STRUCTURAL CHARACTERISTICS OF DEEP CONVECTIVE SYSTEMS OVER TROPICAL AFRICA AND THE ATLANTIC-OCEAN

The structural properties of convective cloud clusters of tropical Africa and the Atlantic Ocean are studied using six summers of Meteosat satellite data in the atmospheric infrared window. A cluster at a given brightness temperature threshold is defined as the area covered by adjacent cloud cells with brightness temperature lower than the threshold. The clusters are classified according to the area they cover and the position of their center of mass. Results show that the convective cluster number can be approximated by a power law of the radius with an exponent around -2. This gives a nearly equal contribution of each cluster size to the mean high cloud cover for a given brightness temperature threshold. Using the visible channel (0.4-1.1-mu-m) of Meteosat, we show that the part of the clusters with reflectance larger than 0.7 also follows a power law. The cluster-size distributions remain similar for different subregions and seasons, even if they are subject to large variations in the mean cloudiness. We further inspect the relatively large diurnal and interdiurnal variations of the cluster-size distribution. We also look at the variations of the cluster-size distribution as a function of a vertical extension, defined as the lowest brightness temperature reached by the cluster. We find preferential sizes that increase as the vertical extension increases. We also show that the distance between clusters, defined as the minimum distance between clouds of the same size, also follows a well-defined distribution.