Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Open Access

Peer-reviewed

Research Article

Spatial variation in food web structure in a recovering marine ecosystem

  • Kyle J. Krumsick ,

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Software, Visualization, Writing – original draft

    Kyle.Krumsick@mail.concordia.ca

    Affiliation Centre for Fisheries Ecosystems Research, Fisheries and Marine Institute of Memorial University of Newfoundland, St. John’s, Newfoundland, Canada

    ⨯
  • Jonathan A. D. Fisher

    Roles Conceptualization, Funding acquisition, Investigation, Project administration, Supervision, Writing – review & editing

    Affiliation Centre for Fisheries Ecosystems Research, Fisheries and Marine Institute of Memorial University of Newfoundland, St. John’s, Newfoundland, Canada

    ⨯

Abstract

Spatial heterogeneity in food web structure and interactions may reconcile spatial variation in population and community dynamics in large marine ecosystems. In order to assess food web contributions to the different community recovery dynamics along the Newfoundland and Labrador shelf ecosystem, we quantified species interactions using stable isotope mixing models and food web metrics within three sub-regions. Representative samples of each species caught in trawls and plankton tows were analyzed for stomach contents and stable isotope ratios (δ15N and δ13C) to parameterize isotope mixing models. Regional variation, highlighted by the diets of three economically important species, was observed such that the southern region demonstrated a variety of trophic pathways of nutrient flow into the higher food web while the diets of fish in the northern regions were typically dominated by one or two pathways via dominant prey species, specifically shrimp (Pandalus sp.) and hyperiids. Food web metrics indicated that the low-diversity northern regions had higher connectance and shorter food chain lengths. This observed regional variation contributes to our understanding of the role of specific forage species to the ecosystem which is an essential contribution towards ecosystem-based management decisions.

Citation: Krumsick KJ, Fisher JAD (2022) Spatial variation in food web structure in a recovering marine ecosystem. PLoS ONE 17(5): e0268440. https://doi.org/10.1371/journal.pone.0268440

Editor: Vitor Hugo Rodrigues Paiva, MARE – Marine and Environmental Sciences Centre, PORTUGAL

Received: September 30, 2021; Accepted: April 30, 2022; Published: May 20, 2022

Copyright: © 2022 Krumsick, Fisher. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.

Data Availability: All relevant data are within the paper and its Supporting information files.

Funding: KK: Research and Development Corporation of Newfoundland and Labrador (RDC, now part of InnovateNL) Ocean Industries Student Research Award (Grant #5201.1006.104). Their website may be accessed at https://www.tcii.gov.nl.ca/innovation/innovateNL/index.html. JF: Natural Sciences and Engineering Research Council of Canada (NSERC) Discovery Grant (Grant # 419957-2012). Their website may be accessed at: http://www.nserc-crsng.gc.ca/index_eng.asp. JF: Canada First Research Excellence Fund (20181014). The website can be accessed at: https://www.cfref-apogee.gc.ca/home-accueil-eng.aspx The funders had no role in the study design, data collection and analysis, decision to publish, or preparation of the manuscript.

Competing interests: The authors have declared that no competing interests exist.

Introduction

Quantifying spatial variation in ecosystem processes can be key to understanding the functioning and dynamics within large marine ecosystems [1, 2]. The collapse of demersal fish stocks off Newfoundland and Labrador (northwest Atlantic Ocean) brought about ecosystem changes by reducing the influence of dominant top predators [3]. One impact of this change was interactions among species through predation and indirect effects [4]. Now, more than two decades after stock collapses and initial fisheries moratoria, the ecosystem has not fully recovered [5, 6]. Factors hypothesized to influence population trajectories include continued fishing [7], depensatory growth [8], climate change [9], and life history changes [10]. However, Newfoundland and Labrador marine ecosystems also exhibit spatial heterogeneity in their structure and function with northern communities displaying lower fish species diversity and diet diversity [6, 11], and lower recovery of marine fish community size-structure [6]. In an ecosystem context, it is important to quantify the extent to which spatial variability in diets on relevant timescales may also contribute to differing population and community trajectories.

Historically the Newfoundland and Labrador region had been managed focusing on single species, contributing to the collapse of groundfish stocks in the regions in the early 1990s [12–15]. It has since become apparent that managing a stock in isolation is not sufficient and that managing the ecosystem with complex interactions with multiple species would help to facilitate future sustainable fishing practices [16–19]. Given interest in the application of ecosystem-based analyses to fisheries management, understanding food-web interactions and their contribution to species and community dynamics has become essential [16, 20, 21]. These trophic interactions represent a key factor which regulates fish populations [22–25]. Management of stocks requires consideration of trophic interactions which may confound indicators used to assess effectiveness of management measures and can influence management decisions when multiple exploited species exist, either through assessing the trade-offs of exploiting each species in the case of direct trophic interactions or determining the ecosystem-level impacts of fishing pressures [26–28]. Quantifying trophic interactions has relied on various methodologies and tools including stomach content analysis [29–31], fatty acid analyses [31–34], pyrosequencing of prey DNA from stomach contents or faeces [35–37], and stable isotope analyses [38].

Stable isotope analyses constitutes the primary technique that has been used to supplement stomach contents analyses for assessing energy flow through an ecosystem over longer periods [39, 40]. Stable isotopes in ecology provide information on the origins and consequent assimilation of organic matter which provides insights into food web [41–44]. Together these two isotope values can be used to determine isotopic niches which are thought to be proxies of trophic niches [45–47]. Importantly, with additional inputs of prey isotopic values, fractionation coefficient estimates, and prey elemental composition, the approximate proportions of the predator diet can be estimated using isotope mixing models [48, 49]. The isotope mixing model used herein is designed to quantify the relative contributions of the pelagic/benthic food chains, determine the relative trophic position of members of the community and to quantify the trophic relationships between organisms. The results of this isotope mixing model will allow for reconstruction of the marine food web to allow for assessment of the direct and indirect interactions between organisms as well as the important species for the passage of energy through the food web.

Three of the dominant, economically important groundfish species found along the Newfoundland and Labrador shelves are Atlantic cod (Gadus morhua), redfish (Sebastes sp.) and Greenland halibut (Reinhardtius hippoglossoides). Between 2013 and 2017 these three species represented 78.1% of the landed value of groundfish species in the Newfoundland-Labrador region [50]. Given their relative importance to the fisheries in this region, several historical diet studies have been conducted using stomach contents analyses. The demersal Atlantic cod are thought to be primarily generalist feeders but do show preferences towards high-lipid forage fish such as capelin [e.g. 11, 51–53]. The largely pelagic redfish has been reported to feed primarily on pelagic invertebrates such as hyperiid amphipods, copepods, euphausiids, and northern shrimp [e.g. 54–57]. The Greenland halibut is primarily a demersal fish but is thought to be an active mid-water predator. As such, their diet consists of a wide variety of pelagic and demersal prey, particularly capelin, shrimp, squid, and zooplankton [e.g. 58–61]. All of these diets have been reported to have changed along with the large-scale ecosystem changes that were observed in this region in the early 1990s [57, 61, 62].

Given the utility of stable isotope analyses to contribute new information on the structure of food webs and trophic dynamics among ecosystems, our objectives are threefold: (a) to construct simplified marine food webs based on results from stable isotope mixing models supplemented by information from stomach contents data for three regions within the northeast coast of Newfoundland and Labrador and then compare food web metrics among regions; (b) to analyze in greater detail spatial variation in the diets of the three abundant and economically important species: Atlantic cod, redfish and Greenland halibut, and; (c) to discuss the isotope mixing model results in the context of recently reported spatial variation in marine fish ontogenetic niche overlap and size-spectrum recovery among regions. Together these results contribute new information towards ecosystem-based decisions by quantifying the relative importance of specific forage species to these communities.

Materials and methods

Study area

Sampling was undertaken within marine research surveys conducted by the Center for Fisheries Ecosystems Research (CFER) aboard the RV Celtic Explorer in May 2015 on the offshore shelves from southern Labrador and eastern Newfoundland, corresponding to Northwest Atlantic Fishery Organization (NAFO) subdivisions 2J, 3K, and 3L (Fig 1). These subdivisions together represent the management unit for the ‘Northern’ Atlantic cod stock [63–65]. Three major channels or corridors within this region had previously been identified as important onshore-offshore cod migration pathways: Hawke Channel, Notre Dame Channel, and the Bonavista Corridor [66]. The following analyses were conducted within each channel separately in order to quantify spatial variation.

thumbnail
Download:
Fig 1. Map of Newfoundland and southern Labrador with sampling locations.

The trawl locations are indicated with open blue symbols and plankton tow locations with closed red symbols for the Hawke Channel (HC, triangles), Notre Dame Channel (NDC, circles), and Bonavista Corridor (BC, squares). The inset map outlines the study domain in eastern Canada. The relevant NAFO subdivisions 2J, 3K, and 3L boundaries are also indicated. Dashed lines represent 300 m depth contours. The bathymetry map is reproduced from GEBCO world map 2020 (www.gebco.net) and NAFO subdivisions reproduced from NAFO (www.nafo.int).

https://doi.org/10.1371/journal.pone.0268440.g001

Sample collection

Animal research was provided written consent by the Animal Care Committee of Memorial University of Newfoundland and Labrador. Field research and sample collection was conducted under a Fisheries and Ocean Canada experimental license (NL-2927-15). The collection of samples used for this study was reviewed and approved by the Memorial University Animal Care Committee under the guidelines set by the Canadian Council of Animal Care. In order to sample zooplankton, bongo nets (33 μm mesh and 60 cm diameter) were deployed obliquely within the surface layer for 10 minutes and at a speed of 2 knots. Fish and invertebrate samples were collected using a combination of Campelen 1800 and mid-water trawls deployed at depths ranging from 244 m to 471 m [47]. All trawl sets were deployed between the hours of 7 am and 10 pm. All fish caught in the trawl sets were sorted by species and standard lengths recorded. In cases where a species was particularly abundant, subsampling for length was initiated involving 100 randomly sampled individuals. Individuals were selected for dissection with efforts to provide representation from the observed size ranges for each species. For the majority of species, only nine specimens were collected within each region with the exception of species protected by the Species at Risk Act (Table 1). From most fish, a transverse sample of dorsal muscle tissue directly posterior to the head was collected at-sea, placed in a 1.5 ml centrifuge vial and frozen at -20 ºC for stable isotope analyses. Stomachs were also collected at sea and preserved frozen. Stomachs from fish that showed signs of regurgitation or stomach eversion were noted but not collected due to the potential of biasing stomach content results. The remaining fish with small, difficult to sample stomachs were individually labelled, bagged whole, and preserved frozen for dissection in the laboratory for their muscle tissue and stomachs.

thumbnail
Download:
Table 1. Species included in this study, their species abbreviation, and the group category to which they were assigned.

https://doi.org/10.1371/journal.pone.0268440.t001

While more species were caught and analyzed for their isotopic signatures (Table 1), nine species represented the most abundant fish species: American plaice (Hippoglossoides platessoides), Atlantic cod (Gadus morhua), Atlantic herring (Clupea harengus harengus), capelin (Mallotus villosus), checker eelpout (Lycodes vahlii), Greenland halibut (Reinhardtius hippoglossoides), lanternfish (Notoscopelus sp.), redfish (Sebastes sp.), and thorny skate (Amblyraja radiata). Based on the species’ observed length distributions, sampled individuals were classified as small, medium, or large, by dividing the observed range of sizes into three length categories of equal width (Table 2). Given the more extensive analyses of these species, sampling efforts were increased to include seven individuals within each size category. These categories are recognized to be arbitrary, but as the exact timing of potential ontogenetic shifts was unknown, this division accounted for variation across the range of observed sizes. Of these, Atlantic cod, Greenland halibut, and redfish were selected to portray regional variability given their abundances and economic importance.

thumbnail
Download:
Table 2. Definition of small, medium and large size categories for the most abundant nine fish species.

Size category definitions were consistent across regions.

https://doi.org/10.1371/journal.pone.0268440.t002

In addition to the fish samples, a variety of invertebrates were collected (Table 1). Invertebrates were also sorted by species and measurements such as carapace width in crabs and carapace length in shrimp were obtained for up to 100 randomly sampled individuals. A sample of up to twenty-one of each invertebrate per region were frozen whole, except for large snow crab, each of which was sampled by removing one leg. Zooplankton were collected from each plankton tow passed through a 140-micron sieve and were preserved frozen for further taxonomical identification in the laboratory.

Stomach content analysis

Stomach samples were analyzed using a dissecting microscope and contents were identified to the lowest feasible taxonomic level. Individual weights and numbers of each prey taxa were quantified. From these measurements, in combination with their frequency of occurrence, the index of relative importance (IRI) was calculated for each prey taxa as follows: (1)

Where %N is the percent contribution of a given taxon to stomach content by numbers, %B is its percent contribution by weight, and FO is the frequency of occurrence, defined as the number of stomachs in which the prey taxon was detected over the total number of stomachs [67]. In combining these three measures into a single index, it should free the analysis from any biases associated with each of the measures independently. The percent IRI is presented as a percentage of the summation of the IRIs of all prey observed and were calculated for each predator species and region combination.

Stable isotope analysis

Muscle tissue samples were oven dried at 75°C for 48 hours and homogenized using an amalgamator. The homogenized samples were weighed, packaged in an airtight container with desiccant packages and shipped to the Cornell University Stable Isotope Laboratory (Ithaca, NY, USA) for analysis. Approximately 1 mg of sample was placed into 7×7 mm tin capsules, then flash combusted using a Carlo-Erba NC2500 elemental analyzer coupled on-line to a Finnigan MAT Delta Plus mass spectrometer for analyses of the resulting carbon dioxide and nitrogen gases.

The stable nitrogen isotope signature (δ15N) typically becomes enriched by approximately 3 ‰ for fish species with each consumption, allowing for approximation of trophic level [68–70]. The stable carbon isotope signature (δ13C) provides an indication of the initial carbon source (pelagic or benthic/detrital in origin) and enriches at typically less than 1 ‰ with fractionation frequently considered negligible [71–73]. Nitrogen and carbon ratios were expressed in delta (δ) notation, being the parts per thousand deviation from the standard material, Pee Dee belemnite limestone for carbon and atmospheric nitrogen for nitrogen, as follows: (2) (3)

Lipids were not removed to avoid the potential influence of derived products on isotopic signatures [74–76]. Therefore, following analysis, the δ13 C values were normalized for lipid bias as recommended by Ricklefs & Travis [77] and Post et al. [78], as follows: (4)

As the majority of fish samples possessed a carbon to nitrogen concentration ratio close to 3.3, as would be expected for muscle tissue of most marine fish [77], minimal modification between the normalized and untreated carbon values was observed for the majority of analyzed fish. This adjustment was only particularly relevant for lipid rich fish, such as capelin, lanternfish and Greenland halibut, where the carbon concentration becomes increased relative to the nitrogen.

Stable isotope mixing model

Although stomach contents provide detailed information on the diet composition of species, it represents the diet on the scale of hours to days [79–81]. In order to assess longer-term prey consumption and food web connections on the scale of months, we analyzed the diets of every caught predatory fish species (a total of 25 species) using Bayesian concentration-weighted stable isotope mixing models. Eight of these 25 species were also found to represent a at least three size categories as defined by Table 2, for a total of 41 separate models. These models approximate the contributions of different prey while assessing their locations in isotope space relative to adjusted predator values. These techniques are partly informed by priors, in our case stomach contents data. These priors were separated by region where possible, but in cases where a given region was represented by three or fewer stomachs combined stomach data for all regions was input as the prior. Species or functional groups were further divided into four categories based on descriptions provided by Sherwood & Rose [82] and Scott & Scott [83] (Table 1): benthic invertebrates, pelagic invertebrates, demersal fish, and pelagic fish. Some species, such as copepod species, were not found to differ in their isotopic signature and therefore were pooled together as a single functional group.

A fractionation coefficient, or discrimination factor, is the change in the isotopic signature from prey to predator that occurs due to partitioning upon consumption and assimilation of respective elements. Determining the exact fractionation coefficients between the predator and each individual food source is often recommended for each element [84]. Such determination of prey-specific fractionation coefficients, however, was not possible in the present study. Historically a nitrogen fractionation of 3.4 has been used in ecological studies of fish populations and 0 for carbon, as if fractionation was assumed to be negligible [69, 71, 85]. However, these carbon estimates may have been underestimated and nitrogen overestimated [86]. As a result, two approaches were taken to estimate these discrimination factors. The coefficient was first estimated from the combination of stomach content analysis and the associated isotopic values of prey per Sherwood and Rose [82]. The bounds of values were 1.4–4.4 for nitrogen and -0.5–2 for carbon, as determined by biologically feasible fractionation coefficients [38, 73, 84, 86–90]. For estimated coefficients outside these bounds, the fractionation was estimated to be 3.4 for nitrogen and 0.4 for carbon [38].

Potential prey items to input into the mixing model were selected through a combination of the results of these stomach contents and published reports of North Atlantic diets (S1 Table). For species with particularly diverse diets, such as Atlantic cod and thorny skate, prey that represented over 5% of the weight and/or numbers were analyzed separately and all other reported prey items were combined into four functional groups (Table 1): pelagic invertebrates, benthic invertebrates, pelagic fish and demersal fish. To acknowledge that any predator is gape-limited, only prey that were less than 24% of the predator’s body mass were included in the model [91]. As not all individual weights were obtained, length-weight relations were used to determine approximate body masses for both predators and potential prey (S2 Table). For individual invertebrates which did not have size information, average species sizes were used.

The stable isotope mixing model to determine the percentage of the diets represented by key prey species was implemented using the MixSIAR package in R [92]. Three basic equations were utilized in the Bayesian isotope mixing model to determine the proportions of the diet occupied by each prey type [93]: (5) (6) (7)

Where δ13Cn/δ15Nn represent the tissue isotopic values for a given prey item, δ13CM/δ15NM the tissue isotopic values for the predator, [C]n the carbon concentration of a given prey, [N]n the nitrogen concentration of a given prey, and fn,B the proportion of the predator’s diet represented by the given prey species. Size category (small, medium, or large) was included as a fixed variable for species-region combinations that demonstrated ontogenetic variation. Region- and species-specific IRIs calculated within this study were provided as a prior to these mixing models [94]. Some species-region combinations did not contain stomach data due to a high percentage of empty stomachs. In these few cases, an average IRI from other available regions was used. For cases where there are three or fewer potential prey sources, this model can provide exact contributions to the predator’s diet of each prey. For greater than three potential prey sources, fifty thousand repetitions of the mixing model were run to determine the approximate proportion of the diet each prey taxon occupies.

Food web metrics

Several food web metrics were calculated to characterize the three constructed food webs. The total number of “nodes” or functional groups represents the network size. The “connectance” of the food web is the fraction of all possible predatory links that are realized and the ratio of trophic links within the food web over the square of the network size [92, 93]. For each species, several metrics were calculated to determine the importance of each prey. The “number of links per node” represents the number of predators feeding on that prey. The “average percent of predator diets” is the mean percentage of a given prey within the linked predator diets. The “relative link strength” is calculated as the sum of all the diet proportions contributed by a given prey taxon over the total summation of diet proportions for all food web links within a given region.

From the outcomes of the isotope mixing models and these food web metrics, we constructed simplified food webs for the predatory fish species. Within isotope biplot space, all the species were plotted with species showing ontogenetic variability in their isotopic signatures represented as a maximum of three points. Links between predatory species and their respective prey, identified through our stomach content analyses and reported predation in the literature, were plotted with the line width proportional to the importance to the predator’s diet. The proportion of the total linkage strength for each prey item as determined by summing all the linkage strengths from the prey item and dividing by the sum of all linkage strengths determined within the food web. Circles were drawn over the respective prey items in biplot space in the reconstructed upper foodweb with radiuses proportional to the magnitude of these total linkage strengths to indicate the relative importance of prey species as conduits of energy flow into the upper food web.

Results

Index of relative importance

Stomach content data from all fish species were collected to use as priors for the stable isotope mixing model. From the stomach contents alone, half of the percent IRIs of the three focus fish were made up of shrimp, snow crab, hyperiids, or capelin (Table 3, S3 Table). Shrimp consumption increased in the northern regions, representing an average of 19.2% IRI in the Bonavista Corridor and 33.4% in the Hawke Channel. Gammarids were also a common prey item of Hawke Channel fish, representing an average of 13.3% IRI. Pelagic invertebrates, particularly hyperiids, were found to be a dominant prey item in the Notre Dame Channel representing 58.6% IRI, particularly hyperiids and copepods representing an average 26.0% and 31.5% IRI. The Bonavista Corridor was noted to have the highest incidence of crab (8.5% IRI), polychaetes (16.7%), and pelagic fish (particularly capelin; 4.3%).

thumbnail
Download:
Table 3. Diet compositions of three highlighted predatory fish species presented as percent index of relative importance (IRI) within three regions: Hawke Channel (HC), Notre Dame Channel (NDC), or Bonavista Corridor (BC).

https://doi.org/10.1371/journal.pone.0268440.t003

Simplified food web model

The mean and standard error of the δ13C and δ15N for each species relative to other members of the community are shown in Fig 2. Ellipses enclosing benthic invertebrates, demersal fish, pelagic invertebrates, and pelagic fish, representing bulk prey categories for species with particularly varied diets, were estimated using the Khachiyan algorithm for the computation of minimum volume enclosing ellipsoids [95]. The most nitrogen depleted values were observed in the pelagic algae and benthic plant material (Fig 2) and the highest trophic level species was the Atlantic cod. While the pelagic components of the ecosystem fell within a relatively small range of δ13C values, the benthic community was found to exhibit a wide spread of δ13C values (Fig 2). Within the benthic component of the food web invertebrates, particularly echinoderms, exhibited a larger range of carbon values than the fish species.

thumbnail
Download:
Fig 2. Stable carbon and nitrogen isotope values (mean ±SE) for dominant species within (A) the Hawke Channel, (B) the Notre Dame Channel, and (C) the Bonavista Corridor.

Ovals represent the group categories of pelagic fish (green), benthic fish (red), pelagic invertebrates (blue), benthic invertebrates (yellow), pelagic algae (purple), and benthic plants (dark green). In cases where a species demonstrated ontogenetic variation in their isotopic signatures, these species are represented by three points connected by linear lines. Species abbreviations are defined in Table 1.

https://doi.org/10.1371/journal.pone.0268440.g002

The food web for the Newfoundland and Labrador shelf regions, even when simplified, displayed numerous links (Fig 3). Each region consisted of between 31 and 43 observed trophic “nodes” with the Notre Dame Channel (Fig 3B) having the fewest nodes, likely lower due to less intensive sampling (Table 4). However, the highest latitude system had an ~10% increase in the number of links per prey species (Table 5), which is associated with increased connectance. Strong latitudinal increases were observed in the linkage strength of shrimp, more than doubling from 0.105 in the south to 0.264 in the north. Similarly, PPD increased from 21.2 to 38.1. In contrast, prey such as copepods (LS increasing from 0.135 to 0.216, PPD from 14.8 to 24), brittle stars (LS increasing from 0.021 to 0.050, PPD from 4.8 to 12) and fish species were more frequently consumed in the southern regions (Fig 3, Table 5). The Notre Dame Channel also saw increases in prey items from the center of the biplot space such as hyperiids (LS of 0.297, PPD of 37.3) and bivalves (LS of 0.098, PPD of 27.7) (Fig 3B).

thumbnail
Download:
Fig 3. Simplified food webs with line widths indicating the linkage strength of the predatory interaction for (A) the Hawke Channel, (B) the Notre Dame Channel and (C) the Bonavista Corridor.

Circles represent the relative linkage strength from the given prey species. Species from which a number of size classes were obtained will appear up to three times (for small, medium, and large size categories). Taxa abbreviations are defined in Table 1.

https://doi.org/10.1371/journal.pone.0268440.g003

thumbnail
Download:
Table 4. Three food web metrics calculated for the food webs in each region.

https://doi.org/10.1371/journal.pone.0268440.t004

thumbnail
Download:
Table 5. Food web link characteristics for the three focus regions.

https://doi.org/10.1371/journal.pone.0268440.t005

Trends among three focus species

While a variety of predatory fishes were analyzed as part of this study, we chose to focus on three abundant and economically important species. For details on additional abundant species, refer to S1–S6 Figs.

Within Atlantic cod diets, in the BC, 75% consisted of shrimp, snow crab, polychaetes, hyperiids and benthic invertebrates; in the NDC: shrimp and hyperiids; in HC: shrimp, snow crab, euphausiids, and pelagic invertebrates (Fig 4). Across all three regions shrimp made up a quite substantial portion of cod diets increasing in the northern regions (Table 1). With increasing cod size, the contribution of zooplankton to the diet decreases and is replaced with benthic prey. Furthermore, snow crab, another economically important species in the region, was observed to make up quite a substantial portion of the diet in the Bonavista Corridor and dominate cod diets within the Hawke Channel (Fig 4). Demersal and pelagic fish make up a small portion of the diet in the Bonavista Corridor that was found to decrease in the other two regions (Fig 4).

thumbnail
Download:
Fig 4. Atlantic cod diet composition percentages as determined from isotope mixing models.

This figure is divided by predator size category (S for small, M for medium, L for large, see Table 2) and region (BC for Bonavista Corridor, NDC for Notre Dame Channel, HC for Hawke Channel). Prey taxa abbreviations are defined in Table 1.

https://doi.org/10.1371/journal.pone.0268440.g004

In the Bonavista Corridor the diet of this fish was comprised mostly of copepods with generous contributions from hyperiids and capelin (Fig 5). The Notre Dame Channel saw a shift from a copepod-dominated diet to one comprised primarily of hyperiids. Finally, shrimp dominated the diets within Hawke Channel. With increasing Greenland halibut size, there was a decrease in zooplankton consumption (Fig 5).

thumbnail
Download:
Fig 5. Greenland halibut diet composition percentages as determined from isotope mixing models.

This figure is divided by predator size category (S for small, M for medium, L for large, see Table 2) and region (BC for Bonavista Corridor, NDC for Notre Dame Channel, HC for Hawke Channel). Prey taxa abbreviations are defined in Table 1.

https://doi.org/10.1371/journal.pone.0268440.g005

Redfish diets were relatively less complex in comparison to the previous two predators, consisting of mostly zooplankton and shrimp (Fig 6). The relative proportions of these two contributions varied among regions with higher contribution of shrimp with increasing latitude. Increases in shrimp consumption were also observed with ontogeny within all regions.

thumbnail
Download:
Fig 6. Redfish diet composition percentages as determined from isotope mixing models.

This figure is divided by predator size category (S for small, M for medium, L for large, see Table 2) and region (BC for Bonavista Corridor, NDC for Notre Dame Channel, HC for Hawke Channel). Prey taxa abbreviations are defined in Table 1.

https://doi.org/10.1371/journal.pone.0268440.g006

Discussion

We report the results of within- and among-region variation in the Newfoundland and Labrador shelf marine ecosystem analyzed using isotope mixing models to examine whether regional variation in community overlap [47] and differential size-based community recovery rates [6] may be explained by differences in energy and nutrient flows through ecosystems recovering from overexploitation [5]. While every food web is inherently a simplification of reality, based on the results of both the stomach content and the stable isotope analyses, the importance of specific forage species changes markedly among the three analyzed regions. These findings have implications for management of both established invertebrate fisheries and recovering groundfish fisheries and further reveal the importance of spatial downscaling species interactions within large marine ecosystems.

Stable isotope ecology has proven to be an effective means of assessing energy flows as it directly attempts to track the path of rarer heavier atoms through the food web [39, 96, 97]. In addition to helping to identify and quantify key trophic interactions within the ecosystem, particularly robust to spatial variation in the importance of shrimp among regions, stable isotopes also provided insights into the underlying processes that are governing these ecosystems by resolving the relative importance of the pelagic and benthic components of the community. To demonstrate the importance of considering this regional variability in species interactions, we separated the food-webs by regions defined by Rose [66] and highlighted the regional variation in reconstructed diets observed in three economically important groundfish species. Although partitioning the entire study area into smaller sub-regions may initially seem unnecessary, if spatial similarity was observed, the three areas could simply be recombined [2]. However, we observed regional variation in both the stable isotope model outcomes and in the composition of the catches and the stomach contents. These findings and consistent, biogeographic differences in food web metrics along this latitudinal gradient parallels results from spatial partitioning of the Barents Sea ecosystem, where species interactions, food web metrics, and environmental conditions varied among four main food-web regions [2].

While observed extremes in the carbon values of invertebrates indicated either pelagic or benthic signatures, the fish species tended towards a central carbon isotope value. This pattern tends to became more pronounced with increasing fish size, suggesting a balance between the consumption of pelagic and benthic productivity. The balanced acquisition of pelagic and demersal productivity has recently been hypothesized to explain the dominance of large benthic demersal fishes over large pelagic fishes in boreal and temperate regions [98], while other comparative studies in this region have focused largely on the dynamics of single prey species [62, 99]. Whether this energy acquisition hypothesis also influences ecosystem recovery is an important question with potential implications for differential ecosystem recovery pathways. For example, recoveries of community size-structure are strongest in the Bonavista Corridor compared to the other two regions [6]. Given both this hypothesis and our reported patterns, further examinations of benthic-pelagic prey contributions to predator diets should be included in analyses of differential recovery within and among marine ecosystems.

To further explain benthic and pelagic differences in prey consumption among regions, each region was characterized by a few key prey species as revealed by the link strength between predators and prey in the food web models. The fish diets in the Bonavista Corridor are known to be the most diverse of these three regions from previous published research [11, 47]. This study further demonstrated that groundfish in this region show increased consumption of a variety of fish species (many of them pelagic forage fish), copepods, and brittle stars. Apart from the increased copepod and brittle star consumption, these observations result in overall higher trophic level of the top-level predators and facilitate strong contributions of both the pelagic and benthic portions of the food webs towards generalist feeders.

In contrast, the Notre Dame Channel was characterized by important prey species with mid-range carbon and nitrogen isotopic values: shrimp, bivalves and hyperiids. Relatively extreme carbon values were observed in this region among copepods and echinoderms. As periphyton develops, the δ13C has been observed to increase into the range which could explain the extreme carbon values observed in echinoderms that likely would at least partially predate on this periphyton [100–102]. Furthermore, spatial variation in these values could be explained with the composition and thickness of this periphyton. Enriched δ13C observed in echinoderms have also been explained as a result of metabolic processes and preferential selection of high δ13C prey such as diatoms [103–105]. Spatial variation in copepod carbon isotope ratios have been observed elsewhere yet have not been fully explained, though it has been proposed that they reflect long-term trends in primary productivity carbon combined with temporal metabolic variability and spatial variation in algal isotopic composition [106–108]. The stable isotope and the stomach content data, however, consistently demonstrated the importance of zooplankton and infauna to the diets of fish in this region.

Finally, the Hawke Channel food web was dominated by the consumption of shrimp, although polychaete consumption was also found to be relatively higher in this region even if it only comprised a small portion of the diet. The vast majority of species in this region were found to at least have shrimp in their stomach contents if not representing the majority of the contents, consistent with numerous other groundfish diet studies in this region [11, 109, 110]. These patterns and evident differences in food web metrics among regions reiterates the importance of spatial downscaling to quantify variation in species interactions within marine ecosystems [1, 2, 4, 5, 47].

Many of the known detritivore invertebrates have been noted to have relatively enriched carbon signatures. This finding is consistent with the results found in other studies of temperate and high-latitude systems [111, 112]. We were not able to obtain isotopic values for the detritus, but we may infer this food source would possess a more enriched nitrogen and carbon signature compared to the benthic plant material in order to account for the wide range of values in the benthic portion of the community [113–115].

Three abundant fish species of socioeconomic importance were analyzed in greater detail due to recent debates regarding their population status and interactions with other fisheries. The biomass of these predators is often greater in the southern region studies than the north [116–119] and represented the most abundant species caught in our bottom trawls [6, 47]. We observed that shrimp and/or crab often comprised a substantial portion of the diet with the proportion increasing with ontogeny (Figs 4–6).

Additionally, as northern regions show less diversity in diet and fish communities [47], shrimp and crab become increasingly important channels of nutrient flow through the food web. Shrimp and crab represent the two most economically dominant fisheries within the entire Newfoundland and Labrador region, with landed values of $196 million and $309 million in 2019, respectively (120These values greatly exceed the recent landed values of Atlantic cod ($21 million), redfish ($10 million), and Greenland halibut ($58 million) [120]. Thus, while shrimp and crab are presently the two most lucrative fisheries they also represent major prey for these groundfish species [e.g. 3, 121, 122], the impact of which depends on the respective species- and size-specific predator population sizes and regions.

Given these food-web links, demersal fish recovery would therefore likely negatively impact the present shellfish industries. Studies have shown that large aggregations of cod and shrimp don’t coexist in this ecosystem, with increases associated with decreases in the other [3, 4, 123–125]. This interaction is exacerbated by each species favoring differing environmental conditions such as ocean temperatures [4, 126–130]. Quantifying spatial variation in these interactions reverses the decades-old question of whether declines of Atlantic cod facilitated the expansion of shrimp fisheries [3, 10] to now reveal regions in which lucrative shrimp fisheries and associated industries and communities may be impacted most by demersal fish recovery [131].

Several food web metrics were presented in order to quantify these interactions. The network size was found to be similar in the Hawke Channel and the Bonavista Corridor, though lower in the Notre Dame Channel. We observed that between the Bonavista Corridor and the northern regions there was a distinct increase in the connectance. This increase in connectance is associated with decreased food chain lengths and high predator-prey mass ratios (PPMR) [132, 133]. The observation that fish consumption tends to be reduced in the northern regions while zooplankton and shrimp become a much more prominent part of the diet could explain these observations, and the recent observation that PPMR is much higher in the Hawke Channel than other regions [6]. The links per prey species were found to be greater in the northern regions. Given these regions show lower diversity, predators would be expected to be more likely to prey upon common species. This metric therefore confirms that food web complexity is reduced in the northern regions, driven primarily by observed decreases local species diversity in these regions [6, 11]. This food web metric, however, has been called into question as it has been noted to incorrectly characterize ecological trends when varying number of nodes are present between food webs [134]. Despite this caveat, we observe the largest difference between the two regions with comparable network size, indicating an underlying ecological process may be responsible for the observations. Numerous limitations exist when constructing isotope mixing models. Firstly, an assumption was made that all potential prey species are represented within the model. However, the survey gear types used inherently limited the species and the size ranges sampled in this study, and as such there are gaps that are not represented. An example of this is the absence of detritus in the analysis. As we did not have access to a grab sample, the detritus anticipated to make up a major component within the benthic food web needed to be left out. We could infer the approximate position within biplot space based on the positions of the detritivores but given the wide range of carbon and nitrogen isotope values among these detritivores we cannot conclusively state the exact range of detrital isotopic values. In regards to the energy input into the upper food web, however, the exclusion of detritus and periphyton is not anticipated to have a significant impact as it is unlikely that detritus makes up a significant portion of these predators’ diets. The absence of components such as detritus prevent us from constructing a complete food web, but should not inhibit our ability to analyze species interactions within the upper food web to understand key prey species within these communities.

Most taxa that were identified in the stomachs had representative isotopic values from the trawl data. While all the major taxa present within the stomachs (representing > 5% of the Atlantic cod stomach contents by weight), minor invertebrate taxa were sometimes not represented in the isotopic signature. The Bonavista Corridor was missing sea cucumbers (3.4% of cod stomach content biomass); Notre Dame Channel was missing gastropods, crabs (Cancer sp.), tunicates, bryozoans and sea urchins (all at less than 0.1% of cod and Greenland halibut stomach content biomass); and Hawke Channel was missing sipunculids (0.1% of American Plaice stomach content biomass) and bryozoans (0.01% of cod stomach content biomass). Given the limitations of trawl sampling of invertebrate communities, it is not unexpected to be missing taxa. The only fish species that was found in a stomach that was not observed in the trawl catches were grenadiers in the Notre Dame Channel (0.3% of stomach content biomass). With the exception of gastropods, all of the prior mentioned taxa would have been grouped in either “other benthic invertebrates” or “other demersal fish” categories. Since these underrepresented species occurred infrequently in limited quantities, it appears unlikely that their omission would impact the results of the isotope mixing model. We also chose not to sample these taxa from the stomachs themselves as digestion has been shown to influence isotopic compositions [135, 136].

The stomach contents were also used in combination with those from other published studies as a means of determining potential prey as inputs for the stable isotope mixing model. However, this method may introduce uncertainty regarding the detectability of different prey [137]. For example, prey types will have variable evacuation rates which could bias the results of stomach contents analysis [138] but combining this analysis with stable isotope analysis helps to resolve this bias. An alternative means of detection within the stomach would be to use stomach content DNA to determine prey diversity [139, 140]. However, such methods are also subject to errors in the detectability of prey due to DNA degradation and as such should be used in combination with other methods [141].

The isotope mixing models are also known to be highly sensitive to discrimination factors [142]. Yet the measurement of these factors can often be difficult for this kind of study. While extensive lab testing is often encouraged to determine these contributions [143], many factors will influence the exact value of these factors within a given tissue type, including temperature [86], feeding rate [86], isotopic values of the prey [84, 144], protein and fat content of the prey [88, 145] and approximate trophic position [146]. Considering this wide range of uncertainty, we initially attempted to estimate a discrimination factor based on our observed stomach content information and isotope values of the prey. However, this estimation sometimes produced unreasonable results likely due to the low sample sizes of stomachs for some species with a sometimes high percentage of empty stomachs. In these cases, we used the measure provided by Post [38] as an approximation with full knowledge that this estimate is likely a simplification of reality, yet robust enough to provide an interpretable outcome.

The number of potential prey sources were often quite high for many of the species analyzed. Previous work in this region identified over 100 different prey species within cod stomachs [11]. This number of potential prey sources is naturally beyond what the model can realistically handle. Group categories were created to account for minor contributions to the diets within stomach data in the present study and previously published studies. For species with particularly varied diets, this resulted in 14 prey categories as inputs into the model [97]. Although this situation is less than ideal, as over 7 prey categories are to be used with caution [92], such an approach would limit the analyses to our 4 group categories and three individual prey items for species with high diet variability. We therefore opted to increase the number of potential inputs at the tradeoff that the outputs are more likely to be confounded and the model less likely to converge. Even so, interpretable results and trends arose from the study with comparable results from repeated test runs.

Despite these uncertainties in the measures of the diet composition, the proportions represented by each prey item provide an indication of the relative importance of each prey item [97]. The isotope mixing models for more than 3 prey sources are based on probabilities of each prey item being selected and as such often have high variation as the resulting proportions are an average of 10,000 runs of the model. Therefore, instead of focusing on determining the exact proportions of each prey, we focused on determining key prey items for the food web as well as spatial variation in these trends. The use of stomach content information used as a prior helped to direct the measure of dietary proportions towards this end. However, these priors were not found to directly determine the outputs of the stable isotope mixing models but merely increased or decreased the likelihood of individual prey species being important to the predator. In several instances the priors presented to the model did not match the long-term isotopic signature of the predator. We interpreted these occurrences as the priors demonstrating short-term diet patterns that were not representative of the long-term diets of these fish. Alternatively, the short-term patterns may represent a switch from a winter to summer diet. This study was primarily focused on describing the winter and spring diets for the various members of the community. A complementary study for the summer and fall diets would help to determine if this mismatch of priors and isotope signatures represent a seasonal shift in diets.

The Newfoundland and Labrador marine ecosystems are still in a state of recovery following overexploitation beginning decades ago [3–6, 61, 62, 65, 82, 147]. Despite significant progress towards recovery, many groundfish populations have not yet reached reference points for commercial exploitation [5, 6, 119, 120], a process that can be rate limited by interspecific interactions [148]. Through consideration of species interactions, we have highlighted interactions between current and recovering fisheries to illustrate likely ecological and economic impacts of groundfish recoveries. If additional groundfish populations approach levels where exploitation may resume in a sustainable manner, ecosystem and trophic dynamic considerations such as those presented should be considered in management decisions in order to facilitate ecosystem productivity and recovery. Such information gaps as trophodynamics and how they vary spatially are essentials inputs to ecosystem-based models that interact with other gaps such as essential fish habitat and population dynamics to help build an ecosystem framework of management [16, 149–151].

In summary, along the Newfoundland and Labrador shelf, spatially variable recovery rates of groundfish stocks are associated with variation in trophic interactions. Through a combination of stable isotope and stomach content analyses, multispecies diets reconstructed the upper food web within three regions. While the southern-most region showed the greatest diversity in diets with multiple species being of near equal contribution as prey, the diets were dominated by only one or two species in the northern two regions. This finding demonstrates that diet diversity within the community is positively associated with recovery rates of groundfish species, if not directly influencing those rates. Recent concerns about the future of economically important Pandalid shrimp fisheries in the context of potential groundfish recoveries illustrates the need to quantify spatial variation in trophic interaction and key pathways of energy flow into the upper food web. Pandalid shrimp were a key food web node as a significant contribution to the diets of multiple fish species. Further comparison of the diets of three economically important fishes confirmed spatial variability in prey consumption, emphasizing the need to appreciate and incorporate fine-scale variation in trophic interactions and consideration of the roles of both target and non-target species when developing management advice.

Supporting information

S1 Table. List of prey species/groups included in the isotope mixing model for each predatory fish species.

Determined from our stomach contents analyses and the presented studies from other researchers.

https://doi.org/10.1371/journal.pone.0268440.s001

(DOCX)

S2 Table. Length-weight relationships of representative fish species of Newfoundland and Labrador.

https://doi.org/10.1371/journal.pone.0268440.s002

(DOCX)

S3 Table. Fractionation coefficients and prey groups analyzed for the isotope mixing models of the remaining species.

Percent IRI is presented in parentheses.

https://doi.org/10.1371/journal.pone.0268440.s003

(DOCX)

S4 Table. Carbon and nitrogen stable isotope values and concentrations for all sampled organisms.

Reported values represent means (± SD).

https://doi.org/10.1371/journal.pone.0268440.s004

(DOCX)

S5 Table. Isotope sample sizes by region and size category with food-containing stomach sample sizes.

https://doi.org/10.1371/journal.pone.0268440.s005

(DOCX)

S1 Fig. American Plaice diet composition as determined from isotope mixing models divided by size category and region.

https://doi.org/10.1371/journal.pone.0268440.s006

(PDF)

S2 Fig. Atlantic Herring diet composition as determined from isotope mixing models divided by size category and region.

https://doi.org/10.1371/journal.pone.0268440.s007

(PDF)

S3 Fig. Capelin diet composition as determined from isotope mixing models divided by size category and region.

https://doi.org/10.1371/journal.pone.0268440.s008

(PDF)

S4 Fig. Eelpout diet composition as determined from isotope mixing models divided by size category and region.

https://doi.org/10.1371/journal.pone.0268440.s009

(PDF)

S5 Fig. Lanternfish diet composition as determined from isotope mixing models divided by size category and region.

https://doi.org/10.1371/journal.pone.0268440.s010

(PDF)

S6 Fig. Thorny Skate diet composition as determined from isotope mixing models divided by size category and region.

Functional group categories are designated by two letter abbreviations: PI for pelagic invertebrates, BI for benthic invertebrates, PF for pelagic fish, DF for demersal fish.

https://doi.org/10.1371/journal.pone.0268440.s011

(PDF)

Acknowledgments

At-sea collections were supported by the Newfoundland and Labrador Department of Fisheries and Aquaculture. Special thanks to our industry partner Captain G. Chidley of G & D Fisheries, Ltd. We thank D. Robert, S. Leroux, G. Sherwood, T. Eddy, and S. Grant for their constructive feedback, and G. Rose and S. Rowe for access to cod isotope samples and stomach content data. We would also thank the laboratory manager at the Cornell University Stable Isotope Laboratory, K. Sparks, for facilitating our analyses.

References

  1. 1. Fisher JA, Casini M, Frank KT, Möllmann C, Leggett WC, Daskalov G. The importance of within-system spatial variation in drivers of marine ecosystem regime shifts. Philos Trans R Soc B Biol Sci. 2015;370(1659): 1–8.
  2. 2. Kortsch S, Primicerio R, Aschan M, Lind S, Dolgov A V., Planque B. Food-web structure varies along environmental gradients in a high-latitude marine ecosystem. Ecography (Cop). 2019;42(2): 295–308.
  3. 3. Lilly GR, Parsons DG, Kulka DW. Was the increase in shrimp biomass on the Northeast Newfoundland Shelf a consequence of a release in predation pressure from cod? J Northwest Atl Fish Sci. 2000;27: 45–61.
  4. 4. Windle MJS, Rose GA, Devillers R, Fortin MJ. Spatio-temporal variations in invertebrate-cod-environment relationships on the Newfoundland-Labrador Shelf, 1995–2009. Mar Ecol Prog Ser. 2012;469: 263–278.
  5. 5. Pedersen EJ, Thompson PL, Ball RA, Fortin MJ, Gouhier TC, Link H, et al. Signatures of the collapse and incipient recovery of an overexploited marine ecosystem. R Soc Open Sci. 2017;4(7): 170215 pmid:28791149
  6. 6. Krumsick KJ, Fisher JAD. Community size spectra provide indicators of ecosystem recovery on the Newfoundland and Labrador shelf. Mar Ecol Prog Ser. 2020;635: 123–137.
  7. 7. Shelton PA, Sinclair AF, Chouinard GA, Mohn R, Duplisea DE. Fishing under low productivity conditions is further delaying recovery of Northwest Atlantic cod (Gadus morhua). Can J Fish Aquat Sci. 2006;63(2): 235–238.
  8. 8. Shelton PA, Healey BP. Should depensation be dismissed as a possible explanation for the lack of recovery of the northern cod (Gadus morhua) stock? Can J Fish Aquat Sci. 1999;56(9): 1521–1524.
  9. 9. Sguotti C, Otto SA, Frelat R, Langbehn TJ, Ryberg MP, Lindegren M, et al. Catastrophic dynamics limit Atlantic cod recovery. Proc R Soc B Biol Sci. 2018;286(1898): 20182877.
  10. 10. Olsen EM, Heino M, Lilly GR, Morgan MJ, Brattey J, Ernande B, et al. Maturation trends indicative of rapid evolution preceded the collapse of northern cod. Nature. 2004;428(6986): 932–935. pmid:15118724
  11. 11. Krumsick KJ, Rose GA. Atlantic cod (Gadus morhua) feed during spawning off Newfoundland and Labrador. ICES J Mar Sci. 2012;69(10): 1701–1709.
  12. 12. Milich L. Resource Management versus sustainable livelihoods: the collapse of the Newfoundland cod fishery. Soc Nat Resour. 1999;12(7): 625–642.
  13. 13. Rose GA. Cod: The ecological history of the north Atlantic fisheries. St. John’s, Breakwater Books Limited; 2007.
    • 14. Bavington D. Managed annihilation: an unnatural history of the Newfoundland cod collapse. Vancouver, UBC press; 2011
      • 15. Pikitch EK, Santora C, Babcock EA, Bakun A, Bonfil R, Conover DO, et al. Ecosystem-Based Fishery Management. Science. 2004;305(5682): 346–348. pmid:15256658
      • 16. Slocombe DS. Implementing ecosystem-based management managing a region. Bioscience. 1993;43(9): 612–22.
      • 17. Worm B, Hillborn R, Baum JK, Branch TA, Collie JS, Costello C, et al. Rebuilding global fisheries. Science. 2009; 325(5940): 578–585. pmid:19644114
      • 18. Hillborn R. Future directions in ecosystem based fisheries management: a personal perspective. Fish Res. 2011;108(2–3): 235–239.
      • 19. Howell D, Schueller AM, Bentley JW, Buchheister A, Chagaris D, Cieri M, et al. Combining ecosystem and single-species modeling to provide ecosystem-based fisheries anagement advice within current management systems. Front Mar Sci. 2021; 7: 1163.
      • 20. Crowder LB, Hazen EL, Avissar N, Bjorkland R, Latanich C, Ogburn MB. The impacts of fisheries on marine ecosystems and the transition to ecosystem-based management. Annu Rev Ecol Evol Syst. 2008;39: 259–278.
      • 21. Fulton EA, Smith ADM, Smith DC, Johnson P. An integrated approach is needed for ecosystem based fisheries management: Insights from ecosystem-level management strategy evaluation. PLoS One. 2014;9(1): e84242. pmid:24454722
      • 22. Hollowed AN, Bax N, Beamish R, Collie J, Fogarty M, Livingston P. et al. Are multispecies models an improvement on single-species models for measuring fishing impacts on marine ecosystems? ICES J Mar Sci. 2000; 57(3): 707–719
      • 23. Casini M, Bartolino V, Molinero JC, Kornilovs G. Linking fisheries, trophic interactions and climate: threshold dynamics drive herring Clupea harengus growth in the central Baltic Sea. Mar Ecol Prog Ser. 2010;413: 241–252.
      • 24. Bundy A, Fanning P. Can Atlantic cod (Gadus morhua) recover? Exploring trophic explanations for the non-recovery of the cod stock on the eastern Scotian Shelf, Canada. Can J Fish Aquat Sci. 2005; 62(7): 1474–1489.
      • 25. Naiman RJ, Latterell JJ. PRinciples of linking fish habitat to fisheries management and conservation. J Fish Bio. 2005; 67: 166–185.
      • 26. Navia AF, Cortés E, Jordán F, Cruz-Escalona VH, Mejía-Falla PA. Changes to marine trophic networks caused by fishing. In: Ali M, editor. Diversity of Ecosystems. Rijeka, Croatia: InTech; 2012. pp. 417–452.
        • 27. Zhang C, Chen Y, Ren Y. An evaluation of implementing long-term MSY in ecosystem-based fisheries management: Incorporating trophic interaction, bycatch and uncertainty. Fish Res. 2016; 174: 179–189.
        • 28. Kadin M, Blenckner T, Casini M, Gårdmark A, Torres M, Otto SA. Trophic interactions, management tradeoffs and climate change: the need for adaptive thresholds to operationalize ecosystem indicators. Front Mar Sci. 2019; 21: 249.
        • 29. Amundsen PA, Sánchez-Hernández J. Feeding studies take guts–critical review and recommendations of methods for stomach contents analysis in fish. J Fish Biol. 2019;95(6): 1364–1373. pmid:31589769
        • 30. Hynes HBN. The food of the freshwater sticklebacks (Gasterosteus aculeatus and Pygosteus pungitius) with a review of methods used in the studies of the food of fishes. J Anim Ecol. 1950;19: 36–58.
        • 31. Hyslop EJ. Stomach contents analysis—a review of methods and their application. J Fish Biol. 1980;17(4): 411–429.
        • 32. Kharlamenko VI, Zhukova N V., Khotimchenko S V., Svetashev VI, Kamenev GM. Fatty acids as markers of food sources in a shallow-water hydrothermal ecosystem (Kraternaya Bight, Yankich Island, Kurile Islands). Mar Ecol Prog Ser. 1995;120(1–3): 231–242.
        • 33. Kiyashko SI, Kharlamenko V. I., Imbs AB. Stable isotope ratios and fatty acids as food source markers of deposit-feeding invertebrates. Russ J Mar Biol. 1998;24: 170–174.
        • 34. Parrish CC, Abrajano TA, Budge SM, Helleur RJ, Hudson ED, Pulchan K, et al. Lipid and phenolic biomarkers in marine ecosystems: Analysis and applications. Mar Chem. 2005;(May 2014): 193–223.
        • 35. Symondson WOC. Molecular identification of prey in predator diets. Mol Ecol. 2002;11(4): 627–641. pmid:11972753
        • 36. King RA, Read DS, Traugott M, Symondson WOC. Molecular analysis of predation: A review of best practice for DNA-based approaches. Mol Ecol. 2008;17(4): 947–963. pmid:18208490
        • 37. Deagle BE, Kirkwood R, Jarman SN. Analysis of Australian fur seal diet by pyrosequencing prey DNA in faeces. Mol Ecol. 2009;18(9): 2022–2038. pmid:19317847
        • 38. Post DM. Using stable isotopes to estimate trophic position: Models, methods, and assumptions. Ecology. 2002;83(3): 703–718.
        • 39. Fry B. Stable Isotope Ecology. New York, New York: Springer; 2006.
          • 40. Lorrain A, Paulet Y, Chauvaud L, Savoye N, Donval A, Saout C. Differential δ13 C and δ15 N signatures among scallop tissues implications for ecology and physiology. Journal of Experimental Marine Biology and Ecology. 2002;275: 47–61.
          • 41. Peterson BJ, Howarth RW, Garritt RH. Multiple stable isotopes used to trace the flow of organic matter in estuarine food webs. Science. 1985;227(4692): 1361–1363. pmid:17793771
          • 42. Owens NJP. Natural variations in 15N in the marine environment. Adv Mar Biol. 1987;24: 389–451.
          • 43. Peterson BJ, Fry B. Stable isotopes in ecosystem studies. Annu Rev Ecol Syst. 1987;18(1): 293–320.
          • 44. Bunn SE, Boon PI. What sources of organic carbon drive food webs in billabongs? A study based on stable isotope analysis. Oecologia. 1993;96(1): 85–94. pmid:28313757
          • 45. Bearhop S, Adams CE, Waldron S, Fuller RA, Macleod H. Determining trophic niche width: A novel approach using stable isotope analysis. J Anim Ecol. 2004;73(5): 1007–1012.
          • 46. Newsome SD, del Rio CM, Bearhop S, Phillips DL. A niche for isotopic ecology. Front Ecol Environ. 2007;5(8): 429–436.
          • 47. Krumsick KJ, Fisher JAD. Spatial and ontogenetic variation in isotopic niche among recovering fish communities revealed by Bayesian modeling. PLoS One. 2019;14(4): e0215747. pmid:30998793
          • 48. Schwarcz HP. Some theoretical aspects of isotope paleodiet studies. J Archaeol Sci. 1991;18(3): 261–275.
          • 49. Phillips DL. Mixing models in analyses of diet using multiple stable isotopes: A critique. Oecologia. 2001;127(2): 166–170. pmid:24577645
          • 50. DFO. Groundfish Newfoundland and Labrador Region NAFO Subarea 2 + Divisions 3KLMNO. 2019. https://www.dfo-mpo.gc.ca/fisheries-peches/ifmp-gmp/groundfish-poisson-fond/2019/groundfish-poisson-fond-2_3klmno-eng.htm
            • 51. Minet JP, Perdou JB. Predation of cod, Gadus morhua, on capelin, Mallotus villosus, off eastern Newfoundland and in the Gulf of St Lawrence. Intern Comm Northw Atl Fish Res Bull. 1978;13: 11–20.
            • 52. Lilly GR. Interannual variability in predation by cod (Gadus morhua) on capelin (Mallotus villosus) and other prey off southern Labrador and northeastern Newfoundland. ICES Mar Sci Symp. 1991;193: 133–146.
            • 53. Link JS, Garrison LP. Trophic ecology of Atlantic cod Gadus morhua on the northeast US continental shelf. Mar Ecol Prog Ser. 2002;227: 109–123.
            • 54. Lambert DG. The food of the Redfish Sebastes marinus (L.) in the Newfoundland area. J Fish Res Board Canada. 1960;17(2): 235–243.
            • 55. Pedersen SA, Riget F. Feeding habits of redfish (Sebastes spp.) and greenland halibut (Reinhardtius hippoglossoides) in west Greenland waters. ICES J of Mar Sci. 1993;50(4): 445–459.
            • 56. González C, Bruno I, Paz X. Food and feeding of deep-sea redfish (Sebastes mentella Travin) in the North Atlantic. NAFO Sci Counc Stud. 2000;(33): 89–101.
            • 57. Perez-Rodriguez A, Saborido-Rey F. Food consumption of Flemish Cap cod Gadus morhua and redfish Sebastes sp. using generic bioenergetic models. NAFO Sci Counc Res Doc. 2012: 15.
            • 58. Chumakov AK, Podrazhanskaya SG. Feeding of Greenland halibut Reinhardtius hippoglossoides (Pleuronectidae) in the Northwest Atlantic. NAFO Sci Counc Stud. 1986;10: 47–52.
            • 59. Bowering WR, Lilly GR. Greenland halibut (Reinhardtius hippoglossoides) off southern Labrador and northeastern Newfoundland (Northwest Atlantic) feed primarily on capelin (Mallotus villosus). Netherlands J Sea Res. 1992;29(1–3): 211–222.
            • 60. Dawe EG, Bowering WR, Joy JB. Predominance of squid (Gonatus spp.) in the diet of Greenland halibut (Reinhardtius hippoglossoides) on the deep slope of the northeast Newfoundland continental shelf. Fish Res. 1998;36(2–3): 267–273.
            • 61. Dwyer KS, Buren A, Koen-Alonso M. Greenland halibut diet in the Northwest Atlantic from 1978 to 2003 as an indicator of ecosystem change. J Sea Res. 2010;64(4): 436–445.
            • 62. Rose GA, O’Driscoll RL. Multispecies interactions capelin are good for cod: Can the northern stock rebuild without them? ICES J Mar Sci. 2002;59(5): 1018–1026.
            • 63. Templeman W. Migration and intermingling of stocks of Atlantic cod, Gadus morhua, of the Newfoundland and adjacent areas from tagging in 1962–66. ICNAF Res Bull. 1979;14: 5–50.
            • 64. Rose GA, deYoung B, Kulka DW, Goddard SV, Fletcher GL. Distribution shifts and overfishing the northern cod (Gadus morhua): a view from the ocean. Can J Fish Aquat Sci. 2000;57(3): 644–663.
            • 65. Smedbol RK, Wroblewski JS. Metapopulation theory and northern cod population structure:interdependency of subpopulations in recoveryof a groundfish population. Fish Res. 2002;55: 161–174.
            • 66. Rose GA. Cod spawning on a migration highway in the north-west Atlantic. Nature. 1993;366(6454): 458–461.
            • 67. Stevens BG, Armstrong DA, Cusimano R. Feeding habits of the dungeness crab Cancer magister as determined by the index of relative importance. Mar Biol. 1982;72(2): 135–145.
            • 68. Macko SA, Lee Wen Yuh, Parker PL. Nitrogen and carbon isotope fractionation by two species of marine amphipods: Laboratory and field studies. J Exp Mar Bio Ecol. 1982;63(2): 145–149.
            • 69. Minagawa M, Wada E. Stepwise enrichment of 15N along food chains: Further evidence and the relation between δ15N and animal age. Geochim Cosmochim Acta. 1984;48(5): 1135–1140.
            • 70. Vander Zanden MJ, Cabana G, Rasmussen JB. Comparing trophic position of freshwater fish calculated using stable nitrogen isotope ratios (δ15N) and literature dietary data. Can J Fish Aquat Sci. 1997;54(5): 1142–1158.
            • 71. DeNiro MJ, Epstein S. Influence of diet on the distribution of carbon isotopes in animals. Geochim Cosmochim Acta. 1978;42(5): 495–506.
            • 72. Hecky RE, Hesslein RH. Contributions of benthic algae to lake food webs as revealed by stable isotope analysis. J North Am Benthol Soc. 1995;14(4): 631–653.
            • 73. Vander Zanden MJ, Rasmussen JB. Variation in δ15N and δ13C trophic fractionation: Implications for aquatic food web studies. Limnol Oceanogr. 2001;46(8): 2061–2066.
            • 74. Pinnegar JK, Polunin NVC. Differential fractionation of δ13C and δ15N among fish tissues: Implications for the study of trophic interactions. Funct Ecol. 1999;13(2): 225–231.
            • 75. Sotiropoulos MA, Tonn WM, Wassenaar LI. Effects of lipid extraction on stable carbon and nitrogen isotope analyses of fish tissues: Potential consequences for food web studies. Ecol Freshw Fish. 2004;13(3): 155–160.
            • 76. Logan JM, Jardine TD, Miller TJ, Bunn SE, Cunjak RA, Lutcavage ME. Lipid corrections in carbon and nitrogen stable isotope analyses: Comparison of chemical extraction and modelling methods. J Anim Ecol. 2008;77(4): 838–846. pmid:18489570
            • 77. Ricklefs RE, Travis J. A morphological approach to the study of avian community organization Auk 1980;97(2): 321–338.
            • 78. Post DM, Layman CA, Arrington DA, Takimoto G, Quattrochi J, Montaña CG. Getting to the fat of the matter: Models, methods and assumptions for dealing with lipids in stable isotope analyses. Oecologia. 2007;152(1): 179–189. pmid:17225157
            • 79. Elliott JM, Persson L. The estimation of daily rates of food consumption for fish. J Anim Ecol. 1978;47(3): 977–991.
            • 80. Jobling M. Mathematical models of gastric emptying and the estimation of daily rates of food consumption for fish. J Fish Biol. 1981;19(3): 245–257.
            • 81. Temming A, Herrmann JP. Gastric evacuation in cod. Prey-specific evacuation rates for use in North Sea, Baltic Sea and Barents Sea multi-species models. Fish Res. 2003;63(1): 21–41.
            • 82. Sherwood GD, Rose GA. Stable isotope analysis of some representative fish and invertebrates of the Newfoundland and Labrador continental shelf food web. Estuar Coast Shelf Sci. 2005;63(4): 537–549.
            • 83. Scott WB, Scott MG. Atlantic Fishes of Canada. Toronto: University of Toronto Press; 1988.
              • 84. Caut S, Angulo E, Courchamp F. Variation in discrimination factors (Δ15N and Δ13C): The effect of diet isotopic values and applications for diet reconstruction. J Appl Ecol. 2009;46(2): 443–453.
              • 85. Fry B, Sherr EB. δ13C Measurements as indicators of carbon flow in marine and freshwater ecosystems. In: Rundel PW, Ehleringer NKA, editors. Stable Isotopes in Ecological Research. NEw York: Springer-Verlag: 1989; 196–229.
                • 86. Barnes C, Sweeting CJ, Jennings S, Barry JT, Polunin NVC. Effect of temperature and ration size on carbon and nitrogen stable isotope trophic fractionation. Funct Ecol. 2007;21(2): 356–362.
                • 87. Hansson S, Hobbie JE, Elmgren R, Larsson U, Fry B, Johansson S. The stable nitrogen isotope ratio as a marker of food-web interactions and fish migration. Ecology. 1997;78(7): 2249–2257.
                • 88. McCutchan JH, Lewis WM, Kendall C, McGrath CC. Erratum: Variation in trophic shift for stable isotope ratios of carbon, nitrogen, and sulfur (Oikos (2003) 102 (378–390)). Oikos. 2005;111(2): 416.
                • 89. Hussey NE, Brush J, McCarthy ID, Fisk AT. δ15N and δ13C diet-tissue discrimination factors for large sharks under semi-controlled conditions. Comp Biochem Physiol—A Mol Integr Physiol. 2010;155(4): 445–53. pmid:19800980
                • 90. Varela JL, Larrañaga A, Medina A. Prey-muscle carbon and nitrogen stable-isotope discrimination factors in Atlantic bluefin tuna (Thunnus thynnus). J Exp Mar Bio Ecol. 2011;406(1–2): 21–28.
                • 91. Barnes C, Maxwell D, Reuman DC, Jennings S. Global patterns in predator-prey size relationships reveal size dependency of trophic transfer efficiency. Ecology. 2010;91(1): 222–232. pmid:20380211
                • 92. Stock BC, Jackson AL, Ward EJ, Parnell AC, Phillips DL, Semmens BX. Analyzing mixing systems using a new generation of Bayesian tracer mixing models. PeerJ. 2018;6: e5096. pmid:29942712
                • 93. Koch PL, Phillips DL. Incorporating concentration dependence in stable isotope mixing models: A reply to Robbins, Hilderbrand and Farley (2002). Oecologia. 2002;133(1): 114–125.
                • 94. Moore JW, Semmens BX. Incorporating uncertainty and prior information into stable isotope mixing models. Ecol Lett. 2008;11(5): 470–80. pmid:18294213
                • 95. Khachiyan LG. A polynomial algorithm in linear programming. Dokl Akad Nauk SSSR. 1979;244(5): 1093–1096.
                • 96. Parnell AC, Phillips DL, Bearhop S, Semmens BX, Ward EJ, Moore JW, et al. Bayesian stable isotope mixing models. Environmetrics. 2013;24(6): 387–399.
                • 97. Phillips DL, Inger R, Bearhop S, Jackson AL, Moore JW, Parnell AC, et al. Best practices for use of stable isotope mixing models in food-web studies. Can J Zool. 2014;92(10): 823–835.
                • 98. Van Denderen PD, Lindegren M, MacKenzie BR, Watson RA, Andersen KH. Global patterns in marine predatory fish. Nat Ecol Evol. 2018;2(1): 65–70. pmid:29180711
                • 99. Koen-Alonso M, Lindstrøm U, Cuff A. Comparative modeling of cod-capelin dynamics in the Newfoundland-Labrador shelves and Barents Sea ecosystems. Front Mar Sci. 2021;8: 579946.
                • 100. Klumpp DW, Salita-Espinosa JS, Fortes MD. The role of epiphytic periphyton and macroinvertebrate grazers in the trophic flux of a trophic seagrass community. Aquat Bot. 1992; 43(4): 327–349.
                • 101. Lauerman LM, Kaufmann RS. Deep-sea epibenthic echinoderms and a temporally varying food supply: results from a one year time series in the NE Pacific. Deep Sea Res II: Tropical Studies Oceanogr. 1998. 45(4–5): 817–842.
                • 102. Hill WR, Middleton RG. Changes in carbon stable isotope ratios during periphyton development. Limnol Oceanogr. 2006; 51(5): 2360–2369.
                • 103. Nadon M-O, Himmelman JH. Stable isotopes in substidal food webs: Have enriched carbon ratios in benthic consumers been misinterpreted? Limnol Oceanogr. 2006; 51(6): 2828–2836.
                • 104. Tamelander T, Renaud PE, Hop H, Caroll ML, Ambrose WG, Hobson KA. Trophic relationship and pelagic-benthic coupling during summer in the Barents Sea Marginal Ice Zone, revealed by stable carbon and nitrogen isotope measurements. Mar Ecol Prog Ser. 2006; 310: 33–46.
                • 105. Gale KSP, Hamel J-F, Mercier A. Trophic ecology of deep-sea Asteroidea (Echinodermata) from eastern Canada. Deep Sea Res I: Oceanogr Res Papers. 2013; 80: 25–36.
                • 106. Saupe SM, Schell DM, Griffiths WB. Carbon-isotope ratio gradients in western arctic zooplankton. Mar Biol. 1989; 103: 427–432.
                • 107. Sato M, Sasaki H, Fukuchi M. Stable isotopic compositions of overwintering copepods in the arctic and subarctic waters and implications to the feeding history. J Mar Syst. 2002; 38(1–2): 165–174.
                • 108. Tamelander T, Søreide JE, Hop H, Carroll ML. Fractionation of stable isotopes in the Arctic marine copepod Calanus glacialis: Effects on the isotopic composition of marine particulate organic matter. J Exp Mar Biol Ecol. 2006; 333(2): 231–240.
                • 109. Bowering WR, Parsons DG, Lilly GR. Predation on shrimp (Pandalus borealis) by Greenland halibut (Reinhardtius hippoglossoides) and Atlantic cod (Gadus morhua) off coastal Labrador (Div. 2H and 2J). NAFO Sci Counc Res. 1983;83/IX/88.
                • 110. Parsons DG. Predators of northern shrimp, Pandalus borealis (Pandalidae), throughout the North Atlantic. Mar Biol Res. 2005;1(1): 48–58.
                • 111. Hobson KA, Fisk A, Karnovsky N, Holst M, Gagnon JM, Fortier M. A stable isotope (δ13C, δ15N) model for the North Water food web: Implications for evaluating trophodynamics and the flow of energy and contaminants. Deep Res Part II Top Stud Oceanogr. 2002;49(22–23): 5131–5150.
                • 112. Nadon MO, Himmelman JH. Stable isotopes in subtidal food webs: Have enriched carbon ratios in benthic consumers been misinterpreted? Limnol Oceanogr. 2006;51(6): 2828–2836.
                • 113. Schlacher TA, Wooldridge TH. Origin and trophic importance of detritus—Evidence from stable isotopes in the benthos of a small, temperate estuary. Oecologia. 1996;106(3): 382–388. pmid:28307326
                • 114. Wooller M, Smallwood B, Jacobson M, Fogel M. Carbon and nitrogen stable isotopic variation in Laguncularia racemosa (L.) (white mangrove) from Florida and Belize: implications for trophic level studies. Hydrobiologia. 2003;499: 13–23.
                • 115. Tewfik A, Rasmussen JB, McCann KS. Anthropogenic enrichment alters a marine benthic food web. Ecology. 2005;86(10): 2726–2736.
                • 116. Healey BP. Greenland halibut (Reinhardtius hippoglossoides) in NAFO Subarea 2 and Divisions 3KLMNO: Stock trends based on annual Canadian research vessel survey results during 1978–2009. NAFO Sci Counc Res. 2010;10/21: 1–67.
                • 117. DFO. Recovery potential assessment of Redfish (Sebastes mentella and S. fasciatus). Can Sci Advis Secr Sci Advis Rep. 2011;2011/044: 18.
                  • 118. DFO.Stock assessment of northern cod (NAFO division 2J3KL) in 2019.. CSAS Sci Advis Rep. 2019;2019/050: 1–20.
                    • 119. Rose GA, Rowe S. Does redistribution or local growth underpin rebuilding of Canada’s northern cod? Can J Fish Aquat Sci. 2018;75(6): 825–835.
                    • 120. Department of Fisheries and Land Resources. Seafood industry year in review 2019. SYIR 2019. 2019; https://www.gov.nl.ca/ffa/files/2019-SIYIR-WEB.pdf
                      • 121. Worm B, Myers RA. Meta-analysis of cod-shrimp interactions reveals top-down control in oceanic food webs. Ecology. 2003;84(1): 162–173.
                      • 122. Mullowney DRJ, Dawe EG, Colbourne EB, Rose GA. A review of factors contributing to the decline of Newfoundland and Labrador snow crab (Chionoecetes opilio). Rev Fish Biol Fish. 2014;24(2): 639–657.
                      • 123. Parsons DG, Crawford RE, Savard L, Dalley EL. Aspects of the distribution and behaviour of northern shrimp (Pandalus borealis) affecting survey design and stock assessment. 1991. CAFSAC Working Paper 91: 1–16.
                        • 124. Parsons DG. Interactions between northern shrimp, Pandalus borealis (Pandalidae), and its key predators within the eastern Newfoundland and Labrador marine ecosystem. Mar Biol Res. 2005; 1(1): 59–67.
                        • 125. Link JS, Bogstad B, Sparholt H, Lilly GR. Trophic role of Atlantic cod in the ecosystem. Fish Fish. 2009; 10(1): 58–87.
                        • 126. Anderson PJ. Pandalid shrimp as indicators of ecosystem regime shift. J Northw Atl Fish Sci. 2000; 27: 1–10.
                        • 127. Colbourne EB, Anderson JT. Biological response of a changing ocean environment in Newfoundland waters during the latter decades of the 1900s. ICES Mar Sci Symposia. 2003; 219: 169–181.
                        • 128. Drinkwater KF. The reponse of Atlantic cod (Gadus morhua) to future climate change. ICES J Mar Sci. 2005; 62: 1327–1337.
                        • 129. Reade D. A cod forsaken place? Fishing in an altered state in Newfoundland. Anthropol Q. 2014: 695–726.
                          • 130. Le Corre N, Pepin P, Han G, Ma Z. Potential impact of climate change on northern shrimp habitats and connectivity on the Newfoundland and Labrador continental shelves. Fish Oceanogr. 2021; 30(3): 331–347.
                          • 131. Curruthers EH, Parlee CE, Keenan R, Foley P. Onshore benefits from fishing: Tracking value from the northern shrimp fishery to communities in Newfoundland and Labrador. Mar Policy. 2019;103: 130–137.
                          • 132. Williams RJ, Berlow EL, Dunne JA, Barabási AL, Martinez ND. Two degrees of separation in complex food webs. Proc Natl Acad Sci U S A. 2002;99(20): 12913–12916. pmid:12235367
                          • 133. Jennings S, Warr KJ. Smaller predator-prey body size ratios in longer food chains. Proc R Soc B Biol Sci. 2003;270(1522): 1413–1417. pmid:12965034
                          • 134. Havens K. Scale and Structure in Natural Food Webs Science 1992;257(5073):1107–9. pmid:17840281
                          • 135. Hwang YT, Millar JS, Longstaffe FJ. 2007. Do δ15N and δ13C values in feces reflect the isotopic composition of diets in small mammals? Can J Zool 85(3): 388–396
                          • 136. Guelinckx J, Dehairs F, Ollevier F. Effect of digestion on the δ13C and δ15N of fish-gut contents. J Fish Biol. 2008;72(1): 301–309.
                          • 137. Baker R, Buckland A, Sheaves M. Fish gut content analysis: Robust measures of diet composition. Fish Fish. 2014;15(1): 170–177.
                          • 138. Rindorf A, Lewy P. Bias in estimating food consumption of fish by stomach-content analysis. Can J Fish Aquat Sci. 2004;61(12): 2487–2498.
                          • 139. Barnett A, Redd KS, Frusher SD, Stevens JD, Semmens JM. Non-lethal method to obtain stomach samples from a large marine predator and the use of DNA analysis to improve dietary information. J Exp Mar Bio Ecol. 2010;393(1–2): 188–192.
                          • 140. Carreon-Martinez L, Johnson TB, Ludsin SA, Heath DD. Utilization of stomach content DNA to determine diet diversity in piscivorous fishes. J Fish Biol. 2011;78(4): 1170–1182. pmid:21463313
                          • 141. Hosseini R, Schmidt O, Keller MA. Factors affecting detectability of prey DNA in the gut contents of invertebrate predators: A polymerase chain reaction-based method. Entomol Exp Appl. 2008;126(3): 194–202.
                          • 142. Bond AL, Diamond AW. Recent Bayesian stable-isotope mixing models are highly sensitive to variation in discrimination factors. Ecol Appl. 2011;21(4): 1017–1023. pmid:21774408
                          • 143. Gannes LZ, O’Brien DM, Del Rio CM. Stable isotopes in animal ecology: Assumptions, caveats, and a call for more laboratory experiments. Ecology. 1997;78(4): 1271–1276.
                          • 144. Caut S, Angulo E, Courchamp F. Discrimination factors (Δ15N and Δ13C) in an omnivorous consumer: Effect of diet isotopic ratio. Funct Ecol. 2008;22(2): 255–263.
                          • 145. Robbins CT, Felicetti LA, Sponheimer M. The effect of dietary protein quality on nitrogen isotope discrimination in mammals and birds. Oecologia. 2005;144(4): 534–540. pmid:15800751
                          • 146. Hussey NE, Macneil MA, Mcmeans BC, Olin JA, Dudley SFJ, Cliff G, et al. Rescaling the trophic structure of marine food webs. Ecol Lett. 2014;17(2): 239–250. pmid:24308860
                          • 147. Lilly G, Kruse GH, Drinkwater K, Ianelli JN, Link JS, Stram DL, et al. The decline, recovery, and collapse of Atlantic cod (Gadus morhua) off Labrador and eastern Newfoundland. Alaska Sea Grant Coll Progr. 2008: 67–88.
                          • 148. Fung T, Farnsworth KD, Shephard S, Reid DG, Rossberg AG. Why the size structure of marine communities can require decades to recover from fishing. Mar Ecol Prog Ser. 2013;484: 155–171.
                          • 149. Molina-ureña H, Ault JS, De Biología E, Rica UDC, José S, Rica C. Towards an ecosystem-based approach to assess non-target tropical reef fishes. Gulf and Caribbean Fisheries Institute 59 2007: 185–190.
                          • 150. Long RD, Charles A, Stephenson RL. Key principles of marine ecosystem-based management. Mar Policy. 2015;57: 53–60.
                          • 151. Perryman HA, Hansen C, Howell D, Olsen E. A review of applications evaluating fisheries management scenarios through marine ecosystem models. Rev Fish Sci Aquac. 2021. 29(4): 800–835.