A research expedition off the coast of Brazil recovered 31 animals from the mesopelagic zone, the vast midwater layer between roughly 200 and 1,000 meters deep that remains one of the most poorly sampled regions on Earth. The voyage deployed a suite of instruments aboard ROV SuBastian, combining 3D imaging, laser-based flow measurement, robotic tissue capture, and shipboard genomic sequencing to document fragile deep-sea organisms as living animals rather than damaged net samples. The effort matters because mesopelagic creatures drive a significant share of the ocean’s biological carbon pump, yet scientists still lack baseline species data for most of the animals that live there.
Why 31 mesopelagic specimens change the carbon-flux conversation
The mesopelagic zone sits below the sunlit surface waters where satellites can track plankton blooms and above the deep seafloor where submersibles have spent decades mapping geology. That middle layer holds enormous biomass, much of it composed of gelatinous animals, small crustaceans, and larval fish that migrate vertically each night to feed near the surface and sink back by dawn. This daily migration is one of the largest movements of living matter on the planet, and it shuttles carbon from the atmosphere into the deep ocean. Yet because mesopelagic animals are soft-bodied and easily destroyed by trawl nets, researchers have lacked the tools to observe their swimming behavior, measure the fluid dynamics around their bodies, and sequence their genomes from the same individual specimen.
The 31-animal haul from this Brazilian expedition represents a new workflow designed to close that gap. By pairing in-water robotics with onboard microscopy and sequencing, the team generated matched datasets linking an animal’s 3D body shape, its swimming kinematics, its surrounding flow field, and its gene expression profile. Those matched datasets are what modelers need to refine estimates of how much carbon each species actually transports downward. If even a handful of the swimming gaits recorded during this voyage differ substantially from the simplified assumptions used in current carbon-flux models, the correction to South Atlantic estimates could be significant. The hypothesis that four or more previously unmeasured gaits might shift modeled flux by double-digit percentages is testable precisely because the expedition produced kinematic and vertical-migration data from the same specimens, a pairing that did not exist before.
Instruments that captured living deep-sea animals intact
The expedition’s instrument chain began underwater and continued on deck. ROV SuBastian carried two systems built by MBARI’s lab: DeepPIV, a laser sheet that illuminates particles around an animal to map the fluid flow its swimming generates, and EyeRIS, a calibrated plenoptic camera that records high-resolution 3D volumes at high frame rates. Together, these tools captured both the geometry and the hydrodynamics of each animal while it was still alive and behaving normally in its natural habitat.
Once a specimen was identified and imaged, the team used a device called RAD-2, a rotary actuated dodecahedron, to gently encapsulate the animal without the crushing forces of a traditional suction sampler. A peer-reviewed methods paper published in Science Advances describes this integrated approach as an “in situ digital synthesis strategy” for discovering and describing ocean life. The paper lays out how quantitative imaging, robotic encapsulation, tissue preservation, and next-generation genomic sequencing can be combined into a single pipeline that produces a complete biological description from one dive.
On the ship’s deck, two instruments developed by Stanford University’s Prakash Lab took over. The Gravity Machine, described in a Nature Methods paper (DOI: 10.1038/s41592-020-0924-7) as a scale-free vertical tracking hydrodynamic treadmill, allowed researchers to observe how each animal moved vertically under controlled conditions that mimic the open water column. The Squid microscope, an open-source modular confocal system, then imaged internal cellular structures in three dimensions while the animal was still alive. A separate peer-reviewed study documenting transcriptome sequencing of seven deep marine invertebrates independently confirmed the deployment of DeepPIV, EyeRIS, and RAD-2 on ROV SuBastian, providing corroboration of the toolchain from a second research group’s published record in the broader NCBI database.
Gaps in the specimen record and what to watch next
For all the technical sophistication of the instrument suite, several pieces of the story are still missing from the public scientific record. No primary dataset accessible through NCBI repositories yet names the specific 31 specimens, lists their collection coordinates off Brazil, or provides formal species identifications. Shipboard logs and ROV dive reports that would confirm exact deployment depths and RAD-2 capture success rates for each animal have not appeared in curated bibliography collections, leaving outside researchers to infer many operational details from methods descriptions and conference presentations rather than from a dedicated cruise report.
That absence does not negate the scientific value of the expedition, but it does shape how its results can be interpreted. Without specimen-level metadata, modelers cannot yet connect an individual animal’s swimming style and gene expression profile to a precise geographic location or to contemporaneous measurements of temperature, oxygen, and particulate organic carbon. The lack of open cruise documentation also makes it harder to assess potential sampling biases, such as whether the ROV targeted visually conspicuous organisms over transparent or cryptic species that may contribute differently to carbon export.
Several developments could change that picture. First, the team could release a structured dataset linking each of the 31 animals to dive numbers, depth ranges, capture timestamps, and preliminary taxonomic identifications. Even a provisional table, clearly labeled as subject to revision, would allow other groups to begin comparing these Brazilian mesopelagic specimens with existing midwater records from the North Pacific and North Atlantic. Second, depositing raw imaging and flow-field data alongside genomic reads would enable cross-disciplinary analyses of how morphology, behavior, and gene expression co-vary across taxa.
Third, independent cruises could adopt the same integrated toolchain to test how generalizable the Brazilian results are. If DeepPIV and EyeRIS consistently reveal similar swimming gaits in unrelated species that occupy comparable depth bands, that would strengthen the case for revising carbon-flux models globally. Conversely, if different ocean basins host distinct behavioral strategies-even among morphologically similar animals-that would argue for region-specific parameterizations of the biological pump.
Finally, future publications will need to grapple explicitly with uncertainty. Quantitative estimates of how much these 31 specimens might alter regional carbon budgets should include sensitivity analyses that explore how results change under different assumptions about unobserved species, diel migration amplitudes, and seasonal variability. Clear statements about what the available data can and cannot support will be essential as policymakers and climate modelers look to deep-ocean biology for improved projections.
For now, the Brazilian expedition stands as a proof of concept for a new way of doing mesopelagic science: one that treats fragile animals as whole organisms with measurable behaviors, hydrodynamics, and genomes, rather than as anonymous fragments in a net. As more datasets emerge and additional cruises adopt similar workflows, those 31 specimens may come to be seen not just as isolated curiosities, but as the first entries in a more complete inventory of the midwater life that quietly helps regulate Earth’s climate.
More from Morning Overview
*This article was researched with the help of AI, with human editors creating the final content.
