Search

Automated camera-based sensors are widely used in vessel-based research to monitor plankton and marine particles. However, current methods suffer from the costly and time-consuming requirement of annotating data for fully supervised learning, especially in plankton grouping tasks characterized by long-tailed datasets. In response, we propose a novel self-supervised learning (SSL) framework that significantly reduces reliance on expensive human annotations by leveraging crucial metadata such as water depth and location. The method comprises three major steps: self-supervised training, innovative sampling, and final classification. It identifies key sample subsets from an unlabelled dataset using hierarchical clustering approach and incorporates an innovative balancing representative subsampling strategy that addresses the challenge of dataset imbalance and enhances generalisability across diverse plankton classes. Our approach prioritises discerning representation features observed in images that exhibit correlations with the patterns found in their associated metadata. Furthermore, our method introduce a novel grouping based on visual perspective selection method, enabling the identification of balanced subset views that depart from traditional class-based categorisation. Our experimental results showcase a significant enhancement in image classification accuracy, with a 23% improvement over methods that do not utilise metadata, and attains a macro F1-score of 54% for 10 populated species from a severely long-tailed dataset. This is achieved with a mere 0.3% of the entire dataset used for annotation.

The biological carbon pump, driven principally by the surface production of sinking organic matter and its subsequent remineralization to carbon dioxide (CO2) in the deep ocean, maintains atmospheric CO2 concentrations around 200 ppm lower than they would be if the ocean were abiotic. One important driver of the magnitude of this effect is the depth to which organic matter sinks before it is remineralised, a parameter we have limited confidence in measuring given the difficulty involved in balancing sources and sinks in the ocean's interior. One solution to this imbalance might be a temporal offset in which organic carbon accumulates in the mesopelagic zone (100–1000 m depth) early in the productive season before it is consumed later. Here, we develop a novel accounting method to address non-steady state conditions by estimating fluxes of particulate organic matter into, and accumulation within, distinct vertical layers in the mesopelagic zone using high-resolution spatiotemporal vertical profiles. We apply this approach to a time series of measurements made during the declining phase of a large diatom bloom in a low-circulation region of the Southern Ocean downstream of South Georgia. Our data show that the major export event led to a significant accumulation of organic matter in the upper mesopelagic zone (100–200 m depth) which declined over the following weeks, implying that temporal offsets need to be considered when compiling budgets. However, even when accounting for this accumulation, a mismatch in the vertically resolved organic carbon budget remained, implying that there are likely widespread processes that we do not yet understand that redistribute material vertically within the mesopelagic zone.

Particle size estimation is key to understanding carbon fluxes and storage in the marine ecosystem. Images of particles provide much information about their size. A subsea digital holographic camera was used to image particles in vertical trajectory in South Georgia. The holograms were processed using a rapid hologram processing suite that extracted focused particle vignettes from these raw holograms. A machine-learning-based method has been developed to analyse the particle size information from these vignettes. To be specific, a structured-forest-based model trained on a group of synthetic holographic particle images is used to detect the particle edges in these vignettes. Following that, a set of pixel-wise morphology operators are used to extract particle regions (masks) from their edge images. Lastly, the size information of the recorded particles can be calculated based on these mask images. The proposed method has been evaluated on a group of synthetic holograms and real holograms, compared with the other ten methods, including four edge-based methods, four region-based methods, a thresholding-based method, and a Kmeans-based method. The results show that our method has the best performance regarding accuracy and processing time. It reaches ∼0.7 of mean IoU and ∼25 s of running time on the 1,000 test vignettes. In terms of qualitative analysis, the regions of the given examples extracted by the proposed method closely match the real particle regions. We also use this method to analyse the size distributions of two profiles, and some generic results are given in this paper.

The TechOceanS project is developing new remote ocean sensing technology supporting wider ocean measurement and a drive to net zero. The project will deliver 5 new sensor classes for biogeochemistry, biology and ecosystems addressing 10 of 19 EOVs, 31 of 73 subvariables, 6 of 9 MSFD targets together with microplastics and a range of biotoxins and contaminants. It will also develop a new image processing workflow for extracting EOVs (9) and MSFD (6) and litter measurements from images. These innovations concentrate on key capability gaps in ocean observing from non-ship systems with a focus on low-cost per measurement through minimised instrument and deployment costs. This paper gives a brief overview of the technologies, and were possible, because of progress or protection of intellectual property, details of our approaches and early results.

Studying the amount, variety, and distribution of microscale particles and plankton at the global scale at different times and seasons is very important to understand ocean environments, and by reference, the global environment. It is important to image them in their natural habitats. Many optical techniques have been developed to meet this requirement. Most techniques suffer from one or another drawbacks, such as limited depth-of-field or low volume. Digital holography is an advanced optical imaging technique to image micro-objects, and it provides high-resolution recording, large depth-of-field and recording volume, and 3D viewing and tracking. This paper describes basic principles of in-line digital holography and provides common image processing methods. Four submersible digital holographic cameras, eHoloCam, RamaCam, weeHoloCam, and LISST-Holo, are introduced, as well as their image processing software. At the end, some limitations and challenges in the underwater holographic imaging systems are mentioned, and some possible solutions are discussed.

Plankton imaging systems supported by automated classification and analysis have improved ecologists' ability to observe aquatic ecosystems. Today, we are on the cusp of reliably tracking plankton populations with a suite of lab-based and in situ tools, collecting imaging data at unprecedentedly fine spatial and temporal scales. But these data have potential well beyond examining the abundances of different taxa; the individual images themselves contain a wealth of information on functional traits. Here, we outline traits that could be measured from image data, suggest machine learning and computer vision approaches to extract functional trait information from the images, and discuss promising avenues for novel studies. The approaches we discuss are data agnostic and are broadly applicable to imagery of other aquatic or terrestrial organisms.

Plankton imaging systems supported by automated classification and analysis have improved ecologists' ability to observe aquatic ecosystems. Today, we are on the cusp of reliably tracking plankton populations with a suite of lab-based and in situ tools, collecting imaging data at unprecedentedly fine spatial and temporal scales. But these data have potential well beyond examining the abundances of different taxa; the individual images themselves contain a wealth of information on functional traits. Here, we outline traits that could be measured from image data, suggest machine learning and computer vision approaches to extract functional trait information from the images, and discuss promising avenues for novel studies. The approaches we discuss are data agnostic and are broadly applicable to imagery of other aquatic or terrestrial organisms.

Recognition of the importance of jellyfish in marine ecosystems is growing. Yet, the biochemical composition of the mucus that jellyfish constantly excrete is poorly characterized. Here we analyzed the macromolecular (proteins, lipids and carbohydrates) and elemental (carbon and nitrogen) composition of the body and mucus of five scyphozoan jellyfish species (Aurelia aurita, Chrysaora fulgida, Chrysaora pacifica, Eupilema inexpectata and Rhizostoma pulmo). We found that the relative contribution of the different macromolecules and elements in the jellyfish body and mucus was similar across all species, with protein being the major component in all samples (81 ± 4% of macromolecules; 3.6 ± 3.1% of dry weight, DW) followed by lipids (13 ± 4% of macromolecules; 0.5 ± 0.4%DW) and carbohydrates (6 ± 3% of macromolecules; 0.3 ± 0.4%DW). The energy content of the jellyfish matter ranged from 0.2 to 3.1 KJ g−1 DW. Carbon and nitrogen content was 3.7 ± 3.0 and 1.0 ± 0.8%DW, respectively. The average ratios of protein:lipid:carbohydrate and carbon:nitrogen for all samples were 14.6:2.3:1 and 3.8:1, respectively. Our study highlights the biochemical similarity between the jellyfish body and mucus and provides convenient and valuable ratios to support the integration of jellyfish into trophic and biogeochemical models.

Zooplankton are fundamental to aquatic ecosystem services such as carbon and nutrient cycling. Therefore, a robust evidence base of how zooplankton respond to changes in anthropogenic pressures, such as climate change and nutrient loading, is key to implementing effective policy-making and management measures. Currently, the data on which to base this evidence, such as long time-series and large-scale datasets of zooplankton distribution and community composition, are too sparse owing to practical limitations in traditional collection and analysis methods. The advance of in situ imaging technologies that can be deployed at large scales on autonomous platforms, coupled with artificial intelligence and machine learning (AI/ML) for image analysis, promises a solution. However, whether imaging could reasonably replace physical samples, and whether AI/ML can achieve a taxonomic resolution that scientists trust, is currently unclear. We here develop a roadmap for imaging and AI/ML for future zooplankton monitoring and research based on community consensus. To do so, we determined current perceptions of the zooplankton community with a focus on their experience and trust in the new technologies. Our survey revealed a clear consensus that traditional net sampling and taxonomy must be retained, yet imaging will play an important part in the future of zooplankton monitoring and research. A period of overlapping use of imaging and physical sampling systems is needed before imaging can reasonably replace physical sampling for widespread time-series zooplankton monitoring. In addition, comprehensive improvements in AI/ML and close collaboration between zooplankton researchers and AI developers are needed for AI-based taxonomy to be trusted and fully adopted. Encouragingly, the adoption of cutting-edge technologies for zooplankton research may provide a solution to maintaining the critical taxonomic and ecological knowledge needed for future zooplankton monitoring and robust evid

The Agulhas Bank on the tip of southern Africa, like other shelf seas, is a relatively productive environment which plays a crucial role in the biology and success of many commercially valuable fish species. Fish and their larvae depend on zooplankton to feed on but, despite their importance, little is known about zooplankton distribution and production on the Agulhas Bank. Here we present results from a survey conducted in March 2019 on the East and Central Agulhas Bank, investigating mesozooplankton abundance, biovolume, taxonomic composition, size distribution (normalised biovolume size spectrum (NBSS) approach) and secondary production. A clear cross-shore gradient was observed with the inner-shelf having higher abundance and biovolume of mesozooplankton dominated by small-size organisms, most likely mirroring higher overall productivity of the coastal waters, while the outer-shelf showed the opposite trend (i.e., low abundance and biovolume; shallow NBSS slopes). This general pattern on the outer-shelf was, however, disturbed in one location (between 24 and 25°E) with a distinguishable mesozooplankton community, most likely linked to the passage of a meander on the inshore side of the Agulhas Current. The Central Agulhas Bank was typified by high mesozooplankton biomass (∼4 g C m-2), comparable to upwelling areas and dominated by copepods and doliolids. High copepod biomass was observed in this region before and was linked to a feature called the “Cold Ridge”. However, during our survey, no such ridge was observed. The mixed layer depth was relatively deep (>20 m) and high Chlorophyll a concentration was measured at depth, despite low net primary production rates. We suggest that this region is prone to other mechanisms such as retention due to cyclonic circulation or processes injecting nutrients in the upper mixed layer that require further research. Secondary production on the Agulhas Bank was in the same range as other shelf seas (0.03–1.55 g C m-2 d-1) and

© The Author(s), 2020. This article is distributed under the terms of the Creative Commons Attribution License. The definitive version was published in Mayers, K. M. J., Poulton, A. J., Bidle, K., Thamatrakoln, K., Schieler, B., Giering, S. L. C., Wells, S. R., Tarran, G. A., Mayor, D., Johnson, M., Riebesell, U., Larsen, A., Vardi, A., & Harvey, E. L. The possession of coccoliths fails to deter microzooplankton grazers. Frontiers in Marine Science, 7, (2020): 562020, doi:10.3389/fmars.2020.569896., Phytoplankton play a central role in the regulation of global carbon and nutrient cycles, forming the basis of the marine food webs. A group of biogeochemically important phytoplankton, the coccolithophores, produce calcium carbonate scales that have been hypothesized to deter or reduce grazing by microzooplankton. Here, a meta-analysis of mesocosm-based experiments demonstrates that calcification of the cosmopolitan coccolithophore, Emiliania huxleyi, fails to deter microzooplankton grazing. The median grazing to growth ratio for E. huxleyi (0.56 ± 0.40) was not significantly different among non-calcified nano- or picoeukaryotes (0.71 ± 0.31 and 0.55 ± 0.34, respectively). Additionally, the environmental concentration of E. huxleyi did not drive preferential grazing of non-calcified groups. These results strongly suggest that the possession of coccoliths does not provide E. huxleyi effective protection from microzooplankton grazing. Such indiscriminate consumption has implications for the dissolution and fate of CaCO3 in the ocean, and the evolution of coccoliths., Mesocosm experiments in 2015 were supported by the Kiel Excellence Cluster “The Future Ocean” (CP1540) and the Leibniz Award to UR, in 2017 the MESOHUX experiment was supported by NSF (OCE-1559179) to KT and KB, NSF (OCE-1537951 and OCE-1459200) to KB, NSF (OCE-1459190, 1657808, and DBI-1624593) to EH, and in 2018 by AQUACOSM (EU H2020-INFRAIA-project No 731065). KM was supported by a NERC Doctoral Training Partnership (DTP) studentship as part of the Southampton Partnership for Innovative Training of Future Investigators Researching the Environment (SPITFIRE, grant number NE/L002531/1) and Research Council of Norway project (#280414) MIXsTRUCT.