loading page

Retrieval of phytoplankton pigments and functional types from underway spectrophotometry in the Fram Strait
  • +4
  • Yangyang Liu,
  • Emmanuel Boss,
  • Alison Chase,
  • Yanqun Pan,
  • Hongyan Xi,
  • Rüdiger Röttgers,
  • Astrid Bracher
Yangyang Liu

Corresponding Author:yangyang.liu@awi.de

Author Profile
Emmanuel Boss
Author Profile
Alison Chase
Author Profile
Yanqun Pan
Author Profile
Hongyan Xi
Author Profile
Rüdiger Röttgers
Author Profile
Astrid Bracher
Author Profile


Gaussian decomposition and Singular Value Decomposition combined with Non-Negative Least Squares (SVD-NNLS) method are compared and combined to estimate the concentration of 18 phytoplankton pigments (names and abbreviations are displayed in Table 2) from phytoplankton absorption spectra. Results show that both methods tend to overestimate pigment concentration. Gaussian decomposition method provides robust estimation of TChl-a, TChl-b, Chl-c1/2, PSC and PPC. The estimates of TChl-a, Fuco, Diato, $\beta$-Caro, Prasino, TChl-b, Zea, Viola and Lut from SVD-NNLS show reasonable estimation accuracy, while the other pigments are subjected to relatively large prediction errors. The estimated pigments concentrations are further exploited based on Diagnostic Pigment Analysis to derive four phytoplankton functional types, i.e. diatoms, prymnesiophytes, green algae and prokaryotes. By the application of these two methods to the particulate absorption spectra collected by underway spectrophotometry during three summer cruises in 2015 - 2017 in the Fram Strait, continuous surface phytoplankton functional types are estimated along the cruise course.