The Analytical Technique For Assessment Of Phytoplankton Class Abundance Based On Fluorescence Excitation Emission Matrix

Phytoplankton are the basis of food webs in marine ecosystems, affecting globalcarbon cycles and marine ecological process. The excessive growth of phytoplanktoncan pose significant threats to local biodiversity and ecosystem functioning, especiallytoxic algal blooms can cause significant threats to the economic and human health.Abundance and community composition studies of phytoplankton are important forevaluating the ecological status of coastal seawater regions. Consequently, It becomesan urgent need for developing a rapid, low-cost taxonomic technique fordifferentiating phytoplankton populations in marine environments.Excitation-emission matrix (EEM) fluorescence spectroscopy, well known as a rapid,simple and sensitive analytical technique for the discrimination of phytoplanktontaxonomic groups, has drawn increased attentions in recent years.Forty-one algae species belonging to28genera of five divisions were studiedbased on the analysis of the phytoplankton community structure in coastal area ofChina sea. The algal pigment extracts were studied using fluorescenceexcitation-emission matrices (EEMs) and EEMs were collected from different culturetemperature, different culture lights and different growth stages. First, theDaubechies7(db7) wavelet analysis was applied to EEM, the scale discriminatingcharacteristic spectra attempts to replace high-performance liquid chromatography(HPLC) pigment data, along with CHEMTAX software to develop the fluorometrictechnique for the differentiation of phytoplankton taxonomic groups, Second, theparallel factor (PARAFAC) model was applied to a fluorescence EEM and14fluorescent components were generated, This paper attempts to replace HPLCpigment data with the14fluorescent components, in combination with CHEMTAXsoftware and nonnegative least squares (NNLS) method to develop the fluorescence differentiation technique (EEM-PARAFAC-CHEMTAX and EEM-PARAFAC-NNLS)for the determination of algae community structure. The fluorometric techniquedeveloped succeeded in the qualitative and quantitative discrimination of five algaltaxonomic groups: Bacillariophyta, Dinophyta, Chlorophyta, Cyanophyta, andCryptophyta at the division level. The main research results were as follows:1. The db7wavelet analysis was used to decompose the EEM fluorescencespectra of algal pigment extracts and different scale vectors were obtained, Bayesiandiscriminant analysis (BDA) was used to select the discriminating characteristicspectra of scale vectors. Then the reference scale discriminating characteristic spectrawas constructed for CHEMTAX, Finally, CHEMTAX method was used to developthe fluorometric method to differentiate algae taxonomic groups. When thefluorometric method was utilized to discriminate307samples of single algae species，the average correct discrimination ratio (CDR) of the five phytoplankton taxonomicgroups was96.3%, with the average relative content of81.0%at the division level.The average correct discrimination ratios (CDRs) for195mixture samples were89.9%for the dominant algae species and77.3%for the subdominant algae species.The results also indicated that approximately half of the samples (data not given)could not be correctly identified at the division level when the relative abundance ofthe subdominant algae species estimated was below20.0%. Therefore, Thediscrimination results of the subdominant algae class were not reliable when therelative abundance estimated was below20.0%. Sixteen of the85field samplescollected from the Yangtze River estuary were analyzed by both HPLC-CHEMTAXand the fluorometric techniques developed. The results of both methods revealed thatBacillariophyta was the dominant algal group in these16samples and that thesubdominant algal groups comprised Dinophyta, Chlorophyta and Cyanophyta.2. The PARAFAC model was applied to fluorescence EEM of41algae speciesand14fluorescent components were identified according to the residual sum ofsquares and specificity of the composition profiles of fluorescent. By the14fluorescent components, the algae species at different growth stages were classifiedcorrectly at the division level using BDA. Then the reference fluorescent component ratio matrix was constructed for CHEMTAX, and the EEM–PARAFAC–CHEMTAXmethod was developed to differentiate algae taxonomic groups. The fivephytoplankton groups with the average CDR when the fluorometric method was usedfor single-species samples were96.3%, with the average relative content was79.2%at the division level. The CDRs for mixtures were above88.6%for the dominantalgae species and above74.9%for the subdominant algae species. However, theCDRs of the subdominant algae species were too low to be unreliable when therelative abundance estimated was <15.0%. The fluorometric method was tested usingthe16samples from the Yangtze River estuary in March2013. The discriminationresults of the dominant algae groups were in good agreement with HPLC-CHEMTAX,as well as the subdominant algae groups which relative abundance estimated wasabove15.0%.3. Forty-one phytoplankton species belonging to28genera of five divisions werestudied. First, the PARAFAC model was applied to a fluorescence EEM, and15fluorescent components were generated. Second,15fluorescent components werefound to have a better discriminatory capability based on Bayesian discriminantanalysis (BDA). Third, all spectra of the fluorescent component compositions for the41phytoplankton species were spectrographically sorted into61reference spectrausing hierarchical cluster analysis (HCA), and then, the reference spectra were used toestablish a database. Finally, phytoplankton biomass in the taxonomic groups waspredicted from the database using multivariable linear regression model resolved bythe NNLS method. The five phytoplankton groups with CDRs when the fluorometricmethod was used for single-species samples were100.0%at the division level and theaverage relative contents estimated were81.1%to94.0%. The CDRs for the mixtureswere91.7%for the dominant phytoplankton species and74.3%for the subdominantphytoplankton species. When the dominant algal taxonomic groups accounted for20.0%of the mixtures, their CDRs ranged from50.0%to84.2%, with the averagerelative contents estimated from18.9%to44.3%. For Bacillariophyta, Chlorophytaand Cryptophyta, the CDRs reached80.0%even when their relative contents were aslow as20.0%. However, the CDRs of the subdominant algae species were too low to be unreliable when the relative abundance estimated was <15.0%. Sixteen of the85field samples collected from the Yangtze River estuary were analyzed by bothHPLC-CHEMTAX and the fluorometric technique developed. The discriminationresults of the dominant algae groups agreed with HPLC-CHEMTAX, as well as thesubdominant algae groups which relative abundance estimated was above15.0%.4. The discrimination results of three fluorometric methods for both thedominant algae groups and the subdominant algae groups revealed thatidentification results of EEM-PARAFAC-NNLS method was better. The fluorometricmethod was tested using the field samples from Yangtze River estuary, Yellow Seaand East China Sea. These results for the dominant algae groups and the subdominantalgae groups which relative abundance estimated was above15.0%were inaccordance with those historical research results in the same sea region.In conclusion, The fluorometric technique could differentiate algal taxonomicgroups accurately at the division level. This work demonstrated that the developedfluorometric technique is a powerful tool for quickly analyzing a large number ofalgal samples in situ.