Comparison of modelled ozone distributions with ozonesonde observations in the Tropics

Modelling as a tool for supporting research on major scientific questions in atmospheric chemistry and climate research, has received a lot of attention during the last decades. Many aspects of modelling atmospheric processes have been investigated.

In this report, the global three-dimensional chemistry transport model of the Royal Netberlands Meteorological Institute (KNMD), TM3, is used to compare ozonesonde profiles in the tropics with corresponding modeled ozone distributions in order to validate the model results for ozone. TM3 uses 6-hourly meteorological data from European Centre for Medium-Range Weather Forecasts (ECMWF). analysis and includes parameterisations for subgrid scale processes such as convection. TM3 includes an ozone chemistry module containing the methane and carbon monoxide oxidation chain for the troposphere and lower stratosphere.

Our study was limited to three ozonesonde stations in the tropics: Ascension Island, Nairobi and San Cristobal. Generally, monthly mean modeled azone profiles compare reasonably weil for all three stations with the natural variability constructed from all the available observations in the corresponding montha, except for the upper stratosphere (above 50 mbar). In the upper stratosphere, an evident systematic deviation is visible. From this, it can be concluded that the prescribed climatology for the upper stratosphere in TM3 is less reliable for the tropics, due to the low number of included observations in the tropics.

Furthermore, the seasonality of ozone for the ozonsonde stations Ascenaion Island and Nairobi has been investigated. For Ascension Island, which is located on an island far away from the continents and therefore should not be affected by local pollution, the seasonality is reproduced quite well by TM3, except for
- again - the upper stratosphere. For Nairobi, which is located at the equatorial East-const of Afrika, the seasonality is only reproduced reasonably well in the lower troposphere.

Finally, an attempt has been made to investigate the possibility to reproduce daily changes in ozone profiles with TM3. For this, individual soundings in September 1998 at the San Cristobal station were simulated. The conclusion is that the modeled ozone profiles are not accurate or sensitive enough to show detailed features. Probably this can be explained by the very high spatial and temporal resolution required to reproduce these fine structures. The used model resolution of 8 x 10 x 19 vertical layers is probably not fine enough.