Global warming caused by the Industrial Revolution has severely impacted terrestrial and marine environments, resulting in serious socio-economic and political issues (IPCC, 2013). Hypoxia caused by global warming (dissolved oxygen concentration of 2 mg L^-1 or less) is one of the most widespread harmful anthropogenic effects on coasts and estuaries (Diaz and Rosenberg, 2008), and is classified as a major global environmental problem, along with ocean acidification, overfishing, and habitat loss (Breitburg et al., 2018). In marine ecosystems, climate change induces an increase in sea levels and water temperature, as well as changes in precipitation and wind patterns, driving environmental changes in coasts and estuaries (Altieri and Gedan, 2015). Additionally, a growing body of evidence has demonstrated that global warming induces hypoxia in coastal regions and estuaries (Bendtsen and Hansen, 2013; Carstensen et al., 2014; Jenny et al., 2014; Altieri and Gedan, 2015).

Dissolved oxygen changes in the ocean are mainly driven by factors such as decreased oxygen solubility, increased stratification, and increased biological respiration (Keeling et al., 2010). Additionally, the increase in hypoxic regions over the past half century is largely due to excessive anthropogenic nutrient loading, which stimulates eutrophication (Kemp et al., 2005). This increases the supply of organic matter to deeper water layers and sediments, causing low oxygenation in the bottom water and resulting in the occurrence of ‘dead zones’ (Diaz and Rosenberg, 2008; Conley et al., 2009). The area range of coastal hypoxia has increased significantly in recent decades due to the increase in coastal eutrophication (Diaz, 2001; Diaz and Rosenberg, 2008; Conley et al., 2009; Rabalais et al., 2010). An assessment of the literature shows that the number of coastal areas with reported hypoxia have increased over time at an exponential growth rate of 5.5% per year (Vaquer-Sunyer and Duarte, 2008). This growth rate may be partly due to increased observational efforts, but the prevalence of hypoxia is expected to increase further due to the combined effects of eutrophication and climate change (Conley et al., 2009).

Hypoxia can be a serious stressor for marine organisms and ecosystems, resulting in community structure changes through increased dominance of less sensitive species or a decrease in the growth of sensitive species (Marcus et al., 2004; Richmond et al., 2006; Roman et al., 2019; Keister et al., 2020). Particularly, mesozooplankton plays a pivotal role in the food web and is a major component of the biological carbon pump. Hypoxia can directly or indirectly affect mesozooplankton community structure, and the tolerance of mesozooplankton to hypoxia varies in a taxon- and stage-dependent manner (Roman et al., 1993; Marcus, 2001; Richmond et al., 2006; Grodzins et al., 2016; Pan et al., 2018). Several studies have focused on the effects of zooplankton in relation to hypoxia. Elliott et al. (2012) found that the abundance of zooplankton was low in the hypoxic bottom waters, but abundant in the upper normoxic waters. Roman et al. (2012) noted that zooplankton shifted the vertical distribution to the upper water column during the day in the Gulf of Mexico where bottom hypoxia occurs. Shi et al. (2019) showed that in regions where hypoxia was observed, the zooplankton biomass was relatively high and the taxonomic diversity was lower. Slater et al. (2020) showed that hypoxia clearly contributed to a decrease in copepods and planktivorous fish, and an increase in gelatinous zooplankton. However, among the various taxonomic groups experiencing hypoxia, some species may respond strongly to these conditions, but the overall effect on the communities is less clear. Therefore, more research is clearly needed to understand the relationship between hypoxia and the various taxonomic groups that inhabit coastal areas.

Mesozooplankton abundance tends to fluctuate greatly within a short period due to the temporal differences in the life strategies of each taxon (Klais et al., 2016). For example, the population of copepods may increase in weeks or months (Hansen et al., 2006), whereas the population of cladocerans may increase in days or weeks (Egloff et al., 1997; Park and Marshall, 2000). Particularly, cladocerans have a very short generation time compared to copepods, and therefore determining their seasonal abundance requires an in-depth understanding of their reproductive characteristics through frequent sampling (Omorii and Ikeda, 1984; Mullin and Onbé, 1992; Atienza et al., 2008). Therefore, to quantitatively evaluate mesozooplankton populations and communities, the survey period and interval must be considered by reflecting the reproductive characteristics of taxonomic groups (Omorii and Ikeda, 1984; Klais et al., 2016).

Keister et al. (2020) reported that the two most common effects of hypoxia in the Hood Canal (Washington, USA) were lower total abundance and biomass of zooplankton and higher abundance of gelatinous zooplankton. However, the hypothesis that hypoxia leads to a dominance of cyclopoid copepods relative to calanoid copepods was not clearly demonstrated. In Masan Bay, Korea, where hypoxia commonly occurs in summer, a cladoceran species was positively correlated with low DO, suggesting that hypoxia could have an effect on the population growth of cladocerans (Jang et al., 2015). Moreover, copepods and cladocerans predominate in eutrophic coastal and estuarine environments depending on the region (Soh et al., 2002; Isinibilir et al., 2008; Isinibilir, 2009; Rakhesh et al., 2013; Li et al., 2022). However, despite their high abundance, few studies have investigated the effect of hypoxia on copepods and cladocerans.

Gamak Bay has been the subject of intensive research in relation to hypoxia in recent decades (Kim et al., 2010; Seo et al., 2012; Shin et al., 2013; Lee et al., 2016; Lee et al., 2019). Unlike detailed studies focusing on the response and abundance of taxonomic groups of mesozooplankton according to hypoxia, very few studies have characterized the relationship between mesozooplankton and hypoxia in Gamak Bay. Moon et al. (2006) demonstrated that mesozooplankton abundance decreased significantly when hypoxia occurred in summer, with the highest abundances occurring in the surface layer. However, the effects of hypoxia on specific taxonomic groups remains unknown. Therefore, our study focused on the composition of mesozooplankton populations (mainly cladocerans and copepods), abundance, and occurrence patterns that respond rapidly to changes in the environment. Furthermore, to determine whether hypoxia decreases or increases the abundance of mesozooplankton and changes its community structure, samples were collected from the inner bay (hypoxic zone) and middle bay (normoxic zone) areas of the study site and then compared. The specific goals of this study were to evaluate the effect of hypoxia on (1) the composition of the mesozooplankton community, (2) the abundance of mesozooplankton, and (3) the occurrence pattern of the dominant species in the hypoxic period.

The unique marine environment of Gamak Bay caused a decrease in DO concentration in summer and, particularly, an oxygen deficiency in the bottom of the inner Gamak Bay, which affected the mesozooplankton community structure. Previous studies have shown that zooplankton exhibits multiple responses when exposed to hypoxia (Marcus et al., 2004; Ekau et al., 2010; Kimmel et al., 2010; Elliott et al., 2013; Keister et al., 2020; He et al., 2021). Hypoxia not only causes the direct mortality of mesozooplankton but also negatively affects metabolism, productivity, growth, vertical migration, and distribution (Auel and Verheye, 2007). However, previous studies have demonstrated that many zooplankton prefer hypoxia as a refuge to tolerate hypoxia or avoid predators, either through vertical or seasonal migration (Ludsin et al., 2009).

Although many approaches to zooplankton collection in situ have been performed with respect to hypoxia, none have been completely satisfactory and their uncritical use may confound the interpretation of the results (Roman et al., 1993; Moon et al., 2006; Jang et al., 2015). In this study, the vertical hauls sampling of mesozooplankton from the bottom to the surface using a conical net without a closing device has methodological problems (avoidance or depth of distribution of mesozooplankton) in understanding the effect of hypoxia (Roman et al., 1993; Jang et al.,2015). Nevertheless, this method can directly and conveniently collect mesozooplankton in coastal waters. Also, direct death under hypoxic conditions will only occur if the organism is unable to move. The apparent avoidance behavior of zooplankton in hypoxic conditions may also respond differently depending on region and population. In the laboratory, the Acartia tonsa population from the Chesapeake Bay exhibited hypoxia avoidance behavior (Stalder and Marcus, 1997), and the A. tonsa population from the Gulf of Mexico did not avoid hypoxia at low 0.1 mg L^-1 (Decker et al., 2003).

In this study, mesozooplankton abundance was higher in the inner bay than in the middle bay during the hypoxic period. This may be due to the high primary production in the inner bay, which increases the grazing and productivity of the zooplankton (Shi et al., 2019). During the hypoxic period in Gamak Bay, phytoplankton is largely generated by abundant nutrients under the thermocline and sufficient solar radiation (Kim et al., 2006). In this study, Chl-a concentration peaked in the inner bay during the hypoxic period. According to previous studies, the response of the zooplankton population was expected to increase with functional grazing and reproductive capacity as primary production increased (Blaxter et al., 1998; Tan et al., 2004). Therefore, the influx of nutrients causing hypoxia can indirectly improve mesozooplankton abundance through an increase in food supply (Shi et al., 2019).

In this study, a total of 45 and 52 taxa were identified in the inner and middle bay, respectively. Water temperature and salinity are important factors that determine the spatial and temporal distribution of zooplankton (Froneman, 2004; Shin et al., 2022). In the inner bay, most of the population was due to the active reproductive ability of cladocerans in summer when the water temperature was high, and copepods dominated in winter. In spring and autumn, copepods, planktonic larvae, and other zooplankton appeared to form the mesozooplankton community. In the middle bay, a higher proportion of copepods appeared compared to the inner bay. During the hypoxic period, cladocerans showed a high composition in July and August in the inner bay, whereas they were higher after September in the middle bay. Our data confirmed that the composition of mesozooplankton species, abundance, and spatiotemporal distribution was different due to differences in the tolerance of taxa in the hypoxic period. Cladocerans may exhibit preferential growth compared to copepods at the inner station of Gamak Bay, where hypoxia frequently occurs in summer. When dissolved oxygen levels are between the minimum and critical levels, copepod growth, development, physiology, and behavior change with dissolved oxygen levels (Brett, 1979; Dinh et al., 2020). Copepod survival and egg production are significantly lower at lower DO concentrations (Lutz et al., 1992; Roman et al., 1993; Keister et al., 2000: Marcus et al., 2004; Sedlacek and Marcus, 2005). The ability of cladocerans to synthesize hemoglobin is common, and this ability has been suggested as an adaptation to an unpredictable environment in which oxygen concentrations may abruptly drop to fatally low levels (Heisey and Porter, 1977; Weider and Lampert, 1985).

Copepods and cladocerans that generally form their populations through dormant eggs, but they have different life history strategies. Most cladocerans can reproduce rapidly through parthenogenesis during periods of favorable conditions (Egloff et al., 1997; Uye et al., 2000). On the other hand, copepods exhibit sexual reproduction, and must undergo a series of developmental stages to reach the adult stage (Dussart and Defaye, 1995). The temporal difference in the life history of these two taxa is an important factor in maintaining adaption and population under favorable or unfavorable conditions.

One of the major characteristics of the seasonal dynamics of zooplankton in temperate waters is their dormancy during their life cycle. Particularly, copepods and cladocerans produce dormant eggs to survive in adverse environmental conditions (Onbé et al., 1996; Marcus, 1996). These dormant eggs go through a refractory period and when favorable conditions are established, they hatch temporarily and join the population. That is, repeated generation maintenance strategies are taken (Marazzo and Valentin, 2004; Choi et al., 2021). In this study, the rapidly increased mesozooplankton abundance in the inner bay in summer and autumn was likely due to the hatching of dormant eggs and provide an environment for temporary mass reproduction. However, dormant eggs in the field were not investigated, so it cannot be predicted easily. Nevertheless, the dormant eggs of copepods and cladocerans clearly hatched at the inner bay, which have been hypoxic for a considerable time, making significant contributions to the food web of the Gamak Bay ecosystem.

In this study, the total mesozooplankton abundance in the inner bay was mainly contributed by cladocerans during the hypoxic period. In the inner bay, cladocerans appeared from May, when the water temperature gradually increased, and the number of more than 5,000 indiv. m^-3 continued to appear throughout August when the average water temperature was 25.6–26.5 °C and the average salinity was 29.3–31. After that, the water temperature gradually decreased and the salinity increased, so they did not appear. In this study, three species of cladocerans dominated at different times (coexisting and separating), and the order of appearance was P. polyphemoides, P. tergestina, and P. avirostris. The optimal environmental conditions for cladocerans are known to vary by region, and this separation is often explained by the optimal temperature and salinity conditions for each species (Egloff et al., 1997). P. polyphemoides appeared in a wider water temperature and salinity range than the other two species. Although P. tergestina and P. avirostris had similar preferred salinity ranges, P. tergestina had a wider range of water temperature than P. avirostris. Therefore, we concluded that the three species of cladocerans were able to maintain a high population by separating or coexisting in the inner bay because each species had different ranges of preferred water temperature and salinity according to their physiological characteristics.

Cladocerans responded to low oxygen conditions depending on their size. Karpowicz et al. (2020) found that small freshwater rotifers and small freshwater cladocerans were more resistant to hypoxia than larger freshwater cladocerans, and that this oxygen stress is one of the most important factors to alter the structure of zooplankton in lakes. In this study, cladocerans were far more dominant than copepods through their rapid reproductive capacity during the hypoxic period in the inner bay. The order of appearance was P. polyphemoides, P. tergestina, and P. avirostris, and the earlier the species appeared, the longer it appeared. P. polyphemoides appeared when the bottom DO concentration was 5.2 mg L^-1, and peaked when a low bottom DO concentration of 1.2 to 1.5 mg L^-1 was formed. When the hypoxia had temporarily disappeared, they did not appear. Even after hypoxia had completely disappeared, they also appeared at a DO concentration of 6 mg L^-1. P. tergestina began to appear at the end of June when hypoxia first occurred, and the abundance increased even when hypoxia had temporarily disappeared. After that, the maximum abundance was shown in the hypoxic condition of 1.9 mg L^-1, and when the hypoxia had disappeared, and decreased rapidly and appeared in small numbers until October. P. avirostris appeared in mid-July, when the hypoxia had temporarily disappeared, and the maximum abundance was in mid-August, when the bottom DO concentration was very low at 0.7 mg L^-1. After that, they did not appear even though the hypoxia was not over. In this study, the average body length of the three species was approximately 0.65 mm for P. polyphemoides, 1.3 mm for P. tergestina, and 1.4 mm for P. avirostris, and the later appearance period, the larger the species appeared. Therefore, it is believed that P. polyphemoides, which are the smallest in size, tolerated the hypoxic period for the longest time, and the P. avirostris, which are relatively large in size, are not likely to survive the hypoxic period for long.

Copepods, cladocerans and planktonic larvae are the most important mesozooplankton taxa in Gamak Bay (Soh et al., 2002; Choi, 2020). Katechakis et al. (2004) suggested that copepods, crustaceans, and dolioida could coexist in summer due to different food selectivity and ingestion and assimilation rates. Acartia clausi reached the highest grazing coefficient for large algae over 70 μm, and P. avirostris covered nanoflagellates and ciliates ranging from 2.5–100 μm. It is a known fact that cladocerans are mostly filter feeders with strong feeding preferences for bacteria, microzooplankton and certain phytoplankton species. The feeding pattern of Podonidae (P. polyphemoides and P. tergestina), which appeared dominant in this study, is known as a raptorial feeder (Jagger et al., 1988), they prey largely on diatoms (Jagger et al., 1988; Kim et al., 1989) and potentially on flagellates, including heterotrophs and autotrophs (Turner et al., 1988). P. avirostris is also known as a filter feeder, feeding on bacteria and nanoflagellates (Paffenhöfer and Orcutt, 1986; Turner et al., 1988) and larger diatoms (Kim et al., 1989). In this study, the reason cladocerans can dominate the mesozooplankton community in Gamak Bay may be due to their broader feeding efficiency. However, only three sizes of phytoplankton were considered in this study, and this may not be sufficient to explain the observed variability.

The RDA results indicated that the dominant species showed a significant correlation with water temperature, salinity, DO concentration, total Chl-a concentration, and size-fractionated Chl-a concentration among environmental variables. In the inner bay, P. polyphemoides showed a positive correlation with water temperature and the < 3 µm categories of Chl-a concentration, and a negative correlation with bottom DO concentration. P. tergestina, P. avirostris, and A. sinjiensis showed a positive correlation with water temperature and the 3–20 µm categories of Chl-a concentration and a negative correlation with salinity. In the middle bay, cladocerans showed a positive correlation with water temperature and a negative correlation with surface and bottom DO concentration, A. sinjiensis showed a positive correlation with water temperature and the < 3 µm categories of Chl-a concentration. In general, in temperate waters with distinct seasonal changes, mesozooplankton is more affected by water temperature and salinity. Therefore, the results of this study are considered valid. Furthermore, hypoxic environments may play a role in reducing zooplankton abundance in eutrophic systems (Park and Marshall, 2000). However, several studies have also shown that zooplankton can exhibit high abundance in hypoxic environments (Roman et al., 2019; Karpowicz et al., 2020). As a result, the environmental factors (non-oxygen) that affect the dominant species in the two regions are different, but considering the characteristics of the sea area where seawater flow was low and high phytoplankton proliferated, our findings indicated that an appropriate habitat for dominant species had been created.

The fluctuations of zooplankton communities in coastal areas where hypoxia occurs frequently depend on the tolerance of taxa and seasonal fluctuations. In this study, we observed the effect of hypoxia on zooplankton community distribution. During the hypoxic period, cladocerans were found to be more tolerant than copepods in the inner bay where hypoxia is formed. Most of them experience a strong hypoxic state rather than an anaerobic state over the years, eventually changing the structure and function of the ecosystem. However, we did not find clear evidence that cladocerans and copepods could advance their appearance according to space in the hypoxia period. These results provide important information for understanding the response of mesozooplankton under hypoxic conditions as hypoxia gradually intensifies due to climate change or environmental pollution in the future.

The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation.

EHL conceived and designed the experiments, performed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. SYC conceived and designed the experiments, performed the experiments, analyzed the data, authored or reviewed drafts of the paper, and approved the final draft. MHS conceived and performed the experiments, analyzed the data, prepared figures and/or tables, authored or reviewed drafts of the paper, and approved the final draft. HYS contributed to funding acquisition, conceptualization of the experiments, data interpretation and discussion, authored or reviewed drafts of the paper, and approved the final draft. All authors contributed to the article and approved the submitted version.

This research was part of a project titled “Research center for fishery resource management based on the information and communication technology” (2022, grant number 20180384), funded by the Ministry of Oceans and Fisheries, Korea, and by the National Marine Biodiversity Institute of Korea (grant number 2022M01100).

We thank the reviewers for their valuable comments and suggestions on the manuscript.

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.