Radiative Forcing and Feedback
Many facets of the earth's climatic system are poorly understood. A significant uncertainty associated with the modeling of future climatic changes is due to deficiencies in the understanding of, and in the incorporation into the climate models of, several interactive climate feedback mechanisms. In this discussion of radiative forcing, the planet's heat balance, and these feedbacks and their consequences, emphasis is given to global mean quantities, since, conventionally, the concept of radiative feedback mechanisms is applied to global mean quantities associated with changes from one equilibrium climate to another. Many aspects of these feedback mechanisms are controversial. The conventional wisdom has been challenged on several points (see, for example, The George C. Marshall Institute, 1989; Lindzen, 1990), and new analyses of key issues are reported regularly (see, for example, Jenkinson et al., 1991; Kirchman et al., 1991; Ramanathan and Collins, 1991).
The Heat Balance
When the planet's climate is not undergoing change, the energy flux to the earth from the sun is in balance with the sum of the reflected portion of that solar flux and the outward infrared radiation that emerges from the top of the atmosphere. Figure 12.1 gives a schematic view of these fluxes and also depicts several of the components of the energy flow that occurs within the climatic system.
In particular, Figure 12.1 indicates that, of the 340 watts per square meter (W/m2) that are incident on the atmospheric envelope from the sun, about 100 W/m2 are reflected (from clouds, glaciated areas, and so on); 80 W/m2 are absorbed within the atmosphere, and 160 W/m2 are absorbed into
the oceans and the continental land masses. Heat is returned from the oceans and the land by infrared radiation from the surface (390 W/m2). Ninety W/m2 are returned by nonradiative processes that lead to the upward transport of water vapor (with its latent heat) and of sensible heat. All of the 90 W/m2 flux and a substantial part (call it B) of the 390 W/m2 radiative flux are deposited in the atmosphere. The total energy flux to the atmosphere (80 + 90 + B) is necessarily in balance with the infrared emission from the atmosphere, 320 W/m2 of which is directed downward into the ocean-land mass interface and the rest of which emerges from the top of the atmosphere (along with that fraction of the interface radiation (390 - B) that is not absorbed in the atmosphere). One can see that these numbers are mutually consistent by noting that the total nonreflected solar input is equal to the total (infrared) output of the system, that the total energy flux downward through the interface balances the upward flux from that interface, and that the energy received by the atmosphere (80 + 90 + B) plus the nonabsorbed upward flux from the interface (390 - B) is in balance with the radiation from the atmosphere that supplies the output of the system (at the top) and the radiative flux downward into the interface.
The interplay among the transmission of infrared radiation through the atmosphere, the absorption of infrared in the atmosphere, and the emission of infrared from the atmosphere is distributed throughout the height of the gaseous envelope. The foregoing, highly oversimplified splitting of these aspects of the infrared interchange into individual macroscopic items is needed only for the schematic presentation within this chapter. One need not choose a particular, necessarily artificial, value for B, because here a chosen value for B would not affect the conclusion drawn.
Note that, since each flux shown in Figure 12.1 is a globally averaged quantity, there is no depiction of the horizontal heat fluxes within the atmosphere and ocean that redistribute heat from one part of the planet to another. Nevertheless, such internal transfers play vital roles in the physical processes that determine the climatic state.
There are two important points to note from Figure 12.1. The first is that the 240 W/m2 emission at the top of the atmosphere (TOA) is 150 W/m2 less than the 390 W/m2 emission from the surface. This radiative flux difference is the greenhouse effect of the earth's present atmosphere, and it is caused by the absorption of infrared radiation by greenhouse gases and clouds. The second important point of Figure 12.1 is that the atmospheric greenhouse gases and the clouds emit infrared radiation downward to the surface, and this direct radiative heating of the surface by the atmosphere (320 W/m2) is twice the direct solar heating (160 W/m2). By itself, the additional 320 W/m2 provided by infrared surface heating produces substantial warming of the surface above the temperature that would otherwise prevail; thus it is the greenhouse effect that makes our planet habitable.
Suppose that, at time t, the state of the climatic system is that characterized by Figure 12.1. Suppose, too, that the greenhouse gas content of the atmosphere at that time is equivalent to a CO2 content of 300 ppmv. Suppose (once more) that, just after time t, the greenhouse gas levels are changed and the subsequent concentrations are equivalent to a CO2 concentration of 600 ppmv. This scenario is adopted here, not because CO2 doubling has some special significance but because it is an atmospheric modification that may occur within the next century and particularly because the CO2 doubling scenario has become the baseline case most commonly used by atmospheric scientists to compare the performance of one model with that of another. Immediately after this augmentation of greenhouse gas content, no changes in interface temperature, atmospheric temperature distribution, atmospheric moisture content, or any other component can yet have occurred. Accordingly, the only heat flux changes at that time would be those implied by an increased fractional absorption of the 390 W/m2 of infrared radiation up from the interface. That implication (for the CO2 doubling scenario) is a reduction of approximately 4.4 W/m2 in the upward infrared flux through the tropopause (not shown in Figure 12.1), which, after a radiatively controlled adjustment of the state of the stratosphere that evolves on a time scale of about one month, leads to a deficit, also of 4.4 W/m2, in the 240 W/m2 radiation out of the top of the atmosphere. This 4.4 W/m2 imbalance in
the overall heat budget of the planet is the ''radiative forcing" that accompanies the CO2 doubling scenario.
Comparisons of the radiative forcing associated with various greenhouse gas emission scenarios provide quantified characterizations of the effectiveness of such scenarios, and Chapter 17 includes a more general quantitative relationship between greenhouse gas emission scenarios and the radiative forcing levels to which they lead.
Radiative Feedback Mechanisms
In order to demonstrate radiative feedback mechanisms, it is convenient to assume initially that climate change is manifested solely by temperature changes within the climatic system and that all other climate parameters remain fixed at their unperturbed values. In this framework, there is no change in the climatic system's 240 W/m2 solar absorption. Moreover, let G denote the 4.4 W/m2 radiative forcing, and let DF be the change in the TOA infrared flux following the imposition of the forcing. Thus,
G = DF = dF / dT · DTs,
where dF/dT is the black body rate of change of the surface radiative flux per unit change of surface temperature, DTs. For the surface temperature of this calculation (288 K), dF/dT is 3.3 W/m2/°C, and therefore
DTs = 4.4 / 3.3 @ 1.3°C.
If it were not for the fact that this warming introduces numerous interactive feedback mechanisms, then DTs = 1.3°C (2.3°F) would be a robust estimate of that global mean quantity. Unfortunately, such feedbacks introduce considerable uncertainty in DTs estimates. Three of the more commonly discussed radiative feedbacks are described in the following sections. Although these phenomena are interrelated in the actual climatic system, they are discussed separately here to clearly identify their respective consequences. Additional feedback mechanisms involving land-surface hydrology are described in Chapter 15. The qualitative tools used to analyze energy transfers in the climatic system are less well developed in these latter mechanisms.
Water Vapor Feedback
The best understood, although still controversial, feedback mechanism is water vapor feedback. This phenomenon is intuitively easy to comprehend:
a warmer atmosphere can contain more water vapor, which itself is a greenhouse gas. Thus an increase in one greenhouse gas (CO2) induces an increase in yet another greenhouse gas (water vapor), resulting in a positive (amplifying) feedback mechanism. Although it has been suggested that the water vapor feedback might be negative (Lindzen, 1990), recent combined observational and model results strongly support the conventional interpretation that water vapor provides positive feedback (Raval and Ramanathan, 1989; Rind et al., 1991). Notwithstanding the connection between water vapor and clouds in the climatic system, they are treated separately for analytic purposes (see also the "Cloud Feedback" section, below).
To be specific on this point, Raval and Ramanathan (1989) have employed satellite data to quantify the temperature dependence of the water vapor greenhouse effect. From their results, it readily follows (Cess, 1989) that water vapor feedback reduces dF/dT from the prior value of 3.3 W/m2/°C to 2.3 W/m2/°C. This in turn increases the global warming, for a CO2 doubling, from 1.3° to 1.9°C (2.3° to 3.4°F).
There is yet a further amplification. Because water vapor also absorbs solar radiation, water vapor feedback leads to an additional heating of the climatic system through enhanced absorption of solar radiation. With Q denoting solar absorption by the climatic system (240 W/m2 for the present climate), this effect produces DQ/DTs = 0.2 W/m2/°C (Cess et al., 1990). To incorporate this into a DTs estimate, extension of the previous analysis to include solar absorption yields DTs = lG, where l is the climate sensitivity parameter defined by
It then follows that the inclusion of the solar component of water vapor feedback results in DTs = 2.1°C (3.8°F), so that the net effect of water vapor feedback is to amplify the initial DTs = 1.3°C(2.3°F) warming by the factor of 1.6.
The progressive forcing and feedback amplifications are summarized in Table 12.1. Here H denotes the greenhouse effect (150 W/m2 for the present climate). The radiative forcing (process 1) simultaneously increases H and reduces F, so that the planet emits 4.4 W/m2 less energy than it absorbs from the sun. It is this imbalance that causes greenhouse warming and results in DTs = 1.3°C (2.3°F) (process 2). Although the climatic system returns to its original radiation balance, with 240 W/m2 both absorbed and emitted, process 2 increases further the greenhouse effect by 2.7 W/m2 (154.4 to 157.1 W/m2) because of enhanced surface emission resulting from surface warming.
TABLE 12.1 Forcing and Response of the Climatic System Caused by a Doubling of Atmospheric CO2
Process 3 incorporates the infrared consequences of water vapor feedback, with a 3.3 W/m2 increase in H (157.1 to 160.4 W/m2) being simultaneously due to the increase in atmospheric water vapor and to enhanced surface emission. The TOA radiation budget is only slightly modified by process 4. An important point is that the combined effects of water vapor feedback and surface warming have amplified the 4.4 W/m2 greenhouse forcing to 11.1 W/m2. As Raval and Ramanathan (1989) have emphasized, this suggests that direct monitoring, from satellites, could help identify future changes in the greenhouse effect.
The most detailed climate models for the purpose of projecting climate change are three-dimensional atmospheric general circulation models (GCMs), and these models seem to depict properly the infrared component of water vapor feedback. In a recent intercomparison of atmospheric GCMs (Cess et al., 1990), it was found that 19 GCMs collectively produced
DF / DTs = 2.3±0.2 W / m2.
Thus the models used by many investigators are mutually consistent and in agreement with the observational result of Raval and Ramanathan (1989). It must be noted, however, that a variety of phenomena, such as the role of clouds, make such interpretation difficult.
The above calculations are not dynamic and thus do not completely characterize all relevant phenomena. A recent analysis is more inclusive (Rind et al., 1991). First, a GCM predicts increased water vapor in the middle and upper troposphere. This prediction is compared to observed satellite-generated data on seasonal water vapor content in the convective western Pacific and the largely nonconvective eastern Pacific. These results suggest that water vapor feedback is not overestimated in models and should amplify the climatic response to increased concentrations of greenhouse gases.
An additional feedback mechanism is snow-ice feedback, by which a warmer earth has less snow and ice cover, resulting in a darker planet that in turn absorbs more solar radiation. While this conventional albedo feedback description is quite obvious, and by itself constitutes a positive feedback, it now appears that the retreat of snow and ice cover might activate other interactive processes.
The same set of GCMs as used by Cess et al. (1990) to investigate cloud feedback (see next section) has recently been used to interpret and intercompare feedback associated solely with a change in snow cover. What that study shows is that clouds can significantly influence the snow-induced change in albedo and that this effect is more subtle than a mere masking of the change in surface albedo by the clouds. For example, in some models the snow-induced planetary albedo change is greater when clouds are present than when they are not. The reason for this is that the cloudiness change induced by the snow retreat causes a shift in clear-sky regions from snow-covered land to snow-free land, and this by itself is a positive feedback. The snow retreat can also induce an infrared feedback, which in some models is positive and in others negative. Thus it is clear that snow-ice feedback is far more complex than the conventional interpretation that it is a direct albedo feedback.
Feedback mechanisms related to clouds are extremely complex. To demonstrate this, it is useful to first consider the impact of clouds on the present climate. Summarized in Table 12.2 are the radiative impacts of clouds on the global climatic system for annual-mean conditions. These radiative impacts refer to the effect of clouds relative to a "clear-sky" earth. The
TABLE 12.2 Infrared, Solar, and Net Cloud Radiative Forcing (CRF)
presence of clouds heats the climatic system by 31 W/m2 through increasing the greenhouse effect. Because of the similarity of this process to trace gas radiative forcing, this impact is referred to as cloud radiative forcing. Through reflection of solar radiation, clouds also result in cooling of the system. As demonstrated in Table 12.2, the latter effect dominates the former, and the net effect of clouds on the annual climatic system is a 13 W/m2 radiative cooling.
Although clouds produce net cooling of the climatic system, this does not mean that clouds will necessarily offset additional global warming due to increasing greenhouse gases. As discussed in detail by Cess et al. (1989, 1990), cloud feedback constitutes the change in net cloud radiative forcing associated with a particular change in climate. To emphasize the complexity of this feedback mechanism, three contributory processes are summarized:
• Cloud amount: If cloud amount decreases because of global warming, as occurs in typical GCM simulations (e.g., Cess et al., 1989), then this decrease reduces the greenhouse effect attributed to clouds and so acts as a negative feedback mechanism. But there is a related positive feedback; the solar radiation absorbed by the climatic system increases because the diminished cloud cover causes a reduction of reflected solar radiation by clouds. There is no simple way of appraising the net sign of this feedback component.
• Cloud altitude: A vertical redistribution of clouds will also induce feedbacks. For example, if global warming displaces a given cloud layer to a higher and colder region of the atmosphere, this will produce a positive feedback because the colder cloud will emit less radiation and will thus enhance the greenhouse effect.
• Cloud water content: There has been considerable recent speculation that global warming could increase cloud water content, thereby resulting in brighter clouds and hence a negative component of cloud feedback. This may oversimplify the situation. Increases in cloud albedo can induce compensating positive infrared feedback (Cess et al., 1990), and in some models the net effect may be positive (Schlesinger, 1988; Cess et al., 1990; Rind et al., 1991). Recent analysis using both satellite and model results showed that highly reflective cirrus clouds are produced in tropical regions when sea surface temperature increases sufficiently (Ramanathan and Collins, 1991). The increased albedo may be sufficient to counter further warming due to infrared feedback.
The above discussion illustrates some of the complexities associated with cloud feedback; indeed, differences in models' depictions of this feedback largely account for the significant differences in climate sensitivity among the 19 GCMs (Cess et al., 1990). This intercomparison employed a perpetual July simulation in which the climate was changed by imposing a 4°C (7.2°F) perturbation on the global sea surface temperature while holding sea ice fixed. Since a perpetual July simulation with a GCM produces little snow cover over land, this effectively eliminates snow feedback. The details of this simulation are given elsewhere (Cess et al., 1989, 1990). The approach was chosen to minimize computer time and thus allow a large number of modeling groups to participate.
Cess et al. (1990) have summarized climate sensitivity parameters (l as defined in the "Water Vapor Feedback" section above) for the 19 GCMs, and these results are reproduced in Figure 12.2. The important point is that
cloud effects were isolated by separately averaging the models' clear-sky TOA fluxes, so that in addition to evaluating the climate sensitivity parameter for the globe as a whole (solid circles), it was also possible to evaluate it for an equivalent "clear-sky" earth (open circles). Note the remarkable agreement of the clear-sky sensitivity parameters; this is due to the agreement of water vapor feedback components, as discussed above. There is, however, a nearly threefold variation of the global (clear plus overcast) sensitivity parameter; clearly, given the clear-sky agreement, most of the variation in the global sensitivity parameters of current models can be attributed to cloud feedback. Certainly, improvements in the treatment of cloud feedback are needed if GCMs are ultimately to be used as reliable climate predictors.
Cess, R. D. 1989. Greenhouse effect: Gauging water-vapour feedback. Nature 342:736–737.
Cess, R. D., G. L. Potter, J. P. Blanchet, G. J. Boer, S. J. Ghan, J. T. Kiehl, H. Le Treut, Z.-X. Li, X.-Z. Liang, J. F. B. Mitchell, J. -J. Morcrette, D. A. Randall, M. R. Riches, E. Roeckner, U. Schlese, A. Slingo, K. E. Taylor, W. M. Washington, R. T. Wetherald, and I. Yagai. 1989. Interpretation of cloud-climate feedback as produced by 14 atmospheric general circulation models. Science 245:513–516.
Cess, R. D., G. L. Potter, J. P. Blanchet, G. J. Boer, A. D. Delgenio, M. Deque, V. Dymnikov, V. Galin, W. L. Gates, S. J. Ghan, J. T. Kiehl, A. A. Lacis, H. Le Treut, Z.-X. Li, X.-Z. Liang, B. J. McAvaney, V. P. Meleshko, J. F. B. Mitchell, J.-J. Morcrette, D. A. Randall, L. Rikus, E. Roeckner, J. F. Royer, U. Schlese, D. A. Sheinin, A. Slingo, A. P. Sokolov, K. E. Taylor, W. M. Washington, R. T. Wetherald, I. Yagai, and M. H. Zhang. 1990. Intercomparison and interpretation of climate feedback processes in 19 atmospheric general circulation models. Journal of Geophysical ResearchAtmospheres 95(10):16,601–16,615.
Jenkinson, D. S., D. E. Adams, and A. Wild. 1991. Model estimates of CO2 emissions from soil in response to global warming. Nature 351:304–306.
Kirchman, D. L., Y. Suzuki, C. Garside, and H. W. Ducklow. 1991. High turnover rates of dissolved organic carbon during a spring phytoplankton bloom. Nature 352:612–614.
Lindzen, R. S. 1990. Some remarks on global warming. Environmental Science and Technology 24(4):424–426.
Ramanathan, V., and W. Collins. 1991. Thermodynamic regulation of ocean warming by cirrus clouds deduced from observations of the 1987 El Niño. Nature 351:27–32.
Ramanathan, V., R. D. Cess, E. F. Harrison, P. Minnis, B. R. Barkstrom, E. Ahmad, and D. Hartmann. 1989. Cloud-radiative forcing and climate: Results from the Earth Radiation Budget Experiment. Science 243:57–63.
Raval, A., and V. Ramanathan. 1989. Observational determination of the greenhouse effect. Nature 342:758–761.
Rind, D., E. -W. Chiou, W. Chu, J. Larsen, S. Oltmans, J. Lerner, M. P. McCormick,
and L. McMaster. 1991. Positive water vapor feedback in climate models confirmed by satellite data. Nature 349:500–503.
Schlesinger, M. E. 1988. Ocean drilling program: Early glaciation of Antarctica. Nature 333:303.
The George C. Marshall Institute. 1989. Scientific Perspectives on the Greenhouse Problem. Washington, D.C.: The George C. Marshall Institute.