海洋遥感数据缺值对EOF和REOF时空分布分析的影响

Analysis of effects of invalid value on the temporal-spatial distribution of ocean remote sensing data with EOF and REOF methods

本文分别采用经验正交函数EOF和旋转经验正交函数REOF两种方法，对中国海（14°-40°N、105°～132°E）1998年1月～1999年8月之间SeaWiFS归一化离水辐射率Lwn（490nm）数据，进行时空分布分析研究．结果表明：在不同时间区段、不同空间范围，利用EOF和REOF分解得到的典型要素场时空分布存在较小差异，在时间、空间分布上较稳定．同时也说明，被云遮掩的部分遥感数据对EOF和REOF分析结果影响是很小的．

Used two methods of Empirical Orthogonal Function （EOF） and Rotated Empirical Orthogonal Function （REOF） the temporal-spatial distribution of normalized leaving water radiance was studied at wavelength 490nm during Jan, 1998 to Aug, 1999 in the region of 14°-40°N and 105° - 132°E. The results show that the difference of temporal-spatial distribution of factor fields decomposed by the EOF and REOF is very small and the distribution of factor fields is relatively stable, even in the different time series and regions. In the same way, it indicates that the effect of remote sensing data covering clouds in the EOF and REOF methods is negleetable.