Measurements and modeling of solar spectral absorption by liquid water clouds

Cloud absorption of solar radiation has been a controversial topic from the earliest attempts at reconciling aircraft measurements with theoretical calculations of cloud absorption. The measurement of cloud absorption from aircraft above and below a cloud layer (flux divergence) is made difficult by issues of cloud heterogeneity, aircraft coordination, and the radiometric accuracy of the instrumentation. Additionally, nearly all previous measurements of cloud absorption have been made with broadband radiometers. Broadband measurements are fundamentally incapable of distinguishing between the various absorbing constituents, namely, cloud liquid water, gases (most importantly, water vapor), and aerosol particles. Thus, the interpretation of the broadband measurements is problematic. Here we report on spectrally resolved (400-2100 nm, 416 channels) cloud absorption measurements from the Solar Spectral Flux Radiometer (SSFR) from two extensive Pacific marine stratus cloud systems encountered during the Tropical Composition, Cloud and Climate Coupling Experiment (TC4). Measurements of cloud spectral absorptance are compared with detailed plane-parallel radiative transfer calculations. For a high optical thickness and spatially uniform cloud case, the agreement between measurement and model absorptance across the spectrum is better than 0.05, and better than 0.01 in the visible. A lower optical thickness case shows slightly larger residuals between measurement and model. Measurements and modeling are used to illustrate and quantify the various mechanisms of cloud layer absorption including the role of aerosols. A novel technique for single aircraft flux divergence measurements based on conditional sampling (Marshak, 1999) is also presented. The technique uses randomized pairs of above- and below-cloud net irradiances matched at a single non-absorbing wavelength to account for cloud net horizontal divergence. Statistics are calculated from these pairs for the complete spectral range, giving an estimate of the entire cloud scene spectral absorptance. Spectral absorptance measurements of the same cloud scene made from two aircraft with large vertical separation are shown to give less robust results.