Phytoplankton in the surface ocean are responsible for a significant portion of Earth's primary production, with a fraction of this carbon exported to the deep ocean for long-term sequestration. This process, known as the biological carbon pump, is poorly understood, particularly regarding which primary producers are exported and the ecological conditions influencing export. Improving models of the ocean's carbon cycle requires a mechanistic understanding of this carbon transfer. Technological advancements, including ocean color satellites, in situ imaging, and high-throughput DNA sequencing, allow for high-resolution observations of plankton communities. However, a framework integrating these observations into biological carbon pump models is currently lacking. Previous studies using sediment trap-collected particles have explored the link between phytoplankton communities and POC export through nucleic acid sequencing, offering advantages in sensitivity and throughput compared to microscopy. While these studies have revealed valuable insights, they haven't yet quantitatively linked taxa to POC flux mechanisms, hindering their integration into biological pump models. This research combined quantitative observations of POC flux with DNA sequence communities within individually isolated particles to identify the surface phytoplankton taxa present in sinking particles and their link to POC flux mechanisms.

Numerous studies have examined the relationship between phytoplankton communities and particulate organic carbon (POC) export using various techniques. Microscopy has long been used to visually examine the phytoplankton content of sinking particles, providing valuable insights. However, molecular approaches, such as high-throughput DNA sequencing, offer advantages in sensitivity and throughput, particularly for detecting fragmented or difficult-to-identify organisms. Several studies using molecular methods have confirmed and extended findings from microscopy, demonstrating the importance of pico- and nanophytoplankton as export producers, the significant contribution of heterotrophic protists to POC flux, and the role of specific organisms in episodic POC pulses to the seafloor. Despite these advances, no study had integrated phytoplankton molecular data into models of the biological pump due to the lack of quantitative links between taxa and POC flux mechanisms. This study aimed to bridge this gap by combining microscopy of individually resolved particles with DNA sequencing to link specific taxa to POC flux mechanisms.

Samples were collected during a research cruise in the North Pacific subtropical gyre and the California Current. Sediment traps were deployed at three locations, with particles collected in polyacrylamide gel layers for individual particle isolation and in other tubes for bulk particle collection. Surface seawater samples were also collected. Individual particles larger than 300 µm were isolated from the gel layers using a pipette, flash-frozen, and stored at -80°C. Bulk particle samples were collected by filtration and processed similarly. DNA was extracted from individual and bulk particles and surface seawater using appropriate methods. Approximately 420 bp of the V4 region of the 18S rRNA gene were amplified by PCR and sequenced using Illumina MiSeq. Amplicon sequence variants (ASVs) were assigned taxonomic identities using a naive Bayes classifier against the PR2 database. Data analysis included classifying ASVs as photosynthetic or heterotrophic, identifying unique and shared ASVs across sample types, assessing the effect of sample size on comparisons, and calculating Bray-Curtis dissimilarities to compare community compositions. Statistical analyses included PERMANOVA and t-tests to assess differences among sample groups and particle types. The effects of particle degradation on the detected 18S rRNA gene community were evaluated by comparing samples incubated for different time periods after collection.

In both oligotrophic and coastal ecosystems, a relatively small proportion (approximately 24-26%) of the surface phytoplankton diversity was detected in sinking particles. However, the exported taxa often dominated the surface community in terms of relative read abundance. The relative abundance of exported taxa varied across locations and functional groups. At Station 2 (oligotrophic), exported ochrophytes and hacrobia comprised a large proportion of surface reads, while at Station 3 (California Current), exported diatoms dominated. Exported dinoflagellate and chlorophyte reads were abundant in both ecosystems. At Station 3, approximately half of the phytoplankton and heterotrophic ASVs were present in large (>300 µm) particles, while the other half were found in smaller particles. The size partitioning of ASVs varied among taxonomic groups. Large diatom genera were detected in small particles, while small diatom cells were found in large particles, suggesting different export mechanisms. Differences in 18S rRNA gene community composition were observed among particle types and ecosystems. Heterotrophic taxa were predominant in all sample types. Particle degradation significantly affected the detection of 18S rRNA gene sequences, with dinoflagellates being most affected, followed by diatoms, while chlorophytes showed less degradation. The study revealed a sequential reduction in the relative abundance of phytoplankton in different particle types, suggesting a common zooplankton source for aggregates and dense detritus, derived from large loose fecal pellets.

The study's findings highlight the importance of considering particle size and degradation in understanding carbon export mechanisms. The size fraction in which POC is exported influences POC flux magnitude, sinking speed, and attenuation rate with depth. The observed partitioning of taxa among particle sizes points to the role of food web interactions in determining export pathways. The rapid and differential impact of particle degradation on various taxa emphasizes the need to account for these biases when interpreting data. The differences in community composition between particle types and ecosystems suggest variations in ecological drivers of export. The study's approach of linking POC flux mechanisms with specific phytoplankton taxa provides a valuable framework for developing advanced models of the biological carbon pump.

This study provides a novel approach integrating genetic diversity into models of the ocean's biological carbon pump. By linking POC flux mechanisms to specific phytoplankton taxa and considering particle size and degradation effects, the research improves our understanding of carbon export processes. Future research should focus on expanding the geographic scope of observations to identify generalizable ecological linkages and constrain the effects of degradation on detecting genetic signatures. The integration of optical signatures from satellite data could provide synoptic information about biological carbon pump processes, leading to more realistically constrained models of the ocean's carbon cycle.

The study's findings are based on samples from a limited number of locations and time points. The effects of particle degradation on the detection of genetic signatures need further investigation. The study focused on eukaryotic communities and did not explicitly address the role of prokaryotes in carbon export. Extrapolating findings to broader spatial and temporal scales requires additional research, and the specific mechanisms of particle formation and aggregation require further study.