全球变化知识资源中心

Ocean acidification negatively affects many marine species and is predicted to cause widespread changes to marine ecosystems. Similarly, freshwater ecosystems may potentially be affected by climate-change-related acidification; however, this has received far less attention. Freshwater fish represent 40% of all fishes, and salmon, which rear and spawn in freshwater, are of immense ecosystem, economical and cultural importance. In this study, we investigate the impacts of CO₂-induced acidification during the development of pink salmon, in freshwater and following early seawater entry. At this critical and sensitive life stage, we show dose-dependent reductions in growth, yolk-to-tissue conversion and maximal O₂ uptake capacity; as well as significant alterations in olfactory responses, anti-predator behaviour and anxiety under projected future increases in CO₂ levels. These data indicate that future populations of pink salmon may be at risk without mitigation and highlight the need for further studies on the impact of CO₂-induced acidification on freshwater systems.

Future predicted increases in CO₂ have been demonstrated to cause a wide range of sublethal effects on a variety of marine fish species. Effects include changes in olfactory responses to predator, prey and substrate cues^{1, 2, 3}, interference with neurotransmitter function^{4, 5}, alterations in behavioural lateralization^{4, 6} and heightened anxiety⁵. Despite the growing body of work on the effects of ocean acidification (OA) on marine systems, less is known about the future patterns and dynamics of CO₂ in freshwater, making it difficult to infer how freshwater ecosystems will respond to climate-change-related acidification. Nevertheless, elevated CO₂ in both marine and freshwater systems is a likely scenario in the future⁷. Although freshwater comprises only 0.8% of the water on the Earth’s surface, freshwater ecosystems support almost 40% of all fish species⁸. Therefore, investigating the effects of CO₂-mediated acidification in freshwater systems may provide important insights into how almost half of the world’s fishes and associated communities will respond to climate change.

Salmon are a keystone species in many marine, freshwater and terrestrial ecosystems because of their role in supporting food webs^{9, 10} and, as such, may help link distinct and relatively isolated systems. All salmon initially rear in freshwater but spend the majority of their juvenile and adult lives in seawater, before returning to their freshwater natal streams to spawn¹¹. Thus, the freshwater environment is crucial to their life history. In comparison to all other Pacific salmon, pink salmon (Oncorhynchus gorbuscha) are the most abundant and widely distributed and, consequently, considered an important indicator of ecosystem health¹². Unlike most other anadromous salmonids, pink salmon migrate to sea soon after emergence and, at approximately 0.2 g, are by far the smallest at the time of seawater entry^{13, 14}. As seawater entry is a time of heightened mortality for all species of salmon^{15, 16, 17}, their small size and high ratio of surface area to volume may make pink salmon especially vulnerable to environmental stressors, including OA.

Here, we show negative effects of CO₂-induced acidification on the growth, metabolism, olfactory responses and anti-predator behaviour in the early life stages of pink salmon during freshwater development and post-seawater entry using predicted future levels of CO₂. Unlike other studies, we also incorporated a fluctuating CO₂ treatment to reflect naturally occurring CO₂ fluctuations in freshwater and coastal ecosystems^{18, 19, 20}.

Two weeks before hatch, eyed embryos were transferred into one of four freshwater treatments (constant 450 μatm (present-day control), constant 1,000 μatm (100-year projection), constant 2,000 μatm, and diurnal fluctuating 450–2,000 μatm) for ten weeks. Growth, routine (RMR) and maximum (MMR) metabolic rate were measured throughout freshwater development and, at the end of ten weeks, their levels of anxiety and olfactory and anti-predator responses to conspecific alarm cues were measured. When fish reached yolk sac absorption (week 10 post-CO₂ exposure), fish reared in freshwater at constant 450 μatm CO₂ were transferred to one of three different seawater treatments of 450, 1,600 (future coastal conditions) and a diurnal 450–1,600 μatm CO₂ treatment. A subset of fish reared at constant 2,000 μatm CO₂ in freshwater were also transferred into 1,600 μatm seawater (Supplementary Information). Growth and metabolic rates were measured over the subsequent two weeks post-seawater transfer.

During freshwater development, there was a significant negative effect of CO₂ on fork length, production efficiencies, total wet mass and total tissue mass (yolk removed) and no significant effect of CO₂ on RMR and MMR (Supplementary Information). At yolk sac absorption (week 10), total wet mass (F₃ = 4.0157, p = 0.0218), total dry mass (F₃ = 4.7993, p = 0.0112), production efficiencies (F₃ = 4.5412, p = 0.0139) and fork lengths (F₃ = 8.0345, p = 0.0010) were significantly different among groups (see Fig. 1). Production efficiencies (% of yolk converted into tissue) near yolk sac absorption were 25% lower in the high-CO₂ group compared to the control (p = 0.0064). Similarly, fork lengths were significantly reduced at 1,000 and 2,000 μatm compared to control levels (p = 0.0292 and p = 0.0004, respectively).

a, Body lengths. b, Production efficiencies (net tissue produced/net yolk consumed × 100). c, Total wet mass. d, Total dry mass. 450–2,000 μatm represents a diurnal cycle and other tensions are constant throughout. Values are means ± s.e.m. (n = 8). Asterisks indicate a statistically significant difference from the 450 μatm control group (p < 0.05).

Our results indicate that future increases in CO₂ in freshwater may have substantial negative impacts on the growth, olfactory responses and anti-predator behaviour of pink salmon during early development, which may have large implications on their outward migration success in freshwater and early seawater survival. Near the end of yolk absorption, total wet mass and length were reduced in fish reared at 1,000 and 2,000 μatm compared to those reared in control conditions. Reductions in yolk-to-tissue conversion efficiencies at elevated p_CO2 may indicate that development at p_CO2 levels projected by the end of the century may incur a greater energetic cost associated with acid–base regulation. Previous work investigating growth in fish at OA-relevant levels demonstrates varying responses^{22, 23, 24}, suggesting that OA impacts on growth may be species specific. Early in embryonic development, Japanese medaka, Oryzias latipes, show stunted growth under constant elevated p_CO2, but this was no longer evident by hatch²². In the case of pink salmon, a shift in developmental time does not seem to be occurring. Near yolk absorption, yolk wet mass is not significantly different among the different CO₂ groups, suggesting that yolk consumption rates were similar (Supplementary Information). Given that gravel emergence in pink salmon occurs at yolk sac absorption, these data indicate that elevated CO₂ in freshwater would probably lead to comparable times of gravel emergence, but in fish that are smaller. Predation, in most cases, is strongly size-selective²⁵; thus, smaller hatchlings may have lower survivorship²⁶. This has large implications on the survival of pink salmon not only in freshwater but during early ocean entry as they make the transition to a more predator-rich marine environment.

In addition to impaired growth, the diminished capacity for pink salmon fry to detect olfactory cues and elicit predator avoidance behaviour under projected increases in CO₂ may have negative consequences on survival by altering predator–prey dynamics. Most of the studies to date investigating the effect of acidification on olfactory and behavioural responses to chemical cues have largely been conducted on marine species^{1, 4, 27, 28} or in freshwater species at relatively low pH (pH 5.0–6.3), with the use of strong acids (H₂SO₄ or HCl) to acidify the environment^{29, 30, 31}. The latter will not affect internal receptors to the same degree because membrane permeability to charged H⁺’s is much lower than that of CO₂. Compared to pink salmon in freshwater, marine fish also exhibited similar impairment in detecting chemical cues at elevated CO₂ levels. For example, short-term exposure to high CO₂ reduced anti-predatory behaviour and increased predation by a common predator in juvenile damselfish, Pomacentrus sp.^{2, 28}. In freshwater, acidification to pH < 6.3 seems to diminish or completely abolish predator avoidance behaviour in various fish species, including other salmonids^{29, 31, 32}. Whereas transgenerational acclimation has been documented to mitigate some of the negative effects of CO₂ on growth and aerobic scope in reef fish³³, it had little effect on restoring olfactory responses to alarm cues and behaviour lateralization in juvenile spiny damselfish, Acanthochromis polycanthus³⁴. This suggests that genetic adaptation would be required to mitigate the effects of elevated CO₂ on sensory and cognitive impairment³⁴. Consistent with our EOG data, olfactory responses in Atlantic salmon parr (Salmo salar) to common odorants were significantly reduced at low freshwater pH (4.5 and 5.5; ref. 35). Below a pH of 6.5 (via H₂SO₄ addition), significantly higher concentrations of odorants were needed to produce responses of similar magnitude to controls³⁵. Reduced EOG responses to alarm cues and amino acids suggest that elevated CO₂ may impair olfactory sensitivity to common odorants in pink salmon. The recognition of the amino acid composition of natal streams is thought to play a large role in imprinting and homing migration of salmon^{36, 37}. If future levels of CO₂ impair the recognition of amino acids, then homing ability of adult salmon may be affected. Additional behavioural studies on salmon homing are required to investigate how our observed reduction in olfactory responses is translated into behaviour. Furthermore, exposure to weak freshwater acidification (pH 6.1–6.3) has been shown to suppress spawning and migration behaviours in sockeye salmon, O. nerka^{38, 39, 40}, thereby impacting various life history traits⁴¹. However, previous freshwater studies on anti-predator behaviour and olfactory responses were conducted at significantly lower pH than our present study and without the use of CO₂ (based on our water composition, a pH of 6.0–6.3 would correspond to a p_CO2 of ~4,000–9,000 μatm), which may not be relevant in the context of climate change. To tease apart pH versus CO₂ effects, research investigating both p_CO2 and pH effects in varying types of freshwater sources is required to better understand the mechanistic basis for impaired olfaction.

Additional behavioural alterations in pink salmon, such as decreased anxiety and increased boldness, may further compound the negative effects of elevated CO₂ on predator avoidance behaviour, and thus survival. In our study, pink salmon fry reared in high CO₂ in freshwater spent significantly more time in the centre zone around the novel object and less time in the thigmotaxis zone than control fish, suggesting that elevated CO₂ may suppress anxiety. Thigmotaxis is a common index of anxiety-like behaviour in zebrafish and in rodents^{21, 42}. Inhibition of the GABAergic system with anxiogenic drugs leads to increases in anxiety-like behaviour^{5, 42, 43, 44}. Similarly, our results indicate significant increases in anxiety in both controls and high-CO₂-reared fish after treatment with gabazine, with complete reversal of the anxiety-reducing effects of elevated CO₂. In ref. 4 it was also shown that gabazine reversed the effects of CO₂ on olfactory ability and behavioural lateralization in larval damselfish and clownfish, and a mechanistic explanation was proposed for the effects of OA; a reversal in chloride and bicarbonate gradients through GABA_A receptors, which shifts some of these receptors from inhibitory to excitatory. Consistent with this altered GABA_A receptor activity, exposure to high CO₂ increased anxiety in juvenile rockfish (Sebastes diploproa) and treatment with gabazine resulted in increased anxiety in the control fish to levels of high-CO₂-exposed fish⁵. In marine stickleback, exposure to elevated CO₂ reduced boldness and curiosity during a novel approach test⁶. Unlike stickleback, pink salmon exposed to high CO₂ spent more time around the novel object, demonstrating an increase in boldness. The observed CO₂-induced differences in anxiety and boldness in seawater compared to the freshwater salmon in this study may be due to a variety of reasons, from completely different mechanisms that regulate ion balance to differences in life history traits.

Our results show that future increases in CO₂ in seawater may impair early seawater survival of migrating pink salmon fry through impacts on subsequent growth and MMR. Growth rates were negative in groups transferred into high-CO₂ seawater, whereas control fish continued to grow throughout our seawater exposure. Fish were not fed; however, internal yolk stores were still present, which were presumably drawn on for growth. This suggests that tissue production may be impaired at high CO₂. However, because growth rates are based on total wet mass, the inability to hypo-osmoregulate efficiently at high p_CO2 could also have contributed to the apparent negative growth rate estimates (due to water loss) during the transition to seawater. If the lower growth rates at high p_CO2 are associated with impaired hypo-osmoregulatory ability, then the osmoregulatory challenge of seawater entry may be further compounded by OA. Following seawater transfer, an increase in CO₂ reduced MMR in pink salmon fry. These results indicate that pink salmon fry may be particularly sensitive to climate-change-related acidification during exercise, which may have implications on the success of their seaward migration and early ocean survival. Pink salmon migrate to sea soon after emergence and this heightened capacity for activity may be crucial to their survival in a predator-dense environment. However, elevated p_CO2 may reduce the capacity for maximal O₂ uptake and exercise of migrating fry, making them more susceptible to predation and reducing their foraging success at a time when yolk reserves become limiting.

This study demonstrates that pink salmon may be faced with numerous sublethal impacts of acidification on their physiology and behaviour under predicted future increases in p_CO2. However, carbonate chemistry data of inland waters suggests that some freshwater ecosystems are at present exposed to temporal elevations in p_CO2 (ref. 45). For example, p_CO2 in the Columbia River, which hosts runs of various salmon species, ranges from 541 to 981 μatm (ref. 46). Similarly, recent oceanic surveys of the West Coast of North America suggest that our coastal ecosystems temporally experience large elevations in CO₂, which may exceed 1,000 μatm and persist for months in some locations^{20, 47, 48, URL:}