In order to obtain the three dimensional (3D) fluorescence spectra directly and rapidly by in situ and in vivo phytoplankton monitoring, a prototype ocean lidar system with the ability of measuring 3D fluorescence spectra is developed, in which an optical parametric oscillator (OPO) and a grating spectrometer are employed. Using the system, the 3D fluorescence spectra of five typical red-tide species in the East China Sea are measured in lab. The preliminary experimental results show the feasibility of monitoring and classifying phytoplankton by this ocean lidar.
The 3D fluorescence discrimination of phytoplankton classes was investigated by SA4 multiwavelet,GHM multiwavelet,and coifman-2(coif2) wavelet analysis.Belonging to 35 genera of 7 major phytoplankton divisions in the inshore area of China Sea,Single species cultures of 51 phytoplankton species were employed.The second scale vector (Ca2) of SA4,Ca2 of GHM and the third scale vector (Ca3) of coif2 were selected as feature spectra by Bayesian discriminate analysis (BDA).The reference spectra were obtained via hierarchical cluster analysis (HCA).With average high correct discrimination ratios (CDRs),reference spectra were representative to phytoplankton species.For one-algae samples,the average CDRs were 95.6% at genus level and 86.7% at division level.For the laboratory mixed samples,the average CDRs (one division accounted for 25%,75% or 100% of the total biomass) were 86.6%,91.4% and 100% at division level.Moreover,the average CDRs of the dominant species (accounted for 75%) was 79.8% at genus level.Results for the in situ samples were coincided with the microscopic ones at division level with the relative contents of 54.3%-96.5%.The fluorometric discriminating technique was further tested during the cruise in Bohai Sea recently.
