应用水表面下辐照度比估测太湖夏季水体叶绿素a浓度

Estimating chlorophyll-a concentration in Lake Taihu in summer by irradiance ratio just beneath water surface

利用便携式光谱辐射计,采用一定的观测角度获取水体表面的光谱,进而提取水表面下辐照度比R(O^-)信息,分析R(O^-)光谱特征与叶绿素a浓度之间的相互关系,结果表明太湖夏季水体叶绿素a浓度与R(O^-)光谱曲线762 nm、727 nm和496 nm处的相关系数较大,分别达到了0.85、0.84、-0.80.通过单波段、波段比值模型分析,认为以R(O^-)_(762)、R(O^-)_(762)/R(O^-)_(496)、R(O^-)_(727)/R(O^-)_(496)为自变量的二次函数模型是利用水表面下辐照度比R(O^-)估算太湖夏季水体中叶绿素a浓度的最佳模型,模型的决定系数R^2分别达到了0.923、0.919、0.916,回归估计的标准误差S分别为0.012、0.013、0.013,F检验值分别为101.241、96.576、92.925.利用剩余10个样本对估算模型进行精度和误差检验,结果表明以R(O^-)_(762)/R(P^-)_(496)为自变量的二次函数模型好于另外两个,对太湖夏季水体叶绿素a浓度估算具有一定的实用性.此外,将光谱微分技术应用到R(O^-)信息分析太湖夏季水体叶绿素a浓度,结果不能获得较高的预测精度.

Water surface spectra, which can work irradiance ratio just beneath water surface R（0ˉ ） out, is obtained by ASD FieldSpec Hand-Held Spectroradiometer. When the in-situ data is collected, a certain special angle has been adopted. After analyzing the relationship between spectral characteristics of R （0ˉ） and chlorophyll-a concentration, the results showed that the strongest correlations appeared at 762 nm,727 nm and 496 nm, the values separately were 0. 85,0. 84 and -0. 80. The model analysis with single band and band ratio revealed that the quadratic models, of which independent variables were R（0ˉ ）762 ,R（0ˉ ）762/ R（0ˉ ）496 ,R（0ˉ ）727/ R（0ˉ ）496, were better than others. R^2 separately reach 0. 923,0. 919,0. 916; std. error of the estimate S separately were 0. 012,0. 013,0. 013 ; F separately were 101. 241,96. 576,92. 925. Make use of the 10 surplus samples to estimate three quadratic models and carry on the accuracy and error margin examinations. Then find that the quadratic model with the independent variable R（0ˉ ）762/ R（0ˉ ）496 was the best, and to a certain extent had some functions on chlorophyll-a concentration estimation in Lake Taihu in summer. In addition, try to apply differential calculus to R （0ˉ） to estimate chlorophyll-a concentration, and can not acquire the better result.