PROSPECT模型的特定吸收系数的测定方法和叶绿素a和b的反演研究

由于缺乏在活体条件下测定叶片的特定吸收系数的实验方法，长期以来PROSPECT模型不能分别反演叶绿素a和b。本研究通过选用低叶绿素b突变体和叶绿素b缺失突变体水稻以及常规水稻品种为材料，在温室条件下采用远红光诱导和低氮处理的方法培养出大量的在叶绿素a和b方面组成特殊的叶片样品，从中筛选出其他生物物理化学参数（干物重、叶面积、木质素、纤维素、类胡萝卜素、水分）相近而只在叶绿素含量有差异的叶片样本，进而找出只有叶绿素a或b含量差异的特殊样本，测定叶片的吸收光谱，通过分析计算这些样本间在叶绿素吸收区域的吸收率差异，计算叶绿素a和b的特定吸收系数。将测得的特定吸收系数输入PROSPECT模型，用最小化代价函数方法获得其他特定吸收系数，建立改进的模型，以能分别反演叶绿素a和b。同时，在提取叶绿素b缺失突变体叶片光谱特性基础上，研究用新构建的高光谱植被指数分别反演叶绿素a和b的方法。

In vivo determination of specific coefficients in PROSPECT model is critical since they link the optical and biochemical properies together. PROSPECT model has long been not successful for retrieval of chlorophyll a and b due to lack of methodology for measurement of the specific coefficients. The main objectives of this study are to establish an experimental method for in vivo measurement of the specific absorption coefficients of chlorophyll a and b in intact leaves and improve the PROSPECT model for retrieval of chlorophyll a and b by input of the determined specific coefficients and those derived by minimizing the merit function.Direct comparison between simulated and measured reflectance (and transmittance) will be made to assess the performance of the improved model in terms of accuracy of fit. The experimental method is based on derivation of a large size of leaf samples with special compositions of chlorophyll a and b by use of chlorophyll b deficient rice mutant and low chlorophyll b rice mutant，induction of far red light and low nitrogen treatment under greenhouse conditions.The in vivo specific absorption coefficients of chlorophyll a and b will be determined by calculating the difference in absorptance of the leaves very close in the biophysical and biochemical parameters (e.g. leaf area, water, dry weight, cellulose, lignin, carotenoids) but only different in chlorophyll a or b.The leaf spectral features of the chlorophyll b deficient rice mutant and low chlorophyll b rice mutant will be extracted for construction of new spectral vegetation indices for retrieval of chlorophyll a and b by using new constructed hyperspectral vegetation indices.