The Albedo of Fractal Stratocumulus Clouds

An increase in the planetary albedo of the earth-atmosphere system by only 10% can decrease the equilibrium surface temperature to that of the last ice age. Nevertheless, albedo biases of 10% or greater would be introduced into large regions of current climate models if clouds were given their observed liquid water amounts, because of the treatment of clouds as plane parallel. Past work has addressed the effect of cloud shape on albedo; here the focus is on the within-cloud variability of the vertically integrated liquid water. The main result is an estimate of the “plane-parallel albedo bias” using the “independent pixel approximation,” which ignores net horizontal photon transport, from a simple fractal model of marine stratocumulus clouds that ignores the cloud shape. The use of the independent pixel approximation in this context will be justified in a separate Monte Carlo study. The focus on marine stratocumulus clouds is due to their important role in cloud radiative forcing and also that, of the wide variety of earth's cloud types, they are most nearly plane parallel, so that they have the least albedo bias. The fractal model employed here reproduces both the probability distribution and the wavenumber spectrum of the stratocumulus liquid water path, as observed during the First ISCCP Regional Experiment (FIRE). The model distributes the liquid water by a cascade process, related to the upscale cascade of energy transferred from the cloud thickness scale to the mesoscale by approximately 2D motions. For simplicity, the cloud microphysical parameters are assumed homogeneous, as is the geometrical cloud thickness; and the mesoscale-averaged vertical optical thickness is kept fixed at each step of the cascade. A single new fractal parameter, 0 ≤ f ≤ 1, is introduced and determined empirically by the variance of the logarithm of the vertically integrated liquid water. In the case of conservative scattering, the authors are able to estimate the albedo bias analytically as a function of the fractal parameter f, mean vertical optical thickness Tν, and sun angle θ. Typical observed values are f = 0.5, Tν = 15, and θ = 60°, which give an absolute bias of 0.09, or a relative bias equal to 15% of the plane-parallel albedo of 0.60. The reduced reflectivity of fractal stratocumulus clouds is approximately given by the plane-parallel reflectivity evaluated at a reduced “effective optical thickness,” which when f = 0.5 is Teff ≈ 10. Study of the diurnal cycle of stratocumulus liquid water during FIRE leads to a key unexpected result: the plane-parallel albedo bias is largest when the cloud fraction reaches 100%, that is, when any bias associated with the cloud fraction vanishes. This is primarily due to the variability increase with cloud fraction. Thus, the within-cloud fractal structure of stratocumulus has a more significant impact on estimates of its mesoscale-average albedo than does the cloud fraction.