| Among all the compositions of atmosphere, ozone is one of the most important trace gases. It distributes in stratosphere mainly and can intensely absorb short wave ultra-violet radiation, which prevents the earth creatures from being hurt. What's more, as a greenhouse gas, ozone can absorb infrared radiation and thus influences the global temperature or climate. Nowadays, a complete and multi-platform ozone observing system, including ground-based observation, aerial observation and satellite remote sensing, have been developed, thus a long-term and continuous observation of ozone is realized. Compared with the ground-based and aerial observing, satellite remote sensing, especially for the limb observing technology, can provide higher spatial and temporal resolution data, which makes the near-global and high vertical resolution ozone retrieval possible. This study aims at the ozone profile retrieval by Differential Optical Absorption Spectroscopy (DOAS) from OSIRIS limb scattering data. The DOAS technology can remove the spectral features caused by Rayleigh and Mie scattering. Therefore, it has a great advantage in the ozone profile retrieval.The radiance observed by satellite is the solar radiation attenuated by particle scattering or gas absorption. The scattering varies slowly with wavelength, while the absorption caused by all kinds of gasses varies rapidly with wavelength. Therefore, the ozone column density can be obtained from the rapidly varying components once the slowly varying effects are removed. In this study, the total optical depth was calculated by the ratio of the chosen reference spectrum and the radiance observed at other tangent heights, and the slowly varying components of it were separated from the rapidly varying components by means of second order polynomial fitting method. Then, through the least-squares fitting method, the column density of ozone was obtained.The column density of ozone was converted to number density by Multiplicative Algebraic Reconstruction Techniques (MART) with SCIATRAN as forward model and the weighting factors calculated by means of perturbation.The OSIRIS limb scattering data used in this study covered eight areas. According to the characteristics of the chosen data, the tangent height near fifty kilometers was chosen as reference tangent height, and the wavelength from 571 to 617 nm was chosen as fitting spectrum. The weighting factors were calculated with a perturbation coefficient of 5%. By taking the absorption of O3, NO2 and O4 into consideration and taking the 66% of the ozone profiles in SCIATRAN as the priori profile, the 10~48 km ozone profiles, with a resolution of 1 km, were obtained. By comparison, the retrieved profiles have a great consistency with the SCIMACHY ozone profile products and the laser radar observing profile from Anhui Laser Researches in tendency and magnitude. With respect to the SCIAMACHY ozone profile product, the relative deviation is less then 10% between 10 and 48 km, and less than 5% between 17 and 47 km. While for the ground-based radar observing profile, the relative deviation is within 22% from 21 to 36 km. Analysis suggested that the main reason causing the deviation is the difference between model parameters and the true state of atmosphere and earth, and the inconsistency of the observing time and location between the OSIRIS data and the contrast data.The sensitivities of aerosol, cloud and NO2 to ozone retrieval were conducted finally, and the results showed that DOAS removed the influence of aerosol and cloud effectively. In the fitting spectrum window, the content of NO2 has no impact on the retrieved ozone profile.
|
PDF Full Text Request