Atmospheric sphericity is an important factor that must be considered in order to evaluate an accurate ozone loss rate in the polar stratosphere. The built-in plane-parallel radiative transfer scheme of a nudging chemical transport model (CTM) and an atmospheric general circulation model (AGCM) with coupled chemistry is modified by a pseudospherical approximation. The plane-parallel atmosphere radiative transfer version (PPA version) is compared with the pseudospherical atmosphere radiative transfer version (SA version) for both the nudging CTM and AGCM. The nudging CTM can isolate the chemical effects for a given dynamical field, while the interaction among the chemical, radiative, and dynamical processes can be studied with the AGCM. The present analysis focuses on Antarctica during an ozone hole period. In the ozone loss period over Antarctica, ozone starts to decrease earlier and minimum value of total ozone becomes lower in the SA versions of both the nudging CTM and the AGCM than in the corresponding PPA versions. The ozone mixing ratio decreases earlier in the SA version because of an earlier increase of ClO concentration initiated by the upward actinic flux at solar zenith angles greater than 90¿. Dynamics plays an important role as well as the chemical processes. During the ozone recovery period, the ozone distribution becomes almost the same in the SA and PPA versions of the nudging CTM, while in the AGCM the ozone amount in the SA version remains at lower values compared to those of the PPA version. In the AGCM, a decrease of ozone over Antarctica enhances the latitudinal gradient of temperature and thus strengthens the polar vortex in the SA version. A resultant delay of the polar vortex breakup causes the delay of the ozone recovery. For the AGCM, ensemble runs are performed. The ensemble experiment exhibits large ozone variances after the middle of December, when the ozone recovery is dynamically controlled. Most ensemble members of the AGCM show a delay of the polar vortex breakup in the SA version, while a few members show opposite results. In the latter members, the polar vortex breakup is strongly affected by the enhanced EP flux from the troposphere around 100 hPa, which causes the variances in the ozone recovery period. Most members, however, do not show large statistical variances; that justifies the conclusions from the ensemble means.