A Study On Sea Surface Meteorological Parameters Retrieval With AMSR-E Satellite Observations

The global climate change, one of the environmental issues nowadays which has attracted universal attention from the scientific communities, governments and the public, has a great impact on human society and will probably be a disaster to the ecosystem and the social economy. The change of the ocean, which takes up 70% of the earth's surface, should take great responsibility for that of the climate, thus, more emphasis on researches into the meteorological parameters are laid. However, the marine science has been largely promoted with the development of the satellite remote sensing technology, by means of which the global ocean can be observed continuously so as to obtain marine meteorological data of long term and large scale for a comprehensive and in-depth research into the global ocean and atmosphere.This dissertation mainly focuses on retrieving real-time and monthly average sea surface meteorological parameters from the EOS satellite AMSR-E data. A new method, genetic algorithm (GA),is proposed to derive instantaneous and monthly mean air temperature (Ta) and specific humidity (Qa) near sea surface from AMSR-E observations between June 2002 and December 2006 .Compared with the in situ observations, the root mean square (rms) of instantaneous Ta and Qa are 1.13℃and 1.28 g/kg, respectively. In the same way, the root mean square of the monthly Ta and Qa are 0.80℃and 1.03k/kg respectively according to the AMSR-E monthly mean data and the in situ observations between July 2002 and October 2007. It is shown that GA is a better choice after comparing the results of GA with those of ANN and multivariate regression technique. Furthermore, through a comparison of the retrieved results from the 2005 annual AMSR-E brightness temperature data and in situ observations from TAO, the rms of Ta and Qa are 0.71℃and 0.91g/kg, respectively.