珠江河口混浊高产水域叶绿素a浓度的遥感估算模型

Remote sensing retrieval of chlorophyll-a in turbid,productive estuaries:A case study of Zhujing River estuary

近年来为提高混浊高产水体叶绿素a浓度的估算精度,Gitelson等提出了一种基于红与近红外3个波段遥感反射率的概念模型。文章基于2004年的实测数据,检验该概念模型在珠江河口水域的适用性,估算该概念模型以及它的特殊形式——两波段模型对应的光谱位置,在此基础上构建利用卫星数据提取珠江河口水域叶绿素a浓度的遥感估算模型。结果表明,珠江河口混浊高产水域叶绿素a浓度与三波段模型和两波段模型均有很强的线性相关性,相关系数分别达到0.91与0.88。构建的三波段模型和两波段模型估算的叶绿素a浓度与实测叶绿素a浓度的均方根差(RMSE)分别为5.82mg.m-3和6.53mg.m-3,精度高于其他常用算法。根据MERIS的波段设置构建的三波段模型估算的叶绿素a浓度与实测叶绿素a浓度的RMSE为6.47mg.m-3,显示了良好的应用潜力。

Recently a conceptual model based on the reflectance in three spectral bands of red and near infrared ranges of the speetrum was suggested for retrieving ehlorophyll a （Chl a） concentration in turbid, productive waters. The objeetive of this study is to validate the applieability of the model by selecting spectral positions of this three-band model and its special case, the two-band model, to evaluate the performance of the models based on the spectral and Chl a data measured in the Zhujiang River estuary. Strong linear relationships were found between the Chl a concentration and the three-band model and between the Chl a concentration and the two-band model in the estuary, with correlation coefficients of 0.91 and 0.88, respec- tively. The root mean square errors （RMSE） between measured Chl a concentration and the Chl a concentrations estimated by the three-band model and the two-band model are 5.82mg · m^-3 and 6.53mg · m^-3, respectively. The RMSE between measured Chl a concentration and the Chl a concentration estimated by the three-band model based on MERIS bands is 6.47 mg · m^-3, demonstrating the robustness of this algorithm for MERIS Chl a retrieval.