Our objective is to examine the potential application of a k-epsilon model (TEMPEST) for simulating flow-field and dispersion patterns around a cubical building under stably stratified approach Row. The results of these numerical simulations are compared with available measurements from towing-tank experiments on dispersion around a cube. The modeled and measured Row and concentration patterns are compared in order to ascertain the effects of changes in the stratification, which is characterized by the Froude number Fr = U/NH (where U is the approach-Row wind speed, N is the Brunt-Vaisala frequency, and H is the building height-large Fr implies weak stability and vice versa). Both the model and experimental results suggest that the Row structure is independent of stratification when Fr greater than or equal to 6. The model predicts a moderate increase in the length L(c) of the lee-side ''cavity'' (recirculation region) as the Froude number is decreased from 6 to 3 (moderate stratification), then a sharp decrease in L(c) as Fr is further decreased to 1 (very strong stratification). The measured results, however, show L(c) to be constant as Fr is decreased to 3, then a sharp decrease as Fr is further decreased to 1. The agreement between the measured and model-predicted concentration field when a source was located within the recirculating cavity behind the building was fair under weakly stratified conditions (Fr greater than or equal to 3), but poor under strongly stratified conditions (Fr = 1). These simulations show that, in strongly stratified conditions, mean advection plays a more important role than does turbulent diffusion because the turbulence is very weak in such flows. And because the Froude number will rarely be less than about 3 in the nighttime stable boundary layer, these results suggest that stratification will rarely be a significant factor influencing the flow structure in the near-vicinity of a building.