On the intensity of atmospheric convection

Abstract

Several runs with a numerical cloud model are discussed in order to quantify the effect of latent heat release on the development of convective activity' averaged over some region. A scheme of the numerical results is constructed by means of a simple gradient system, with surface heat flux and moisture content as control parameters. For some range of these control parameters, two stable steady states occur. Internal forcing due to latent heat release creates the bifurcation. On the basis of Monte Carlo integrations with the gradient system it is shown that for moist conditions convective activity tends to have a bimodal probability distribution. The practical consequence is that the shortrange predictability of the weather is substantially reduced.