Ozone Profile Retrieval And Aerosol Classification Based On The Satellite Measurement

Firstly, the studies of the thesis are based on simulated AtmosphericInfraRed Sounder (AIRS) radiance by Stand-Alone Radiative Transfer Algorithm(SARTA). An ensemble 10000 profiles selected from the whole globe were used astraining data to generate regression coefficients. An independent samplesconsisting of 2200 profiles was used to verify the accuracy achieved by presentalgorithm. To test the effect of different number of eigenvectors on the ozoneretrieval accuracy, 30, 50 and 80 eigenvectors were used for whole AIRSchannels, respectively. There is a little impact in ozone error using differenteigenvectors. Finally, 80 eigenvectors are used to reconstruct most AIRSchannels. The first step is to perform a cloud detection procedure for each AIRSfootprint. The Moderate Resolution Imaging Spectroradiometer (MODIS) pixelswith 1-km spatial resolution are collocated within an AIRS footprint. Only AIRSclear-sky footprints are availability. Whereafter, The retrieval of ozone profilesand ozone total column was performed using the eigenvector regression and AIRSLIB data. The results were compared with the AIRS L2 ozone production andEuropean Center of Medium-range Weather Forecasts (ECMWF) model. Theozone of retrieval can provide better spatial resolution with respect to the AIRS L2ozone production. Compared with ECMWF, the retrieved profilesroot-mean-square error (RMS) is about 0.3ppmv.Finally, Distribution and characteristics of global total column ozone(60°N-60°S) data retrieved from AIRS, The TIROS Operational VerticalSounder (TOVS) and Total ozone mapping spectrometer (TOMS) during2003～2005 are investigated and compared, showing that the total column ozonein North hemisphere has an evident seasonal variation with a maximum of322.25DU in spring and a minimum of 277.83DU in autumn, decreasing by45DU. Whereas it is unapparent in South Hemisphere. Comparison among threedatasets indicates the global mean total column ozone retrieved from the AIRS islarger by 3-5DU than these from TOVS and TOMS, and significant abnormal inAntarctic continent and deserts. A comparison of satellite ozone data Withground-based data found that TOMS retrieved value is lower than ground-basedmeasurements. Further evidence depicts that AIRS/TOMS retrieved values havethe moderately good agreement and well relationship with the ground-based ozonedata. In section 2, the temporal and spatial distribution of four major aerosol types(dust, carbonaceous, sea salt and sulfate) retrieved by MODIS-OMI (OzoneMonitoring Instrument) Algorithm over East-China area. Compared to SpectralRadiation-Transport Model for Aerosol Species (SPRINTARS) model, the resultsshow reasonable consistency in the distribution by aerosol type.