Abstract
The paper describes a procedure for predicting the luminance distribution of a cloudy sky. It can be used to relate ground-based observations to cloud amount and cloud reflectance seen from satellites, and to give a relationship between cloud amount and sunshine hours. The shape of clouds is modelled stochastically with probability distributions of cloud length and intercloud distance. The effect of multiple scattering within clouds is then examined using Monte Carlo simulation; empirical functions are fitted to the results from this. The resulting equations are used in a calculation procedure which begins with the luminance distribution of the upper sky and then computes in turn the luminance pattern beneath successive cloud layers, including effects of interreflection. The model is based on cumulus clouds; it can, by varying the parameters, be used to predict sky luminance with other types of cloud.
Get full access to this article
View all access options for this article.
