The upper branch of the Brewer-Dobson circulation (BDC) is only partly represented in state-of-the-art chemical climate models which mostly do not include the mesopause region, similar to chemical transport models which depend on the vertical domain of the applied analyses. As the atmospheric lifetime of some long-lived trace gases as for example CH4 or N2O is influenced by mesospheric processes, the representation of the upper branch in the model may have an impact on derived trends of such gases. The thermal and wind structure of this height region in the model strongly depends on the gravity wave drag parameterization and the chosen tuning parameters. Here we study the upper branch of the BDC with the KASIMA model which includes the mesopause region up to the lower thermosphere, when nudged to ERA-Interim analyses in the stratosphere. We adapt our Holton-Lindzen type GW drag parameterization in order to improve cross mesopause transport which is often observed after strong mid-winter sudden stratospheric warmings and analyze its impact on trends of mean age of air and long-lived tracers in the middle atmosphere.