School of Marine Sciences, Nanjing University of Information Science & Technology, Nanjing 210044, China.
School of Marine Sciences, Nanjing University of Information Science & Technology, Nanjing 210044, China.
Sci Total Environ. 2021 Oct 1;789:147846. doi: 10.1016/j.scitotenv.2021.147846. Epub 2021 May 20.
Marine phytoplankton absorption plays an important role in oceanic biological productivity and ecological environmental dynamics. Understanding the optical absorption variability associated with phytoplanktonic groups still remains a challenge. In this study, samples (n = 206) were collected for the marginal seas of the northwest Pacific Ocean from six cruise surveys that covered different seasons. Using in situ parameters, including phytoplankton absorption coefficients and concentrations of the phytoplanktonic groups derived from phytoplankton pigments collected with high-performance liquid chromatography (HPLC), we developed a Gaussian model to characterize the specific absorption spectra of eight phytoplanktonic groups, including diatoms, chlorophytes, cryptophytes, cyanobacteria, prymnesiophytes, prasinophytes, dinoflagellates, and chrysophytes, without the package effect. The model was established by accurately identifying for the numbers and locations of the Gaussian peaks and their corresponding half-wave widths. The proposed model produced promising results, and a leave-one-out cross validation generated R values exceeding 0.7 for the whole visible light range and above 0.85 (correspondingly MAPE <40%) for the simulated wave bands, excluding the range of 550-650 nm. Meanwhile, a comparison with several spectra observed in the lab showed a high degree of similarity, indicative of the superior performance of our model. Applying the documented specific absorption spectra to the investigated water bodies (whether water surface or profiles) enabled us to quantify the absorption coefficients from different phytoplanktonic groups and characterize their relative contributions to the total. The findings of this study support our understanding of the dynamics of phytoplankton community structure with optical data.
海洋浮游植物的吸收在海洋生物生产力和生态环境动态中起着重要作用。了解与浮游植物群相关的光吸收可变性仍然是一个挑战。本研究共采集了西北太平洋边缘海的 206 个样本，这些样本来自于涵盖不同季节的六次航次调查。利用原位参数，包括浮游植物吸收系数和用高效液相色谱（HPLC）采集的浮游植物色素中浮游植物群的浓度，我们开发了一个高斯模型，以描述八个浮游植物群（包括硅藻、绿藻、隐藻、蓝藻、甲藻、金藻、鞭毛藻和金藻）的特有吸收光谱，该模型不考虑包裹效应。通过准确识别高斯峰的数量和位置及其对应的半峰宽，建立了该模型。该模型产生了令人满意的结果，整个可见光范围的留一法交叉验证生成的 R 值超过 0.7，模拟波段的 R 值超过 0.85（相应的 MAPE<40%），除了 550-650nm 波段。同时，与实验室中观察到的几个光谱进行比较，显示出高度的相似性，表明我们的模型性能优越。将记录的特有吸收光谱应用于研究水体（无论是水面还是剖面），使我们能够量化来自不同浮游植物群的吸收系数，并描述它们对总吸收的相对贡献。本研究的结果支持我们通过光学数据了解浮游植物群落结构动态的理解。
