Large uncertainties still remain about the mechanisms of genesis and interaction of clouds in a large scale environment. Through a semi-empirical dynamical model using hourly data from TOGA-COARE IOP (130E--180E and 15S--15N, 01.11.91--28.02.92), we show that the convective activity leading to clouds is a self-regulating, threshold process embedded in a large-scale environment. The convective activity is prescribed through a of hourly fractional cloud cover while the environmental forcings are defined in terms large-scale moist state defined in terms of the average brightness temperature. A prognostic model is then developed for cloud cover as the dynamical variable. To validate the model we integrate it with a given initial condition from observed data. The model reproduces the observed spatio-temporal structure of the cloud with a remarkable degree of success. We identify and quantify two thresholds for large-scale forcing for genesis and intensity of high clouds.