Linking phytoplankton absorption to community composition in Chinese marginal seas
Introduction
In marine waters, phytoplankton assemblages play essential roles in regulating the food web, ecological and biogeochemical processes, and carbon export and cycling (Field et al., 1998, Le Quéré et al., 2005, Guidi et al., 2016, Mouw et al., 2016). It is thus desirable to assess variations in phytoplankton assemblages. While field and laboratory techniques (e.g., microscopic taxonomy, flow cytometry, and high performance liquid chromatography or HPLC) provide accurate measurements of phytoplankton assemblages or pigment composition (Jeffrey, 1997, MacIsaac and Stockner, 1993, Roy et al., 2013, Sosik et al., 2010), they are limited in both space and time. Ocean color remote sensing provides more synoptic and frequent measurements of the surface ocean's optical properties, yet it is not trivial to infer information on phytoplankton assemblages from optical properties.
In the past decades, progress has been made towards retrievals of information on phytoplankton assemblages through remote sensing (Sathyendranath et al., 2004, Brewin et al., 2014, Brewin et al., 2015, Wang et al., 2016, Sun et al., 2019). Chlorophyll a (Chla) pigment is usually regarded as an indicator of phytoplankton (Falkowski et al., 1998, Yoder and Kennelly, 2003, Dierssen, 2010, Harding et al., 2016). While there are correlations in most ecosystems between Chla and other accessory pigments, it is well established that Chla does not carry taxonomic information (Chase et al., 2013). Other methods have been developed for this purpose. For instance, detecting phytoplankton size classes (PSCs) enables us to divide the phytoplankton community into different classes specific to different cell size ranges; three size classes, including pico- (<2 μm), nano- (2 – 20 μm), and microphytoplankton (>20 μm), are generally detected by using those pigment-based methods (Brewin et al., 2015, Devred et al., 2011, Uitz et al., 2006, Hirata et al., 2011). In order to quantify different size classes of phytoplankton, phytoplankton assemblages are strictly divided into three size ranges. Yet, some species are distributed across wide size ranges and even simultaneously belong to two size classes. For instance, diatoms are usually located in micro- and nanophytoplankton classes and prymnesiophytes coexist in nano- and picophytoplankton classes (Uitz et al., 2008, Brewin et al., 2014, 2015). Besides, the same size class sometimes contains multiple phytoplankton communities (Uitz et al., 2008, Brewin et al., 2014, Brewin et al., 2015, Sun et al., 2017, Sun et al., 2019, Sun et al., 2018). The detection of one phytoplankton size class cannot always clearly denote the existence of specific phytoplankton communities in marine waters. Thus, PSCs show a limited capacity in characterizing phytoplankton community composition, despite their superiority to a single Chla indicator.
Another recent work estimates concentrations of phytoplankton pigments through optical data, as pigment data have been used widely to characterize phytoplankton community composition. The pigments or pigment groups have been modeled using phytoplankton absorption signatures. For instance, Moisan et al. (2011) made use of in situ data (phytoplankton absorption spectra and HPLC pigments) in a linear inverse calculation to extract pigment-specific absorption spectra, and then established a second linear inverse calculation with the total phytoplankton absorption spectra, and estimated seven pigments with good correlations (R2 > 0.5). Chase et al. (2013) reported that the total Chla, chlorophyll b (Chlb), and chlorophyll c (Chlc), as well as photosynthetic carotenoids (PSC) and photoprotective carotenoids (PPC), could be predicted via decomposing the in situ particulate absorption spectra into absorption by the five groups of pigments and non-algal particles, and generated percent errors between 30% and 57%. Note that the pigments predicted here were not individual pigments, just five summed-pigment groups, such as total Chla (representing a sum of Chla, divinyl chlorophyll a (DV-Chla), and chlorophyllide a (Chlide-a)), total Chlb (Chlb and divinyl chlorophyll b (DV-Chlb)), Chlc (chlorophyll c1 (Chlc1) and chlorophyll c2 (Chlc2)), PSC (19′-hexfucoxanthin (19′-hex), fucoxanthin (Fuco), 19′-butfucoxanthin (19′-but), and peridinin (Per)), and PPC (α-carotene (α-Car), β-carotene (β-Car), zeaxanthin (Zea), alloxanthin (Allo), and diadinoxanthin (Diadino)). Liu et al. (2019) compared the performances of two approaches, namely Gaussian decomposition and matrix inversion, and demonstrated that the Gaussian decomposition showed good estimates of total Chl-a, Chlb, Chlc, PPC, and PSC, while the matrix inversion algorithm enabled robust retrievals of specific carotenoids, including Fuco, Diadino, and 19′-hex, unavailable to Gaussian decomposition. However, despite the progress made on absorption-based pigment retrievals, the retrievals are for the five discrete pigment groups rather than all individual pigments, making it difficult to estimate phytoplankton community composition.
Compared with the HPLC analysis, remote sensing algorithms to estimate pigments have an advantage of large-scale applications through satellite data. Pan et al. (2010) established regression models between band ratios of remote sensing reflectance (Rrs) and pigment concentrations. Those models were then applied to MODIS images to map the spatial concentration distributions of Chla, Fuco, Per, and Zea on the northeast coast of the United States. Using in situ hyperspectral Rrs measurements, Chase et al. (2017) developed an inversion algorithm that defined phytoplankton pigment absorption as a sum of Gaussian functions, where amplitudes of the Gaussian functions would be associated with concentrations of Chla, Chlb, Chlc, and PPC. For coastal waters, Craig et al. (2012) retrieved the Chla concentrations and phytoplankton absorption spectra (aph) based on the empirical orthogonal function (EOF) analysis and found that the spectral normalization was the key to the models' success. Bracher et al. (2015) applied an EOF analysis to Rrs data to predict a suite of pigments and pigment groups with high quality, which developed multiple linear regression models with measured pigment concentrations and EOF loadings. The EOF method was also utilized later in Palacios et al., 2015, Xi et al., 2015, Xi et al., 2020), Bracher et al. (2020). In recent work, a neural network classifier was used to estimate ten pigments from satellite ocean color observations (El Hourany et al., 2019). However, similar to the existing absorption-based analyses, these Rrs-based analyses also focused on major pigment groups rather than individual pigments, thus facing the same difficulty when estimating phytoplankton community composition. On the other hand, the variations in phytoplankton pigments directly impact phytoplankton absorption signals in water bodies. Therefore, absorption-based methods have a direct advantage in revealing the variability of phytoplankton pigment concentrations.
Thus, this study aims to establish a regionally applicable approach to determine the phytoplankton pigments of Chinese marginal seas through a series of analyses of phytoplankton absorption and pigments. To achieve this goal, the study analyzes an extensive data set of HPLC-derived concentrations of twenty phytoplankton pigments (including Chla and nineteen biomarker pigments), as well as phytoplankton absorption from each of the water samples. The use of phytoplankton absorption spectra is to seek potential remote sensing applications in the future as the former can, at least in principle, be derived from the latter for each remote sensing image pixel. Once individual pigments of a given water sample are estimated from its absorption spectrum, the water sample's phytoplankton community composition can be determined from a hierarchical cluster analysis of HPLC pigment data collected from the study region.
Section snippets
Study areas
Our study areas are located in the Chinese marginal seas, including the Bohai Sea (BS), Yellow Sea (YS), and East China Sea (ECS). The BS is located to the northeast of the other Chinese marginal seas and has a total area of approximately 77,000 km2 and an average depth of 18 m. Adjacent to the BS, the YS has an area of approximately 380,000 km2 and an average depth of approximately 44 m (Feng et al., 1999, He et al., 2004). The ECS has more open water areas and a mean depth of 370 m, with most
Hierarchical cluster analysis of measured HPLC pigments
Our study aims to model 20 measured HPLC pigments from the absorption spectra. Pigment classification was done based on the measured datasets and published methods to interpret different clusters into different community compositions (Aiken et al., 2009, Jeffrey et al., 2011, Moisan et al., 2013; Roy et a., 2013; Uitz et al., 2006, Vidussi et al., 2001). As shown in Fig. 3, several pigments may be grouped into a biological set dominated by diatoms, given that (1) the Fuco pigment is a
Discussion
Phytoplankton pigments naturally exist in phytoplankton species groups (Alves-De-Souza et al., 2008, Sarmento and Descy, 2008, Udovič et al., 2015, Emanuele et al., 2017). However, it is difficult to directly link those pigments to phytoplankton community composition. One pigment may occur in more than one phytoplankton group, while one phytoplankton group may contain multiple pigments with variable proportions (Salmaso et al., 2013, Salmaso and Padisák, 2007, Kruk et al., 2010). Despite such
Conclusion
Characterizing phytoplankton pigments through their absorption properties is technically challenging for the Chinese marginal seas where waters are optically complex. This work addresses this challenge using an extensive dataset together with derivative analysis and optical modeling. The pigment inversion is successfully acquired from the linkage between phytoplankton absorption spectra (first-order and second-order derivative) and pigments. The classification result from hierarchical
Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgments
This research was jointly supported by the National Natural Science Foundation of China (No. 41876203, 41576172, and U1901215), the National Key Research and Development Program of China (No. 2016YFC1400901 and 2019YFD0901305), the Jiangsu Six Talent Summit Project (No. JY-084), the Qing Lan Project, CEReS Oversea Joint Research Program, Chiba University (No. CI19-103 and CI20-104), and also sponsored by the NSFC Open Research Cruise (Cruise No. NORC2018-01), funded by Shiptime Sharing Project
References (84)
- et al.
Phytoplankton pigments and functional types in the Atlantic Ocean: A decadal assessment, 1995–2005
Deep Sea Research Part II: Topical Studies in Oceanography
(2009) - et al.
Biomarker pigment signatures in cochin back water system-a tropical estuary south west coast of India
Estuarine, Coastal and Shelf Science
(2012) - et al.
Distributions of suspended sediment concentration in the Yellow Sea and the East China Sea based on field surveys during the four seasons of 2011
J. Mar. Syst.
(2013) - et al.
An intercomparison of bio-optical techniques for detecting dominant phytoplankton size class from satellite remote sensing
Remote Sensing of Environment
(2011) - et al.
Influence of light in the mixed-layer on the parameters of a three-component model of phytoplankton size class
Remote Sensing of Environment
(2015) - et al.
Decomposition of in situ particulate absorption spectra
Methods in Oceanography
(2013) Turbidity as a control on phytoplankton biomass and productivity in estuaries
Continental Shelf Research
(1987)- et al.
Deriving optical metrics of coastal phytoplankton biomass from ocean colour
Remote Sens. Environ.
(2012) - et al.
A three component classification of phytoplankton absorption spectra: Applications to ocean-colour data
Remote Sensing of Environment
(2011) - et al.
Phytoplankton distributions and their relationship with the environment in the Changjiang Estuary
China. Mar. Pollut. Bull.
(2005)
Shallow water tidal constituents in the Bohai Sea and the Yellow Sea from a numerical adjoint model with TOPEX/POSEIDON altimeter data
Cont. Shelf Res.
Using geostationary satellite ocean color data to map the diurnal dynamics of suspended particulate matter in coastal waters
Remote Sensing of Environment
A comparison of phytoplankton populations of the Arabian Sea during the Spring Intermonsoon and Southwest Monsoon of 1995 as described by HPLC-analyzed pigments
Deep Sea Research, Part II
Retrieval of phytoplankton pigments from underway spectrophotometry in the Fram Strait
Remote Sensing
Effects of suspended sediments on coral growth: evidence from remote sensing and hydrologic modeling
Remote Sensing of Environment
Algorithm development for predicting biodiversity based on phytoplankton absorption
Continental Shelf Research
Remote sensing of phytoplankton functional types in the coastal ocean from the hyspiri preparatory flight campaign
Remote Sensing of Environment
Remote sensing of phytoplankton pigment distribution in the United States northeast coast
Remote Sensing of Environment
Remote sensing of spatial and temporal patterns of phytoplankton assemblages in the Bohai Sea, Yellow Sea, and east China sea
Water Research
Computer-assisted high-performance liquid chromatography method development with applications to the isolation and analysis of phytoplankton pigments
Journal of Chromatography A
Global retrieval of phytoplankton functional types based on empirical orthogonal functions using cmems globcolour merged products and further extension to olci data
Remote Sensing of Environment
Retrieval of total suspended matter concentration in the Yellow and East China Seas from MODIS imagery
Remote Sensing of Environment
Functional groups in marine phytoplankton assemblages dominated by diatoms in fjords of southern chile
Journal of Plankton Research
Controls on temporal patterns in phytoplankton community structure in the Santa Barbara Channel
California. Journal of Geophysical Research
Evaluating the importance of phytoplankton community structure to the optical properties of the Santa Barbara Channel
California. Limnology and Oceanography: Methods
Using empirical orthogonal functions derived from remote-sensing reflectance for the prediction of phytoplankton pigment concentrations
Ocean Sci.
High resolution water column phytoplankton composition across the atlantic ocean from ship-towed vertical undulating radiometry. Frontiers in Marine
Science
A multicomponent model of phytoplankton size structure
Journal of Geophysical Research
Phytoplankton pigment communities can be modeled using unique relationships with spectral absorption signatures in a dynamic coastal environment
Journal of Geophysical Research: Oceans
Estimation of phytoplankton accessory pigments from hyperspectral reflectance spectra: Toward a global algorithm
Journal of Geophysical Research: Oceans
Phytoplankton Community Structure in the Transects across East China Sea and Northern South China Sea Determined by Analysis ofHPLC Photosynthetic Pigment Signatures
Advances in Earth Science (Chinese).
The trophic status of various oceanic provinces as revealed by phytoplankton pigment signatures
Limnology and Oceanography
Dataset on the absorption characteristics of extracted phytoplankton pigments
Data in brief
Quantifying absorption by aquatic particles: a multiple scattering correction for glass-fiber filters
Limnol. Oceanogr.
Perspectives on empirical approaches for ocean color remote sensing of chlorophyll in a changing climate
Proceedings of the National Academy of Sciences of the United States of America
Estimation of Secondary Phytoplankton Pigments From Satellite Observations Using Self-Organizing Maps (SOMs)
J. Geophys. Res. Ocean.
On the discrimination of multiple phytoplankton groups from light absorption spectra of assemblages with mixed taxonomic composition and variable light conditions
Applied Optics
Phytoplankton across tropical and subtropical regions of the Atlantic
Indian and Pacific oceans. PLoS One
Biogeochemical Controls and Feedbacks on Ocean Primary Production
Science
Primary production of the biosphere: Integrating terrestrial and oceanic components
Science
Plankton networks driving carbon export in the oligotrophic ocean
Nature
Cited by (10)
Composition, distribution, and source of organic carbon in surface sediments of Erhai Lake, China
2023, Science of the Total EnvironmentCitation Excerpt :MBC content increases with increasing water content (Li et al., 2020a; Wu et al., 2015; Wu et al., 2012). Chlorophyll a, as an indicator of the total phytoplankton biomass (Feng et al., 2021; Lugoli et al., 2012; Sun et al., 2021), exhibited different relationships with the various SOC fractions. Overall, W-Chl a had a modest effect on the SOC pool, with only LFOC being significantly correlated with W-Chl a (p < 0.01).
Bio-optical depiction of a polar ocean under global change: Exploring the regional absorption traits
2022, Global and Planetary ChangeCitation Excerpt :The in-situ optical characterization of oceanic waters at a regional scale is essential to parameterize the light absorption coefficients of the optical constituents. This would help in checking their applicability in global biogeochemical and remote sensing studies (Johnson et al., 2009; Sun et al., 2021). However, most of the studies focussing on light absorption budget usually have stressed upon the fine resolution of physical processes and have left the biological understanding less attended (Asselot et al., 2021).
Two-decadal estimation of sixteen phytoplankton pigments from satellite observations in coastal waters
2022, International Journal of Applied Earth Observation and GeoinformationCitation Excerpt :Finally, aph(λ) was calculated as the difference between ap(λ) and ad(λ) (spectral range of 400–700 nm, 1 nm intervals) (Cleveland and Weidemann, 1993). More details can be found in Sun et al. (2021). Data quality control of the aph(λ) was carried out according to the method described by Hirata et al. (2008).