Search for: All records

Eastern Boundary Systems support major fisheries whose early life stages depend on upwelling production. Upwelling can be highly variable at the regional scale, with substantial repercussions for new productivity and microbial loop activity. Studies that integrate the classic trophic web based on new production with the microbial loop are rare due to the range in body forms and sizes of the taxa. Underwater imaging can overcome this limitation, and with machine learning, enables fine resolution studies spanning large spatial scales. We used theIn-situIchthyoplankton Imaging System (ISIIS) to investigate the drivers of plankton community structure in the northern California Current, sampled along the Newport Hydrographic (NH) and Trinidad Head (TR) lines, in OR and CA, respectively. The non-invasive imaging of particles and plankton over 1644km in the winters and summers of 2018 and 2019 yielded 1.194 billion classified plankton images. Combining nutrient analysis, flow cytometry, and 16S rRNA gene sequencing of the microbial community with mesoplankton underwater imaging enabled us to study taxa from 0.2µm to 15cm, including prokaryotes, copepods, ichthyoplankton, and gelatinous forms. To assess community structure, >2000 single-taxon distribution profiles were analyzed using high resolution spatial correlations. Co-occurrences on the NH line were consistently significantly higher off-shelf while those at TR were highest on-shelf. Random Forests models identified the concentrations of microbial loop associated taxa such as protists,Oithonacopepods, and appendicularians as important drivers of co-occurrences at NH line, while at TR, cumulative upwelling and chlorophyllawere of the highest importance. Our results indicate that the microbial loop is driving plankton community structure in intermittent upwelling systems such as the NH line and supports temporal stability, and further, that taxa such as protists,Oithonacopepods, and appendicularians connect a diverse and functionally redundant microbial community to stable plankton community structure. Where upwelling is more continuous such as at TR, primary production may dominate patterns of community structure, obscuring the underlying role of the microbial loop. Future changes in upwelling strength are likely to disproportionately affect plankton community structure in continuous upwelling regions, while high microbial loop activity enhances community structure resilience.

 

The small sizes of most marine plankton necessitate that plankton sampling occur on fine spatial scales, yet our questions often span large spatial areas. Underwater imaging can provide a solution to this sampling conundrum but collects large quantities of data that require an automated approach to image analysis. Machine learning for plankton classification, and high-performance computing (HPC) infrastructure, are critical to rapid image processing; however, these assets, especially HPC infrastructure, are only available post-cruise leading to an ‘after-the-fact’ view of plankton community structure. To be responsive to the often-ephemeral nature of oceanographic features and species assemblages in highly dynamic current systems, real-time data are key for adaptive oceanographic sampling. Here we used the new In-situ Ichthyoplankton Imaging System-3 (ISIIS-3) in the Northern California Current (NCC) in conjunction with an edge server to classify imaged plankton in real-time into 170 classes. This capability together with data visualization in a heavy.ai dashboard makes adaptive real-time decision-making and sampling at sea possible. Dual ISIIS-Deep-focus Particle Imager (DPI) cameras sample 180 L s -1 , leading to >10 GB of video per min. Imaged organisms are in the size range of 250 µm to 15 cm and include abundant crustaceans, fragile taxa (e.g., hydromedusae, salps), faster swimmers (e.g., krill), and rarer taxa (e.g., larval fishes). A deep learning pipeline deployed on the edge server used multithreaded CPU-based segmentation and GPU-based classification to process the imagery. AVI videos contain 50 sec of data and can contain between 23,000 - 225,000 particle and plankton segments. Processing one AVI through segmentation and classification takes on average 3.75 mins, depending on biological productivity. A heavyDB database monitors for newly processed data and is linked to a heavy.ai dashboard for interactive data visualization. We describe several examples where imaging, AI, and data visualization enable adaptive sampling that can have a transformative effect on oceanography. We envision AI-enabled adaptive sampling to have a high impact on our ability to resolve biological responses to important oceanographic features in the NCC, such as oxygen minimum zones, or harmful algal bloom thin layers, which affect the health of the ecosystem, fisheries, and local communities. 

As the basis of oceanic food webs and a key component of the biological carbon pump, planktonic organisms play major roles in the oceans. Their study benefited from the development of in situ imaging instruments, which provide higher spatio-temporal resolution than previous tools. But these instruments collect huge quantities of images, the vast majority of which are of marine snow particles or imaging artifacts. Among them, the In Situ Ichthyoplankton Imaging System (ISIIS) samples the largest water volumes (> 100 L s -1 ) and thus produces particularly large datasets. To extract manageable amounts of ecological information from in situ images, we propose to focus on planktonic organisms early in the data processing pipeline: at the segmentation stage. We compared three segmentation methods, particularly for smaller targets, in which plankton represents less than 1% of the objects: (i) a traditional thresholding over the background, (ii) an object detector based on maximally stable extremal regions (MSER), and (iii) a content-aware object detector, based on a Convolutional Neural Network (CNN). These methods were assessed on a subset of ISIIS data collected in the Mediterranean Sea, from which a ground truth dataset of > 3,000 manually delineated organisms is extracted. The naive thresholding method captured 97.3% of those but produced ~340,000 segments, 99.1% of which were therefore not plankton (i.e. recall = 97.3%, precision = 0.9%). Combining thresholding with a CNN missed a few more planktonic organisms (recall = 91.8%) but the number of segments decreased 18-fold (precision increased to 16.3%). The MSER detector produced four times fewer segments than thresholding (precision = 3.5%), missed more organisms (recall = 85.4%), but was considerably faster. Because naive thresholding produces ~525,000 objects from 1 minute of ISIIS deployment, the more advanced segmentation methods significantly improve ISIIS data handling and ease the subsequent taxonomic classification of segmented objects. The cost in terms of recall is limited, particularly for the CNN object detector. These approaches are now standard in computer vision and could be applicable to other plankton imaging devices, the majority of which pose a data management problem. 

Blooms of the colonial pelagic tunicate Pyrosoma atlanticum in 2014–2018 followed a marine heatwave in the eastern Pacific Ocean. Pyrosome blooms could alter pelagic food webs of the northern California Current (NCC) by accelerating the biological pump via active transport, fecal pellet production and mortality events. Although aggregations of P. atlanticum have the potential to shape marine trophic dynamics via carbon export, little is known about pyrosome vertical distribution patterns. In this study, we estimated the distribution of P. atlanticum in the NCC along transects off of Oregon (45°N and 124°W) and northern California (41°N and 124°W), USA during February and July 2018. Depth-stratified plankton tows provided volume-normalized pyrosome abundance and biovolume estimates that complemented fine-scale counts by a vertically deployed camera system. Pyrosomes were numerous offshore during February, especially off Oregon. Colonies were distributed non-uniformly in the water column with peak numbers associated with vertical gradients in environmental parameters, notably density and chl-a. Vertical distributions shifted over the 24-h period, indicative of diel vertical migration. Understanding the vertical distribution of these gelatinous grazers in the NCC gives insight to their behavior and ecological role in biologically productive temperate ecosystems as conditions become more favorable for recurring blooms.

 

Abstract Cnidarian jellyfish can be dominant players in the food webs of highly productive Eastern Boundary Currents (EBC). However, the trophic role of inconspicuous hydromedusae in EBCs has traditionally been overlooked. We collected mesozooplankton from five stations along two cross-shelf transects in the Northern California Current (NCC) during winter and summer of 2018–2019. We analyzed gut contents of 11 hydromedusan species and the prey community to (i) determine prey resource use by hydromedusae and (ii) determine temporal shifts in the trophic niche of hydromedusae, focusing on the two most collected species (Clytia gregaria and Eutonina indicans). Hydromedusae in the NCC fed mostly on copepods, appendicularians and invertebrate larvae. Nonmetric multidimensional scaling of hydromedusan diets showed seasonal shifts in prey resource driven by the abundant C. gregaria, which fed mostly on copepod eggs during winter and fed mostly on appendicularians and copepods during summer. Prey selectivity for copepod eggs increased during winter for C. gregaria and E. indicans. Intriguingly, theoretical ingestion rates show that both species acquire similar amounts of carbon during upwelling and nonupwelling conditions. Hydromedusae’s consistent presence and predation impact across seasons may lead to significant effects in carbon and energy transfer through the NCC food web. 

Doliolids are common gelatinous grazers in marine ecosystems around the world and likely influence carbon cycling due to their large population sizes with high growth and excretion rates. Aggregations or blooms of these organisms occur frequently, but they are difficult to measure or predict because doliolids are fragile, under sampled with conventional plankton nets, and can aggregate on fine spatial scales (1–10 m). Moreover, ecological studies typically target a single region or site that does not encompass the range of possible habitats favoring doliolid proliferation. To address these limitations, we combined in situ imaging data from six coastal ecosystems, including the Oregon shelf, northern California, southern California Bight, northern Gulf of Mexico, Straits of Florida, and Mediterranean Sea, to resolve and compare doliolid habitat associations during warm months when environmental gradients are strong and doliolid blooms are frequently documented. Higher ocean temperature was the strongest predictor of elevated doliolid abundances across ecosystems, with additional variance explained by chlorophyllafluorescence and dissolved oxygen. For marginal seas with a wide range of productivity regimes, the nurse stage tended to comprise a higher proportion of the doliolids when total abundance was low. However, this pattern did not hold in ecosystems with persistent coastal upwelling. The doliolids tended to be most aggregated in oligotrophic systems (Mediterranea and southern California), suggesting that microhabitats within the water column favor proliferation on fine spatial scales. Similar comparative approaches can resolve the realized niche of fast‐reproducing marine animals, thus improving predictions for population‐level responses to changing oceanographic conditions.