The Ross Sea region marine protected area (RSR MPA), which is a large Antarctic embayment, hosts the most productive and highest phytoplankton biomass in the Southern Ocean (Smith et al., 2003; Smith et al., 2011; Kaufman et al., 2014; Cummings et al., 2021). In the western Ross Sea, the shelf water (SW) which is mostly from the Polynya region, flows northward and then mixes with modified circumpolar deep water (MCDW), which intrudes northward into the western Ross Sea (Smith et al., 2014; Davis, 2016). As a result, nutrients are brought in by the MCDW, resulting in high levels of primary production (Smith et al., 2007; Smith et al., 2014). The Ross Sea is also an ecologically important area that supports numerous top Antarctic predators such as penguins, seals, and cetaceans (Nelson and Smith, 1986; Picco et al., 2017; Smith et al., 2017). In addition, the Ross Sea is considered a biodiversity hotspot and is one of the least human-impacted marine environments in the world because of its remoteness, extreme environments, and extensive ice cover (Smith et al., 2007; Cummings et al., 2021). The largest marine protected area (MPA) was established in the Ross Sea region in December 2017. Basic information on the ecological communities in this region is needed to better understand the ecological structure of the area; therefore, scientific surveys and monitoring programs are required (CCAMLR, 2016; Jabour and Smith, 2018; Cummings et al., 2021). We conducted several oceanographic investigations in the western RSR MPA to improve our knowledge of the mesozooplankton community.

Zooplankton is the main intermediate component in food webs connecting the primary producers with higher trophic levels (Picco et al., 2017; Smith et al., 2017; Pakhomov et al., 2020). While many studies have focused on macrozooplankton in the Ross Sea, including Euphausia superba and E. crystallorophias (Smith et al., 2003; Schmidt and Atkinson, 2016; Smith et al., 2017; Leonori et al., 2017; Vereshchaka et al., 2019; Kang et al., 2020), mesozooplankton have received relatively little attention. According to previous studies, the mesozooplankton community of the Ross Sea is dominated by copepods and the pteropods (Limacina helicina antarctica) (Hopkins, 1985; Hopkins et al., 1987; Pane et al., 2004; Elliott et al., 2009; Stevens et al., 2015; Smith et al., 2017). The dominant copepods include Metridinidae, Euchaetidae, Calanidae, Oithonidae, Clausocalanidae, and Oncaeidae (Hopkins, 1987; Pane et al., 2004; Stevens et al., 2015; Grillo et al., 2020; Smith et al., 2017). Although large species tend to be more frequent at higher latitudes, small copepods including the Oithonidae, Clausocalanidae, and Oncaeidae, are dominant in the northwestern Ross Sea (Stevens et al., 2015; Evans et al., 2020; Grillo et al., 2020). Previous surveys of the mesozooplankton community structure in relation to environmental factors in the Ross Sea have been conducted mainly in McMurdo Sound and Terra Nova Bay where major support facilities such as McMurdo Station and Mario Zucchelli Station are located (Hopkins, 1985; Carli et al., 2000; Smith et al., 2007; Elliott et al., 2009; Smith et al., 2017). Although Stevens et al. (2015) and Grillo et al. (2020) studied the copepod community in the western Ross Sea, including the open sea, they detailed the community composition of copepods and not their distribution or relationship with environmental factors. In order to build on the information provided by previous studies, the mesozooplankton community structure in the western Ross Sea still requires further study.

In this paper, we investigate the spatiao-temporal variations in the mesozooplankton community structure in the western RSR MPA from 2018 to 2020. Additionally, we examine how the mesozooplankton community is related to environmental factors based on the reported community structure. Based on these results, the interrelationship between mesozooplankton and environmental factors such as temperature or chlorophyll concentration is discussed. Transitions in the community structure in response to temporal changes are also explained in this study. In particular, the community structure in the late summer coinciding with the summer bloom and decline in the water temperature is examined in this study. Given the role of the mesozooplankton as an intermediate component in food webs, the results of this study are helpful for understanding the transfer of the trophic dynamics and the future change in the mesozooplankton community in the RSR MPA caused by global warming.

Oceanographic and biological data were collected in the western RSR MPA during three cruises on the Korean icebreaking research vessel (IBRV) Araon ( Figure 1 ). ANA08C was undertaken at twelve sampling stations from 25 February to 1 March 2018. ANA09B was conducted at 15 sampling stations from 16 to 21 January 2019. ANA10C was performed at 13 sampling stations from 14 to 31 March 2020.

Among the 2018, 2019, and 2020 surveys, the surface water temperature was highest in the 2019 survey whereas lowest in the 2020 survey. The timings of the surveys, specifically, February 2018, January 2019, and March 2020, are likely to be relevant for this finding (Smith et al., 2003; Smith et al., 2014; Cau et al., 2021). Given that MCDW is characterized by a subsurface high temperature and low dissolved oxygen, the 2019 survey’s high surface water temperature and low DO were likely influenced by the MCDW (Budillon et al., 2003). Unlike surface water temperature, average water temperature showed greater variability in 2020 than in other surveys. This is thought to be due to the sampling stations in 2020 having a wider geographic range with respect to latitude than that in other years. Regarding the Chl-a concentration, the concentrations in 2018 were about two times higher than those in other surveys. In particular, the southern stations (subgroup B2) in 2018 indicated much higher concentrations than other sampling stations. This high concentration is well corresponding with the finding of Jo et al. (2021) that much higher diatoms concentration in the southern stations. Furthermore, the nutrients such as NO₃+NO₂, PO₄, and SiO₂ were negatively correlated with B2 in the PCoA plot. This negative correlation was related to the phytoplankton biomass because these nutrients can be removed from the environment by autotrophic assemblages (Nelson and Smith, 1986; Smith et al., 2003). Taken all together, given summer bloom can occur in February by diatoms and the timing of the bloom is very consistent, the high Chl-a concentration in 2018 could be related to the summer bloom by diatoms (Edwards and Richardson, 2004; Peloquin and Smith, 2007; Smith et al., 2011; Kaufman et al., 2014; Smith et al., 2014).

Considering the earlier studies covered different regions and employed different methodologies, it is difficult to compare our data that from prior studies in the Ross Sea (Stevens et al., 2015; Smith et al., 2017). However, when a small mesh plankton net (200 µm) was used instead of a coarse plankton net (over 330 µm), the amount of the mesozooplankton was higher (Pane et al., 2004; Elliott et al., 2009; Stevens et al., 2015; Smith et al., 2017). Similar to findings in prior studies, the abundance recorded in this study using a coarse mesh-sized plankton net was lower than that reported in the survey using a fine mesh-sized plankton net (Pane et al., 2004; Stevens et al., 2015; Smith et al., 2017). The mesozooplanklton community, however, was dominated by small species (e.g. Oithona spp., Ctenocalanus sp., etc.). The composition of the mesozooplankton community was similar to that in previous studies (Pane et al., 2004; Stevens et al., 2015; Grillo et al., 2020). Therefore, the mesozooplankton sampling region and methodology are regarded as significant factors influencing the survey’s conclusion (Stevens et al., 2015; Smith et al., 2017).

Given the mesozooplankton community of the present study, our analyses of data from 2018, 2019, and 2020 surveys showed that the mesozooplankton community in the western RSR MPA could be separated by spatio-temporal variability. Taxa observed in 2019 and 2020 could be divided into four and five significantly different groups, respectively. Although there were no significantly different groups among the taxa observed in 2018, the community structure could be distinguished by the feeding strategy of the dominant species. Additionally, the subgroups of 2018 and groups of 2019 and 2020 showed significant differences according to environmental factors.

In 2018, Euphausia spp., P. antarctica, Sagitta spp., and unidentified Ostracoda were strongly correlated with subgroup B1, whereas C. acutus, Ctenocalanus sp., M. gerlachei, Oithona spp., Oncaea spp., and Polychaeta larvae were strongly correlated with subgroup B2. The dominant species in B1 had omnivorous or carnivorous feeding modes, whereas the dominant species in B2 were mostly herbivorous (Hopkins, 1987; Lopez and Huntley, 1995; Pasternak and Schnack-Schiel, 2001; Bocher et al., 2002; Ward and Hirst, 2007; Elliott et al., 2009; Grillo et al., 2020). According to the LINKTREE analysis, the stations included in subgroup B2 had much higher Chl-a _total, Chl-a _micro, and Chl-a _nano concentrations than those included in subgroup B1. In accordance with the previous studies and the present study, the relatively high Chl-a _micro concentration of subgroup B2 recorded in 2018 was likely to result of summer blooms by diatoms (Peloquin and Smith, 2007; Smith et al., 2011; Kaufman et al., 2014; Smith et al., 2014; Jo et al., 2021; Saggiomo et al., 2021). As a result, herbivorous zooplankton increased in B2, while B1 was characterized by relatively high abundances of omnivorous or carnivorous species. Thus, the Chl-a concentration gradients throughout the region could be considered the main factor regulating the mesozooplankton community in 2018. This finding is consistent with the previous reports, which found that foods such as Chl-a are the main factors regulating mesozooplankton distribution (Liu and Dagg, 2003; Chen et al., 2017; Smith et al., 2017).

In the 2019 survey, Ctenocalanus sp. and Oithona spp., and Oncaea spp. were correlated with group A; C. acutus, M. gerlachei, P. antarctica, Sagitta spp., and Unidentified Ostracoda were positively associated with group C; and S. thompsoni and Unidentified cnidarians 3 were positively correlated with group D. Even though group D had a much lower abundance than the other groups, the SIMPER analysis indicated that the abundances of S. thompsoni and Unidentified cnidarians 3 were higher than the abundances observed in the other groups. Small species such as Ctenocalanus sp. and Oithona spp. were dominant in group A, whereas relatively large species such as C. acutus and M. gerlachei were more common in group C. In other words, small species were common in group A, while relatively larger species were more common in group C. Regarding environmental factors, group D was separated from the other groups at a much lower temperature. Group A was separated from group C by a lower NH₄ concentration. The food resources (e.g., Chl-a) did not have a significant relationship with the groups in 2019, unlike 2018. Taken all together, these results indicate that the dietary components, such as Chl-a, did not affect the mesozooplankton community structure in 2019. Considering that MCDW intrudes over the Ross Sea at the beginning of January and that water temperature in which group D was found was significantly lower than that of the water in which the other groups were found, it is possible that group D was disturbed by aperiodic intrusion (Smith et al., 2007; Castagno et al., 2017). As a result, the very low abundance at the corresponding stations was confirmed to distinguish them from other groups. Additionally, Biggs (1982) report that ammonium excretion is highly correlated with zooplankton body size corresponds well with the discrimination between groups A and C according to the body size of the dominant species. Large herbivorous species, C. acutus, and large carnivorous taxa such as P. antarctica and Sagitta spp. represented a relatively higher abundance in group C than in group A, whereas small omnivorous species such as Oithona spp. were more abundant in group A than in group C. In summary, regardless of the feeding strategy, the large species were more abundant in group C, whereas relatively small species were common in group A. This differentiation based on size would be related to the predation of small species by large species because trophic processes in the plankton community are highly dependent on the food size (Nelson and Smith, 1986; Lancelot et al., 1993; Hansen et al., 1997; Bocher et al., 2002; Tang et al., 2008; Chen et al., 2017; Sun et al., 2021).

In 2020, there have been more divided groups than in other surveys. This result was thought to be that the 2020 survey included open sea and Terra Nova Bay stations which were not contained in the 2018 and 2019 surveys. Groups B and C, which were mostly located in the open western Ross Sea, had similar species compositions. The two groups shared the most abundant species, but the ranking of those species differed. Groups D and E, which were mainly located on the continental shelf of the western Ross Sea, had similar species compositions; however, group D had a much higher abundance than group E. According to LINKTREE analysis, groups D and E could be separated from groups B and C by having a lower Temp_average. This distinction is likely a result of the geographic distributions of these groups since groups B and C can be separated from groups D and E according to their latitudinal location. Consequently, the Temp_average mainly determined these significant differences between the open sea and continental shelf regions of the western Ross Sea in 2020.

When comparing of the mesozooplankton community structures according to the interannual variations ( Figure 8 ), this comparison can be summarized in a few points: (1) the sampling stations were well clustered according to the survey periods; (2) the stations of 2018 clustered into a single group; (3) the stations in 2019 were separated into two groups, and one of the groups was more similar to the group including the 2018 stations than to another group of 2019; (4) the Terra Nova Bay stations were clustered far from other stations; (5) the stations located in the open Ross Sea in 2020 were much more similar to the stations included in 2018 than to the stations in 2020 on the continental shelf. PERMANOVA indicated that there were significant differences between the groups (pseudo-F = 10.367, P < 0.05). These significant differences were the same in the comparison between the stations excluding open sea stations (pseudo-F = 9.4362, P < 0.05). By the dominant taxa, Oithona spp., which are known to be affected by water temperature, showed a wide distribution range according to the surveys in this study (Ward and Hirst, 2007). Specifically, this species showed a strong correlation with high-latitude stations in 2018 and 2019. However, Oithona spp. was strongly correlated with the open sea, including the low-latitude stations in 2020. Limacina helicina antarctica, which has a high abundance in the southern Ross Sea, had a significantly high abundance at the continental shelf stations in 2020, especially near Terra Nova Bay (Hopkins, 1987; Pasternak and Schnack-Schiel, 2007; Elliott et al., 2009). Furthermore, L. helicina antarctica showed relatively low abundance at the stations included in 2019, which indicated comparatively warm temperatures, whereas its abundance in 2018 was intermediate among the three surveys. The greatest abundance of Ctenocalanus sp., which prefers cold water, was found from February to April namely in 2018 and 2020 surveys, while this species showed relatively low abundance in 2019 (Pasternak and Schnack-Schiel, 2007).

Taken together, these results indicate that the mesozooplankton community of the western RSR MPA was altered according to the survey period and differed spatiotemporally. Considering the food resource such as the Chl-a was among the best explanatory factors only in 2018, but not in 2019 and 2020, and the mean total abundance in the 2018 survey is higher than other surveys except for the Terra Nova Bay region, the relatively high Chl-a concentration in late summer induced high mesozooplankton abundance in the western Ross Sea. In other words, if prey, such as phytoplankton or diatoms, were abundant (2018), then the community was influenced by the prey distribution. However, when the abundance of phytoplankton had less influence (2019), the zooplankton community structure was affected by aperiodic intrusions or predation pressure that induced trophic cascades. Predation pressure causing trophic cascades rather than abiotic factors was assumed to be the key factor determining the mesozooplankton community structure in 2019. As the water temperature decreased with latitude and sea ice covered the whole continental shelf (2020), the mesozooplankton community of the western RSR MPA was replaced with species suitable for lower water temperatures. Although the correlation was not significant at the 0.05 level, the Chl-a concentration and mesozooplankton abundance indicated a negative correlation in 2020 (r = -0.547, P > 0.05). Considering the spatial mismatch between phytoplankton and mesozooplankton means both populations asynchrony, this negative correlation is regarded as a typical spatial mismatch between phytoplankton and zooplankton (Edwards and Richardson, 2004; Durant et al., 2013; Mcginty et al., 2014; Sun et al., 2021). This spatial mismatch would be related with the increase of the species that prefer cold water such as L. helicina antarctica which shows a significantly high abundance at the continental shelf stations in 2020.

The results of this study suggest that the mesozooplankton community of the western RSR MPA was influenced by prey if food resources were abundant but affected by predation pressure or environmental factors when food resources were less influential. Therefore, the occurrence of the summer bloom in late summer, mainly in February, is likely to affect the mesozooplankton community in the western RSR MPA. The continental shelf community present in January and February was replaced by species that preferred cold water by March, as the water temperature decreased in the western RSR MPA. These results suggest that climate change will affect the mesozooplankton community of the western RSR MPA in the late summer, because climate change could result in warmer water temperatures and the expansion of sea ice-free periods in the RSR MPA, although the concentration of sea ice has recently increased (Smith et al., 2007; Smith et al., 2014). As a result, the distributions of the species that are mainly found in the western RSR MPA could expand southward. The distribution of the species that prefer cold water, such as L. helicina antarctica and Ctenocalanus sp., could be limited in the southern region. Eventually, the mesozooplankton community currently found in the western RSR MPA would move southward, and a novel zooplankton community could invade the RSR MPA. Considering that the Ross Sea, which is the largest MPA, is an essential feeding area of higher predators such as penguins and whales, these changes in the mesozooplankton community could impact the predation abilities of higher predators, for which zooplankton are the main food resources (Nelson and Smith, 1986; Smith et al., 2012; Smith et al., 2014; Picco et al., 2017; Smith et al., 2017). However, the three surveys in this study were conducted in different months in 2018, 2019, and 2020. Interannual variations can be substantial within the Southern Ocean, and it is possible that these interannual differences were overlooked in this study (Smith et al., 2003; Smith et al., 2014; Cau et al., 2021). Furthermore, the Ross Sea ecosystem has been dramatically influenced by climate change (Smith et al., 2014). Thus, further studies and constant monitoring of the mesozooplankton community in the RSR MPA are required.

The original contributions presented in the study are included in the article/ Supplementary Material . The zooplankton and environmental data are available at the Korea Polar Data Center (https://dx.doi.org/doi:10.22663/KOPRI-KPDC-00001925.1). Further inquiries can be directed to the corresponding author.

HL conceived of the study. BK, BL, NJ, WS, and HL collected and analyzed the environmental data. SK analyzed zooplankton community composition and wrote the manuscript. SK, BK, SL, S-YH, J-HK, and HL contributed to the study design and data discussion. All authors contributed to the article and approved the submitted version.

This study was a part of a project titled “Ecosystem Structure and Function of Marine Protected Area (MPA) in Antarctica” project (PM21060) funded by Ministry of Oceans and Fisheries, Korea (20170336).

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.