【作者】 陈坤；

【导师】 张前前；

【作者基本信息】 中国海洋大学 ， 海洋化学， 2007， 硕士

【摘要】 利用荧光技术研究浮游植物主要基于色素的荧光光谱,Rayleigh(瑞利)散射、二次散射等光散射一直被做为干扰因素而加以降低、排除或避开。事实上,光散射光谱,尤其是Rayleigh散射光谱中蕴藏着关于藻细胞结构与数量的信息,将其与色素的荧光光谱相结合,获取更为全面的活体浮游植物光谱信息,这对于进行浮游植物识别和数量估计具有重要的意义。本文选取了东海赤潮多发区的11种浮游植物进行研究,共计甲藻3种,硅藻6种,绿藻1种,金藻1种,包括:塔玛亚历山大藻(Alexandrium tamarense)、等鞭金藻(Isochrysis galbana 3012)、岛国大扁藻(Platymonas helgolanidica)、东海原甲藻(Prorocentrum micans)、中肋骨条藻(Skeletonema costatuma)、尖刺拟菱形藻(Pseudo-nitzschia pungens (PS0201-01))、新月菱形藻(Nitzschia closterium)、旋链角毛藻(Chaetoceros Curvisetus)、纤细角毛藻(Chaetoceros Debilis)、双突角毛藻(Chaetoceros Didymus)和裸甲藻(Gymnodinium sp)。通过对3个温度水平(15℃、20℃、25℃)、3个光照水平(7000 lux、4100 lux、1180 lux)、一个生长周期(1-8天)的活体荧光光谱和Rayleigh散射光谱进行了研究。本论文的主要研究成果如下:1.光谱特征的研究.对11种浮游植物的全波段三维荧光光谱进行了研究,共得到了3类谱:原始Rayleigh散射谱、组合谱和重构Rayleigh散射谱。其中,组合谱是色素荧光特征谱char_Dv1与原始Rayleigh散射谱首尾相连组成( char_Dv1是去除Rayleigh散射后,取λem在670-690nm之间的三维光谱进行奇异值分解,取出的相应于激发光谱的第一主成分);重构Rayleigh散射谱是指取主成分分析中,取70%贡献率的前n个主成分重构所得的光谱。分别对三者进行了Bayesian判别分析,结果表明重构Rayleigh散射谱的判别分析效果最佳,组合谱次之,原始Rayleigh散射谱最差。2.浮游植物的光谱相似性的比较.分别用平均相对误差和平均相关系数两种指标考察了上述3类谱的相似性。结果表明:两种指标衡量三类谱的相似性基本一致。Rayleigh散射光谱的相似性Pl、Sk和Cu最好,Ps、Cl、Di的最差;重构Rayleigh散射光谱的相似性Is、Pl、Sk和Cu的最好,Ps的最差;组合谱Is、Pl、Sk和Cu相似性依然是最好,Ps相似性最差;本文选取组合谱进行浮游植物种类识别和数量估计。3.混合光谱的定性定量分析.首先,寻找浮游植物数量与Rayleigh散射谱的最高峰的定量关系。对原始Rayleigh散射谱进行主成分分析,得到11种浮游植物的标准谱,找出各自的最高散射峰λmax。取20℃,7000 lux培养条件下、生长状态良好、处于指数生长期的11种浮游植物,进行不同浓度层次的梯度实验,发现对实验选用的11种浮游植物,细胞密度与其各自的最高散射峰λmax强度和活体叶绿素荧光强度均呈较好的线性关系。其次,依据组合谱,采用遍历非负最小二乘法,对实验设计的59个浮游植物混合样品进行了定性定量分析。进而讨论了物种数的确定与定性正确率的规定,选定拆分系数阈值为0.25。在属的层次上拆分,拆分正确率为76%,在纲的层次上拆分,正确率为90%;2物种混合样品有46个,在属的层次上拆分,拆分正确率为76%,在纲的层次上拆分,正确率为91%;甲藻硅藻混合样品有28个,在属的层次上拆分正确率为78%,在纲的层次上拆分,正确率为93%。最后,根据拆分系数、混合样品Rayleigh散射峰λmax的强度,以及不同培养条件下细胞密度与散射峰λmax强度的线性方程,估计混合样品中不同浮游植物的细胞密度。4.对经lugol染色固定浮游植物的Rayleigh散射光谱进行初步研究,通过主成分分析得到各浮游植物的标准谱各一条,发现其强度与活体Rayleigh散射光谱相比,强度增大很多、且谱型差异很大,但是随存放时间的增长强度迅速下降。考察染色后Rayleigh散射光谱的相似性,发现与活体光谱的相似性相当。总之,本文对活体浮游植物的Rayleigh散射光谱进行了研究,将其与活体色素荧光特征相结合,得到了具有定性定量区分能力的组合谱。在本实验条件下,浮游植物数量与其荧光强度及Rayleigh散射强度三者具有良好的相关性,运用遍历非负最小二乘法定性定量识别实验设计的混合样品正确率较单独使用浮游植物的色素荧光光谱有所提高。更多还原

【Abstract】 Three dimensional fluorescence spectrum is used for monitoring phytoplankton classes and their abundance in previous study. Avoiding or eliminating Rayleigh scatter signals is often a key procedure for improving detection sensitivity in phytoplankton fluorescence analysis in the previous study. However, it is found that there are much useful information of cell contents and abundance in the scattering signals. As a result, Rayleigh scatter spectra of phytoplankton were studied in this paper. It was predicted that by combining both Rayleigh scatter spectra and three dimensional fluorescence spectra, more information could be provided for better identification of phytoplankton classes and assessment of their abundance. This paper aims to study the characteristics of both Rayleigh scatter spectra and three dimensional fluorescence spectra of phytoplankton in order to obtain better combined spectra for identification of phytoplankton. Eleven red-tide blooming algae and dominant phytoplankton species in the East China Sea, including Alexandrium tamarense, Isochrysis galbana 3012,Platymonas helgolanidica, Prorocentrum micans, Pseudo-nitzschia pungen (PS0201-01), Skeletonema costatuma, Nitzschia closterium, Chaetoceros curvisetus, Chaetoceros gracilis, Chaetoceros didymus and Gymnodinium sp, which belong to five divisions of phytoplankton, were chosen and cultivated in lab cultures under different temperatures and illumination intensities.The main research work is as follow:1 Three kinds of spectra were studied: the Rayleigh scatter spectrum, the combined spectrum and the restructured Rayleigh scatter spectrum. The combined spectrum refers to the Rayleigh scatter spectrum combined with the pigment characteristic spectrum char_Dv1, while the restructured Rayleigh scatter spectrum refers to the spectrum of 70% of principle components. The discriminant analysis results show that the restructured Rayleigh scatter spectrum has best discriminant ability, then it is the combined spectrum.2 The similarity of the above three kinds of spectra of living phytoplankton was studied. Two indicators including the average relative error and the average correlation coefficient were used to evaluate the similarity of the above three kinds of spectra. For the combined spectrum, Is, Pl, Sk and Cu have better similarity than the others, Ps has the worst. Similar results were obtained by the other two kinds of spectra. The results indicate that the similarity of the three kinds of spectra of the eleven phytoplankton species is good enough for analysis. In this paper, the combined spectra were chosen for qualitative and quantitative analysis.3 Qualitative and quantitative analysis. Firstly, for each phytoplankton, highest peak of Rayleigh spectra were found and the linearity between cell density and peak value was studied in the cultivating condition. Then, qualitative and quantitative analysis of 59 mixtures of phytoplankton was carried out using the non-negative least square method (NNLS). The recognition rate for all the sample was 90% in class, 76% in genus. At last, cell density for each phytoplankton was estimated by the linearity equation and the result of NNLS.4 Rayleigh spectra of lugol staining phytoplankton were primarily studied. Hierarchical clustering was processed to classify the Rayleigh scatter spectrum of each phytoplankton. Each species have one standard fluorescence spectrum. The measuring precision of the spectra was studied.In summary, according to our knowledge this paper takes the lead in studying the Rayleigh spectrum of in vivo phytoplankton combined with the pigment characteristic spectrum char_Dv1. Using the combined spectra and NNLS method, dinoflagellates and diatoms can be distinguished from each other in laboratory cultures, and their abundance was estimated.更多还原