Metabolic rate in fishes: definitions, methods and significance for conservation physiology

There is a growing interest in the potential for physiological tools and know-how to be applied in conservation research, the emerging field of conservation physiology. For fishes, there is a long-standing hypothesis whereby their bioenergetics and respiratory physiology should be of particular significance for their ecological performance (Fry, 1971). Water can be challenging as a respiratory medium, because it only contains a few milligrams of oxygen per litre (Schmidt-Nielsen, 1990). The physiology and oxygen demands of ectothermic fishes can be profoundly affected by environmental temperature. Oxygen can also easily become depleted in aquatic habitats, and such episodes of hypoxia are becoming more prevalent because of human effects (Diaz Breitburg et al., 2009). It has been argued, therefore, that the capacity of fishes to provide oxygen for vital processes and important activities, such as swimming or digesting and assimilating food, may define the habitats they can colonize successfully (Fry, 1971). As such, there is interest in how traits of metabolism and energetics may be applied in conservation physiology. Such applications can be direct, in experimental contexts, and also indirect, to parameterize models to project the potential effects of environmental changes, such as eutrophication, contamination and ongoing climate change. A core element of energetics is, of course, the metabolic rate that, in fishes, is typically measured as rate of oxygen uptake. There has been a surge in interest for measuring oxygen uptake in fishes but newcomers to the field may well find the terminology and definitions for major metabolic traits confusing. Further, methods and protocols are dispersed throughout the literature, without much integration and comparison. This special issue (SI) arose to respond to these issues, following conversations among fish ecophysiologists at conferences organized within the context of the European Cooperation in Science and Technology (COST) Action Conservation Physiology of Marine Fishes (COST FA1004). The first objective of this SI was to provide literature reviews for the major metabolic traits and recommend appropriate techniques to measure and calculate different metabolic rates. This includes instructions to construct appropriate experimental setups and software to analyse the data. That is, the aim was to make this SI the essential guide that ecophysiologists can consult before planning experiments to measure the metabolic rate of fishes. To this effect, the SI provides 13 papers that explain why the metabolic rate of fishes can be estimated by measuring oxygen uptake, and discuss the different levels of metabolic rate that are of interest, and terminology to deal with metabolic capacity and shortages of dissolved oxygen (hypoxia). These papers provide detailed instructions for proper design of equipment and protocols to measure these metabolic traits accurately and reliably in fishes in the laboratory, including fish larvae and fishes with bimodal respiration (that can extract oxygen from both water and air). Techniques to estimate the metabolic rate of free-ranging fishes are also reviewed. A second objective was to provide perspectives on the ecological significance of metabolic rate in fishes and review some examples of how the respiratory physiology and metabolism of fishes can be used in a conservation context. This comprises five papers, starting with a review of why the metabolic rate of fishes is of ecological relevance, followed by a perspective arguing for the continuing relevance of the oxygen capacity limited thermal tolerance (OCLTT) hypothesis. There are reviews of aspects of how the respiratory physiology and performance of salmonids can be of significance for conservation research, and the section concludes with a discussion on integration of fish respiratory metabolism into models. The SI concludes with three research articles applying respiratory metabolism in studies with conservation objectives. Authors should note that the SI does not follow this Journal's conventions for some abbreviations used. The SI Editors chose to use the most widely used terms, such as standard metabolic rate (SMR) or active metabolic rate (AMR) that, in Journal convention, would be Rs and Ra, respectively. This decision was made because many of the review papers dealt with changes in terminology over time and needed to track these using the original authors' terms. Physiologists typically use oxygen uptake when they want to measure the metabolic rate of fishes, without a second thought, generally not bothering to convert their results into energy units. Nelson (2016) gives a thorough historical account of the study of metabolic rate, particularly in fishes. This is a must-read for anyone interested in measuring energy-use in animals in general. Aerobic energy production is so much more efficient than anaerobic production and usually makes up the bulk of energy production in animals, including fishes. Further, respirometry is much easier to perform than direct calorimetry, particularly in water, and almost all studies of metabolic rate in fishes used respirometry as a proxy for metabolic rate. Nelson (2016) then discusses the error when assuming that all of the energy used by fishes is produced by aerobic processes, and foresees gains in the understanding of the relative contribution of aerobic and anaerobic energy production, in different situations, as better equipment becomes available to perform direct calorimetry on fishes. Svendsen et al. (2016a) compare three methods used to measure oxygen uptake (O2) in fishes and argue that intermittent-flow (or stop-flow) respirometry offers the most advantages and should always be preferred. The rest of this review paper makes recommendations for the design of respirometers and computation of O2. The assumptions of the method are considered. The authors discuss the sources of noise and errors and common problems. In particular, the importance of mixing the water in the respirometer well and the need to control and correct for background respiration (including the methods to do so). The authors provide detailed instructions to build a low-cost but high-performance intermittent-flow respirometry system and describe free software, written by one of the authors, to run respirometry experiments and compute O2 in real-time. Two further papers deal with the design of intermittent-flow respirometers. First, Svendsen et al. (2016b) deal with the issues of respirometer size and of the relative contributions of experimental error v. biological variability to the variability observed in consecutive measurements of O2. Obviously, the decrease in dissolved oxygen caused by a small fish inside a very large respirometer will be more difficult to measure accurately than that of the same small fish in a small respirometer. The authors compared different ratios of respirometer volume with fish volume and make recommendations for which ratios of respirometer volume to fish volume provide reasonable precision in O2 measurements. In addition, Svendsen et al. (2016b) designed a system that replicates the O2 of a fish in a respirometer, without the need for a fish! This allowed them to control the rate of oxygen decrease (i.e. the rate of O2 by the simulated fish) and compare this with results from a real live fish with a similar O2. Their results suggest that most of the variability observed in the studies of O2 in fishes is real biological variability with only a small proportion being technical experimental error. Poor methodology can, however, increase experimental error. Thus Rogers et al. (2016) demonstrate that insufficient mixing of water inside the respirometer is a major source of problems. O2 was underestimated in the absence of a mixing device, demonstrating that the presence of a fish in a respirometer is not always sufficient to prevent water stratification. It was not possible to measure background respiration of an empty chamber accurately without mixing. This paper confirms the need to use an efficient mixing device in intermittent-flow respirometry and to correct for background respiration. Chabot et al. (2016b) review the concept and terminology of minimum sustainable level of O2 in fishes, which is properly called the standard metabolic rate (SMR). SMR is central to fish energetics. It is an obligatory expense, on top of which all other costs are added. It is required to calculate the metabolic aerobic scope (AS), a measure of metabolic capacity that is useful to understand the effects of the environment on fish physiology, e.g. within the OCLTT paradigm (Farrell, 2016). The authors describe the different experimental conditions and approaches that can be used to measure SMR. One of the most frequently used methods, because of simplicity and low cost, is to place fishes in a respirometer, maintain conditions that should result in calm inactivity, but without any simultaneous measurement of locomotor activity. Such experiments result in variable levels of O2 and authors then typically use a relatively small number of the lowest values to estimate SMR. Chabot et al. (2016b) point out that with the very short durations over which O2 can be measured with modern instruments, and considering potential physiological variability within the animal, a proportion of measurements of O2 will in fact be below SMR. The paper compares many different calculation methods to estimate SMR with this type of data and makes recommendations as to which ones should be used or avoided. They provide a script in the R language that calculates all the methods discussed in the paper. Along with SMR, maximum metabolic rate (MMR) is the other variable required to assess the AS of fishes. Norin & Clark (2016) offer a critical historical review of the methods used to measure MMR in fishes. They demonstrate that the methods have to be adapted to the fishes' propensity to swim for extended durations in swim flumes (tunnels) and to simultaneously digest food and swim. The importance of accurately measuring MMR is emphasized. The review summarizes the results from a large number of studies on many species of marine fishes. In addition, the authors address issues such as the effects on MMR of body size, temperature, oxygen availability and even less often studied variables, such as salinity, reproduction and circadian and seasonal cycles. Additional topics include variability in MMR between individuals and species and the possible link between MMR and SMR. Finally, the authors relate MMR of fishes to lifestyle and ecology, anticipating progress in this area with the advent of accelerometer and heart rate telemetry (Metcalfe et al., 2016a). The postprandial increase in oxygen uptake, known as specific dynamic action (SDA), represents the cost of food assimilation. Measuring SDA makes it possible to compare the assimilation cost of different food items or different rations, in different environmental conditions (e.g. temperature and dissolved oxygen). SDA is relatively easy to measure in fishes that remain quiescent in respirometers. Many fishes, however, can be active at some point during the long experiments that are required to measure SDA. In the past, the best that could be done was to manipulate the experimental conditions with the aim of reducing such activity, but any activity nonetheless increased O2 in a manner that was inextricable from the SDA. Chabot et al. (2016a) report on a statistical technique, non-parametric quantile regression, which makes it possible to track the increase in O2 following the ingestion of a meal without considering high values that are caused by stress or activity. Removing the influence of activity not only provides a more accurate estimate of SDA but also reduces any variation among individuals that is caused by different levels and patterns of activity, thereby improving the sensitivity of comparisons between, for example, different treatment groups. The authors provide R scripts facilitating the use of this technique. As with older fishes, it is necessary to measure the metabolic rate of larval fishes to improve the understanding of the relationships between environmental factors and their physiology. Peck & Moyano (2016) show that measuring the O2 of fish larvae is fraught with specific difficulties, owing to their small size. The authors provide a historical review of the methods and results of 63 studies involving over 50 fish species, indicating which type of metabolic rate was measured (standard, routine or active), the effect of larval size (allometric relationships) and the effect of temperature (Q10). Methods have to be adapted, and respirometer design and materials are of the outmost importance. The authors make recommendations on methods and quality control that will facilitate inter-studies comparisons in the future. One of the more unusual aspects of fish respiratory physiology is the existence of many species with bimodal respiration. Lefevre et al. (2016a) review measurements of O2 by such fishes. Measuring O2 from aquatic and aerial phases, simultaneously, poses a number of challenges and, until recently, this involved direct intervention and manipulation by the experimenter. This is unsatisfactory for bimodal fishes, because disturbance inhibits spontaneous surfacing behaviours. The review describes two methods to perform automated intermittent-flow respirometry based around a straightforward design of bimodal respirometer. A simple method is described that requires little equipment but offers less control over potential errors, compared with a more technically refined method that requires some knowledge of electronics and digital control systems. The review also describes the challenges of measuring bimodal respiration by air-breathing fishes during aerobic exercise, with guidance for potential pitfalls. It concludes with perspectives on how advances in probe technology might lead to new research avenues in the study of this fascinating group of fishes. Hypoxia is often treated as a presence or absence variable. Either dissolved oxygen is below a threshold and the habitat is hypoxic, or it is above and the habitat is normoxic. Although categorization can be useful in assessing the progression of hypoxia in time, it greatly underestimates the true effect of hypoxia on fishes, by ignoring differences in sensitivity among species and individuals, and also the many non-lethal effects of chronic hypoxia. Claireaux & Chabot (2016) use the AS concept of Fry (1971) to show that the metabolic capacity can be reduced even at mild levels of hypoxia, with progressively more debilitating effects as hypoxia progresses. This is illustrated by the limiting oxygen level (LOL) curve. This curve shows that when fishes are exposed to hypoxia when their oxygen uptake is relatively elevated, they will appear to be oxygen conformers, such that the difference between oxygen conformers and oxygen regulators may be an artefact of the method used, in particular, the success with which the experiments achieved calm, quiescent fishes before manipulating dissolved oxygen. The authors also present a commonly used measure of hypoxia tolerance in fishes, the critical oxygen level or O2crit, providing scripts in the R language to calculate and plot the results. Finally, the review considers how the concept of degree days, so common in plant ecology, can be used to understand how long fishes survive once oxygen levels drop below O2crit. Snyder et al. (2016) compare two methods commonly used to decrease dissolved oxygen, within a respirometer, during experiments to measure O2crit. Some studies simply keep the respirometer closed for the entire duration of the experiment, which obviously results in progressive hypoxia and hypercapnia. The alternative is to continue to flush the respirometer, but to control the oxygen content of the flushing water, reducing it in a stepwise manner. The authors provide an experimental comparison of both methods, with the shiner perch Cymatogaster aggregata Gibbons 1854. They recommend stepwise control of flush water, because C. aggregata were less hypoxia tolerant when they decreased the oxygen content by themselves, probably because of the more rapid rate of decrease of oxygen and of the inevitable accumulation of metabolites, mostly carbon dioxide, in the closed respirometers. Training early stage researchers to use adequate equipment and techniques to measure the oxygen uptake of fishes is of paramount importance. Rosewarne et al. (2016) provide step-by-step instructions to teach respirometry techniques tailored to the measurement of oxygen uptake and the impact of hypoxia in fishes. The authors provide a list of materials required to perform intermittent-flow respirometry, a detailed protocol to measure SMR, MMR and AS, and explain the calculations involved in transforming the changes in oxygen content in the respirometer into oxygen uptake, including how to correct for background respiration. Finally, they explain how the students can investigate the effect of declining oxygen levels on these metabolic variables. Altogether, the paper constitutes an excellent tutorial allowing readers to put together the techniques and recommendations from many of the papers in the SI. At present, it is a major challenge to estimate the metabolic rate of fishes in their natural habitat, especially marine fishes in their vast underwater realm. Metcalfe et al. (2016a) review current methods for estimating metabolic rate indirectly, by calibration against other physiological variables. They focus on electronic devices that record body acceleration in three dimensions (accelerometry). Correlations between dynamic body acceleration and rate of oxygen uptake can be developed, to then estimate activity-specific energy expenditure of fishes in semi-natural or natural habitats, over extended periods. This technique can, therefore, be applied in conservation physiology studies aimed at understanding the behavioural and energetic decisions that wild fishes may make as a function of environmental conditions. This is an area of conservation physiology of fishes that is expected to expand rapidly. Metcalfe et al. (2016b) review current knowledge of the extent and causes of individual variations in metabolic phenotype (SMR, MMR and AS) of fishes and the links between metabolic phenotype, performance and behaviour. For instance, SMR can be related to behavioural traits, such as dominance, risk taking and growth rate, and fishes with high SMR may be advantaged when food is abundant, but disadvantaged during periods of food shortage. Such trade-offs are probably involved in the maintenance of diversity in metabolic phenotypes through genetic diversity. Considering possible effects of MMR on the ability to catch prey and avoid predators and of AS with performance in hypoxic conditions, the authors emphasize the need for more research relating MMR and AS to behaviour and fitness of fishes. Farrell (2016) provides a perspectives paper in support of the Fry (1971) AS paradigm, and its derivative the OCLTT hypothesis, as universal frameworks to explain how temperature constrains physiological performance in fishes. The Fry and OCLTT paradigms are both based upon respiratory metabolism, and the notion that the capacity to provide oxygen for metabolic demands is constrained at thermal extremes. The paper argues that competition among activities, and the vascular beds that support these, makes cardiac performance capacity a key mechanism underlying thermal constraints on aerobic performance, which requires more careful study. The objective is that, through better understanding of its underlying mechanisms, the OCLTT hypothesis can support predictive ecological applications for fishes in a rapidly changing world. Eliason & Farrell (2016) provide an authoritative review of the research into the anadromous migratory life cycle of Pacific salmon Oncorhynchus spp., in the Fraser River catchment. This constitutes what is perhaps the world's most emblematic example of research into the conservation physiology of fishes, which also happens to focus upon respiratory physiology and energy metabolism. Among the earliest studies of fish respiratory performance during aerobic exercise, and how this was affected by temperature, this research was initiated over 50 years ago on these species. The review discusses how traits of energy metabolism relate to a migratory lifestyle, in particular how the semelparous spawning strategy has put strong selection on adaptations for powerful and efficient upriver migration. Also, how homing to specific natal streams has led to strong local adaptation, of cardiorespiratory capacity and temperature tolerance of populations, in response to the specific challenges of their migration route. Ongoing climate change is raising serious conservation concerns and, despite the long history of study of these species, the review concludes by identifying critical knowledge gaps that remain. Enders & Boisclair (2016) take Atlantic salmon Salmo salar L 1758 parr as a case study to review evidence that variations in environmental conditions have significant effects on energy metabolism. They focus on water temperature, presence of shelter and variations in current speed. They argue therefore that the accuracy of models of fish metabolism will be improved if such environmental conditions are considered. This will increase the accuracy of energy budget calculations, which is valuable for conservation actions and for predicting effects of future climate change. The integration of physiological information into models is a major objective in conservation physiology. The review by Jørgensen et al. (2016) describes how models may put fish bioenergetics in a broader context, with the objective of making fish ecophysiological research more useful for science and society, in an era of climate change. Bioenergetics is the mechanistic foundation of many models. The paper argues that the predictive capacity of physiologically based models will be improved if they complement bioenergetics with consideration of important potential trade-offs that might influence behaviour and performance of the animal, particularly trade-offs linked to survival. The research article by Lea et al. (2016) investigates how water temperature affects the mechanical efficiency of swimming and aerobic performance of brown trout Salmo trutta L. 1758. The results indicate that S. trutta modulated the relationship between tailbeat frequency and amplitude to optimize swimming efficiency as they were acutely warmed from 11 to 24° C. Despite this, there was evidence that AS for activity declined at temperatures above 18° C, which may indicate energetic trade-offs when temperatures rise above thermal optima. Very little is known about the ability of air-breathing fishes to tolerate future climate changes. Lefevre et al. (2016b) studied the significance of air breathing for sensitivity to acute warming in a South-east Asian tropical species, the Asian swamp eel Monopterus albus (Zuiew 1793). The species relied on air breathing to meet the metabolic challenges of acute warming. Acclimation to a temperature above current seasonal averages improved upper temperature tolerance, but this was not linked to increased reliance on air breathing or improved cardiac The authors that in this species, air breathing will not provide particular advantages in warming The study to the evidence that oxygen is not the only limiting temperature tolerance in The by et al. (2016) that contamination by has effects on AS and swimming performance of is a of conservation physiology techniques of fish respiratory metabolism have been applied In particular, to investigate how might fish performance by maintenance costs or by reducing scope for activity. in is important to avoid in information effects of aquatic It is that this SI will to be a useful guide for researchers want to measure fish metabolic rate. The SI are to all authors for their support of this and to the at for their with with the They are also very to the for the in particular that more than one often so rapidly when these contributions were The Action Conservation Physiology of Marine the conferences and that led to this SI.