Cod spatial dynamics and vertical movements in European waters and implication in fisheries management

To compare the behavioural pattern of cod in the different ecosystems and raising this to population level the task of etho-typing includes extracting average habitat characteristics and activity pattern at scales comparable to otolith macrostructure analysis was undertaken. It was clear from our results that, although daily increments are visible in cod up to five years old, the otoliths of adult cod are unlikely to provide a reliable record of short-term (weekly) changes in behaviour because otolith properties could change dramatically at a time when neither temperature nor depth varied in any significant way. The cause of such changes was not clear. However, longer term changes in optical density and accretion rate do appear to provide a proxy for seasonal differences in behaviour that are associated with a long-term change in somatic growth or temperature experience. For example, the pattern of annual zone formation in cod from the southern and northern North Sea was clearly different and this is marked enough to be used a means of distinguishing between fish from each region. In cod from the Baltic Sea and the Faroe plateau, significant decreases in opacity were observed as cod changed their dominant behaviour from feeding to spawning, suggesting that otoliths could be used to provide reliable indications of spawning events and maturity ogives in otolith collections from these regions. The otolith accretion model developed in the IBACS project (QLRT-2001-01610) was applied to DST tagged individuals of the Baltic Sea and the North Sea to evaluate output of otolith macrostructure with actual observations of these individuals. Major biological drivers were identified that primarily influence otolith macrostructure between ecosystems. This modelling of otolith macrostructure using the environmental experience data from each ecosystem produced promising results.

Two individual-based models (IBMs), one representing the North Sea and the other representing the Baltic Sea, were developed to investigate the links between cod distribution and the environment. In the North Sea model, a detailed model of seabed temperature was used to analyse the influence of differing thermal optima on the predicted distribution of cod over a full seasonal cycle. In the Baltic Sea model, a detailed model of oxygen saturation was used to determine the movement rules used by cod to select habitat during the spawning season. The results of our IBMs have demonstrated how simple behavioural rules can result in complex distributions of individuals, and have provided insights into how much weight these rules have in the habitat selection decisions of individual cod. The models are not yet mature enough to provide comparable horizontal or vertical distributions to those observed in the wild, and so cannot yet be taken forward to predict the consequences of climate change or changes in the distribution of important prey species. However, the optimal strategies that they embody provide a valuable yardstick against which to measure real behaviours, and will help to refine behavioural hypotheses as the models themselves are refined and improved.

A statistical method was developed to determine the way that individual cod use their vertical environment. This method uses depth time-series to estimate empirical cumulative density functions of the difference between tag-measured and 24h maximum depths. The purpose of this analysis has been to develop an operational method to quantify the vertical distribution of cod, enabling to assess the probability of catch for an individual cod at a known location when gear opening and trawling depth are known. For acoustic surveys the method can aid to assess the probability that an individual cod is in the acoustic dead zone. The vertical range above maximum depth for which capture probability was 0.5 was used as a reference for system comparison, and ranged between 1.5 m in the Baltic and 84.9 m in the Barents Sea. This methodology allows in principal to estimate the capture probability for a cod when positioning of the gears known, both in the case of trawls or acoustic devices. Considering individuals that do not differ behaviourally, the aforementioned probabilities translate directly to the fractions of the population, which are accessible to a trawl or acoustic survey.

Atlantic cod is a highly adaptive boreal species with a broad distribution through different ecosystems. The behaviour of cod will always be constrained by physiology, be it temperature tolerance, swimming speeds or buoyancy control. However, diet, environmental factors and population dynamics vary enormously throughout its geographic range, and behaviours are therefore extremely plastic. The vertical movements observed in different ecosystems reflected this plasticity and so vertical movements varied significantly in frequency and scale as well as between individuals, season, years and ecosystems. Clear patterns were therefore difficult to detect from the comparative analyses: the data appear more chaotic than ordered. However it was clear that, in all ecosystems, cod vertical movements were influenced by the migratory strategy, tidal flow and level of illumination. Coastal or resident cod displayed a smaller range of behaviours than migratory cod. Cod in all ecosystems typically varied their behaviour seasonally, as they switched from reproductive behaviour to feeding behaviour. For example, cod in the North Sea appear to reduce their vertical movements in summer to feed on benthic prey mainly at night. In contrast, migratory cod in the Barents Sea and Iceland undertake large vertical migrations throughout the day in the vicinity of frontal systems, where pelagic preys are abundant.

Prior to the CODYSSEY project the only experimental work on otolith accretion had been conducted on larval and early juvenile stages of fish. These studies demonstrated that otolith structure responded rapidly to extrinsic changes such as temperature and food availability. The work undertaken in CODYSSEY is the first to link experimental and field investigations of otolith formation in adult fish, but our findings suggest that the otoliths of adult cod are unlikely to provide a reliable record of short - term (weekly) changes in behaviour unless the behaviour leads to a very large change in temperature. It is therefore not surprising that whilst there were some cases of a clear change in otolith properties that corresponded with a change in depth and temperature recorded by the DST it was more common not to find such a match. Our findings therefore showed that clear relationships between environmental and biotic factors and otolith microstructure were difficult to realise.

Environmental data from each ecosystem included in the CODYSSEY project were used to test initial geolocation methodologies and enable the development of an expert query system for reconstruction of cod migratory pathways. Individual migrations were reconstructed from the North Sea, the Barents Sea, the Icelandic plateau and the Baltic Sea using the new methodology with the environmental data most relevant to each ecosystem e.g. tidal and water temperature data in the North sea, water temperature and salinity in the Baltic Sea. However developing a standardised geolocation query system for all four ecosystems within the CODYSSEY project has proved to be a difficult, if not impossible, task, for several reasons. First, environmental databases for each ecosystem were not always available in the same formats, with the same resolution, and with the same precision. Second, each ecosystem is very different in terms of the dynamic range of data, and therefore the discriminatory power of a standardised geolocation query system would vary considerably. It is therefore the case that applying a standard method was neither appropriate nor necessary, and geolocation methods for each ecosystem were refined to optimise results in those ecosystems. Thus, we developed a number of innovative statistical techniques that were not only capable of estimating the position of cod with greater precision than before, but could also output estimates of certainty associated with these position. This is a step-change in the capabilities of geolocation of demersal fish, and the methods are also transferable to other species and locations.

Most cod stocks in the CODYSSEY project are assessed and managed using the virtual population assessment (VPA) methodology. This method can be modified to take account of stock structure and changes in catchability between seasons, but most often doesn't and instead uses broad and largely untested assumptions about the rates of population movements between management areas. Stock hypotheses and model assumptions have been reviewed in each ecosystem and tested against the observations from the CODYSSEY project. The data from the CODYSSEY project show that cod stocks are often partitioned into a migratory and resident component, and that the characteristics of the migratory component vary considerably between stocks (Iceland, Barents Sea, North Sea) and also between years (Baltic). Catchability of cod depends critically upon ecosystem, time of year and migratory strategy, and can therefore have consequences for the relative mortality of different stock components. It is therefore recommended that stock assessments take more account of population movements, and their consequences for determining abundance, sub-population identity and mixing.

The purpose of this work was to test whether structural features in the otolith can be used to identify and understand behaviours recorded by DST tags. The basic premise is that significant changes in the fish's behaviour are likely to involve changes in the extrinsic and intrinsic factors that influence otolith microstructure. To achieve our objectives, patterns of behaviour of DST-tagged cod were identified and quantified to determine the frequency and timing of key behavioural phases such as migration, spawning and residence on feeding grounds. In combination with the DST temperature record, the daily increments of recovered otoliths were examined to relate changes in growth conditions with changes in behaviour at scales from days to months. Experiments were conducted with adult cod to parameterise the effects of temperature, size and age for comparison with the long-term natural otolith signal from returned DST-tagged cod. A model predicting how signature behaviour influences otolith accretion to produce structures that are identifiable at the macrostructure level was applied to DST tagged individuals of the Baltic Sea and the North Sea to evaluate output of otolith macrostructure with actual observations of these individuals. Correlation of otolith microstructure with cod behaviour did not help to identify consistent patterns within ecosystems, although major biological drivers were identified that primarily influence otolith macrostructure between ecosystems.

The CODYSSEY web site was designed and published at http://www.codyssey.co.uk in October 2002 and has since been updated at 6 month intervals throughout the duration of the project. The dedicated Internet web site is intended as a facility for CODYSSEY team members, the general public, fisheries scientists and fishermen as user groups. The web site facilitates communication between project partners, informs on recent activities by project partners, provides information on tag rewards and updates on tag returns relevant to the fishing community. The detailed information obtainable from the site and the means by which the CODYSSEY project aims have been achieved are intended to be of interest to both the general public and fisheries scientists working on similar projects. A wide range of press and publicity material (including good quality illustrations introducing a non-specialist audience to the project's objectives and results) has been generated throughout the project period. This has been achieved by each partner institute submitting articles and advertisements to regional and national newspapers in their respective countries throughout the course of the project, with the aim of increasing awareness of CODYSSEY among the general public and, in particular, among the fishing communities. The specific purpose has been to maintain interest in the project and inform fishermen of the importance of the tagging programme and the rewards offered for tag returns. Details of these efforts can be found in the annual and final reports of the project. Further actions taken towards the effective communication of project results to the targeted audience involving dedicated workshops and seminars. Posters detailing the appearance of DSTs, what to do when a tagged fish is captured, and how to claim the reward were produced and distributed to Fisheries Offices, Harbour masters, Fisheries agencies and Fish Markets throughout the UK and Europe. Details of these efforts can be found in the annual and final reports of the project. A poster and leaflet to draw the public's attention to cod biology "Cod amazing even without chips" has now been designed, and will be disseminated to the public by each partner institute in printed and electronic form.

The results of the CODYSSEY field programme demonstrate that cod can occupy habitat over the temperature range -1.5°C to 19°C and, while the majority of temperature experience was centred between 6° to 10°C in all ecosystems, individuals can tolerate extended periods of time at the extreme temperatures and endure thermal shocks as great as 10°C in a short period of time as they cross the boundaries between different water masses. Although individual cod are known to adapt to changes in temperatures, they are thought to behave in a way that reduces thermal stress (behavioural thermoregulation, Claireaux et al 1995) and to be unable to tolerate large, abrupt changes in temperature. As a result, it is believed that thermoclines of only a few degrees can constitute fairly impenetrable boundaries and that this aspect of their biology may contribute to the segregation of cod populations in nature (Claireaux et al 1995). Our observations represent a significant development in our understanding of cod biology because they suggest that thermal tolerance of wild cod is considerably greater than that suggested by laboratory studies. Simulation experiments have explored the extent to which the behaviour of cod can be explained by simple behavioural rules. In the Baltic Sea, cod appear to use a combination of behavioural rules to ensure that they remain within preferred habitat. Simple depth or oxygen preference rules do not accurately recreate their observed behaviours. In the North Sea, simulations of cod response to water temperature suggested that simple hypotheses that relate cod movements to optimal temperature alone create highly unrealistic seasonal distributions. Simulating cod movements within a tolerance range provide more realistic results, and support the results of the electronic tagging experiments that described a tolerance of higher temperatures than expected. Together, these results suggest that the minimisation of physiological stress and the maximisation of growth rate are not be the only factors to consider as the basis of, or inputs to, predictive models. Overall the assessment of thermal and anoxic tolerance of cod, and the analysis of depths occupied, has demonstrated that cod are tolerant to a wider range of environmental conditions than previously thought. These results help to explain the widespread distribution of cod in the Atlantic, and also provide evidence that, while human-induced environmental change will inevitably shift the distribution of cod, the capacity of stocks and individuals to tolerate thermal stress offers hope for stock recovery if fishing mortality can be controlled, and if recruitment of young cod occurs at historical levels.

It became obvious very early on in the project that the cod swim bladder imposed major constraints on vertical migration both by limiting vertical movement distances and on the energy expenditure budget of the cod. It was therefore necessary to develop a complete dynamic bio-energetic model of the cod swim bladder. Further developmental work has taken place on a vertical movement model for cod in the Barents Sea and for parameterising this model for other ecosystems. The model is a detailed bioenergetic model aimed at investigating the effect of different factors on the vertical positioning strategies in adult cod. The model has been completed successfully for cod in the Barents Sea, and is being parameterised for cod in the North Sea and the Baltic Sea.

Atlantic cod is a highly adaptive boreal species with a broad distribution through different ecosystems. The behaviour of cod will always be constrained by physiology, be it temperature tolerance, swimming speeds or buoyancy control. However, diet, environmental factors and population dynamics vary enormously throughout its geographic range, and behaviours are therefore extremely plastic. The vertical movements observed in different ecosystems reflected this plasticity and so vertical movements varied significantly in frequency and scale as well as between individuals, season, years and ecosystems. Clear patterns were therefore difficult to detect from the comparative analyses: the data appear more chaotic than ordered. However it was clear that, in all ecosystems, cod vertical movements were influenced by the migratory strategy, tidal flow and level of illumination. Coastal or resident cod displayed a smaller range of behaviours than migratory cod. Cod in all ecosystems typically varied their behaviour seasonally, as they switched from reproductive behaviour to feeding behaviour. For example, cod in the North Sea appear to reduce their vertical movements in summer to feed on benthic prey mainly at night. In contrast, migratory cod in the Barents Sea and Iceland undertake large vertical migrations throughout the day in the vicinity of frontal systems, where pelagic preys are abundant. Many publications have been generated from analysis of results within each ecosystem. Several manuscripts of the comparison of cod spatial and vertical dynamics between different ecosystems are now anticipated.

The purpose of this area of work has been to test whether structural features in the otolith can be used to identify and understand behaviours recorded by DST tags. The basic premise is that significant changes in the fish's behaviour are likely to involve changes in the extrinsic and intrinsic factors that influence otolith microstructure. Patterns of behaviour of DST-tagged cod were identified and quantified to determine the frequency and timing of key behavioural phases such as migration, spawning and residence on feeding grounds. In combination with the DST temperature record, the daily increments of recovered otoliths were examined to relate changes in growth conditions with changes in behaviour at scales from days to months. Experiments were conducted with adult cod to parameterise the effects of temperature; size and age for comparison with the long-term natural otolith signal from returned DST tagged cod. The otolith accretion model developed in the IBACS project (QLRT-2001-01610) has been applied to DST tagged individuals of the Baltic Sea and the North Sea to evaluate output of otolith macrostructure with actual observations of these individuals. However, correlation of otolith microstructure with cod behaviour did not help to identify consistent patterns within ecosystems, although major biological drivers were identified that primarily influence otolith macrostructure between ecosystems. Otolith representative of the cod populations in the different ecosystems are being analysed to identify indicators of habitat characteristics, migration, activity and seasonal growth cycles in order to establish within and among ecosystem variations. Results will have management implications e.g. bearings for identification of stock structure, large scale seasonal migrations as well as trends in environmental impact on population behavioural characteristics.

The results of the CODYSSEY field programme demonstrate that cod can occupy habitat over the temperature range -1.5°C to 19°C and, while the majority of temperature experience was centred between 6° to 10°C in all ecosystems, individuals can tolerate extended periods of time at the extreme temperatures and endure thermal shocks as great as 10°C in a short period of time as they cross the boundaries between different water masses. Although individual cod are known to adapt to changes in temperatures, they are thought to behave in a way that reduces thermal stress (behavioural thermoregulation, Claireaux et al 1995) and to be unable to tolerate large, abrupt changes in temperature. As a result, it is believed that thermoclines of only a few degrees can constitute fairly impenetrable boundaries and that this aspect of their biology may contribute to the segregation of cod populations in nature (Claireaux et al 1995). Our observations represent a significant development in our understanding of cod biology because they suggest that thermal tolerance of wild cod is considerably greater than that suggested by laboratory studies. Simulation experiments have explored the extent to which the behaviour of cod can be explained by simple behavioural rules. In the Baltic Sea, cod appear to use a combination of behavioural rules to ensure that they remain within preferred habitat. Simple depth or oxygen preference rules do not accurately recreate their observed behaviours. In the North Sea, simulations of cod response to water temperature suggested that simple hypotheses that relate cod movements to optimal temperature alone create highly unrealistic seasonal distributions. Simulating cod movements within a tolerance range provide more realistic results, and support the results of the electronic tagging experiments that described a tolerance of higher temperatures than expected. Together, these results suggest that the minimisation of physiological stress and the maximisation of growth rate are not be the only factors to consider as the basis of, or inputs to, predictive models. Overall the assessment of thermal and anoxic tolerance of cod, and the analysis of depths occupied, has demonstrated that cod are tolerant to a wider range of environmental conditions than previously thought. These results help to explain the widespread distribution of cod in the Atlantic, and also provide evidence that, while human-induced environmental change will inevitably shift the distribution of cod, the capacity of stocks and individuals to tolerate thermal stress offers hope for stock recovery if fishing mortality can be controlled, and if recruitment of young cod occurs at historical levels.

Two individual based models (IBMs) have been developed. The first is an IBM of cod vertical migration strategies, which incorporates a bioenergetic swim-bladder model. The model finds optimal tilt angles and swimming speed given the buoyancy of the fish, and implements the effect of behavioural drivers using a genetic algorithm. Simulations and sensitivity analysis with different environmental forcing have been performed and the predicted vertical migration patterns have been compared to various observations and data obtained from DTS tagged cod in the Barents Sea. Work is ongoing to parameterise the model for other ecosystems to provide a comparative analysis. A second predator-prey IBM was developed to investigate the effects of prey density and relative value on the proportion of time that cod spend in the pelagic vs benthic water layers. The description of the IBM in general and the first major findings has now been submitted for publication in Bulletin of Mathematical Biology.

The CODYSSEY project is a comparative evaluation of the behavioural characteristics of cod in the NE Atlantic to improve understanding of the horizontal migrations and vertical movements of cod in relation to environmental and biological factors in order to provide fisheries managers with information on the horizontal availability, vertical accessibility and individual vulnerability of cod to fishing activities. The environmental data from the CODYSSEY tagging programme has demonstrated that, in the wild, cod can tolerate a wide range of environmental conditions that are outside of the range considered tolerable in laboratory experiments. These data make a tremendous contribution to our understanding of the results and analyses from fishing surveys, which sometimes make simplifying assumptions about the way that cod are expected to select habitat. Thus, while environmental temperature undoubtedly has a powerful controlling influence on metabolic demand in fish, our studies suggest that the role of temperature in determining movements and distribution of adult cod in the open sea may not be critical and other environmental factors (e.g. prey distribution, suitable flow refuges etc.) may be equally, or even more, important. Similarly, in trying to understanding the potential impact of climate change at the population level, the effect of temperature on adult cod may not be as important as understanding the effect of temperature on other biological mechanisms such as the link between temperature and survival during the early life history stages.

The development of geolocation methods for all ecosystems has enabled the comparison of cod migrations and distribution. Two different movement strategies were observed: migration and residency. Migratory cod travelled different distances to different areas dependent upon their release position, and their stock identity. For example, cod tagged in the southern North Sea in spring would often move northwards to summer feeding grounds. Cod tagged in the English Channel in spring would often move west. Cod in both areas would move back towards the southern North Sea in the months prior to spawning. i.e. migrations were not random, but instead directed towards feeding or spawning grounds. Resident cod were caught close to their site of release and did not visit feeding or spawning sites at a significant distance from their point of release, even after long periods at liberty. Reconstructions of their likely horizontal movements suggested that they remained within their release locality for the period of their liberty. Examples of resident cod were observed in each CODYSSEY ecosystem except the Baltic. The proportion of resident to migratory cod varied between 30% in the Skagerrak and Kattegat, to 100% in the Faroe plateau. Migratory cod undertook directed migrations at specific times of year that were consistent with congregation on spawning grounds and with dispersal to feeding grounds. In the northern ecosystems (Iceland plateau, Barents Sea), post-spawning migrations to offshore feeding grounds occurred in June and July, and fish began their migration back to spawning grounds as early as February, ready to spawn in April and May. In the southern ecosystems, spawning takes place between January and March, so cod migrate to feeding areas between April and June, and start their pre-spawning migrations as early as November. In the Baltic Sea, conditions for spawning are often very limited geographically or temporally due to the unique hydrographic conditions in the deep, low salinity basins of the Bornholm Basin or the Gdansk Deep. Migrations into deeper water and below the halocline in spring therefore indicated the onset of spawning migrations. Cod tagged in the CODYSSEY project undertook directed migrations towards these basins at very similar times each year in each of three years.

Deliverables

Share this page

Download