Preliminary Research On The Algae Fluorescence Auto-analyzer For Phytoplankton Population

The frequency and intensity of harmful algal blooms (HABs) have increased in coastal areas. It is imperative to develop a rapid, and reliable observations technique to determine the composition of phytoplankton population and to discriminate algal species when they cause HABs. Among the existing methods that used to delineate phytoplankton community composition, fluorescent methods, which developed in recent years, acquire great popularity because they are sensitive, rapid to perform, and in situ.Based on the research of the composition of phytoplankton population in coastal area of China, 53 phytoplankton species which belonged to seven spectral algal groups Bacillariophyta, Chrysophyta, Chlorophyta, Dinophyta, Xanthophyta, Cryptophyta, and Cyanophyta, were cultured at different illumination, temperature, and period of growth for the fluorescence measurements. In order to develop an algae fluorescence auto-analyzer for phytoplankton population which used a series of LEDs as the light source, 13 excitation wavelengths were chosen to form the new 3D discrete fluorescence spectra. Then, the wavelet (db7, coif2, and sym7), Bayes Classifier and non-negative least squares were applied to establish the fluorescence differentiation method for phytoplankton populations. The results are as follows:1. According to the stepwise discriminant analysis and considering the availability of monochrome LEDs, 13 excitation wavelengths (400, 430, 450, 460, 470, 490, 500, 510, 525, 550, 570, 590, and 610nm) were chosen to form the new 3D discrete fluorescence spectra.2. Three wavelets (db7, coif2, and sym7) were selected to decompose the 3D discrete fluorescence spectra respectively, and some fluorescence feature spectra specific to the algae classes were obtained respectively. Simulatively mixed algae samples were used to test the fluorescence feature spectra and the complementarities among the feature spectra were analyzed. Then the reference spectra database were constructed for the division level and the genus level.3. For laboratory simulatively mixed samples, samples mixed from 53 algal species in which the algae of one divisions accounted for 25%, 50%, 75%, 85%, and 100% of the gross biomass, yielded average discrimination rates at the level of division of :62.7%, 84.4%, 95.9%, 97.1% and 99.9%, respectively, with average relative contents of 17.5%, 41.3%, 68.9%, 78.4% and 87.0%, respectively. When the algae of one division accounted for 50% of the gross biomass, the discrimination rate at the level of division can be greater than 85%, except for Xanthophyta and Chlorophyceae. At the genus level, when the proportion of the dominant species reaches 80%, the dominant algae species can almost be recognized and the discrimination rate can reach 85% and above for 28 algae species. Especially, the two-rank database can significantly improve the differentiation of Thalassiosira rotula, Rhizosolenia setigera Brightwell, Amphidinium carterae, Gymnodinium simplex and Gymnodinium sp.4. For the laboratory mixed samples, when the algae of one division accounted for 75% of the gross biomass, the discrimination rates at the level of division were over 80% for six division except for Chrysophyta; when the algae of one division accounted for 100% of the gross biomass, the discrimination rates at the level of division of others were 100% except for Chrysophyta. when the proportion of the dominant species reaches 75%, the average discrimination rate at genus level was 80.6%, and most of the dominant species can be discriminated by over 80% and even up to 100%, except for Odontella cf_sinensis, Thalassiosira rotula, Phaeocystis globosa Gymnodinium sp. and Prococentrum marinum.5. For the 12 samples obtained from Jiaozhou Bay, the dominant algae groups were recognized for all samples at the division level and for 2 of the three samples whose dominant algae species accounted for more than 80% of the gross biomass at the genus level; for the 12 samples collected from the mesocosm experiment in Maidao Bay of Qingdao, the dominant algae groups were recognized for the 11 samples at the division level and for 4 of 5 samples whose dominant species accounted for more than 80% of the gross biomass at the genus level. 6. A fluorescence auto-analyzer for algae population was devised using monochrome LEDs as light source, which would be a promising tool for in situ differentiation of the algae population.The innovation of this paper is: to choose 13 excitation wavelengths to form the new 3D discrete fluorescence spectra for developing a in situ algae population auto-analyzer using monochrome LEDs as light sources; the fluorescence feature spectra specific to the algae classes were extracted by three wavelets, and the complementary relationship among them were analyzed, then a two-rank fluorescence reference spectra database was constructed and a fluorometric discrimination technique was developed; at the division level, the fluorometric technique could not only identify the dominant phytoplankton group, but also identify the subdominant phytoplankton group and give the relative contents of the dominant and subdominant phytoplankton divisions, respectively; at the genus level, it can recognize the dominant algae species when its dominance reaches a certain extent (causing HAB).