Arctic cod (Boreogadus saida) is the dominant pelagic fish in Arctic seas and a staple food of many arctic predators including several seabird species. Marginal ice zones are known as important feeding locations for seabirds. The hypothesis that thick-billed murre (Uria lomvia), northern fulmar (Fulmarus glacialis) and black-legged kittiwake (Rissa tridactyla) congregate in areas of high Arctic cod food resource and low ice concentration was tested at different spatial scales. Arctic cod biomass was estimated by hydroacoustics as a resource proxy, and seabirds were counted and sampled for stomach analysis along eight longitudinal transects across the marginal ice zone in southern Baffin Bay in June–July 2016. With increasing length, the epipelagic age-0 Arctic cod migrated from open waters to ice-covered areas. Subsequently, age-1 and age-2 Arctic cod tended to concentrate in a subsurface layer (40–100 m) within the epipelagic layer. Arctic cod 5.7–16.1 cm long (late age-0 to age-5) were the main fish prey of the three seabird species, which preferentially captured age-1 cod (6–11.5 cm). At large spatial scale (western versus eastern Baffin Bay), thick-billed murre, northern fulmar and their Arctic cod resource proxy were generally more abundant on the western ice-covered side of Baffin Bay. No clear spatial match was found, however, when comparing seabird abundances and their food-resource proxy in different ice concentrations across the marginal ice zone or at small scale (5 km). At medium scale (12.5 km), only murre density was influenced positively by its Arctic cod resource. A lack of schooling behavior and a successful strategy to avoid predation by hiding under the ice could explain the absence of any strong spatial match between Arctic cod and its seabird predators at these different scales.

Echosounders are used by fisheries and ocean observatories, but significant manual effort is required to classify species of interest within multifrequency echograms. This article investigates the use of modified U-Net convolutional neural networks for the pixel-level classification of biological and physical data in echogram images with accurate classification of herring and salmon schools, bubbles, and the sea surface. Data were collected on the coast of British Columbia, Canada, over two years using an Acoustic Zooplankton and Fish Profiler at four frequencies (67, 125, 200, 455 kHz). In addition, simulated data (water depth and solar elevation angle) provide spatial and temporal context to improve the quality of predictions. Redundancy is built into the model by using a tiling strategy during training and classification. During training, using a limited set of annotated data, translational augmentation encodes the U-Nets with robust features that enable applications for alternate deployment configurations (lower sampling rates or alternate water depths). To ensure broad applicability, these networks were trained to classify echograms with noise left intact. The best-performing model classifies herring, salmon, and bubble classes with $\rm {F_{1}}$ scores of 93.0%, 87.3%, and 86.5%, respectively. The results are accurate even when multiple classes are in close proximity, thus, retaining biological data that would otherwise be discarded due to surface bubble noise.

This paper addresses the detection of finfish and krill in echograms. Finfish, in particular Pacific hake, are used both as human food and fish meal. Krill, harvested for aquaculture and aquariums, are a primary food source for finfish, including hake. Thus, spatial distributions of hake follow that of krill. Stock assessments need an accurate differentiation of krill from finfish (hake) in acoustic echograms. This paper proposes a semantic segmentation paradigm for the pixel-level classification of multi-frequency information to detect co-occurring finfish and krill. This paradigm is highly relevant for identifying cloud-like, diffuse krill aggregations that are intertwined with small, often sparse and sometimes dense schools of finfish. We propose U-MSAA-Net, a deep learning U-Net-like framework with novel multi-scale additive attention (MSAA) modules. MSAA modules allow us to leverage all contextual and local information from feature maps available at any given level of the decoding phase of the network, yielding an efficient suppression of the feature responses from regions with lesser semantic value. Experimental results on a new finfish and krill data set spanning across nine months of acoustic data and covering various situations show that U-MSAA-Net outperforms both traditional, texture-based machine learning methods, and deep learning methods based on state-of-the-art semantic segmentation networks. Additional experiments on a data set containing schools of herring and salmon confirm the versatility of U-MSAA-Net and its superiority in terms of accuracy and ability to detect schools of varying sizes. U-MSAA-Net is the first step in creating a comprehensive tool for stock and ecosystem assessments.

In the California Current Ecosystem, the California Undercurrent (CU) is the predominate subsurface current that transports nutrient-rich water from southern California poleward. In this study, we used a large dataset of spatially explicit in situ observations of Pacific hake ( Merluccius productus) and the CU (36.5–48.3°N) to estimate relationships between northward undercurrent velocity and hake distribution and determine whether these relationships vary across space or life-history stage. We found that both hake occurrence and density had strong spatially complex relationships with the CU. In areas north of 44°N (central Oregon), the CU effect was spatially consistent and opposite for occurrence (negative) and density (positive), indicating that hake may aggregate in areas of high northward velocity in this region. In areas south of 44°N, the CU effect showed a cross-shelf gradient for both occurrence and density, indicating a more nearshore hake distribution when northward velocity is higher in this region. Together, our results suggest that future changes in the CU due to climate change are likely to impact hake differently in northern and southern areas.

Humpback whales ( Megaptera novaeangliae) use southern British Columbia waters to feed, but the type and quantity of prey in many areas used for feeding is unknown. We conducted active acoustic prey mapping in 55 small grid-surveys in two regions off Vancouver Island. We quantitatively compared fish and zooplankton-dominated biomass in known feeding areas with and without foraging humpback whales, and qualitatively described the prey characteristics of the foraged areas. Surveys of the water column suggest that, on average, humpback whale foraging was associated more with increased zooplankton than fish biomass. Prey characteristics varied between the two regions (∼500 km apart), but there was no significant difference in mean backscatter strength in the actively foraged areas between the two regions. Frequency differencing discriminated between the dominant taxa in the water column, but potential epipelagic prey (<10 m) would have been omitted from analysis. However, average depth at the maximum acoustic prey detections was significantly deeper when whales were present (84 m) versus absent (60 m), suggesting predominantly subsurface foraging opportunities suitable to prey mapping.