Tropical deforestation is a current anthropogenic change to the land surface. Tropical forests are known to be changing rapidly, but the rate and extent of deforestation and the resulting secondary cover, as would be needed for quantitative projections of future change, are difficult to establish now. Also necessary for climate simulations is a description of the micrometeorological processes within the forest canopy, including especially the processes of evapotranspiration and interception. Past General Circulation Model sensitivity studies have established the potential major significance of perturbation of surface energy processes. Climate modeling sensitivity studies and ecological systems studies have highlighted the importance of the surface energy balance and the hydrological cycle over continental areas. However, predicting even the local, immediate effects of replacing tropical broadleaved forest with impoverished grassland is difficult because the land-surface parameterization schemes used in most climate models have been inadequate. A wide range of parameterization schemes in past studies of the climate change with tropical deforestation have led to divergent conclusions as to the consequences. The most recent such study has attempted to include more realistic processes than past studies, including separate submodels for soil and forest canopy. Results from the simulations of this study are reviewed. With attempts to simulate realistically the climatic impact of tropical deforestation, several issues must be considered. What is the rate and extent of the deforestation? What land cover replaces the forest and what physical properties of the forest and its replacement are most important for the simulation? Are the parameterizations of these properties plausibly correct and are the model 'predictions' explicable in terms of these properties? A simulation methodology that treats these latter questions is proposed.