【作者】 伍玉梅；

【导师】 何宜军；

【作者基本信息】 中国科学院研究生院（海洋研究所） ， 物理海洋学， 2006， 博士

【摘要】 全球气候与人类活动密切相关，气候研究一直是现代科学家关注的重点问题之一。过去海面气象参数的研究主要借助于浮标和站点的观测，但是它们稀少的观测资料极大地限制了海面气象参数的研究。现在借助卫星观测大大地提高我们的认识能力，卫星可以对全球海洋进行连续观测，获取长期大范围的海洋气象资料，为全面深入地了解大洋甚至全球大气活动提供可能。 本文的工作就是利用卫星资料进行月平均和实时的近海面气象参数的反演及应用研究。利用近十八年SSM／I和AVHRR卫星资料与实测资料进行结合，建立神经网络(ANN)模型反演近海面月平均气温和湿度，与实测资料相比气温的均方根差为0.87℃，相关系数为0.99，相对湿度的均方根差为3.73％，相关性为0.65。利用同步物理方法从TOVS资料反演中国海区上空1000mb到10mb之间的温湿廓线，再利用神经网络方法和基于Bowen比的方法从温湿廓线的结果反演出近海面处的实时气温和露温参数，取得了比较合理的结果，气温和露温结果的均方根差分别是1.85K和2.59K(与实测数据相比)。利用2005年1月的AMSR-E亮温资料对实时气象参数反演进行探讨，分析AMSR-E的各个探测通道与海表温度、近海面气温、湿度和风速等参数的相关性，把12个通道分为四种情况并在每种情况下分别进行试验，选择最合适的组合通道并进行气象参数的反演，结果与TAO资料进行比较，海表温度的均方根差是0.55℃，近海面气温的均方根差是0.74℃，海面湿度的均方根差是3.24％，海面风速的均方根差是1.11m／s。目前，与其它结果相比该结果的精度是最好的。 把以上反演得到的近海面气象参数结果应用于海气界面热通量的计算，以更好地研究海气相互作用。分别采用神经网络和Bulk公式两种方法计算月平均潜热和感热通量，结果与GSSTF2资料进行比较，Bulk方法反演的感热和潜热的均方根差分别为9.05±4.6W／m~2和23.7±4.0W／m~2，ANN模型得到的分别是7.54±3.0W／m~2和20.1±3.2W／m~2，结果表明ANN模型得到的结果明显的好于Bulk更多还原

【Abstract】 Global climate is closely linked to human being, which is a hot topic. Traditional research on climate parameters near sea surface was based on buoy and in situ observations, which restrict the development because of small observed region and limited observation data. Now satellite is considered to be a good alternative to obtain knowledge of ocean. Satellites could attain long-time and large-scale oceanic climatic data, which help us understand global ocean and atmosphere activities clearly.The two main tasks of this paper are to retrieve climatic parameters over sea surface from satellites data and to apply the retrieval results into heat flux estimation and climatic study. A new method is proposed to derive long-scale monthly mean air temperature (Ta) and relative humidity (RH) near sea surface from satellite data with artificial neural networks (ANN). Compared with the in situ observations, the root mean square (rms) of Ta and RH are 0.87℃ and 3.73%, respectively. The TOVS data are used to obtain the outline of Ta and dew-point temperature (Td) from the 1000mb to 10mb with the synchronous physical method. The method based on Bowen ratio and ANN model are adopted to get instantaneous Ta and Td parameters near sea surface from the outline. The rms of Ta and Td are 1.85K and 2.59K, respectively. The results are reasonable.One month of AMSR-E data is analyzed used as a case to retrieve real time climate parameters of SST, Ta, RH and wind speed. The twelve channels of AMSR-E are divided into four cases, under which the four parameters are test. The best case is chose to retrieval the four parameters. Compared the results with TAO data, the rms of SST, Ta, RH and wind speed are 0.55℃, 0.74℃, 3.24% and 1.11m/s, respectively. Up to now, the accuracies of them are best.All the above results are applied to study interaction on sea-air surface. Both更多还原