The global distribution of ozone in the troposphere and lower stratosphere calculated with a three-dimensional chemistry transport model driven by European Centre for Medium-Range Weather Forecasts (ECMWF)-analyzed meteorological fields has been compared with observed ozonesonde profiles. This comparison is presented in a new graphical format, which shows in a single panel the vertical and seasonal dependence. The modeled ozone profiles compare reasonably well with climatological ozonesonde data for various stations all over the world, especially if the variability of the ozone concentrations is taken into account. However, the ozone mixing ratios in the upper troposphere and lower stratosphere at midlatitudes are generally overestimated by the model. This is probably caused by a combination of an overestimation of the stratosphere-troposphere exchange and the absence of heterogeneous reactions in the lower stratosphere which reduce ozone. The latitudinal dependence and seasonal dependence of the observations are reproduced by the model calculations, except for the ozone concentrations at the surface. This might be due to the neglect of nonmethane hydrocarbons, which give rise to photochemical ozone production during summer, although other factors such as emissions and deposition cannot be ruled out. The ozone column density obtained by combining calculated ozone distributions up to 50 hPa with climatological zonal mean data for ozone above 50 hPa compares reasonably well with total ozone mapping spectrometer (TOMS) observations. Particularly, the variability caused by synoptic features observable in modeled total ozone shows a high degree of correspondence to the observations. This indicates that rapid variations in the ozone column density are mainly the result of corresponding variations in ozone concentrations near the tropopause due to transport. Modeled total ozone is generally underestimated in the tropics and overestimated elsewhere compared to the TOMS observations. This overestimation, which is large in spring at northern midlatitudes and increased toward the pole, can partly be ascribed to differences between the prescribed climatological ozone concentrations above 50 hPa and the actual values in 1990. Budget calculations of tropospheric ozone showed that reducing these prescribed ozone concentrations lowers the ozone input from the stratosphere, which is largely compensated by a higher photochemical ozone production in the troposphere such that deposition, which depends on the ozone concentrations in the lower troposphere, remains almost unaffected. The reason for the overestimation needs to be investigated further using in situ measurements of several trace gases simultaneously in order to better understand the chemical processes involved. In this study a methane and carbon monoxide oxidation chemistry scheme has been employed without stratospheric chemistry. Furthermore, the comparison of TOMS total ozone observations in the tropics with model calculations seems to suggest that the treatment of ozone precursors such as the NOx emissions by lightning and biomass burning needs to be improved. The reduced correlation between observed and modeled total ozone fields at southern midlatitudes can probably be ascribed to the lower quality of the analyzed fields in that area. Because of the above mentioned discrepancies the calculated atmospheric impact of anthropogenic emissions needs to be interpreted with care. ¿ 1998 American Geophysical Union