Decomposition Of Phytoplankton Absorption Spectra Based On Phytoplankton Community Structure Characteristics In The East China Sea

Phytoplankton specific absorption coefficient is an important parameter to correlate phytoplankton biomass and absorption spectrum.The quantitative analysis of the specific absorption coefficient of phytoplankton population makes up for the deficiency of the study on the specific absorption spectrum of phytoplankton population in the coastal waters of China,and provides a new sight to monitor the concentration of phytoplankton population.Compared with the traditional site observation method,satellite can provide large area synchronous observation data,which is an ideal method to study the phytoplankton population structure.Taking the East China Seas?Yellow Sea,Bohai Sea and East Sea?as the study area,based on the field measurement data,this paper developed a phytoplankton absorption decomposition model and an phytoplankton absorption inversion model,and established a remote sensing inversion model of absorption at 675nm based on the satellite remote sensing data,and explored and analyzed the population absorption characteristics.The main research contents are as follows:According to the additivity of phytoplankton absorption spectrum,the multiple linear regression model and gaussian decomposition model between the concentration of dominant algae species and the absorption coefficient of phytoplankton were established,and then the field measured data were used to verify the two models.The determination coefficient R²was greater than 0.8,MAPE was less than 50%and RMSE was less than 0.04 m^-1.In addition,the absorption decomposition models were compared by three precision evaluation indexes.The MAPE of this two models was very close,but the R² of the multivariate linear regression model was higher than that of the gaussian decomposition model,and the RMSE of the gaussian decomposition model was much higher than that of the multivariate linear regression model,which mean the multiple linear regression model is confirmed to be the optimal model.Based on the specific absorption spectra of living algae in laboratory culture obtained from previous studies,the specific absorption spectra obtained by inversion in this study were verified.The results show that the shape and magnitude of the absorption spectrum obtained in this study are similar to the spectral characteristics of previous studies.Based on a_ph???and the measured chlorophyll data in the East China Sea,the power function models of Chla and a_ph???and the parametric models of a_ph?675?and a_ph???were established respectively,then the measured data were used to validate models.The determination coefficient R² were greater than 0.7,the algorithm precision was higher.And,the parametric model based on 675nm band is superior to the power function model based on chlorophyll concentration and phytoplankton absorption coefficient.Exploring the intrinsic relationship between phytoplankton absorption and remote sensing reflectance,and the spectral characteristic characterization factors of phytoplankton absorption at 675nm were extracted.The X1 band had the highest correlation coefficient with phytoplankton absorption coefficient,with the result of 0.99.And the results fitted with five functions show that the phytoplankton absorption derived from the model fitting of cubic polynomial function has the best performance.The cross validation of the model with one method based on the measured data shows that the accuracy of inversion was high,with the R²of 0.999,RMSE of 0.0001 m^-1.Based on the GOCI satellite data matched up with the data measured in situ,the phytoplankton absorption decomposition model,the phytoplankton absorption inversion model and the cubic polynomial inversion model of a_ph?675?were used to carry out the inversion of the phytoplankton population concentration,which had good performance.Except Cyanobacteria,the R² of the other algae wear all greater than 0.7 with smaller MAPE and RMSE.The absorption coefficient of phytoplankton population was obtained by combining with the obtained population specific absorption coefficient.Combined with the 2016 satellite image data,the population absorption contribution rate of the study area was obtained,and the change law of population absorption contribution rate at different spatial and time scales was obtained.Chlorophytes has a large absorption contribution in offshore areas,while Cryptophytes absorption has a greater contribution in near-shore areas.From the perspective of seasonal changes,the absorption of Chrysophytes in the spring has a greater contribution to the total absorption,Chlorophytes has a larger absorption contribution in the summer,at the same time,the absorption contribution of Dinoflagellates in the autumn is relatively high,Judging from the monthly changes,the absorption contribution of Prasinophytes and Cryptophytes will show two high values in summer and winter,Cyanobacteria have a“strong peak”in winter and a“weak peak”in autumn,the five study areas have curves in March to October.The shapes are similar,and there are differences in the contribution.overall,Prasinophytes absorption contributes the least to total absorption.