UC Santa Barbara

Phytoplankton are microscopic protists that are ubiquitous in the sunlit global ocean. These organisms form the base of the marine food web and are essential to biogeochemical cycling as sources and sinks of elemental compounds and nutrients. Carbon sequestration from the atmosphere to ocean sediments is facilitated through biological production by phytoplankton, and phytoplankton produce half of the oxygen in Earth’s atmosphere. Distinct phytoplankton taxa differentially impact these essential ecosystem processes. Thus, a complete understanding of the role of phytoplankton in the Earth system can only be achieved through a complete description of the distribution and abundance of phytoplankton communities in the global ocean. Existing methods to characterize phytoplankton community composition (PCC) using in situ measurements are limited by the scales of observation. However, satellites provide unprecedented coverage of the global surface ocean. While existing global ocean color sensors are limited to multi-spectral sampling resolution, future satellites (such as NASA’s Plankton, Aerosol, Cloud, ocean Ecosystem [PACE] sensor) will have hyperspectral resolution, providing observations across the visible spectrum of light at wavelengths sensitive to absorption and scattering by phytoplankton. Satellite ocean color approaches can therefore be leveraged to distinguish between phytoplankton groups. The major goal of this thesis is to characterize patterns of PCC in the global surface ocean using a combination of existing chemotaxonomic, molecular, and imaging methods with newly-developed remote sensing approaches. In chapter 1, I used quality-controlled, consistent measurements of high performance liquid chromatography (HPLC) phytoplankton pigments collected across the global surface ocean to characterize the distributions of phytoplankton groups from co-variability in phytoplankton pigment concentrations. In both the global dataset and regional time series datasets, the number of phytoplankton groups that could be separated from HPLC pigments was limited across statistical methods to maximum 4-6 distinct pigment-based groups. In chapter 2, the statistical methods employed in chapter 1 were applied to a dataset of HPLC pigments and flow cytometry collected as part of the North Atlantic Aerosols and Marine Ecosystems Study (NAAMES) to describe the evolving surface ocean PCC in the western North Atlantic Ocean across distinct bloom phases. Pigment-based phytoplankton communities revealed a transition from diatoms and dinoflagellates in spring and early summer to haptophytes and cyanobacteria in early fall, followed by green algae and mixed pigment assemblages in early winter. In chapter 3, I modeled phytoplankton pigment concentrations in the global surface ocean from measured and modeled remote sensing reflectance spectra. The concentrations of thirteen pigments were retrieved by the model, and these results were validated with measured HPLC phytoplankton pigment concentrations. The relationships between and among groups of phytoplankton pigments remained consisted between measured and modeled pigment datasets, separating five distinct pigment communities. Finally, in chapter 4, multiple in situ methods were compared to better constrain and quantify the information content of HPLC pigment data using samples collected as part of NAAMES and the first EXport Processes in the Ocean from RemoTe Sensing (EXPORTS) field campaign in the North Pacific Ocean. The eukaryotic phytoplankton community was compared from HPLC pigments, 18S rRNA metabarcoding, and quantitative cell imagery from the Imaging FlowCytobot. The prokaryotic and eukaryotic phytoplankton communities were both compared from HPLC pigments, 16S rRNA metabarcoding, and flow cytometry. While broad group-level trends were consistent between methods, inconsistencies between methods arose at higher taxonomic resolution and when environmental and physiological impacts were taken into account (e.g., sea surface temperature; mixotrophy). Taken together, these four chapters describe PCC from in situ and remote sensing approaches across the global surface ocean. Chapters 1 and 2 use HPLC pigments to describe broad trends in phytoplankton pigments across regions (chapter 1) and seasons (chapter 2). Chapter 3 demonstrates that pigments can be modeled with reasonable accuracy from hyperspectral ocean color data. Chapter 4 then describes the strengths and limitations of HPLC pigment data when compared to methods with higher taxonomic resolution and in the context of a changing ocean environment.