Radiative processes constitute the ultimate source and sink of energy in the climate system. They are generally well known, and particularly the clear-sky long-wave transfer, including the absorption properties of most greenhouse gases. Their treatment in atmospheric general circulation models relies on several approximations: the fluxes are computed as averaged quantities over a few spectral intervals, the propagation is limited to the vertical upwelling or downwelling directions, and the role of sub-grid scale features of the clouds or aerosols is essentially neglected. Although this methodology is believed to have only a marginal impact on the accuracy of computed long-wave fluxes, the analysis of satellite, ground-based or aircraft measurements (Cess et al., 1995; Pilewskie and Valero, 1995; Ramanathan et al., 1995) has generated a concern that the radiative algorithms used in climate models could significantly underestimate the atmospheric shortwave absorption. Similar results have been obtained as part of the ARM/ARESE experiment (Valero et al., 1997; Zender et al., 1997). The excess, or "anomalous" observed absorption may reach typical values of about 30 to 40 Wm-2. Its very existence, however, remains controversial and is at odds with other investigations. Li et al. (1995) analysed large global data sets from satellite (ERBE) and ground observations (GEBA). They did not find the anomalous absorption except for a few tropical sites over a short period of time. The exception appears to be induced by enhanced absorption due to biomass burning aerosols (Li, 1998). Limited accuracy of the measurements (Imre et al., 1996) and the methodology of analysis (Arking et al., 1996; Stephens, 1996; Barker and Li, 1997) have been raised as possible contributing factors to the finding of the anomalous absorption.
A comparison of the ARM-ARESE measurements with a state of the art radiation code, which uses measured atmospheric quantities as an input, and relies on the same physical assumptions as used in climate models, indicated that the anomalous absorption is larger for cloudy than for clear conditions, and also tends to be larger for visible rather than for near-infrared fluxes (Zender et al., 1997). These two characteristics are consistent with the results of a comparison between the output of the CCM3 general circulation model and observations from Nimbus 7 (Collins, 1998). However, Li et al. (1999) analysed all the data sets collected during the ARM/ARESE experiment made by space-borne, air-borne and ground-based instruments and did not find any significant absorption anomaly. They traced the source of controversial findings to be associated with inconsistent measurements made by some air-borne radiometers as used in the studies of Valero et al. (1997) and Zender et al. (1997) with those from all other instruments. However, other studies (Cess et al., 1999; Pope and Valero, 2000; Valero et al., 2000), that do find a significant absorption anomaly, have further demonstrated that the ARESE data do indeed satisfy a number of consistency tests as well as being in agreement with measurements made by other instruments. Meanwhile, evidence of enhanced cloud absorption has been found from measurements of the MGO aircraft laboratory (Kondratyev et al., 1998). Inclusion of anomalous absorption has been found to improve the representation of atmospheric tides (Braswell and Lindzen, 1998). There finally exists evidence of an effect associated with clear-sky conditions - in the presence of aerosols: models tend to overpredict clear-sky diffusion to the surface (Kato et al., 1997).
The three dimensional nature of the solar radiation diffusion by cloud (Breon, 1992; Cahalan et al., 1994; Li and Moreau, 1996) is unaccounted for by present climate models and may play some role in anomalous absorption. In a recent study using a Monte-Carlo approach to simulate three-dimensional radiative transfer, O'Hirok and Gautier (1998a; 1998b) show that cloud inhomogeneities can increase both the near-infrared gaseous absorption and cloud droplet absorption in morphologically complex cloud fields. The enhancement is caused when photons diffused from cloud edges more easily reach water-rich low levels of the atmosphere and when photons entering the sides of clouds become trapped within the cloud cores. The amplitude of those effects remains limited to an average range of 6 to 15 Wm-2, depending on the solar angle. The inhomogeneities significantly affect the vertical distribution of the heating, though, with potential consequences on cloud development. Incorporating the effect of cloud inhomogeneities in radiative algorithms may become a necessity, and recent efforts have been made along that path (Oreopoulos and Barker, 1999).
If anomalous absorption turns out to be real, it is an effect that will need to be incorporated into radiation schemes. Evaluation of its importance is hampered by lack of knowledge of a physical mechanism responsible for the absorption, and hence lack of a physical basis for any parametrization. Modelling studies by Kiehl et al. (1995) have demonstrated the sensitivity of the simulated climate to changes in the atmospheric absorption. As a radiative forcing, anomalous absorption is fundamentally different from water vapour or CO2 in that it does not significantly alter the Earth's net radiation budget. Instead, it shifts some of the deposition of solar energy from the ground to the atmosphere (Li et al., 1997), with implications for the hydrological cycle and vertical temperature profile of the atmosphere. Anomalous absorption may not, however, appreciably affect climate sensitivity (Cess et al., 1996).
Model validation is also affected. Many of the data which are used to tune or validate model parametrizations, such as the Liquid Water Path (LWP) or the droplet equivalent radius, are obtained from space measurements by the inversion of radiative algorithms, which ignore cloud inhomogeneities and anomalous absorption. This gives a strong importance to satellite instruments such as POLDER which provide measurements of the same scene at a variety of viewing angles, and provides a good test of the plane-parallel hypothesis in retrievals of cloud quantities.
Other reports in this collection