Dissostichus spp. are commercially valuable marine fish. Dissostichus mawsoni is mainly distributed south of the polar front. Antifreeze proteins allow them to survive low water temperatures. Their range differs from Dissostichus eleginoides, with some exceptions (Gon and Heemstra, 1990; Eastman and Hubold, 1999; Goldsworthy et al., 2002). Dissostichus mawsoni primarily feeds on fish and cephalopods. They represent a high ecological niche as apex predators and scavengers and serve as food sources for seals, cetaceans, and giant squids (Fenaughty et al., 2003; Stevens et al., 2014). Furthermore, based on ontogenetic changes, this species inhabits shallow waters during the early stages of growth and then moves to deeper waters up to 2,200 m during the adult stage (Hanchet et al., 2015); thus, its ecological role in the Southern Ocean ecosystem is critical throughout its life cycle (Mormede et al., 2014). However, there are uncertainties of information on the biology and life cycle of D. mawsoni has been primarily obtained from fishery samples (Mormede et al., 2014; Ainley et al., 2016).

The geographical distance and inaccessible natural environment of the Southern Ocean make it challenging to systematically study species richness and biomass within its waters (Caccavo et al., 2021). Antarctic toothfish fisheries occur only in convention areas of the Commission for the Conservation of Antarctic Marine Living Resources (CCAMLR) and are managed through ecosystem-based resource management and a precautionary approach. Most of the data used were collected by scientific observers onboard vessels, and sampling followed strict regulations of the CCAMLR (e.g., random sampling). Antarctic toothfish catches averaged 4,157 metric tons per year during the 2011–2020 fishing season, with the highest average being 2,871 metric tons per year in subarea 88.1 (Ross Sea). The study area, subarea 88.3, is a CCAMLR convention area located in the Bellingshausen Sea and eastern Amundsen Sea. The Antarctic toothfish fishery in subarea 88.3 began in 1998 and was conducted intermittently until 2012, when exploratory fisheries were reinstated in 2016 to obtain information on biological and ecological species relationships for ecosystem-based management, with an average annual catch of 85 metric tons by 2020. There is also a lack of information on the environmental characteristics, species composition of the ecosystem, and trophic interactions of Antarctic toothfish in subarea 88.3.

Traditional single-species management methods must be reconsidered as the sole methods for preserving and managing fishery resources worldwide. Recently, an ecosystem-based resource management method has been developed that considers the relationships between various species as well as the interactions between target species and the environment (Slocombe, 1993; Berkes, 2012; Kadin et al., 2019; Cucuzza et al., 2021). A precautionary approach is used for risk management to address situations with incomplete data (Karim et al., 2020). Feeding ecology studies can help identify predators-prey relationships within ecosystems, assess the ecological status and role of target species, and predict how environmental changes caused by fishing and climate change affect ecosystems (Link, 2002; Roessig et al., 2004; Melnychuk et al., 2017). Such studies can also estimate the biomass and abundance of prey species (Sallaberry-Pincheira et al., 2018; Ng et al., 2021). Similarly, it is important to understand the function of the target species in the food web structure through dietary studies, such as exploring the effects of population structure changes and population decline of prey species through fisheries. However, diet composition data can differ depending on the temporal and spatial factors related to the sample and size class (Boyd, 1996; Holt et al., 2019). Diet composition studies can be conducted using various size classes and samples with a wide spatiotemporal range, which helps mitigate limitations, and specific trophic interactions within the target area can be better understood. A long-term dietary study of D. mawsoni can reveal its specific ecosystem functions in the Southern Ocean. Furthermore, it can help inform ecosystem-based management and precautionary approaches to better predict resource fluctuations caused by fisheries and changes in the marine environment.

Among the studies of D. mawsoni diets in area 88 of the Southern Ocean (Pacific Ocean sector), Kokorin (2010) analyzed differences in the composition of prey organisms based on the growth and habitat of D. mawsoni collected in the Ross and Amundsen Seas. In addition, Stevens et al. (2014) analyzed the differences in diet composition based on the growth, region, and fishing season of D. mawsoni collected in the Ross Sea. Both studies have demonstrated that the diet of D. mawsoni differs depending on its growth and area (Kokorin, 2010; Stevens et al., 2014). There was no difference in the dietary composition of D. mawsoni living in the Ross Sea between 2003 and 2010 (Stevens et al., 2014). Similarly, Queirós et al. (2022) reported differences in diet composition by region via morphological and isotopic analyses of D. mawsoni collected from the Amundsen Sea and Dumont D’Urville Sea. Yoon et al. (2017) and Lee et al. (2022) used metabarcoding to compare the diets of D. mawsoni collected in subareas 88.3 and 58.4 and identified various prey species that could not be accurately classified morphologically post-digestion. However, these authors could not analyze specific dietary ecology based on factors such as water depth and size class.

The present study analyzed the contents of D. mawsoni stomach samples collected over five years from 2016 to 2020 in the subarea 88.3 of the Southern Ocean. The study aimed to analyze feeding ecology in detail regarding temporal and spatial factors and size variations. We conducted quantitative analyses of differences in diet composition among research blocks and fishing seasons within subarea 88.3 and changes in D. mawsoni diet composition over time (fishing season), space (depth), and size class at each site (88.3_3 and 88.3_4) during each year of the study. This study will help improve our understanding of the feeding ecology, ecological niche, and functional role of D. mawsoni by identifying the main prey by area and the pattern of dietary changes in subareas of 88.3 of the Southern Ocean. In addition, this study provides ecological information for the sustainable use and management of D. mawsoni resources by predicting the impacts of fishing and environmental change.

It is presumed that C. dewitti, Macrourus spp., and Nototheniidae were the dominant prey species in this study because they are similar to D. mawsoni in their depth ranges and appear abundantly in the study area. Chionobathyscus dewitti, which appeared predominantly in this study, is the deepest among the 16 species belonging to Channichthyidae and has been reported to live at depths of up to 2,025 m (Kock, 2005; Balushkin and Prut’ko, 2006; Eastman et al., 2013). Macrourus caml occurred at depths of 350–2,080 m and was previously identified as M. whitsoni; however, McMillan et al. (2012) suggested the existence of two species. Lee et al. (2022) identified M. caml and M. whitsoni in the stomach contents of D. mawsoni from subarea 88.3 through metabarcoding. It is assumed that these two species are sympatric in subarea 88.3. Lepidonotothen spp. were assumed to be Lepidonotothen squamifrons based on comparisons with stomach content metabarcoding analysis by Lee et al. (2022). Lepidonotothen squamifrons, T. loennbergii, and T. pennellii are all known demersal fish, and they inhabit depths of 5–651 m (Koubbi et al., 2000; Eastman et al., 2013), 65–1,191 m (Gon and Heemstra, 1990; Eastman and Hubold, 1999), and 63–675 m (Hanchet et al., 2013) respectively. In addition, L. squamifrons, Macrourus spp., and C. dewitti are demersal fish predominantly observed in the western Southern Ocean (Amsler et al., 2016) and are the main bycatch species among the toothfish fisheries in subarea 88.3 (Arana and Vega, 1999).

The differences in diet composition among the research blocks resulted from a combination of factors, including habitat type, prey abundance, and water depth. Fish feeding differs markedly according to spatial factors, such as habitat type (Platell et al., 2007; Lek et al., 2011), differences in food resources (Mittelbach et al., 1992; Francis and Schindler, 2009), temporal factors, such as seasonal and annual changes in prey species (Hovde et al., 2002; Ginter et al., 2012), and ontogenetic changes (Werner and Gilliam, 1984; Park et al., 2021). Toothfish are top predators that feed on fish and cephalopods. They have been reported to show differences in diet composition depending on area, depth, and body size (Arkhipkin et al., 2003; Roberts et al., 2011; Stevens et al., 2014). The main prey of D. mawsoni are Channichthyidae and cephalopods in the Lazarev Sea (Petrov and Tatarnikov, 2011), and Macrouridae and cephalopods in the South Sandwich Islands (Roberts et al., 2011). In addition, Macrouridae dominates the diet composition of D. mawsoni in the Weddell Sea (Petrov and Gordeev, 2015) and subarea 58.4 (Seong et al., 2021). Furthermore, D. mawsoni, which inhabits the continental slopes of the Ross Sea, feeds mainly on Channichthyidae and Macrourus spp. However, in habitats such as ridges, hills, and seamounts, they mainly feed on Macrourus spp. (Stevens et al., 2014). D. mawsoni localized in the continental slope of the Amundsen Sea mainly feed on cephalopods, Macrourus spp., and Trematomus spp.; however, in the seamounts, they mainly feed on cephalopods, Macrourus spp., and Antimora rostrata (Queirós et al., 2022). Therefore, it is assumed that the food web structure in the Southern Ocean ecosystem differs depending on the local environment. Additional studies are required to understand the structure of the food webs in each area.

Fish cannibalism is affected by many factors, including the quantity and quality of prey and population density (Pereira et al., 2017). Cannibalism regulates population density (Mammel et al., 2022), and individuals grow faster than in populations without cannibalism (Berg et al., 2010; Huss et al., 2010). Toothfish are voracious predators that feed on their offspring. In this study, cannibalism of D. mawsoni was observed, with 163.8 cm and 158.8 cm individuals feeding on 58.1 cm and 54.3 cm individuals, respectively. Cannibalism of D. mawsoni in the past was reported by Queirós et al. (2022), and D. eleginoides was reported by Arkhipkin et al. (2003) and Troccoli et al. (2020).

Stomach emptiness in fishes is influenced by a variety of factors, including feeding ecology, environmental conditions, and sampling procedures (Vinson and Angradi, 2011). At higher trophic levels, piscivorous fishes consume larger, more energy-dense prey and have higher empty rates, on average, than omnivores, invertivores, planktivores, and herbivores (Malmquist et al., 1992; Arrington et al., 2002). Previous studies on Antarctic toothfish diets have reported emptying rates ranging from 9.8–69.1%, with an average of approximately 32% (Fenaughty et al., 2003; Kokorin, 2010; Petrov and Tatarnikov, 2011; Roberts et al., 2011; Stevens et al., 2014; Seong et al., 2021; Queirós et al., 2022), and these high emptying rates are similar to our results. Moreover, both Antarctic and Patagonian toothfish from South Sandwich Island have been shown to have decreased empty rates with increasing depth and size classes (Roberts et al., 2011), which is similar to our findings. These results are contrary to the hypotheses that deeper depths are more likely to cause stress-induced prey regurgitation (Pilling et al., 2001) and that relatively large fish should have higher empty rates because they feed on larger, more energy-dense prey (Vinson and Angradi, 2011). Therefore, the variation in emptying rates among Antarctic toothfish by depth and size group is likely the result of factors such as feeding ecology, physiological characteristics, and environmental conditions; however, further research is needed to clarify this.

Changes in the Antarctic toothfish prey composition with depth were closely linked to the depth range and abundance of each prey species. At shallow depths below 550 m, Nototheniidae such as Trematomus scotti and L. squamifrons are predominantly observed in the Bellingshausen Sea (Matallanas and Olaso, 2007). However, with increasing depth, the abundance of Nototheniidae decreased significantly, and Zoarcidae, Macrouridae, and Muraenolepidae dominated at depths > 550 m. In addition, Nototheniidae, such as L. squamifrons and T. loennbergii were dominant at depths of 400–599 m on Anvers Island and the Western Antarctic Peninsula, with a significant decrease in abundance observed at depths deeper than 599 m (Amsler et al., 2016). Macrourus spp. and C. dewitti were abundant at depths of 700–1,499 m and 900–1,499 m, respectively. In Marguerite Bay, Macrourus spp. and C. dewitti proportions were 5% and 6%, respectively, at depths of< 600 m and 30% and 12% at depths of 600–1,000 m, respectively. At depths > 1,000 m, the Macrourus spp. and C. dewitti proportions were 65% and 82%, respectively (Eastman et al., 2013).

In general, fish show variations in diet composition as they grow because they improve their swimming ability, increase their mouth size, become more food-demanding, and can consume various prey species (Barnes et al., 2010; Holt et al., 2019), and we believe that the changes in diet composition among size classes in this study are related to this. Differences in prey composition with growth rate are associated with an increased ability of larger individuals to consume a broader range of prey species and are known to reduce within-species competition for food (Park et al., 2022). Dissostichus mawsoni consumed Channichthyidae, Nototheniidae, Anotopteridae, and cephalopods in the<100 cm size class, whereas in the ≥100 cm size class, the diet was dominated by Macrouridae (Stevens et al., 2014; Seong et al., 2021; Pérez-Pezoa et al., 2023). This may be because D. mawsoni migrates to deeper depths due to ontogenetic shifts, thus feeding primarily on Macrouridae, which are abundant in deeper waters. Nevertheless, the observed differences in preferred prey sources among size classes in this study, even within the same depth range, suggest that they coexist in a way that avoids competition for food among the same species.

Data on most fish and cephalopod species in the Southern Ocean are inherently biased owing to fishery-dependent sampling (Caccavo et al., 2021). Similarly, data on the diet of D. mawsoni are largely fishery-dependent, which means that relatively large individuals are likely to be collected as a feature of fishing gear (Ashford et al., 2005), and most are caught on continental slopes at depths greater than 600 m in accessible summer sea ice-free areas, limiting their ability to understand the diet of D. mawsoni which inhabits a wide range of depths owing to ontogenetic variation. In this regard, at depths of 300-350 m in the Cosmonaut Sea, D. mawsoni fed primarily on fishes such as Trematomus eulepidotus and Chaenodraco wilsoni (Pakhomov and Tseytlin, 1992), in the McMurdo Sound sea ice region at depths of 500-600 m they fed primarily on Pleuragramma antarcticum; and on the continental shelf of the Ross Sea, D. mawsoni fed primarily on Trematomus spp. and Channichthyidae (Parker et al., 2019). Therefore, if a study is conducted through systematic sampling in space and time, it will be possible to collect specific life cycle characteristics, such as the feeding ecology of D. mawsoni according to seasonal and ontogenetic changes, which is important for ecosystem-based management.

Antarctic toothfish play an important role in the ecosystem as apex predators by influencing the structure of prey species populations and regulating the abundance of organisms at lower trophic levels. However, in recent years, many changes in the marine ecosystem have been observed in the Southern Ocean due to the direct and indirect effects of climate change and fisheries, and in the face of these challenges, CCAMLR is working to manage the species through an ecosystem-based, precautionary approach. As generalist feeders and opportunistic predators that feed on various fish and cephalopods, depending on prey availability in the area, we expect that Antarctic toothfish will flexibly adapt to changes in food abundance that may result from future environmental changes. However, Antarctic toothfish have been reported to respond negatively to future increases in water temperature due to their low tolerance to water temperatures above 2°C (Cheung et al., 2008; Reid, 2019), and little information is available on life cycle characteristics such as reproductive ecology and early life history stages; therefore, the impact of future climate change on Antarctic toothfish stocks remains unclear (Pérez-Pezoa et al., 2023). Therefore, to preserve marine ecosystems and conduct fisheries that consider species diversity and trophic interactions, it is necessary to continue dietary studies to obtain detailed information on the food web structure by area. In this study, we determined the ecological niche and role of the Antarctic toothfish in the Bellingshausen Sea and the eastern Amundsen Sea, where ecological data on this species are scarce. This allowed us to predict future changes in ecosystems, such as the trophic cascade of prey species (Channichthyidae, Macrouridae, and Nototheniidae) in response to changes in Antarctic toothfish abundance. In addition, no significant temporal differences in the diet composition of Antarctic toothfish were observed between 2016 and 2020 in each research block in subarea 88.3, suggesting that these data could be used as a baseline for identifying future environmental changes in the region. However, these results may be subject to inherent bias due to fishery-dependent summer opportunistic sampling, and further research is needed on the food web structure of the entire region for ecosystem-based precautionary management.

The original contributions presented in the study are included in the article/ Supplementary Material . Further inquiries can be directed to the corresponding author.

Ethical approval was not required for the study involving animals in accordance with the local legislation and institutional requirements because The Antarctic toothfish individuals were collected during exploratory fisheries, which does not violate animal ethics because the stomachs of dead animals were used for the dietary analysis.

GB and SC designed this study. GS, D-GK, DK, SJ, and GB performed the experiments. GS and GB wrote the manuscript and prepared the figures and tables. GS and GB did the final editing of the manuscript. All authors discussed, read, and approved the final manuscript.