National Center for Atmospheric Research (NCAR)
Boulder, CO 80307
Recent indications suggest that clouds may be absorbing more solar radiation than was previously thought. This research investigates some of the evidence for this hypothesis; potential physical mechanisms are briefly discussed as well. The climatic implications of the enhanced absorption are investigated using the NCAR Community Climate Model (CCM). It is found that the model's heat budget in the tropical warm pool agrees more closely with observations when enhanced absorption is included.
Observational evidence: Cess et al. correlated top of the atmosphere (TOA) albedo as a function of normalized surface insolation. Points lie approximately on a line with a slope of ~-0.6 for observations in four locations (Barrow, Boulder, Samoa, and Wisconsin), compared to a slope of ~-0.8 for two GCMs (CCM2 and ECMWF). This implies that the model atmospheres are absorbing too little solar radiation. This result is independently confirmed by comparing the mean short wave cloud radiative forcing (SWCF) at the surface to that at the TOA. The ratio of these two values is 1.5, which is the reciprocal of -0.6, corrected for the surface albedo.
Indirect evidence is provided by Ramanathan et al .who evaluate the annual mean surface heat budget for the tropical warm pool, using many years of data, and infer a SWCF at the surface of -105Wm-2 and at TOA of -66Wm-2, which yields a ratio of 1.5.
In comparison to this direct and indirect observational evidence for a ratio of 1.5, current radiative transfer models, incorporating very little cloud absorption, produce a ratio close to 1. Apparently, the models are depositing too much solar flux at the surface compared to the measurements. Observations suggest that a further 40Wm-2 should be absorbed in the model atmosphere.
Solar absorption by clouds in the models is parameterized via the single scattering albedo, which is a function of cloud particle size. One way to correct the models is to adjust this particle size. (Larger particles absorb more radiation than smaller ones.) An effective particle size of 60-70 microns would be needed to achieve the correct absorption, but this is unrealistically large. In practice, for the purpose of the following climate simulations, the increased absorption is introduced by increasing the co-albedo in the visible region and near-infrared region.
Possible mechanisms for the observed enhanced cloud absorption include vapor-droplet 'overlap,' finite cloud effects, continuum absorption, and aerosols.
Implications for accuracy of GCMs Figure 9.1 shows the solar flux budget of the global annual mean atmosphere without enhanced cloud absorption. Including enhanced cloud absorption by changing the cloud single scattering albedo leads to an increased TOA solar absorption of 12 W/m2 (Figure 9.2). This reduction in planetary albedo from 0.32 to 0.29 would need to be tuned out in a GCM. Using cloud amount and cloud liquid water path to tune the TOA solar budget back to the control leads to a 40 W/m2 change in the partitioning of solar flux between the surface and the atmosphere (Figure 9.3). This amounts to roughly half the global mean latent heat flux. Possibly, this large amount of heat flux was missed because in past model runs SST was typically fixed, and many models overestimated surface latent heat flux.
Including the additional cloud absorption had the following effects: latent heat flux decreases by about 25W/m2 in the tropics, precipitation decreases, solar heating in the atmosphere is roughly doubled, the upper troposphere is warmed by about 6° C, upper tropical easterlies accelerate by about 16 m/s, the troposphere is significantly moistened, total average global cloud cover decreases by 4%, the Hadley cell is slowed down by about 10% (hence, surface winds decrease, consistent with the latent heat decrease). The model with enhanced cloud absorption agrees far better with Ramanathan et. al.'swarm pool heat budget discussed above. Incidentally, studies of the warm pool region indicate the need to include aerosol effects to eliminate 20 W/m2 bias in the surface clear sky flux.
On the whole, the addition of enhanced absorption improves the model's performance in the tropics and degrades it in the extra- tropics. Before enhanced absorption can be included operationally in GCMs, the cause of this absorption should be determined.
