Skip to Main Content

Breadcrumb

EBSNoteworthy

Eastern Bering Sea 2024 Noteworthy Topics

Here we present items that are new or noteworthy and of potential interest to fisheries managers.

Environmental DNA: Poised to Transform How We Track Fish Populations

Background
Environmental DNA (eDNA) consists of skin, scales, cells, and DNA from organisms released into the environment. In marine ecosystems, researchers collect and filter water to concentrate biological and genetic material. DNA is then extracted, and DNA sequencing provides information about species near the sampling location. This technology can yield presence and relative abundance data for individual species-of-interest or community composition and biodiversity (Thomsen et al., 2012). Each water sample can be analyzed for multiple taxonomic groups: for example, DNA primers can be selected to target fishes, zooplankton, bacteria, or marine mammals.

The two primary approaches for analyzing eDNA samples are: 1) single-species quantitative PCR (qPCR) and 2) multi-species metabarcoding. Single-species qPCR data provide quantification, whereas metabar-coding results in compositional data that can be analyzed for relative abundance of the species present (Figure 4).

eDNA can reliably generate presence/absence data with appropriate DNA primers and reference DNA libraries. eDNA concentration is correlated with abundance or biomass (Spear and Andrews III, 2021), but multiple factors influence this relationship, most notably body size (Rourke et al., 2022; Yates et al., 2023) and distance between the organism and sampling location (Baetscher et al., 2024). Both presence/absence data and eDNA concentration could be paired with other survey approaches to identify species, for example, in acoustic trawl surveys, particularly in the absence of catch data. Similarly, eDNA could be collected alongside camera images and used to verify species identities or potentially increase the detection distance beyond the camera field-of-view.

Applications
The relative ease of collecting eDNA from water compared to more resource-intensive methods makes eDNA an efficient way to add biological observations that could be used for species distribution models (Riaz et al., 2020), generating indices of abundance (Shelton et al., 2022), filling in spatial data gaps between trawl stations, or extending a survey footprint into habitats poorly suited to other collection methods. Further, because sampling is non-lethal, eDNA is a viable sampling option within protected areas (Gold et al., 2021). Water samples for eDNA may be collected from fisheries survey vessels at stations using cast rosettes of Niskin bottles or using passive collectors deployed alongside or independent of nets (i.e., Maiello et al., 2022). Samples may also be collected autonomously from moorings and uncrewed vessels, including Saildrone using flow-through systems while underway (Preston et al., 2024).

image

Figure 4: Schematic of the process from water filtration to taxonomic assignment used in eDNA metabarcoding.

Although in the early stages of its field applications, eRNA may provide information about life-history stages (juveniles vs. adults) and physiology (Parsley and Goldberg, 2024; Yates et al., 2023), expanding the potential utility of environmental genetic data for contributing to survey data products.

The potential for eDNA to provide data for fisheries assessments and management is recognized by both eDNA researchers and assessment scientists (Kasmi et al., 2023; Ram´irez-Amaro et al., 2022; Rourke et al., 2022; Stoeckle et al., 2020). However, discrepancies between eDNA and more traditional survey methods are perceived as limitations that undermine the utility of eDNA (Jerde, 2021). Some of these limitations include information about abundance or biomass; size/age class; or questions about the spatial and temporal area sampled by eDNA (Jerde, 2021). Furthermore, eDNA typically detects more species than surveys, which could be an asset (e.g., detecting species that can avoid nets), but also complicates comparisons between data sources. While outstanding questions remain, the field of eDNA has done considerable work to better characterize the dynamics of eDNA, including understanding shedding and degradation rates (Collins et al., 2018), dilution and transport (Baetscher et al., 2024; Shea et al., 2022), and the relationship between eDNA concentration and biomass (Figure 5).

image

Figure 5: Results from an eDNA aquarium study with Pacific and Arctic cods and walleye pollock that shows a strong correlation between proportional biomass and eDNA sequencing reads. Aquarium tanks included fish from one or more species and each point in the figure corresponds to the biomass in a given trial. Solid lines represent linear models for each species. eDNA data were transformed using a quantitative metabarcoding model that accounts for amplification bias across the three species. Aquarium experiments were performed by the AFSC Fisheries Behavioral Ecology Program at the Hatfield Marine Science Center.

In order to validate the data generated from eDNA, species composition has been compared to other survey methods, including trawl catches (Kasmi et al., 2023; Maes et al., 2024; Salter et al., 2019), acoustic data (Shelton et al., 2022), beach seines (Shelton et al., 2019), and angling catches (Ogonowski et al., 2023).

Generally there is agreement that in its present form, eDNA cannot replace some of the biological data collected from traditional survey methods (i.e., size, fecundity, body condition, etc.). Despite these limitations, there are numerous opportunities to expand eDNA collections and leverage eDNA data for fisheries research and management. eDNA collected from moorings and autonomous platforms can expand spatial and temporal sampling, particularly in hard-to-access environments, including the Arctic.

Collaborations Through a collaboration between PMEL and the NOAA AFSC, eDNA sampling is being performed throughout the Chukchi Sea to identify changes in distributions of arctic cod and walleye pollock. The AFSC Genetics Program has also sampled eDNA alongside bottom trawl surveys, the BASIS surface trawl ecosystem survey in the northern and southeastern Bering Sea, Marine Mammal Lab (MML) northern fur seal and Steller sea lion diet studies in the Aleutian Islands, and ice seal surveys along the ice edge during spring break-up.

Specifically, in collaboration with AFSC’s NBS surface trawl survey, the Genetics Program is using paired trawls/eDNA to analyze concordance between the fish communities identified by each method. This research will be used to compare species distribution models from eDNA and trawls and provide insight into future applications for eDNA alongside surveys.

Contributed by
Diana Baetscher and Kimberly Ledger
Auke Bay Laboratories, Alaska Fisheries Science Center, NOAA Fisheries

Alaska Department of Fish & Game Nearshore Survey

In September 2024, the Alaska Department of Fish & Game surveyed nearshore marine waters of the southeastern Bering Sea (SEBS). The aim of this survey is to study the early marine life stage, known as the juvenile stage, of salmon originating from southwestern Alaska systems (primarily Kuskokwim River and Bristol Bay). The juvenile life stage (i.e., the first summer in the ocean) is a critical period within the salmon life cycle. Juvenile salmon surveys provide valuable insights on the early marine ecology (e.g., abundance, distribution, size, diet, and condition) of juvenile salmon, which are needed to understand which parts of the salmon life cycle are most critical for survival. This project used trawl gear fished at the surface to sample the shallow juvenile salmon habitat on the SEBS shelf. The goals of this project are to estimate the juvenile abundance of SEBS stocks of salmon and to evaluate their life history and health characteristics, such as size at capture, diet, and energetic status.

Survey operations were conducted onboard a chartered commercial fishing vessel between August 25 and September 22 at stations spread across the SEBS shelf between Nunivak Island and Bristol Bay. At each of 55 stations successfully sampled in 2024, a Conductivity-Temperature-Depth instrument was deployed to measure oceanographic characteristics of the water column, like salinity and temperature; a Calvet net (vertical tow) and bongo net array (oblique tow) were deployed to assess the distribution and abundance of various zooplankton species; and a surface trawl net (Nordic 264, NET Systems, Bainbridge, WA, USA) was towed for one hour to collect epipelagic species. The average vertical and horizontal dimensions of the net were 15.6 m and 21.5 m, respectively. Following each tow, juvenile salmon were counted, measured for length and weight, and sampled for various biological analyses (e.g., genetics, diet, and energetic density). Other pelagic species caught in the trawl were also enumerated, measured, and retained for various analyses.

Juvenile salmon were captured at all but two stations in the survey grid. Chum salmon (Oncorhynchus keta) were the predominant salmon species captured (n=1,302) followed by sockeye salmon (O. nerka; n=471), Chinook salmon (O. tshawytscha; n=310), coho salmon (O. kisutch; n=215), and pink salmon (O. gorbuscha; n=95). Juvenile pink salmon catches were expected to be relatively low given that pink salmon runs in western Alaska are predominantly even-year dominant and thus juveniles would be expected to be present in higher abundances in odd years. Juvenile Chinook and coho salmon catches were higher in nearshore stations, especially around Cape Newenham. Juvenile chum salmon catches were spread across the survey grid but absent in the furthest offshore stations. Juvenile pink salmon were concentrated in the western portion of the grid while sockeye salmon were concentrated in the middle of the grid (Figure 6).

Catches of non-salmonid fish species were variable (Figure 7). Capelin (Mallotus villosus; all age-0 individuals) were only found at four stations throughout the survey (n=35). Pacific cod (Gadus macro-cephalus; predominantly age-0) were present at the most offshore stations (n=527). Pacific herring (Clupea pallasii; predominantly age-1+) were caught in relatively low numbers at both nearshore and offshore stations (n=324). Pacific sand lance (Ammodytes hexapterus; various ages) were primarily caught at the nearshore most stations (n=2,255). Walleye pollock (Gadus chalcogrammus; predom-inantly age-0) were the most numerous and commonly encountered non-salmonid fish species in the survey grid (n=66,087). In total, 23 fish species were caught during survey operations.

image

Figure 6: Juvenile salmon (Oncorhynchus spp.) catch-per-unit-effort (number/hour) at 55 stations sampled during the ADF&G nearshore southeastern Bering Sea salmon survey. Solid line on the map denotes the 50 m isobath.

image

Figure 7: Juvenile salmon (Oncorhynchus spp.) catch-per-unit-effort (number/hour) at 55 stations sampled during the ADF&G nearshore southeastern Bering Sea salmon survey. Solid line on the map denotes the 50 m isobath.

This survey is intended to be the start of a long-term project to assess the early marine ecology of southwestern Alaska stocks, understand factors that influence population dynamics, and make progress towards the long-term goal to develop a forecasting tool for Kuskokwim River salmon. Funding has been secured to continue survey operations into 2027. Zooplankton abundance and distribution, juvenile salmon genetic stock compositions, and juvenile salmon diet and energetic density results should be available in Spring 2025.

Contributed by
Sabrina Garcia, Dr. Kathrine Howard, and Benjamin Gray
Alaska Department of Fish & Game
Division of Commercial Fisheries

A Borealization Index for the Southeastern Bering Sea

Borealization – the transition from an Arctic physical state supporting a cold-adapted species assemblage to a subarctic (boreal) physical state supporting a warm-adapted assemblage – is one of the most consequential impacts of climate change for the Bering Sea and Arctic marine ecosystems generally. While “borealization” sometimes references only changes in species composition, we use the term to refer to the broad reorganization of Arctic ecosystems to a more temperate physical and biological state.

Here, we present an index of borealization for the southeastern Bering Sea. This index was developed to help understand, attribute, and project the causes of the snow crab collapse in 2019–2021, and the index outperforms bottom temperature as a predictor of snow crab abundance. Because the physical and ecological changes associated with borealization involve nearly every component of the ecosystem, the index may be useful for summarizing climatic and ecological changes to a wide range of species of management interest beyond snow crab.

To measure the progression of borealization, we analyzed nine time series that reflect the difference between Arctic and boreal conditions in the southeastern Bering Sea between 1972 and 2024, cov-ering changes in ice cover, bottom temperature, primary production, and community composition for phytoplankton, zooplankton, and groundfish. To create an overall index of borealization from the indi-vidual time series we used Dynamic Factor Analysis (DFA), a state-space approach for identifying shared variability across multiple time series. The DFA model identified a single shared trend that combines negative loadings for time series associated with Arctic conditions, and positive loadings for time series associated with boreal conditions (Figure 8a). The DFA trend provides a clear index of borealization with transition from the most Arctic-like conditions in the 1970s and 2007–2013 to the most boreal con-ditions during the warm, low-ice years of 2018–2019 (Figure 8b). The borealization index has reverted to values similar to the time series mean during 2022–2024.

Contributed by
Mike Litzow1, Erin Fedewa1, David Kimmel2, Jens Nielsen2,3, and Emily Ryznar1
1NOAA Alaska Fisheries Science Center, Kodiak, AK
2NOAA Alaska Fisheries Science Center, Seattle, WA
3Cooperative Institute for Climate, Ocean, and Ecosystem Studies,
University of Washington, Seattle, WA

image

Figure 8: A borealization index for the southeastern Bering Sea. a) Loadings for nine time series in Dynamic Factor Analysis (DFA) model. Negative (positive) loadings indicate time series associated with Arctic (boreal) conditions. b) The borealization index, defined as the shared trend from the DFA model. Negative (positive) values indicate more Arctic (boreal) conditions in the southeastern Bering Sea. Error bars and ribbon indicate 95% confidence interval. For information on methods for individual time series and their relationship with borealization, see Litzow et al. (2024).