Award: OCE-1703664

Right now, in all parts of the ocean, a rain of carbon-containing particles is sinking from the surface and into the deep ocean, effectively transporting carbon away from the atmosphere and sequestering it at depth. The amount of carbon transported by these "marine snow" particles is among the most poorly quantified components of the global carbon cycle and hinders the ability to accurately model current and future changes in the carbon cycle. The uncertainty is largely caused by the lack of data on the complex ecological interactions that eventually result in the vertical export of phytoplankton production. Specific members of the surface phytoplankton community are packaged into different sinking particle types. Each of these particle types contain different quantities of carbon, and differ in their tendencies to sink out of the surface and through the "twilight zone" depths of the ocean. To better quantify the ocean's carbon cycle, this project directly measured 1) which phytoplankton were packaged into specific particle types and 2) how much carbon each particle type exported from the surface ocean. Because these direct observations have traditionally been difficult to collect, we developed new analytical techniques that provide large quantities of biological data about individually-resolved particles. These techniques including sequencing the DNA contents of individually isolated particles, automated image processing of all sinking particle types, and machine learning classifiers to assign particles into ecological groups. We observed these ecological mechanisms of carbon export during a research cruise that transited between Hawaii and Oregon in February-March 2017. At three location we deployed various instruments, including sediment traps that collected particles sinking 150 meters deep in the ocean. These sediment traps drifted freely from the ship for between 1 and 3 days before we recovered them once again and retrieved the particle samples. These samples included particles collected in a way that allowed us to measure the bulk quantity of organic carbon raining down and also to quantify and isolate individual particles. Particles that settled into a jar containing a viscous gel layer were imaged under the microscope and also individually extracted from the gel for DNA analysis. We identified a shift in the ecological mechanisms of carbon export between locations in the subtropical ocean and the California Current. Particles sinking in the subtropical locations contained organisms typical of subtropical environments and most of the carbon was packed into aggregates and microzooplankton fecal pellets. In contrast, a larger percentage of the carbon sinking in the California Current was composed of crustaceous and gelatinous zooplankton fecal pellets. These pellets contained a different phytoplankton community than the cooccurring detrital aggregates, and were enriched in diatoms and green algae. The zooplankton in the California Current appear to have provided a distinct export mechanism for a specific subset of the surface phytoplankton community, especially the smallest phytoplankton cells. These data demonstrate how these particle-resolving optical and genetic approaches can generate the data needed to inform the next generation of ocean carbon models that are based on ecological mechanisms. The approaches that this project enabled us to develop are already being incorporated into other large-scale field campaigns (NASA EXPORTS) for the purposes of improving global carbon cycle models. In addition to these intellectual impacts, this work supported the interdisciplinary collaboration among oceanographers and a computer science student and also supported an early career PI. All data are publicly available in BCO-DMO and NCBI data repositories. Last Modified: 03/24/2020 Submitted by: Colleen A Durkin