Purpose of Review The changes or updates in ocean biogeochemistry component have been mapped between CMIP5 and CMIP6 model versions, and an assessment made of how far these have led to improvements in the simulated mean state of marine biogeochemical models within the current generation of Earth system models (ESMs). Recent Findings The representation of marine biogeochemistry has progressed within the current generation of Earth system models. However, it remains difficult to identify which model updates are responsible for a given improvement. In addition, the full potential of marine biogeochemistry in terms of Earth system interactions and climate feedback remains poorly examined in the current generation of Earth system models. Summary Increasing availability of ocean biogeochemical data, as well as an improved understanding of the underlying processes, allows advances in the marine biogeochemical components of the current generation of ESMs. The present study scrutinizes the extent to which marine biogeochemistry components of ESMs have progressed between the 5th and the 6th phases of the Coupled Model Intercomparison Project (CMIP).

Among marine organisms, gelatinous zooplankton (GZ; cnidarians, ctenophores, and pelagic tunicates) are unique in their energetic efficiency, as the gelatinous body plan allows them to process and assimilate high proportions of oceanic carbon. Upon death, their body shape facilitates rapid sinking through the water column, resulting in carcass depositions on the seafloor ("jelly-falls"). GZ are thought to be important components of the biological pump, but their overall contribution to global carbon fluxes remains unknown. Using a data-driven, three-dimensional, carbon cycle model resolved to a 1°global grid, with a Monte Carlo uncertainty analysis, we estimate that GZ consumed 7.9-13 Pg C y −1 in phytoplankton and zooplankton, resulting in a net production of 3.9-5.8 Pg C y −1 in the upper ocean (top 200 m), with the largest fluxes from pelagic tunicates. Non-predation mortality (carcasses) comprised 25% of GZ production, and combined with the much greater fecal matter flux, total GZ particulate organic carbon (POC) export at 100 m was 1.6-5.2 Pg C y −1 , equivalent to 32-40% of the global POC export. The fast sinking GZ export resulted in a high transfer efficiency (T eff) of 38-62% to 1,000 m and 25-40% to the seafloor. Finally, jelly-falls at depths >50 m are likely unaccounted for in current POC flux estimates and could increase benthic POC flux by 8-35%. The significant magnitude of and distinct sinking properties of GZ fluxes support a critical yet underrecognized role of GZ carcasses and fecal matter to the biological pump and air-sea carbon balance. Plain Language Summary Marine ecosystems play a critical role in the global carbon cycle through food web regulation of air-sea carbon fluxes and the transfer of organic carbon from the upper oceans to the deep sea. The carcasses of gelatinous zooplankton (GZ), which include jellyfish and salps, have been found in mass seafloor depositions ("jelly-falls") in many locations. These jelly-falls are thought to be a fast mechanism for carbon sequestration, yet no global studies on their overall impact have been done. Using a database of GZ observations, we suggest that the inclusion of previously unaccounted for GZ carbon in seafloor carbon deposition could increase current estimates by 8-35%. This previously unconsidered flux represents a substantial amount of carbon sequestered in the deep sea.

The rise of in situ plankton imaging systems, particularly high-volume imagers such as the In Situ Ichthyoplankton Imaging System, has increased the need for fast processing and accurate classification tools that can identify a high diversity of organisms and nonliving particles of biological origin. Previous methods for automated classification have yielded moderate results that either can resolve few groups at high accuracy or many groups at relatively low accuracy. However, with the advent of new deep learning tools such as convolutional neural networks (CNNs), the automated identification of plankton images can be vastly improved. Here, we describe an image processing procedure that includes preprocessing, segmentation, classification, and postprocessing for the accurate identification of 108 classes of plankton using spatially sparse CNNs. Following a filtering process to remove images with low classification scores, a fully random evaluation of the classification showed that average precision was 84% and recall was 40% for all groups. Reliably classifying rare biological classes was difficult, so after excluding the 12 rarest taxa, classification accuracy for the remaining biological groups became > 90%. This method provides proof of concept for the effectiveness of an automated classification scheme using deep-learning methods, which can be applied to a range of plankton or biological imaging systems, with the eventual application in a variety of ecological monitoring and fisheries management contexts.