University of California Santa Barbara
Santa Barbara, CA 93106
Clouds act as the primary modulator of the Earth's radiation at the top of the atmosphere, within the atmospheric column, and at the Earth's surface. They interact with both shortwave and longwave radiation, but it is primarily in the case of shortwave where most of the uncertainty lies because of the difficulties in treating scattered solar radiation. To understand cloud-radiative interactions, radiative transfer models portray clouds as plane-parallel homogeneous entities to ease the computational physics. Unfortunately, clouds are far from being homogeneous, and large differences between measurement and theory point to a stronger need to understand and model cloud macrophysical properties. In an attempt to better comprehend the role of cloud morphology on the 3-dimensional radiation field, a Monte Carlo model has been developed. This model can simulate broadband shortwave radiation fluxes while incorporating all of the major atmospheric constituents. The model is used to investigate the cloud absorption anomaly where cloud absorption measurements exceed theoretical estimates and to examine the efficacy of ERBE measurements and cloud field experiments.
The Monte Carlo method is essentially a direct simulation of the physical processes involved in radiative transfer, in which the path of a photon is described by probability functions. These functions describe the distance a photon travels before an interaction with an atmospheric particle, if a photon is then absorbed or scattered, and if scattered, the direction the photon is scattered. The Monte Carlo model is designed around a 3-dimensional spatially dynamic structure. This structure provides for large low-variability cells to reduce computations and nested higher-resolution cells to allow the mixing of various spatial scales. Each cell can contain any combination and concentration of gases, aerosols and cloud droplets. The cloud fields used in the model are based on modified multifractal fields of liquid water. Broadband fluxes representing upwelling, downwelling, and absorbed shortwave radiation are derived from a few unique wavelengths selected through a principal component analysis and a stepwise regression process.
Results from a tropical scenario with cloud optical depths ranging from 40 to 160 and cloud tops from 1 to 8 kms are displayed in Figure 13.1. The surface downwelling image shows surface fluxes approaching the solar constant from the effect of radiation being reflected from the sides of the towering cumulus. The upwelling fluxes are shown for the top of the cloud and at the top of the atmosphere (TOA). The flux at the TOA is diffused to such an extent that it is difficult to make out features of the cloud. Although ERBE claims to be an estimate of TOA flux, in reality it is a flux derived from TOA radiances where spatial features are easy to distinguish. The absorption image shows the highest column absorption occurs in cloud "valleys" where photons are focused causing enhanced water vapor absorption.
Figure 13.2 is a comparison of the Monte Carlo model results for a stratus cloud and independent pixel results using a plane-parallel model. As shown, higher surface downwelling and lower cloud top upwelling fluxes occur using the Monte Carlo model. Additionally, enhanced atmospheric absorption on the order of 30Wm-2 obtained from the Monte Carlo model, suggesting the role of cloud morphology as a possible source for the cloud absorption anomaly. Figure 13.3 displays the difference between calculated atmospheric column absorption using a single- column budget approach and the "measured" absorption derived directly from the Monte Carlo model. As can be readily seen, there are large discrepancies between the two results, caused by the budget approach being unable to account for the redistribution of radiation in the horizontal plane. This result suggests that field experiments which use simultaneous measurements above and below a cloud to compute cloud absorption may produce erroneous results.
