of atmosphere radiation per degree of global surface temperature change) are due to responses in water vapour, lapse rate, clouds and surface albedo. Two types of radiative feedback in particular are considered: those arising from century scale ‘transient’ warming (from a 1% per annum compounded CO₂ increase), and those operating under the model’s own unforced ‘natural’ variability. The time evolution of the transient (or ‘secular’) feedbacks is first examined. It is found that both the global strength and the latitudinal distributions of these feedbacks are established within the first two or three decades of warming, and thereafter change relatively little out to 100 years. They also closely approximate those found under equilibrium warming from a ‘mixed layer’ ocean version of the same model forced by a doubling of CO₂. These secular feedbacks are then compared with those operating under unforced (interannual) variability. For water vapour, the interannual feedback is only around two-thirds the strength of the secular feedback. The pattern reveals widespread regions of negative feedback in the interannual case, in turn resulting from patterns of circulation change and regions of decreasing as well as increasing surface temperature. Considering the vertical structure of the two, it is found that although positive net mid to upper tropospheric contributions dominate both, they are weaker (and occur lower) under interannual variability than under secular change and are more narrowly confined to the tropics. Lapse rate feedback from variability shows weak negative feedback over low latitudes combined with strong positive feedback in mid-to-high latitudes resulting in no net global feedback—in contrast to the dominant negative low to mid-latitude response seen under secular climate change. Surface albedo feedback is, however, slightly stronger under interannual variability—partly due to regions of extremely weak, or even negative, feedback over Antarctic sea ice in the transient experiment. Both long and shortwave global cloud feedbacks are essentially zero on interannual timescales, with the shortwave term also being very weak under climate change, although cloud fraction and optical property components show correlation with global temperature both under interannual variability and transient climate change. The results of this modelling study, although for a single model only, suggest that the analogues provided by interannual variability may provide some useful pointers to some aspects of climate change feedback strength, particularly for water vapour and surface albedo, but that structural differences will need to be heeded in such an analysis.     

Why radiative forcing might fail as a predictor of climate change

Radiative forcing has been widely used as a metric of climate change, i.e. as a measure by which various contributors to a net surface temperature change can be quantitatively compared. The extent to which this concept is valid for spatially inhomogeneous perturbations to the climate system is tested. A series of climate model simulations involving ozone changes of different spatial structure reveals that the climate sensitivity parameter is highly variable: for an ozone increase in the northern hemisphere lower stratosphere, it is more than twice as large as for a homogeneous CO₂ perturbation. A global ozone perturbation in the upper troposphere, however, causes a significantly smaller surface temperature response than CO₂. The variability of the climate sensitivity parameter is shown to be mostly due to the varying strength of the stratospheric water vapour feedback. The variability of the sea-ice albedo feedback modifies climate sensitivity of perturbations with the same vertical structure but a different horizontal structure. This feedback is also the origin of the comparatively larger climate sensitivity to perturbations restricted to the northern hemisphere extratropics. As cloud feedback does not operate independently from the other feedbacks, quantifying its effect is rather difficult. However, its effect on the variability of for horizontally and vertically inhomogeneous perturbations within one model framework seems to be comparatively small.This revised version was published online March 2005 with corrections to table 5.     

Climate feedback efficiency and synergy

Earth’s climate sensitivity to radiative forcing induced by a doubling of the atmospheric CO₂ is determined by feedback mechanisms, including changes in atmospheric water vapor, clouds and surface albedo, that act to either amplify or dampen the response. The climate system is frequently interpreted in terms of a simple energy balance model, in which it is assumed that individual feedback mechanisms are additive and act independently. Here we test these assumptions by systematically controlling, or locking, the radiative feedbacks in a state-of-the-art climate model. The method is shown to yield a near-perfect decomposition of change into partial temperature contributions pertaining to forcing and each of the feedbacks. In the studied model water vapor feedback stands for about half the temperature change, CO₂-forcing about one third, while cloud and surface albedo feedback contributions are relatively small. We find a close correspondence between forcing, feedback and partial surface temperature response for the water vapor and surface albedo feedbacks, while the cloud feedback is inefficient in inducing surface temperature change. Analysis suggests that cloud-induced warming in the upper tropical troposphere, consistent with rising convective cloud anvils in a warming climate enhances the negative lapse-rate feedback, thereby offsetting some of the warming that would otherwise be attributable to this positive cloud feedback. By subsequently combining feedback mechanisms we find a positive synergy acting between the water vapor feedback and the cloud feedback; that is, the combined cloud and water vapor feedback is greater than the sum of its parts. Negative synergies surround the surface albedo feedback, as associated cloud and water vapor changes dampen the anticipated climate change induced by retreating snow and ice. Our results highlight the importance of treating the coupling between clouds, water vapor and temperature in a deepening troposphere.     

Towards constraining climate sensitivity by linear analysis of feedback patterns in thousands of perturbed-physics GCM simulations

A linear analysis is applied to a multi-thousand member “perturbed physics" GCM ensemble to identify the dominant physical processes responsible for variation in climate sensitivity across the ensemble. Model simulations are provided by the distributed computing project, climate prediction.net . A principal component analysis of model radiative response reveals two dominant independent feedback processes, each largely controlled by a single parameter change. The leading EOF was well correlated with the value of the entrainment coefficient—a parameter in the model’s atmospheric convection scheme. Reducing this parameter increases high vertical level moisture causing an enhanced clear sky greenhouse effect both in the control simulation and in the response to greenhouse gas forcing. This effect is compensated by an increase in reflected solar radiation from low level cloud upon warming. A set of ‘secondary’ cloud formation parameters partly modulate the degree of shortwave compensation from low cloud formation. The second EOF was correlated with the scaling of ice fall speed in clouds which affects the extent of cloud cover in the control simulation. The most prominent feature in the EOF was an increase in longwave cloud forcing. The two leading EOFs account for 70% of the ensemble variance in λ—the global feedback parameter. Linear predictors of feedback strength from model climatology are applied to observational datasets to estimate real world values of the overall climate feedback parameter. The predictors are found using correlations across the ensemble. Differences between predictions are largely due to the differences in observational estimates for top of atmosphere shortwave fluxes. Our validation does not rule out all the strong tropical convective feedbacks leading to a large climate sensitivity.     

Climate sensitivity and response

Results from climate change simulations indicate a reasonably robust proportionality between global mean radiative forcing and global mean surface air temperature response. The "constant" of proportionality is a measure of the overall strength of climate feedback processes and hence of global climate sensitivity. Geographically, however, temperature response patterns are generally not proportional to, nor do they resemble, their parent forcing patterns. Temperature response patterns, nevertheless, exhibit a remarkable additivity whereby the sum of response patterns for different forcings closely resembles the response pattern for the sum of the forcings. The geographical distribution of contributions to the climate sensitivity/feedback are obtained diagnostically from simulations with the Canadian Centre for Climate Modelling and Analysis (CCCma) coupled global climate model (GCM). There is positive feedback over high-latitude oceans, over northern land areas, and over the equatorial Pacific. The remaining regions over oceans and tropical land areas exhibit negative feedback. The feedback results are decomposed into components associated with short-and longwave radiative processes and in terms of cloud-free atmosphere/surface and cloud feedbacks. While the geographic pattern of the feedbacks may generally be linked to local processes, all feedback processes display regions of both positive and negative values (except for the solar atmosphere/surface feedback associated with the retreat of ice and snow which is positive) and all vary from place to place so that there is no simple physical picture that operates everywhere. The stable geographical pattern of the feedback is a consequence of the balance between local physical processes rather than the dominance of a particular process. The feedback results indicate that, to first order, temperature response patterns are determined by the geographical pattern of local feedback processes. The feedback processes act to localize forcing changes and to generate temperature response patterns which depend firstly on the pattern of feedbacks and only secondarily on the pattern of the forcing. The geographical distribution of feedback processes can be regarded as a feature of the climate model (and by inference of the climate system) and not (or only comparatively weak) functions of forcing and climate state. An illustrative model is able to reproduce qualitatively the kinds of forcing/temperature response behavior seen in the CCCma GCM including the quasi-independence of forcing and response patterns, the additivity of temperature response patterns, and the resulting "non-constancy" of the global climate sensitivity.     

The role of ecosystem-atmosphere interactions in simulated Amazonian precipitation decrease and forest dieback under global climate warming

Summary A suite of simulations with the HadCM3LC coupled climate-carbon cycle model is used to examine the various forcings and feedbacks involved in the simulated precipitation decrease and forest dieback. Rising atmospheric CO₂ is found to contribute 20% to the precipitation reduction through the physiological forcing of stomatal closure, with 80% of the reduction being seen when stomatal closure was excluded and only radiative forcing by CO₂ was included. The forest dieback exerts two positive feedbacks on the precipitation reduction; a biogeophysical feedback through reduced forest cover suppressing local evaporative water recycling, and a biogeochemical feedback through the release of CO₂ contributing to an accelerated global warming. The precipitation reduction is enhanced by 20% by the biogeophysical feedback, and 5% by the carbon cycle feedback from the forest dieback. This analysis helps to explain why the Amazonian precipitation reduction simulated by HadCM3LC is more extreme than that simulated in other GCMs; in the fully-coupled, climate-carbon cycle simulation, approximately half of the precipitation reduction in Amazonia is attributable to a combination of physiological forcing and biogeophysical and global carbon cycle feedbacks, which are generally not included in other GCM simulations of future climate change. The analysis also demonstrates the potential contribution of regional-scale climate and ecosystem change to uncertainties in global CO₂ and climate change projections. Moreover, the importance of feedbacks suggests that a human-induced increase in forest vulnerability to climate change may have implications for regional and global scale climate sensitivity.     

The role of ENSO in global ocean temperature changes during 1955–2011 simulated with a 1D climate model

Global average ocean temperature variations to 2,000 m depth during 1955–2011 are simulated with a 40 layer 1D forcing-feedback-mixing model for three forcing cases. The first case uses standard anthropogenic and volcanic external radiative forcings. The second adds non-radiative internal forcing (ocean mixing changes initiated in the top 200 m) proportional to the Multivariate ENSO Index (MEI) to represent an internal mode of natural variability. The third case further adds ENSO-related radiative forcing proportional to MEI as a possible natural cloud forcing mechanism associated with atmospheric circulation changes. The model adjustable parameters are net radiative feedback, effective diffusivities, and internal radiative (e.g., cloud) and non-radiative (ocean mixing) forcing coefficients at adjustable time lags. Model output is compared to Levitus ocean temperature changes in 50 m layers during 1955–2011 to 700 m depth, and to lag regression coefficients between satellite radiative flux variations and sea surface temperature between 2000 and 2010. A net feedback parameter of 1.7Wm^?2 K^?1 with only anthropogenic and volcanic forcings increases to 2.8Wm^?2 K^?1 when all ENSO forcings (which are one-third radiative) are included, along with better agreement between model and observations. The results suggest ENSO can influence multi-decadal temperature trends, and that internal radiative forcing of the climate system affects the diagnosis of feedbacks. Also, the relatively small differences in model ocean warming associated with the three cases suggests that the observed levels of ocean warming since the 1950s is not a very strong constraint on our estimates of climate sensitivity.     

Inferring climate sensitivity from volcanic events

The possibility of estimating the equilibrium climate sensitivity of the earth-system from observations following explosive volcanic eruptions is assessed in the context of a perfect model study. Two modern climate models (the CCCma CGCM3 and the NCAR CCSM2) with different equilibrium climate sensitivities are employed in the investigation. The models are perturbed with the same transient volcano-like forcing and the responses analysed to infer climate sensitivities. For volcano-like forcing the global mean surface temperature responses of the two models are very similar, despite their differing equilibrium climate sensitivities, indicating that climate sensitivity cannot be inferred from the temperature record alone even if the forcing is known. Equilibrium climate sensitivities can be reasonably determined only if both the forcing and the change in heat storage in the system are known very accurately. The geographic patterns of clear-sky atmosphere/surface and cloud feedbacks are similar for both the transient volcano-like and near-equilibrium constant forcing simulations showing that, to a considerable extent, the same feedback processes are invoked, and determine the climate sensitivity, in both cases.     

Evaluation of a component of the cloud response to climate change in an intercomparison of climate models

Most of the uncertainty in the climate sensitivity of contemporary general circulation models (GCMs) is believed to be connected with differences in the simulated radiative feedback from clouds. Traditional methods of evaluating clouds in GCMs compare time–mean geographical cloud fields or aspects of present-day cloud variability, with observational data. In both cases a hypothetical assumption is made that the quantity evaluated is relevant for the mean climate change response. Nine GCMs (atmosphere models coupled to mixed-layer ocean models) from the CFMIP and CMIP model comparison projects are used in this study to demonstrate a common relationship between the mean cloud response to climate change and present-day variability. Although atmosphere–mixed-layer ocean models are used here, the results are found to be equally applicable to transient coupled model simulations. When changes in cloud radiative forcing (CRF) are composited by changes in vertical velocity and saturated lower tropospheric stability, a component of the local mean climate change response can be related to present-day variability in all of the GCMs. This suggests that the relationship is not model specific and might be relevant in the real world. In this case, evaluation within the proposed compositing framework is a direct evaluation of a component of the cloud response to climate change. None of the models studied are found to be clearly superior or deficient when evaluated, but a couple appear to perform well on several relevant metrics. Whilst some broad similarities can be identified between the 60°N–60°S mean change in CRF to increased CO₂ and that predicted from present-day variability, the two cannot be quantitatively constrained based on changes in vertical velocity and stability alone. Hence other processes also contribute to the global mean cloud response to climate change.     

Climate feedbacks in a general circulation model incorporating prognostic clouds

 This study performs a comprehensive feedback analysis on the Bureau of Meteorology Research Centre General Circulation Model, quantifying all important feedbacks operating under an increase in atmospheric CO₂. The individual feedbacks are analysed in detail, using an offline radiation perturbation method, looking at long- and shortwave components, latitudinal distributions, cloud impacts, non-linearities under 2xCO₂ and 4xCO₂ warmings and at interannual variability. The water vapour feedback is divided into terms due to moisture height and amount changes. The net cloud feedback is separated into terms due to cloud amount, height, water content, water phase, physical thickness and convective cloud fraction. Globally the most important feedbacks were found to be (from strongest positive to strongest negative) those due to water vapour, clouds, surface albedo, lapse rate and surface temperature. For the longwave (LW) response the most important term of the cloud ‘optical property’ feedbacks is due to the water content. In the shortwave (SW), both water content and water phase changes are important. Cloud amount and height terms are also important for both LW and SW. Feedbacks due to physical cloud thickness and convective cloud fraction are found to be relatively small. All cloud component feedbacks (other than height) produce conflicting LW/SW feedbacks in the model. Furthermore, the optical property and cloud fraction feedbacks are also of opposite sign. The result is that the net cloud feedback is the (relatively small) product of conflicting physical processes. Non-linearities in the feedbacks are found to be relatively small for all but the surface albedo response and some cloud component contributions. The cloud impact on non-cloud feedbacks is also discussed: greatest impact is on the surface albedo, but impact on water vapour feedback is also significant. The analysis method here proves to be a␣powerful tool for detailing the contributions from different model processes (and particularly those of the clouds) to the final climate model sensitivity. Received: 15 June 2000 / Accepted: 10 January 2001     

A further assessment of vegetation feedback on decadal Sahel rainfall variability

The effect of vegetation feedback on decadal-scale Sahel rainfall variability is analyzed using an ensemble of climate model simulations in which the atmospheric general circulation model ICTPAGCM (“SPEEDY”) is coupled to the dynamic vegetation model VEGAS to represent feedbacks from surface albedo change and evapotranspiration, forced externally by observed sea surface temperature (SST) changes. In the control experiment, where the full vegetation feedback is included, the ensemble is consistent with the observed decadal rainfall variability, with a forced component 60 % of the observed variability. In a sensitivity experiment where climatological vegetation cover and albedo are prescribed from the control experiment, the ensemble of simulations is not consistent with the observations because of strongly reduced amplitude of decadal rainfall variability, and the forced component drops to 35 % of the observed variability. The decadal rainfall variability is driven by SST forcing, but significantly enhanced by land-surface feedbacks. Both, local evaporation and moisture flux convergence changes are important for the total rainfall response. Also the internal decadal variability across the ensemble members (not SST-forced) is much stronger in the control experiment compared with the one where vegetation cover and albedo are prescribed. It is further shown that this positive vegetation feedback is physically related to the albedo feedback, supporting the Charney hypothesis.     

On the vertical extent of atmospheric feedbacks

This study addresses the question: what vertical regions contribute the most to water vapor, surface temperature, lapse rate and cloud fraction feedback strengths in a general circulation model? Multi-level offline radiation perturbation calculations are used to diagnose the feedback contribution from each model level. As a first step, to locate regions of maximum radiative sensitivity to climate changes, the top of atmosphere radiative impact for each feedback is explored for each process by means of idealized parameter perturbations on top of a control (1?×?CO₂) model climate. As a second step, the actual feedbacks themselves are calculated using the changes modelled from a 2?×?CO₂ experiment. The impact of clouds on water vapor and lapse rate feedbacks is also isolated using `clear sky' calculations. Considering the idealized changes, it is found that the radiative sensitivity to water vapor changes is a maximum in the tropical lower troposphere. The sensitivity to temperature changes has both upper and lower tropospheric maxima. The sensitivity to idealized cloud changes is positive (warming) for upper level cloud increases but negative (cooling) for lower level increases, due to competing long and shortwave effects. Considering the actual feedbacks, it is found that water vapor feedback is a maximum in the tropical upper troposphere, due to the large relative increases in specific humidity which occur there. The actual lapse rate feedback changes sign with latitude and is a maximum (negative) again in the tropical upper troposphere. Cloud feedbacks reflect the general decrease in low- to mid-level low-latitude cloud, with an increase in the very highest cloud. This produces a net positive (negative) shortwave (longwave) cloud feedback. The role of clouds in the strength of the water vapor and lapse rate feedbacks is also discussed.     

Quantifying contributions of climate feedbacks to tropospheric warming in the NCAR CCSM3.0

In this study, a coupled atmosphere-surface “climate feedback-response analysis method” (CFRAM) was applied to the slab ocean model version of the NCAR CCSM3.0 to understand the tropospheric warming due to a doubling of CO₂ concentration through quantifying the contributions of each climate feedback process. It is shown that the tropospheric warming displays distinct meridional and vertical patterns that are in a good agreement with the multi-model mean projection from the IPCC AR4. In the tropics, the warming in the upper troposphere is stronger than in the lower troposphere, leading to a decrease in temperature lapse rate, whereas in high latitudes the opposite it true. In terms of meridional contrast, the lower tropospheric warming in the tropics is weaker than that in high latitudes, resulting in a weakened meridional temperature gradient. In the upper troposphere the meridional temperature gradient is enhanced due to much stronger warming in the tropics than in high latitudes. Using the CFRAM method, we analyzed both radiative feedbacks, which have been emphasized in previous climate feedback analysis, and non-radiative feedbacks. It is shown that non-radiative (radiative) feedbacks are the major contributors to the temperature lapse rate decrease (increase) in the tropical (polar) region. Atmospheric convection is the leading contributor to temperature lapse rate decrease in the tropics. The cloud feedback also has non-negligible contributions. In the polar region, water vapor feedback is the main contributor to the temperature lapse rate increase, followed by albedo feedback and CO₂ forcing. The decrease of meridional temperature gradient in the lower troposphere is mainly due to strong cooling from convection and cloud feedback in the tropics and the strong warming from albedo feedback in the polar region. The strengthening of meridional temperature gradient in the upper troposphere can be attributed to the warming associated with convection and cloud feedback in the tropics. Since convection is the leading contributor to the warming differences between tropical lower and upper troposphere, and between the tropical and polar regions, this study indicates that tropical convection plays a critical role in determining the climate sensitivity. In addition, the CFRAM analysis shows that convective process and water vapor feedback are the two major contributors to the tropical upper troposphere temperature change, indicating that the excessive upper tropospheric warming in the IPCC AR4 models may be due to overestimated warming from convective process or underestimated cooling due to water vapor feedback.     

On the contribution of local feedback mechanisms to the range of climate sensitivity in two GCM ensembles

Global and local feedback analysis techniques have been applied to two ensembles of mixed layer equilibrium CO₂ doubling climate change experiments, from the CFMIP (Cloud Feedback Model Intercomparison Project) and QUMP (Quantifying Uncertainty in Model Predictions) projects. Neither of these new ensembles shows evidence of a statistically significant change in the ensemble mean or variance in global mean climate sensitivity when compared with the results from the mixed layer models quoted in the Third Assessment Report of the IPCC. Global mean feedback analysis of these two ensembles confirms the large contribution made by inter-model differences in cloud feedbacks to those in climate sensitivity in earlier studies; net cloud feedbacks are responsible for 66% of the inter-model variance in the total feedback in the CFMIP ensemble and 85% in the QUMP ensemble. The ensemble mean global feedback components are all statistically indistinguishable between the two ensembles, except for the clear-sky shortwave feedback which is stronger in the CFMIP ensemble. While ensemble variances of the shortwave cloud feedback and both clear-sky feedback terms are larger in CFMIP, there is considerable overlap in the cloud feedback ranges; QUMP spans 80% or more of the CFMIP ranges in longwave and shortwave cloud feedback. We introduce a local cloud feedback classification system which distinguishes different types of cloud feedbacks on the basis of the relative strengths of their longwave and shortwave components, and interpret these in terms of responses of different cloud types diagnosed by the International Satellite Cloud Climatology Project simulator. In the CFMIP ensemble, areas where low-top cloud changes constitute the largest cloud response are responsible for 59% of the contribution from cloud feedback to the variance in the total feedback. A similar figure is found for the QUMP ensemble. Areas of positive low cloud feedback (associated with reductions in low level cloud amount) contribute most to this figure in the CFMIP ensemble, while areas of negative cloud feedback (associated with increases in low level cloud amount and optical thickness) contribute most in QUMP. Classes associated with high-top cloud feedbacks are responsible for 33 and 20% of the cloud feedback contribution in CFMIP and QUMP, respectively, while classes where no particular cloud type stands out are responsible for 8 and 21%.     

A new framework for isolating individual feedback processes in coupled general circulation climate models. Part I: formulation

This paper proposes a coupled atmosphere–surface climate feedback–response analysis method (CFRAM) as a new framework for estimating climate feedbacks in coupled general circulation models with a full set of physical parameterization packages. The formulation of the CFRAM is based on the energy balance in an atmosphere–surface column. In the CFRAM, the isolation of partial temperature changes due to an external forcing or an individual feedback is achieved by solving the linearized infrared radiation transfer model subject to individual energy flux perturbations (external or due to feedbacks). The partial temperature changes are addable and their sum is equal to the (total) temperature change (in the linear sense). The decomposition of feedbacks is based on the thermodynamic and dynamical processes that directly affect individual energy flux terms. Therefore, not only those feedbacks that directly affect the TOA radiative fluxes, such as water vapor, clouds, and ice-albedo feedbacks, but also those feedbacks that do not directly affect the TOA radiation, such as evaporation, convections, and convergence of horizontal sensible and latent heat fluxes, are explicitly included in the CFRAM. In the CFRAM, the feedback gain matrices measure the strength of individual feedbacks. The feedback gain matrices can be estimated from the energy flux perturbations inferred from individual parameterization packages and dynamical modules. The inter-model spread of a feedback gain matrix would help us to detect the origins of the uncertainty of future climate projections in climate model simulations.     

Natural variability of summer rainfall over China in HadCM3

Summer rainfall over China has shown decadal variability in the past half century, which has resulted in major north–south shifts in rainfall with important implications for flooding and water resource management. This study has demonstrated how multi-century climate model simulations can be used to explore interdecadal natural variability in the climate system in order to address important questions around recent changes in Chinese summer rainfall, and whether or not anthropogenic climate change is playing a role. Using a 1,000-year simulation of HadCM3 with constant pre-industrial external forcing, the dominant modes of total and interdecadal natural variability in Chinese summer rainfall have been analysed. It has been shown that these modes are comparable in magnitude and in temporal and spatial characteristics to those observed in the latter part of the twentieth century. However, despite 1,000 years of model simulation it has not been possible to demonstrate that these modes are related to similar variations in the global circulation and surface temperature forcing occurring during the latter half of the twentieth century. This may be in part due to model biases. Consequently, recent changes in the spatial distribution of Chinese summer rainfall cannot be attributed solely to natural variability, nor has it been possible to eliminate the likelihood that anthropogenic climate change has been the driving factor. It is more likely that both play a role.