Characterizing Phytoplankton Community Composition in the Open and Coastal Waters of the Subarctic Northeast Pacific Using In situ and Remote Sensing Techniques
Date
2024
Authors
Perumthuruthil Suseelan, Vishnu
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Phytoplankton are the foundation of the marine food web and play a central role in global biogeochemical cycles. The overarching phytoplankton community contains a diverse array of species with distinct ecological and biogeochemical traits and niches. As a result, variations in the composition of phytoplankton communities can significantly affect energy transfer through the marine food web, fisheries production, and the vertical export of carbon into the deep ocean. In the subarctic northeast Pacific (SNEP), phytoplankton plays a distinct role in sustaining fisheries through match-mismatch dynamics. Previous studies in this region have mainly relied on in situ measurements, such as HPLC pigment data, Chemical taxonomy software (CHEMTAX), light microscopy, and physical-biological models, to study phytoplankton communities, but lacking the utilization of satellites. This study employs a comprehensive approach, integrating in situ observations, Earth Observation satellite data, and above-water radiometry, to thoroughly examine the composition of phytoplankton communities. To achieve this goal, first we provide evaluation of the Ocean Land Color Instrument (OLCI)-derived products for Case-1 waters, specifically total chlorophyll-a concentration (TChla; mg/m³) and remote sensing reflectance (Rrs; sr⁻¹), obtained from the POLYMER (POLYnomial-based algorithm applied to MERIS) and the European Space Agency’s (ESA) Baseline Atmospheric Correction (BAC) algorithms. Here, we determined that POLYMER performed better than the BAC algorithm across all bands, with the latter resulting in a loss of approximately 60% of valid data points due to the inability to retrieve products under winter cloud conditions. Additionally, POLYMER successfully captured the regional dynamic range of TChla and aligns well with in situ HPLC TChla data. Next, the study characterizes the composition of phytoplankton communities in Case-1, iron-poor open ocean waters and highly dynamic Case-2 coastal waters in the northeast Pacific using data from the Sentinel 3 OLCI satellite. Utilizing an empirical orthogonal function (EOF) approach, a Sentinel 3 OLCI-based algorithm trained with a regional matchup dataset consisting of phytoplankton groups Chlorophyll-a (Chla) and EOF scores derived from Rrs is employed. Additionally, the robustness of the algorithm is statistically assessed through cross-validation. The Sentinel 3 retrievals from 2019 indicate low biomass for all groups in open ocean Case-1 waters, with minimal variability observed across multiple composites. However, increased biomass is observed for all groups except Hapto/Pelago/Cyano towards the northern Gulf of Alaska (GoA), which aligns with CHEMTAX analysis. Conversely, the 2020 composites reveal a longitudinal gradient from the open ocean to the continental shelf/coast for Diatoms/Dino/Green Algae, but an inverse trend is observed for the Hapto/Pelago/Cyano group. Redundancy analysis (RDA) demonstrates that open ocean flagellate groups are positively correlated with Mixed Layer Depth (MLD) and sea surface salinity (SSS), while Diatoms/Dino/GA are negatively correlated with MLD and SSS. In contrast, satellite-based retrievals for the Case-2 water of the west coast of Canada illustrate seasonal spring and fall blooms dominated by diatoms in the Strait of Georgia (SoG), with the highest biomass concentrated in the central, northern, and Juan de Fuca Strait. During the summer, a similar diatom-dominated trend is observed, with the highest biomass recorded on the southwestern coast of Vancouver Island. Interestingly, raphidophytes exhibit the highest average biomass during the summer, with blooms observed off the southwest coast of Vancouver Island and near the Fraser River mouth. The study further explores the spatial-temporal dynamics of phytoplankton composition within the SoG using EOF retrievals from high-resolution hyperspectral Rrs obtained from autonomous platforms. The results indicate better retrievals for all groups than those obtained using Sentinel 3 OLCI data. Overall, the algorithm successfully predicts the localized, diatom-dominated spring bloom and reveals high variability in flagellate communities, with the highest average biomass occurring in the summer. Furthermore, our algorithm captured strong raphidophyte blooms composed of the harmful algae species, Heterosigma akashiwo. Additionally, our high-resolution transect data derived from the algorithm highlighted peaks in raphidophyte and diatom coinciding with a sharp decline in SSS suggesting a frontal region within the Fraser River plume. In summary, using data from in situ measurements, satellites, and autonomous platforms, along with supporting environmental driver data, provides a more comprehensive understanding of the dynamics of individual phytoplankton groups, including harmful algal bloom dynamics with significant implications for water quality monitoring and fisheries management in this region.
Description
Keywords
Phytoplankton Community Composition, Remote Sensing, Hyperspectral radiometry, Northeast Pacific