This study investigates aerosol source functions and the evolution and optical properties of biomass burning smoke plumes by using an aerosol transport and microphysical model. Since the most extensive biomass burning occurs in the Southern Hemisphere, this work focuses on smoke plumes over Africa and South America. The most extensive field studies that took place in these regions include the Southern African Regional Science Initiative (SAFARI 2000) and the Smoke Aerosols, Clouds, Rainfall and Climate (SMOCC) campaign. In general, observations collected during these campaigns can be used as model input parameters to reproduce the measured optical properties. On a monthly basis, the simulated optical properties compare well to satellite, airborne, and ground-based observations. The simulated spatial distribution of smoke aerosol loading over these regions also compares well to satellite observations, suggesting that the transport processes in the model are adequate. The results also suggest that the emissions provided by the Global Fire Emissions Database are too low, especially over South America. However, wet deposition may be too aggressive in the model over this region. Since the model wet deposition does not include the feedback smoke may have on cloud formation and precipitation suppression, too many aerosols may be removed. This work also allows conclusions to be drawn on the distinction between the properties of smoke aerosols from savanna and forest fires in climate models. In particular, similar initial particle size distributions and aerosol optical properties can be used to simulate smoke plumes produced by both vegetation fires. However, to reproduce the observed smoke optical properties over South America, humidification of smoke aerosols needs to be considered. Model results and observations both suggest that the typical single scattering albedo of smoke over South America and Africa differ because of relative humidity, not vegetation type. The optical properties of smoke aerosols where the particles are assumed to be homogeneous or heterogeneous internal mixtures are also found to be consistent with field studies. Additionally, treating more realistic microphysical processes in a climate model produce different direct radiative forcing estimates than a model that does not. Overall, this work suggests strategies for improving the treatment of smoke plumes in climate and microphysical models.