| 摘要: |
| 利用现场实测的表观光学量和固有光学量数据,得到了我国黄海、东海近岸二类水体多个波段的总吸收系数的统计反演模式。此反演模式采用412/555、490/555两个波段遥感反射比比值的二项式,得到波长412、440、488、510、532、555nm处的总吸收系数,其反演值和实测值的平均相对误差不大于25.8%,相关系数R2达到了0.75到0.85。水体总吸收系数几种统计模型的误差敏感性分析表明,反演模式对±5%的遥感反射比输入误差导致结果增加误差最大为24.0%,因此反演模式是可用的。
同时给出了412、488、510、532、555nm各波段的总吸收系数同波段440nm的总吸收系数之间的关系。结果表明,在400—600nm波段范围内,每一个波段的总吸收系数与440nm波段的总吸收系数的相关性均较高,相关系数R2都超过了0.99。通过对拟合直线的斜率与波长进行回归,得到斜率和波长的关系,其相关系数R2为0.99,这样利用本文中建立的各波段总吸收系数关系模型,可以从一个已知波段的总吸收系数反演出任何另外一个波段的总吸收系数,这就在水色反演与应用中大大减少了未知因子的个数。 |
| 关键词: 二类水体 总吸收系数 统计反演模型 |
| DOI： |
| 分类号: |
| 基金项目:HY-1卫星地面应用系统资助项目:“863模块化二类水体相关遥感信息提取技术”,2001AA636010号;国家海洋局青年海洋科学基金资助项目,2005415号;国家海洋局青年海洋科学基金资助项目,课题编号:2003422;国家卫星海洋应用中心青年科技基金资助项目:“863HY-1/CCD海洋信息提取及应用技术研究”,2002AA639160号和“863带偏振响应的精确Rayleigh散射算法研究”,2002AA639200号;国防科工委卫星重点应用项目,“HY-1卫星水色因子信息提取算法研究”,2001—2005年。 |
|
| THE STATISTIC INVERSION ALGORITHMS AND SPECTRAL RELATIONS OF TOTAL ABSORPTION COEFFICIENTS FOR THE HUANGHAI SEA AND THE EAST CHINA SEA |
|
WANG Xiao-Mei, TANG Jun-Wu, SONG Qing-Jun, DING Jing, MA Chao-Fei
|
|
National Satellite Ocean Application Service, Beijing, 100081
|
| Abstract: |
| The Yellow Sea and East China Sea are typical coastal Case-II waters with high sediment load and complicate optical properties. A total absorption coefficient of sea water is one of the main parameters influencing the calculation of optical distribution. As an inherent optical property (IOP), it has been widely applied in marine researches, such as heat flux, primary production, subsurface light intensity, water type classification, and so on. The authors proposed a group of statistical algorithm for calculating the total absorption coefficient based on in situ apparent optical property and inherent optical property data of waters collected in spring 2003 with strict quality assurance according to NASA ocean bio-optic protocols. The band-ratios of Rrs412 /Rrs555, Rrs490/Rrs555 are used in the algorithm to derive the total absorption coefficients (at) at 412, 440, 488, 510, 532 and 555nm bands, respectively. The average relative errors between inversed and measured values are less than 25.8%, the correlative coefficients (R2) are 0.75–0.85. Error sensitivity analysis shows that the maximum retrieve error increased by 24.0% at ±5% error in Rrs’s. So the statistical algorithms of this paper are usable. At the same time, the relations between total absorption coefficients at 412, 488, 510, 532, 555nm and that of 440nm are also given in this paper. The results show that the relations between total absorption coefficients of 400–600nm and that of 440nm are most correlated and their correlation coefficientsR2 are all greater than 0.99. Hence a regression is also done for the slope of the linear relations and wavelengths, and theR2 is also 0.99. Thus it is able to retrieve other bands' total absorption coefficients with only one band absorption value, which mostly reduce the number of unknown parameters for other ocean color related problems. The results in this paper need to be validated. Although a group of statistical algorithms of total absorption coefficients are quite satisfactory, more and deep-going researches are needed to develop in the future, such as the specific absorption coefficient of each component of waters. |
| Key words: Case-II water algorithms, Total absorption coefficients, Statistic mode |
