基于固有光学特性的北部湾透明度遥感反演及其检验  

An IOP-based remote-sensing algorithm for Secchi depth and its validation for the Gulf of Tonkin

作　　者：魏国妹[1] 商少凌[1] 李忠平 兰健[3] 

机构地区：[1]厦门大学近海海洋环境科学国家重点实验室,厦门361005 [2]Northern Gulf Institute,Mississippi State University,ms39529 [3]中国海洋大学海洋系,青岛266100

出　　处：《高技术通讯》2009年第9期977-982,共6页Chinese High Technology Letters

基　　金：国家863计划(2006AA09A302;2008AA09Z108);国家自然科学基金(40521003);国家908专项908-01-ST09项目;厦门大学新世纪人才计划联合资助。

摘　　要：综合了固有光学量的反演和透明度与固有光学量的关系,构成了基于固有光学特性的透明度遥感算法。将此算法应用于北部湾的MODIS遥感数据,得到了该区域的透明度,并利用实测透明度对遥感反演结果进行了验证。通过对时间窗口为±48h,空间窗口为1km的30组实测与遥感匹配数据的统计分析(实测透明度在1.8～26.0m范围),得到平均百分比误差(ε)为22%,对数均方根误差(1og_RMSE)为0.121。而基于叶绿素的经验算法反演的透明度,ε为42%,log_RMSE为0.185。上述结果说明,至少对于北部湾这类近岸水体,海色遥感结合基于固有光学量的算法,可以实现较高准确度的透明度遥感反演。By combining the retrieval of inherent optical properties （IOP） with the relationship between IOP and Secchi depth （Zsd）, an IOP-based algorithm was formulated to estimate Zsd from remote sensing of water color. This algorithm was further applied to the moderate-resolution imaging spectmradiometry （MODIS） data over the Gulf of Tonkin to get Zsd from satellite observations, and the calculated results were evaluated with the data from the in situ measurements. With a time window of ＋ 48h and a spatial range of lkm, 30 match-up points were found （the measured Zsd was in a range of 1.8 -26.0m）. By comparing the Zsd from the IOP-based algorithm with the Zsd from in situ measurements, the average percentage error （ε） was 22%, and the root-mean-square-error in log scale （log RMSE） was 0.121. When the satellite Zsd is calculated with the traditional chlorophyll-based approach, however, e is 42% and the log _ RMSE is 0.185. Clearly, these results demonstrate that, at least for waters in the Gulf of Tonkin, much more accurate Zsd estimation is achieved when it is calculated with the IOP-based algorithm.

关 键 词：北部湾 透明度 半分析算法 MODIS 

分 类 号：TP79[自动化与计算机技术—检测技术与自动化装置]