Natural ventilation is an effective measure to save energy consumed in buildings and to improve indoor air quality. This investigation studied single-sided natural ventilation by using a computational fluid dynamics (CFD) model, together with analytical and empirical models. The CFD model was applied to determine the effects of buoyancy, wind, or their combination on ventilation rates and indoor conditions. For buoyancy-driven flow, the CFD results are within a 10% difference from the semi-analytical results. For combined wind- and buoyancy-driven flow, CFD may have underpredicted the empirical model results by approximately 25%. This investigation also studied the effects of opposing buoyancy and wind forces. (C) 2002 Elsevier Science B.V. All rights reserved.