FIGURE 14.52 (a) Absorptance and (b) transmittance measured on days with varying degrees of cloudiness using aircraft colocated above and below the clouds where broadband (224 nm to 3.91 /im), near-IR (680 nm to 3.30 jum), and spectral band (10-nm width centered at 500 nm) measurements were made (adapted from Valero et al., 1997a).
no evidence of increased absorption with increasing cloudiness; whatever the absorbing species is (if indeed there is one; see below) it does not appear to absorb light at 500 nm.
The difference between the broadband measured absorption (224 nm to 3.91 yu,m) and that from the near-IR radiometer (680 nm to 3.3 ^m) was used as a measure of the visible light absorption (Zender et al., 1991). On the cloudiest day during these studies, about 25% of the shortwave absorption was attributable to visible light, compared to 10% predicted by a model (Zender et al., 1997). These studies also report values for the ratio of the shortwave cloud forcing at the surface to that at the top of the atmosphere in the range of 1.36-1.65, in agreement with the warm pool energy balance studies, but again significantly larger than model-predicted values of 1.12-1.14. Zender et al. (1997) also show that several other measures of cloud absorption, such as the slope of plots of albedo versus transmittance, are consistent with excess absorption of solar radiation that is not included in the models. Measurements of the single-scattering albedo (the fraction of incident energy that appears as scattered radiation) inside a marine stratocumulus cloud implied a small contribution from enhanced absorption as well (King et al., 1990).
(Wiscombe and Welch, 1986). Cloud absorption is related to droplet size in a complex way (Stephens and Tsay, 1990) so that errors in droplet size measurements can alter the model-predicted absorption by a cloud. The treatment of absorption due to water vapor is another possibility. As discussed by Crisp (1997), the treatment of water vapor in models is simplified and may not properly reflect, for example, continuum absorptions between major bands in the near-IR. Model calculations suggest that the presence of a thin, saturated layer of water vapor above the clouds, for example, leads to increased absorption by 2-6% (Davies et al., 1984; Podgorny et al., f998). However, the Crisp calculations indicate that this cannot account for all of the observed excess absorption.
In addition, it has been suggested that the excess absorption is really due to radiation escaping from the sides of clouds, which would not be observed in measurements carried out above and below an isolated cloud (Ackerman and Cox, 1981). However, Valero et al. (1997a) argue that this should not give significant errors for continuous measurements made over clouds and analyzed with sufficiently long averaging times. Finally, whether the models properly capture the radiative effects, e.g., due to the heterogeneity of drop sizes in clouds and the cloud shape, is not clear (e.g., see Chou et al., 1995; Lubin et al., 1996; Byrne et al., 1996; and Loeb and Davies, 1996). For example, while many models assume plane parallel cloud geometry, the absorption for other shapes such as wavy, broken clouds can be different by as much as 10—15%, depending on the solar zenith angle (Podgorny et al., 1998).
One important aspect of radiation and clouds that may ultimately prove to be important in this issue of excess cloud absorption is the very long effective path lengths for light inside clouds due to multiple scatter ing processes. For example, Pfeilsticker et al. (1997) measured the absorption of solar radiation by the oxygen collision complex (02)2 under clear and cloudy sky conditions. The absorption is sufficiently weak that it is in the linear, "weak-absorber" regime (see Section A.3). Based on the Beer-Lambert law, the effective path length can be calculated from the known absorption cross section and atmospheric concentrations, combined with measurements of its atmospheric ab-sorbance. For one cloud, for example, this approach gave an extra effective path length of 135 km, about an order of magnitude larger than the clear-sky geometrical path length! The presence of even a relatively weak absorption that is not accounted for in the models could have a disproportionate effect with such large effective path lengths (e.g., see Kondrat'ev et al., f996a,b).
If this "excess absorption" by clouds is ultimately shown to be a real phenomenon, then an increased cloud formation and extent due to anthropogenic emissions may alter the radiative balance of the atmosphere not only through increased reflectance but also through increased absorption of solar radiation. Such an effect could impact atmospheric temperatures, their vertical distribution, and circulation, as well as surface wind speeds and the surface latent heat flux (Kiehl et al., 1995). Hence establishing if this is truly excess absorption, and if so, its origins, is a critical issue that remains to be resolved.
2. Feedbacks: Water Vapor, Clouds, and the "Supergreenhouse Effect"
Given the complexity of the ocean-atmosphere-biosphere system, it is not surprising that there are a number of feedbacks that greatly complicate the accurate prediction of the effects of anthropogenic or natural emissions. For example, warming is expected to lead to decreased amounts of clouds and hence to changes in cloud contributions to radiative forcing [see a comparison of model predictions for cloud feedbacks in Cess et al. (1996b)]. The DMS-CCN-cloud formation/ reflectance discussed earlier is another such example.
Another important example involves water vapor in the atmosphere. Water vapor is the most important greenhouse gas, and its concentration in the atmosphere is a function of temperature as given by the Clausius-Clapeyron equation:
Was this article helpful?