The modeling of an event such as an ozone lamina requires reproducing both the global and the small scales. In this study we report on a specific model capable of resolving such scale issues: the COMMID model, which has been developed by coupling a mechanistic model, MSDOL, with a high-resolution advection model, MIMOSA. MSDOL, which is forced toward National Centers for Environmental Prediction (NCEP) reanalyses below 100 hPa, provides a consistent picture of the stratospheric large-scale circulation from which MIMOSA simulates the fine-scale filaments generated by breaking planetary waves in the stratosphere. To evaluate the performances of the model, we present results for a particular event of tropical-air intrusion at midlatitudes across the southern subtropical barrier observed in July 2000 and described in part 1 (Portafaix et al., 2003). The model is used to examine the contribution of each wave to the structure and the development of that event. The methodology consists in filtering the NCEP tropospheric forcing by zonal wave number and by phase speed. Our results show that mixing is significantly reduced precisely at the locations where the phase speeds of the filtered waves are close to the speed of the mean zonal wind, thus confirming the findings of previous studies. However, what is important here is that they validate the use of an approach based on the coupling of two models. The next step will consist in using the COMMID model in a more general way for further investigations of the impact of the tropospheric circulation on the isentropic transport in the stratosphere for climate sensitivity purposes.