To develop a parameterization of shallow convective clouds and precipitation induced by land-surface processes in large-scale atmospheric models, a set of simulations was performed with a high-resolution 3D atmospheric model. Results indicate that landscape discontinuities could significantly affect the timing of onset of clouds, and the intensity and distribution of precipitation. In most cases, landscape discontinuities enhance precipitation. Current parameterizations of clouds and precipitation in atmospheric models do not account for such effects. Similarity theory was used to develop such a parameterization. For this purpose, Buckingham Pi Theory, a systematic method for performing dimensional analysis, was used to derive a set of dimensionless groups, which describes the large-scale atmospheric background conditions, the spatial variability of surface sensible heat flux, and the characteristic structure of the landscape. These dimensionless groups were used to calculate the coefficients of a Chebyshev polynomial, which represents the self similar vertical profiles of dimensionless mesoscale heat fluxes obtained for a broad range of large-scale atmospheric conditions and different landscapes. The numerous 3D numerical experiments performed to evaluate this similarity relationship suggest that the parameterization provides reasonable results