20年时间序列海表温度数据集反演和评价

Retrieval and Evaluation of 20-Year Sea Surface Temperature Sequence Data Sets

海表温度是研究海洋表面海气相互作用的一个至关重要的物理参数。为建立长时间序列海表温度数据集,在海表温度反演模型建立过程中还需要考虑卫星之间的差异、卫星仪器随时间的衰减、所采用反演算法模型建立过程中人的主观因素的影响等。为此,针对1989—2008年间NOAA/AVHRR数据特点,发展了统一的海表温度反演模型,并生成了20年时间序列海表温度数据集,空间分辨率达到1km。利用船舶浮标资料和OISST数据对该数据集进行了验证,验证结果表明该数据集均方根误差在1℃左右。最后针对1997/1998年厄尔尼诺事件,用反演的海表温度分析了西太平洋暖池区海表温度距平对这次事件的响应。

Sea Surface Temperature（SST） is a vital physical parameter in studying sea-atmosphere exchange on the surface.To build the data sets of long-time sea surface temperature sequences,the differences between satellites,attenuation of the sensors,subjective factors of researchers and so on should be considered in establishing the retrieval model.According to the features of NOAA/AVHRR data from 1989 to 2008,a unified SST retrieval model is devised,and the 20-year sequence data sets are established with the spatial resolution reached 1 km.The ship and buoy data,as well as OISST data,are used to evaluate the data sets,with the precision being around 1 ℃（RMSE）.The response of SST anomalies over the West Pacific warm pool to 1997/1998 El Nino is analyzed by using the retrieval data.