Introduction

The unprecedented pace of ocean warming in regions of the Southern Ocean poses risks to notothenioid fishes (Mintenbeck et al. 2012; Beers and Jayasundara 2015), which are regarded as the most stenothermal group of marine fishes globally (Dahlke et al. 2020). The five notothenioid families that are endemic to the Southern Ocean, collectively forming the Cryonotothenioidei, are also important members of Antarctic coastal and shelf ecosystems, dominating fish diversity and accounting for nearly 90% of the fish biomass in the region (Gon and Heemstra 1990; Eastman 2005). During the latter half of the twentieth century, mid-depth Southern Ocean temperatures rose on average 0.17°C (Gille 2002), with sea surface temperature (SST) along the highly impacted western Antarctic Peninsula (WAP) warming over 1 °C (Meredith and King 2005). Precise estimates of future warming scenarios in the Southern Ocean are currently limited by interdecadal variability, geographically asymmetrical responses, and data paucity (Kennicutt et al. 2019). Despite this uncertainty, the Southern Ocean will likely continue to warm in the coming decades, primarily due to anthropogenic influences (Rintoul et al. 2018; Swart et al. 2018). In order to understand the effects of this warming, it is important to determine the thermal tolerance of cryonotothenioids at different life stages to better predict how these fish populations may respond to current and future environmental change.

Most research on cryonotothenioid thermal tolerance is focused on adult notothenioids from the families Nototheniidae, Channichthyidae, and Bathydraconidae (Bilyk and DeVries 2011; Beers and Sidell 2011; Bilyk et al. 2012). Initial studies suggested that these fishes have remarkably low Upper Incipient Lethal Temperatures, ranging from 5 to 7 °C in Pagothenia borchgrevinki, Trematomus bernacchii, and Trematomus hansoni (Somero and DeVries 1967). Subsequent studies suggest that nototheniid thermal tolerance may be influenced by habitat, with nototheniids inhabiting the northerly Seasonal Pack Ice Zone (SPZ) having higher thermal tolerance than closely related species inhabiting the more southerly, cooler High-Antarctic Zone (HAZ) (Bilyk and DeVries 2011). During austral summer, shallow waters (< 20 m) of the SPZ can reach temperatures exceeding 2 °C for nearly a month (Cárdenas et al. 2018), while shallow-water temperatures rarely exceed − 0.5 °C in the HAZ (Hunt et al. 2003; Bilyk and DeVries 2011; Cziko et al. 2014). Several nototheniids from SPZ areas show a higher capacity for metabolic acclimation under warming conditions (Todgham and Mandic 2020).

Studies on the thermal tolerance and physiology of larval and juvenile cryonotothenioids are limited to just two Antarctic nototheniid species (Evans et al. 2012; Davis et al. 2018; Naslund et al. 2021) and one bathydraconid species (Flynn et al. 2015; Flynn and Todgham 2018), all of which were collected from the McMurdo Sound HAZ region. We summarize current information on thermal tolerance for juvenile, larval, and egg stages of cryonotothenioids and one sub-Antarctic larval notothenioid (Bruno et al. 2022) in Table 1.

Table 1 Prior studies on the thermal tolerance of cryonotothenioid fishes during early life stages
Full size table

A method commonly used to describe the thermal tolerance of adult cryonotothenioid fishes is the difference between the critical thermal minimum (CTmin) and critical thermal maximum (CTmax), which are the temperatures below and above, respectively, at which an individual loses its ability to swim (Lutterschmidt and Hutchison 1997; Beers and Jayasundara 2015). The temperature that cryonotothenioids freeze, which ranges from − 1 to − 2.2 °C depending on species and life stage (Cziko et al. 2006), is defined as their CTmin (Bilyk and DeVries 2011). To determine CTmax, fishes are exposed to rising temperatures in aquaria at an acute, fixed rate, usually 0.3 °C min−1 based on the recommendation of Becker and Genoway (1979), and monitored for behavioral changes. The temperature at which a fish cannot maintain its ability to right itself, an indicator of loss of swimming capability, is recorded as the CTmax (Morgan et al. 2018).

In this study, we follow the methods of Bilyk and DeVries (2011) to determine the CTmax of larvae of several species of cryonotothenioids. These results are compared to the CTmax of adults determined by Bilyk & DeVries (2011). We aim to determine if larval cryonotothenioids exhibit lower thermal tolerances compared to adult stages and if CTmax correlates with regional ocean temperatures. We then discuss potential future research priorities to better understand the thermal limits and physiology of several groups of cryonotothenioids.

Materials and methods

Collection of larval fishes

Larval cryonotothenioids were collected during January 2020 and January 2023 aboard the ASRV Laurence M. Gould and November–December 2021 aboard the ASRV Nathaniel B. Palmer as part of the Palmer Antarctica Long-Term Ecological Research (Palmer LTER) program. Sampling was conducted in the Bellingshausen Sea along the western Antarctic Peninsula (WAP) continental shelf (Fig. 1a). Larvae were captured opportunistically from epipelagic oblique net tows targeting zooplankton, using a 2 m2 frame (700 µm mesh; sampling depth interval 0–120 m) or 1 m2 frame (333 µm mesh; 0–300 m) Metro net (Steinberg et al. 2015). Tow contents were immediately and gently transferred from the cod end to a large holding tank, where larval cryonotothenioids were identified and evaluated for condition. Post-flexion larvae swimming normally with no signs of injury during capture were chosen either for control or CTmax trials.

Fig. 1
figure 1

Larval fish collection sites and critical thermal maxima (CTmax). a Map of the western Antarctic Peninsula (WAP) region indicating where larvae were collected for experiments during austral summer research cruises from 2020 to 2023. The point color corresponds to the key in 1b. All artedidraconids and one C. antarcticus were collected in the same region in Dalgleish Bay over two cruises (2020 and 2023) and are represented by a tricolor circle. b Cryonotothenioid CTmax values with means (black diamonds) and standard deviation (vertical black lines). Each circle represents an individual experimental fish, while the color corresponds to the species in the key below

Full size image

CTmax and control trials

Selected larvae were transferred to 11.4-L polycarbonate tanks filled with seawater supplied by a hull-mounted flow-through system. Tanks were aerated with an air pump to maintain high oxygenation and prevent thermal stratification. The control tank was incubated in a flow-through seawater bath to maintain ambient surface seawater temperature (− 1 to 3 °C). Larval behavior was monitored in the control tanks to ensure the results of the CTmax trials were not artifacts of trauma sustained during collection.

Fishes in the CTmax tank were immediately warmed at a fixed rate of 0.3 °C min−1. Moyano et al. (2017) suggested that a slower heating rate is potentially more suitable for larval fishes. However, the primary goal of our study was to compare the CTmax of larvae with adults reported in Bilyk & DeVries (2011); therefore, we chose to replicate their heating rate of 0.3°C min−1, as changes in heating rate can impact CTmax in notothenioids (Peck et al. 2014). Water temperature in the control and experimental tanks was recorded using a NIST-traceable thermistor thermometer (Digi-Sense, USA).

Larvae were monitored continuously for the duration of each CTmax trial. Consistent heating continued until a fish was unable to right itself for one minute after rolling on its side, indicating a persistent loss of equilibrium (Bilyk and DeVries 2011). After the CTmax temperature was recorded, larvae were transferred to a tank held at the environmental temperature, where their recovery was evaluated. Larvae were then measured (standard length [SL] mm), euthanized by overdose of methanesulfonate (MS-222), preserved in 95% ethanol, and cataloged into the Virginia Institute of Marine Science (VIMS) Nunnally Ichthyology Collection.

We used dot plots (with mean and standard deviation) to illustrate and interpret the CTmax data. Larvae with similar characteristics were grouped to analyze the correlations between CTmax and SST at collection location (determined by an SBE38 Digital Thermometer) and standard length using linear regressions. All plots were created in R (R Core Team 2022). We also plot collection sites for larvae along the WAP using ArcGIS Pro Version 3.1.1. (Environmental Systems Research Institute Inc.).

Ethical approval

This research involved contact with live animals. All experiments complied with William & Mary Institutional Animal Care and Use Committee (IACUC) protocols 2019-11-04-13894, 2021-10-05-15162, and 2022-09-15-15848.

Results

We collected 37 larvae that were in suitable condition for this study. Of these, 31 were used for CTmax trials (Table 2), and six larvae were monitored as controls (i.e., no warming). The group of experimental fishes included three cryonotothenioid families: three species of Channichthyidae (n = 13 larvae), two species of Artedidraconidae (n = 14), and two species of Nototheniidae (n = 4) (Table 2). Neodraco skottsbergi (Lönnberg 1905) (Artedidraconidae) possessed the highest average CTmax (16.8 °C ± 1.0, n = 6), with one individual attaining a CTmax of 17.8 °C; this is among the highest values reported for any adult or larval cryonotothenioid (Bilyk and DeVries 2011; Beers and Sidell 2011). Three Chionodraco rastrospinosus DeWitt and Hureau 1980 (Channichthyidae) larvae had CTmax values ranging from 8.6 to 9.2 °C (Fig. 1b), which are among the lowest values of any cryonotothenioid tested to date (Bilyk and DeVries 2011; Beers and Sidell 2011). Of all study species, only N. skottsbergi could resume normal swimming behavior following experimental trials. This behavior was briefly monitored (< 10 min) before larvae were euthanized; therefore, it is unclear how long this behavior would have persisted and the long-term effects of exposure to acute thermal stress on these larvae. All control larvae (four artedidraconids, one nototheniid, and one channichthyid) maintained normal swimming behavior for at least 7.2 h, while the longest CTmax trial was 59 min.

Table 2 Mean critical thermal maxima (CTmax), sample sizes (n), standard lengths (SL), and standard deviations (SD) of larval cryonotothenioids collected near the western Antarctic Peninsula during this study. Previously determined adult CTmax are reported for C. rastrospinosus and L. squamifrons. *The standard lengths of larvae from control trials are reported, but they are not included in the summed sample size for each family
Full size table

There is a positive linear relationship (coefficient [coef] = 0.76, mean squared error [MSE] = 2.6, p = 0.058) between length and CTmax among grouped channichthyid and nototheniid larvae (Fig. 2a). There is also a positive relationship (coef = 0.74, MSE = 2.6, p = 0.067) between tow location SST and CTmax among the grouped channichthyids and nototheniids (Fig. 2b). Due to their anomalously high CTmax, artedidraconids were excluded from both regressions.

Fig. 2
figure 2

Relationship between (a) standard length and (b) sea surface temperature at collection location and CTmax. Linear regression (black dashed line) fitted only to fishes from Channichthyidae and Nototheniidae, with species from Artedidraconidae (light and dark green circles) excluded. Grey shading represents 95% confidence intervals

Full size image

Discussion

With increasing environmental pressures from climate change in Antarctic ecosystems, it is important to better understand the impacts of a warming ocean on larval stages. In this study, we determine the CTmax of cryonotothenioid larvae collected near the western Antarctic Peninsula, representing seven species of three cryonotothenioid families—Nototheniidae, Channichthyidae, and Artedidraconidae. For channichthyid and nototheniid species, CTmax values were positively correlated with body length, whereas those of the artedidraconids that were tested did not. Below, we further discuss these results in the context of climate change and evolutionary history.

Artedidraconids

The high-Antarctic Artedidraco orianae Regan 1914 appears to have limited metabolic capacities at elevated temperatures based on cellular energy budgets; however, more data are needed to draw firm conclusions (Mark et al. 2005). The impact of acclimation time on oxygen consumption in Pogonophryne scotti Regan 1914 has also been tested (Saint-Paul et al. 1988). We are unaware of any studies on the whole-body thermal tolerance of any artedidraconid species. Therefore, it is unclear if the abnormally high thermal tolerances we observed in N. skottsbergi, and to a lesser degree Dolloidraco longedorsalis Roule 1913 (Table 2; Fig. 1b), are unique to larval stages. Also, the rapid recovery of all larval N. skottsbergi following CTmax trials may indicate that the physiological damage of the acute heat exposure is, at least temporarily, reversible (Ern et al. 2023).

The biology of adult N. skottsbergi does not explain the unusually high thermal tolerance of their larval stages. This species has a circumpolar distribution, found in both SPZ and HAZ regions, and possesses antifreeze glycoproteins (Miya et al. 2016; Baalsrud et al. 2018). Like other artedidraconids, N. skottsbergi is a benthic, generalist feeder consuming primarily amphipods and polychaetes (Gon and Heemstra 1990; La Mesa et al. 2015). Small (10.0–14.5 mm SL) N. skottsbergi larvae with remnants of their yolk sac have been found around Joinville Island, WAP, from early November to early December, which suggests that hatching in this area occurs from October–November (Kellermann 1990). They are only captured in the epipelagic zone during the austral summer (January to late February); afterward, they likely begin their descent to the benthos (Kellermann 1990). Thus, their biogeographic and ecological traits do not generally differ from other more stenothermal cryonotothenioids; however, one aspect of their larval biology is unusual and described below.

All N. skottsbergi larvae we captured possessed a unique vibrant green color (Fig. 3a). We are unaware of any other cryonotothenioid larvae with similar green pigmentation, and there is no mention of this trait in Kellermann (1990) or any other publication. Although there is little information on the coloration of post-larvae and juveniles, N. skottsbergi are dark brown with a yellowish hue as adults (Gon and Heemstra 1990). Most cryonotothenioid larvae have a pelagic stage lasting at least two months during the austral summer, overlapping with peak phytoplankton production (Kock and Kellermann 1991). All N. skottsbergi in this study were collected in neritic areas near marine-terminating glaciers in Dalgliesh Bay (Fig. 1a). The freshwater layer associated with glacial meltwater stratifies the upper water column during the summer, which leads to warmer sea surface temperatures, while the macro- and micro-nutrients in the stable meltwater promote the growth of several Antarctic phytoplankton species (Pan et al. 2020). The green coloration of the larvae could help them blend in with phytoplankton concentrated in near-surface (< 1 m depth) waters, providing cover to escape predation (Coston-Clements et al. 1991). During daylight and marginal wind conditions, the upper meter of the water column can be twice as warm as at 10 m depth (Minnett and Kaiser-Weiss 2012). N. skottsbergi would routinely experience warmer temperatures at these shallow depths compared with other notothenioid larvae typically found in deeper areas (20–500 m; Loeb et al. 1993), which could partially explain their elevated CTmax. This explanation aligns with an analysis of eight adult nototheniid species that found cardiorespiratory control correlated with ecotype (e.g., benthic, epibenthic, or pelagic) (Campbell et al. 2009). All larval N. skottsbergi in our study were captured by oblique tows from 0 to 120 m; thus, we are unaware of their exact collection depth. It is also not possible to determine the distribution of freshwater input from marine-terminating glaciers without a detailed analysis occurring concurrently with net tows. Future research should evaluate the precise depth horizon larval N. skottsbergi occupy, their prey field, regional oceanographic conditions, and intraspecific variability in their green coloration.

Fig. 3
figure 3

Post-flexion larva of Neodraco skottsbergi live (a) and preserved in 95% ethanol (b) VIMS 43504, 18.6 mm SL

Full size image

Recent research on artedidraconids has focused on their evolutionary history and phylogenetic relationships, which can drive species trait patterns and thermal tolerance (Felsenstein 1985; Todgham and Mandic 2020). In their phylogenetic analysis, Parker and Near (2022) recovered artedidraconids as a group within Harpagiferidae as the subfamily Artedidraconinae (Andriashev 1965). One harpagiferid species, Harpagifer bispinis (Forster in Bloch and Schneider 1801), is found in the Magellan region of South America (Duhamel et al. 2014), where summer ocean temperatures can exceed 11 °C. Although experimental fishes were acclimated to aquaria before trials, H. bispinis has a CTmax of 23.6 °C (Llompart et al. 2020; Giménez et al. 2021). Several harpagiferid species are also routinely found in shallow tide pools in the SPZ (Duarte and Moreno 1981; Casaux 1998; Eastman 2017), where temperatures can exceed 5 °C depending on sunlight intensity (Mintenbeck et al. 2012; Clarke and Beaumont 2020). Evolutionary history correlates strongly with CTmax in aquatic organisms (Bennett et al. 2021), including freshwater fishes, with the retention of plesiomorphic characters contributing to thermal tolerance (Comte and Olden 2017). Therefore, occupying habitats such as the warm tide pools inhabited by closely related harpagiferids could also contribute to the CTmax of N. skottsbergi; whether this reflects a plesiomorphic habitat association for artedidraconids is unknown. Future study of their physiological responses to warming will contribute to understanding the evolution of this outlying condition within cryonotothenioids.

Channichthyids and nototheniids

Channichthyids are among the most stenothermal of all cryonotothenioids (Beers and Sidell 2011). These fishes completely lack hemoglobin, the protein in red blood cells that binds to oxygen, and facilitates its transport throughout the circulatory system. The absence of respiratory pigments (hemoglobin, and in some cases myoglobin), in white-blooded channichthyids limits their oxygen-carrying capacity to 10% that of red-blooded nototheniids (Holeton 1970; Beers and Sidell 2011). Beers and Sidell (2011) evaluated the CTmax of several channichthyids and nototheniids, finding hematocrit (i.e., the proportion of red blood cells within a circulatory system) positively correlates with CTmax. Two channichthyid species, Chionodraco rastrospinosus and Chaenocephalus aceratus, (Lönnberg 1906) possess the lowest mean CTmax in that analysis (Beers and Sidell 2011).

Results of our linear regression suggest that thermal tolerance in some larvae of channichthyids increases with ontogeny and environmental SST. Although larval nototheniids were included in the regression, we hesitate to draw similar conclusions regarding their relationships without a larger sample size of individuals in experimental trials. However, our results (mean of 13.9 °C) closely align with the only known CTmax for larval nototheniids (14.6 °C; Table 1) (Evans et al. 2012). The positive correlation with SST is consistent with the results obtained by Bilyk and DeVries (2011), who report that several adult cryonotothenioids have a level of plasticity in their upper thermal limit depending on the ocean temperature of their environment. In addition to acclimation to regional conditions, the apparent correlation between environmental temperature and CTmax could also be impacted by the evolution of intraspecific variability in thermal tolerance among cryonotothenioids, potentially due to underlying genetic diversity or transgenerational effects (McKenzie et al. 2021). Previous studies report CTmax values for the adult counterparts for two of our study species: C. rastrospinosus (Beers and Sidell 2011) and L. squamifrons (Bilyk and DeVries 2011). The average CTmax for both species is lower in larvae than that reported for adults (Table 2).

Ontogeny impacts thermal tolerance in fishes, although few experiments have been conducted on the early life stages of cryonotothenioids (Table 1). Studies of other marine fishes suggest heat tolerance is positively correlated with the development of the cardiorespiratory system (Dahlke et al. 2020). For example, Boreogadus saida (Lepechin 1774), or Arctic cod, larvae have an upper thermal tolerance that is 40% less than adults based on cardiac response (Drost et al. 2016). The limited aerobic capacity of early life stages can compromise the ability to deliver oxygen to tissues during unusually warm temperatures (Rombough 1988; Pörtner et al. 2006). Given that channichthyids are already functioning with limited oxygen-carrying capacity, the larval stages of this group may be especially vulnerable to thermal extremes as their cardiorespiratory systems develop.

Higher stenothermy during early life stages has been proposed as a possible trade-off, such that larval fishes are more dependent on ideal environmental temperatures and thus may invest more energy into growth (Pörtner et al. 2006). Pre-flexion larvae are also less motile than older fishes, which limits their ability to escape incursions of warm water (Downie et al. 2020). In addition to the long-term ocean warming observed in the WAP region of 0.1 to 0.3 °C per decade (Schmidtko et al. 2014), increasing frequency of extreme heatwaves may pose a serious risk to pre-flexion channichthyid larvae (Robinson et al. 2020; Robinson 2022; Morrison et al. 2022; González-Herrero et al. 2022). The long-term abundance of the larval nototheniid Pleuragramma antarctica Boulenger 1902 is negatively correlated with sea surface temperature (Corso et al. 2022). Still, there are no long-term analyses of channichthyid abundance and ocean temperature during any life stage. Such analyses are needed to model population dynamics accurately.

In conclusion, the results of this study indicate that additional research on the physiology and ecology of artedidraconids and channichthyids should be prioritized to determine the potential so-called winners and losers of climate change in the Southern Ocean. Future in vivo research is vital to understand the impact that warming may have on all cryonotothenioid fishes, particularly during sensitive larval stages, to help project the future of the Southern Ocean ecosystem.