富营养化的太湖水体叶绿素a浓度模型反演  

Retrieval of Chlorophyll-a Concentration in the Turbid and Eutrophic Taihu Lake

作　　者：杨煜[1] 李云梅[1] 王桥[1] 乐成峰[1] 孙德勇[1] 黄昌春[1] 

机构地区：[1]南京师范大学虚拟地理环境教育部重点实验室,南京210046

出　　处：《地球信息科学》2009年第5期597-603,共7页Geo-information Science

基　　金："十一五"国家科技支撑计划项目(2008BAC34B05);江苏省2008年度普通高校研究生科研创新计划项目(CX08B_014Z);江苏省2008年度普通高校研究生科研创新计划项目(CX08B_015Z)共同资助

摘　　要：半经验模型反演叶绿素a浓度是目前遥感反演水体叶绿素a浓度的主要方法。但是,大量研究结果表明,太湖水体浑浊,富营养化严重,各种半经验模型的反演精度和模型适用性有较大差异。因此,研究一种既满足一定精度要求,又具有时间普适性的叶绿素a浓度反演算法,对提高模型适用性,促进遥感的反演应用具有重要意义。本研究通过2005年6-10月地面实测数据,建立太湖叶绿素a浓度一阶微分反演模型、波段比值反演模型和三波段反演模型,对比各模型反演效果,认为波段比值模型与三波段模型具有较好的反演效果。运用2006年11月和2007年11月实测数据对这三种模型加以检验,结果表明,三波段模型反演高富营养化的太湖水体,不仅精度高,平均误差仅为实测浓度差的8.3%,而且适用性较强,不同年份数据的检验结果证明平均误差均低于实测浓度差的20%。因此,三波段模型是这三种反演模型中效果最好的一类模型。Semi-empirical models are the main models to retrieve chlorophyll-a concentration in remote sensing.The water is so turbid and eutrophic in Taihu Lake that the accuracy of some semi-empirical models varies highly.The datasets were collected over a considerable range of optical properties,trophic status,and geographical locations in Taihu Lake.So,it is very useful to find a model which can retrieve chlorophyll-a concentration in a good accuracy and has the best ability to adapt time changing.This study applied a recently developed model of the form 1/Rrs（λ1）-1/Rrs（λ2）×Rrs（λ3） where Rrs（λi） is the remote-sensing reflectance at the wavelength λi.The objectives of this paper are（a） to build the reflectance first-derivative model、reflectance ratio model and three-band model using the data measured from June to October in 2005 to find which model is the best,（b） to find out the special bands which are suitable in Taihu Lake for the three-band model,and（c） to examine these three models by the dataset measured in November in 2006 and 2007.Strong linear relationships were established between analytically measured chlorophyll-a concentration and both the three-band model 1/Rrs（λ666）-1/Rrs（λ693）×Rrs（λ734） and the two-band model R（λ705）/R（λ680）.Accuracy for all these models were measured by R2 and RMSE.The result is that the three-band model is the best one of them.Its normal RMSE is less than 20% of the change which was shown in the dataset.And it has the best ability to adapt time changing that the normal RMSE of 2006 and 2007 is 11.17 and 9.74 respectively.We can also find that the bands used in the three-band model are available from some satellites.So,the three-band model has a good future to retrieve chlorophyll-a concentrations.

关 键 词：叶绿素浓度 反演模型 太湖 富营养化 

分 类 号：X52[环境科学与工程—环境工程] X87