Estimating phytoplankton pigments in the changing Arctic Ocean
|Estimating phytoplankton pigments in the changing Arctic Ocean.pdf||47.03 MB||Adobe PDF||View/Open|
|Authors:||Liu, Yangyang||Supervisor:||Wolf-Gladrow, Dieter||1. Expert:||Wolf-Gladrow, Dieter||Experts:||Zielinski, Oliver||Abstract:||
Human-induced climate change is amplified in the Arctic. At the root of these amplifications are changes in air temperature and sea ice. The sea-ice cover is dramatically receding in the Arctic Ocean. In the study region of the thesis, the Fram Strait (the largest and only deep gateway to the Arctic Ocean) and its vicinity, changes have been observed in sea-ice conditions and water temperatures due to Arctic warming. This has consequences for phytoplankton. Phytoplankton are one of the main primary producers in the Arctic Ocean. Arctic warming induced alterations in light and nutrient regimes impact phytoplankton seasonality, biomass, community composition and distribution. Phytoplankton biomass and community composition are often indicated by their pigment composition and concentrations. To study the response of phytoplankton to the changing climate, this thesis aims to estimate phytoplankton pigments using observations from the shipboard underway flow-through AC-S spectrophotometer system and the Regulated Ecosystem model version 2 (REcoM2) (Hohn, 2008; Schartau et al., 2007) implemented with phytoplankton growth and photoinhibition models.
In the first part of the thesis, an underway flow-through AC-S system was set up onboard R.V. Polarstern during two Fram Strait cruises, PS93.2 and PS99.2. Hyperspectral particulate absorption coefficient was derived from the underway AC-S measurements. Particulate absorption line height at 676 nm calculated from particulate absorption coefficient was empirically related to high performance liquid chromatography (HPLC) chlorophyll a (Chl a) concentrations for PS93.2 and PS99.2, respectively. Both relationships were applied to high frequency (4 Hz) AC-S data to estimate Chl a concentrations along the cruise tracks. In total, 24424 and 16110 Chl a data points were generated for PS93.2 and PS99.2, respectively. The reconstructed AC-S Chl a data sets were used to evaluate seven satellite Chl a algorithms. The number of AC-S-satellite match-ups is over one order of magnitude greater than HPLC-satellite match-ups. AC-S-satellite match-ups show that all algorithms were characterized by an overestimation of satellite Chl a. Two algorithms based on Polymer atmospheric correction processors (Steinmetz et al., 2011) generated data products with relatively high estimation precision and small error. The Polymer atmospheric correction processors account for sun glint and thin clouds in their reflectance models to derive atmospheric corrected remote sensing reflectance, allowing a much larger spatial coverage of data than using standard atmospheric correction processors. In the Arctic Ocean where operational satellite ocean color data have relatively low space-time resolution, Polymer algorithms are promising candidates in enlarging satellite ocean color data sets, e.g., for Sentinel-3/OLCI satellite sensor, given more validation activities are performed in the future.
In the second part of the thesis, the underway flow-through AC-S system was set up onboard R.V. Polarstern during the Fram Strait cruise PS107, in addition to PS93.2 and PS99.2. AC-S derived hyperspectral particulate absorption coefficient were matched with HPLC pigments data. In total, 298 match-ups were used as the pigment retrieval data set. Two pigment retrieval algorithms, Gaussian decomposition (Chase et al., 2013) and the singular value decomposition combined with non-negative least squares (SVD-NNLS) inversion method (Moisan et al., 2011) were compared and optimized for estimating various phytoplankton pigments or pigment groups from the particulate absorption coefficient data. The Gaussian decomposition method provides good estimates (median absolute percentage error, MPE 21-34%) of Chl a, chlorophyll b, chlorophyll c1 and c2, photosynthetic carotenoids and photoprotective carotenoids (PPC). This method outperformed the SVD-NNLS method in retrieving chlorophyll b, chlorophyll c1 and c2, photosynthetic carotenoids, and PPC. However, SVD-NNLS enables robust retrievals of specific carotenoids (MPE 37-65%), i.e., fucoxanthin, diadinoxanthin and 19-hexanoyloxyfucoxanthin, which is currently not accomplished by Gaussian decomposition. More robust predictions are obtained using the Gaussian decomposition method when the observed spectral phytoplankton absorption coefficient is normalized by the package effect index at 675 nm. The latter is determined as a function of ”packaged” phytoplankton absorption coefficient at 675 nm and Chl a concentrations, which shows potential for improving pigment retrieval accuracy by the combined use of phytoplankton absorption coefficient and Chl a data. Both approaches provide useful information on pigment distribution, and hence phytoplankton community composition indicators, at a spatial resolution much finer than can be achieved with discrete HPLC samples.
Xanthophyll pigments provide one of the most important photoprotective mechanisms to dissipate the excess light energy and prevent photoinhibition. In the third part of the thesis, phytoplankton growth models of Geider et al. (1998), the Geider model, and Marshall et al. (2000), the Marshall model, were implemented into REcoM2 to predict the photoprotective needs of phytoplankton and their attributions from phytoplankton PPC, physiological state, and community composition. Assume that photoinhibition is negligible in phytoplankton communities acclimated to ambient light (Cullen et al., 1992). The difference between the photosynthesis–irradiance (P-E) curves with (Marshall) and without photoinhibition (Geider) is considered a measure of photoprotective needs in order to minimize such photoinhibition. The degree of phytoplankton photoprotection is represented by the difference of the initial slopes of the P-E curves between the Marshall and Geider models. It was then related to the HPLC PPC/Chl a data, producing a 4-D global map of PPC/Chl a estimates. These estimates were in agreement with field observations in most of the surface ocean, at depth and even across seasons, suggesting the role of PPC in photoprotective activities in the global ocean. However, at higher latitudes, discrepancies between predictions and observations suggested PPC content was insufficient to satisfy phytoplankton protective needs the community and thus other mechanisms of nonphotochemical quenching were relevant. Furthermore, at higher latitudes, changes in PPC content can result from both physiological acclimation and shifts in community composition while in the rest of ocean taxonomic changes played a main role.
A comprehensive view of the phytoplankton community pigment signature is crucial for modeling the coupling of light absorption to carbon fixation in the ocean. Future validation of the above model can use the combined HPLC observations and pigment estimates from underway flow-through AC-S system. Furthermore, this work provides insights on how much of the variability in community PPC ratios is attributable to changes in community composition or changes in physiological state. This may allow an improvement of the match between satellite ocean color data and the underlying phytoplankton community. In addition, these insights may contribute to a better understanding of the effect of phytoplankton photoacclimation on the accuracy of satellite ocean color products in the Arctic Ocean.
|Keywords:||phytoplankton; pigments; ocean color; light absorption; photoprotection||Issue Date:||20-May-2022||Type:||Dissertation||DOI:||10.26092/elib/1657||URN:||urn:nbn:de:gbv:46-elib60521||Institution:||Universität Bremen||Faculty:||Fachbereich 02: Biologie/Chemie (FB 02)|
|Appears in Collections:||Dissertationen|
checked on Aug 12, 2022
checked on Aug 12, 2022
This item is licensed under a Creative Commons License