Studies On Absorption Spectra For Identifying Phytoplankton

Monitoring phytoplankton classes and their abundance is a rountine task in marine scientific. With the frequent occurances of red-tide in coastal water of our country at present, more and more scientists have began to focus on the study of monitoring phytoplankton classes and their abundance. Because UV-Spectrophotometer is easy, steady, and have good repetition, can be used to analyse phytoplankton. If the classes of phytoplankton can be identified fast and their abundance can be confirmed exactly, that will make great significance for red-tide. This paper aims to study the characteristics of the absorption spectra of phytoplankton and make sure which species is in the mixed phytoplankton and its concentration.Single species cultures of fifteen common phytoplankton, representing six divisions, were used. Mixtures of phytoplankton were also prepared for analysis. They were cultivated under three different temperatures (25℃,20℃,15℃) and four illuminations(18000lux,11000lux,7000lux,1100lux). Then their absorption spectra at different growth period were measured. The fingerprint features which are extracted by some chemometrics methods, was used to identify these different phytoplankton species. The main research work is as follows:1.Seeing about the similarity of different phytoplankton. All of the absorption spectra of phytoplankton were put together, the similaritis were obtained with eyes. We found that some phytoplankton belonging to the same division are very similar, and between different division they have respective characteristic. For example, though Isochrysis and galbana Rhodomonas sp. have the same absorption at the same place, but the absorption shape of galbana Rhodomonas sp. is wider than Isochrysis between 340nm-500nm; The absorption spectrum of Synechocoocus sp. and Platymonas helgolanidica respectively have three deep absorption and two deep absorption between 400nm-500nm; Phytoplankton between Bacillariophyta and Dinophyta are more sililar, but under different situation the absorption of Dinophyta will have deep absorption at the wavelength less than 370nm. This is a clear difference between Bacillariophyta and Dinophyta. This availed for their identification.2. Feature extraction methods were studied. Firstly the matrix made up by primitive spectrum data was processed by Cluster analysis, the ratio value of S1 and∑S was used to compare the similarities of spectra. According to the cluster result and the similarity, we obtained representative absorption spectra, which have the most information of the primitive spectrum data. The wavelengths that have great ability to differentiate species were abstracted by mathematic methods and 2 incontinuous wavelengths were received. They are 443.5-537.5nm;654-678nm. Finally, the 2 incontinuous wavelengths of all the representative spectra make up the standard spectra.3. Using different method to identify phytoplankton. Firstly, distance discriminant analysis was processed on the fifteen phytoplankton. Both the representative spectra and the standard spectra, the recognition correct rate was more than 80% on the species. If on the division level, the recognition correct rate was more than 95%. Then we also used the ratio method to identify diatoms and dinoflagellate, The ratio analysis indicated that A440/A463 can distinguish diatoms from dinoflagellates effectively.4. Assessment of phytoplankton class abundance using the Non-negative Least Square (NNLS) method. In total, 137 mixtures of phytoplankton were examined on the division level. The recognition rate was 86.1%. We still deal with the mixtures which can be recognized correctly, and assess the abundance of each pure species in the mixture. The result show that: the minimum of the determining range was 0.05mg/L chlorophyll-a/L. More than 0.05mg/L Chl-a can be obtained the assessed abundance.In a word, this paper take the preliminary research to the absorption spectrum of the phytoplankton. the standard spectra were obtained. The ratio method was used to identify diatoms and dinoflagellate, and the ratio analysis indicated that A440/A463 can distinguish them effectively. At last, the class and abundance of phytoplankton was assessed, and we have got some good result.