EU FP6 project RECLAIM (Resolving CLimAtic IMpacts on fish stocks) studied the impact of climate change on the productivity and distribution of fish and shellfish populations. RECLAIM reviewed the strengths and weaknesses of current scientific knowledge with the aim to detect critical knowledge-gaps and to provide sound advice, recommendations and working hypotheses for future research. The study focussed on the north-east Atlantic and comprised of a structured literature review and specific case studies. Nine institutes, mainly form northwest European states collaborated closely in the project (Table 1). To provide a basis for the mechanistic understanding of how climate change may impact fish and shellfish populations. a framework for the study of climate on fish populations was developed that was based on first principles of physiology and ecology. The framework allowed us to structure the literature review on ocean climate, marine ecosystems and fish populations, select the environmental variables and oceanographic features that are relevant for fish and likely to be affected by climate change, and derive working hypotheses on the mechanisms determining the response of fish to climate change. In a second step, patterns of change were explored by analysing biological time-series in relation to climatic variables. Specific mechanisms for change were examined using biophysical models, to increase our understanding of the processes involved. Finally, the implications of future climate on ecosystem processes and the dynamics of fish populations were explored. Achievements, working hypotheses and recommendations for future research are summarised below.
Table 1 List of contractors and subcontractors
IMARES | Institute for Marine Resources and Ecosystem Studies | Netherlands |
FRS | Fisheries Research Services | UK |
Cefas | Centre for Environment, Fisheries & Aquaculture Science | UK |
IFREmer | Institut Français de Recherche pour l’Exploitation de la mer | France |
UniH | Institute of Hydrobiology and Fishery Science, University of Hamburg | Germany |
DTU-Aqua | National Institute of Aquatic Resources, Technical University of Denmark | Denmark |
IMR | Institute of Marine Research | Norway |
NIOZ | Royal Netherlands Institute for Sea Research | Netherlands |
UiB-GFI | University of Bergen , Geophysical Institute | Norway |
1 DMI | Danish Meteorological Institute | Denmark |
1 NERI | National Environmental Research Institute | Denmark |
1 Subciontractors to DTU-Aqua.
Ocean climate
Our expectations of how increased concentrations of green house gasses will affect our climate rely the outputs of Global Circulation Models (GCMs). GCMs consistently predict an increase in both air and sea temperature (larger changes in northern areas), changes in precipitation (increase in northern Europe and a decrease in southern Europe) and a decrease in the sea ice. With higher salinity in the tropics due to higher evaporation, some of the water will be transported northward and will lead to an increased salinity in several sea areas in the northeast Atlantic, including the Barents Sea. In the Baltic, salinities will decrease because of increased precipitation. The Meridional Overturning Circulation (MOC) is expected to weaken by approximately 25% by 2100 but the waters in the north-eastern Atlantic are expected to continue to warm in spite of reduced MOC. The predictions about changes in wind fields and upwelling processes are more variable and in some cases models offer conflicting and diametrically opposite predictions for the future. The increase in CO2 concentrations in the atmosphere will result in increased CO2 concentrations in the ocean and a reduction in the pH. RECLAIM did not specifically address the problem of ocean acidification. Nevertheless, increased acidification is expected to have significant effects on marine ecological processes and may affect fish in unexpected (and indirect) ways such as neurological changes impacting odour recognition leading to problems in natal homing and, hence changes in connectivity of life stages to essential habitats, and changes in the fertilisation of eggs and the development of early fish larvae.
There is a large body of scientific evidence showing that changes in fish populations (recruitment variation, growth, distribution) are correlated to variations in ocean climate, such as the North Atlantic Oscillation (NAO) and the Atlantic Multidecadal Oscillation (AMO). A problem in the study of climate change impacts on marine ecosystems is that the Global Circulation Models, which are the key tools to model the changes in meteorology and oceanography, are not capable of hindcast the observed patterns in NAO and AMO. Finer resolution GCMs should be explored. A concerted research effort towards improving the GCMs, including resolving these two major modes, is required.
Further research is required to develop downscaling strategies for dynamic regional shelf sea models. Such models are critical for improving and providing quantitative estimates of the impacts of climate change on the regional biology. This research needs to include careful assessment of the underlying GCMs used for downscaling to regional models in terms of the GCM’s ability to reproduce past and present climate. Also, downscaling should be carried out using several GCMS to avoid model specific problems. Until this is done, the downscaling will not provide realistic future projections for regional impact studies. Downscaling strategies need to include both atmospheric forcing and open ocean boundary condition issues.
Research is also needed to address uncertainty in regional projections arising from both the regional physical and ecosystem models, as well as global projections. Also, simple ways to express these uncertainties are needed for those using the results of the models.
Marine ecosystem
Marine ecosystems are considered to be primarily regulated by bottom-up processes (i.e. climate, influencing phytoplankton productivity which in-turn influences ‘higher’ trophic levels). Only cold water ecosystems characterised by relatively few species are thought to be regulated primarily by top-down processes (top predators and fishing). Top-down regulated ecosystems will be more sensitive to fishing effects. A first exploration of the relative roles of fishing and changes in primary production in a number of ecosystems including the North Sea, Irish Sea and western Mediterranean suggested that fishing was found to be the primary forcing factor in the North Sea, however primary production was found to be more important in the Irish Sea.
There is ample evidence that climate change will affect the primary and secondary productivity of ecosystem, although the impacts will differ across geographic regions. Primary production will likely increase in the Barents Sea (where sea ice will disappear) and the Norwegian Sea due to increased temperature. In more southerly regions, such as the Bay of Biscay, the increased stratification caused by warming will likely reduce overall primary production. There will also be changes in plankton distribution and community structure. For example with the decreased salinities in the Baltic Sea, there will be an increase in neritic species and they will extend their distribution farther seaward. In other European waters, many species are expected to expand their distribution northward with the warming, although the rate of movement will not necessarily be the same for all species and hence this could potentially change community structure as well as ecosystem function.
Figure 1 Changes in primary production (gCm-2yr-1) in response to an 3°C increase in temperature (upper left), a 30% increase in wind (upper right) and a 20% increase or decrease in radiation (Sh.R. bottom left and right) as predicted by the Ecosmo bio-physical model (Drinkwater et al., 2009. Deliverable 4.2).
Marine primary production is strongly affected by solar radiation, upwelling as well as by the development and breakdown of stratification. A number of coupled bio-physical models have been used to explore the possible impact of climate driven changes in temperature, short wave radiation and winds (Figure 1). Despite numerous differences in model setup and parameterisation, surprisingly coherent results were produced by the different models. Model projections indicated that: (i) changes in phytoplankton production is most strongly affected by changes in solar radiation, with smaller effects caused by changes in wind speed and the weakest effects caused by changes in air temperature; (ii) the combined effect of an increase in solar radiation, wind speed and air temperature showed the largest potential for increased primary productivity; (iii) the sensitivity of lower trophic level production to changes in river nutrient loads is restricted to the coastal waters (particularly in the southern North Sea and at the mouths of major rivers) with little impacts in the larger North Sea, while in the Bay of Biscay, such changes produce a negative response similar to that due to changes in air temperature; (iv) secondary production and zooplankton biomass appear to be more sensitive to changes in climate forcing compared to primary production. Application of a 1-D model in the Kattegat suggested that (v) primary production increase due to higher remineralisation rates although the spring bloom will be lower due to higher grazing pressure and increased pelagic heterotrophy; (vi) sedimentation of organic matter will likely decrease resulting in a reduction of benthic productivity. Benthic filter feeders will be less affected while deposit feeders and sub-surface (within the sediment) feeders are predicted to suffer the largest declines in biomass.
Long-term monitoring programs of various ecosystem components are an essential basis for the study of climate change impacts on marine ecosystems. There is a clear gap in monitoring data and process understanding with regard to the benthos component.
Figure 2 Change in abundance and distribution of John Dory (Zeus faber), a Lusitanian species (Pinnegar et al., 2009. Deliverable 1).
Fish populations: distribution and productivity
Climate change affects a multitude of environmental factors, that may affect various processes at different levels of biological organisation (organism, population, ecosystem). Fish production is first of all dependent on the production of algae in the ocean. On the other hand, the response of commercially-exploited fish species cannot solely be due to changes in their food resources, but may also be affected by direct effects on their physiology, or changes in ecosystem interactions (competitors and predators). Since fishing heavily impacts marine ecosystems through removal of individuals and physical abrasion of benthic communities, fisheries and climate will interact and both drivers need to be taken into account when attempting to resolve the impact of climate change on fish populations.
There is ample evidence for changes in distribution of fish populations in relation to climate. Figure 2 provides an example of the changes in distribution and abundance of a southern species, Zeus faber. Species that have historically had a low- to mid-latitude distribution tend to increase at the high latitudinal range of their distribution as temperatures warm, while species with a higher latitudinal distribution exhibit declines at their lowest latitudes. Also, changes in the bathymetric distribution have been reported with species shifting their distribution to deeper cooler waters. The underlying mechanisms of the distribution shifts, however, are less well known. RECLAIM has revealed that throughout the 20th Century, the centre of fish distribution, as reported by commercial fishermen, changed appreciably in the northeast Atlantic, for example in the North Sea cod have moved to the northeast, plaice have moved northwestwards, sole have moved south towards the Channel (where conditions are now habitable throughout the winter), and haddock occupied the same area on average but their southern limit of distribution moved 130 km to the North over the past 80-90 years.
In analyzing climate impacts, RECLAIM has taken into account the fact that fish have complex life cycles that include different stages (egg, larval, juvenile and adult) that may rely on different, spatially-separated habitats. Moreover, individuals undergo massive developmental and physiological changes as they progress through these stages (e.g., individuals often increasing in body size by a factor of 105 from egg to adults). RECLAIM utilised a variety of methods to reveal critical life stages (those most sensitive to climate impacts) including analyses of the abundance across successive life history stages. There are very few species where there are sufficient data with which one can examine ‘critical life history stages’. It is clear that different factors are affecting each species e.g. in cod cannibalism on age 0 fish, or larval starvation can be significant and density-dependent factors can affect juvenile flatfish on an essentially two dimensional nursery ground. A closer examination of North Sea herring highlighted the necessity to examine temporal changes in mortality schedules in relation to ‘periods’ i.e. there could be shifts in the underlying ‘productivity’ of the system that need to be recognized to be able to interpret changes in mortality.
Knowledge on the eco-physiology of the various life history stages provide a solid foundation to evaluate the likely impact of climate change on fish more generally, and may provide insight into the relative sensitivity of different life stages (Figure 3). There is some evidence that life stages differ in their sensitivity to environmental factors such as temperature. In terms of identifying critical periods, there was no common pattern among species with respect to changes in thermal tolerance among eggs, larvae, juveniles and adults. However, prey availability, particularly during the late larval and early juvenile period, appears to be one of the most important environmental factors creating density-dependent fluctuations in populations in both the northwest Pacific and Northeast Atlantic.
Figure 3 Diagram of the changes in suitable habitats (based on water temperature) with (a) latitude of the species and/or population and (b) by life stage. The arrow (a) denotes a range of tolerable temperatures measured for adults during maturation and spawning (from Rijnsdorp et al., 2009. ICES JMS 66: 1570-1583. Deliverable 1).
There is a lack of basic information on temperature tolerance and sensitivity for many species not to mention information on the impacts of “multiple stressors” (e.g., temperature x salinity, x dissolved O2 for Baltic Sea organisms). The limited data indicate that the ranges in preferred temperatures and tolerable temperatures were positively related with optimal temperatures for growth suggesting that species with a large tolerance range had lower temperature sensitivity. No clear differences between pelagic and demersal species were observed.
Finally, changes in the adaptive capacity of fish (and other organisms) to changes in temperature and other abiotic factors is largely unknown. Physiological adaptation is possible and could mitigate (offset) the impacts associated with environmental change. Climate change may evoke concomitant evolutionary change as has been demonstrated in a variety of taxa. The situation is further complicated by a haemoglobin polymorphism in species such as cod that is expressed differently across the geographical distribution of the species. Studies examining the adaptive capacity of physiology of fish species and populations would clearly benefit ongoing efforts to project climate-driven changes in distribution and productivity of fish populations by using techniques such as bio-envelope climate modelling.
It is uncontested that simulation models that couple the physics of the ocean and the biology and ecological interactions among the species are an indispensable tool in the study of climate change. State-of-the art in bio-physical models typically include the nutrients, phytoplankton, zooplankton and detritus dynamics (NPZD-models). Within RECLAIM, these NPZD-models were used to explore the impact on fish larvae (growth, mortality, transport) and model results were compared with field data of fish larvae and plankton. Results suggests that climate driven bottom-up processes can have multiple, interacting impacts on the survival of marine fish early life stages including: (1) species-specific changes in the match-mismatch dynamics of first-feeding marine fish larvae and their prey (zooplankton), and (2) changes in transport patterns of developing larval cohorts. Overall, the results indicate that continued warming will make vast areas of the North Sea unsuitable for species such as cod (winter-spring spawners) but will have little impact on species such as sprat (spring summer spawners). Also simulations suggested that the survival of Baltic sprat larvae depend upon climate-forced changes in both temperature and prey populations. A study of larval herring in the North Sea, showed that the intensity and spreading of the autumn zooplankton bloom might be an important driver influencing maximum larval length and consequently their overwinter survival. NPZD modelling within RECLAIM also showed that response to increasing temperatures was stronger for zooplankton than for phytoplankton causing higher grazing impact on phytoplankton, faster recycling of nutrients and less sedimentation of organic matter to the sea floor. Climate warming is therefore expected to be more detrimental for benthic animals than for pelagic species. In another study requested by OSPAR, the interaction of increased temperatures and nutrient reductions as (50%-70% for N and P, respectively) were analysed on zooplankton and larval survival of different marine fish species. Model results suggest that changes in zooplankton production and, more importantly, species composition caused by changes in nutrient loading could have marked impact on trophic coupling between zooplankton and fish early life stages in the future.
Bio-physical (NPZD) models are generally restricted to the early life history stages of fish, and do not include juvenile and adult stages allowing an examination of how climate change could impact life cycle closure. Many biophysical lower trophic level models do not include benthic components of the ecosystem, making estimates of changes in energy flow (nutrient cycling) impossible. Within the benthic community, organisms with different feeding ecology will experience a varying degree of food limitation in a warmer climate. NPZD modelling showed that benthic suspension feeders that have access to suspended particulate organic matter in the bottom water were less influences by the reduced sedimentation whereas deposit feeders living in the sediment were severely reduced. Hence, the longer the time - lag, the larger the reduction in the food supply due to respiratory loss and food competition with organisms closer to the source of the organic matter in the ecosystem.
For a population to survive there must be life history closure, i.e., the whole life cycle must be able to be completed. Current coupled bio-physiological models are not equipped to deal with life-cycle closure in fish as they only address egg and larval life phases. In a first attempt to extend the model to include later life history stages, a Dynamic Energy Budget (DEB) model of adult anchovy was combined with a NPZD-model to estimate spatial habitat suitability based upon energy available for spawning in the Bay of Biscay and the North Sea. The maps agreed well with the long-term patterns of anchovy presence, critical seasons and core habitats. A climate scenario suggested that the index of fronts did not change but the bottom temperature increased dramatically by approximately 2.5°C on the shelf in comparison to the reference. As a result, the mean predicted anchovy distribution was guided by the overall increase in temperature and showed anchovy dispersed everywhere with maximum concentrations along the coast (Figure 4). In particular, the shelf north of 46°30´N that was empty in the past is potentially opened to the anchovy under climate change.
Figure 4 Modelled distribution of anchovy in the Bay of Biscay in 2050 under climate change (right hand panel) in comparison with the current distribution (left hand panel) (Petitgas et al., 2009. Deliverable 4.3).
The implications of climate change for fisheries management was explored in Baltic cod (Figure 5). Climate change is expected to result in a warmer and less saline Baltic. The maximum sustainable fishing mortality (Fmsy) is shown decreasing with both salinity and increasing temperature. This can be explained by a direct salinity effect on cod recruitment (i.e., through egg and larvae survival) and an indirect temperature effect channeled through species interactions; due to increased recruitment of sprat and a competition driven decline in herring eventually affecting the dominance and availability of prey for cod. Hence, the degree to which species interactions may either buffer or accentuate the cod stock response to climate change depends on the nature of both positive and negative feedback loops within the food-web.
RECLAIM has provided considerable insight into the likely implications of climate change for marine fish and ecosystems, but it has thus-far provided only limited comment on the likely consequences of such changes for commercial fisheries and regional economies. We anticipate that climate change will have the following implications fisheries, and we urge that further research is needed in this area: (i) Modelling of how fishers will respond to changes in fish distribution, including impacts on distance that fishers will need to travel to maintain catches (additional fuel costs, days at sea etc.), the potential impact on stocks that span national boundaries and consequences for quota allocation, the potential impact on marine-protected-area effectiveness if protected species move outside the boundaries of the closed area; (ii) The potential impact of climate change on fish ‘catchability’, e.g. whether fish behave differently in warmer waters (whether they respond differently to an incoming trawl), whether the geometry of fishing gears will change as species and fisheries move into deeper waters, and the potential impact of ocean acidification on noise-transmission in the ocean, which could impact on the ability of fish to detect incoming trawls; (iii) The socio-economic implications of changed fisheries yields associated with climate change, yields are predicted to increase in the north but decrease in the south of Europe) and possible mitigation methods for fisheries and fishery policy makers.
Figure 5 A sustainable management strategy for Baltic cod under different scenarios of climate change. In (A) and (B) the theoretical carrying capacity (Kt) and corresponding maximum sustainable fishing mortality (Fmsy) for Baltic cod is shown for each combination of projected changes in salinity and sea surface temperature (SST). In order to address the indirect effects of species interactions on cod stock response to climate change, the lower panels (C, D) include fishing pressures for sprat and herring, given a projected increase in SST by 3.5°C. Fishing mortalities range from mean historical levels (Fmean) to the recommended precautionary levels (Fpa) for sprat and herring respectively. The middle planes represents mean K and Fmsy while upper and lower planes respectively illustrate the upper and lower confidence levels of each scenario (Drinkwater et al. 2009; Deliverable 4.4).
Working hypothesis
RECLAIM has formulated six (sets of) working hypothesis based on a priori mechanistic arguments. These working hypothesis will be useful to analyse the processes that underly the response of fish and shellfish populations to climate change and will improve our capabilities to predict the implications of future climate change.
(1) The response of fish populations to climate change will differ between species as well as between stocks across the geographical distribution area. This leads to the first set of working hypotheses:
H1a. Populations at the limits of their latitudinal range will show stronger responses than those occurring within habitats in the centre of their latitudinal distribution;
H1b. Northerly species at the southern limits of their distribution will decrease in abundance and southerly species will increase at their northern limits;
H1c. Species distributions will shift to deeper, cooler waters in response to an increase in water temperature.
H1d. Climate change will result in an increase in the biodiversity (species richness).
(2) We expect that the response of species to climate change will be influenced by their habitat requirements (pelagic, demersal, deepwater species), life history characteristics (short-or long-lived; specialist or generalist), and trophic position within the ecosystem (apex predators or forage fish). This leads to the following hypotheses:
H2a. Species with habitat requirements for fixed geographic locations will differ in their distributional responses from species that are linked to open water masses.
H2b. Deep-water fish species will be less impacted by climate change in comparison with shelf or coastal species;
H2c. Fish species with narrow dietary preferences will be more sensitive to climate change, compared with generalists;
H2d. Short-lived species will show stronger responses and will be better equipped to adapt to changes in their environment than long-lived species.
(3) Populations can survive in systems where suitable habitats for the different life history stages are available and are connected, allowing life cycle closure. If the habitat for a certain life stage is spatially restricted, a change in habitat suitability of this stage will make the species more sensitive to climate change than species which do not have spatially restricted habitat requirements. This leads to the following hypotheses:
H3a. Species with spatially restricted habitat requirements during part of their life history will be more sensitive to climate change than species without specific habitat requirements.
H3b. Fish populations in oceanographic systems with a high variety of mesoscale features will show less influence of climate change.
(4) Fishing will reduce the size- and age-structure of a population and reduce its bet-hedging capabilities that would allow it to successfully contend with variability in suitable conditions for the survival of eggs and larvae. Also, fishing may lead to a reduction in genetic variability that would negatively impact the possibilities of an evolutionary response to climate change and the ability of depleted stocks to recover.
H4. Fish stocks under intense exploitation will be more vulnerable to climate change than those experiencing low fishing pressure.
(5) The ecosystem response to climate change will depend on the response of the individual species and the resulting effect on trophodynamic interactions among species.
H5a. Ecosystems with simple trophic structure will show more rapid responses to climate change than ecosystems with more complex trophic structure;
H5b. Changes in ecosystem structure caused by climate change will be non-linear and abrupt.
H5c. Under climate change we expect the pelagic productivity to go up and the demersal productivity to go down.
H5d. Increased species richness will result in an increased importance of bottom-up processes.
(6) Improved environmental conditions, as well as new shipping routes, will facilitate the spread of warm-water fish species and pathogens.
H6. With improved local conditions, an increased number of exotic warm-water fish species (and fish pathogens) will become established in European waters.
Recommendations
- Global Circulation Models should be improved to capture the decadal (NAO) and multi-decadal (AMO) scale variations in ocean climate.
- Regional downscaling models need to be developed based on different GCM to provide a realistic future projection for regional impact studies.
- Research is needed to address uncertainty in regional projections arising from uncertainty in the regional physics and ecosystem models, as well as GCM.
- Bio-physical modelling of single species should extend beyond the egg and larval stages in order to better project climate-driven changes on marine fish populations stemming from processes acting in various life history stages (from eggs to adults) and on the life cycle closure.
- Lower trophic level ecosystem models need to be improved including more emphasis on pelagic-benthic coupling of marine systems.
- Additional research on the growth physiology of key species and life stages, as well as on the physiological effects of acidification and multiple stressors is needed.
- The evolutionary adaptive capacity of marine fish species and populations to climate-driven change in environmental factors is poorly understood and needs to be examined.
- Monitoring research of the major biotic components (plankton, benthos, fish) should be continued and expanded. Increasing temporal and spatial coverage of data sets and inclusion less well covered ecosystem components (benthos) is required. Such time series are invaluable for analyses of climate impacts and for the formulation and validation of ecosystem models.
- Models of the upper trophic web need to be better linked to biogeochemical and NPZD models in order to predict the consequences of climate change for ‘higher’ consumers in ecosystems.
- Models which attempt to predict predator-prey overlap in the future, are needed to establish how changed communities might interact and develop.
- Better understanding and modelling of implications for fishing fleets and local economies are required, since it can be very difficult to determine how fisheries might look in the future, and what conditions they will need to adapt to.
