Below answers are provided to the four questions given in the introduction: 1. Can Kelvin waves be detected in total ozone column and ozone pro??le measurements from GOME? 2. Is the vertical resolution of the vertical ozone pro??les from GOME su??cient to provide information on the vertical structure of the Kelvin wave activity? 3. How well do the Kelvin wave signals in the GOME measurements agree with Kelvin wave signals in temperature and wind data from the European Centre for Medium-Range Weather Forecasts (ECMWF)? 4. Can the diabatic descent or ascent rates in the polar vortices be determined from ILAS measurements of N2O? Question 1 has been addressed in chapters 3 and 5, question 2 in chapter 5, question 3 in chapter 4 and question 4 in chapter 6. 1. The variability of total ozone columns and pro??les in the tropics measured by GOME, has been analysed by using a bi-dimensional spectral method. The method has been developed by extending an unequally spaced data technique ??rst developed by Lomb from one to two dimensions. The method calculates which frequencies are dominant in a 60-day period and allows an easy determination of the statistical significance of an observed signal. In 7 years of GOME ozone column data three 60-day periods of high Kelvin wave activity are detected that correlate with westward equatorial zonal winds at 30 hPa: 15 July to 13 September 1996, 17 July to 15 September 1998 and 19 September to 18 November 2000. The three periods show signi??cant signals at frequencies corresponding to eastward propagating waves with zonal wavenumbers 1 and 2 and wave periods of 12-15 days. The induced ozone perturbations of 1-2 DU can be attributed to 'slow' Kelvin waves in the lower stratosphere. Furthermore 96 Summary and outlook three di??erent datasets of GOME ozone pro??les have been investigated: the OPERA ozone pro??les derived using an algorithm based on optimal estimation, the NNORSY ozone pro??les derived using a neural network and a dataset consisting of assimilated OPERA ozone pro??les. All three datasets show Kelvin wave signals similar to those in the GOME ozone columns for the three high Kelvin wave activity periods identi??ed in the GOME ozone columns. 2. Concerning the vertical distribution of the Kelvin wave activity, the OPERA and NNORSY ozone pro??les provide some additional information compared to the GOME ozone columns. Due to their limited vertical resolution, both datasets cannot fully resolve the vertical structure of the Kelvin waves but do provide some heightresolved information. They show that the Kelvin wave induced ozone uctuations have their largest amplitudes between 20 and 30 km, which is in agreement with theory. The assimilated OPERA ozone pro??les that make use of a global chemistrytransport model driven by ECMWF meteorological ??elds show clearer height resolved information on the Kelvin wave activity identi??ed in the GOME ozone columns. This information probably originates from the ECMWF meteorological ??elds and shows Kelvin wave characteristics that agree with other studies. Maximum Kelvin wave induced uctuations with an amplitude of 0.65 DU/km (or 0.2 ppmv in mixing ratio) are found around 35 hPa, where the vertical gradient in ozone is largest. This is expected when the kelvin wave perturbations are induced by transport e??ects. The ozone mixing ratio also shows large Kelvin wave induced uctuations between 10 and 1 hPa. These uctuations diminish when making the conversion to DU and hence do not contribute largely to the total column perturbations. Vertical integration of the ozone perturbations leads to total ozone column perturbations between 1 and 2 DU in agreement with the perturbations in the GOME ozone columns. 3. ECMWF Re-analysis (ERA-40) ??elds of zonal wind and temperature are also analysed by using the same spectral method as is used for the GOME ozone measurements. For all three periods where the GOME total ozone columns display high Kelvin wave activity, the ECMWF ??elds manifest corresponding Kelvin wave activity. Spectral features associated with Kelvin waves of zonal wavenumber 1 and 2 and wave periods of 15-20 days are identi??ed. The features are present between 100 and 10 hPa. The correlation between the Kelvin wave signals in the ECMWF ??elds and GOME ozone columns is good. The Kelvin wave induced perturbations in zonal wind and temperature are respectively up to 8 m/s and 2K. Assuming only transport e??ects play a role, these perturbations could lead to total ozone column perturbations of around 1 DU and maximum ozone perturbations around the height of maximum vertical gradient of ozone. This is in agreement with the results from the GOME ozone columns and assimilated OPERA data. The amplitude of the calculated maximum ozone perturbations also shows a good agreement with the amplitudes determined from the assimilated OPERA ozone pro??les. The maximum ozone perturbations around 35 hPa in the assimilated ozone pro??les are nearly in phase with the ERA-40 temperature perturbations. This indicates that the ozone perturbations in this altitude region are mainly caused by transport e??ects. The large ozone mixing ratio uctuations between 10 and 1 hPa are out of phase with temperature perturbations. Further studies will be needed to investigate how large the contributions from trans7.2 Outlook 97 port and photochemical e??ects are on these ozone perturbations. 4. Diabatic descent and ascent rates inside and outside the polar vortices have been inferred from the time evolution of N2O measured by the ILAS instrument. The measurements cover the period of November 1996 to June 1997. Before the calculations are performed the data is regridded onto an equivalent latitude, potential temperature and time coordinate frame. Isopleths of N2O are tracked to determine the descent or ascent rates inside, outside or at the edge of the polar vortices. In the Southern Hemisphere vortex strong diabatic descent is visible preceding the vortex formation with values up to 8 K per day. The descent continues inside the polar vortex after its formation with values around 0.4 to 2 K per day. The values are in agreement with other studies. In the Northern Hemisphere diabatic descent is visible inside the polar vortex below the potential temperature level of 700-800 K. The values are smaller than in the Southern Hemisphere, from 0 to 0.8 K per day, which lies in the range of values calculated in other studies. Following vortex break up diabatic ascent rates are seen with values up to 1 K per day. Concluding, the work described in this thesis shows the potential of GOME measurements to contribute to a global description of Kelvin wave activity The 3- dimensional structure can be analysed by using a combination with ERA-40 data or using assimilation of the GOME ozone pro??les. Furthermore Kelvin wave signatures provide a tool to check the quality of the GOME ozone data in the tropics. This thesis also shows that satellite measurements of N2O are a useful mean for the determination of diabatic descent and ascent rates. Overall the thesis demonstrates the ability of satellite measurements to improve the monitoring and understanding of dynamical processes in the atmosphere on a global scale.
|Kwalificatie||Doctor in de Filosofie|
|Datum van toekenning||6 apr 2005|
|Plaats van publicatie||Eindhoven|
|Status||Gepubliceerd - 2005|