Climate change effects on fishes and fisheries: towards a cause-and-effect understanding


  • H. O. Pörtner,

    Corresponding author
    1. Integrative Ecophysiology, Alfred-Wegener-Institute, Am Handelshafen 12, 27570 Bremerhaven, Germany
      Tel.: +49 47148311307; fax: +49 47148311149; email:
    Search for more papers by this author
  • M. A. Peck

    1. Center for Marine and Climate Research, Institute for Hydrobiology and Fisheries Science, University of Hamburg, Olbersweg 24, 22767 Hamburg, Germany
    Search for more papers by this author

Tel.: +49 47148311307; fax: +49 47148311149; email:


Ongoing climate change is predicted to affect individual organisms during all life stages, thereby affecting populations of a species, communities and the functioning of ecosystems. These effects of climate change can be direct, through changing water temperatures and associated phenologies, the lengths and frequency of hypoxia events, through ongoing ocean acidification trends or through shifts in hydrodynamics and in sea level. In some cases, climate interactions with a species will also, or mostly, be indirect and mediated through direct effects on key prey species which change the composition and dynamic coupling of food webs. Thus, the implications of climate change for marine fish populations can be seen to result from phenomena at four interlinked levels of biological organization: (1) organismal-level physiological changes will occur in response to changing environmental variables such as temperature, dissolved oxygen and ocean carbon dioxide levels. An integrated view of relevant effects, adaptation processes and tolerance limits is provided by the concept of oxygen and capacity-limited thermal tolerance (OCLT). (2) Individual-level behavioural changes may occur such as the avoidance of unfavourable conditions and, if possible, movement into suitable areas. (3) Population-level changes may be observed via changes in the balance between rates of mortality, growth and reproduction. This includes changes in the retention or dispersion of early life stages by ocean currents, which lead to the establishment of new populations in new areas or abandonment of traditional habitats. (4) Ecosystem-level changes in productivity and food web interactions will result from differing physiological responses by organisms at different levels of the food web. The shifts in biogeography and warming-induced biodiversity will affect species productivity and may, thus, explain changes in fisheries economies. This paper tries to establish links between various levels of biological organization by means of addressing the effective physiological principles at the cellular, tissue and whole organism levels.


Ocean warming is currently one of the main driving forces causing changes in species abundance and distribution and, thus, in species composition in marine ecosystems (Perry et al., 2005). This appears as a consequence of the fact that temperature defines the large-scale geographical distribution of marine water-breathing animals, within conditions set by geomorphology, ocean currents, water depth and stratification or salinity. The fish fauna has been widely investigated in this context with long-term data series compiled due to fisheries interests. Such studies have focused on statistical analyses of how stock size or population structure, recruitment or spawning events are influenced by changes in climatic indices such as the North Atlantic Oscillation (NAO; Alheit et al., 2005). Shifts in geographical distribution occur in response to climate change and are generally most evident near the northern or southern boundaries of the geographic range of a species, where warming or cooling theoretically drives marine fishes to higher and lower latitudes, respectively. A number of studies have documented such changes within particularly well-studied ecosystems of the world's oceans including the North Sea and other parts of Europe (Beare et al., 2004; Perry et al., 2005; Rose, 2005). Furthermore, analyses have investigated whether there have been shifts in the seasonal timing (phenology) of crucial events such as spawning by fishes (Sims et al., 2005) and the spring blooms of phytoplankton and zooplankton (Wiltshire et al., 2008). Out-of-phase shifts between the former and latter can have large consequences for match and mismatch phenomena in, for example, food availability for larval and juvenile fishes (Beaugrand et al., 2002), possibly leading to regime shifts (Beaugrand, 2004).

It is important to recognize that the observed effects of global warming on fishes at the various levels of biological organization (organismal, population and community–ecosystem) result from physiological changes at molecular, cellular and whole organism levels and that the ultimate effects of global warming at the ecosystem level will build on species-specific responses (Pörtner, 2001, 2002). As a consequence, species interactions change with interactions at the community level (Pörtner & Farrell, 2008). Such a cause-and-effect understanding is needed to reliably project the effects of global warming on commercially important marine fish species and to disentangle these effects from the synergistic effect of fishing pressure on such populations. A fundamental concept which links the various levels of biological organization and also provides a matrix for integrating various environmental stressors is the concept of oxygen and capacity-limited thermal tolerance (OCLT).

Although purely statistical analyses of climate effects have provided evidence for climate-induced effects, only a few examples exist which demonstrate the relationship between climate sensitivity of particular species using physiological principles. In the following, appropriate examples are provided by discussing changes in specific populations of eelpout Zoarces viviparus (L.), Japanese anchovy Engraulis japonicus Temminck & Schlegel, Japanese sardine Sardinops melanostictus (Temminck & Schlegel) and Atlantic cod Gadus morhua L.

This review paper focuses attention on the need for a cause-and-effect understanding of climate interactions with marine fishes and the ecosystems in which they live. In the first section, a review is given of broad-scale patterns observed in (1) the effects of temperature on species and (2) climate-driven ecosystem-level changes. In the next section, coverage focuses on (3) an elaboration of the cellular and organismal-level physiological underpinning of species-specific responses to temperature and changes in other climate-driven factors (pH and hypoxia). Finally, in (4) a discussion is presented on current modelling methods that are based (at least in part) on physiological first principles and that are being utilized to understand historical and project future climate-driven changes in fish populations.

Temperature tolerance and optima and climatic effects on species

The literature is replete with studies that have measured lethal limits and, thus, tolerance to extreme temperatures by fish species (Appendix). Meta-analyses of the available data have demonstrated clear species-specific differences in both preferred and tolerable extreme temperatures (Jobling, 1981). These differences appear not only between species but also within species, for example related to differences in thermal windows due to acclimation or permanent population differences.

Measurements of the upper and lower lethal temperatures or critical thermal maxima or minima (Becker & Genoway, 1979) of fish species (TCmax and TCmin or TLmax and TLmin) can be used to illustrate the degree of specialization of species and populations on specific thermal environments. As a note of caution, these extreme limits may not be directly relevant in the ecosystem. They are, however, correlated with limits reached sooner and operative at ecosystem level. These limits, which are sublethal on shorter time scales, are not (yet) as widely available in the literature.

Extreme temperature limits and associated tolerance ranges within and among fish species change with latitude of the field population [Fig. 1(a)]. The range in tolerable temperatures is most narrow for fishes inhabiting high latitudes and relatively narrow for species at low latitude. In contrast, the tolerance range tends to be widest for fishes inhabiting mid-latitudes where seasonal differences in temperatures are, on average, largest. Figure 1 also depicts the preferred temperatures of such species. Naturally, these thermal endpoints are not static but depend to some degree on acclimation temperature. When a fish species can be acclimated to vastly different temperatures such as a range of 20° C, large differences in the upper and lower limits can be observed [Fig. 1(b)]. Compared to critical or lethal extremes, acclimation to different water temperatures has less effect on preferred temperatures. Within a species, preferred temperatures are often closely related to TCmax (Tsuchida, 1995) and often correspond to optimal growth temperatures (Jobling, 1981). Caution must be taken, however, not to oversimplify estimates of thermal limits and preferences (and their relation) since other environmental factors (e.g. prey availability and salinity) can modify each of these thermal values. Moreover, measurements of TCmax need to be interpreted with caution since experimental protocols may not have adequately accounted for the time dependence of such lethal thresholds. Nevertheless, overall, patterns of changes in tolerance may give a first approximation of the relative effects of increased temperatures due to global warming. The relationship also allows geographical distributions of species (Arctic, boreal, Lusitanian, tropical and Atlantic) to be used as a classification scheme to evaluate expected responses. In general, the analysis of physiological patterns on large scales, e.g. latitudinal clines as in macrophysiology (Gaston et al., 2008), will help to unravel specialization on climate and associated limitations.

Figure 1.

(a) Thermal tolerance in marine fish species including upper (TU, inline image, 108 species), lower (TL, inline image, 61 species) and preferred (TPR, inline image, 21 species) temperatures and latitude of the fish population. Note that values of TU or TL are lethal or critical thermal maxima and minima (TCmax and TCmin). (b) Values of TL, TU or TPR shift for fish acclimated to different water temperatures (grouped according to the magnitude of the difference between upper and lower acclimation temperatures). Note the increase (mean ±s.d.) in TL (inline image) and TU (inline image) ranges but not TPR (inline image) for fishes studied over the broadest range of acclimation temperatures. Data were taken from published literature sources (see Appendix).

A second important feature of thermal physiology within fish species is ontogenetic change in the width of thermal tolerance windows. Two examples of this phenomenon are presented in Fig. 2, one based on a literature synthesis presented by Brewer (1976) for northern anchovy Engraulis mordax Girard inhabiting the California current, and the second from Rijnsdorp et al. (2009) based on a compilation of results for sole Solea solea (L.) in the North Sea. Patterns of thermal tolerance by life stage in both these highly phylogenetically different species indicate thermal bottlenecks during the early life stages or in spawning stage. In the future and once available, such critical and lethal limits as depicted in Figs 1 and 2 will need to be replaced by more realistic limits (more narrow thermal ranges) beyond which ecological interaction sets in (Pörtner & Knust, 2007). Available knowledge of mechanistic principles already indicates that thermal windows are narrow in early life stages, due to developmental constraints and insufficient capacity of central organs in the larvae (Pörtner et al., 2006). Thermal windows widen in juveniles and young adults in line with rising performance capacity at small body size. Larger individuals then become more thermally sensitive, due to progressively falling oxygen supply capacity in relation to demand. Adult spawners need to provide oxygen to their large egg or sperm masses, at the expense of a narrowing of thermal tolerance windows and a lowering of tolerated extremes (Pörtner & Farrell, 2008).

Figure 2.

Ontogenetic changes in thermal habitats or tolerance for Engraulis mordax in the California current (from Brewer, 1976) and Solea solea in the North Sea (from Rijnsdorp et al., 2009). In both cases, the earliest life stages were more sensitive (had a more narrow range in thermal tolerance) than later larvae and early juveniles that can exploit the largest range in thermal habitats compared to earlier or later (adult) life stages. In the S. solea example, maturation refers to the body size at which fish become sexually mature (first enter the adult stage) and does not refer to adult fish that are in spawning condition (displayed for E. mordax data).

Ecosystem effects are not only related to shifts in seasonal or annual mean temperatures. Currently, such shifts and their effects are prominent at high latitudes. Temperature maxima or minima, as opposed to mean temperatures, however, represent strong driving forces for ecosystem-level changes in population structure and in the community composition of marine areas. This is true on both the cold and the warm side of the temperature spectrum and, accordingly, on both sides of the thermal window of species. Here, ambient temperature exceeds the capacity of the individuals to acclimatize to lower or higher temperatures and, thereby, shift thermal limits. Long-term sustenance of species below their lower thermal limit is not possible. Among adult fishes, a prominent example is the sea bream Sparus aurata L. which cannot be cultured at ≤10° C as it develops symptoms called ‘winter syndrome’ (Domenech et al., 1997). This species cannot acclimate to those low temperatures and therefore suffers from fitness decrements and disease (Gallardo et al., 2003; Ibarz et al., 2010; Kyprianou et al., 2010).

Sensitivity to cold appears as a very important characteristic in shaping community composition. Accordingly, winter exposure of fish larvae is seen as a key selective factor in setting the productivity of a species. The occasional occurrence of cold stress is probably effective in setting limits to the biogeography of cold-sensitive species. Low temperatures during winter may increase mortality, either because temperatures fall outside the thermal window or because energy reserves become limiting, especially in smaller individuals that have relatively fewer reserves compared to larger conspecifics (Post & Evans, 1989; Sogard, 1997). Extreme winter events cause reductions in species abundance and ecosystem changes, for example in the German Wadden Sea (Woodhead, 1964). The frequency of such events, however, has decreased during recent decades and the alleviation of winter cold has been paralleled by an increase in the frequency of warmer summers.

Conversely, cold-tolerant species will suffer from increasing exposure to summer heat. Observed shifts in geographical distribution of species such as G. morhua, snake blenny Lumpenus lumpretaeformis (Walbaum) and anglerfish Lophius piscatorius L. in the North Sea may involve the shifting frequencies of such weather conditions (Perry et al., 2005). Specific life stages, however, may be especially sensitive during well-defined seasonal time windows as seen for G. morhua in the southern North Sea. Here, winter warming is closely correlated with the population shift (Perry et al., 2005), indicating that adult spawners as well as eggs and larvae are the putative critical life stages sensitive to winter warming (Pörtner et al., 2008; Pörtner & Farrell, 2008). As a result, recruitment of G. morhua in the North Sea falls during warming, but rises during cooling. In contrast, recruitment of the Barents Sea (north-east Arctic) G. morhua population decreases during cooling periods and increases upon warming. These trends emphasize that G. morhua in the southern North Sea live close to the upper and those in the Barents Sea to the lower thermal limits of the species, respectively. In line with these observations, correlations between climate-associated temperature change and G. morhua recruitment are stronger at the southern and northern borders of the geographical distribution area of the species in the north-east Atlantic Ocean. As a corollary, temperature defines the southern and northern distribution limits of the species.

A second example of the consequences of different thermal windows for growth and spawning productivity can be found in the oscillations between E. japonicus and S. melanostictus in the Pacific Ocean. In the Kuroshio–Oyashio current, alternating periods of relatively cold and warm years are closely associated with alternating dominance of S. melanostictus and E. japonicus, respectively. These shifts are linked to the limited and different thermal windows of growth and of spawning productivity of the two species (Takasuka et al., 2007, 2008). Performance capacity within a limited thermal window thus appears key to the successful survival and competition of a species.

Such climate-dependent functional specialization and differentiation even occurs between populations of the same species. Gadus morhua populations between the southern North Sea and the Arctic North Atlantic display different thermal windows of growth (Pörtner et al., 2001, 2008). The key observation is that high latitude populations display an earlier reduction in growth during warming and are more cold tolerant. Their overall growth rates appear reduced, probably due to the metabolic cost associated with being cold tolerant and eurythermal at the same time. Such climate-driven functional differentiation may support the divergence of species from a common ancestor. For example, along the North American Atlantic coast between Nova Scotia, Canada, and Florida, U.S.A., the killifish Fundulus heteroclitus (L.) has split into two species with hybrids occurring in the middle of the range. To what extent adaptation to local or regional differences in climate has supported speciation is an active field of research (Fangue et al., 2006). In this context, identifying physiological changes such as adaptations enhancing energy efficiency depending on local and regional variability in temperatures (Pörtner, 2006) will help clarify how metabolic adaptations are involved in specialization on the climate regime and may support speciation.

The success or fitness of a species is best measured in terms of its productivity in the field which, in terms of biomass, is determined by rates of growth and mortality and recruitment success. Most commercial fish species produce millions of eggs (Rothschild, 1986), and mortality rates of early life-history stages are very high and variable, generating large fluctuations in the survival of year classes destined to enter the adult population (Houde, 2008). Relatively small changes in rates of growth and mortality during the egg and larval phases can have a great influence on the recruitment success of populations. As early life-history stages are likely to be more sensitive to altered environmental conditions, climate change is expected to have a major effect on the distribution and abundance of fishes through its influence on recruitment. Apart from direct mortality of early life stages due to intolerable conditions, differences in rates of survival will be attributable to match–mismatch dynamics between the timing of reproduction relative to the production of larval food (e.g. G. morhua) and predators (Cushing, 1990; van der Veer et al., 2000; Platt et al., 2003; Temming et al., 2007) or the connectivity (retention or transport) between spawning sites and required nursery areas of early life stages (Sinclair, 1988; Wilderbuer et al., 2002).

Inter-stock comparisons often indicate dome-shaped relationships between recruitment strength and water temperature experienced during the spawning season, with maximum recruitment at an intermediate temperature in both demersal and pelagic fishes (Brander, 2000; MacKenzie & Köster, 2004). The effects of climate on recruitment, however, are not limited to processes acting during the egg and larval stages, but may extend to later (juvenile and adult) life stages and finally adult reproductive performance. For example, survival through the first, critical year of life in some species appears linked to feeding and temperature conditions experienced during the postlarval or juvenile period (Baumann et al., 2007), particularly in fish species that feed exclusively on zooplankton and that can exert strong top-down control and exhaust that prey resource.

Productivity will also be influenced by the effect of temperature on growth rate (Brander, 1995; Teal et al., 2008). In a comparative study of 15 G. morhua stocks, seven-fold differences in the productivity among stocks corresponded to the differences in the temperature of the environment (Dutil & Brander, 2003). This effect adds to the finding of different performance rates (e.g. in growth) of G. morhua populations at the same temperature due to specialization on local or regional climate regimes. Fish species are generally plastic in the age and size at which they become sexually mature. An increase in juvenile growth as well as an increase in temperature may result in a decrease in the length and age at first maturation, affecting the growth of adults as surplus energy is channelled into reproduction at an earlier age and smaller size (Heino et al., 2002). The data available for E. japonicus and S. melanostictus in the Japan Sea indicate that recruitment processes and growth occur within the same species-specific range of temperatures (Takasuka et al., 2007, 2008).

As a consequence of thermal specialization, low productivity and high mortality will occur when temperature conditions reach extreme values. Further factors, however, interfere and interact with temperature. Mortality will be elicited or enhanced during hypoxic, or anoxic conditions, as has been reported for the Kattegat (Diaz, 2001). In some ecosystems, mass mortalities during summer have been reported in relation to harmful algal blooms (Yin et al., 1999; Heil et al., 2001). Climate change may have dramatic (negative) effects on the productivity of fish populations by increasing the frequency of these episodic extreme events that have acute, physiological consequences.

For some broadcast spawning fish species, there is evidence that the size of populations is determined by the size and availability of spawning and nursery habitats (Rijnsdorp et al., 1992; Gibson, 1994; Sparholt, 1996; MacKenzie et al., 2000). Limits on the availability of these habitats may act as a bottleneck for population size (and productivity). In these cases, the focus should be on the effect of climate change on the critical life-history stages. Whether this relationship is applicable will depend on the relative size of the habitat in relation to that of other life-history stages and is determined by the specific geographic setting. Finally, life cycle closure may be affected if climate change influences the connectivity between the habitats of successive life-history stages (Sinclair, 1988; Rijnsdorp et al., 2009). For instance, changes in ocean climate may affect the transport of eggs and larvae between spawning grounds and nursery areas (Corten, 1986; van der Veer et al., 2000; Wilderbuer et al., 2002) or may change the timing of spawning migrations and arrival of adults at the spawning grounds (Sims et al., 2005).

Extending beyond species: community and ecosystem effects

The principles behind species specialization and sensitivity also extend to the community level and ecosystem functioning. Different thermal windows of species probably influence the quality and intensity as well as the seasonal timing of their interactions in an ecosystem (Pörtner & Farrell, 2008). In general, species coexist where their thermal windows overlap and their thermal windows are not necessarily identical. This physiological feature explains why climate sensitivity differs among species and may be one principal reason for climate-induced changes in community composition and food-web interactions. The limited window of whole organism performance capacity will then directly affect the productivity of a species in an ecosystem, as well as indirectly by influencing its interactions with other species (in predator–prey and competitive interactions) in that ecosystem. Alterations in community composition may result as direct effects of temperature on individual species and lead to higher level effects in the ecosystem. For example, a regime shift among copepods in the Calanus genus, from colder water Calanus finmarchicus to warmer water Calanus helgolandicus was correlated with the timing of the decrease in North Sea G. morhua stocks due to reduced food availability for juvenile G. morhua (Beaugrand et al., 2002). In the adjacent Baltic Sea, a similar regime shift occurred among copepods with a reduction in Pseudocalanus acuspus and the increase in smaller, warmer water Acartia and Temora spp. The resulting alteration in predator–prey relationships favours the early (larval) survival of sprat Sprattus sprattus (L.), one of the two dominant clupeid species in the Baltic, over that of larval cod. Adult G. morhua are the major predator of Baltic S. sprattus (Möllmann et al., 2009).

These regime shifts occurring in copepods represent a similar phenomenon as the one described for the regime shift between S. melanostictus and E. japonicus. These examples emphasize that the differential physiological effects of temperature on individual species in relation to their limited window of whole organism performance capacity are key to understanding and projecting climate-induced changes in species interactions and, furthermore, in community composition (Pörtner & Farrell, 2008). The fundamental insight here is that higher level processes (at population, community or ecosystem levels) are driven by the similarities, differences and relationships between niches of the various species that make up a community.

Finally and as exemplified above, climate change is expected to affect a number of different attributes of ecosystems including the dynamics of marine fish prey resources. When the mismatch between predators and prey is not so severe, food-limited (lower) growth rates may be observed which may shift thermal sensitivity or make mortality due to predation more likely. The processes are indeed complex and highly interactive. As a first step, the effects of global warming on marine fish species can be inferred from empirical comparisons made among populations inhabiting environments with different mean characteristics. This would then be the basis for elaborating the contributing mechanisms to patterns observed among populations of a species (Brander, 1995; Drinkwater, 2005).

As a corollary, an overarching pattern in this complexity is the specialization of marine fauna on climate-related temperature windows. This causes sensitivity to temperature extremes, due to decrements in performance (i.e. the capacities to forage, migrate, grow or reproduce). Temperature change is currently the main driving force causing shifts in the geographical distribution of species and in the species composition of marine ecosystems. Increasingly, temperature interacts with other stressors operating at large scales such as carbon dioxide induced ocean acidification and hypoxia in warming, more stratified oceans. As outlined below and elsewhere (Pörtner, 2010), the thermal window comprising temperature-dependent performance and passive resistance may act as a suitable matrix to address such relationships and then build a foundation on which ecosystem-level complexities can be addressed.

Physiological effects of climate change on ecosystems

Oxygen and capacity-limited thermal tolerance

Specialization of marine fauna and their life stages on climate-related thermal windows emerges as one overarching reason for the sensitivity of marine fauna to temperature extremes. As previously discussed, thermal windows differ between life stages of one species as well as between species and thereby co-define changes in higher level processes at ecosystem level in response to climate change (Pörtner & Farrell, 2008). For animals, the concept of oxygen and capacity-limited thermal tolerance (OCLT; Fig. 3) provides a mechanistic explanation of why and how thermal limitation occurs. It also provides access to the differentiation between the thermal range of permanent, active survival through temperature-dependent performance and the range of time-limited passive tolerance to temperature extremes. This concept also explains the transition between those ranges. It is a suitable candidate to define the borders of the niche of functional capacity of a species with respect to temperature and to species interactions (Pörtner et al., 2010).

Figure 3.

Conceptual model of how ocean warming, hypoxia and acidification, as interacting stressors, shape the thermal window of performance of a species at a specific life stage, based on the concept of oxygen and capacity-limited thermal tolerance (OCLT). Oxygen supply to tissues is (a) optimal between low and high pejus temperatures and (b) supports temperature-dependent performance capacity. Critical temperatures (TC) characterize the onset of anaerobic metabolism and denaturation temperatures (Td) the loss in integrity of molecular structures. Critical thermal maxima and minima (TCmax and TCmin) delineate the outermost limits where whole organism function ceases. The kinetic stimulation of performance by warming leads to a tilted curve with a functional optimum close to the upper pejus temperature. This performance curve reflects the role or functional capacity and associated aerobic scope in ecosystem-level processes such as competition, foraging, immune response, growth and behaviours. Ambient hypoxia and elevated carbon dioxide levels both cause a narrowing of thermal windows and possibly, lower performance optima through lower functional capacities and reduced systemic oxygen tensions (inline image). The graph depicts acute performance levels and limitations in response to short-term temperature fluctuations. Seasonal acclimatization causes a shift and changing widths of acute performance windows within the limits of the thermal niche of a species (inline image) (after Pörtner, 2010).

The concept implies that oxygen supply to tissues is optimal between limits called lower and upper pejus temperatures (pejus means getting worse). Between pejus limits, oxygen supply can also be increased to levels above maintenance demand at which point aerobic metabolism can sustain performances such as growth, foraging, migration and reproduction that support the fitness of species. The excess in oxygen availability supporting such performances is reflected in an animal's aerobic scope (the difference between the lowest and highest rates of aerobic respiration) with an optimum close to the upper pejus temperature. These considerations match earlier definitions by Fry (1971) who classified environmental factors based upon their influence on aerobic metabolism and aerobic scope. The latter represents the energy available for all life processes and is, more precisely, the difference between standard levels (or costs) and the rate caused by environmental stimuli or activity increments. According to Fig. 3, temperature co-defines the scope, due to limitations on capacity, and other factors also act to increase or decrease (limit) the aerobic scope.

The kinetic stimulation of performance by warming causes the curve of the relationship to be tilted to the right. The kinetic slowing of performance by cooling below the optimum leads to a lower pejus limit, which is characterized by performance significantly below optimum levels. Some polar fishes live permanently below their acclimated optimum (Lannig et al., 2005; Brodte et al., 2006), indicating that the temperature range below the optimum but above lower pejus temperature is still ecologically viable. Beyond upper pejus limits, oxygen supply capacity becomes limiting, maintenance demand rises and aerobic scope starts to decrease in due course. Hypoxaemia develops and whole organism functional capacity (e.g. exercise or growth performance) declines. Once critical temperatures are reached, extreme hypoxaemia causes a transition to anaerobic metabolism. Finally, thermal damage to the integrity of molecular structures (denaturation) may occur.

Figure 3 distinguishes between the temperature range associated with long-term survival, growth and reproduction due to maintained performance (the active range) and the subsequent time-limited endurance of temperature extremes (passive range). The transition between the two and the loss in performance is progressive, leading to a progressively severe time limitation of thermal tolerance. The point where this transition starts is defined as the pejus temperature (Tp, the temperature at which aerobic scope begins to decline). Organisms would not survive long term at temperatures higher than the upper Tp or below the lower Tp; time constraints become more and more severe as temperatures approach and reach beyond TC where aerobic scope approaches zero. The combination (product) of time and heat load beyond Tp (Fig. 4) would then set the lethal temperature, which indicates the ultimate limits of passive tolerance (Pörtner, 2010). The lethal temperature is thus time dependent, such that short-term heating protocols will lead to higher lethal temperatures than long-term heating protocols. In laboratory studies, lethal tolerance limits of fishes have traditionally been determined as the onset of spasms at critical thermal maxima or minima (TCmax and TCmin). These limits have been found beyond TC values as defined by the OCLT concept, but again, are strongly influenced by the rate of temperature change to which the organism is exposed (Fig. 4). TCmax and TCmin are the temperatures at which fish are definitely unable to escape conditions that will ultimately lead to thermal death. This view already emphasizes the time component, which needs to be systematically considered during evaluation of such limits. During exposures to extreme temperatures beyond Tp, animals will progressively enhance their exploitation of protective mechanisms such as the capacity of anaerobic metabolism (beyond TC), of antioxidative defence and of the heat-shock response. All these contribute to extend the period of passive tolerance. The relevance of these processes varies between species and would be under strong selection in some environments where organisms are regularly exposed to thermal extremes. For example, such passive extension of the range of tolerance appears crucial in thermally variable environments such as the intertidal zone.

Figure 4.

Projection of how lethal tolerance limits depend on the time duration of heat exposure building on the oxygen and capacity-limited thermal tolerance (OCLT concept) (see Fig. 3). The upper pejus limit delineates the onset of time limitation in thermal tolerance. The tolerable heat budget is shaped by the capacity of protective mechanisms and equals the integral of excess temperature or heat experienced above Tp over time [(TTp) × time (h, days or weeks)]. This integral is assumed to remain constant between short-term exposures to high temperatures and long-term exposures to mild temperature changes. The integral may change with acclimation in protective capacity (after Pörtner, 2010). inline image, heat excess tolerated over time [(TTp) h].

Pejus limits reached acutely on short time scales may shift over time when the organism has the capacity to acclimate to temperature change (Lucassen et al., 2006). Acclimation capacity is limited, however, such that during exposure to temperature on long time scales (as in growth studies), performance decrements occur at the limits of thermal acclimation capacity. Beyond these long-term thermal limits, the loss in fitness remains uncompensated and leads to various phenomena at the ecosystem level, i.e. a loss of growth performance and reduced population abundance. Such effects were observed in Z. viviparus in the German Wadden Sea (Pörtner & Knust, 2007), where abundance decreases upon exposure to extreme summer temperatures, in the same range of temperatures where growth performance decreases and the capacity of the circulatory system to supply oxygen becomes limiting. Large individuals are lost first from the population since they are affected most by the heat stress, in line with the exacerbation of oxygen limitation at larger body sizes. The combination of physiological studies in the laboratory with results from field surveys indicates that heat exposure in these fish leads to decreased abundance due to reductions in performance. It remains to be explored whether the largest individuals exposed to temperature extremes exploit their capacity of passive resistance.

While these phenomena identify the physiological processes limiting thermal tolerance at ecosystem levels, these observations clearly indicate the importance for an understanding of why animals specialize on limited thermal environments. Such specialization on a climate regime seems to involve savings of metabolic energy. The benefits of such specialization are so large that temperate zone species shift their thermal windows between winter, spring and summer or between life stages according to their season of existence. Permanent differences between thermal windows have even been detected between adults from various populations of the same species (fishes including G. morhua and marine invertebrates) in a latitudinal cline along the east Atlantic coast between the North Sea and the Arctic. The complexity increases further when considering that thermal specialization differs among life stages and influences the key processes shaping population size and structure, fecundity, growth and mortality.

Available data suggest that aerobic scope, which is limited to the window of tolerance, sets the environmental niche either acutely, e.g. as the niche of the respective life stage, or as a longer-term average of the species. A minimum level of aerobic performance characterizes the limits of this niche and supports minimum functional rates and competitiveness within an ecosystem. The concept of aerobic scope has a long-standing history in comparative physiology and has only recently been demonstrated to underpin climate-induced ecological shifts observed in the field. The concept of OCLT provides explanations of thermally limited aerobic scope. It relates the level of aerobic scope to climate and lifestyle; it explains the width and the tilted shape of the temperature-dependent aerobic performance curve (Fig. 5). Such mechanistic knowledge specifies the more empirical descriptions of thermally limited performance in evolutionary thermal biology (Angiletta, 2009).

Figure 5.

Conceptual model of how standard and maximal metabolic rates fit the thermal window of oxygen and capacity-limited thermal tolerance (projected from mechanistic knowledge as in Fig. 2). The exponential phase of standard metabolism RS is bordered by critical temperatures (TC). The performance curve as in Fig. 3 results from the difference between maximum (Rmax) and RS rates.

Performance patterns and energy budget

Thermal specialization attempting to match the local climate conditions occurs at molecular, cellular and whole organism levels. This has performance implications (e.g. due to changes in metabolic rates and capacity for acid–base regulation) at the level of the whole organism. Species-specific rates of body growth, feeding and special features of behaviour and activity can be expected as overarching consequences of specialization according to climate, leading to species-specific responses to environmental changes associated with climate change. The answers relate to how the available excess in aerobic energy is allocated to specific functions and processes at the whole organism level. Following the second law of thermodynamics, these physiological effects at the organismal level have been addressed using balanced bioenergetics budgets that include specific pathways for energy gain and loss:


where C = energy gained from food consumption, G = the energy gained in growth, R = energy lost via respiration, E = energy lost via nitrogenous excretion and F = the energy lost in egested faecal material (Brett & Groves, 1979). All variables have common energetic units (e.g. joules day−1). Rates of R (and E) can be subdivided to include different levels of energy loss due to basal (standard RS), feeding (specific dynamic action, RSDA), sometimes referred to as apparent heat increment (Blaikie & Kerr, 1996), and active (RA) metabolism. The scope (maximum–minimum) of these rates thus follows the temperature-dependent window of aerobic scope, but the relative magnitude of these rates may change over time (Fig. 6). Growth can also be subdivided to partition between gonadal (GG) and somatic (GS) growth. This balanced equation is most commonly used to estimate either growth or food consumption after the other variables are determined (Hanson et al., 1997). Rates of these various variables are affected by a number of factors including water temperature, fish body size and feeding level (Brett & Groves, 1979). Thus, balanced budgets provide a route for exploring possible effects of environmental changes at the level of the individual. For a true cause-and-effect understanding, however, the question requires the consideration of which variables are dependent variables and, thus, follow from the relationships between climate specialization and climate sensitivity.

Figure 6.

Energy budget for processes beyond maintenance comprising fractions of overall performance fuelled by the temperature-dependent window of aerobic scope (Fig. 3). inline image, shifting fractions of various processes in the energy budget over time (SDA, specific dynamic action) (order and timing arbitrary) (after Pörtner et al., 2010).

The effect of temperature on growth rate (G) follows the aerobic performance curve with net growth performance only possible above and below critical thresholds (Fig. 3). Sensitivity to temperature extremes rises when energy gained from C is much less than that lost via R, E and F (metabolic costs cannot be met by feeding). Above the lower threshold, when C is maximal (well-fed fishes), growth rates increase with increasing temperature until an optimal temperature is reached. With further increases in temperature, G often rapidly declines until mortality occurs. At sub-optimal high temperatures, fishes cannot consume enough food to meet increasing metabolic costs; available evidence suggests that there is not enough aerobic scope available to cover the postprandial increase in oxygen demand (Fig. 3). Losses in appetite occur at the highest temperatures, rapidly leading to mortality. Existing evidence suggests that the availability of aerobic power and its budgeting are crucial for the fitness of a species at the ecosystem level and for co-defining the niche of a species (Pörtner et al., 2010; Guderley & Pörtner, 2010).

For some marine fishes, particularly traditional aquaculture species, relationships between growth rate and temperature have been determined at optimal (ad libitum) food supply (e.g. temperature-growth curves; Fonds et al., 1992; Peck et al., 2003). In the wild, however, some fishes may not always experience unlimited prey resources, food supply may even become limiting to growth, particularly in young larvae (Buckley & Durbin, 2006) that have relatively poor foraging ability or in young juveniles that have high foraging ability and may occur at high enough abundances to exert top-down control of their prey, leading to reduced rates of growth and starvation (Baumann et al., 2007). Food-limited growth, however, may be difficult to detect since these slower growing individuals of early life stages have an increased risk of predation mortality and tend to be removed from fish populations. Furthermore, some fishes may not live at their optimum temperature for the whole year or not even for their whole life as indicated by maximum growth at higher than ambient temperature in an Antarctic fish Pachycara brachycephalum (Pappenheim) (Brodte et al., 2006). Therefore, observed growth rates in the field are quite dynamic both spatially and temporally (particularly in temperate environments), for reasons associated with temperature-dependent kinetic rate limitations or since other energy budget components take precedence and restrict energy allocation to growth. In the case of adult G. morhua, however, food limitations did not become visible in field samples collected in a latitudinal cline of the eastern North Atlantic Ocean. A reduced growth capacity in sub-Arctic compared to southerly populations was rather related to the metabolic costs associated with eurythermal cold adaptation (Pörtner et al., 2001).

Insight from Fig. 3 suggests that growth rate depends on the availability of aerobic scope. Using the terminology of the energy budget (equation 1), a key process related to differences in environmental (e.g. thermal) tolerance is the rate of energy loss due to metabolism (R). The rate of metabolic loss tends to increase exponentially with increasing temperature, between limits set by critical temperatures of the OCLT window (Figs 3 and 5). Climate specialization and the width of thermal window cause R to vary if compared at specific temperatures. Specifically, if standard metabolic rate changes in response to thermal acclimatization or adaptation, then maximum metabolic rate and thus aerobic scope and in consequence R also vary accordingly. This might cause growth rate to vary depending on climate as described above for G. morhua. For a deeper understanding, the mechanistic background of such relationships requires consideration such as that indicated in Fig. 3.

Fry (1971) classified environmental factors based upon their influence on aerobic metabolism and aerobic scope. Thermal or other stimuli either act to increase (unload or release) or decrease (e.g. limit) the aerobic scope. The reasons for these changes can be linked to patterns and processes mentioned in Fig. 3 as well as to the role of temperature in reaction kinetics. At the most basic level, suitable environments are those where an organism's metabolic scope is positive and various controlling factors (e.g. temperature, pressure or water pH and carbon dioxide levels) determine the range of metabolic scope and hence reflect the active tolerance range (Figs 3 and 6). Thus, other factors interact with the controlling effect of temperature and either increase or decrease the metabolic rate and metabolic scope of an organism. Factors like carbon dioxide (ocean acidification) or hypoxia interact with the thermal window of OCLT (Pörtner, 2010). Existing knowledge suggests that these factors elicit metabolic depression that would support passive tolerance to thermal extremes but, at the same time, cause a reduction in aerobic scope. These mechanisms thereby cause a narrowing of thermal performance windows as in Fig. 3 and lead the organism earlier to the limits of its thermal acclimation capacity. In this context, several authors have stated that marine fishes are likely to be less affected by an increase in oceanic carbon dioxide and a corresponding decrease in pH than invertebrate groups such as molluscs and corals (Pörtner, 2008). Synergistic effects of temperature extremes and carbon dioxide, however, have rarely been studied and require consideration when assessing climate sensitivity. Studies in coral reef fishes have confirmed the hypothesis that carbon dioxide levels as expected from anthropogenic emissions would exacerbate the effects of temperature extremes and cause further reductions in aerobic scope (Munday et al., 2009a).

As a corollary, the relationships between the levels of energy turnover, the capacities of activity and other functions and the widths of thermal windows support an integrative understanding of specialization on climate and of sensitivity to climate change. The balance between oxygen demand for maintenance and the level of aerobic scope available to crucial performances between critical temperatures therefore appears essential for defining sensitivity and responses to abiotic and biotic environmental factors (Fig. 6). Pejus limits encompass the optimal range in temperatures (i.e. the temperatures where metabolic scope and thus performance is highest) and therefore delineate the window of highest competitiveness. In line with the predictions resulting from Fig. 3, aerobic scope closely describes the range of performance sensitivity to environmental changes.

As outlined above, the range of tolerable temperatures is both species and often life-stage specific and is a key attribute defining the observed distribution (niches and habitats utilized) by marine fish species. Global warming is not expected to affect all life stages of marine fishes in the same manner. For example, evidence for S. solea and G. morhua suggests that the optimum temperatures for growth decrease with increasing body size of juveniles and adults (Pörtner et al., 2008). Decreases in oxygen-limited tolerance to warm temperatures with increasing body size can be used to explain the ontogenetic shift in habitats observed in several fish species where young stages occur in shallower (warmer) waters than larger stages. In some species, egg and larval stages have a narrower range in tolerable temperatures than other life-history stages (particularly juveniles), making these early life stages more vulnerable to the effects of global warming (Figs 2 and 7). As outlined above, the sensitivity of larval stages to global warming probably relates to the ontogenetic development of convective oxygen transport during a period of when diffusive oxygen supply becomes limiting (Pörtner et al., 2006). In larvae, behavioural compensation is constrained by limited swimming abilities (particularly in temperate and polar species), making individuals less capable of selecting and migrating towards more suitable habitats. Furthermore, their relatively high (mass-specific) metabolic rates and lower energy reserves make these young life stages more vulnerable to mortality due to starvation during periods of adverse environmental condition (i.e. periods of food shortage). Another sensitive life stage is the mature spawner, where the egg mass is the largest tissue to be supplied with oxygen, which will narrow its tolerance range to temperature fluctuations due to the allometry of oxygen limitation. This may contribute to setting relatively low temperature limits to G. morhua during their winter and spring spawning (Pörtner et al., 2008) and also contribute to limiting the heat tolerance of Pacific salmon Oncorhynchus spp. during their spawning migrations (Farrell et al., 2008). In line with Fig. 2, Fig. 7 outlines these patterns expected from the principles of the OCLT concept.

Figure 7.

Ecosystem-level perspective of how (a) thermal windows develop during ontogeny in relation to seasonal temperature dynamics (as in Gadus morhua, for further examples see Fig. 3). Furthermore, the synergistic effects of shifting temperatures, carbon dioxide and hypoxia levels affect species interactions, starting from (b) undisturbed conditions through (c) changes in temporal and thermal overlap as well as in relative performance (after Pörtner & Farrell, 2008).

In juvenile and (young) adult life stages, individual fishes can clearly respond to a change in temperature and in many cases exhibit active temperature preference. Some elasmobranchs respond to quite subtle changes in temperature (i.e. to changes of <0·001° C; Brown, 2003). Naturally, even for individuals in these later life stages that have strong swimming capacity, information concerning the location of optimal habitats will not be perfect. Moreover, behavioural changes may have unexpected consequences. For example, as a result of an increase in temperature, the swimming speed of G. morhua and other species increases (Peck et al., 2006) and animals may behave differently in response to oncoming fishing gear making them more (or less) vulnerable to capture. Finally, over longer time periods, climate-driven changes in temperature can modify the phenology of annual migrations to feeding and spawning grounds as observed and predicted for temperate marine species. It can be inferred that the behavioural response to changes in environmental conditions brought on by global warming or ocean acidification (Munday et al., 2009b) will depend on the rate of change and the spatial scale over which conditions change relative to fish body size or developmental stage since the capacity for avoidance responses will increase with increasing body size.

Changes in species interactions and competitiveness at the ecosystem level may also reflect these principles and may result from relative changes in performance due to different thermal sensitivities (Fig. 7; Pörtner & Farrell, 2008). The regime shifts described then not only relate to the principle effects of temperature according to thermal windows. Shifts may be accelerated by the relative changes in performance and the associated shift in competitiveness once species compete for the same resource or ecological niche.

Physiology-based projections of effects

Projections have been made of the effects of global warming on marine fishes using a variety of methods that range from simple to complex, each of which has its own caveats and only some of which incorporate physiological mechanisms. Most projections of the effects of global warming on fish populations, aside from simple what if hypothesis testing, rely on the outputs of global climate models (GCM).

In the most simple example, outputs from GCMs are used to adjust environmentally sensitive stock–recruitment relationships which are essentially functions estimating year class survival (recruitment) from key population and environmental factors (e.g. spawning stock biomass and temperature), as in the dome-shaped relationships described above. Such predictions do not reveal mechanisms but, nonetheless, provide short-term estimates of population strength in the face of climate effects, all other elements of the ecosystem being equal.

‘Bioclimatic envelope modelling’ represents a second approach which is more physiology-based, in that it makes projections based upon changes in environmental conditions correlated with the survival (or current distribution) and productivity of individual species (Beaumont et al., 2005). Projected global or regional changes in key physical features of marine habitats (water temperature and circulation patterns) in various ecosystems have thus been translated into broad-brush projections of changes in distribution (and productivity) based upon a species' bioclimate envelope. Recent modelling results for >800 commercially important marine fish species (Cheung et al., 2009) suggested that climate change could lead to numerous local extinction events by the year 2050, particularly in subpolar regions, the tropics and semi-enclosed seas (e.g. the Mediterranean Sea), with the distribution of pelagic fish species (Clupeiformes) and demersal species [such as G. morhua, haddock Melanogrammus aeglefinus (L.) and flatfishes] moving pole-wards by an average of 600 and 223 km, respectively. Similar to the simple stock–recruitment relationships described above, however, such predictions do not take into account other changes that may occur in marine habitats that place constraints on the survival and persistence of populations of marine fishes.

Biophysical individual-based models (IBM) represent a third example of methods available for climate model-based projections. To date, most IBMs constructed for marine fishes have been used to help understand the processes affecting the survival and growth of early life stages and recruitment success (Miller, 2007). Many IBMs are physiology-related and estimate growth and survival using balanced bioenergetics and foraging potential (Peck & Daewel, 2007; Hinrichsen et al., 2010). Recently coupled ecosystem-larval fish IBMs have been developed that allow climate-driven changes in factors that may directly (water currents and temperatures) and indirectly (prey productivity) affect fish survival and growth to be included (Daewel et al., 2008, in press). Finally, end-to-end biophysical models are now being applied that include juvenile and adult behaviour and growth bioenergetics to assess climate-driven effects in zooplanktivorous species such as Pacific herring Clupea pallasii Valenciennes (Kishi et al., 2007). The reliability of this modelling approach to project future results, however, depends upon obtaining robust estimates of hydrographic changes resulting from global warming. The latter is a rapidly advancing field of climate research. Furthermore, such modelling efforts would benefit from a robust understanding of the mechanisms shaping climate-dependent evolution, specialization on climate and the associated trade-offs and constraints in adaptation.

As a corollary, future studies need to move beyond correlative approaches and should attempt to reveal a cause-and-effect understanding of the underlying mechanisms; otherwise, consensus on the ecological implications of climate change will remain elusive. Such a cause-and-effect understanding needs to be improved. Ideally, investigations need to be integrative and address how mechanisms interact across various levels of biological organization, from genome to molecule, cell, tissue, organism and ecosystem. For an assessment of fitness in the natural habitat, these approaches should include an assessment not only of performance (and performance variability) in individual species but also of relative changes in performance of interacting species and of the respective consequences at ecosystem levels (Pörtner & Farrell, 2008). This will advance a mechanism-based understanding of the constraints and trade-offs in adaptation as well as of the limits in acclimatization and adaptation processes, which are closely linked to ecosystem-level phenomena and responses to climate change (Pörtner & Knust, 2007; Farrell et al., 2008). Such knowledge would need to be extended and applied to many more examples than are currently available. This conceptual framework also needs to be considered and, when possible, integrated within modelled projections, both for improved reliability of those projections and for developing further hypotheses for a deepening of the integrative understanding.

This work is a contribution to the European Project on Ocean Acidification (EPOCA), which received funding from the European Community's Seventh Framework Programme (FP7/2007–2013) under Grant Agreement Number 211384. It is was also supported by the BIOACID programme funded by the German Ministry of Research and Education and ECODRIVE (Ecosystem Change in the North Sea: Processes, Drivers, Future Scenarios), an E.U. MarinERA project.


Table APPENDIX..  Table summarizing the critical or lethal thermal maxima and minima for various species of teleosts, together with other relevant information including acclimation temperature (acclim.). Life stages include larvae (L), juvenile (J) and adult (A). Fish types include freshwater (F), marine (M), estuarine (E) and anadromous (A)
TeleostLife stage (L, J, A)TypeLatitude of field collection (° N or ° S)Acclim.Temperature (° C)Preferred meanReference
SpeciesCommon nameMini-mumMaxi-mumMini-mumMaxi-mum
  1. np, not provided.

Apogon pacificiCardinalfishAM3.026.510.835.2   Graham (1971), Mora & Ospina (2001, 2002)
Plagiotremus azaleusSabertooth blennyAM3.026.513.438.2   Mora & Ospina (2001, 2002)
Cirrhitichthys oxycephalusPixy hawkfishAM3.026.511.435.4   Mora & Ospina (2001, 2002)
Bathygobious ramosusGoby sp.AE3.026.512.035.8   Mora & Ospina (2001, 2002)
Coryphopterus urospilusRedlight gobyAE3.026.512.336.0   Mora & Ospina (2001, 2002)
Haemulon steindachneriChere-chere gruntAM3.026.513.238.1   Mora & Ospina (2001, 2002)
Halichoeres dispilusChameleon wrasseAM3.026.515.639.7   Mora & Ospina (2001, 2002)
Thalassoma lucasanumCortez Rainbow WrasseAM3.026.516.240.1   Mora & Ospina (2001, 2002)
Malacoctenus zoniferGlossy blennyAM3.026.512.537.3   Mora & Ospina (2001, 2002)
Lutjanus guttatusSpotted rose snapperAM3.026.512.035.9   Mora & Ospina (2001, 2002)
Eucinostomus gracilisMojarraAM3.026.512.536.0   Mora & Ospina (2001, 2002)
Mugil curemaWhite mulletAM3.026.510.734.7   Mora & Ospina (2001, 2002)
Chromis atrilobataScissortail damselfishAM3.026.511.235.3   Mora & Ospina (2001, 2002)
Stegastes acapulcoensisAcapulco damselfishAM3.026.512.637.5   Mora & Ospina (2001, 2002)
Apogon doviiTailspot cardinalfishAM3·0 and 9·026.513.137.8   Mora & Ospina (2001, 2002)
Apogon novemfasciatusNine-banded cardinalfishnpM5.517–3215.438.0   Eme & Bennett (2009)
Bathygobius fuscusDuskyfrill gobynpE5.520–3511.141.2   Eme & Bennett (2009)
Bathygobius sp.Sandflat gobynpE5.517–3412.441.4   Eme & Bennett (2009)
Liza vaigiensisSquaretail mulletnpM5.520–3513.042.2   Eme & Bennett (2009)
Dascyllus aruanusWhite-tailed humbugnpM5.517–3114.338.0   Eme & Bennett (2009)
Horabagrus brachysomaYellow catfishnpF8.015–3615.240.0   Dalvi et al. (2009)
Apogon maculatusSpotted cardinalfishAM9.026–2813.037.7   Graham (1971)
Bathygobius ramosusPanamic frillfinAE9.017–289.539.5   Graham (1971)
Bathygobius soporatorFrillfin gobyAE9.026–2810.140.9   Graham (1971)
Abudefduf saxatillisSergeant majorAM9.026–2811.639.3   Graham (1971)
Abudefduf troscheliiPan. sergeant majorAM9.026·5–2911.939.4   Graham (1972)
Ambasis commersoniGlass fishAE11.528 40.5   Rajaguru & Ramachandran (2001)
Lates calcariferBarramundiAE11.528 44.5   Rajaguru & Ramachandran (2001)
Liza dussumeriMullet sp.AM11.528 44.5   Rajaguru & Ramachandran (2001)
Etroplus suratensisPearlspotAM11.520–35 42.3   Rajaguru & Ramachandran (2001)
Scatophagus argusSpotted scatAM11.528 42.5   Rajaguru & Ramachandran (2001)
Siganus javusStreaked spinefootAE11.528 39.5   Rajaguru & Ramachandran (2001)
Therapon jarbuaThree-striped tigerfishAM11.520–35 42.0   Rajaguru & Ramachandran (2001)
Ambassis kopsiiFreckled hawkfishnpM13.322·8 and 26·0  38.638.6 Menasveta (1981)
Apogon auteusRing-tailed cardinal fishnpM13.323·0–27·3   38.4 Menasveta (1981)
Chaetodon rostratusButterfly fishnpM13.323 and 29  37.837.8 Menasveta (1981)
Lutianus melabaricusSnappernpM13.323–28  37.037.0 Menasveta (1981)
Lutianus vittaBlack-striped snappernpM13.323–28  37.037.0 Menasveta (1981)
Monocanthus chinensisFile fishnpM13.322·5–28·0  35.235.2 Menasveta (1981)
Mugil dussumeriiMulletnpM13.322·5–29·5  38.440.2 Menasveta (1981)
Plotosus anguillarisSea catfishnpM13.325 and 28  38.238.2 Menasveta (1981)
Halichocres nigreceusParrot fishnpM13.322·7–28·0  37.635.0 Menasveta (1981)
Epinephelus tauvinaGreasy groupernpM13.322·7–28  37.637.6 Menasveta (1981)
Siganus oraminSpinefootnpM13.322·7–25·0  35.935.9 Menasveta (1981)
Sillago sihamaSilver sillagonpM13.322·8 and 29·5  37.837.8 Menasveta (1981)
Therapon therapsLargescaled teraponnpM13.322.7   41.2 Menasveta (1981)
Cyprinodon dearborniOcellated killifishJ and AE18.0   14.037.0 Brett (1970)
Ocyurus chrysurusYellowtail snapperJM18.020–32   32.827.4Wallace (1977)
Poecilia sphenopsMollyJ and AF18.0   15.037.0 Brett (1970)
Rivulus marmoratusMangrove rivulusJ and AE18.0   14.037.0 Brett (1970)
Haemulon flavolineatumFrench gruntnpM18.027   36.0 Sylvester (1973)
Labeo rohitaRohoJF18.925–3514.041.3   Chatterjee et al. (2004)
Cyprinus carpioCommon carpJF18.925–359.141.0   Chatterjee et al. (2004)
Anabas testudineusClimbing perchnpF18.925–3513.141.1   Sarma et al. (2010)
Canthigaster jactatorWhitespotted tobyJ/AM19.4np    27Casterlin & Reynolds (1980)
Kuhlia sandvicensisHawaiian flagtailJ and AM20.023  13.834.4 Brett (1970)
Catla catlaCatlaJF21.030–3815.242.4   Das et al. (2004)
Pagrus majorRed seabreamnpM22.09–28  11.027.8 Woo & Fung (1980)
Pangasius pangasiusYellowtail catfishJF22.630–3814.743.5   Debnath et al. (2006)
Tilapia mossambicaMozambique tilapianpF25.022–36  16.037.6 Allanson & Noble (1964)
Xiphophorus maculatusPlatyfishJ and AF25.415–3012.940.4   Prodocimo & Freire (2001)
Centropomus undecimalisSnookJ and AF26.015–2512.7 10.1  Shafland & Foote (1983), Howells et al. (1990)
Sardinops sagaxPacific sardineJM28·0 and 34·015·5 and 20·56.331.3  17.9Martínez-Porchas et al. (2009)
Heteropneustes fossilisIndian stinging catfishAF29.016 and 29  6.038.630.2Vasal & Sundararaj (1978)
Fundulus heteroclitus heteroclitusCommon mummichogAE30.52–343.238.2   Fangue et al. (2006)
Atherinops affinisTopsmelt silversideL, J and AE33.014·5–25·5  10.631.125.2Doudoroff (1945), Ehrlich et al. (1979)
Clinocottus analisWoolly sculpinnpM33.011–23 (1970)
Fundulus parvipinnisCalifornia killifishJE33.020  3.0  Doudoroff (1945)
Girella nigricansOpaleyenpM33.012–28  8.630.1 Doudoroff (1942)
Scomber japonicusChub mackerelJ and AM33.016.5  6.029.0 Schaefer (1986)
Sebastes dalliCalico rockfishnpM33.013 and 16 25.9  13Shrode et al. (1982)
Rypticus nigripinnisBlackfin soapfishAE33.017–288.640.0   Graham (1971, 1972)
Hypsoblennius gilbertiRockpool blennyLM34.011–23  6.528.822.2Graham (1970)
Engraulis mordaxNorthern anchovyLM34.012–24  10.029.2 Brewer (1976)
Engraulis mordaxNorthern anchovyJ and AM34.08–24  12.033.5 Brewer (1976)
Citharichthys stigmaeusSpeckled sanddabJ and AM34.010–19    11.1Ehrlich et al. (1979)
Cheilotrema saturnumBlack croakerJ and AM34.017    27.6Ehrlich et al. (1979)
Ictalurus punctatusChannel catfishJF34.520–306.338.5   Currie et al. (1998)
Micropterus salmoidesLargemouth bassJF34·5 and 45·88–327.136.9   Currie et al. (1998), Fields et al. (1987)
Fundulus heteroclitusCommon mummichogJ and AE37.05–36 38.2 32.325Bulger & Tremaine (1985), Garside & Chin-Yuen-Kee (1972), Garside & Morrison (1977)
Oncorhynchus mykissRainbow troutJA37.010–200.729.0   Currie et al. (1998)
Gillichthys mirabilisLongjaw mudsuckerJ and AE38.0 13or27 38.2  21de Vlaming (1971)
Brevoortia tyrannusAtlantic menhadenLM38.07–15  3.8  Lewis (1965)
Menidia menidiaAtlantic silversideAE39.07–28 Hall et al. (1982), Brett (1970)
Micropogonias undulatusAtlantic croakerL and JE39.08  3.0  Lankford & Targett (2001)
Sphoeroides maculatusNorthern puffernpM40.010–28  10.330 Brett (1970)
Lucania parvaRainwater killifishnpF42.035   37.5 Brett (1970)
Apeltes quadracusFourspine sticklebacknpE42.025–30   31.5 Brett (1970)
Gobiesox strumosusSkilletfishnpM42.020   31.5 Brett (1970)
Tautogolabrus adspersusCunnernpE42.01–22  2.827.5 Brett (1970)
Oncorhynchus tshawytschaChinook salmonJA42.020 and 24 25.1  13Brett (1970)
Pseudopleuronectes americanusWinter floundernpM42.07–28  2.725.618.7Brett (1970)
Alosa pseudoharengusAlewifeJE42.45–25 31.07.0 22.8Otto et al. (1976)
Clinocottus globicepsMosshead sculpinnpM43.021–37   26.0 Brett (1970)
Leptocottus armatusPacific staghorn sculpinnpM43.021–37   29.5 Brett (1970)
Oligocottus maculosusTidepool sculpinnpM43.021–37   26.5 Brett (1970)
Fundulus heteroclitus macrolepidotusCommon mummichogAE44.32–343.836.4   Fangue et al. (2006)
Coregonis clupeaformisLake whitefishJF44.95–22·5   24.5 Edsall & Rottiers (1976)
Anguilla rostrataAmerican eelJ and AA45.05–20    18.5Haro (1991)
Gobius paganellusRock gobynpE45.022–25   31.8 Brett (1970)
Crenilabrus ocellatusNo common namenpE45.022–25   32.8 Brett (1970)
Mullus barbatusRed mulletnpM45.022–25   32.0 Brett (1970)
Mullus surmuletusStriped red mulletnpM45.022–25   30.7 Brett (1970)
Scorpaena porcusBlack scorpionfishnpM45.022–25   32.5 Brett (1970)
Box salpaSalemanpM45.022–25   32.5 Brett (1970)
Sargus vulgarisTwo-banded seabreamnpM45.022–25   32.8 Brett (1970)
Morone saxatilisStriped bassJE45.55–30  6.931.624·5–36·1° NCook et al. (2006), Coultant et al. (1984)
Gadus morhuaAtlantic codJ and AM47.2     5.5Despatie et al. (2001)
Lepidopsetta bilineataRock soleJ and AM48.06·9–13   24.9 Ames et al. (1978)
Parophrys vetulusEnglish soleJ and AM48.06·9–13   26.1 Ames et al. (1978)
Oncorhynchus kisutchCoho salmonJA49.25–233.324.2  11Brett (1952), Konecki et al. (1995)
Oncorhynchus ketaChum salmonJA52.05–233.823.0  14Brett (1952)
Oncorhynchus nerkaSockeye salmonJA52.05–233.124.2  13Brett (1952)
Zoarces viviparusCommon eelpoutnpM54.012 28.8   Zakhartsev et al. (2003)
Pleuronectes flesusEuropean floundernpM55.012–14   29.0 Brett (1970)
Pleuronectes platessaEuropean plaicenpM55.012–14   26.514.2Jobling (1981)
Clupea harengusAtlantic herringL and AM56.07.5–15  −1.2520.1 Blaxter (1960), Brawn (1960)
Pomatoschistus minutusSand gobyAE60.05–17 28.8  14.3Hesthagen (1979)
Salvelinus alpinusArctic charrnpA63·5–70·00·5–20 22.5−1.0 11·4 and 10·2Baroudy & Elliott (1994), Larsson (2005), Mortensen et al. (2007)
Pagothenia borchgrevinkiBald notothennpM78.0−1.9  <−2.56.0 Somero & DeVries (1967)
Trematomus bernacchiiEmerald rockcodnpM78.0−1.9  −2.56.0 Somero & DeVries (1967)
Trematomus hansoniStriped rockcodnpM78.0−1.9  <−2.56.0 Somero & DeVries (1967)